Junggarsuchus Sloani, an Early Late Jurassic Crocodylomorph with Crocodyliform Affinities

Home , Dibothrosuchus, Dromicosuchus, Edentosuchus, Hallopus, Mesoeucrocodylia, Pissarrachampsa

Junggarsuchus sloani, an Early Late Jurassic Crocodylomorph with Crocodyliform Affinities

by Alexander Altieri Ruebenstahl

B.S. in Biology, May 2018, The George Washington University B.S. in Geology, May 2018, The George Washington University

A Thesis submitted to

The Faculty of The Columbian College of Arts and Sciences of The George Washington University in partial fulfillment of the requirements for the degree of Master of Science

August 31st, 2019

Thesis directed by

James M Clark Professor novitas of Biology

Acknowledgments

I would like to give my sincere thanks and appreciation to James Clark, Catherine

Forster, Andrew Moore, Joseph Stiegler for their help in this project and my education. I would also like to thank the George Washington University, the Institute of Vertebrate

Paleontology and Anthropology, the University of Berkeley at California, the Carnegie

Museum of Natural History and the American Museum of Natural History for access to specimens. Thanks to the Hennig Society for providing TNT. The skull of

Junggarsuchus was prepared by Wang Hai-jun. The skulls of Dibothrosuchus and

Junggarsuchus were scanned by Yi Hongyu. This work was supported by NSF grant

EAR 1636753.

iii

Abstract of Thesis

Junggarsuchus sloani, an Early Late Jurassic Crocodylomorph with Crocodyliform Affinities

The holotype of Junggarsuchus sloani, from the early Late Jurassic of Xinjiang,

China, consists of a nearly complete skull and the anterior half of an articulated skeleton, including the pectoral girdles, nearly complete forelimbs, vertebral column, and ribs. This taxon shares many features with a cursorial group of crocodylomorphs, known as

‘sphenosuchians’ whose relationships are poorly understood. However, it also displays several derived crocodyliform features that are not found among members of this group.

A phylogenetic analysis corroborates the hypothesis that Junggarsuchus is closer to

Crocodyliformes than Dibothrosuchus and Sphenosuchus, but not as close to crocodyliforms as Almadasuchus and Macelognathus, which includes extant crocodylians, and that the “Sphenosuchia” are a paraphyletic assemblage. Two other species of “sphenosuchians,” Dibothrosuchus elaphros and Sphenosuchus acutus, are hypothesized to be more closely related to Crocodyliformes while the rest of the

‘Sphenosuchia’ form several smaller groups and are largely unresolved. We find that

Dibothrosuchus is not as closely related to crocodyliforms as Junggarsuchus, but also possesses several unique autapomorphies. We also report an elongate blade-like coronoid in Junggarsuchus that is also widely present in ‘sphenosuchians’.

Table of Contents

Acknowledgements……………………………………………………………………...iii

Abstract of Thesis……………………………………………………………………….iv

List of Figures……………………………………………………………………………vi

List of Tables……………………………………………………………………………vii

Institutional Abbreviations…………………………………………………………....viii

Chapter 1: Introduction……………………………………………………………………1

Chapter 2: Methods and Materials……………………………………………………….10

Chapter 3: Systematic Paleontology…………………………………………………...... 16

Chapter 4: Description of Junggarsuchus and comparison with Dibothrosuchus………21

Chapter 5: Phylogenetic Results……………………………………………………...... 105

Chapter 6: Discussion…………………………………………………………………..117

Chapter 7: Conclusion……………………………………………………………….....125

Literature Cited:………………………………………………………………………..127

Appendices……………………………………………………………………………..142

List of Figures

Figure 1: Material of Junggarsuchus sloani holotype IVPP 14010……………………218

Figure 2: Generalized tree of crocodylomorph relationships…………………………..219

Figure 3: Life reconstruction of Junggarsuchus sloani………………………………...220

Figure 4: Skulls of Junggarsuchus and Dibothrosuchus in left lateral view…………...221

Figure 5: Skulls of Junggarsuchus and Dibothrosuchus in dorsal view……………….223

Figure 6: Skulls of Junggarsuchus and Dibothrosuchus in ventral view……………...224

Figure 7: Skulls of Junggarsuchus and Dibothrosuchus in occipital view……………..225

Figure 8: Braincase of Junggarsuchus …………………………………………………227

Figure 9: Ear region of Junggarsuchus…………………………………………………229

Figure 10: Braincase of Dibothrosuchus……………………………………………….231

Figure 11: Mandible of Junggarsuchus……………………………………………..….232

Figure 12: Mandible of Dibothrosuchus………………………………………………..234

Figure 13: Shoulder bones of Junggarsuchus and Dibothrosuchus…………………....235

Figure 14: Left forelimb of Junggarsuchus and Dibothrosuchus………………………236

Figure 15: Left wrist bones of Junggarsuchus and Dibothrosuchus…………….……..237

Figure 16: Left manus of Junggarsuchus and Dibothrosuchus………...………………238

Figure 17: Axis and atlas of Junggarsuchus………………………………………..….239

Figure 18: Cervical vertebrae of Junggarsuchus and Dibothrosuchus ………………...240

Figure 19: Dorsal vertebrae and ribs of Junggarsuchus………………………………..241

Figure 20: Phylogenetic tree with Tennant’s characters and equal weights……………242

Figure 21: Phylogenetic tree with Tennant’s characters and implied weights…………243

Figure 22: Phylogenetic tree without Tennant’s characters and with implied weights...244

List of Tables

Table 1: Table for CI, RI, and Steps from Phylogenetic analyses without Tennant’s characters……………………………………………………………………………….245

Table 2: Table for CI, RI and steps from Phylogenetic analyses with Tennant’s characters……………………………………………………………………………….245

Table 3: Relationships and node support for groups in analyses without Tennant’s characters……………………………………………………………………………….246

Table 4: Relationships and node support for groups in analyses with Tennant’s characters……………………………………………………………………………….248

Table 5: Table of Synapomorphies for group found……………………………………249

vii

Institutional Abbreviations

AMNH: American Museum of Natural History (Fossil Reptiles), New York, United States

BP: Evolutionary Studies Institute (formerly Bernard Price Institute for Palaeontological Research), University of the Witwatersrand, Johannesburg, South Africa

CM: Carnegie Museum of Natural History, Pittsburg, PA United States

CUP: Fujen Catholic University of Peking (Beijing) collection in Field Museum of

Natural History

IGM: Institute of Geology, Mongolian Academy of Sciences

IVPP: Institute of Vertebrate Paleontology and Paleoanthropology, Chinese Academy of Sciences, Beijing, China

UCMP: University of California Museum of Paleontology, Berkeley, United Stat

viii

INTRODUCTION

The Shishugou Formation of Xinjiang, China is a continuous series of sediments spanning the poorly known late Middle to early Late Jurassic (Eberth et al., 2001; Clark et al., 2006). The lower part of the formation has yielded a variety of turtles, brachyopoid amphibians, a mammaliaform, and theropod and sauropod dinosaurs while the upper contains a more diverse fauna of dinosaurs and non-dinosaurian amniotes. An expedition in 2001 recovered the skull and part of the postcranial skeleton of an unknown crocodylomorph, briefly described and named Junggarsuchus sloani by Clark et al.

(2004), from the lower part of the Shishugou, now dated to the early Late Jurassic

(Choiniere et al, 2010; see below). This taxon shares many similarities—such as adaptations toward cursoriality and an elongate, rod-like coracoid—with basal crocodylomorph archosaurs known as ‘sphenosuchians.’ However, it also shares features—such as modifications to the braincase that solidify the skull—with

Crocodyliformes, the group that includes modern crocodylians. Furthermore,

Junggarsuchus possesses autapomorphies not seen in any other crocodylomorph, such as a surangular foramen and a massively dorso-ventrally expanded and pneumatic basisphenoid, which is farther expanded than that in Sphenosuchus (Walker, 1990).

The ‘Sphenosuchia’ (Bonaparte 1971, 1982) are archosaurs known from the Late

Triassic to the Late Jurassic (Clark et al. 2000, Göhlich et al. 2005; Leardi et al, 2017) that fall within the Crocodylomorpha (Hay 1930 emend Walker 1970) but outside the

Crocodyliformes. The Crocodyliformes (includes modern crocodylian species and their fossil relatives that possess specializations that solidify the skull); ‘sphenosuchians’ are

1 the most basal crocodylomorphs. At least 13 valid monotypic genera are considered potential sphenosuchians: Sphenosuchus (Haughton 1915; Walker 1990), Saltoposuchus

(Heune 1921; Sereno and Wild, 1992), Hallopus (Marsh 1877; Walker 1970),

Terrestrisuchus (Crush 1984), Dibothrosuchus (Simmons 1965; Wu and Chatterjee

1993), Hesperosuchus (Colbert 1952), Pseudhesperosuchus (Bonaparte 1971),

Litargosuchus (Clark and Sues 2002; Pedeticosaurus of Gow and Kitching 1988),

Kayentasuchus (Clark and Sues 2002), Dromicosuchus (Sues et al. 2003), Macelognathus

(Marsh 1884; Göhlich et al. 2005), Almadasuchus Pol et al. (2013) and Junggarsuchus

Clark et al. (2004). Phyllodontosuchus (Harris et al., 2000), Trialestes (Nesbitt et al.,

2013); and Redondavenator (Nesbitt et al., 2005), are ‘sphenosuchians’ but known from incomplete or poorly preserved material and their affinities are not well understood.

Terrestrisuchus has been considered a junior synonym of Saltoposuchus (e.g., Benton and Clark, 1988), or as distinct taxa (e.g., Sereno and Wild, 1992); Allen (2003) considered Terrestrisuchus material to be juvenile individuals of Saltoposuchus. Nesbitt

(2011) tentatively recognized Terrestrisuchus, but considered the specimen assigned to

Hesperosuchus by Clark et al. (2000), Carnegie Museum 29894, to potentially belong to a different taxon due to its younger age within the Chinle Formation and the lack of autapomorphies shared by this specimen and the holotype of Hesperosuchus agilis.

Many of the features shared among ‘sphenosuchians’ are related to an upright posture and terrestrial lifestyle, unlike modern semi-aquatic crocodylians (Walker 1970,

Crush 1984, Parrish 1989, Sereno and Wild 1992). However, there are few putative synapomorphies; thus, it is unclear whether or not these taxa comprise a monophyletic group. Various analyses have shown the group either to be monophyletic (Sereno and

Wild 1992; Wu and Chatterjee 1993; Clark et al. 2000; Sues et al. 2003) or paraphyletic, with some taxa being more closely related to Crocodyliformes (Benton and Clark 1988;

Parrish 1991; Clark and Sues 2002; Clark et al. 2004; Pol et al., 2013; Leardi et al. 2017).

Because of this lack of agreement, the relationships and hence the membership of the

‘Sphenosuchia’ have been difficult to establish. Early studies were not consistent with the use of their characters, however, and a critical review by Clark et al. (2000) revealed numerous problematic characters in earlier publications diminishing support for their results. A subsequent analysis (Clark et al. 2004), including new characters and

Junggarsuchus sloani, found in favor of a paraphyletic ‘Sphenosuchia’ but with weak support, and in an analysis without Junggarsuchus Nesbitt (2011) also found a paraphyletic Sphenosuchia, and more resolution among them. Similar results have been reproduced by Pol et al. (2013) and Leardi et al. (2017).

Here, the holotype specimen of Junggarsuchus sloani, IVPP 14010 (Figure 1), is described in detail in comparison with other ‘sphenosuchians’ including detailed comparison to Dibothrosuchus. The characters used in previous analyses are again critically reviewed and reanalyzed, and the results support a paraphyletic Sphenosuchia.

The clade Archosauria is represented by two extant lineages, crocodylians and birds. Within the crocodylian lineage, one of the larger inclusive clades, which includes both extant crocodylians and extinct, stem crocodylians, is referred to as

Crocodylomorpha. Crocodylomorpha is defined as the most inclusive clade containing

Crocodylus niloticus Laurenti, 1768, but not Rauisuchus tiradentes Huene, 1942,

Poposaurus gracilis Mehl, 1915, Gracilisuchus stipanicicorum Romer, 1972c,

Prestosuchus chiniquensis Huene, 1942, or Aetosaurus ferratus Fraas, 1877 (Sereno,

2005) (Nesbitt, 2011). (Figure 2)

Extinct, crocodylomorphs outside the crocodylian crown group have been known for over two centuries, from several countries (Chapman, 1759; 1826; Marsh, 1881).

Despite the strong representation of early crocodylomorphs in the fossil record the evolution of many key features of the group remained unclear until nearly three decades ago, when primitive crocodylomorphs were described in more detail, older specimens were re-described and as more members of the transition have been found (Walker, 1990;

Wu and Chatterjee, 1993; Clark, et al, 2000, 2004). Basal members of Crocodyliformes, a group defined as crocodylomorphs that acquired many of the distinctive features of living crocodylians (Figure 1), have been known since 1904, from fairly complete if poorly preserved skeletons (Broom, 1904). These early crocodyliforms are sometimes placed in a group referred to as Protosuchia (Mook, 1934). Basal members of Crocodylomorpha known as sphenosuchians (Figure 2, 3) have been known for over a century (Marsh,

1881), but their anatomy and relationships were not well understood until the work of

Walker (1970, 1990). These sphenosuchians have played a major role in shaping our understanding of crocodylian evolution as the members of ‘Sphenosuchia’ make up the closest outgroups to Crocodyliformes (Walker, 1990; Wu and Chatterjee, 1993, Clark et al, 2004; Sues et al, 2003).

The importance of the taxa known as ‘sphenosuchians’ is twofold. First, as some of the sphenosuchians are widely considered the closest outgroups to Crocodyliformes

(Leardi et al, 2017), they help to polarize characters used in analyses of crocodylian evolution. Alternatively, some analyses have found the marine crocodylomorph group

Thalattosuchia as the sister group to Crocodyliformes (Figure 2), but even in these analyses ‘Sphenosuchia’ is found as a paraphyletic grade near Crocodyliformes (Figure

2,) (Wilberg, 2015). The differences in anatomy between sphenosuchians and

Crocodyliformes detail what the primitive condition for various features of

Crocodyliformes are as well as giving us a general idea about the lifestyle of early crocodylomorphs (Clark et al, 2004; Wu and Chatterjee, 1993).

The second point of importance is a subject of some debate. During the early study of the group, ‘Sphenosuchia’ was found to be a monophyletic clade, sister to

Crocodyliformes in some analyses (Wu and Chatterjee, 1993) but other researchers found that ‘Sphenosuchia’ was a paraphyletic, grade group, with some forms closer to

Crocodyliformes than others. In this case, sphenosuchians are a transitional series of forms leading into Crocodyliformes (e.g., Benton and Clark 1988). More recent studies of the group, with additional taxa and more extensive character and taxon sampling have supported this finding, of sphenosuchians as a paraphyletic grade (Leardi et al 2017;

Wilberg et al, 2015; Pol et al, 2013). With sphenosuchians as a transitional grade, the morphology of individuals along that grade can be studied, from forms closer to

Crocodyliformes to those more basal, and so the evolution of crocodyliform features can be elucidated as a step by step process. In this regard sphenosuchians have provided important information on the early evolution of the crocodylian braincase, showing the stepwise strengthening of the braincase, palate and skull. The tight suturing of previously mobile elements in the skull is one of the essential features that allows extant crocodylians to have such a powerful bite (Pol et al, 2013; Leardi et al, 2017).

When ‘Sphenosuchia’ is recovered as a transitional grade, the sphenosuchians

Junggarsuchus sloani, Macelognathus vagans and Almadasuchus figarii are found to be the closest relatives of Crocodyliformes and species such as Sphenosuchus acutus,

Dibothrosuchus elaphros, Terrestrisuchus gracilis and Litargosuchus leptorhynchus are usually found to be more basal members of Crocodylomorpha (Leardi et al, 2017; Benton and Clark, 1988; Pol et al, 2013; Wilberg et al, 2015). However, Kayentasuchus walkeri, a sphenosuchian from the Early Jurassic of Arizona (Clark and Sues, 2002), has been found in some analyses (e.g., Nesbitt, 2011) to be the sister taxon to Crocodyliformes, although Nesbitt did not include Junggarsuchus, Almadasuchus, or Macelognathus. This placement for Kayentasuchus as the sister taxon to Crocodyliformes was also found by later authors including Wilberg (2015), who included Almadasuchus and Junggarsuchus, and Zanno et al. (2016), who included Junggarsuchus though not Almadasuchus or

Macelognathus. Though Wilberg’s sampling included more sphenosuchians and crocodylomorph outgroup taxa, he discusses that this position is not well supported, and two-character state changes would place Junggarsuchus and Almadasuchus as sister to

Crocodyliformes+Thalattosuchia. The posterior section of the skull of Kayentasuchus is not well preserved, and as many key elements in the evolution of the crocodylomorph skull are located in the braincase, Kayentasuchus is missing critical information that would be essential in better resolving its relationships (Wilberg, 2015). An analysis of

Nesbitt’s characters found that one of the characters that supported his placement of

Kayentasuchus, the posterior process of the maxilla (his character 2), was problematic.

Leardi et al. found that the character states were poorly defined, and that taxa with similar morphologies of the posterior process of the maxilla, were scored for different character

6 states. This character was omitted, and Leardi et al. (2017) found Kayentasuchus in a more basal position.

Of the over a dozen species that are referred to as sphenosuchians several are known from exceptionally preserved fossils that preserve the majority of the postcranial skeleton and nearly complete skulls. Two Chinese ‘sphenosuchians’, Dibothrosuchus elaphros and Junggarsuchus sloani, are known from excellent material, including near complete skulls and at least half of the postcranial skeleton. Dibothrosuchus elaphros is a sphenosuchian from the Zhangjiawa member of the Lower Lufeng Formation in

Yunnan, China, which has been biostratigraphically dated as Early Jurassic, potentially

Sinemurian (Luo and Wu, 1994). Several specimens are known, though the most complete is IVPP V 7907, comprising a complete skull, the anterior portion of the axial column, the forelimbs and includes some elements of the hind limbs and pubis (Wu and

Chatterjee, 1993). Junggarsuchus sloani (Figure 1,3) is known from the Shishugou

Formation of Xinjiang China, from strata below a tuff dated as very early Late Jurassic

(Choiniere, 2010) It was originally considered to be late Middle Jurassic (Clark et al.,

2004) based on the Gradstein (2004) time scale, but later revisions (Gradstein, 2012) moved this boundary older and indicate Junggarsuchus is earliest Late Jurassic (see

Horizon and Locality below). The holotype specimen, IVPP 14010, consists of an exceptionally preserved skull and the anterior portion of the body, with a few disarticulated elements of the posterior portion of the skeleton associated with the holotype (Figure 1). Junggarsuchus sloani has several features that seem to reduce the morphological gap between more basal crocodylomorphs and crocodyliforms, including

7 the ventral contact of the quadrate to the braincase that is a key step in the beginning of the solidification of the skull.

The best specimen of Dibothrosuchus elaphros (IVPP V 7907) was described by

Wu and Chatterjee (1993), who described in detail the skull and postcranial anatomy. The holotype and only significant specimen of Junggarsuchus sloani was described briefly in a paper by Clark et al in 2004, which did not include a full description of the skeletal anatomy and focused specifically on the features of the braincase and forelimbs. In addition to this description, a Masters student working with Clark analyzed and prepared a detailed description of much of the known material of Junggarsuchus sloani (Klein,

2007) but it was not published. This description did not include detailed discussion of the braincase internal anatomy, as CT scans were not available and the skull had not been fully prepared. The two taxa share several features, such as fused parietals, straight exoccipitals, a frontal not underlain by the prefrontal, a supratemporal fenestra with a horizontal shelf and that the mastoid antrum enters the prootic. However, several features were found in Junggarsuchus sloani that are not found in more basal crocodylomorphs but are found extensively in crocodyliforms. The exoccipitals meet on the midline and a ventrolateral extension of the exoccipital contacts the quadrate, the jugal arches dorsally, the quadrate is fenestrated and the occipital portion of the parietal is narrow. Several of these features are involved in the solidification of the skull and reduction of cranial kinesis which is found in more basal crocodylomorphs. Other important features of

Junggarsuchus, such as short transverse processes of the dorsal vertebrae, its digitigrade stance and a lack of dorsal osteoderms, have been interpreted as features related to active, terrestrial lifestyle (Wu and Chatterjee, 1993; Clark et al, 2004).

Several analyses recover a similar pattern of relationships within ‘Sphenosuchia’, in which Dibothrosuchus elaphros and Junggarsuchus sloani are found to be closer to

Crocodyliformes than are most other sphenosuchians (Leardi et al, 2017; Benton and

Clark, 1988; Pol et al, 2013; Wilberg et al 2015), though not as close as Macelognathus,

Almadasuchus and potentially Kayentasuchus (Wilberg, 2015). Though Dibothrosuchus and Junggarsuchus are two of the best represented members of this lineage of crocodylomorphs, there is limited comparison of these two, relatively closely related taxa.

Here we describe all elements of Junggarsuchus sloani IVPP 14010 (Figure 1, 3), in comparison to Dibothrosuchus elaphros IVPP V7907. This includes a redescription of the cranial and postcranial material of Dibothrosuchus elaphros, IVPP V 7907, and specifically the internal skull anatomy which we were able to analyze in CT scans. The description of Junggarsuchus sloani, IVPP 14010, is a revised version of the earlier

Master’s thesis description by Klein, with an in depth description of the internal anatomy of the braincase and the sutures of the bones in the skull, which earlier authors were unable to do due to the lack of CT data (Leardi et al, 2017; Clark et al, 2004; Wu and

Chatterjee, 1993; Benton and Clark, 1988). We ran a Parsimony analysis with TNT

(Goloboff and Catalano 2016) on 47 crocodylomorphs, focusing largely on sphenosuchians to test the relative placement of the two taxa, and clarify the relationships of sphenosuchians within the paraphyletic grade. Of particular interest to our analysis is the placement of Junggarsuchus relative to more advanced sphenosuchians like

Almadasuchus and relative to Dibothrosuchus, which is often found as a crocodylomorph basal to Junggarsuchus (Leardi et al., 2017; Wilberg, 2015). In regards to

Dibothrosuchus we also describe several unique features in Dibothrosuchus than are not found in other sphenosuchians. This description will compare the two taxa, determining important traits that the two taxa share, and major transitional steps between

Dibothrosuchus and Junggarsuchus, essentially the step between the more basal sphenosuchians and the taxa closest to Crocodyliformes, with a focus on traits that are believed to be important in the evolution of crocodylians.

METHODS AND MATERIALS

The limb bones of Junggarsuchus sloani are described as if the animal was standing erect. Hence, anterior = cranial and posterior = caudal. The terms ‘ventral’ and

‘dorsal’ are used in describing the digits but not the more proximal elements, assuming a digitigrades stance. The skull was separated from the rest of the skeleton and was transported to The George Washington University, Washington, DC for study and further prepared after the publication of Clark et al. (2004), and nearly all of the matrix was removed. Only the left side of the postcranial skeleton has been prepared out of its plaster jacket, and the right side of the specimen has not been viewed except for the right forelimb (Figure 1 a). The holotype skull was first CT scanned before extensive internal preparation on 9/28/2004 on a GE Lightspeed 16 CT scanner at Stony Brook

University. The specimen was scanned at 140 kV and 160 mA with a slice spacing of

0.31 mm. Slices were reconstructed at a diameter of 96.0 mm using the GE BonePlus algorithm. The holotype skull was rescanned after nearly all of the matrix was removed in a Mi-CT 225kV micro-computerized tomography (developed by the Institute of High

Energy Physics, Chinese Academy of Sciences) CT at the Key Laboratory of Vertebrate

Evolution and Human Origins, Institute of Vertebrate Paleontology and

Paleoanthropology, Chinese Academy of Sciences. Slices were spaced at .0459 millimeters for a total of 3402 slices along a 143.32 millimeter long skull. . We used

Mimics software(https://www.materialsise.com/en/medical/software/mimics) to segment and analyze CT scans of the skull of Junggarsuchus sloani. Mimics takes the micrometer spaced slices of a CT scan and displays cross sections of the skull in a format in which colored “masks” can be designated to individual bones. These masks are determined by the identification of the extent of bones in a dorso-ventral view, left lateral to right lateral view and an anterior-posterior view of each CT slice. Once the extent of each bone is determined, colored masks are applied to each bone. These masks can then be isolated as a 3D image, which can then be manipulated in 3D space. The detailed anatomical work that Mimics allows for the scoring of new characters. All non-photograph illustrations of

Junggarsuchus in this paper are detailed 3D models of the masks reconstructed in

Mimics.

Dibothrosuchus (IVPP V7907) was also CT scanned at the IVPP in Beijing using the same scanner. The rostrum and jaw were segmented in a single file, and the braincase was segmented separately, as the braincase was disarticulated from the rest of the skull.

The non-photo illustrations of the skull of Dibothrosuchus used in our description were all also reconstructed from MIMICs files.

Characters: The character set of 466 characters is a combination of data sets from Wilberg (2015), Leardi et al. ( 2017 ), and Young and Andrade (2009) Following this initial set of analyses 120 of the 329 characters from Tennant et al.’s 2016 data set were incorporated into this matrix for a total of 586 characters. The selected characters

11 greatly increase the relative number of postcranial characters, including more characters for the axial skeleton and osteoderms. These characters were used for an analysis focused on neosuchians. For each data set we looked at overlap between the characters and omitted repetitive, semantically dependent characters between the data sets.

Taxa: Coding for most taxa used characters taken from existing matrices, including Wilberg (2015), Leardi et al (2017) and Young and Andrade (2009). Young and Andrade (2009) outlined the synapomorphies of Thalattosuchia. Junggarsuchus sloani (IVPP V14010), and Dibothrosuchus elaphros (IVPP V7907) were studied in person and in CT segmentation, while Nominosuchus matutinus (IVPP V14392),

Protosuchus haughtoni (BP-14746), ‘Gomphosuchus wellsi’ (UCMP 97638/125871),

‘Edentosuchus’ (UCMP125358), Zaarasuchus shepardi (IGM 100/1321), Shamosuchus djadochtaensis (IGM 100/1195), Crocodylus niloticus, and Alligator mississippiensis were all studied in person, ).

Our sampling of sphenosuchians and more basal crocodylomorphs includes 14 species. Two crocodylomorphs of uncertain relationship, Trialestes and

Phyllodontosuchus were also included. Our sampling of outgroup taxa outside of crocodylomorphs includes Carnufex, Dyoplax, Erpetosuchus two additional gracilisuchids, Yonghesuchus and Turfanosuchus and Postosuchus. Our ingroup sampling was more limited, with 10 species of basal crocodyliforms, 4 basal mesoeucrocodylians,

5 thalattosuchians, 2 notosuchians, 3 tethysuchians, 1 extinct paralligatorid and 2 extant taxa: Alligator mississippiensis and Crocodylus niloticus. This selection of taxa was adapted from the Leardi et al, 2017 data set. For the complete list of taxa and their scorings see (Supplementary Data- Appendix 2).

Rooting:-- This study uses two alternative rooting schemes (Tables 1-4). The first roots our phylogeny on Gracilisuchus stipanicicorum, a presumed relative of early crocodylomorphs. With this scheme, two other gracilisuchids were included,

Yonghesuchus sangbiensis and Turfanosuchus dabenensis, both the from Middle Triassic of China (Wu et al, 2001, Wu and Russel, 2001). However, gracilisuchids may be more distantly related to crocodylomorphs than assumed in several early analyses (Nesbitt,

2011), so in second set of analyses the three gracilisuchids were excluded and theroot was the rauisuchian Postosuchus kirkpatricki. If gracilisuchids are much more basal than the paraphyletic series of 7 taxa of ‘rauisuchians’ found by Nesbitt 2011, they may not provide accurate character states for the outgroup of Crocodylomorpha (Wilberg, 2015).

Furthermore, the gracilisuchids may be convergent in their limb structure with the similarly gracile sphenosuchians. In the analyses where gracilisuchids were omitted, both

Dyoplax and Erpetosuchus were also omitted. These two taxa are members of the family

Erpetosuchidae, which has several proposed positions outside Postosuchus +

Crocodylomorpha (Nesbitt and Butler, 2012), including some positions well outside the node that unites even Gracilisuchidae with other stem crocodylomorphs (Nesbitt, 2011).

Excluded Taxa:- Of the eight analyses run, 4 have two taxa omitted from them a priori. In earlier rounds of our analysis Trialestes romeri and Phyllodontosuchus lufengensis were identified as potentially problematic ‘rogue’ taxa. The two taxa jumped around in our early trees, being found both outside Crocodylomorpha and among

‘Sphenosuchia’. This is believed to be due to the incomplete nature of Trialestes

(Lecuona et al, 2016) and the very poor preservation of Phyllodontosuchus. The latter has been weathered in such a way that any suture of the skull are nearly totally obliterated

13 and it is very difficult to score any characters confidently (Harris et al 2000). Two sets of analyses were run, 4 with these taxa included and 4 without.

Parsimony Analysis: For this project eight phylogenetic trees were constructed with parsimony analysis using 464 characters applied to 47 taxa with Gracilisuchus as the outgroup and 40 with Postosuchus as the outgroup. TNT v1.5 (Goloboff and

Catalano, 2016) was used to find the most parsimonious trees and the strict consensus. 44 multistate characters were treated as ordered. Unlike Wilberg’s analysis, additional analyses were not run where these 44 characters were assumed to be unordered, as ordering of characters is justified by the similarities among the states (Lipscomb, 1992).

Eight different analysis were run. Four of our phylogenetic analyses were rooted on Gracilisuchus. Of these four analyses two of the tests used equal weighting. For these two equal weight analyses, one was performed with all taxa included, and the other performed with the omission of Phyllodontosuchus and Trialestes. Two analyses with implied weights (k=12) were also performed, one for each of the taxon distributions discussed above. The other 4 analyses were rooted on Postosuchus, and the three gracilisuchids and two erpetosuchids were omitted. Two of these tests included

Phyllodontosuchus and Trialestes and the other two did not. For each of the aforementioned groups, both a weighted (k=12) and unweighted analysis was performed.

For our Equal Weights analysis minimal tree lengths were first found using New

Technologies searches. We set the search to look for the minimum tree length 5 times and set the initial addition sequences at 50. This search was carried out with drift, tree fusing and sectorial search set at the default settings. Ratchet was also included in our new technologies search for all analyses with 100 total number of iterations. For our implied

14 weight analysis, tree fusing, sectorial search, drift and ratchet were maintained, but instead of finding minimum tree length, we looked for the stabilized consensus two times, with a factor of 75, the default. The weighting function for the implied weight was set at k=12. For both equal weight and implied weight analyses, to ensure that all minimum tree lengths were discovered, all analyses were subjected to traditional search with TBR (tree bisection reconnection) branch swapping. Following the use of the TBR algorithm, a strict consensus was found for the set of trees retained from the analysis.

For trees with equal weights the Consistency index and retention index were calculated for each set of most parsimonious trees, using the maximum, minimum and best step numbers. The Consistency Index was found by dividing the minimum step over the observed tests for each consensus tree. The retention index was calculated by subtracting the observed steps from the number of maximum steps. This number was then divided by the number of maximum steps minus the number of minimum steps. (Table 1)

The same methodology was then later applied to a new data matrix that included

588 characters applied to 47 taxa. 134 additional characters were added from the Tennant et al. (2016) data set. The same rooting schemes, variations in weights and omission of rogue taxa were applied for another 8 analyses. In the end we performed 16 analyses total.

Synapomorphies:-- Synapomorphies were first mapped along trees using TNT.

The synapomorphies were then checked against a tree built in Mesquite from our data set to visualize the evolutionary history of the character and its various states.

Synapomorphies for the various groups are shown in Table 5.

Node Support:-- Support for the nodes was found using symmetric resampling

(Goloboff et al, 2003), with a .33 change probability, which is the default. The results for the topologies were output as both absolute frequencies and frequency differences.

Frequency differences tend to give slightly lower numbers, but are considered more accurate as they compare the frequency of a given group versus the frequency of the next most likely group to be found. This tests the assumed group against possible contradictory groups (Goloboff et al, 2003). This resampling was run with 100 replicates and was set to collapse any node with a support number lower than 1. Trees were searched with a New Technology search which used sectorial searches, ratchet, tree fusing and drift and inserted an additional 10 sequences as the starting point for each analysis prior to a new technology search. The minimum length was calculated only once.

SYSTEMATIC PALEONTOLOGY

CROCODYLOMORPHA Hay 1930 (emend Walker, 1970)

New Taxon to be Named in Formal Publication: In the publication of this work we will name a new taxon that we will informally term here “Solidocrania”.

Definition: All taxa more closely related to Junggarsuchus and Crocodylus than to

Dibothrosuchus.

Diagnosis: exoccipitals contact below supraoccipital, which is reversed in Almadasuchus

(20); greatly expanded basisphenoid with pneumatic cavities (29); the exoccipital contacting the quadrate ventrolaterally (30); reduction in the size of the antorbital fenestra

(48); a developed anterior process of the ectopterygoid projecting along the surface of the jugal (58); quadrate, squamosal and otoccipital enclose the cranioquadrate canal laterally

(62); the presence of additional quadrate fenestra has been inferred as a synapomorphy of this group, and the loss of the additional fenestra in Almadasuchus and Macelognathus may be secondary losses (78); the primary head of the quadrate approaches the laterosphenoid (109); a pneumatized pterygoid (313); the anterior process of the pterygoid ramus of the quadrate is firmly sutured to the pterygoid (322); dorsal edge of surangular arched dorsally (376); anterior edge of the scapular blade is larger than the posterior (122); the radius is longer than the humerus (138); the olecranon process of the ulna is very low (144); the glenoid surface of the coracoid is extended on a vertical plane

(442), like Protosuchus.

Ambiguous characters that may support group: squamosal contacts the posterodorsal surface of the quadrate enclosing the otic recess posteriorly (61); supratemporal fossa anterior margin is posterior to the postorbital (211); two large palpebrals (254); the quadrate ramus of the pterygoid is broad in ventral view (316); the supraoccipital may contribute to the medial border of the posterior temporal fenestra (59).

JUNGGARSUCHUS Clark et al., 2004

Type Species: Junggarsuchus sloani Clark et al., 2004, by original designation

JUNGGARSUCHUS SLOANI Clark et al., 2004

Holotype: IVPP14010, a nearly complete skull and mandible and the anterior part of the postcranial skeleton.

Horizon and Locality: Upper part of lower Shishugou Formation, Wucaiwan, Altay

Prefecture, Xinjiang, China. A tuff approximately 30 meters stratigraphically above this specimen has been dated at 162.2 + /-0.2 million years (Choiniere et al., 2014), which places it younger than the 163.5+4 mya estimated age of the Middle-Late Jurassic boundary (albeit with a larger error; Gradstein et al., 2012). With an estimated sedimentation rate of ~4.6 cm/ka (Eberth et al., 2006) the fossil is estimated to be about

652,000 years older than the dated tuff, placing it at approximately 162.85 mya, still slightly younger than the boundary estimate.

Revised Diagnosis: Jugal below infratemporal fenestra is strongly arched dorsally (31); the ventral edge of the jugal possess a longitudinal concavity (32); presence of a surangular foramen (37); M.pterygoideus ventralis insertion on angular extends well onto the angular- seen in later crocodyliforms (38); two quadrate fenestrae (78), ; pit between premaxilla and maxilla for lower caniniforms not exposed laterally (85); presence of prefrontal overhang (247); premaxilla’s ventral edge dorsal to ventral edge of maxilla

(488); basisphenoid body greatly expanded and pneumatized more so than

Sphenosuchus (29); quadrate ramus of pterygoid extends to posterior end of quadrate; laterally facing basispterygoid process is situated posteriorly on the basisphenoid unlike the condition in Sphenosuchus where the opening is more complex (26) ; postorbital contacts frontal anteromedially (211); first metacarpal slender (126); radiale more elongate than other sphenosuchains (128); first manus digit faces laterally (125); two distal carpal bones forming straight joint at wrist (124); well-developed hypapophyses present on cervical and anterior dorsal vertebrae (119); procoelous vertebral centra (118);

18 dorsal , cervical and caudal osteoderms absent (121, 477, 488). It is possible that additional fenestrations in the palate (598) are present.

Dibothrosuchus Simmons, 1965

Type Species: Dibothrosuchus elaphros Simmons, 1965, by original designation.

Comments: IVPP V7907 was originally described as a second species of D. elaphros, D. xingsuensis Wu 1986, but it was synonymized with D. elaphros by Wu and Chatterjee

(1993) and there is currently only the type species recognized as valid in this genus.

Dibothrosuchus elaphros Simmons, 1965

Holotype: CUP 2081, a partial skull and skeleton.

Referred Specimens: IVPP V7907, a nearly complete skull and mandible and partial postcranial skeleton; Wu and Chatterjee (1993) referred three other, incomplete specimens (CUP 2106, 2084 and 2489) to this species.

Horizon and Localities: The holotype and referred specimens were collected near Dawa village, about 10 Km northeast of Lufeng, Yunnan (Wu and Chatterjee, 1993). They are from the Zhangjiawa Formation in the Lower Lufeng Group (Luo and Wu, 1994).

Revised Diagnosis: Of the original characters diagnosed by Wu & Chatterjee (1993), the following remain correct: frontals with three parasagittal ridges converging at 2 ends; frontal-postorbital contact forming a crescentic ridge; and a transversely broad supratemporal fenestra; pronounced ovular depression on anterior surface of the humerus

(may be present in Junggarsuchus but smaller). Wu and Chatterjee identified potential autapomorphies as uncertain due to the unknown conditions in other sphenosuchians at

19 the time and we find support for the following: squamosal curves sharply medially anterior to the supratemporal fenestra; squamosal separated from quadratojugal by quadrate; elongate antorbital fenestra; ventral process of the postorbital covers the posteromedial surface of the jugal and a small triangular mandibular fenestra.

The full sheathing of the basioccipital condyle by the otoccipital is not supported as an autapomorphy due to our uncertain reconstruction of that region in Dibothrosuchus and

Junggarsuchus. The condition of the anterior temporal fenestra is seen in Junggarsuchus and so rejected as an autapomorphy. The autapomorphies of the coracoid are also reported in other sphenosuchians (Clark, 2004). The trigeminal recess is also not supported as an autapomorphy and may be widely present in sphenosuchians.

In our analysis we also identified several other potential autapomorphies in

Dibothrosuchus. While Wu and Chatterjee identified the trigeminal recess as an autapomorphy of Dibothrosuchus, it was later identified in other ‘sphenosuchians’ such as Almadasuchus. We identify additional potential autapomorphies for Dibothrosuchus including a massively enlarged and pneumatic prootic and potentially ventrally closed facial antrum; the basipterygoids are massively ventrally expanded and pneumatic, though Sphenosuchus demonstrates a similar, if less extreme condition (26); the descending process of the prefrontal contacts the palatine, unlike other sphenosuchians

(6); the articular possesses a dorsomedial projection (27); the lateral border of the orbit is medial to the lateral border of the suborbital fenestra (234); the lateral temporal fenestra is over 50% the size of the orbit (484).

DESCRIPTION OF JUNGGARSUCHUS AND COMPARISON WITH

DIBOTHROSUCHUS

Nearly all of the matrix has been removed from the skull and the bone has been glued where it had separated along several large cracks. The largest of these is between the braincase and the rest of the skull, where the dorsal part of the braincase is now rotated 5mm to the left and the ventral part was rotated anteriorly. The right posterolateral part of the skull and mandible were eroded before discovery, and missing or too fragmentary to identify are the quadratojugal, squamosal, postorbital, all but the anterior tip of the jugal, the paroccipital process lateral to the quadrate, much of the angular, most of the surangular except its most anterior end, and the posterior end of the splenial; the right articular and a fragment of the angular and posterior dentary are preserved separately. The ventral portion of the right quadrate has been dislocated and was preserved in the right orbit. The left ventrolateral part of the basisphenoid is missing, and both pterygoids are fragmentary. A large piece is missing from the dorsum of the rostrum just anterior to the antorbital fenestra, and another from the right laterosphenoid. The sclerotic ossicles were preserved in the right orbit and were removed in articulation, and a portion of the hyoid skeleton and a portion of the right postorbital and palpebral were also removed along with numerous fragments. A fragment of a large tooth was collected on the surface. During preparation, part of the right palatine was broken off and mistakenly glued to the anterior palatal process of the pterygoid.

The postcranial skeleton was preserved largely in articulation, and was prepared lying on its right side. The right side of the vertebrae and ribs and most of the right

21 shoulder girdle are not exposed, but the incomplete right forelimb and the left, dorsal and ventral surfaces of most vertebrae are visible. Nearly all of the elements of the left forelimb were preserved in articulation, and these were removed from the skeleton. Only three complete and four partial phalanges are preserved on the left side. The right ulna and radius are preserved with their proximal ends articulated with the humerus on the main block and the remainder in pieces separately. The disarticulated atlas elements were preserved with the skull, the axis and following two cervicals were removed from the block when the skull was separated, and a cervical and three posterior thoracic vertebrae were collected separately in the field. 15 cervical and thoracic vertebrae and nearly the entire rib cage are preserved in articulation. Osteoderms and gastralia are not preserved, an interclavicle, clavicles and sternum are not evident but the ventral midline of the skeleton has not been completely prepared. Collected on the surface is a distal caudal vertebra and what may be a sacral rib.

Cranium

Skull Openings:

The antorbital fenestra of Junggarsuchus (Figure 4a) is half the length of the maxilla, and triangular in shape. It lacks a well-developed antorbital fossa, like the one in

Dibothrosuchus (Figure 4b). The maxilla boarders the anterior, ventral and antero-dorsal sides of the fenestra. The lacrimal boarders the fenestra posteriorly and postero-dorsally and the jugal is fully excluded from the fenestra’s border. Relative to the orbit the antorbital fenestra is smaller than the fenestrae in Dibothrosuchus and Sphenosuchus, which have fenestrae nearly as large as their orbits. The antorbital fossa is partially

22 developed and fills much of the fenestral space in the latter two taxa. However, the openings including the fossa are relatively larger than those Junggarsuchus. The antorbital fenestra of Junggarsuchus is still large (half the size of the large orbit), but this relative size is similar to the reduced size of the antorbital fenestra in Protosuchus and early Crocodyliformes, though relative to the orbit the fenestra in Protosuchus and

Orthosuchus is much smaller (Nash, 1975; Brown, 1934).

The orbit of Junggarsuchus is circular and large (Figure 4a), at 3.5 cm long it is more than twice the size of the 1.5 cm long antorbital fenestra, and nearly 1/4th the length of the 15 cm skull. The orbit faces laterally and has no dorsal aspect, like Dibothrosuchus and other sphenosuchians, but unlike modern crocodylians which have dorsally facing orbits. Anteriorly the orbit is bordered by the lacrimal, in which the posterior process contributes slightly to the medial wall of the orbit. The antero-dorsal boarder of the orbit is formed by the prefrontal and the postero-dorsal border is formed by the frontal and overlain by the palpebral. The posterior border of the orbit consists nearly entirely of post orbital, with the ventral most portion made up by the jugal. The jugal forms nearly all the ventral boarder of orbit, except the anterior most part. The postero-ventral process of the lacrimal makes up this anteroventral boarder of the orbit.

The orbit of Dibothrosuchus (Figure 4b, 5b) is smaller relative to the size of the skull than in Junggarsuchus, only roughly 1/6th the length of the skull, small relative to other sphenosuchians like Sphenosuchus (Walker, 1990). In Dibothrosuchus the prefrontal contributes far more to the anterior portions of the orbit than Junggarsuchus.

The prefrontal also contributes more to the medial wall of the orbit as well. It also appears that the jugal makes up the posterior border of the orbit as opposed to the post

23 orbital (Wu and Chatterjee, 1993). The orbit of Dibothrosuchus lacks the prefrontal overhang seen in Junggarsuchus.

The supratemporal fenestra is long, nearly 1/4th the length of the skull and is triangular in shape (Figure 5a). The fenestra narrows anteriorly along with the skull table, like in Almadasuchus. In Junggarsuchus, the lateral and posterior border of the supratemporal fenestra is the parietal while the squamosal contributes to the posterolateral corner and most of the lateral border. The frontal contributes slightly to the fossa but does not contribute to the fenestra. The anterior border of the supratemporal fenestra is comprised nearly entirely of the postorbital, if the first interpretation of the postorbital is accurate. If not, the squamosal borders the anterior edge and the postorbital is not involved at all.

In Dibothrosuchus the fenestra is smaller relative to the skull roof and oval with a similar axis (Figure 10d). Overall, the borders of the fenestra are largely similar, though the parietal contributes more to the postero medial edge of the fenestra, the postorbital comprises the anterior border and the frontals clear contribute to the border anteriorly. In addition, Dibothrosuchus has a more developed supratemporal fossa. A dorsally expanded pneumatic portion of the prootic floors the posterior half of the fenestra.

The infratemporal fenestra of Junggarsuchus, though incompletely defined, appears similar in shape to Sphenosuchus (Walker, 1990), as it is anteroposterior narrow and dorsoventrally tall (Figure 4a). The borders of the infratemporal fenestra in

Junggarsuchus are not all clearly defined due to unclear sutures and crushed and incomplete jugal, quadratojugal and postorbitals. The postorbital appears to form the anterodorsal border of the fenestra. The posterior dorsal border of the fenestra appears to

24 be comprised of the squamosals. The quadratojugal forms the posterior border and some of the posterior ventral border of the fenestra. The ventral border and ventral portion of the anterior edge of the fenestra may be formed by the jugal. Unlike in Sphenosuchus and

Protosuchus the jugal does not appear to extend posterior beyond the infratemporal fenestra.

The borders of the infratemporal fenestra of Dibothrosuchus are not well preserved, but the reconstruction by Wu and Chatterjee based on available material reconstructs the fenestra as longer that the orbit, unlike Junggarsuchus. The postorbital contributes to the entire anterior border of the fenestra and the anterior half of the dorsal border, unlike the condition in Junggarsuchus. Other differences include that the ventral border is comprised only of the jugal and the quadratojugal only contributes to the posterior border. The ventral border of the infratemporal fenestra is flat, unlike the narrow, rounded ventral border seen in Junggarsuchus and Protosuchus. Like

Junggarsuchus the posterior half of the dorsal border is comprised of the squamosal.

The choanae of Junggarsuchus are narrow and slit like, 6 times as long as they are wide at the center. They narrow anteriorly and posteriorly. The maxilla borders the choanae anteriorly and anterolaterally, the vomer comprises the entire medial border and the palatines comprise the lateral and posterior edges of the choana (Figure 6a).

The borders of the choanae of Dibothrosuchus are the same as seen in

Junggarsuchus. The choanae themselves are slightly shorter and wider than in

Junggarsuchus. The pterygoids are not involved with the choanae as they are in crocodyliforms (Figure 6b).

The suborbital fenestra of Junggarsuchus is not clearly defined. The incomplete nature of the palatine makes it difficult to determine the exact size of the fenestra. The anterior boarder of the fenestra appears to be made up by the palatine exclusively (Figure

6a). A thin extension of the palatine encloses the anterior half of the lateral boarder of the fenestra. The antero-medial boarder is also likely comprised of the palatine. It is not clear how the pterygoid bordered the medial and posterior boarder of the suborbital fenestra.

The posterolateral edge and part of the posterior edge are enclosed by the ectopterygoid.

The fenestra appears to narrow posteriorly.

Dibothrosuchus has a large suborbital fenestra relative to the orbit, that appears more oval than in other sphenosuchians (Figure 6b). The borders of the fenestra are similar to those hypothesized above in Jungarsuchus. The pterygoid appears to contribute more to the posterior border of the fenestra. The lateral posterior process of the palatines is not preserved, so the extent of the palatines contribution to the lateral border is unclear.

Wu and Chatterjee reconstruct the lateral process of the palatine as totally bordering the lateral edge of the fenestra.

Bones of the Cranium:

The premaxilla is mostly complete except for the median dorsal process that would have divided the narial opening (Figure 4a, 5a, 6a). The anterior end of the right premaxilla has been pushed slightly towards the left, so that the narrow base of the broken internarial bar is a few mm left of the skull midline, and the facial portion of the left premaxilla has been displaced slightly medially where it contacts the maxilla. The premaxillas’ suture with the maxilla is vertical, which is not seen in other

26 sphenosuchians, but is comparable to the suture between the two bones seen in

Protosuchus and other protosuchians. In both Dibothrosuchus and Sphenosuchus the sutures are more postero-dorsally directed (Figure 4a). In ventral view the suture is straight. The preserved portion forms the ventral and posteroventral borders of the external nares, which faced anterolaterally (Figure 6a). The facial process posterior to the nares is approximately equal in length to the portion anterior to the posterior border of the nares. The shorter posterior process of the premaxilla is comparable to some crocodyliforms, like Fruitachampsa and Dyrosaurus. Like other ‘sphenosuchians’, the premaxilla widens posteriorly and the a posterodorsal process of the premaxilla extends from the facial process between the anterior portions of the maxilla and nasal. Its nearly flat dorsal edge abuts the lateral edge of the nasal, and the process extends posteriorly to above the level of the first preserved maxillary tooth. The dorsal edge of this process on the right premaxilla has a small indentation on its medial surface close to the narial border, but it is absent on the left side (Figure 5a). Assuming the internarial process was similar to other crocodylomorphs, the openings were narrow and elliptical in lateral view with the long axis running anteroventral-posterodorsal, which is also seen in

Dibothrosuchus. The external surface of the facial process immediately ventral to the narial opening anteriorly has a shallow narial fossa but a distinct border is lacking. There is no notch between the maxilla and premaxilla laterally where the fourth dentary tooth occludes, like the one seen in Dibothrosuchus and some other basal crocodylomorphs, but an internal notch is present within the maxilla (see below), the same kind that is seen in Dibothrosuchus and Sphenosuchus (Figure 5a). The palatal portion of the premaxilla is very brief due to the anterior extent of the maxillary palatal process, and does not meet

27 medially, similarly to other sphenosuchians. A small, undivided incisive foramen is present on the midline where the maxilla and premaxilla meet opposite the posterior end of the narial opening. The ventrolateral edge of the premaxilla is gently convex ventrally, so that there is a gentle ventral concavity along the premaxillary symphysis and at the premaxilla-maxilla contact. Two faint impressions are preserved on the antero-lateral face of the premaxilla, which may be neuro-vascular foramina.

The premaxilla of Dibothrosuchus is largely similar to that of Junggarsuchus, however the two premaxillae of Dibothrosuchus are separate, likely due to post mortem deformation. The premaxilla of Dibothrosuchus is taller, wider and shorter than that of

Junggarsuchus and the postero-dorsal process of the premaxilla is shorter than that of

Junggarsuchus, being less than half the length of the premaxilla anterior to the nares, but still separates the maxilla and nasals (Figure 4b, 5b). The nares face anterolaterally as in

Junggarsuchus. On the anterior end of the premaxilla, anterior to the opening for the nares, there is a similar break to that of Junggarsuchus, which suggests the presence of an internarial bar that separated the naris. The ventral edge of the premaxilla is in line with the ventral edge of the posterior process of the maxilla, unlike Junggarsuchus, in which the premaxilla’s ventral edge is located dorsal to the maxillas ventral edge. The palatal portion of the premaxilla is comparably brief, but on the right premaxilla, a deep notch is present, medial to the 4th premaxillary tooth (Figure 6b). This notch may have occluded the anterior dentary teeth, but is not observed on the left premaxilla. Both elements also have a single small foramen on the anterior edge of the facial portion of the premaxilla, likely the same as that seen on Junggarsuchus (Figure 4a, b). Dibothrosuchus also possesses a slight depression on the facial portion of the premaxilla, but it is less concave

28 than that in Junggarsuchus. Dorsal to the tooth row, the ventral most part of the lateral face of the premaxilla has a slight ridge that runs the length of the premaxilla and separates the tooth row from the rest of the bones lateral face. In Dibothrosuchus this ridge is totally missing, and the bone above the tooth row is smooth. The greatest difference between the premaxilla of Dibothrosuchus and Junggarsuchus is the presence of the notch for the occlusion of the 4th dentary tooth. The notch between the premaxilla and maxilla in Dibothrosuchus is wide and ovate, nearly the length of the naris and more than half as wide (Figure 4b). This notch is totally missing in Junggarsuchus, where the suture is vertically directed, not postero-dorsally, as it is in Dibothrosuchus.

Each premaxilla has five tooth positions, but the fifth tooth is preserved only on the right side. The anterior two right teeth were in the process of replacement as indicated by their small size relative to the teeth in the left premaxilla. Based on alveoli, which all occur as separate alveoli, the relative tooth sizes are 1<2,5<3<4. The posterior but not the anterior edge of at least the 3rd-5th teeth is serrated, but the anterior teeth are too poorly preserved or exposed to examine this. Serrations are similar in size to those of the maxillary teeth. The posterior third and fifth, and probably the fourth, teeth are slightly recurved, but are only slightly compressed labiolingually (Figure 4a, 6a).

Dibothrosuchus has 5 teeth in its premaxilla, with relative sizes 1<2<5<3<4, just as observed in Junggarsuchus. None of the teeth are preserved in their entirety, and what teeth are seen lack serrations (Wu and Chatterjee, 1993) They do however, have circular- ovate cross sections similar to Junggarsuchus teeth (Figure 4b, 6b).

Both maxillae are nearly complete but both are missing a small portion dorsally just anterior to the antorbital fenestra. The facial process (Figure 4a) anterior to the antorbital fenestra is approximately 50% longer than it is tall. Posteriorly, the maxilla divides into two processes that make up most of the dorsal and ventral borders of the antorbital fenestra. The posterodorsal process meets the lacrimal approximately halfway along the dorsal edge of the antorbital fenestra; the suture between them is poorly preserved but the lacrimal is lateral to the maxilla. The posterodorsal process is longer than those observed in other sphenosuchians, and appears to nearly totally separate the medial surface of the lacrimal from the lateral edge of the nasal. This posterodorsal process underlays the anterior edge of the lacrimal. The posteroventral process makes up the entire ventral border of the antorbital fenestra. The fenestra is approximately triangular with corners anteriorly, posterodorsally and posteroventrally and is taller and less elongate than in some ‘sphenosuchians,’ such as Terrestrisuchus. The ventral edge of the fenestra slopes downward posteriorly relative to the ventral edge of the maxilla. The posteroventral process of the maxilla tapers gradually posteriorly, where the lacrimal broadly overlaps its posterior end. The tapered anterior end of the jugal inserts into the lateral surface of the posteroventral process of the maxilla to end above the last maxillary tooth and below the middle of the lacrimal’s ventral edge; the maxilla-jugal overlap extends for 10mm. The premaxillary contact is extensive and nearly vertical anteriorly, and the maxilla’s anterior edge is slightly convex on the left side but not the right. The maxilla quickly curves posteriorly dorsally, and it covered dorsally by the nasal along their straight, poorly preserved contact. The ventral edge of the maxilla is gently convex beneath the enlarged anterior teeth and becomes straight posterior to the sixth tooth.

Anterior to the antorbital fenestra the maxilla forms a very brief fossa, preserved on the left side, and dorsal to the fenestra the fossa continues into a groove along the ventral edge of the maxilla’s posterodorsal process. There is no evidence for an extensive medial wall to the antorbital fossa, as in Dibothrosuchus and Hesperosuchus. Small ventrolaterally opening nutrient foramina pierce the ventrolateral surface of the maxilla dorsal to the tooth row, 12 on the right maxilla and 14 on the left (Figure 4A). They are not evenly spaced, or evenly sized and are slightly fewer than the number of teeth, and are denser above the third tooth. Along the medial face of the dorsoposterior process of the maxilla there is a groove, which continues onto the anterior medial surface of the lacrimal.

The two maxillae (Figure 4b) of Dibothrosuchus are nearly complete and broadly similar to those in Junggarsuchus. Only the right maxilla is missing the posterior most process of the maxilla that contacts the jugal. The maxillae are wider than those in

Junggarsuchus and bow outward posteriorly, though this lateral displacement is likely due at least in part to post mortem crushing (Figure 5b). Unlike Junggarsuchus, the anterior end of the maxillae is laterally concave due to space for the enlarged 4th caniniform tooth between the maxilla and premaxilla. The ventral edge of the maxilla is even more gently concave near the enlarged maxillary teeth than Junggarsuchus. The maxilla overlaps any lateral exposure of the nasal on the skull. As discussed above, the medial wall of the antorbital fossa is far more extensive in Dibothrosuchus than it is in

Junggarsuchus. The fossa also extends father posteriorly. Like Junggarsuchus, several ventrolaterally opening nutrient foramina pierce the ventrolateral surface of the maxilla.

They are smaller and fewer than the ones on Junggarsuchus, with 8 to 9 occurring dorsal

31 to the tooth row. There also appears to be an additional row of small foramina on the dorso-lateral face of the ventral posterior process of the maxilla (Figure 4b).

On the dorsal face of both maxillae, there are two dorsal openings. The more anterior one (illustrated, but not described by Wu and Chatterjee, 1993), is located in line with the second maxillary tooth. The more posterior one is smaller, and in line with the

4th maxillary tooth. These openings are not seen in any other sphenosuchians or crocodyliform. They appear to be due to post mortem crushing as they are associated with the roots of the tooth they correspond to. During deformation, as the dorsal surface of the skull was compressed, the roots of these teeth punctured the lateral wall of the maxilla, making weak spots.

The palatal processes of the maxillae in Junggarsuchus (Figure 6a) meet medially to form a secondary bony palate. The secondary bony palate begins far anteriorly, between the premaxillae, and extends posteriorly to the position of the fourth maxillary tooth. Anteriorly, the maxilla forms a pocket medial to the premaxillary contact into which the fourth dentary tooth inserted. This pocket is open dorsally, visible from the narial opening. Posterior to this, the secondary palate becomes flatter in cross section and appears to thicken in CT scans, especially along the medial surface of the maxilla, where the two bones form a low midline ridge dorsally. The maxilla forms only the lateral and anterior boarders of the choanae.

The palatal process of Dibothrosuchus (Figure 6b) is slightly wider than

Junggarsuchus, though the partial separation is due to compression. The palatal shelf extends back to the position of the 5th maxillary tooth. A small medial extension of the palatal shelf forms the anterior and anterior most medial edge of the choanae. This

32 process may be present in Junggarsuchus, but is broken. However, there is a concavity on the ventral surface on the vomer that indicates its potential presence.

We infer fourteen tooth positions in each maxilla in Junggarsuchus (Figure 4a,

6A), the fourteenth is represented by an apparent tooth fragment in this position on the left side and alveolus on the right. On both sides, the second and fourth teeth have been lost. The labial edge of the maxilla bulges laterally between the first and third teeth, indicating an alveolus, but the right maxillary edge extends inward at this position, possibly a pathology. The third tooth is the largest and the alveolus for the fourth is smaller, while the first and fifth teeth are of similar size; apparently the teeth gradually increased in size up to the third tooth, after which they decreased posteriorly. The third tooth is nearly twice as long as the fifth tooth, and the size of the alveoli and the ventral excursion of the maxilla in this region indicate that the second to fourth teeth were unusually large compared to other basal crocodylomorphs and may have formed a functional unit separate from the posterior teeth. The fifth tooth is slightly smaller than the first tooth, and all teeth posterior to this become gradually smaller in size, until the last and smallest tooth is only 3mm long. All of the teeth, except possibly the posterior one, are recurved, and the 6th and 7th are strongly recurved. The posterior edge of each maxillary tooth is serrated in a similar manner, and the seventh tooth of the left side has 6 serrations/mm. The anterior edge is also serrated on its distal half from the sixth tooth posteriorly, but the first and third teeth lack serrations anteriorly. A small, loose tooth is preserved on the lateral surface of the right dentary beneath the posterior end of the tooth row, similar in size to the 13th preserved maxillary tooth, and therefore it is possibly the

14th tooth.

Dibothrosuchus has positions for 15 maxillary teeth on both sides (Figure 4b, 6b).

On the left side only alveoli 8, 10 and 15 are empty, and on the right side only alveoli 4,

14 and 15 are empty. The first two maxillary teeth are small, the second slightly larger than the first, but both are barely exposed laterally. The largest teeth and alveoli are the

3rd and 4th teeth. Both the 4th alveolus and tooth are slightly larger, but neither the 3rd or

4th tooth are preserved entirely, those seen are missing the ends of the crown. The 5th tooth is smaller than the 3rd and 4th, but larger than the others. Relative to the height of the maxilla the enlarged maxillary teeth are not as large as those of Junggarsuchus. Like

Junggarsuchus the rest of the teeth decrease in size posteriorly. The maxillary teeth of

Dibothrosuchus are recurved, but less recurved than those of Junggarsuchus. The 7th tooth is the most recurved. Like Junggarsuchus the enlarged lanceolate caniniform teeth lack anterior serrations, and from the 6th tooth back are serrated posteriorly and anteriorly.

The nasals of Junggarsuchus are paired, long, narrow bones that widen posteriorly along with the rest of the skull and make up the anterior half of the skull roof, similar to Protosuchus and other crocodyliforms (Figure 5a). A large, central area where the nasal would have contacted the maxilla on both the right and left sides is missing, and the left nasal is also damaged anterior to this gap. Anterior to the prefrontal their lateral edge bends ventrally, dividing the bone into dorsally and laterally facing parts. The medial part is slightly convex dorsally in the anterior half of the bone, resulting in a dorsal midline groove. Posteriorly, the nasals are nearly flat and rise medially to form a low midline ridge. The anterior end of the nasals makes up a short posterodorsal rim of the external nares, and extends between the premaxilla. The nasal contacts the

34 posterodorsal process of the premaxilla lateroventrally, and the nasal widens slightly anterior to this process. The nasal ends anteriorly in the broken base of the internarial process, which is broad and dorsoventrally flattened. The posterior end of the nasal does not extend between the frontals where they meet opposite the anterior end of the orbit.

The nasal extends between the prefrontal and frontal, where it overlies, mostly the frontal and some of the prefrontal. The posterior process is similar to those in Dibothrosuchus and Hesperosuchus, but the portions of the bone extending between the prefrontals and frontal are much shorter, though the contact is not as nearly transverse as that seen in

Sphenosuchus. Anterior to the prefrontal the nasal has a short contact laterally with the anterodorsal process of the lacrimal. Unlike other sphenosuchians the lateral edge of the nasals does not contact the medial edges of the maxilla. They are separated by an elongate posterior dorsal process of the maxilla and an anterior process of the prefrontal.

This is comparable to the conditions seen in Protosuchus and Orthosuchus.

The paired nasal of Dibothrosuchus is preserved more completely (Figure 5b), with the exception of the anterior end that extends between the premaxilla that would meet the internarial bar. The nasals have been displaced slightly ventrally, due to post mortem crushing. The nasal is largely similar to that in Junggarsuchus, it contacts the dorsal edge of the maxilla anteriorly, and contacts the medial edges of the prefrontals dorsally. There is also a brief contact with the medial edge of the lacrimals dorsally.

Unlike the nasals of Junggarsuchus, the nasals of Dibothrosuchus do not widen posteriorly, and are relatively wider than those of Junggarsuchus relative to length, unlike Protosuchus. The two bones are flatter, lacking the slight dorsal ridges present in other sphenosuchians. The nasals also lack the lateral aspects seen in Junggarsuchus, and

35 possess a more distinct fork of the posterior nasals, which extend between the prefrontals and the frontals. These twin posterior processes are separated by an anterior process of the frontals, and are wider than those seen in Junggarsuchus.

The lacrimal is in the shape of an inverted L with a long anterodorsal process, and is approximately as long as it is high (Figure 4a, 5a). Its ventral process is nearly vertical anteriorly, forming the posterior edge of the antorbital fenestra. The anterior edge of this process has a deep longitudinal groove that becomes open laterally near the base, forming a narrow antorbital fossa. The posterior edge of this process curves posteroventrally, forming the anteroventral edge of the orbit, and is longer than the process seen in Dibothrosuchus. On the right side of the skull, the bone has been partially crushed, the left element is better preserved. The lacrimal is very narrow on the skull roof, and contacts the prefrontal medially and posteriorly above the preorbital bar. The posterior contact with the prefrontal is abrupt, and the prefrontal covers the posterior edge of the entire dorsal part of the lacrimal. In Dibothrosuchus the contact with the prefrontal is more extensive ventrally, and the suture is vertical. The tapering anterodorsal process overlies the maxilla approximately at the midpoint of the antorbital fenestra, but this suture has been damaged on both sides of the skull. The dorsal part of the lacrimal has a rough lateral surface, while the descending process is smooth. Medial, within the skull, a large but shallow pocket is visible investing the body of the lacrimal, and the wall of the lacrimal body is relatively thin, rather than being a solid bone. This is where the paranasal sinus is housed. The descending process of the lacrimal has an anteriorly opening foramen on the lateral side, which is referred to as the lacrimal foramen (Figure

4a). Unlike Dibothrosuchus, the foramen is entirely enclosed by the lacrimal.

The lacrimal of Dibothrosuchus is similar to Junggarsuchus (Figure 4b). The antero-dorsal process of the lacrimal is far longer than the process seen in Junggarsuchus, and forms the posterior half of the dorsal boarder of the antorbital fenestra. The lacrimal is longer that the prefrontal antero-posteriorly, which is similar to the relative length of the lacrimal to the prefrontal seen in Protosuchus and crocodyliforms, but not

Junggarsuchus. The groove that extends into an anterior antorbital fossa is far less developed in Dibothrosuchus. The lacrimal appears to lack the posterior projection that overlays the anterior portion of the prefrontal, but the bone is crushed in this region on both sides, obscuring potential sutures. As in Junggarsuchus, the lacrimal is thin walled and possesses an enlarged hollow space in the anterior body of the bone, larger than that seen in Junggarsuchus. The lateral surface of the lacrimal, halfway down the dorsoventral suture with the prefrontal, has a small opening for a lacrimal foramen, which is only enclosed anteriorly by the lacrimal, as opposed to Junggarsuchus in which the foramen is fully enclosed by the lacrimal.

The rhomboidal prefrontal of Junggarsuchus overhangs the orbit anteriorly

(Figure 4a, 5a). Its broad descending process extends into the anterodorsal region of the orbit, where its robust lateral part borders the lacrimal posteriorly. This process forms the dorsal half of the orbit anteriorly, and medially it curves posteriorly to form a posterolaterally facing fossa. The medial part of the descending process is much thinner than the lateral part, its ventral edge is horizontal and its posterior edge is vertical (Figure

4a). Anterodorsally, the prefrontal narrows and extends between the lacrimal and the nasal. The contact with the frontal is approximately as long as the contact with the nasal.

Posteriorly, the prefrontal does not appear to extensively underlie the frontal, as it does in

37 other ‘sphenosuchians’ except Dibothrosuchus. Its dorsal surface is shallowly concave posteriorly and becomes slightly convex in the area where it contacts the lacrimal.

Junggarsuchus has a laterally expanded prefrontal which forms a prefrontal overhang which is not observed in Dibothrosuchus, Sphenosuchus or other sphenosuchians. The overhang is enlarged and oblique, similar to the overhang seen in thalattosuchians. The posterior face of the orbital fossa of the prefrontal has a small foramen that is directed posteriorly.

The rhomboidal prefrontal of Dibothrosuchus (Figure 4b, 5b) is largely similar to that seen in Junggarsuchus. More of the descending medial and posterior processes of the prefrontal are preserved in Dibothrosuchus. The antero-lateral section of the prefrontal does not feature a concavity for the posterior process of the lacrimal, and lacks the two anterior processes of the prefrontal found in Junggarsuchus and other sphenosuchians.

The antero-dorsal process of the prefrontal is longer and has a longer contact with the nasal than it does in Junggarsuchus. The posterior dorsal suture of the prefrontal with the frontal is not very clear, but it appears that not much of the prefrontal extends under the frontals. As in Junggarsuchus the dorsal surface in concave medially and convex and ridge like laterally, though this may have been exaggerated by crushing. Dibothrosuchus lacks a prefrontal overhang.

The descending process of the prefrontal in Dibothrosuchus (Figure 4b) extends farther ventrally and posteriorly than the process in Junggarsuchus. The descending process is more extensive and forms the anterior-medial wall of the orbit than any other sphenosuchian, including Sphenosuchus. The medial contact of the prefrontals to form a

“transverse-brace” reported by Wu and Chatterjee is not observed in the CT scans of the

38 skull. The descending process contacts the palate at the point that the posterior edges of the palatine meet lateral edges of the pterygoid. The contact between the prefrontal and palate is not observed in any other sphenosuchians, but is present in later crocodyliforms.

This contact is not as extensive, as in neosuchians and would not have provided the support it does in those forms. The contact may be due to dorsoventral crushing of the skull.

In Junggarsuchus the frontals make up the skull roof medial to the orbits, posterior to the nasals and prefrontal, and anterior to the parietal and postorbital (Figure

4a, 5a). The frontal forms the dorsal margin of the posterior part of the orbit. The orbital margin is only preserved on the left side, where the palpebral covers it, and the frontal appears to be laterally concave. However, this concavity may be accentuated by the palpebral, which has been pressed strongly and unnaturally onto the surface of the frontal. Inside the orbit, the crista cranii forming the lateral margin of the olfactory tract is thin but descends much further than in living crocodylians (Figure 4a). The crista is incomplete, but its anterior end is preserved on both sides where it contacts the medial surface of the prefrontal’s descending process. Posteriorly a fragment of the left crista is preserved on the lateral surface of the braincase. A broad, low longitudinal ridge, similar to that of Sphenosuchus and Hesperosuchus, runs along the midline of the frontals the entire length. Anteriorly, the frontals taper to fit bluntly between the posterolateral processes of the nasals, and the frontals are overlain by the nasals on the midline.

Anterolaterally the frontal contacts the prefrontal along a straight suture.

Posterolaterally, the frontals have a gently concave contact with the dorsal part of the postorbital. The posterior end of the frontals is broken, corresponding to a large fracture

39 in the specimen, the posterior end of the frontal does not extend as far laterally as in

Hesperosuchus, Sphenosuchus and crocodyliforms, giving the supratemporal fenestra its triangular rather than oval shape. The frontal subtly extends into the supratemporal fossa, but not the fenestra which is demonstrated as a mixed color mask in Figure (5c,d).

The frontal of Dibothrosuchus is similar in position to other sphenosuchians

(Figure 4b, 5b). However, the frontal is wider than that of Junggarsuchus and lacks the deep lateral concavity seen in Junggarsuchus. The concavity is far shallower, which may be related to the lack of a palpebral. The crista cranii in Dibothrosuchus is less ventrally developed than in Junggarsuchus, though the crista cranii may be broken. The frontal is also wider laterally, giving the supratemporal fenestra a more oval shape, but lacks the posterolateral processes seen in Sphenosuchus. The dorsal parasagittal ridges of the skull are better developed in Dibothrosuchus than they are in Junggarsuchus, which has low dorsal ridges. The median ridge of the frontal is divided by a wide groove along the midline resulting in 2 midline ridges around the central ridge along the suture. These ridges converge anteriorly and posteriorly into a lanceolate shape (Wu and Chatterjee,

1993). Posteriorly the parasagittal ridges are laterally separated from the postorbitals by another deep groove and ridge along the suture. This is a feature unique to

Dibothrosuchus (Wu and Chatterjee, 1993). Anteriorly, the frontals narrow and have an extensive anterolateral contact with the nasals, which is more extensive than the contact seen in Junggarsuchus.

The left palpebral is seen only in Junggarsuchus and is preserved in contact with the frontal and postorbital bones above the left orbit, its lateral edge displaced slightly ventromedially from its presumed sub-horizontal position (Figure 4a, 5a). It is ovoid in

40 dorsal view, with an anteromedial-posterolateral long axis that divides the bone nearly symmetrically. It is dorsally convex and covered with a low, rugose sculpturing. Its posterior edge is preserved contacting the anterior edge of the postorbital and roughly reflects the latter’s shape. This edge is only gently curved, less so than other edges. The posterolateral end of the bone is more acute than the anteromedial end. The medial part of the bone, which overlies the frontal, has a small notch. In the only other basal crocodylomorph for which a palpebral is known, Hesperosuchus agilis (CM 29894), is more circular in shape, thicker, and has very fine, extensive sculpturing. No palpebral is known from Dibothrosuchus but the lack of firm suturing of this bone to the skull could explain its absence.

The anterior tip of the triradiate jugal in Junggarsuchus inserts into the posterior end of the maxilla, where the ventral process of the lacrimal borders it dorsally (Figure

4a). Posteriorly the jugal widens mediolaterally, where it forms the ventral border of the orbit; this region is marked by a concave longitudinal depression along its entire ventrolateral surface. The bone then curves posterodorsally to form the posteroventral border of the orbit and the ventral half of the postorbital bar. Thus, its ventral edge is not flat, as in other sphenosuchians like Dibothrosuchus and in many crocodyliforms but ventrally concave beneath the postorbital bar, opposite the dorsal convexity of the surangular. The medial surface beneath the orbit also possesses a longitudinal groove, bordered ventrally by a horizontal ridge along the ventral part of the bone (Figure 6a).

The dorsal process of the jugal lies directly medial to the postorbital, unlike other sphenosuchians and most crocodyliforms, but as in thalattosuchians.

Posterior to the postorbital bar, the concavity on the ventrolateral surface of the jugal opens into a broad, thin, medially depressed lower temporal bar (Figure 4a). It is not clear which part of this region is formed by the jugal and which by the quadratojugal.

The jugal most likely continues dorsally posterior to the postorbital, where there is a distinct suture, but this could also be the quadratojugal, as in some sphenosuchians (Clark et al., 2000). The quadratojugal appears not to extend that far anteriorly, as discussed below, and the anterior half of the lower temporal bar appears to be mostly jugal. The posterior process of the jugal arches dorsally, similar to Protosuchus, posterior to the postorbital bar and is slightly shorter than the anterior process, unlike other sphenosuchians (Wu and Chatterjee, 1993; Walker, 1990). Posteriorly, the jugal likely ended where the quadratojugal curves dorsally to form a vertical anterior edge, and a flake of bone there is probably jugal. There is possibly extensive contact between the jugal and quadratojugal, like the contact seen in protosuchians that reduces the size of the infratemporal fenestra. The lower temporal bar is very broad dorsoventrally and very thin mediolaterally.

Only the anterior end of the left jugal is known from Dibothrosuchus IVPP V

7907 (4B). The jugal is better known from CUP 2981, which allows for comparison to

Junggarsuchus. The anterior tip of the jugal has two anterior processes, and the dorsal anterior tip just barely participates in the posterior boarder of the antorbital fenestra. The jugal narrows posteriorly, unlike Junggarsuchus. The posterior process of the jugal of

Dibothrosuchus, unlike Junggarsuchus is straight, like most other sphenosuchians, not dorsally arched. The dorsal part of the jugal that contacts the post orbital bar lies lateral to

42 the postorbital, as in most crocodyliforms. As in crocodyliforms the supratemporal fenestra does not narrow anteriorly like Protosuchus and unlike it does in Junggarsuchus,

The parietal in Junggarsuchus lacks any trace of a midline suture, unlike in some other ‘sphenosuchians’ such as Litargosuchus, Hesperosuchus and Dromicosuchus

(Figure 4a, 5a). The parietal is defined by a sharp T-shaped crest that comprises the sagittal crest and is continuous posteriorly with a supraoccipital crest that runs mediolaterally along the entire occipital portion of the skull roof. Anteriorly the sagittal crest continues onto the posterior end of the frontals, but the contact between the frontal and parietal is obscured by a large crack. In dorsal view, the crest along the posterior margin of the parietal is straight, as opposed to the V-shaped crest seen in almost all other

‘sphenosuchians’ except Dibothrosuchus, Sphenosuchus and Almadasuchus. The occipital crest continues laterally and curves anterolaterally at the posterolateral corner of the supratemporal fossa, and then continues onto the squamosal. The lateral edge of the body of the parietal is dorso-laterally convex and forms the medial and posteromedial border of the supratemporal fenestra. The parietal meets the squamosal in a vertical suture approximately midway around the posterior edge of the fenestra. A small anterior opening to the temporo-orbital foramen is situated between the parietal and squamosal near the ventral end of their contact within the fenestra, and the parietal appears to form the medial and dorsal edge of the foramen. The prootic forms the ventral edge. The posterodorsal part of the supratemporal fenestra faces anterodorsally, and forms only a brief shelf rather than the much more extensive one in crocodyliforms, Dibothrosuchus and Almadasuchus. The parietal sends a small process onto the occipital surface, rhomboidal in posterior view, wedging into the dorsal edge of the supraoccipital (Figure

7a). The parietal also extends onto the occipital surface between the supraoccipital and squamosal and resting on the paroccipital process. It forms the medial extent of the dorsal border of the posterior opening of the post-temporal foramen, like in

Sphenosuchus. The occipital portion is triangular, with a low, gently convex ventral end and a broad dorsal base. It extends dorsolaterally as a slender process over the squamosal to reach the posterodorsal corner of the supratemporal fossa. It is less extensive than in

Dibothrosuchus and Sphenosuchus, though not as small dorsoventrally as in basal crocodyliforms. The supraoccipital covers the parietal on the occiput medial to this except for the midline process. The dorsal roof of the braincase is formed by the parietals, which is the condition seen in all crocodylomorphs. As in Dibothrosuchus and other sphenosuchians the parietal of Junggarsuchus is not pneumatic and no parietal diverticulum is known. Though the parietals contact with the frontals is not well preserved in Junggarsuchus due to a break, it appears that a small portion of the parietal may project between the potential posterior extension of the frontal and the laterosphenoid. This is similar to the condition seen in some thalattosuchians, however the process is not elongate and does not participate in the supratemporal fossa as seen in some thalattosuchians.

The parietals of Dibothrosuchus (Figure 4b, 5b, 10a, 10e) have a lower T shaped sagittal crest than Junggarsuchus and appear to have a visible midline suture anteriorly, though it is only visible due to a break in the sagittal crest. The parietals anterior contact with the frontal is blunt and rectangular, though there is a slight anteromedial process that projects anteriorly. The lateral expansions of the occipital ridge do not extend as far laterally as those of Junggarsuchus and contribute to less than half of the medial posterior

44 border of the supratemporal fenestra. The post-temporal foramen is much larger in

Dibothrosuchus, similar to Sphenosuchus rather than Junggarsuchus. The parietal does not contribute to the edges of the foramen in Dibothrosuchus, where the medial and ventral edge is formed by the prootic and the dorsal edge by the squamosal. The parietal and prootic contribute to a broader supratemporal shelf than seen in Junggarsuchus. The parietal of Dibothrosuchus is involved in the occipital portion of the skull, which has a medial rhomboid projection into the supraoccipital and expanded rectangular expansions that separate the squamosal and supraoccipital. The post temporal fenestra is slightly larger than that in Junggarsuchus, and unlike Junggarsuchus, the parietals do not contribute to the medial or dorsal edge of the fenestra. The parietals end dorsal to a thin process of the squamosal that forms the boarder of the posterior temporal fenestra. The parietals involvement in the posterior temporal fenestra is seen in more basal sphenosuchians, but in Junggarsuchus the supraoccipital approaches the medial margin of the fenestra and very subtly contributes to the medial margin of the posterior temporal fenestra, like in some protosuchians, and similar to the condition in crocodyliforms, in which the supraoccipital forms the medial and ventral edge of the fenestra.

The ventral process of the postorbital in Junggarsuchus makes up the dorsal half of the postorbital bar and has a broad dorsal portion (Figure 4a). The postorbital overlies the jugal anteriorly, making up the posterior border of the orbit as in other

‘sphenosuchians’, but unlike the unusual condition of Dibothrosuchus in which the postorbital is posterior to the jugal and the jugal forms the posterior border of the orbit.

However, the condition in Dibothrosuchus is similar to the condition seen in crocodyliforms. This descending process extends medially as a broad sheet that meets

45 the laterosphenoid (Figure 8c). A descending process along the lateral surface of the laterosphenoid is preserved on the left side, an unusual condition if true. Dorsally, the suture between the postorbital and the frontal is semicircular in shape, with the convex area directed anteriorly (Figure 5a). The frontal lies medial to the postorbital and the concave posterolateral edge of the frontal articulates to a convex medial edge of the post orbital. A narrow lateral expansion of the frontal boarders the postorbital anteriorly. The projection of postorbital limits the lateral expansion of the frontal. The posterior extent of the postorbital is difficult to determine, and two possible interpretations exist though one would be unusual. The first is that the post orbital has a relatively short posterior process and the squamosal extends far forward. This process is directed postero-medially and is diamond shaped. Its medial edge is bordered by the parietal, and potentially a thin portion of the frontal. The lateral edge of the process is sutured to the medial edge of the anterior process of the squamosal. The posterior process of the postorbital reaches the anterolateral edge of the supratemporal fenestra in this interpretation. The contribution to the anterior and lateral edge of the fenestra is brief, and most of the lateral border is made up by the squamosal (Figure 5a).

The more unusual interpretation is that a longitudinal suture between the postorbital and squamosal in the anterior part of the supratemporal bar indicates that the postorbital forms the anterolateral part of the bar and does not border the supratemporal fossa (Figure 5c). Thus, rather than being medial to the squamosal, as in Saltoposuchus and Dibothrosuchus, or forming the anterior half of the bar as in Hesperosuchus, it lies lateral to the squamosal as a long rectangular process, half the length of the squamosal and unlike the post orbital of any known sphenosuchian or basal crocodyliform. The

46 posterior extent of the postorbital is unclear, but it apparently ended about half way along the bar. The postorbital is strongly concave ventrally where it overhangs the lateral temporal fenestra, continuous with the concavity in the squamosal.

Only the dorsal portion of the postorbital is preserved in Dibothrosuchus (Figure

4b, 5b). The ventral portion of the postorbital bar is preserved on the CUP holotype. The dorsal portion of the postorbital has a medial ridge that contacts the frontal along a smoothly concave contact. The more lateral surface of the dorsal portion of the postorbital is slightly convex, unlike the smooth dorsal portion of the postorbital in

Junggarsuchus. The dorsal part of the postorbital is also wider. A broad medial expansion of the postorbital that contacts the laterosphenoid is not found in

Dibothrosuchus. The postorbital portion of the postorbital bar is posterior to the ascending process of the jugal, which is unlike the condition seen in other sphenosuchians, but similar to Protosuchus and other crocodyliforms.

In Junggarsuchus the squamosal is a kidney-shaped bone that broadly overhangs the infratemporal fossa (Figure 4a, 5a, 7a). It is very broad posteriorly, more similar to

Saltoposuchus than to the narrower squamosal of Dibothrosuchus and Sphenosuchus. It tapers anteriorly along the lateral edge of the supratemporal fenestra, reaching the anterior edge of the fenestra where it may contact the postorbital laterally and, unusually, the frontal anteromedially. The exact contact between the squamosal and postorbital is unclear, so there are two interpretations of the anterior portion of the squamosal, which have been outlined in the discussion of the postorbital. The first possible condition, which is similar to the conditions seen in sphenosuchians, is a laterally extensive squamosal. In this case, the squamosal still narrows anteriorly, but the postorbital contributes

47 anterolaterally to the supratemporal fenestra and is not excluded from the border by the squamosal (Figure 5a). The lateroventral surface of the anterior portion of the squamosal, which appears as a thinner sheet of bone, anteroposteriorly extensively overhanging the lateral fenestra. The anteromedial edge of the bone contacts the potential posterior projection of the postorbital. The alternative, with a long posterolateral process of the postorbital fully separated from the supratemporal fenestra by a narrow anterior portion of the squamosal is unknown in other sphenosuchians (Figure 5c,d). In this case the squamosal would widen substantially posteriorly, contributing to the last third of the laterosphenoids overhang, and the anterior portion contacts a potential posterior process of the frontal anteromedially. The portion of the squamosal anterior to the occiput is ventrally concave.

Unlike most other sphenosuchians, the dorsal edge of the squamosal lacks a sharp ridge along the lateral edge of the supratemporal fossa (Figure 5a). This lack of a ridge along the dorsal surface of the squamosal is similar to the condition seen in protosuchians and other crocodyliforms except thalattosuchians. The squamosal forms the dorsal and posterodorsal portion of the articulation surface for the dorsal head of the quadrate. As shown on the right side where the dorsal part of the squamosal is missing, the quadrate has a broad, stout posterodorsal contact with the occipital portion of the squamosal (Figure 8b). The contact continues anterolaterally along a thin anterolateral process of the quadrate, and the quadratojugal briefly contacts the ventral portion of the squamosal anterior to the quadrate. The quadrate articulation with the squamosal is more limited both mediolaterally and anteroposteriorly than in sphenosuchians such as

Dibothrosuchus (Figure 6a). The contact with the parietal within the supratemporal fossa

48 is obscured by breakage and glue, but appears to extend from the posterodorsolateral corner of the fossa ventromedially to end at the temporo-orbital foramen (the anterior temporal foramen of Walker, 1990), with the squamosal forming the entire dorsal and lateral edges of the foramen (Figure 5a). The occipital ridge on the parietal is continuous laterally with a much shorter ridge on the squamosal extending anterolaterally while quickly diminishing in size. Another similarly brief ridge rises from just ventral to the lateral end of the occipital ridge and continues posteroventrally on the squamosal and onto the posterior edge of the expanded distal edge of the paroccipital process.

The occipital portion of the squamosal is bordered medially by the parietal and ventrally by the paroccipital process of the otoccipital (Figure 7a). Medially on the occiput the squamosal extends ventromedially beneath the parietal, ending as a slender process which forms the lateral and much of the ventral edge to the temporo-orbital foramen and nearly reaching the supraoccipital. The contact with the parietal is thus dorsomedial, unlike the strictly lateral contact in Sphenosuchus, and the occipital surface of the squamosal is triangular like Dibothrosuchus rather than square as in these taxa. As in other ‘sphenosuchians’, a ventral process of the squamosal extends along the anterior edge of the paroccipital process and terminates at the ventral edge of the latter process. It is slightly concave posterolaterally, similar to Almadasuchus and some longirostrine neosuchians but unlike protosuchians. The posterior contact of the squamosal with the dorsal head of the quadrate encloses the otic recess posteriorly, like Kayentasuchus, many protosuchians, later Crocodyliformes and unlike other sphenosuchians like

Dibothrosuchus(Wu and Chatterjee, 1993). It forms the anterior part of a broad vertical

49 depression. Similar to Almadasuchus, a sub triangular concavity is located on the posteroventral process of the squamosal, which contacts the paroccipital.

The right squamosal of Dibothrosuchus is well preserved (Figure 10a). The squamosal forms the entire lateral border, more than half the posterior border and the anterolateral edge of the supratemporal fenestra. The anterior body of the squamosal curves medially, which gives the supratemporal fenestra a circular shape, unlike

Junggarsuchus. The postorbital overlaps the anterior edge of the squamosal in a short triangular process, which is bordered on both the lateral and medial edges by the squamosal.

Unlike Junggarsuchus, but similar to Sphenosuchus there is a ridge along the lateral edge of the supratemporal fenestra along the dorsal surface of the squamosal

(Figure 10d). As in other sphenosuchians the dorsal head of the quadrate contacts the vertical edge of the squamosal, though the mediolateral and anteroventral contact is more extensive than in Junggarsuchus (Figure 10b). In Junggarsuchus the quadrate is involved in the medial wall of the temporo-orbital foramen, but in Dibothrosuchus the prootic is more involved in the lateral wall posteriorly, though the quadrate is involved posteriorly.

The occipital dorsal ridge is longer on the squamosal than the parietal and curves anteriorly. Like Junggarsuchus there is also a posteroventral ridge that rises from the dorsal ridge. The squamosal descends anterior to the paraoccipital process. Part of this process is visible lateral to the paraoccipital process in posterior view (Figure 7b). A ventral extension of the squamosal is also present in Junggarsuchus, but is more laterally extensive in Dibothrosuchus. The ventral portion of the squamosals exposed on the occiput do not contact the parietal laterally as in Sphenosuchus, but sends a thin

50 triangular medial projection of bone beneath the ventral edge of the parietal and forms the entire dorsal border of the posterior temporal fenestra. A very shallow concavity is present on the descending process of the squamosal, in s similar location to where the deeper groove is present in the expanded posteroventral region of the squamosal of the

Junggarsuchus and Almadasuchus. The occipital surface of the squamosal lateral to the process overlaying the fenestra is anteriorly concave, similar to Junggarsuchus.

Dibothrosuchus lacks the posterior concavity of the squamosal seen in Almadasuchus that also appears to be shallowly present on the posterior end of the squamosal in

Junggarsuchus. In Dibothrosuchus’ the squamosal extends far posterior to the quadrate condyle in lateral view, which is a condition found in some sphenosuchians like

Kayentasuchus as well as crocodyliforms like Protosuchus. The quadrate condyle is in line with the squamosals’ posterior edge in Junggarsuchus, most other sphenosuhcians and a number of protosuchians.

The left quadrate of Junggarsuchus is not well preserved—the central area connecting the proximal and distal ends is missing—but many of the details are visible dorsally and ventrally (Figures 4A, 6A). Moreover, the dorsal part of the right quadrate is preserved in place and the middle part of the right quadrate has been displaced intact to the posterodorsal part of the right orbit and rotated so that its dorsal end faces posteroventrally (Figure 8b). The anterior surface of the dorsal part of the quadrate in the posterior end of the supratemporal fenestra is anterodorsally concave. The quadrate narrows dorsally, and at its termination it is much narrower mediolaterally than in

Dibothrosuchus and Sphenosuchus. Posteriorly the quadrate head rests against the anterior surface of the occipital portion of the squamosal, but as in all crocodylomorphs

51 the dorsal process does not appear to widely contact the otoccipital (Figure 8a). In lateral view the articulation surface on the quadrate is gently convex posterodorsally. Medially the dorsal head has a long, firm contact with the prootic (Figure 8c), and ventrally the dorsal head overlies the posterodorsal portion of the prootic enclosing the mastoid antrum, similar to Sphenosuchus and Dibothrosuchus. The trigeminal opening is enclosed by the prootic posteriorly and the laterosphenoid anteriorly. It appears that the trigeminal opening is a single large circular opening similar to that of Dibothrosuchus, enclosed between the laterosphenoid and prootic (Figure 8a). The quadrate extends lateral to the prootic to overhang the otic region slightly. A small, anteroposteriorly elliptical dorsal (superior) tympanic recess may be present within a white matrix-filled area between the right quadrate and prootic posterior to the mastoid antrum, similar to

Dibothrosuchus (Figure 9b). On the right side the quadrate approaches the laterosphenoid, but does not contact it as in Almadasuchus, and, more broadly, in crocodyliforms including Protosuchus (Figure 8b).

At least two fenestrae are present within the quadrate, partially preserved on both elements (Figure 7a, 8b). Both fenestrae pass posteromedial to the pterygoid ramus of the quadrate to connect with an extensive space forming the middle ear region. A complete, dorsoventrally ovoid fenestra is preserved in both elements at about the same level as the jugal’s suborbital ramus, well above the mandibular articulation. There is no evidence for a siphonium that would have passed from the quadrate into the articular, as in living crocodylians. The ventral part of a second fenestra dorsal to the first is better preserved on the right element, and is slightly more elongate than the ventral foramen.

This second fenestra is slightly offset medially from the first one and its long axis is

52 oriented. Like Dibothrosuchus and other sphenosuchians, and unlike protosuchians,

Junggarsuchus lacks the extensive pneumaticity for a quadrate diverticulum, and the bone is sheet like. Below the quadrate fenestrae the quadrate is vertically oriented, and its anterior surface is shallowly concave. The contact with the quadratojugal occurs along the lateral edge and continues dorsally where it briefly contacts the squamosal. Laterally the contact narrows ventrally (Figure 7a). The quadrate condyles on its mandibular articulation do not appear unusual. The condyles are low and the similar length on the lateral and medial sides.

The medial surface of the quadrate is complex, and its contact the quadrate ramus in this area is difficult to determine (Figure 8b). An anteromedial projection from the edge of the quadrate body is anteroventrolaterally convex, and the quadrate forms only the dorsal half of this projection, the pterygoid forming the other half. On the anterolateral surface of this projection the ventral edge of the quadrate’s pterygoid ramus is evident where it overlies the pterygoid laterally (Figure 6a). This horizontal contact is traceable posterolaterally where it turns ventrolaterally and then, at a point dorsal to the quadrate condyles, turns ventrally to end 1 mm above the medial condyle. Three mm further dorsally a stout process extends posterodorsally from the quadrate, broken off after 2mm. A faint suture is present extending dorsally and slightly medially from the ventrolateral part of this process, but it is unclear whether the quadrate forms the dorsolateral part of this process. It seems more likely that it was formed by the ventrolateral process of the otoccipital, but the base of the process is broken and reglued, and whether it is continuous with the quadrate is unclear. It appears to be a proper suture between the ventral portion of the otoccipital to the medial edge of the quadrate. This is

53 also seen in Almadasuchus (Pol, et al, 2013) as well as Protosuchus and other crocodyliforms but not in other sphenosuchians. Whichever bone formed this process, it likely projected dorsally to contact the otoccipital. The quadrate-pterygoid suture continues anteriorly from this process, extending along the medial edge of a dorsal process of the quadrate that forms the lateral border of the ventral fenestra, but anterior to this process the suture is obscure. The dorso-medial portion of the quadrate contributes to the lateral wall of the temporal-orbital fenestra, similar to Dibothrosuchus(Figure 5a). In

Junggarsuchus, the cranio quadrate canal is enclosed by the exoccipital and quadrate on the lateral to the foramen magnum (Figure 7a).

Both quadrates of Dibothrosuchus are preserved, though the scans we have are missing the midsection and the articular heads of the quadrate have been crushed against the articulars (Figure 12d). The quadrate of Dibothrosuchus has a posteriorly projected ventral body and possesses an elongate dorsomedial process the overlays the ascending posterior process of the pterygoid, neither of which are seen in Junggarsuchus and

Sphenosuchus (Walker, 1990) (Figure 10a,c). The dorsal process of the quadrate is also more anteroposteriorly expanded in Dibothrosuchus than Junggarsuchus, which lacks the

“T” shape of the dorsal portion of the bone in lateral view. The quadrate does not contact the laterosphenoid at all, unlike in Junggarsuchus, where a narrow anterodorsal process approaches the laterosphenoid, however, the dorsal overhang of the mastoid antrum and cranial nerve 5 (trigeminal) are comparable to Junggarsuchus..Ventrally, the pterygoid ramus of the quadrate is broad and overlaps the anteromedial surface of the quadrate ramus of the pterygoid as opposed to the lack of overlap and tight suturing seen in

Junggarsuchus.

The medial curve for the otic aperture is less medially concave than in

Macelognathus, and more similar to Junggarsuchus. Dibothrosuchus appears to have only one quadrate fenestra unlike Junggarsuchus and basal crocodyliforms which have 2 or more. The body of the quadrate is solid, and not as pneumatic as in crocodyliforms,

Junggarsuchus, and Macelognathus. The anterior concavity and crest on the ventral portion of the quadrate are not observed in other sphenosuchians. There is no contact between the quadrate and exoccipital in Dibothrosuchus like Junggarsuchus and

Protosuchus (Figure 7b). The pterygoid ramus of the quadrate is similar in ventral view in both taxa, but the quadrate ramus of the pterygoid is not as medially expansive in

Dibothrosuchus as in Junggarsuchus. Dibothrosuchus lacks the posterior medial contact with the pterygoid and posterior dorsal projection of the otoccipital or quadrate relative to the quadrate condyle seen in Junggarsuchus

The quadratojugal in Junggarsuchus is preserved only on the left side and is poorly preserved (Figure 4a, 6a, 7a). Extensive breakage and the unusual shape of the quadratojugal and jugal—which limits comparison with other crocodylomorphs—do not allow definitive determination of their contact below the infratemporal fenestra the contact with the quadrate may correspond to a vertical crack just lateral to the fenestra in the ventral part of the quadrate (Figure 7a). Assuming this to be the contact, the quadratojugal is a thin, antero-ventro-medially convex bone. Its ventral edge is obscured by fractures, and may be underlain anteriorly by a portion of the jugal. The quadratojugal extends anterodorsally from the quadrate articulation, in which it does not appear to participate. About one-third of the way across the infratemporal fenestra the quadratojugal curves dorsally, and the jugal appears to have articulated here. A large

55 piece of bone anterior to a break may be part of the quadratojugal, as it has a depression anteroventrally on its lateral surface that would have articulated laterally with the jugal.

However, this posterior lateral piece of bone appears continuous with the jugal and laterally overlaps the quadratojugal (Figure 4a). Dorsally, the quadratojugal may continue along the anterior edge of the quadrate to reach the squamosal. A narrow ribbon of bone is visible covering most of the lateral surface of the dorsal part of the quadrate on the left side, and meeting the squamosal. If it is the quadratojugal this dorsal portion contrasts with the lack of a dorsal continuation in Hesperosuchus (CM 29894), and, as interpreted by Clark et al. (2000) Sphenosuchus and Pseudhesperosuchus, but is similar to Terrestrisuchus (Crush, 1984) and basal crocodyliforms such as Protosuchus.

Dibothrosuchus appears to have small slivers of bone that ascend dorsally along the lateral margins of quadrate, but do not appear to reach the squamosal.

Very little of the quadratojugal is preserved in IVPP V 7906. Two small slivers of bone on the lateral surfaces of the two dorsal ascending process are observable in posterior view. A slightly larger portion of the right quadratojugal is preserved on the lateral surface of the descending process of the quadrate (Figure 10b, 12d). Overall, the general placement of the quadratojugal appears similar to that seen in Junggarsuchus and

Sphenosuchus. However, it is unclear if the dorsal expansion of the quadratojugal is as extensive as it is in other sphenosuchians.

The supraoccipital of Junggarsuchus occupies the dorsomedial area of the occiput, from its ventral contact with the otoccipital dorsal to the foramen magnum to the dorsal edge of the occiput. (Figure 7a). Except for the small exposure of the parietal along the midline extending into a midline notch in the supraoccipital, the supraoccipital

56 is roughly a square plate on the occiput, of approximately equal height and width, although it widens on the left side towards the ventral edge. This is more comparable to the wider supraoccipitals seen in Almadasuchus and crocodyliforms. Ventrolaterally it is bounded by the otoccipital, which excludes it from the dorsal margin of the foramen magnum. This region is damaged, and has been re-attached after separation of the left and right sides, but a dorsoventrally broad medial process of the otoccipital is preserved on the right side below the straight ventral edge of the supraoccipital, and this process is broken on the left side, so that a dorsal extension of the foramen magnum is due to this break and the ventral dislocation of the left side of the braincase relative to the right. The supraoccipital is fully separated from the foramen magnum by the contact between the otoccipitals (exoccipitals), which is the condition seen in Protosuchus and other crocodyliforms although it is narrower in Junggarsuchus. The supraoccipital approaches the medial margin of the small posterior temporal fenestra, similar to Dibothrosuchus.

However, the posterior temporal fenestra does approach the supraoccipital more so than in Dibothrosuchus. Based on CT scans 61-80 the dorsal tympanic recess did not extend through the supraoccipital as it does in most crocodyliforms. In the anteromedial portion of the supraoccipital the space for the third semicircular canal is enclosed, where the bone contacts the prootic.

The supraoccipital of Dibothrosuchus is more pentagonal in shape than the square bone in Junggarsuchus (Figure 7b). The parietal sends three ventral extensions down the occipital surface of the supraoccipital, two on each side and a short one along the midline. The supraoccipital contacts the otoccipitals ventrally, and a ventral midline projection separates the otoccipitals and forms the dorsal border of the foramen magnum,

57 which is similar to some other sphenosuchians, like Sphenosuchus. The posterior surface of the supraoccipital is more concave than in Junggarsuchus. The anterior process of the supraoccipital, that ventrally contacts the otoccipital, is expanded anteromedially. The supraoccipital is a solid bone in CT scans and the dorsal tympanic recesses do not connect through the body of the bone as hypothesized by Wu and Chatterjee (1993:69).

The exoccipital is apparently fused with the opisthotic to form an otoccipital and forms the paroccipital process as well as part of the lateral braincase wall (Figure 4a, 7a,

8, 9b). The wing-shaped, dorsoventrally broad paroccipital process is fully preserved only on the left side of the skull. Its distal end is expanded and the lateral surface is concave, forming a broad dorsoventral groove along with the squamosal, which overlies it anteriorly. This unusual groove may have been the site of origin if the M. depressor mandibulae, although in living amniotes it does not arise from the lateral surface of the paroccipital process (Diogo, 2008). However, alternative interpretations suggest that groove may be related to the ear muscles, potentially related to the groove on the lateral surface of the squamosal, which serves as another ear muscle attachment site. The ventral edge of the process is slightly concave, and a poorly preserved ventrolateral projection of the otoccipital is present ventral to the paroccipital process (Figure 8a). A small part of this projection is preserved near the lateral end of the paroccipital process, where it encloses a foramen that was identified as for the internal carotid artery by Clark et al. (2004). However, given its lateral position we infer that the foramen for the carotid foramen is not preserved and that this foramen is for the cranioquadrate canal (Figure 7a).

The otoccipital, squamosal and quadrate meet lateral to this canal. Another part of it is preserved on the right side of the occiput between the basioccipital and basisphenoid

58 lateral to the basioccipital and dorsal to the basisphenoid (Figure 7a). Other pieces of this process were removed during preparation. The otoccipital may contact the ventral edge of the occipital portion of the parietal very briefly lateral to the temporo-orbital foramen, separating the squamosal from the supraoccipital, and form part of the ventral edge of this mediolaterally ovoid foramen. The contact between the otoccipitals on the midline is dislocated, and the left side is moved ventrally about 2 mm relative to the right.

The contact between the otoccipital and basioccipital ventrolateral to the foramen magnum is obscured on the right side due to breakage, and on the left a suture or crack is evident ventrally on the lateral edge of the occipital condyle but not dorsally. If it is a suture, the otoccipital contributed a small anterior portion to the lateral surface of the condyle, however, in CT scan these inferred sutures are difficult to interpret and could be cracks. The occipital condyle is reconstructed as being all basioccipital, though some of the lateral dorsal aspects may be parts of the basioccipital (Figure 7a). The hypoglossal foramina are not evident lateral to the foramen magnum. On the right side a smooth, a flat laterally facing surface on the ventral part of the otoccipital lateral to the basioccipital formed the medial border of a passage through the otoccipital. This opening may be for the inner carotid artery, but no comparable structure is seen on the more complete left side. It may be comparable to the foramen on the left side interpreted as for the cranioquadrate canal, although the left opening lies slightly further laterally There appears to be a finished surface on the medial surface of the fragment with the foramen on the left, which may indicated the presence of two foramina. No basisphenoid otoccipital suture is evident, though what may be the descending process of the right otoccipital may contact the basisphenoid. The exact nature of this rectangular section of

59 bone is unclear and it may be basioccipital (Figure 7a, 8d) In Junggarsuchus a laterally and ventrally extensive otoccipital diverticulum is present. While not figured here, the diverticulum is shown to be more extensive in the otoccipital in Junggarsuchus than more basal sphenosuchians. It also appears to be more widespread than the diverticulum in Dibothrosuchus, though the nature of the scan limits exact determination of relative space size.

The cranioquadrate canal appears similar in position to Almadasuchus, though more ventrally directed, however the latter is possibly due to breakage. The otoccipital, quadrate and squamosal contact lateral to the canal, similar to Almadasuchus and crocodyliforms. However, due to breakage, whether this contact is broad is unclear. This contact is not seen in Dibothrosuchus or Sphenosuchus (Walker, 1990) or more basal crocodylomorphs. The ventrolateral contact of the otoccipital with the quadrate appears to be broad as in Almadasuchus (Pol et al, 2013), but is incomplete. The posterior tympanic recess is bordered anteriorly by the prootic and posteriorly by the otoccipital and is angled sub-vertically as in Almadasuchus (Pol et al, 2013) (Figure 8d, 9b). In

Dibothrosuchus (Wu and Chatterjee, 1993) this recess is set in a deep depression at the same level of the mastoid antrum, which is similar to neosuchians. Junggarsuchus may have a rhomboidal recess on the anterior face of the otoccipital, medial to the paroccipital process, however it is located more posteriorly and its diagnosis is uncertain. The subscapular buttress has a dorsally convex dorsal lip and medial to this is dorsal lamina of the otic capsule. Posterior to this is an opening that may be the metotic fissure and posteriorly a groove for the semicircular canal. This general anatomy appears similar in

Dibothrosuchus. The metotic fissure, for the vagus nerve and the opening for XI appear

60 the to be directed mediolaterally as in Dibothrosuchus and unlike crocodyliforms (Figure

8a, 9b). Within the otic region the fenestra ovalis overlays the narrow crista interfenestralis, which separates the ventral fenestra ovalis from the dorsal vestibule. This region of the opisthotic continues anterior to an enlarged region that contacts the prootic crista (Figure 9a).

The otoccipitals of Dibothrosuchus are well preserved on both sides of the skull.

Dorsoventrally the otoccipitals are shorter than in Junggarsuchus and terminate dorsal to the most ventral extension of the occipital condyle (Figures 7a, 10). The medial borders of the foramen magnum are medially concave around the foramen magnum. Dorsally the otoccipitals are separated by a ventral projection of the supraoccipital, unlike in

Junggarsuchus and crocodyliforms. Along the posterior midline of the edge of the occipital face of the otoccipitals there is a groove that runs from the lateral to the medial edge. The paroccipital processes are not as wing shaped as in Junggarsuchus, but instead have symmetrical dorsal and ventral edges, giving the paroccipital process a broad, spade-shaped lateral edge. They are laterally concave, not convex as in Junggarsuchus and more like crocodyliforms. The ventral edge of the otoccipital contact with the basioccipital has more extensive contact with the lateral edges of the condyle. In prior reconstructions (Wu and Chatterjee, 1993), the otoccipitals have been reported to form the dorsal portion of the basioccipital and the posterior floor of the braincase. Our CT scans show that these portions of the otoccipital have unclear sutures with the basioccipital, and we reconstruct the otoccipitals as less extensive than interpreted by Wu and Chatterjee (1993). In this interpretation the basioccipital forms most of the posterior floor of the braincase. The anteroventral process of the otoccipital on the occiput has two

61 foramina, a smaller anterior one and a slightly larger posterior one, both for branches of cranial nerve XII (7b). The dorsal contact of the otoccipital with the squamosal and supraoccipital are straight, unlike the ventral depression above the foramen magnum in

Junggarsuchus. Like Junggarsuchus the posterior tympanic recess is bordered posteriorly by the otoccipital as in Dibothrosuchus (Figure 10a, c). Ventral to this is the opening for cranial nerves IX-XI and anterior to it is the fenestra ovalis. The exit for cranial nerve XII and the vagus nerve are ventrolateral and similar in position to Sphenosuchus but not through a single opening like Almadasuchus (Walker, 1990; Pol et al, 2013) and crocodyliforms. The basisphenoid does not contact the otoccipital in Dibothrosuchus unlike to Protosuchus and potentially unlike Junggarsuchus.

The otic region in Junggarsuchus is best exposed on the right side (Figure 8b, 9b).

The otoccipital forms the posterior portion of the otic region, and the horizontal crista interfenestralis is preserved extending anteriorly to separate the fenestra ovalis dorsally from the fenestra pseudorotundum ventrally, which is similar to Dibothrosuchus. The prootic and otoccipital contact dorsal to the fenestra ovalis, and the contact underlies what may be a small dorsal tympanic recess. Posterior to the subscapular process the otoccipital forms a broad ventrolateral groove leading to the passage through the otoccipital preserved on the right side, presumably for the vagus and accompanying posterior cranial nerves. Ventral to the subcapsular process a large opening leads to a sinus in the basioccipital (Figure 9a). This condition is similar to other sphenosuchians like Dibothrosuchus and Almadasuchus. The endocranial pneumaticity in Junggarsuchus is extensive laterally and ventrally in Junggarsuchus, though diverticulae do not extend above the braincase dorsally as in Crocodyliformes. This condition is seen in most

62 sphenosuchians. However, Dibothrosuchus has very pneumatized prootics (Figure 10), which nearly reach the dorsal edge of the parietal. This dorsal reach of pneumatic space is unknown in other sphenosuchian. Dibothrosuchus lacks the extensive diverticulae in the parietal and quadrate. These dorsally expanded diverticulae are known in protosuchians and other crocodyliforms (Young et al, 2019).

In Junggarsuchus the basioccipital is roughly circular in posterior view (Figure

7a), except at its dorsal margin where it becomes concave along the ventral margin of the foramen magnum and along its ventral margin where it is slightly arched dorsally along the midline. It extends anteriorly into the braincase where it forms the posterior part of the floor. Anterior to the foramen magnum, the dorsal surface of the basioccipital slopes slightly ventrally. Below the occipital condyle, the posterior surface of the basioccipital is gently concave. The basioccipital diverticulum is extensively ventrally expanded in

Junggarsuchus (Figure 8a,b). Internal to this concavity, as exposed by a break on the left side, the basioccipital recess occupies a large space, larger than in Sphenosuchus

(Walker, 1990). However, as noted in Leardi et al, 2019 (in prep) the basioccipital recess in Junggarsuchus exists as a reduced structure and only exists as the ‘basioccipital recess in the loose sense. They are interpreted as not homologous to the basioccipital recesses seen in other crocodylomorphs and crocodyliforms, but similar to the general recess seen in Postosuchus. Dorsolaterally, the contact with the otoccipital is difficult to identify.

However, it appears that the dorsolateral edge of the basioccipital contacts the medioventral surface of the posterior end of the otoccipital, which form the posterior lateral walls of the braincase. Ventrally, the basioccipital has a relatively flat posterior surface, and pocketing and tubera are lacking unlike other sphenosuchians like

Dibothrosuchus and Sphenosuchus (Figure 6a). The orientation of the ventral part of the basioccipital depends on the correct orientation of the skull; if the ventral edge of neurocranial cavity is horizontal then the basioccipital is nearly perpendicular to it, if the sagittal crest of the parietal is horizontal then the basioccipital faces posteroventrally.

The lateral eustachian opening may be present where the basioccipital meets the otoccipital and basisphenoid ventrolaterally, forming a mediolaterally elongate slit. The eustachian tubes appear to be enclosed between the basisphenoid and the basioccipital.

The opening is not obvious but appears to be located posterior to the basipterygoid process (Figure 8f).

The basioccipital of Dibothrosuchus is similar to the basioccipital found in

Sphenosuchus. Unlike Junggarsuchus, there are more obvious basioccipital tubera present ventral to the occipital condyle and two ventrally opening foramina for eustachian tubes (Figure 7b, 10c,d). These foramina are oval and separated along the midline. They are located in a ventral concavity and the foramina are fully enclosed by the basioccipitals. The anterior portion of the basioccipitals vertical surface is overlain by two posterior projections of the basisphenoid. Lateral to the occipital condyle the otoccipitals limit the lateral extent of the basioccipital relative to Junggarsuchus. The dorsal concavity of the foramen magnum in Dibothrosuchus is also less pronounced than it is in Junggarsuchus. Dibothrosuchus possesses the basioccipital recesses present in most other sphenosuchians, including Macelognathus (Leardi et al, 2017),

Junggarsuchus and crocodyliforms like Protosuchus. The basioccipital appears to contact the basisphenoid posteriorly as in crocodyliforms like Protosuchus

The basisphenoid in Junggarsuchus is expanded ventrally compared with other

‘sphenosuchians’ (Figure 8b), similar to some basal crocodyliforms such as Protosuchus and to the therizinosaurid theropod Erlikosaurus (Lautenschlager et al., 2014). It is therefore visible on the anterovental part of the occipital surface, where the basioccipital overlies it posteromedially. The body of the basisphenoid houses a large pneumatic cavity, comparable in size to the neurocranial cavity. The ventral midline portion of the basisphenoid was separated from the skull, and showed that the ventral surface of the basisphenoid is convex ventrally. A robust lateral process from the ventral surface of the basisphenoid just anterior to its contact with the otoccipital on the right side is presumably the basipterygoid process, although it is in an unusually posterior position and laterally directed (Figure 6a, 8d). The process is solid, with a broad, flat, circular surface that faces laterally and only slightly ventrally. Medial to it the basisphenoid is robust and solid nearly to the midline. The basisphenoid is not preserved at its midline contact with the basioccipital, but a notch in the ventral edge of the basioccipital and another on the posterior edge of the midline fragment are the edges of a basisphenoid recess (“middle Eustachian opening”). The eustachian tubes are enclosed by the basisphenoid and basioccipital, the opening potentially posterior to the basipterygoid process. The lateral surface of this recess is preserved, and it is separated from the

‘basioccipital recess’ posteriorly and from the larger pneumatic cavity anteriorly, but may have been connected along the midline. The ventral surface of the basisphenoid medial to the process is very slightly concave ventrally, lateral to the midline expansion. The basisphenoid appears to contact the otoccipital posteriorly along the ventromedial edge of the otoccipitals ventral process though a suture is not evident (Figure 8b). The identity of

65 the rectangular portion of bone ventral to the otoccipital on the right makes diagnosis of this contact difficult. The medial surface of the solid portion of the basisphenoid forms a medially open concavity of unknown function along the lateral wall of this recess. On the right side the anterior extent of the basisphenoid and its relationship with the pterygoid is unclear. Post mortem crushing has fragmented the anterior portion of the basisphenoid and potential overlap from posterior process of the pterygoid onto the basisphenoid is unclear. Elements in this region of the skull have also likely been shifted during fossilization as the quadrate ramus of the pterygoid appears nearly in level with the ventral portion of the basisphenoid. If the medial expansion of the pterygoid ramus of the quadrate continues to ascend dorsally and medially, and if complete, it may contact the basisphenoid. Anteriorly, at the contact between the laterosphenoid and basisphenoid there is an opening which may correspond to the opening for the palatine artery seen in

Almadasuchus (Figure 8a).

The large pneumatic cavity within the basisphenoid is open laterally, and in ventral view would be covered broadly by the quadrate ramus of the pterygoid in the latter were complete. Anteriorly the recess is very deep, but it is unclear which parts of the anterior wall are formed by the basisphenoid and which by the pterygoid. The basisphenoid forms more than half of the neurocranial floor anteriorly, and the hypophyseal fossa is poorly preserved at its anterior end. Lateral to the large dorsal midline opening of the pituitary fossa, there are two small foramina in the dorsal surface of the portion of the basisphenoid that forms the anterior portion of the bottom of the braincase. These two opening continue through the body of the basisphenoid, but are unfortunately lost with the posterior cracking of the bone. These foramina are most likely

66 the exits of the inner carotid arteries, which may exit through a medial opening on the otoccipital identified earlier (Figure 8a, e) A delicate strut traverses the pneumatic space within the basisphenoid along the midline, extending posteroventrally from below the hypophyseal fossa and dividing the space into dorsal and ventral parts that are confluent laterally. The strut houses the pituitary fossa, which is expanded anteriorly. -

The basisphenoid of Dibothrosuchus (Figure 10a,b,c) is not as dorsoventrally tall as Junggarsuchus, and while pneumatic, lacks the extensive pneumaticity seen in

Junggarsuchus. As in Junggarsuchus the basioccipital overlays the basisphenoid, but the posterior two lateral processes of the basisphenoid seen here are not observed in

Junggarsuchus. The basisphenoid is enlarged relative to the basioccipital, though not as much as it is in Junggarsuchus and Protosuchus, in which the expansion is more dorsoventral. The basipterygoids of Dibothrosuchus are radically different than the posterior, laterally directed knobs seen in Junggarsuchus. The basipterygoid process are enlarged, bulbous and pyramidal and extend substantially ventrally (Figure 10a,b,c).

These two processes are very pneumatized and anteriorly overlain by the quadrate ramus of the pterygoid. The contact with the quadrate ramus does not seem to contact the basipterygoid process in Junggarsuchus, however this depends on the anterior extent of the basipterygoid.

The anterior break in the basisphenoid reveals the foramina for the internal carotid artery on the anterodorsal surface, above the openings for the pre carotid recess (Leardi et al, 2019 in progress), which had previously been interpreted as the pituitary fossa (Wu and Chatterjee, 1993) (Figure 10b). It is difficult to compare these passages to

Junggarsuchus, which is missing much of the medial portion of the basisphenoid. The

67 basisphenoid-exoccipital suture is absent in Dibothrosuchus, though the exoccipital ventrally approaches the posterior extent of the basisphenoid. This suture is possibly present in Junggarsuchus on the posterior surface of the braincase. While the basisphenoid itself is more extensively expanded in Junggarsuchus the basisphenoid recess in Dibothrosuchus appears more expanded than it is in Junggarsuchus, though this region is incomplete.

The prootic is visible within the supratemporal fossa between the parietal and quadrate (Figure 5a, 8a,b,f). It faces anterodorsally, and has a gently concave surface.

Posterodorsally it forms the ventral edge of the temporo-orbital artery foramen. The canal for this vasculature is lateral to the dorsal most extent of the braincase sinuses lateral to the inner ear and endosseous labyrinth. It appears to continue to just anterior to the posterior temporal fenestra (Figure 8f). The prootic diverticulum is the most dorsally extensive of the diverticuluae of the braincase. The prootic extends anteriorly to meet the laterosphenoid midway in the supratemporal fenestra, bordering the trigeminal opening posteriorly, and dorsally. The trigeminal opening is directed antero-laterally, and the prootic forms much of the dorsal, ventral and posterior boarder (Figure 8a, b). Ventral to the dorsal head of the quadrate it encloses the opening of the mastoid antrum extending dorsomedially, and may border a small superior tympanic recess. The mastoid antrum is anteroposteriorly long and more oval than the openings in Almadasuchus, though not as enlarged as in Dibothrosuchus. The facial antrum is anterior to and enlarged and ovular relative to the mastoid antrum, and opens ventrally. Both foramina for cranial nerves VII and VIII are preserved on the anteroventral face of the prootic, with the foramin for VII located on the dorsal part of the crista prootica. The crista prootica is extremely robust,

68 extending vertically ventral to the anterior end of the otic recess. The crista descends ventrally to contact the basisphenoid and the anterior portion of the basioccipital and does not appear to widely contact the laterosphenoid. A fenestra appears to open on the ventral surface of the crista prootic (Figure 8d, Figure 9b, c). The fenestra ovalis and fenestra rotunda appear to be preserved on the ventromedial surfaces of the prootic. The endosseous labyrinths are preserved on both sides within the prootic (Figure 9b,c,d).

They are medial to the dorsally expanded extent of the open space in the prootic diverticulum and the openings are partially exposed on the lateral side of the prootic..

The depression for the posterior tympanic recess is located posterior to the fenestra ovalis on the anterior surface of the paroccipital process of the otoccipital, similar to

Almadasuchus and Sphenosuchus. Opening though the medial surface of the prootic, the semicircular canals in the prootic are well preserved on both sides of the skull (Figure

9b,c,d). Young, 2019 have generated endocasts of the arteries and openings of the skull of Junggarsuchus and were able to fully reconstruct all three semicircular canals. These canals are clear in CT sections and the entire length of canals can be followed the first located within the medial body of the prootic, the next following and extending partially through the otoccipital and the final dorsal canal extending from the prootic through the anterior end of the supraoccipital.

Both of the prootic in Junggarsuchus preserve an unusual anteriorly opened foramen on the dorso-lateral surface of the prootic (Figure 8c). This small foramen is the beginning of a canal that continues through the dorsal body of the prootic. While it appears that this canal may be entirely enclosed by the prootic, the lateral edge of this canal may be formed by the medial extension of the quadrate. While the contact between

69 the prootic and quadrate is clear ventrally, with the posterior head of the quadrate extending posteriorly ventrally convexly, it is less clear dorsally, due to a crack. It appears that a thin sheet of the quadrate overlays the lateral dorsal edge of the prootic and forms the lateral borders of the canal (Figure 9c).

The extremely pneumatic and expanded prootic of Dibothrosuchus is unique among sphenosuchians. The prootic is bulbous and dorsally convex and is dorsal to the suture between the parietal and prootic (Figure 10). The prootic is far more pneumatic than observed in other crocodylomorphs and the pneumatic spaces are divided into three regions by thin sheets of bone that expand dorsolaterally from the front of the expanded region to the midsection and back. The prootic forms the ventral border of the foramen for the post temporal artery, as seen in other crocodylomorphs (Walker, 1990). The exit of cranial nerve VII and the fenestra ovalis are exposed on the lateral surface of the prootic (Figure 10a). It encloses the superior tympanic recess posteriorly which is larger in Dibothrosuchus than the oval recess in Junggarsuchus. The anterior most enlarged aspect of the prootic is interpreted here as corresponding to the opening for the facial antra. On the right side, an opening appears present, though this may be due to the incomplete nature of the prootic. On the left side this enlarged region is floored for by a thin sheet of bone, though a narrow slit oriented lateromedially bisects this region of the bone (Figure 10a,c). The mastoid antrum is about the same size as the posterior tympanic recess and larger than the facial antra and is far wider mediolaterally than the openings in

Junggarsuchus, and unlike Junggarsuchus the mastoid antrum is divided along the midline by a thin sheet of bone. The posterior tympanic recess is anteriorly bounded by the prootic, but posteriorly bounded by the otoccipital. The depression for this recess is

70 similar to Maceloganthus and Protosuchus, in that it deeply penetrates the otoccipital and prootic, unlike Junggarsuchus and other sphenosuchians. The prootic is anteroposteriorly longer than Junggarsuchus and has a robust crista prootica, ventrally descending pillars of the prootic that contact the lateral surface of the laterosphenoid and posterodorsal edge of the basisphenoid comparable to that found in Junggarsuchus (Figure 10a,c).

The laterosphenoid is incompletely preserved ventral to the parietal and frontal within the supratemporal fenestra (Figure 5a, 8f). The right element is more complete but is fragmented. As preserved, it is a largely flat, vertical bone that makes up the anterolateral sidewall of the braincase. Few features are evident, other than a round, undivided trigeminal opening posteriorly and the capitate process anterodorsolaterally, visible within the orbit. A ribbon of bone extends anteroventrally from the anterior end of the laterosphenoid on the left side lateral to the hypophyseal fossa (Figure 8a, b, e), extending farther ventrally than is usual for this bone. There is a small anteriorly directed opening on the ventral posterior process of the bone, similar to those seen in

Macelognathus (Leardi et al, 2017). On the right side, the anterior process of the laterosphenoid has a lateroventrally directed foramen, likely for cranial nerve 4 (Figure

8b). We do not find evidence of an epipteryygoid (Holliday and Witmer, 2009). The laterosphenoids may meet anteroventrally but it is unclear on the CT scans. The lack of the anterior portion of the laterosphenoid also does not allow for inference of the anterior extent of the dural envelope chamber. Where the laterosphenoid meets with the prootic, enclosing the trigeminal opening, there is a recessed space divided between the prootic and laterosphenoid through which the trigeminal runs that is likely the trigeminal recess reported in Dibothrosuchus and Almadasuchus.

The laterosphenoid of Dibothrosuchus is better preserved in the specimens than is observed in the CT scans. The laterosphenoid is extensive, though in CT scans the posterior walls of the laterosphenoid are poorly preserved. However, the laterosphenoid is better preserved anteriorly and the capitate process projects dorsally and contacts the postorbital (Figure 4b, 5b). If the anterior left side of the laterosphenoid has been accurately reconstructed (not part of the postorbital) than this joint appears to be present.

The laterosphenoids meet ventrally and forms the anterior portion of the ventral surface of the braincase, which is not preserved in Junggarsuchus. The laterosphenoid is longer in Dibothrosuchus and meets the postorbital at the postorbital frontal suture. The laterosphenoid in Junggarsuchus extends anterior to this contact, but the bone relative to the rest of the braincase is shorter. The laterosphenoid is also more extensive posteriorly, where the descending process of the prootic contacts the posterior end of the latersophenoid. The anterior border of the trigeminal nerve foramina is circular, as in

Junggarsuchus, but different from the bilobate opening of Sphenosuchus. The trigeminal is enclosed by the laterosphenoid. The trigeminal recess consists of the most dorsal opening in the laterosphenoid and is larger and complex. The anterior end bifurcates in a

T-shape and the dorsal cavity opens to the middle ear space. The ventral opening is much smaller (Pol et al., 2019). The trigeminal recess, has been considered an autapomorphy of

Dibothrosuchus, but is also present in Almadasuchus, Kayentasuchus Maceloganthus and is present less extensively than in Almadasuchus and in Junggarsuchus. (Leardi et al,

2017; Clark and Sues, 2002; Pol et al, 2013).

The paired vomers extend along the midline from the palatal shelves of the maxilla anterior to the pterygoids (Figure 4a, 6a). The vomers divide the elongate

72 choanae, and the vomers are rod-like between the anterior half of the choanae. The portion of the rod like vomer anterior to the choana sits on the dorsal side of the posterior extent of the palatal portion of the maxilla. The vomers expand posteriorly into a rhomboid shelf, and contact the anterior medial edges of the palatines. Potential contact with the pterygoid is not preserved, though is possible, as seen in the anterior extension of the pterygoid of Dibothrosuchus. The expanded portion of the vomer is separated down the midline by a tall ventral septum. On the ventral surface of the vomer, lateral to the septum on both sides is a deep elongate groove, that runs the entire length of the expanded process. This central bony wall is far more developed than the septum seen in

Dibothrosuchus. The dorsal side of the expanded region of the vomer is a smooth surface that slopes gently ventrolaterally. The dorsal midline is marked by a deep groove that runs nearly the entire length of the bone (Figure 6a). The vomers rise slightly around the lateral margins of this canal. The same structure is seen in Dibothrosuchus, and houses a narrow anterior process of the pterygoids. It is likely that a similar process of the pterygoid existed in Junggarsuchus, but was not preserved.

The vomers of Dibothrosuchus are similarly paired, rod-like bones that divide the choana and expand to contact the medial edges of the palatines (Figure 6b). In

Dibothrosuchus the vomer is not as well preserved, and is missing the anterior portion dorsal to the palatal shelves of the maxilla. However, the contact with the pterygoid is preserved. The pterygoid extends as a thin anterior process in the dorsal groove between the two vomers, seen in Junggarsuchus. Relative to the width of the palate and skull, the expanded region of the vomer of Dibothrosuchus is smaller than that of Junggarsuchus.

The tall, ventral septum along the midline is also far less developed than the one in

Junggarsuchus, and the depressions on the ventral surface of the vomer of

Dibothrosuchus are very faint. The portion that seperates the choana is taller, and less rod like than that of Junggarsuchus.

The dorsal surface of the palatine is visible through the antorbital fenestra, indicating that it may have been displaced dorsally somewhat (Figure 4a). The medial part of its dorsal surface is convex and smooth, and as in Sphenosuchus (Walker, 1990) there are no apparent depressions or pockets in the dorsal surface, but is not completely preserve dorsally. Any midline contact between the palatines is not preserved, although they come close. In ventral view the palatine is triangular in shape, expanding posterolaterally, where it borders the ventral edge of the expanded portion of the vomer

(Figure 6a). The body of the palatine is ventrally concave, forming with the vomer a longitudinal depression roofing a broad passage medial to the posterior end of the choana.

The depression is divided along the midline by a longitudinal septum formed at least partly by the vomers. The septum is separated from the palatine by a longitudinal crack or suture, suggesting it is formed mainly by an anterior projection of pterygoid that projects into a groove on the dorsal surface of the vomer. Laterally the palatine narrows as it reaches the maxilla. The anterior edge of the palatine is expanded ventrally where it forms the posterior end of the choana, separating the choana from the depression. This expansion forms the anterior edge of a pocket on the ventral surface of the palatine that is continuous anteromedially with the depression. This pocket corresponds in position to a small opening on the palatine of Dibothrosuchus (Wu and Chatterjeee, 1993, figure 2b) and Sphenosuchus (Walker, 1990, figures 3a, 10b) Anterior to the posterior end of the choana the palatine descends ventrally to separate the choana and the medial depression.

The medial section of the left posterior bar of the palatine has been mistakenly moved into the space between the two palatine bars, where it contacts the anterior process of the pterygoid (Figure 6a).

Posteriorly the palatine is damaged, and its contacts with the pterygoid are unclear. A strut is preserved on each side extending posteriorly from the maxillary contact to the preserved portions of the pterygoid. A fragment of bone medial to the level of the ectopterygoid on the right side may be the posterior end of the palatine. As preserved the palatine is separated from most of the pterygoid to form an elongate fenestra paralleling the suborbital fenestra, but this unusual feature may be due to damage to the palatine. However, some parts of the palatine have finished surfaces, and this additional medial fenestra may be real (Figure 6a). The palatine forms the medial and anterior borders of the suborbital fenestra, but its precise contribution to the medial edge is not clear. The contact with the pterygoid near the midline is also obscure, and the pterygoid may contribute to the midline septum. Laterally the palatine extends very briefly posteriorly along the lateral border of the suborbital fenestra, and extends dorsally to contact the jugal and lacrimal where they contact. The anteromedial end of the palatine is obscure, but an apparent contact on the right side suggests a slender process of the palatine may have extended anteriorly a great distance along the medial surface of the maxilla, to opposite the 6th maxillary tooth.

The palatines of Dibothrosuchus are more completely preserved and less dorsally arched than those in Junggarsuchus (Figure 6b). A thin portion of bone overlaps the lateral dorsal surface of the vomer, but the dorsal contact of the palatines is not preserved.

Like in Juggarsuchus, the ventral surface of the anterior medial surface of the palatine

75 has a concavity posterior to the choana, though it is shallower than the depression seen in

Junggarsuchus. Anteriorly the lateral surface of the palatines extends a long way along the medial surface of the maxilla, to the 9th maxillary tooth. The lateral posterior projection is not present, if preserved. The palatines are separated ventrally by an anterior projection of the pterygoid. The palatines narrow posteriorly on either side of the posteriorly expanding pterygoid. No additional palatal fenestra are preserved medial to the sub orbital fenestra, though laterally the palatines have a concavity on the medial border of the suborbital fenestra. This is similar to the rod shaped posterior process of the palatines seen in Junggarsuchus and suggests that existing medial border of the suborbital fenestra in Junggarsuchus is accurate, but that what has been interpreted as an additional palatal fenestra may actually be covered by a thin sheet of the palatine that extended from the lateral edge of the anterior process of the pterygoid to the preserved ridge like processes of the palatines. The palatines form the medial and the posterior portions of the medial and lateral borders of the choana.

The pterygoid is poorly preserved in Junggarsuchus, but its remains indicate that it had an unusually broad, long quadrate ramus for sphenosuchians. The posterior part of the quadrate ramus is preserved on the left side, and pieces of the anterior part are preserved on the right side (Figure 4a, 6a). The posterior part extends posteriorly to just dorsal to the quadrate’s mandibular condyles, where it appears that it is firmly attached to the quadrate, like crocodyliforms including Protosuchus, and to the ventrolateral end of the otoccipital (Figure 6a, 8d). This small rectangular piece of bone posterior to the medial edge of the quadrate condyle may not be part of the pterygoid but a dorso-lateral fragment of the otoccipital (Figure 7a). This portion of the ramus forms a thin,

76 anteroventrolaterally convex shell that bordered the basisphenoid pneumatic cavity laterally. Dorsally within this shell a longitudinal shelf extends medially, but it is broken and its medial extent is unclear. The surface for the basipterygoid articulation is not preserved. The preserved fragments of the anterior part of the quadrate ramus indicate that this portion was also dorsoventrally broad, and contributed to the anterolateral wall of the basisphenoid pneumatic space. The transverse flanges are both preserved separate from the main body of the pterygoid, adhering to the mandible laterally and with the posterior edge of the ectopterygoid, and are robust and solid. The flange is oriented subvertically as preserved, extending slightly anterodorsally. Posteriorly on the palatal midline the pterygoids are poorly preserved, and ventrally form a gently concave plate with a low midline ridge. They continue anteriorly along the midline as a long, slender process, and the smooth lateral surface of this process might support the interpretation that the fenestra lateral to them is real. However, the bone is thin here and may have been broken, if not, this would be the only sphenosuchian known to have this accessory palatal fenestra. The midline process rises anteriorly where it would meet the palatine, and is dorsal in position to the rod formed by the pterygoid and palatine laterally. The ventral surface of this midline anterior process has a small ventrally directed notch. The anterior extent of the pterygoid is unclear, but it may contribute to the midline septum between the palatine bodies, although there is a gap between the posterior end of the septum and the midline process of the pterygoid. Anterolaterally the pterygoid extends along a rod- like structure formed with the palatine, but its extent is limited to the posterior fourth of the palatine rods. Posterior to the concave midline portion, two pieces of the pterygoid are preserved ventrolateral to the portion discussed above (Figure 6a). The portion on the

77 right side is more completely preserved and is ventrally and anteriorly convex.

Posteriorly the bone is concave and may overlap the anterior portion of the basisphenoid.

This portion may be part of the basisphenoid.

The pterygoid is more completely preserved in Dibothrosuchus than it is in

Junggarsuchus (Figure 6b). The pterygoid expands posteriorly, from a narrow anterior process that begins in the posterior groove in the vomers, widens gradually between the palatines, then abruptly widens laterally into the transverse flanges of the pterygoid and the lateral edges articulate to the ectopterygoid. No evidence for an additional medial fenestra in the palate is preserved and the region appears to be solid bone. Along the midline of the pterygoid, there is a shallow groove that expands posteriorly, and the ridges curve laterally along the transverse flange. The quadrate ramus ascends along the posteromedial edge of the quadrate, with which it forms a joint and contacts the anterior end of the basisphenoid, overlapping part of the basipterygoid process (Figure 10). It approaches the prootic dorsally. The ascending process is more extensive dorsally than in other sphenosuchians (Wu and Chatterjee, 1993). It is far different from the pterygoid of

Junggarsuchus. It features broader pterygoid transverse flanges and a dorsally tall posterior quadrate ramus of the pterygoid. The ramus does not extend as far posteriorly as in Junggarsuchus and lacks the pneumatic medial expansion seen in Junggarsuchus as well. The pneumaticity suggested by the medial concavity of the pterygoid of

Junggarsuchus (Figure 8b) is not observed in Dibothrosuchus which has a more sheet like pterygoid. The posterior projections of the pterygoid between the basisphenoid and quadrate is also seen in Sphenosuchus (Walker, 1990) and thalattosuchians.

The small ectopterygoid of Junggarsuchus is completely preserved on both sides.

It contacts the jugal along the longitudinal ridge on the jugal’s medial surface beneath the orbit (Figure 6a). This anterior process is long and expanded, similar to the anterior processes of Protosuchus and early crocodyliforms. In dorsal view the lateral portion of the ectopterygoid is triangular, with a lateral base, and the base extends further anteriorly than posteriorly. The ectopterygoid narrows medially and twists to face posteromedially as it descends the posterior surface of the transverse flange of the pterygoid. The medial portion is also triangular, with a ventral base. It covers the dorsal third of the posterior surface of the flange, and contributes briefly to the dorsal part end of the flange’s lateral surface.

The ectopterygoids of Dibothrosuchus are relatively larger than those of

Junggarsuchus, though only the right one is more completely preserved (Figure 5b, 6b).

The posterior body of the ectopterygoid extensively contacts the lateral edge of the transverse flange of the pterygoid. Like Junggarsuchus, the anterior process that contacts the medial edge of the jugal is longer than the posterior process, but in Dibothrosuchus the anterior processes are reduced in size relative to the posterior body of the ectopterygoid. The angle of contact between the ectopterygoid and pterygoid in

Dibothrosuchus is far gentler, at 35 degrees (Wu and Chatterjee, 1993) than the near vertical contact between the ectopterygoid and pterygoid flange of Junggarsuchus. While the anterior end arches dorsally similar to Sphenosuchus and Junggarsuchus, the posterior end’s contact with the pterygoid is rather odd, and demonstrates a condition not seen in other sphenosuchians. A dorsal process of the ectopterygoid overlays a brief portion of the pterygoid laterally, and the main, elongate posterior body of the

79 ectopterygoid underlays the same transverse process. In Junggarsuchus the contact does not underlay the pterygoid in the same way.

Portions of seven thin scleral ossicles were preserved within the right orbit and later removed. These ossicles are small, square bones which partially overlap one another. They appear to be from the ventral part of the series, and they curve anterodorsally, originally extending from the back of the orbit to the ventrolateral edge of the descending process of the prefrontal.

No scleral ossicles are known from Dibothrosuchus.

Mandible

In Junggarsuchus the dentary is a long, relatively thick bone that makes up the anterior two-thirds of the mandibular ramus (Figure 11a, b, c). Nearly all of its medial surface is overlain by the splenial. Posteriorly, it extends approximately to the midpoint of the orbit, where it forks dorsally and ventrally around the mandibular fenestra. The posterodorsal process tapers gradually as it underlies the surangular on a sloping suture, eventually ending at the posterior margin of the mandibular fenestra, as indicated by a fragment of dentary adhering to the surangular in this region. This dorsal process is three times the length of the posteroventral process. The posteroventral process is broken but appears to be very short, and contributes little to the ventral border of the mandibular fenestra. Anteriorly, the dentaries meet in a symphysis that extends to the fifth dentary tooth, opposite the third maxillary tooth. The symphyseal region faces anteroventrally, and is not as flattened as in Macelognathus. Small “nutrient” foramina are present on the lateral surface of the anterior half of the dentary but not the posterior half, about mid-way

80 on its lateral surface. The ventral surface of the anterior portion of the mandible is moderately pitted.

Most of the dentary teeth are not exposed, but CT scans show at positions for 17 teeth in each dentary (Figure 11a, b). The teeth gradually enlarge from the first dentary tooth to the fourth. The fourth dentary tooth projects into a pocket between the premaxilla and maxilla. The dentary is only slightly expanded laterally beneath the fourth tooth. The teeth posterior to the fourth tooth are half or less the height of the fourth two or less.

Dentary teeth 5 through 13 are similar in size and teeth 14-17 are smaller than these. -

Dentary teeth 1-13 are conical opposite the premaxilla and recurved and mediolaterally compressed opposite the maxillary teeth. The anterior dentary teeth exposed between the third and fifth right maxillary teeth are serrated posteriorly but not anteriorly, and the serrations are similar to those on the maxillary teeth (Figure 4c). The dentary teeth are not enlarged opposite the enlarged maxillary teeth, and the teeth exposed anteriorly on the right side are similar in size to the teeth in the middle of the maxillary tooth row. The posterior four dentary teeth (14-17) differ slightly in morphology from the anterior teeth, they are constricted at the base, with concavities on the anterior and posterior edges of the tooth root. These posterior dentary teeth are less recurved than the anterior ones. The holotype shows serrations on the posterior maxillary some teeth.

The dentary of Dibothrosuchus is nearly fully preserved, though it has been dorsoventrally crushed which has separated the dentaries at the symphysis, which would have been articulated in life (Figure 12a, b). The dentary symphysis extends to the fourth dentary tooth and is shorter than in Junggarsuchus. The dentary of Dibothrosuchus is slightly narrower and longer than Junggarsuchus. The posterior dorsal and ventral

81 processes of the dentary are not well preserved, but like Junggarsuchus the dorsal process appears slightly longer, but the extent that it borders the mandibular fenestra is unclear, though the posterodorsal end is overlain by part of the surangular. The ridge along the posterior dorsal process, ventral to the surangular is longer and more concave than

Junggarsuchus. Like Junggarsuchus the ventral anterior portion of the bone is pitted and the lateral anterior dentaries have several foramina. Both pitting and foramina continue farther posteriorly and ventrally in Dibothrosuchus.

Each dentary of Dibothrosuchus appears to have 16 teeth (Figure 12b). Like

Junggarsuchus the teeth increase in size anteriorly and the fourth is the largest, though the second tooth is nearly as tall as the fourth. The 5-14th teeth are half the height of the anterior dentary teeth and become less recurved farther back. The alveoli for 2-3 more teeth posterior to the teeth preserved are observed, but information on these teeth is unavailable. Constriction below the root of the posterior dentary teeth is not present in any anterior preserved teeth. Serrations are difficult to determine from CT data, though previous analysis have shown that the enlarged fourth tooth had posterior serrations (Wu and Chatterjee, 1993).

The Junggarsuchus splenial lies medial to the dentary and makes up the medial surface of the mandible anterior to the mandibular fenestra (Figure 11 c,f). It does not reach the mandibular symphysis anteriorly, ending about 2 mm posterior to it. Its anterior edge is forked, forming an opening for the mandibular portion of the trigeminal ganglion (George and Holliday, 2013) (Figure 11f). Similar structures are known in

Dibothrosuchus and Sphenosuchus; (Walker, 1990). The suture between the splenial and the dentary is clearly visible along the ventral surface of the mandible, and the splenial

82 forms the medial one third of the ventral surface (Figure 11c). The medial surface of the splenial is flat and smooth. Posteroventrally the splenial’s contact with the angular is obscure, but the splenial apparently extended posteriorly to about midway beneath the mandibular fenestra. The splenial forms the anterior border of the mandibular fenestra on the medial surface of the mandible. Posteriorly the ventral extent of the splenial forks, for what may be the anterior opening of the infra Meckelian fenestra. This fenestra is also seen in Dibothrosuchus (Wu and Chatterjee, 1993), and Sphenosuchus (Walker, 1990).

Both splenials of Dibothrosuchus are preserved (Figure 12b, c,e), contacting the medial surface of the dentary and are widely similar in their ventral suture to the dentary and anterior extent to Junggarsuchus. Like Junggarsuchus, the splenial contributes to the medial wall of the mandibular fenestra. Though incomplete it appears that posteriorly the splenial forks into a dorsoposterior process and the longer posteroventral process, which contacts the medial edge of the angular more extensively than it does in Junggarsuchus.

The posterior extent of the splenial has been reconstructed to have a medial fenestra, in which the splenial forms the anterior borders and the prearticular forms the posterior border. However, the bone is fragmentary here, and we cannot report that opening from

IVPP V7906 with confidence. No anterior fork is preserved (Wu and Chatterjee, 1993).

The coronoid of Junggarsuchus was an elongate splint like bone that runs medial to the intradentary plate and dorsal to the splenial that is more than half the length of the mandible (Figure 11b,f). The coronoid is antero-posteriorly long, extending from the second dentary tooth to the midpoint of the anterior dorsal process of the surangular. It widens dorsoventrally around its mid-point, and posterior to this it constricts faintly and then expands slightly, at the end of the dentary. The coronoid arches dorsally along with

83 the dorsal process of the dentary. Laterally it contacts the anterodorsal portion of the dentary anteriorly and the dorso-anterior process of the surangular on its medial surface.

The ventral edge of the coronoid contacts the splenial from the 6th dentary tooth to the end of the bone. On the left coronoid, the posterior end of the bone contacts a broken fragment of the splenial ventrally that had been misidentified as the coronoid. A similar dorsoventrally tall plate of bone is not observed in the same posteriorly concave shape on the right side, so it is unlikely this is a massive ventral extension of the coronoid.

Long blade-like coronoids in sphenosuchians have been reported in Sphenosuchus

(Walker, 1990) and Dromicosuchus (Sues et al, 2003). A short coronoid similar to those seen in Protosuchus is reported in Terrestrisuchus (Crush, 1984). Following our analysis, we report that an elongate coronoid dorsal to the splenial is a trait that occurs widely across sphenosuchians. We report its presence in Litargosuchus, Kayentasuchus and

Dibothrosuchus as well based on new observations. The coronoid of Dibothrosuchus was originally reported as a short crescentic bone (Wu and Chatterjee, 1993). However, in our segmentation of the skull, the coronoid is a long blade-shaped bone. In overall shape the coronoids are similar to those of Junggarsuchus, though they are dorsoventrally shorter relative to the dentary (Figure 12e). Anteriorly the coronoid contacts the splenial ventrally and the laterally contacts the intradentary plate on the plates medial side. The posterior portion of the coronoid contacts the anterior dorsal process of the surangular.

The coronoids do not arch dorsally, but bow out laterally, likely due to post mortem crushing. Much of the bone is obscured by crushing which has moved the midsection deep to the splenial.

These elongate coronoids are not present in crocodyliforms except thalattosuchians, but have been reported in distantly related groups. An elongate bone similar to the coronoids reported here has been reported in thalattosuchians (Wilberg,

2015), sauropodomorphs, primitive theropods, tyrannosaurs and abelosaurid (Sampson &

Witmer, 2007; Hurum and Currie, 2000) theropods and referred to as a ‘supradentary’ in tyrannosaurs (Sampson and Witmer, 2007), but this identification is likely inaccurate

(Walker, 1990). The function of this bone is poorly understood, though it has been suggested that this bone plays a role in reinforcing the mandible (Hurum and Currie,

2000).

In Junggarsuchus the angular is nearly fully preserved on the left side and only the anterior most portion is preserved on the right side (Figure 11a,c). In lateral view, it extends anteriorly to meet the dentary near the anterior end of the mandibular fenestra, and proceeds anteriorly between the dentary and splenial to the 17th dentary tooth. The angular contacts the splenial medially beneath the fenestra. The dorsal surface of the angular is only slightly concave beneath the fenestra. The posterior half of the angular does not expand dorsally, and makes a longitudinal dorsal contact with the surangular at the level of the ventral edge of the mandibular fenestra. A distinct longitudinal ridge near the dorsal edge of its lateral surface begins anteriorly beneath the posterior end of the mandibular fenestra and ends posteriorly beneath the quadrate articulation. This marked the dorsal edge of the insertion area for the M. pterygoideus ventralis muscle (Figure

11a,d). This surface is seen in neosuchian crocodiles and not in sphenosuchians or protosuchians. The angular continues posteriorly to the end of the mandible, forming the dorsal part of the lateral surface of the retroarticular process, and may contribute to the

85 lateral edge of the quadrate articulation surface. The suture between the surangular and angular posteriorly is unclear. This posterior dorsal surface of the angular may be a continuation of the surangular. A break in the articular makes this suture difficult to determine. If it is part of the surangular then the angular does not reach the end of the mandible and terminates abruptly, with a small medial extension on the medial portion of the bones lateral face (Figure 11d). Medially the angular is covered by the splenial and posteriorly by the prearticular (Figure 11c). The angular tapers in medial view on the ventral surface. Any contact with the prearticular is difficult to interpret.

Both angulars are nearly completely preserved in Dibothrosuchus (Figure 12a,c) and are longer and dorsoventrally shorter than in Junggarsuchus. The angular contacts the prearticular on the ventromedial contact. Unlike Junggarsuchus, the angular does not extend to the end of the mandible. Instead it narrows into a short projection that arches slightly dorsally on the lateral face of the surangular, which has a substantial ventral process. The angular also lacks the distinct ridge for the insertion of the pterygoideus muscle, seen in Junggarsuchus. The angular forms the ventral border of the mandibular fenestra, but the exact size and shape of the fenestra is unclear due to the fragmentary nature of the posterior mandible. Wu and Chatterjee reconstructed the fenestra as a small circular fenestra, but the CT scans do not preserve the surangular and angular completely, which limits confident reconstruction of the size and shape of the mandibular fenestra.

The anterior process of the surangular overlies the dentary above the mandibular fenestra (Figure 11 a, b). On the lateral surface the dentary separates the surangular from the dorsal edge of the fenestra, but the surangular forms the dorsal roof of the mandibular fossa. Its anterior tip is narrow and gradually widens posteriorly above the fenestra. The

86 main body of the remaining bone has a flat dorsal surface that faces posterodorsally. The surangular forms the entire posterior edge of the mandibular fenestra and at least the posterior most ventral border. Of particular interest is a small fenestra near the posterior end of the surangular. The anterior edge of this opening is broken but otherwise the edges are smooth, and it is a natural feature. A surangular foramen does not appear in any other crocodylomorph, but one is present in some theropod dinosaurs, specifically the

Tyrannosauridae and Dromaeosauridae. The posterior ventral suture of the surangular to the angular is unclear. The surangular may end abruptly anterior to the potential dorsal projection of the angular onto the lateral face of the articular (Figure 11a). However, if the surangular forms this lateral cover of the articular, the angular ends abruptly. If the latter is the case, this is similar to the condition of the surangular in Dibothrosuchus and

Sphenosuchus (Figure 11d). CT scans suggest that the latter is more accurate, with a posteriorly extensive surangular. Unfortunately, even in CT scans the suture interpreted here is not fully evident.

Both surangulars of Dibothrosuchus are preserved, though the mid sections of both bones are incomplete and so the exact border of the mandibular fenestra and fossa is difficult to determine (Figure 12a,b,f). They are long narrow bones, lacking the dorsal arch seen in Junggarsuchus. The crushed nature of the surangulars make inferring the size of the mandibular fenestra difficult. It appears that it was dorsoventrally shorter than in Junggarsuchus, and Wu and Chatterjee reconstructed the fenestra as a small circular opening, unlike the larger ovule fenestra in Junggarsuchus. The anterior process of the surangular, overlaying the posterior process of the dentary is broader than in

Junggarsuchus. The posterior portion of the surangular differs from Junggarsuchus,

Dibothrosuchus has a tall surangular, that is dorsoventrally as tall as the mandible. The surangular extends ventrally along the medial surface of the angular and forms the ventral and posterior border of the medial face of the mandibular fenestra, unlike in

Junggarsuchus. The surangular covers the entire lateral surface of the articular and also expands slightly laterally as a ridge on the posterior end of the mandible. Ventral to this posterior ridge the laterally expanded face of the surangular is concave. The lateral posterior cover of the articular is more extensive in Dibothrosuchus than any other sphenosuchiam, but is comparable to Sphenosuchus. Dibothrosuchus, like all other sphenosuchians except Junggarsuchus, lacks a surangular fenestra.

The condition of the retroarticular process in Junggarsuchus, preserved well on the both sides, is unusual among basal crocodylomorphs (Figure 11, b,c). Whereas in other ‘sphenosuchians’ there is often a dorsomedial projection, as in Dibothrosuchus, or a posterior one, like in Pseudhesperosuchus, the process is ventrally directed. The retroarticular process has a concave posterodorsomedial surface, for the insertion of the

M. depressor mandibulae. This concavity gradually opens up posteroventrally. All of this surface is composed of the articular. The posterior edge is rimmed laterally by a vertical lip that extends outwards from the main body of the articular, and forms the posterior border of the insertion area for the M. pterygoideus. A triangular lateral exposure of the articular anterior to this lip forms part of this insertion area ventral to the angular. The medial edge of the articular also forms a lip along the dorsomedial edge of the retroarticular process. The quadrate articulation is not antero-posteriorly elongated as in some notosuchians. The articular is more dorsally expanded on the medial portion of the anterior edge of the articulation surface than on the posterior edge, suggesting the

88 primary forces of mandibular adduction had an anterior component. The articular covers the medial surface of the angular, but a displaced rib and neural spine cover the articular medially. Mid way along the ventral portion of the medial face of the articular there is a small groove that appears to open in a foramina, which may be for the chorda tympani.

At the anteromedial end of the mandibular fenestra, an acute process inserts into the splenial and borders the ventral edge of the splenial; this process may belong to a prearticular if one is present covering the medial surface of the articular. CT scans suggest this narrow anterior process that runs along the dorsal side of the angular might be part of the articular, but identification of the prearticular remains indeterminate, even with CT scans. A very thin sheet of bone projects along the ventral end of the articular and appears continuous with the projection of bone dorsal to the angular which may be prearticular (Figure 11e). If this is the prearticular, the contact with the articular and angular is very similar to that in Dibothrosuchus. The issue is that dorsally it becomes difficult to identify any suture with what is the articular. A thin rectangular wall of bone is located medial to the anterior process of the articular. It is possible that this thin wall of bone may be part of the prearticular.

The retroarticular process of the articular in Dibothrosuchus expands dorsomedially, unlike Junggarsuchus (Figure 12b, c,d). Like Junggarsuchus, there is an anterior process of the articular than appears to extend towards the posterior end of the splenial. However, this may be a portion of the prearticular, the exact suture isn’t fully apparent due to some crushing. The ventral exposure of the articular is limited to the posterior end of the mandible, with much of the bone covered by the prearticular, surangular and angular. Whether this condition is present in Junggarsuchus is unclear

89 due to the uncertainty of the extent of the prearticular. No medial groove is preserved in

Dibothrosuchus, though the foramen aerum is present on the dorsal face of the posteromedial portion of the articular, unlike in Junggarsuchus. The articular fossa of

Dibothrosuchus is mediolaterally wide and shorter dorsoventrally than in Junggarsuchus.

The medial walls of the articular fossa are also lower in Dibothrosuchus than

Junggarsuchus.

Both of the prearticulars of Dibothrosuchus are preserved, though the right prearticular is better preserved (Figure 12 b,c). The bone extends anteriorly along the dorsal groove of the angular. The medial and dorsal surface of the prearticular contact the articular. The prearticular contacts the ventral edge of the angular anteriorly and the surangular posteriorly and reaches the end of the articular. The bone is very thin and there may be more exposed on the left and right sides along the medial face of the articular, but the sutures are not clear.

Pectoral girdle and forelimb

The scapula of Junggarsuchus is completely preserved on both sides of the skeleton but is only exposed on the left side. It is similar in overall shape to the scapula of other “sphenosuchians” but is slightly broader and more triangular, becoming extremely broad dorsally (Figure 13a,b). The dorsal portion is much wider anteroposteriorly than the ventral portion, and the middle of the bone is slightly constricted. The glenoid fossa at the anteroventral edge is unusual among crocodylomorphs in its orientation: whereas the fossa of most other crocodylomorphs has a lateral component to its direction, the glenoid of Junggarsuchus has no lateral component. It is directed ventrally and slightly

90 posteriorly, if the scapula-coracoid articulation was horizontal. This results in an unusual area for articulation with the humerus, leaving little space between the edge of the scapula and the coracoid. The ventral orientation of the glenoid fossa is more comparable to protosuchians than to other sphenosuchians. The glenoid fossa comprises a circular concavity that accommodated the rounded head of the humerus.

The anterior edge of the scapula is anteriorly convex in its ventral part, forming a rounded surface. In Junggarsuchus, the anterior edge of the scapula is more strongly concave than the posterior (Figure 13a,b). This condition may also be present in

Dibothrosuchus (Figure 13c), though the anterior most edge of the scapula is broken (Wu and Chatterjee, 1993). In Sphenosuchus the anterior and posterior edges of the scapula are narrow (Walker, 1990), and the condition seen in Junggarsuchus and Dibothrosuchus is more comparable to the scapula of Crocodyliformes including Protosuchus. Dorsal to this, the anterior edge becomes anteriorly concave, continuing to the dorsal edge of the scapula. The most anterodorsal area of the scapula is anteriorly convex while the dorsal margin of the scapula is slightly concave. The dorsal edge is gently sinuous, with a broad depression in the middle and an extended posterodorsal corner. The posterior margin does not curve, but a long, narrow ridge along the posterodorsal half of the margin marks the anterior edge of the insertion area of the large M. serratus ventralis thoracis. The lateral surface of the scapular blade is mostly flat and smooth, with a low but distinct ridge near to and paralleling its anterior edge that would have served for the attachment of the M. deltoideus scapularis, or possibly the M. deltoideus clavicularis. The surface of the bone is also raised immediately dorsal and slightly posterior to the glenoid fossa, where the M. coracobrachialis brevis dorsalis would have attached (Meers 2003). A

91 raised area in the posterodorsal part of the lateral surface may be due to deformation from the underlying vertebra.

Both coracoids in Junggarsuchus are preserved in articulation with the scapula; the left element is completely exposed while the right is not (Figure 13a,b). The coracoid is nearly the same length as the scapula, as seen in crocodylomorphs like Sphenosuchus,

Dibothrosuchus, and early crocodyliforms like Protosuchus. The left scapula and coracoid have been slightly disarticulated so that the coracoid is bent ventromedially and shifted slightly posterior relative to the scapula. The anterior edge of the glenoid fossa of the coracoid is displaced to lie beneath the posterior edge of the fossa on the scapula. The coracoid is elongate ventrally and a smooth rod-like process extends posteroventrally, as in Dibothrosuchus (Figure 13c), which lacks the ventral groove found in Sphenosuchus,

Terrestrisuchus and earlier crocodylomorphs and pseudosuchians (Chatterjee, 1985). Its articulation with the scapula occurs entirely anterior to the glenoid fossa. The coracoid portion of the glenoid fossa is nearly perpendicular to the scapular portion, and faces posteriorly and somewhat dorsally, similar to Protosuchus and more crownward

Crocodyliformes. The orientation of the glenoid fossa of the coracoid is more vertical than the subhorizontal fossa of most other crocodylomorphs, like Dibothrosuchus and

Sphenosuchus, and earlier pseudosuchians like Postosuchus. Anteriorly, the coracoid is relatively thick and square shaped. The postglenoid process of the coracoid is a long rod of bone that is approximately the same length as the remainder of the coracoid; it extends to the posterior margin of the second thoracic vertebrae as preserved specimen. The process lacks any indication of the articulation with the sternum and a biceps tubercle, such as are present in Dibothrosuchus (Wu and Chatterjee, 1993), so the extent of the

92 articulation between the coracoid and sternum, if present, cannot be determined. The elongate posteromedial process of the coracoid found in Junggarsuchus, Dibothrosuchus

(Wu and Chatterjee, 1993), and Sphenosuchus (Walker, 1990) is believed to represent a transitional stage from the shorter expanded process found in other sphenosuchians, like

Terrestrisuchus (Crush, 1984) to the ventrally expanded ventromedial process of

Crocodyliformes.

An interclavicle and clavicles are not evident, but the ventral midline of the skeleton has not been completely excavated. An ossified sternum has also not been identified.

Both humeri are preserved (Figure 14a,b). The right element remains in articulation with the shoulder girdle but much of the rest of the forelimb was lost to erosion. Nearly all of the elements of the left forelimb have been preserved in articulation and removed from the skeleton. The left humeral shaft shows a slight concavity medially at mid-shaft and laterally near the distal end, but some of this may be due to distortion.

The humeral shaft is gently but distinctly curved, with the shaft distal to the deltopectoral crest forming an anteriorly concave arc. This curvature is comparable to that of

Sphenosuchus and more pronounced than that of Dibothrosuchus, Hesperosuchus (CM

29894) and Terrestrisuchus. The proximal articular surface of the bone is proportionally large when compared to other ‘sphenosuchians’ it projects posteriorly perpendicular to the rest of the articulating surface and forms a semi-spherical head. This expansion extends onto the proximal surface of the articulation. A small depression lies laterally opposite the head on the left humerus but this may have been caused by postmortem crushing, as its presence on the right side was undetectable. The remaining articulating

93 surface is mediolaterally broad and slightly concave anteriorly, and the part medial to the expanded head is slightly narrower than the part lateral to it, and tapers medially. A thin ridge descends from the medial edge of the articulation, gradually declining into the shaft distally about ¼ of the way down the bone. Laterally, the proximal articulating surface is continuous with the dorsal surface of the deltopectoral crest and there is no separation, as in some crocodylomorphs (Nesbitt, 2011). The deltopectoral crest does not project as far from the shaft as in Sphenosuchus, and is similar in size to Dibothrosuchus,

Hesperosuchus (CM 29894), and Terrestrisuchus but extends further down the shaft than in the latter. Its distal edge is straight for three quarters of its length beginning on the proximal humeral end, and it then quickly descends to the shaft. The flat portion has a rough surface similar to that of the proximal articulating surface and continuous with it.

Together the deltopectoral crest and the medial ridge enclose a fairly deep depression on the proximal end of the humerus on the posterior side.

The distal end of the humerus is approximately the same width as the proximal end, which is seen in other sphenosuchians like Dibothrosuchus (Figure 14e) and

Sphenosuchus. the lateral and medial condyles are similar in size and separated by a very shallow intercondylar groove posteriorly; the anterior surface is not exposed. The lateral condyle is slightly narrower and deeper than the medial condyle, and forms a sharper ridge lateral to the intercondylar groove. However, the disparity in size is not as much as in Dibothrosuchus (Figure 14e) (Wu and Chatterjee, 1993, fig. 12), Terrestrisuchus

(Crush, 1984, fig. 7), and Sphenosuchus (Walker, 1990, fig. 43).

The left and right ulnae are complete, the left element is preserved intact with the rest of the arm (Figure 14a,b). The right ulna and radius are preserved in pieces, except

94 for the proximal ends articulated with the humerus on the main block. The left ulna is nearly 10mm longer than the radius, and only a small part of this disparity is due to the olecranon process on the ulna. The proximal end of the ulna is subtriangular, similar to that of Dibothrosuchus but with a more concave posterior edge (Figure 14e). The olecranon process is low, broad and gently convex. It is lower than in Dibothrosuchus,

Terrestrisuchus, and Hesperosuchus (holotype and CM 29894), and less distinct than in the latter two taxa This lower process may represent a transition towards the nearly nonexistent olecranon process of protosuchians. A low ridge continues medially from the olecranon process along the anterior edge but does not reach the medial edge. Medially the proximal articulating surface is flat anteriorly and has a small but pronounced process posteromedially. The shaft of the ulna is widest at the proximal end, where it has been anteroposteriorly crushed on the left element; it becomes thinner and more circular in cross section distally, and the distal end is only slightly thicker than the shaft. The distal end of the ulna has an anterior expansion that is found in nearly all Crocodyliformes, with the exception of thalattosuchians. As preserved the radius and ulna have only a short contact proximally, and a more extensive one near the distal end. As in Dibothrosuchus the distal end of the ulna’s medial facet for articulation is not confluent with the distal articulation for the ulnare (Figure 14d). This contact occurs along a flat medial surface of the distal ulna, seen best on the right element. This surface may also have contacted the lateral edge of the radiale. The ulna ends in a rounded process that may have contacted a pisiform, as in Dibothrosuchus and Terrestrisuchus, but one is not preserved. The articulation surface extends laterally from this process, tapering in anteroposterior width and in distal length. The posterior surface of the ulna’s distal end forms a broad groove

95 on the right element, but the left element appears to be flatter in this region, although it is largely unexposed.

The radius is a cylindrical bone about half the diameter of the ulna (Figure 14a,).

Unlike other sphenosuchians, the radius is shorter than the humerus, similar to the condition seen in Protosuchus. The radius is also much shorter than the ulna, due in part to the greater proximal length of the radiale. The proximal end of the radius is preserved on the left side ending far below the ulna’s proximal articulating surface, due presumably to distal dislocation. The proximal end is expanded anteroposteriorly, but the proximal surface is not exposed. A faint ridge is present running along the anterior edge of the proximal end of the bone. The distal end is expanded anteroposteriorly but is only slightly wider than the shaft. Its lateral surface is flat and abuts against the medial surface of the ulna.

The radiale and ulnare of Junggarsuchus are preserved only on the left side. They are elongate, as in all crocodylomorphs, although to a greater degree than most. The radiale is more than half the length of the radius (Figure 15a), unlike other sphenosuchians or crocodyliforms, like Dibothrosuchus (Figure 15b) and Protosuchus, which have an elongated radiale, but not to the degree seen in Junggarsuchus The convex distal end of the radius articulated with a concavity on the proximal end of the radiale in a concavo-convex joint. The articulation surface on the radiale is hemicircular, with a flat posterior edge, and the curved edges are raised while the flat one is not. In anterior view the proximal end of the radiale is hatchet-shaped, with a broad expansion laterally. This expanded area has two longitudinal ridges on its medial surface descending from the proximal end, with a surface similar to that of the articulation. The

96 anterior ridge is about half the length of the posterior ridge. The ulna would have articulated with the medial surface along these ridges, but it is unclear what surfaces were in contact. The proximal end of the ulnare contacts the radiale at the base of this medial process. The dorsal surface of the radiale is smooth and slightly convex; it increases gradually in breadth to the distal end. The distal end abuts a medial distal carpal, but it is not exposed. The distal end of the radius is convex, at least anteriorly, but the more posterior part of this surface is not exposed. The distal end of end of the radiale also has a lateral contact with the mediolaterally broad distal end of the ulnare.

The ulnare in Junggarsuchus is shorter and narrower than the radiale, but with a much broader distal end (Figure 15a). Its proximal end is concave, and there is no pisiform preserved proximal to it, as in Dibothrosuchus (Figure 15b) and Terrestrisuchus.

The anterior edge of the proximal surface extends proximally in the middle, forming an anterior buttress to the articulation of the ulna or pisiform. The proximal end of the ulnare is slightly more than half the width of the distal end of the ulna. The middle of the shaft, where it does not contact the radiale is thin and straight. Its dorsal surface is rounded while the ventral side appears to be flatter, although it is mostly covered by matrix. The distal end of the ulnare flares distally to become twice as broad as the proximal end, and is flattened anteroposteriorly. Relative to the proximal end it is twisted about 45 degrees, so the anterior surface faces anteromedially.

One distal carpal is present and another may be present but is unexposed (contra

Clark et al., 2004, Fig. 3d) (Figure15a). The lateral distal carpal is flattened and articulates broadly with the distal end of the ulna and the proximal ends of metacarpals

II-III. It is unclear whether this carpal is a fusion of carpals 3 and 4, as may be present in

Dibothrosuchus (Figure 15b). A space between the radiale and the proximal ends of metacarpals IV-V suggests the presence of a medial distal carpal, but it is not exposed.

Four metacarpals are present, and the lack of articulation surfaces medially on the medial element and laterally on the lateral element suggest that another one was not present. Contrary to Clark et al. (2004), digit I is interpreted here as reduced and digit V absent, as in Dibothrosuchus (16b). The third metacarpal is the longest, the second and fourth are slightly shorter and approximately the same length (Figure 16a). The first metacarpal does not reach as far proximally or distally as the other metacarpals, is about half the breadth and width of mcII, and adheres to the shaft of mc II over its entire length.

The metacarpals of digits II-IV are similar in breadth, but mc IV is slightly thicker than the other two. Metacarpals II-IV are preserved compacted together, similar to the preservation of Saltoposuchus (Sereno and Wild, 1992). The proximal ends of metacarpals II-III are mediolaterally flattened, but that of mc IV is convex laterally. The distal ends of metacarpals II-IV have two well developed condyles, the medial condyle is slightly larger than the lateral on mc III and the lateral condyle slightly larger than the medial on mc II. The distal surface of mc I is simple, lacking distinct condyles.

Three entire phalanges, the proximal part of two more, and the distal part of two others are preserved. A phalange preserved with the manus is attached near the end of the first metacarpal. It is a proximal phalange, and is too large to be from digit I so may be from digit II. Its proximal end is gently concave, it is about four times as long as wide, and distally it has a single convex condyle with only a very gentle indentation where the intercondylar groove usually lies. Distinct ligament pits indent the lateral and medial surfaces, and the condyle is slightly flared ventro-laterally, so that the ventral part

98 of the condyle is wider than the dorsal part. The fourth digit retains its natural articulation with the proximal end of a phalange. The proximal end of the phalange has a very short posterior process on its ventral surface, which fits into the intercondylar groove of mc IV. The proximal end preserved separately is similar to the phalange articulated near II, and one of the distal pieces probably belongs with it as it is very similar to the articulated phalange. The distal piece differs from the articulated phalange in having an intercondylar groove, although it is very broad and shallow. A small isolated complete phalange is about 2/3 as long as the articulated phalange, and its distal end is not flared ventro-laterally. The other distal phalange has a broad, shallow intercondylar groove but no ligament pits, is not flared ventrolaterally, and expands medially and laterally more than does the articulated phalange. The phalanges that can be observed appear mostly similar to those present in Dibothrosuchus in overall form. No unguals are preserved in Junggarsuchus, unlike Dibothrosuchus.

An isolated phalange is unusual in shape and similar to phalange IV-2 of

Dibothrosuchus (contrary to Wu and Chatterjee 1993, figure 13F). It is about as long as its proximal end is wide. Its proximal end is greatly expanded, about twice as broad as at mid-shaft, and flat proximally. It abruptly narrows distally, and two distal condyles are well developed and confluent. The condyles are slightly beveled and face slightly medially and laterally. Shallow ligament pits are present.

Axial skeleton

The disarticulated intercentrum and neural arch of the atlas were preserved adjacent to the medial surface of the left angular and surangular (Figure 11c). The

99 intercentrum is flattened dorsoventrally and dorsally concave. It is concave anteriorly where it contacted the occipital condyle. Anterolaterally, the atlas centrum projects anteriorly to form a flange on either side of the anterior concavity where the neural arch would have articulated dorsally.

The right half of the atlas neural arch is preserved and exposed adjacent to the mandible (Figure 17a). The postzygapophysis projects laterally from the flat anterior surface of the neural arch and does not have a long posterior component. There is a notch immediately medioventral to the postzygapophysis that separates it from the ventral process. This process would have made up part of the articulation with the odontoid process as well as the intercentrum. Posterodorsally, a subtriangular, flat surface would have articulated with the axis prezygapophysis. The atlas of Dibothrosuchus is largely similar to what is preserved of Junggarsuchus.

The axis is preserved disarticulated from the rest of the vertebral column (Figure

17a). Anteriorly, the odontoid process is semicircular in anterior view, with a flat dorsal surface and a rounded ventral surface. The diapophysis and parapophysis are represented by two small, round projections from the ventrolateral surface of the process. They are almost directly adjacent and are only separated by a slight concavity on the ventral surface of the centrum. Three low projections are present on the dorsal part of the odontoid process: an anteromedial one and a lateral one on either side. These would have inserted at the junction of the atlas centrum and neural arch. The centrum body is both narrower and anteroposteriorly longer than the cervical vertebrae and it is strongly concave ventrally. The posterior end of the axis centrum has a concave dorsal surface and slopes posteroventrally. It has a convex condyle posteriorly. The neural arch is smooth,

100 due to the lack of rib articulations. Anteriorly, there are two very small pre-zygapophyses that would have articulated with the atlas neural arch, similar to the pre-zygapophyses seen in Dibothrosuchus (Figure 18b) and Sphenosuchus. The post-zygapophyses are larger and more horizontally oriented. The neural spine appears to have been mostly broken off, leaving only a low ridge that slopes downward towards the posterior end. The axis does not have a hypophysis ventrally.

The post-axial vertebrae of Junggarsuchus are procoelous, unlike the amphicoelous or amphiplatyan vertebrae of other basal crocodylomorphs. Four post-axial cervical vertebrae are still articulated with the skeleton (Figure 18a) and another one was collected separately. The centrum and neural spines of the anterior cervical are both taller than those of the dorsals, and the more posterior cervical vertebrae become progressively smaller until they grade smoothly into the dorsal vertebrae. The neural spines are broken off most of the vertebrae but when present they appear to be equal in height to the entire centrum and neural arch complex. They are flattened mediolaterally and taller than they are anteroposteriorly long, resembling the vertebrae of living crocodiles, although the neural spines of Junggarsuchus do not become rod-like in posterior cervical vertebrae as in the latter. The pre-zygapophyses are pronounced and almost completely vertical, with angles of close to 90o. These interlock with the vertebrae in front of them across a flat contact. Such a configuration likely constricted horizontal movement of the spine in Junggarsuchus but would have allowed for greater vertical motion. The lateral surfaces of the vertebrae are depressed compared to the anterior and posterior ends, particularly between the parapophyses and diapophyses. On the three exposed cervicals, the parapophysis is situated anteroventrally on the lateral side of the

101 centrum and faces laterally, the diapophysis is nearly in the center of the lateral surface and facing ventrolaterally.

Hypapophyses are present on the ventral surface of at least the four most posterior cervical vertebrae. These are also present on extant crocodylians but are unknown in other ‘sphenosuchians’ and basal Crocodyliformes. The hypapophysis on the anterior articulated cervical vertebra is restricted to the anterior end of the centrum. On the following three vertebrae the hypapophysis becomes more centered and vertically oriented. They are relatively long anteroposteriorly compared to living crocodylians.

The condyle developed on the posterior end of the centrum is hemispherical peripherally, but a central depression occupies more than half the diameter of the condyle. The depression on the anterior end of the centrum is smoothly concave, lacking any expansion corresponding to the depression in the condyle.

All 9 cervical vertebrae are preserved in Dibothrosuchus, including the neural arches and most of the neural spines (Figure 18b). The cervicals are similar to

Junggarsuchus in size, shape and in the form of the neural spine. The cervical zygapophyses are sub-vertical, as in Junggarsuchus. In general anatomy cervicals are similar, but Dibothrosuchus lacks hypapophysis on the cervical and dorsal vertebra.

Eleven dorsal vertebrae remain in articulation and four more are separated from the skeleton in Junggarsuchus (Figure 19a). The parapophyses and diapophyses are mostly smooth and concave except for the parapophyses on the anterolateral edge of the vertebral centrum. On their lateral surface, the neural arches have a low ridge located posterodorsally, which makes up part of the diapophysis. This ridge extends from the diapophysis and ends just anterior to the post-zygapophyses. A short transverse process is

102 present, but it is only exposed on the last articulated vertebra where it is inclined posterodorsally. The relationship of the parapophysis and diapophysis to the transverse process is unclear, as the distal end of the process is broken on the last vertebra. The pre- zygapophyses are less vertically oriented than those of the cervical vertebrae, with angles closer to 65o. They are also smaller in size. The neural spines are slightly taller than half the height of the centra, and their anteroposterior length is nearly equal to that of the centra. Their height is consistent throughout the entire dorsal section of the vertebral column, and the dorsal edge is horizontal throughout. The posterior edge of the neural spines of the most anterior dorsals have a short posterior projection dorsally, while more posterior dorsals develop a short anterior projection on the dorsal part of the anterior edge of the neural spine. The transition from cervical to dorsal vertebrae is hidden by the scapula, so the dorsal progression of the parapophysis on more posterior vertebrae is not visible. The four most anterior dorsal vertebrae have well developed hypapophyses, which curve posteroventrally, unlike those of the cervicals. The anterior depression and posterior condyle on the centra are similar in shape to those on the cervicals.

The first three dorsal vertebrae and last three dorsals are known from

Dibothrosuchus (Figure19b). The cervico-dorsals are amphicoelous, which is standard for sphenosuchians other than Junggarsuchus. The anterior three dorsal vertebrae are similar in form to those in Junggarsuchus. The parapophyses migrate dorsally posterior to become confluent with the parapophysis. The neural arch is as tall as it is long, and the angle of the pre-zygapophyses appears comparable to that seen in Junggarsuchus. The neural spine is preserved on the first dorsal vertebrae and is narrower and slightly taller

103 than those in Junggarsuchus. They also lack the posterior and anterior projections seen in

Junggarsuchus.

The ribs are preserved only on dorsal vertebrae, except for one cervical rib partially exposed to the right of the second preserved cervical on the main block (Figure

19a). Many dorsal ribs have been preserved undistorted and articulated but they are badly fractured at many points along the bone. The cervical rib is straight with a short anterior process and a long posterior process, about the length of the centrum of the cervical above it. The cervical ribs are double-headed, as indicated by the parapophyses and diapophyses of the vertebrae. The preserved dorsal ribs are all double-headed, gracile and distally become nearly cylindrical. Dorsally the ribs are flattened dorsoventrally, and distally they become thinner and curve medially. Particularly well-preserved examples, such as the ribs on dorsal vertebrae 2-5, show an anterior longitudinal depression in the proximal 1/3 to 1/2 of the rib. The anterior dorsal ribs have likely been displaced and are propped up on their heads, sitting perpendicular to the vertebrae and not preserving the natural contact. The more posterior dorsal ribs lie flat on their heads against the ridges on the lateral surfaces of the vertebrae, with the parapophyses placed more dorsally on the centrum and the tuberculum meeting the diapophysis anteroventral to the postzygapophysis. The diapophysis is further posterior when compared with the cervical vertebrae. Although they are partially flattened, the ribs of Junggarsuchus appear to have only a small horizontal component dorsally, as in other sphenosuchians (Crush, 1984,

Walker, 1990) but unlike in extant crocodylians. This would have made the animal’s body dorsoventrally taller, unlike the mediolaterally wide bodies of modern crocodylians.

104

The condition in Junggarsuchus is inferred from the sharp curvature of the proximal ends of the ribs in which the rib is directed ventrally almost immediately lateral to the head.

Two dorsal ribs are known from Dibothrosuchus, unfortunately the proximal ends of the ribs are not preserved. In general anatomy the bodies of the ribs appear similar to the dorsal ribs of other sphenosuchians (Wu and Chatterjee, 1993).

The sacral vertebrae are all preserved in Dibothrosuchus, but not in

Junggarsuchus, though a disarticulated sacral rib is known from the latter.

A single disarticulated caudal vertebra is known from Junggarsuchus. It is a small long centrum, 3 times long as it is wide. It is likely from the mid to posterior section of the tail. The caudals preserved in Dibothrosuchus are from the more anterior series of the caudal and cannot be directly compared to the disarticulated caudal.

Though osteoderms are well known in crocodyliforms and in some other sphenosuchians, like Dibothrosuchus (Figure 18b), Dromicosuchus and Hesperosuchus, no osteoderms are preserved in Junggarsuchus. Junggarsuchus seems to be one of the only sphenosuchians known from relatively complete, articulated material that lacks osteoderms. In Dibothrosuchus (Wu and Chatterjee, 1993) the osteoderms are comparable to other known sphenosuchian osteoderms, like Dromicosuchus and

Hesperosuchus, arranged in a single, longitudinal pair of rows.

PHYLOGENETIC RESULTS

From our initial set of analyses, we produced 8 trees using the various rooting schemes and implied and equal weights as discussed above. The 4 equal weights analyses resulted in 4 trees, with steps ranging from 2005 to 1799 steps. CI ranges from .35 to .39

105 while the RI ranges from .62 to .68. The full set of CI, RI and relevant information on our trees can be found in Tables 1 and 2. Our results generally agree with the results of Clark et al. (2004), Nesbitt (2011), Pol et al. (2013), Wilberg (2015) and Leardi et al. (2017) who found a paraphyletic ’Sphenosuchia’. All of our equal weight analyses produced partial polytomies along a paraphyletic ‘Sphenosuchia’. Our implied weight analyses

(K=12) shows greater resolution among ‘Sphenosuchia’, however the support given at many of these nodes, with the exception of nodes subtending Junggarsuchus,

Almadasuchus, Macelognathus, Dibothrosuchus and Sphenosuchus, are very low, ranging from 24 to 3 and are collapsed in the equal weight’s analyses. In 6 of these 8 analyses we find Junggarsuchus as the sister taxon to ‘Almadasuchus+

Macelognathus+Crocodyliformes’ with the weakest support being a score of 59 (in an equal weight analysis with Gracilisuchus as the outgroup and the omission of

Phyllodontosuchus and Trialestes) and the strongest support for ‘Solidocrania’ being a score of 84, found using implied weights with Postosuchus as the outgroup and the two

‘rouge taxa’ omitted (Figure 22). (Tables 3,4). In our secondary analysis with Tennant’s

(Figures 20, 21) characters we found Solidocrania in 6 of 8 analyses with node support scores ranging from 44 to 81 (Postosuchus as outgroup with implied weights and rogue taxa omitted) (Figures 20, 21, and 22).

This group is supported by several unambiguous synapomorphies: exoccipitals contact below supraoccipital, which is reversed in Almadasuchus (20); greatly expanded basisphenoid with pneumatic cavities (29); the exoccipital contacting the quadrate ventrolaterally (30); reduction in the size of the antorbital fenestra (48); a developed anterior process of the ectopterygoid projecting along the surface of the jugal (58);

106 quadrate, squamosal and otoccipital enclose the cranioquadrate canal laterally (62); the presence of additional quadrate fenestra has been inferred as a synapomorphy of this group, and the loss of the additional fenestra in Almadasuchus and Macelognathus may be secondary losses (78); the primary head of the quadrate approaches the laterosphenoid

(109); a pneumatized pterygoid (313); the anterior process of the pterygoid ramus of the quadrate is firmly sutured to the pterygoid (322); dorsal edge of surangular arched dorsally (376); anterior edge of the scapular blade is larger than the posterior (122); the radius is longer than the humerus (138); the olecranon process of the ulna is very low

(144); the glenoid surface of the coracoid is extended on a vertical plane (442), like

Protosuchus.

Some of these synapomorphies are more ambiguous. In part this is due to the presence of several character states in a basal polytomy of sphenosuchians (Figure 20, 21,

22), which makes it difficult to reconstruct the exact history of these characters across these taxa. The squamosal contacts the posterodorsal surface of the quadrate enclosing the otic recess posteriorly (61) in Junggarsuchus, but this is also present in

Kayentasuchus. The supratemporal fossa anterior margin is posterior to the postorbital

(211), though this condition is also present in the more basal Terrestrisuchus. Two large palpebrals (254) are found in Junggarsuchus, but due to the ease of which the palpebrals can be disarticulated, some absences of these bones may incorrectly inferred. The quadrate ramus of the pterygoid is broad in ventral view (316), similar to later crocodyliforms, but this condition is not reported in Hallopodidae. The last ambiguous synapomorphy is that the supraoccipital may contribute to the medial border of the

107 posterior temporal fenestra (59). While this is the character state found in crocodyliforms it is not seen in Hallopodidae.

In the analyses that recover ‘Solidocrania’ Almadasuchus and Macelognathus are found closer to Crocodyliformes than to Junggarsuchus. The most common of the relationships finds Almadasuchus as the sister taxon to Crocodyliformes, and

Macelognathus as the sister taxon to Almadasuchus + Crocodyliformes (Figure 20, 22).

In our analyses without Tennant’s characters we found this topology in 7 of the 8 analyses we ran. The support for Almadasuchus + Crocodyliformes ranges from 51 to 59, the higher values found in implied weights with Postosuchus as the outgroup (Figure 20,

22). The support for (Macelognathus + (Almadasuchus+ Crocodyliformes)) range from

55 to 77, with the higher values found using implied weights (Table 3, 4). Several synapomophies support this paraphyletic assemblage. For (Macelognathus +

(Almadasuchus + Crocodyliformes)): (7) prefrontal not underlying anterior portion of frontal; (9) dorsal surface of frontal flat; (73) supraoccipital wide; (341) quadrate contacts the basisphenoid; (352) ventrolateral contact of the otoccipital and quadrate is broad; and

(418) loss of tooth serrations. For the more specific (Almadasuchus + Crocodyliformes) we find the following 3 unambiguous synapomorphies: (49) dorsal cranial bones pitted;

(108) postorbital, squamosal and parietal in the same plane and (343) the ventral surface of quadrate is concave. The following may also be synapomorphies but the relevant regions are not preserved in Macelognathus so the condition is unknown: (13) lateral edge of squamosal with longitudinal groove; (25) loss of basipterygoid process; (51) dorsally flat skull table; (74) quadrate laterosphenoid contact present; (80) IX-XI cranial nerves exits through common foramen; (110) squamosal groove for the external ear flap

108 muscles ventral, placed lateral to dorsal rim of squamosal; (231) and the squamosal portion of the supratemporal rim in level with rest of skull table. The unknown character states here may be breaking up Hallopodidae.

It is in these two taxa that more disparate results are seen with the inclusion of

Tennant’s characters, as Almadasuchus and Macelognathus form a group (Hallopodidae)

(Figure 21). In the 8 analyses run 4 recover this group using only implied weights, with low support ranging from 36 to 44. This grouping is based on 5 synapomorphies: (78) one quadrate fenestra; (114) proximal head of femur with medially directed head 55% of the width; (131) the long axis of the femoral head and axis that joins the fibular and medial condyles at the distal femoral end are parallel to each other; (132) lesser trochanter present as a long ridge; and (580) the proximal development of the greater trochanter is short and lacks a distinct ridge. (Almadasuchus + Macelognathus) +

Crocodyliformes) is found to be supported by several of the unambiguous synapomorphies that support the paraphyletic assemblage discussed above.

Dibothrosuchus and Sphenosuchus are also more closely related to

Crocodyliformes than are other sphenosuchians other than Junggarsuchus and hallopodids. In other analyses Dibothrosuchus is often found closer to Junggarsuchus than Sphenosuchus (Wilberg, 2015; Leardi et al., 2017) and some of our results support this, with a few synapomorphies. Dibothrosuchus+Solidocrania is found in only 2 of the analyses with Tennant’s characters and none with them excluded (Figure 21). The support values for these nodes are very low, ranging from 22 to 33. Characters supporting this placement are not found in Sphenosuchus and include: (34) postorbital with oblique, anterolaterally facing edge so that skull roof and supratemporal fenestra narrow

109 anteriorly; (35) well developed horizontal shelf in the supratemporal fossa; (60)

Subtriangular concavity located on the posterolateral surface of the squamosal, located posteriorly to the otic shelf recess and anterolaterally from the paroccipital process; (256) postorbital bar medial or posterior to jugal; (305) edentulous posterior portion of the maxilla long; (531) ectopterygoid posterior process reduced; (141) smooth posteroventral edges of the coracoid;

The more common relationships is Dibothrosuchus and Sphenosuchus form a group that is the sister group to ‘Solidocrania’ (Figure 22). Dibothrosuchus is found to group with Sphenosuchus in all 8 of the analyses run without Tennant’s characters, even when other sphenosuchians are found in a polytomy, with support values ranging from 39 to the highest at 60 from the implied weight analysis using Gracilisuchus as the outgroup and including all taxa (Tables1-4). In the analysis with Tennant’s characters this organization was found only twice, with support values of 33. Seven characters were identified as unambiguous synapomorphies of Dibothrosuchus+Sphenosuchus: (12) a squamosal with a ridge along the supratemporal fenestra; (26) a greatly expanded basipterygoid; (94) a weakly developed dentary symphysis; (99) portion of the premaxilla posterior to nares 36-45% of length posterior; (318) presence of pterygoid-basisphenoid

‘wings’ (436) dicephalous axis rib.

Dibothrosuchus + Sphenosuchus form the sister group to Solidocrania in 3 of our

8 analyses without Tennant’s characters (the remaining ones found in a basal polytomy), with support values of 32 to 46 all with implied weights. With Tennant’s characters, in the two implied weight schemes where this relationship was found support was similarly low, at 28 to a high score of 40. Despite these low scores a number of synapomorphies

110 united Dibothrosuchus+Sphenosuchus to Solidocrania: (22) depression for antrum entering deeply into prootic, though this may be present in Kayentasuchus; (16) interparietal suture absent; (19) occipital portion of parietals straight; (77) quadrate fenestra fully bounded by quadrate; (184) posterior process of the maxilla terminates anterior to orbit; (216) frontals expand into the supratemporal fossa; (398) five premaxillary teeth;

(481) antorbital fossa rounded and tall; and (112) coracoid with elongate posteromedial process. The presence of the dorsomedial process of the quadrate was not a character in our analysis, but is also only known in these two sphenosuchians.

In addition to the above groupings, which were similar to those found by Clark et al. (2004), relationships among other ‘sphenosuchian’ species were found.

Pseudhesperosuchus, Redondaventor, Dromicosuchus, Hesperosuchus, and

Kayentasuchus are found in a basal polytomy.In a few implied weights analyses

Hesperosuchus is found as the most basal ‘sphenosuchian’ with the other three taxa in various weakly supported (scores of less than 15) topologies. Despite several derived features of the braincase similar to sphenosuchians closer to crocodyliforms we do not find Kayentasuchus as the sister taxon to Crocodyliformes and instead it falls in the poorly resolved polytomy at the base of crocodylomorphs (Figures 20, 21, 22). While it appears that even though ‘Sphenosuchia’ may not be a monophyletic group of crocodylomorphs, there may be smaller, specialized groups within it. This is seen in the relationships of Dibothrosuchus and Sphenosuchus as well as Almadasuchus and

Macelognathus.Another group that finds some support in our analysis is Terrestrisuchus

+ Litargosuchus, which in turn form a group sister to ((Sphenosuchus+Dibothrosuchus)+

111

Solidocrania). When found as a paraphyletic assemblage node support is often below 15 and never exceeds 35 (Figure 21). Positions of taxa vary widely when not found in a polytomy, though Litargosuchus is most often found as the sister to (Dibothrosuchus +

Spehnosuchus + Solidocrania) (Figure 22). These results are found in 3 of 8 analyses performed without Tennant’s characters with weak node support (14-30). They are found as a group in 4 of the 8 analyses using Tennant’s characters with weak scores ranging from 29 to 51, with stronger scores found when Gracilisuchus is the outgroup. This is one of the few groups that has better scores with the Gracilisuchus outgroup. The group of, Terrestrisuchus, and Litargosuchus is notable because these are often referred to as the more ‘gracile’ forms and may be assigned similar states due simply to size (Clark et al. 2000). Specifically, they have one unambiguous synapomorphy: and a (518) narrow postorbital bar. There are an additional 6 ambiguous synapomorphies that are found to unite these two taxa, but are also found in other sphenosuchians: (9) the dorsal surface of the frontal is flat; (18) that the medial margins of supratemporal fossae on lateral surfaces of parietals are separated on the midline by broad, flat area; (93) the dorsal portion of the posterior dentary over the mandibular fenestra is as long as the ventral portion; (171) external nares are longer than wide; and (482) supratemporal fenestra are narrow and slit like. In addition to these characters, digitgrade forelimbs, with long, slender metacarpals are considered to be shared among these taxa, while other taxa have somewhat stouter digits that may have contacted the ground differently when walking (Sereno and Wild

1992). Overall, their skeletons are all in the lower range of sizes for ‘sphenosuchains’, yet

Junggarsuchus is also small compared to some of the larger taxa like Hesperosuchus so

112 the similarity between these gracile taxa may not simply be the result of allometric differences, as proposed by some authors (Clark et al. 2000)

Other noteworthy relationships we found included the grouping of Dyoplax with

Erpetosuchus when included, which was found in 3 of the 4 analyses in which it was included (with Gracilisuchus as the outgroup) with scores ranging from 72-80 (Tables 3-

4) (Figure 21).

The inclusion of the Tennant et al, (2016) characters did not affect the relationships of sphenosuchians much with the exception of Macelognathus and

Almadasuchus, but did provide consistently different topologies within Crocodyliforms including variations in the monophyly or paraphyly of ‘Protosuchia’ and in the position of Thalattosuchia (Table, 4) (Figures 20, 21). In all of our analyses Crocodyliformes is found well supported with values from 87 to 92, with the exception of an equal weight analysis excluding Tennant’s characters with Postosuchus as the outgroup where

Thalattosuchia is found as the sister group to Crocodyliformes. In this case the

Crocodyliformes node has a low score of 24. Overall, we found 35 characters that support

Crocodyliformes that are recovered in more than one of 8 analyses (Table 5). One newly identified synapomorphy of Crocodyliformes we found was the loss of elongate blade like coronoids in all crocodyliforms with the exception of thalattosuchians. While

Crocodyliformes is found in all our analyses the monophyly of ‘Protosuchia’ varies between data sets.

When the Tennant (2016) characters were omitted, we found a weakly supported monophyletic ‘Protosuchia’ in 7 of the 8 analyses with scores ranging from 24 to 48

(Figure 22). The exceptions include a paraphyletic ‘Protosuchia’ found when implied

113 weights were used with Gracilisuchus as the outgroup. Within this monophyletic group two distinct groups form. In the first, Orthosuchus is found to be closest to the two

Protosuchus species, with ‘Edentosuchus’ and ‘Gomphosuchus’ either in a polytomy outside or the prior three taxa or in a weakly supported branching pattern. Then on the next node are (Gobiosuchus + Zaraasuchus), with support between 91 and 98. Sister to this group are the protosuchians referred to as ‘shartegosuchids’ (Clark, 2011).

Shantungosuchus and Sichuanosuchus are found to form a group with node support between 55 and 74s, while (Fruitachampsa + Nominosuchus) has support between 50 and 64 and sister to them is Zosuchus with similar supports (Figure 22). Potential synapomorphies for this monophyletic group are listed in Table 5 and include features of the braincase, palate and dentary.

With the inclusion of over 130 characters for an analysis designed to explore metasuchian relationships (Tennant, 2016) a paraphyletic ‘Protosuchia’ was found as in all 4 of the implied weight analyses, though not in any equal weight analyses. The low CI number relative to the higher RI (Tables 1 and 2) suggests that this while there are multiple convergent characters that show more than one-character change, they provide important grouping information and when convergent characters are downweighted these characters, some from the Tennant data set break up ‘Protosuchia’. ‘Edentosuchus’ is found as the most basal member of this paraphyletic protosuchian grouping, with high support for crocodyliforms (88-92). Relationships are similar between protosuchians in the monophyletic grouping even as a paraphyly and found consistently in all 4 implied weight analyses. The next node supports two groups though the support for this low, between 29 and 34. This basal group consists of the well support Protosuchus node, with

114 support of 98 to 100 and far more weakly supported ‘Gomphosuchus’ + Orthosuchus

(16-20) which are united by nodes in the high 50s and low 60s. The nextgroup,

Gobiosuchus + Zaraasuchus (96 to 98), are united with the rest of Crocodyliformes with a low score of 30. The following clades of ‘Protosuchians’ are weakly supported as well, with scores in the 20s and 40s (Figures 20,21,22). They include (Shantungosuchus +

Sichuanosuchus) and then the group sister to (Hsisosuchus + Mesoeucrocodylia),

Shartegosuchidae, which is (Zosuchus (Fruitachampsa + Nominosuchus)) in our analysis.

The other major difference between the results from the two data sets is the position of Thalattosuchia. While thalattosuchians have been found as the sister group of

Crocodyliformes in some prior studies (Wilberg, 2015) in the 16 analyses we performed

Thalattosuchia was found as the sister group of Crocodyliformes only once, with the omission of Tennant’s characters, Postosuchus as the out group and equal weights applied. The support for Thalattosuchia + Crocodyliformes is 79 and is supported by several synapomorphies (Table 5). This position also explains some of the primitive features of Thalattosuchia, (79) such as dorsally exposed prootics. This position suggests a complex origin of the crocodylian secondary palate, involving two origins of the structure and at least one loss in Protosuchia, but does explain several primitive features of the thalattosuchian braincase (Wilberg, 2015). However, the other 15 analyses consistently found Thalattosuchia within Crocodyliformes, though with lower node support. Of these 15, four of these find a polytomy among the relationships of neosuchians, thalattosuchians and notosuchians, two from analyses with Tennant’s

115 characters with equal weights and Gracilisuchus as the outgroup and the other two from equal weights without Tennant’s characters with both of the outgroup situations.

In four of the analyses without Tennant’s added characters Thalattosuchia is recovered as a basally branching group of mesoeucrocodylians, forming the sister group of (Notosuchia + Neosuchia) in our analysis (Figure 22). Three of our implied weight analyses recovered this position, the only exception being the implied weight analysis rooted on Gracilisuchus with rouge taxa removed and one equal weight analysis finds this placement, though it has the weakest support of the four, at 49 while the other positions have support scores of 60 to 66. 19 potential synapomorpheis were found to support this position (Table 5). This position, along with other placements of

Thalattosuchia in Crocodyliformes supports a simpler origin for the secondary palate (a single origin with variation among ‘protosuchians’), but also requires multiple reversals of characters, like the lack of contact between the lateropshenoid and quadrate and dorsal exposure of the prootics. The single exception mentioned above recovers Thalattosuchia in a clade termed ‘Longirostrine’ (Wilberg, 2015) where thalattosuchians are found as the sister group to long snouted neosuchians like the dyrosaurs and pholidosaurs.

The longirostrine clade is only found once, with the exclusion of the selected

Tennant characters in an analysis that omits rouge taxa, uses implied weights and has

Gracilisuchus as the outgroup. Thalattosuchia is united with (Dryosaurus (Pholidosaurus

+ Sarcosuchus)) with a score of 36 at the node. There are 6 potential synapomorphies for this group which are listed in table 5. They include: (42) a lack of nasal contribution to the nares; (104) a long symphysis of the dentary- 40% or greater; (169) rostrum twice as long as skull roof; (335) pendulous rugose tuberosities of the basioccipital present; (403)

116 over 20 teeth in the maxilla; and (405) no enlarged maxillary tooth. All of these characters, with the exception of the enlarged tooth and tubera of the basioccipital are related to the long snouts seen in the taxa present in the clade. We only find this group once and it is poorly supported. Wilberg, (2015) pointed out that this group is likely convergent as most synapomorphies they found were related to the elongated rostrum and dentary and we find support for this position here.

The final position for Thalattosuchia we found is not one of those outlined in

Wilberg (2015). In this position Thalattosuchia forms the sister group to Neosuchia

(Figure 20 and 21). This differs from the longirostrine clade as Thalattosuchia is not in

Neosuchia. We found this position in 6 of the 8 analyses performed with the inclusion of the selected Tennant characters. When not found (with equal weights and Gracilisuchus as the outgroup) Thalattosuchia fell out in a crocodyliform polytomy. The support for this position is slightly stronger than the mesoeucrocodylian position, with a range of scores from 63 to 77, with the highest support found in the Postosuchus rooted implied weight analysis. This position is supported by 17 potential synapomorphies (Table 5). This includes characters for the quadrate, skull roof, dentary, lateral temporal fenestra, ulna and osteoderms which alongside higher node support does suggest that this placement is not dependent on convergent ‘longirostrine’ characters.

DISCUSSION

The features that unite Junggarsuchus with Crocodyliformes are largely related to the reorganization of the skull, specifically strengthening it for more powerful bite force

(Clark et al. 2004). When compared to the conditions in Dibothrosuchus, they

117 demonstrate a gradual shift from the more ‘primitive’ sphenosuchian skull seen in

Dibothrosuchus to the more advanced forms closer to Crocodyliformes like

Almadasuchus. Whereas the quadrate is free ventrally in other ’sphenosuchians’ in

Junggarsuchus the otoccipital contacts the ventral part of the quadrate (character 30), as in basal Crocodyliformes and Almadasuchus. This contact with the quadrate and squamosal laterally encloses the cranioquadrate canal and this contact helps form a larger more solid occiput. Character 29 is related to this; the expansion of the basisphenoid moves the otoccipital laterally. Although the basisphenoid is expanded and filled with airspaces in Sphenosuchus and Dibothrosuchus, it is even larger in Junggarsuchus as in basal Crocodyliformes such as Protosuchus. This has caused the occipital exposure of the bone to increase in size, forming yet another buttress to the skull along the medial surface of the quadrate. The strongly arched jugal (Character 31) may be related to an overall strengthening of this part of the jaw while chewing, although no specific muscles attach here. However, the jugal differs from extant crocodylians in that it is concave below the orbit and then dramatically rises to meet the postorbital; in most species living crocodylians, the jugal is already convex and arching below the orbit. It is also thicker and more robust than in other taxa. Hence, its unique form in Junggarsuchus sloani may simply be a result of the other changes occurring throughout the skull. Fitting the arched jugal is a dorsally arched surangular unlike that seen in other sphenosuchians but comparable to those in some protosuchians, though the slope of the arch is not as steep.

This arch may be related to increased action of the jaw muscles on the mandible as the bite strength increased.

118

The contact between the otoccipitals (Character 20) and the width of the parietal on the occiput (Character 43) may be correlated with one another, as particular states are found in conjunction in both ’sphenosuchians’ and Crocodyliformes. In Junggarsuchus as well as Crocodyliformes, the otoccipitals make contact directly ventral to the supraoccipital, forming the dorsal margin of the foramen magnum. In all of these taxa, the occipital region of the parietal is reduced, becoming narrow between the squamosal and supraoccipital. With the otoccipitals closing in to reinforce the mid-occipital region of the skull, they may be constricting the other bones in that area, namely the parietal and the supraoccipital. In other’sphenosuchians’, the supraoccipital alone borders the foramen magnum. Moreover, other species also have a relatively wider occipital contribution from the parietal, which contacts the supraoccipital on a vertical median suture. The reduction of the size of the antorbital fenestra in the rostrum may also be related to the solidification of the skull. Within ‘Protosuchia’ as the secondary palate becomes more developed the antorbital fenestra becomes reduced, limiting the open space in the skull.

The anterior process of the ectopterygoid projecting along the surface of the jugal can also be interpreted as a step towards the solidification of the palate, which is not seen in

Dibothrosuchus. We also observe that the quadrate approaches, but does not contact the laterosphenoid, another important feature in the restriction of cranial kinesis in spehnosuchians. Several features of the ptyergoid are related to the strengthening of the skull. First the pneumatized nature of the pterygoid may have acted similarly to the expanded nature of the basisphenoid, helping to form an additional medial buttress to the quadrate. This inference is supported by the broad quadrate ramus of the pterygoid. The conditions above are not seen in Dibothrosuchus or Sphenosuchus and unlike them the

119 anterior process of the pterygoid ramus is firmly sutured to the pterygoid, helping to further strengthen the palate and the braincase relative to the quadrate.

Not all of the morphological shifts from more basal sphenosuchians to

Solidocrania are obviously involved in the solidification of the skull though, such as

Solidocrania synapomorphies including the presence of additional quadrate fenestra, the lack of a supratemporal fossa anterior margin posterior to the postorbital; and the presence of two large palpebrals. In addition to these cranial characters several of the changes in sphenosuchians occur in the post cranial skeleton. These are inferred to be related to an increase in cursorial locomotion seen in protosuchians and Junggarsuchus.

The first of these changes are that the anterior edge of the scapular blade is larger than the posterior, which may provide more room for the attachment of arm retractor muscles.

The radius is also slightly longer than the humerus; and the olecranon process of the ulna is very low. The glenoid surface of the coracoid is extended on a vertical plane, like

Protosuchus, which was interpreted as an adaptation that moved the arms under the body and allowed for a more cursorial lifestyle (Clark, 2004).

While the skull of Junggarsuchus had begun to resemble the strongly reinforced skull of modern crocodylians, as indicated by the features uniting it to Crocodyliformes in the phylogenetic analysis, it has a number of interesting autapomorphies not seen in any other crocodylomorph archosaur or outgroup. In the skull some of these features seems to be related to further strengthening the skull like the loss of the foramen for the eustachian tubes and laterally directed basipterygoid processes which may help to further buttress the skull. The area for the insertion of the M. pterygoideus ventralis on the lateral surface of the angular present in Junggarsuchus is not seen in other sphenosuchians and

120 is seen in lateral crocodyliforms. This expanded area allowed for increased area for the muscles of the jaw, likely related to an increase in jaw strength. However, the function of other autapomorphies is less clear, including a strange potentially additional palatine fenestration and a laterally closed opening for the caniniform tooth of the dentary. The post cranial autapomorphies may be related to its increased cursoriality. These include absence of osteoderms (121), first manus digit flexing towards digit II, facing laterally as opposed to ventrally as in other crocodylomorphs (125), and first metacarpal more slender than second metacarpal (126). The lack of osteoderms may have allowed a greater range of movement while the reduced digits are indicative of animals which are highly cursorial (Hildebrand and Goslow 2001). Other features that are related to this form of locomotion are a mammal-like ventrally or posteroventrally facing glenoid fossa; the large, perpendicularly facing humeral head, allowing the humerus to be held in a nearly-vertical position; and the flattened distal end of the ulna, which forms a straight joint along with two flattened distal carpals that place the wrist in line with the rest of the forelimb (Clark et al. 2004). Vertical zygapophyses may also be indicative of highly vertical body movements, as opposed to modern crocodylians whose locomotion has a larger horizontal element.

Though Junggarsuchus, Macelognathus and Almadasuchus show the gradual acquisition to traits in the skull related to the strengthening of the braincase, the age of these three taxa is substantially younger than the oldest Crocodyliformes. These three taxa are all from the Late Jurassic, Junggarsuchus and Almadasuchus from the Oxfordian

(Eberth et al., 2006; Leardi et al., 2017; Pol et al., 2013) while the oldest crocodyliforms are known from the late Triassic, no younger than 213 mya (Martinez et al., 2019; Kent

121 et al., 2014). As these taxa form the immediate outgroup of crocodyliforms, this creates a nearly 50-million-year long ghost lineage in both Hallopodidae and Junggarsuchus. This suggests that these changes in morphology evolved in the Late Triassic, and that

Hallopodidae and Junggarsuchus are late surviving members of the organisms in which these traits arose. The discovery of Late Triassic forms similar to Junggarsuchus or

Hallopodidae would help to support this idea (Figures 20, 21, 22).

While Dibothrosuchus is more basal than Junggarsuchus, Dibothrosuchus is not the most basal sphenosuchian and in many of our analyses is found right outside

‘Solidocrania’. In addition to the curious unique traits it possesses, such as its massively expanded prootic, Dibothrosuchus also demonstrates several important traits in the transition from even more basal sphenosuchians to ‘Solidocrania’ which may be related to the strengthening of the skull such as the absence of an inter-parietal suture and a straight occipital portion of parietals. The restriction of the quadrate fenestra to the quadrate might also help provide integrity to the structural quadrate. Some of these features are also found in Sphenosuchus and we found limited support in resampling for nodes and in synapomorphies for both Dibothrosuchus + Sphenosuchus (Figure 22)

(Table 5) and alternatively (Sphenosuchus + (Dibothrosuchus + Solidocrania)) (Figure

21).

One additional trait that may be related to the strengthening of the skull found widely in sphenosuchians is the elongate coronoid medial to the dentary. Studies in theropod dinosaurs have included the ‘supradentary’ (elongate coronoid) as one of the important factors in the mandible of tyrannosaurs that helps reinforce the mandible when biting (Hurum and Currie; 2000). This elongate coronoid has been reported widely in

122 sphenosuchians, including Junggarsuchus, Dibothrosuchus, Sphenosuchus,

Kayentasuchus and Dromicosuchus. However, it is lost in Crocodyliformes in which later forms the secondary bony palate is more complete and the braincase closer to modern forms. It is possible that this bone reinforcing the mandible was reduced or lost once the palate and braincase became more solid, however this is strictly correlative data.

Conflicting with this suggestion is that elongate coronoids are present in some thalattosuchians, which do possess an elongate secondary bony palate. The elongate coronoid is another convergent feature, in addition to the prefrontal overhang, that thalattosuchians share with Junggarsuchus. While the prefrontal overhang is not found to support the single instance of thalattosuchia as the crocodyliform outgroup, the presence of the elongate coronoid is, in addition to other more basal members of the braincase

(Table 5).

We found continued and relatively strong support (Table 3,4) for Macelognathus and Almadasuchus as the sister taxon to Crocodyliformes, supporting Leardi et al 2017 and Pol et al 2013, 2019. We find some conflicting in topology in whether these two taxa formed a monophyletic group or if Almadasuchus is the sister taxon to Crocodyliformes and Macelognathus is immediately outside this node. With the use of implied weights

Hallopodidae is retained, which suggests that the paraphyletic assemblage is affected by homoplastic theories. Other relationships of sphenosuchians remains poorly resolved.

The position of Kayentasuchus was not clarified and despite several more derived traits, like a mastoid antrum deep in the prootic, a trigeminal recess and a posteriorly closed otic recess not enough of the rest of the anatomy is known to clarify this taxon’s relationships beyond the basal polytomy of ‘Sphenosuchians’. The only other relationship we find any

123 support for in sphenosuchian is Litargosuchus + Terrestrisuchus, but the support for this group is low and based on only a few synapomorphies. This groups relationship relative to other sphenosuchians is poorly understood.

Our use of additional characters originally intended for an analysis of neosuchian

Crocodyliformes did not change sphenosuchian relationships much, as many of the characters were not applicable. However, these characters did give alternative topologies for ‘protosuchians’ and the relationships of thalattosuchians to other Crocodyliformes.

The support for Crocodyliformes is strong in both topologies (Figures 20, 21, 22). While we found Protosuchia as monophyletic in 6 of 8 analyses omitting these characters that number falls to one when the characters are included. The support found via resampling for a monophyletic ‘Protosuchia’ is very low in the 30s (Figure 20 and 21). While found less frequently, a paraphyletic ‘Protosuchia’ is found in 5 of the 18 analyses, 4 with the inclusion of Tennant’s characters. Within the paraphyletic topology ‘Edentosuchus’ is found as the most basal crocodyliform. This position is also found in the single monophyletic ‘Protosuchia’ recovered with Tennant’s characters, where ‘Edentosuchus’ is found as the most basal crocodyliform and outside Protosuchia (Figure 20). Despite these different results there are some consistent relationships found with Protosuchia, including (Orthosuchus + Protosuchus), a strongly supported Gobiosuchidae and a weakly supported Shartegosuchidae (Figures 20, 21 and 22).

We find Thalattosuchia within Crocodyliformes in the majority of our analyses

(14/16) with similar support for the two major placements of thalattosuchians within

Crocodyliformes. This position implies a single origin for the secondary bony palate, but also several reversals in the braincase to more primitive conditions such as dorsally

124 exposed prootics and a quadrate that does not contact the laterosphenoid. Without the inclusion of Tennant’s characters Thalattosuchia is found as a member of

Mesoeucrocodylia with node support in the 60s. The use of Tennant’s characters finds

Thalattosuchia as the sister group to Neosuchia with comparable node support values (in the 60s). Several of the synapomorphies that support his position are from Tennant’s characters. The synapomorphies for this group, which include elements of the quadrate, brain case, mandible, skull roof, ulna and osteoderms do not suggest that this is a position that relies on convergent ‘longirostrine’ conditions detailed by Wilberg (2015). The consistency of the placement of Thalattosuchia outside Neosuchia in comparison to the trees that exclude Tennant’s characters suggests that the characters Tennant had selected for an analysis of mesoeucrocodylians allows for better resolution of Crocodyliformes.

However, the node support for these two different positions is similar and both are supported by several synapomorphies.

CONCLUSIONS

We find that’ Solidocrania’ (Junggarsuchus + Hallopodidae + Crocodyliformes) is well supported by a number of cranial synapomorphies related to the strengthening of the skull on the way to the crocodyliform condition. Junggarsuchus exhibits derived traits shared by hallopodids and Crocodyliformes as well as autapomorphies similarly related to the skull and also related to a cursorial lifestyle. The features of the skull demonstrate the transition of the skull from more basal sphenosuchians to that of

Crocodyliformes. We also report that the elongate coronoid reported in some sphenosuchians is present widely through the group as well as in thalattosuchians, but is

125 absent in other Crocodyliformes. Dibothrosuchus is found to be closer to Junggarsuchus than other sphenosuchians, though it lacks many of the synapomorphies of Solidocrania.

Dibothrosuchus may also form a group with Sphenosuchus. We find limited support for

(Litargosuchus + Terrestrisuchus) and (Dibothosuchus + Sphenosuchus). Hallopodidae is also recovered in some of our analyses, with the use of implied weights, which suggests homoplastic traits were breaking the group apart. We do not find compelling support for a monophyletic ‘Protosuchia’ but the paraphyletic assemblage is also poorly supported. The use of implied weights appears to break up the monophyletic

‘Protosuchia’, which suggests homoplastic characters support this monophyletic group.

We also conclude that thalattosuchians are Crocodyliformes, though their exact relationships to neosuchians is unclear.

126

LITERATURE CITED

Allen D. When Terrestrisuchus gracilis reaches puberty, it becomes

Saltoposuchus connectens! Journal of Vertebrate Paleontology

supplement to 23: 29A. 2003.

Benton M J and Clark JM. Archosaur phylogeny and the relationship of the

Crocodylia; pp. 295-338 in Benton M. H, ed. The Phylogeny and Classification of

the Tetrapods, Vol. 1: Amphibians, Reptiles, Birds. Systematics Association

Special, Vol. 35A. 1988.Clarendon Press, London.

Bonaparte JF. Dos nuevas 'faunas' de reptiles Triasicos de Argentina. Gondwana

Stratigraphy (IUGS Symposium, Buenos Aires). 1969; 2:283-306.

Bonaparte J F. Los tetrapodos del sector superior de la Formacion Los Colorados,

La Rioja, Argentina. Opera Lilloana (Fund. Miguel Lillo). 1971; 22:1-183.

Bonaparte J F. Classification of the Thecodontia. Geóbios, Mémoire Spéciale.

1984; 6:99-112.

Buckley GA., Brochu, CA, Krause DW, Pol D. A pug-nosed crocodyliform from

the Late Cretaceous of Madagascar. Nature. 2000; 405 (6789): 941–944.

127

Busbey AB and Gow CE. A new protosuchian crocodile from the Upper Triassic

Elliot Formation of South Africa. Palaeontologia Africana. 1984; 25:127-

149

Butler R J, Sullivan C, Ezcurra MND, Liu J, Lecuona A, Sookias RB. New clade

of enigmatic early archosaurs yields insights into early pseudosuchian

phylogeny and the biogeography of the archosaur radiation. BMC

Evolutionary Biology. 2014; 14 (1): 128

Chapman W. An account of the fossile bones of an Alligator, found on the

seashore, near Whitby in Yorkshire. Philosophical Transactions of the

Royal Society of London. 1759; 50(2): 688-691.

Chatterjee S. Postosuchus, a new thecodontian reptile from the Triassic of Texas

and the origin of tyrannosaurus. Philosophical Transactions of the Royal Society

of London B. 1985; 309:395-460.

Choiniere JN, Clark JM, Norell MA, Xu X. Cranial Osteology of Haplocheirus

sollers. American Museum Novitates. 2010; 3816 :1-44.

Clark JM. Phylogenetic relationships of the crocodylomorph archosaurs.

Unpublished PhD Dissertation, The University of Chicago, Chicago. 1986.

128

Clark, J M. A new shartegosuchid crocodyliform from the Upper Jurassic

Morrison Formation of western Colorado. Zoological Journal of the

Linnean Society. 2011; 163(s1): S152–S172.

Clark, JM, Sues HD and Berman DS. A new specimen of Hesperosuchus

agilis from the Upper Triassic of New Mexico and the interrelationships of basal

crocodylomorph archosaurs. Journal of Vertebrate Paleontology. 2001; 20:683-

Clark JM, and Sues HD. Two new basal crocodylomorph archosaurs from the

Lower Jurassic and the monophyly of the Sphenosuchia. Zoological Journal of the

Linnean Society. 2002; 136:77-95.

Clark JM, Xu. X, Eberth DE, Forster CA, Machlus M, Hemming S, Wang Y and

Hernandez R. The Middle-to-Late Jurassic terrestrial transition: new

discoveries from the Shishugou Formation, Xinjiang, China; pp. 26-28 in 9th

International Symposium Mesozoic Terrestrial Ecosystems and Biota, P. M.

Barrett and S. E. Evans eds., Manchester, U.K. 2006

Clark JM, Xing X, Forster CA, and Wang Y A. Middle Jurassic

‘sphenosuchian’ from China and the origin of the crocodylian skull.

Nature. 2004; 430:1021-1024.

Colbert EH, A pseudosuchian reptile from Arizona. Bulletin of the American

129

Museum of Natural History. 1952; 99:561-592.

Crush PJ. A late upper Triassic sphenosuchid crocodylian from Wales.

Paleontology. 1984; 27:131-157.

Diogo, R. Comparative anatomy, homologies and evolution of mandibular, hyoid

and hypobranchial muscles of bony fish and tetrapods: a new insight.

Animal Biology. (2008); 58: 123–172

Eberth DA. et al. Sequence stratigraphy, paleoclimate patterns and vertebrate

fossil preservation in Jurassic–Cretaceous strata of the Junggar Basin,

Xinjiang Autonomous Region, People’s Republic China. Can. J. Earth

Sci. 2001; 38: 1627–1644

George ID, Holliday CM. Trigeminal nerve morphology in Alligator

mississippiensis and its significance for crocodyliform facial sensation and

svolution, The Anatomical Record. 2013; 296:670–680

Göhlich, UB, Chiappe LM, Clark JM, and Sues HD. The systematic

positions of the Late Jurassic alleged dinosaur Macelognathus (Crocodyliforma:

Sphenosuchia). Canadian Journal of Earth Sciences. 2005; 42:307-321.

130

Goloboff PA, Farris JS, Källersjö M, Oxelman B, Ramírez MJ, Szumik, CA.

Improvements to resampling measures of group support, Cladistics, 2003;

19: 324-332

Gow CE, and Kitching JW. Early Jurassic crocodilomorphs from the

Stormberg of South Africa. Neues Jahrbuch für Geologies und Paläontologie,

Monatshefte 1988:517-536.

Gradstein FM, Ogg JG, and Smith AG. A geologic time scale 2004.

University Press, Cambridge, UK, 589 pp. 2004

Gradstein, F.M., Ogg, J.G., Schmitz, M.D., and Ogg, G.M., eds, 2012, The

Geological Time Scale 2012, Amsterdam, Elsevier, 2 vols., 1144 p.

Harris JD, Lucas SG, Estep JW, Jianjun L. A new and unusual sphenosuchian

(Archosauria: Crocodylomorpha) from the Lower Jurassic Lufeng

Formation, People's Republic of China. Neues Jahrbuch für Geologie und

Paläontologie, Abhandlungen. 2000; 215 (1): 47–68.

Haughton SH. A new thecodont from the Stromberg beds. Annals of the South

African Museum. 1915; 12:98-105.

Haughton, S. H. The fauna and stratigraphy of the Stormberg Series. Annals of the

131

South African Museum. 1924. 12:323-497.

Hawkins, J. A. A survey of primary homology assessment: different botanists

perceive and define characters in different ways; pp. 22-53 in R. Scotland and R.

T. Pennington (eds.) Homology and Systematics. Taylor and Francis, London.

2000.

Hay OP. Second Bibliography and Catalogue of the Fossil Vertebrata of North

America. Vol 2. Publ. Carngie Institute, Washington 390. 1930

Hendy M and Penny D. Branch and bound algorithms to determine minimal

evolutionary trees. Mathematical Biosciences. 1982; 59:277-290.

Hildebrand M and Goslow G. Analysis of Vertebrate Structure, 5th Edition. John

Wiley and Sons. New York, New York, 2001; 660pp.

Holliday CM. and Witmer LM. The epipterygoid of crocodyliforms and its

significance for the evolution of the orbitotemporal region of

eusuchians, Journal of Vertebrate Paleontology, 2009; 29(3): 715-

733,DOI: 10.1671/039.029.0330

Huene FV. Neue Pseudosuchier und Coelurosaurier aus dem

württembergischen Keuper. Acta Zoologica, Stockholm. 1921; 2: 329-403.

132

Hurum JH, and Currie PJ. The crushing bite of tyrannosaurids. Journal of

Vertebrate Paleontology. 2000; 20(3): 619-621.

Jouve, S. A new description of the skull of Dyrosaurus phosphaticus.(Thomas,

1893) (Mesoeucrocodylia: Dyrosauridae) from the Lower Eocene of North

Africa. Canadian Journal of Earth Sciences. 2005; 42 (3): 323-337.

Jouve S. The skull of Teleosaurus cadomensis (Crocodylomorpha;

Thalattosuchia), and phylogenetic analysis of Thalattosuchia. Journal of

Vertebrate Paleontology. 2009; 29: 88-102

Kent DV, Santi-Malnis P, Colombi C, Alcober OA and Martinez RN. Age

constraints on the dispersal of dinosaurs in the Late Triassic from

magnetochronology of the Los Colorados Formation (Argentina).

Proceedings of the National Academy of Sciences of the United States of

America 111:7958–7963. 2014; doi: 10.1073/pnas.1402369111.

Kley, N.J.; Sertich, J.J.W.; Turner, A.H.; Krause, D.W.; O'Connor, P.M.; Georgi,

J.A. Craniofacial morphology of Simosuchus clarki (Crocodyliformes:

Notosuchia) from the Late Cretaceous of Madagascar. Journal of

Vertebrate Paleontology. 2010; 30 (6, Supplement): 13–98.

133

Lautenschlager S, Witmer LM, Altangerel P, Zanno LE and Rayfield EJ. Cranial

anatomy of Erlikosaurus andrewsi (Dinosauria, Therizinosauria): new

insights based on digital reconstruction. Journal of Vertebrate

Paleontology. 2014; Volume 36(6): 1263-1291

DOI: 10.1080/02724634.2014.874529

Leardi JM, Pol D, Clark JM. Detailed anatomy of the braincase of

Macelognathus vagans Marsh, 1884 (Archosauria, Crocodylomorpha)

using high resolution tomography and new insights on basal

crocodylomorph phylogeny. PeerJ5:e2801

Lecuona A, Ezcurra MD and Irmis RB. Revision of the early crocodylomorph

Trialestes romeri (Archosauria, Suchia) from the lower Upper Triassic

Ischigualasto Formation of Argentina: one of the oldest know

crocodylomorphs. Peerj. 2016; 4 (2): 585-622

Lipscomb, D.L. Parsimony, homology and the analysis of multistate characters,

Cladistics, 1992, vol. 8 (pg.45-65)

Maddison WP and Maddison DR. Mesquite: a modular system for

evolutionary analysis. 2005; Version 1.06 http://mesquiteproject.org.

134

Maisch MW, Andreas T M, Rathgeber T. Re-evaluation of the enigmatic

archosaur Dyoplax arenaceus O. Fraas, 1867 from the Schilfsandstein

(Stuttgart Formation, lower Carnian, Upper Triassic) of Stuttgart,

Germany. Neues Jahrbuch für Geologie und Paläontologie -

Abhandlungen. 2013; 267 (3): 353–362.

Marsh OC. Notice of some new vertebrate fossils. American Journal of Science.

1877; 14(3):249-256.

Marsh OC. A new order of extinct Jurassic reptiles (Macelognatha). American

Journal of Science, 1884; 27(3):341.

Martínez R N, Alcober OA, and Pol D. A new protosuchid crocodyliform

(Pseudosuchia, Crocodylomorpha) from the Norian Los Colorados

Formation, northwestern Argentina. Journal of Vertebrate Paleontology.

2019 DOI: 10.1080/02724634.2018.1491047

Meers MB. Crocodylian forelimb musculature and its relevance to Archosauria.

The Anatomical Record Part A 274A891-916.Mimics Software, https://www.materialsise.com/en/medical/software/mimics. 2003.

Nascimento PM, Hussam Z, . A new species of Baurusuchus (Crocodyliformes,

Mesoeucrocodylia) from the Upper Cretaceous of Brazil, with the first complete

135

postcranial skeleton described from the family Baurusuchidae. Papeis Avulsos de

Zoologia. 50. (2010); (21): 323-361.

Nesbitt SJ. The Early Evolution of Archosaurs: Relationships and the Origin of

Major Clades. Bulletin of the American Museum of Natural History. 2010.

Nesbitt SJ, Butler R J. Redescription of the archosaur Parringtonia gracilis from the

Middle Triassic Manda beds of Tanzania, and the antiquity of Erpetosuchida.

Geological Magazine: 2012; 1: 1-14

Nesbitt SJ, Irmis RB, Lucas SG, Hunt AP. A giant crocodylomorph from the Upper

Triassic of New Mexico. Palaeontologische Zeitschrift. 2005; 79 (4): 471–478.

Newton ET. Reptiles from the Elgin Sandstone. Description of two new genera.

Philosophical Transactions of the Royal Society of London B. 1894; 185:573-607.

Olsen P E, Sues HD., and Norell MA... First record of Erpetosuchus (Reptilia:

Archosauria) from the late Triassic of North America. Journal of Vertebrate

Paleontology. 2000; 20:633-636.

Parrish, JM. A new specimen of an early crocodylomorph (cf. Sphenosuchus sp.)

from the Upper Triassic Chinle Formation of Petrified Forest National Park.

Journal of Vertebrate Paleontology. 1991; 11:198-212.

136

Pol D, Gasparini Z. Skull anatomy of Dakosaurus andiniensis (Thalattosuchia:

Crocodylomorpha) and the phylogenetic position of Thalattosuchia, Journal of

Systematic Paleontology. 2009; 7: 163-197

Pol, D. and Norell, MA. (2004). A new gobiosuchid crocodyliform taxon from the

Cretaceous of Mongolia. American Museum Novitates 3458: 1-31.

Pol D and Norell MA. A new crocodyliform from Zos Canyon, Mongolia. American

Museum Novitates. 2004; 3445: 1-36.

Pol D and Norell MA. A new gobiosuchid crocodyliform taxon from the Cretaceous of

Mongolia. American Museum Novitates. 2004; 3458: 1-31.

Pol D, Rauhut OWM, Lecuona A., Leardi JM, Xu X, Clark JM. A new fossil from the

Jurassic of Patagonia reveals the early basicranial evolution and the origins of

Crocodyliformes. Biological Reviews. 2013; 88 (4): 862-872

Romer, A. S. 1972. The Chañares (Argentina) Triassic reptile fauna. XIII. An early

ornithosuchid pseudosuchian, Gracilisuchus stipanicicorm, gen. et sp. nov.

Brevoria 389:1-24.

137

Sereno PC, Larson H C E, Sidor CA, Gado B. The Giant Crocodyliform

Sarcosuchus from the Cretaceous of Africa. Science. 2001; 294 (5546):

1516–9.

Sereno PC and Wild R. Procompsognathus: theropod, ‘thecodont’, or both?

Journal of Vertebrate Paleontology. 1992; 12:435-458.

Simmons DJ. The non-therapsid reptiles of the Lufeng Basin, Yunnan, China.

Fieldiana, Geology. 1965; 15:1-93.

Storrs GW, Efimov MB. Mesozoic crocodyliforms of north-central Eurasia. In

Benton, Michael J.; Shishkin, Mikhail A.; Unwin, David M.; Kurochkin,

Evgenii N. The Age of Dinosaurs in Russia and Mongolia. Cambridge:

Cambridge University Press. 2000; 402–419.

Strong E and Lipscomb DL. Character coding and inapplicable data. Cladistics. 1999;

15:363-372.

Sues HD Olsen JG, Carter JG, Scott DM. A new crocodylomorph archosaur from the

Upper Triassic of North Carolina. Journal of Vertebrate

Paleontology. 2003; 23:329-343.

Swofford D. PAUP*, Phylogenetic Analysis Using Parsimony, version 4b.10

138

Smithsonian Institution. Washington, DC. 2002

Walker A D. Triassic reptiles from the Elgin area Stagonelepis, Dasygnathus and

their allies. Philosophical Transactions of the Royal Society of London B. 1961;

244:103-204.

Walker AD. A revision of the Jurassic reptile Hallopus victor (Marsh) with

remarks on the classification of crocodiles. Philosophical Transactions of the

Royal Society of London B. 1970; 257:323-372.

Walker AD. A revision of Sphenosuchus acutus Haughton, a crocodylomorph

reptile from the Elliot Formation (late Triassic or early Jurassic) of South Africa.

Philosophical Transactions of the Royal Society of London B. 1990; 330:1-120.

Wilberg EW. What's in an Outgroup? The Impact of Outgroup Choice on the

Phylogenetic Position of Thalattosuchia (Crocodylomorpha) and the

Origin of Crocodyliformes, Systematic Biology. 2015; 64 (4): 621–637.

Wilkinson, M. Coping with abundant missing entries in phylogenetic inference

using parsimony. Systematic Biology. 1995; 44:501-514.

Wu XC and S Chatterjee. Dibothrosuchus elaphros, a crocodylomorph from

the Lower Jurassic of China and the phylogeny of the Sphenosuchia. Journal of

139

Vertebrate Paleontology. 1993; 13:58-89.

Wu XC, Donald BB, Lu JC. A new species Shantungosuchus from the Lower

Cretaceous of Inner Mongolia (China) with comments on S. chuhsiensis

Young, 1961, and the phylogenetic position of the genus. Journal of

Vertebrate Paleontology. 1994; 14(2):210-229.

Wu XC, Russell A. Redescription of Turfanosuchus dabanensis (Archosauriformes) and

new information on its phylogenetic relationships. Journal of Vertebrate

Paleontology. 2001; 21 (1): 40–50

Wu XC, Liu J, Li JL, The anatomy of the first archosauriform (Diapsida) from the

terrestrial Upper Triassic of China (PDF). Vertebrata PalAsiatica. 2001; 39 (4):

251–265.

Young MT. [On a new pseudosuchian from Turfan, Sinkiang]. Reports of Paleontological

Expedition to Sinkiang (I): Permo-Triassic Vertebrate Fossils of Turfan Basin.

Memoirs of the Institute of Vertebrate Paleontology and Paleoanthropology

Academia Sinica. 1973; 10:15-37

Young MT, and Andrade MB. What is Geosaurus? Redescription of Geosaurus

giganteus (Thalattosuchia: Metriorhynchidae) from the Upper Jurassic of Bayern,

140

Germany. Zoological Journal of the Linnean Society. 2009; 157(3), 551-585.

doi:10.1111/j.1096-3642.2009.00536.x

Young MT, Schwab JA, Walsh S, Witmer LM, Herrera Y, Zanno L, Clark J, Ruebenstahl

A, Xing X, Choiniere J, Dollman K, Brusatte S. Intracranial Soft Tissue

Adaptations in the land-to-sea transition: Shift in marine crocodylomorph

(Crocodylomorpha: Thalattosuchia) vasculature and Paratympanic Sinus. ICVP

Abstract. 2019.

Zanno LE, Drymala S, Nesbitt SJ, Scheider VP. Early crocodylomorph increases top tier

predator diversity during rise of dinosaurs. Scientific Reports. 2015; 5: 9276.

141

Appendices

Appendix 1: List of characters

(1): (Leardi et al., 2017) Posterodorsal process of the premaxilla overlaps anterodorsal surface of the maxilla (0) or dorsal process of premaxilla vertical, strongly sutured to maxilla (1).

(2): (Leardi et al.,2017) Facial portion of maxilla anterior to anterior edge of antorbital fenestra equal in length or longer than portion posterior to anterior edge of fenestra (0) or shorter than posterior portion (1).

(3): (Leardi et al., 2017) Maxillae do not meet on palate (0) or meet on palate to form secondary bony palate anterior to choana (1).

(4): (Leardi et al., 2017) Jugal participates in posterior margin of antorbital fenestra (0) or is excluded by lacrimal or maxilla (1).

(5): (Leardi et al., 2017) Descending process of prefrontal absent (0) or present (1).

(6): (Leardiet al., 2017) Descending process of prefrontal does not contact palate (0) or contacts palate (1).

(7): (Leardi et al., 2017) Prefrontal not underlying anterolateral edge of frontal to a significant degree (0) or with distinct posterior process underlying frontal dorsal to orbit

(1).

(8): (Leardi et al., 2017) Postfrontal present (0) or absent (1).

(9): (Leardi et al., 2017) Dorsal surface of frontal flat (0) or with longitudinal median ridge (1).

(10): (Leardi et al., 2017) Squamosal not significantly overhanging lateral temporal region (0) or with broad lateral expansion overhanging lateral temporal region (1).

142

(11): (Leardi et al., 2017) Descending process of squamosal anterior to quadrate present

(0) or absent (1).

(12): (Leardi et al., 2017) Squamosal without ridge on dorsal surface along edge of supratemporal fossa (0) or with ridge (1).

(13): (Leardi et al., 2017) Lateral edge of squamosal without (0) or with longitudinal groove (1).

(14): (Leardi et al., 2017) Quadratojugal extending anterodorsally to contact postorbital

(0) or not contacting postorbital (1).

(15): (Leardi et al., 2017) Quadrate not in contact with prootic (0) or contacting prootic

(1).

(16): (Leardi et al., 2017) In presumed adults, parietals separate (0), interparietal suture partially obliterated (1), or interparietal suture absent (2). (Ordered).

(17): (Leardi et al., 2017) Posteroventral edge of parietals extending more than half the width of the occiput (0) or less than half the width of the occiput (1).

(18): (Leardi et al., 2017) Medial margins of supratemporal fossae on lateral surfaces of parietals separated on midline by broad, flat area (0) or by sagittal crest (which may be divided by median sulcus) (1).

(19): (Leardi et l., 2017) Occipital margin of parietals V-shaped in dorsal view (0) or straight (1).

(20): (Leardi et al., 2017) Exoccipitals broadly separated dorsal to foramen magnum (0), approaching midline without contacting (1), or contacting below supraoccipital (2).

(Ordered).

143

(21): (Leardi et al., 2017) Prootic broadly contacting anterior surface of paroccipital process (0) or not in broad contact (1).

(22): (Leardi et al., 2017) Depression for mastoid antrum: absent (0), present on lateral surface of prootic dorsal to otic capsule (1), or entering deeply into prootic or connecting with each other through supraoccipital (2). (Ordered).

(23): (Leardi et al., 2017) Depression for posterior tympanic recess: absent (0), depression posterior to fenestra ovalis on anterior surface of paroccipital process (1), or penetrating prootic and paroccipital process (2). (Ordered).

(24): (Leardi et al., 2017) Paroccipital process dorsoventrally tall and distinctly expanded distally (0) or process narrower dorsoventrally, distal end only slightly expanded (1).

(25): (Leardi et al., 2017) Basipterygoid processes of basisphenoid present (0) or absent

(1).

(26): (Leardi et al., 2017) Basipterygoid processes simple, without large cavity (0) or greatly expanded, with large cavity (1).

(27): (Leardi et al., 2017) Articular without dorsomedial projection posterior to glenoid fossa (0) or with dorsomedial projection (1).

(28): (Leardi et al., 2017) Posterior edge of maxillary and more posterior dentary teeth concave or straight (0) or distinctly convex (1).

(29): (Leardi et al., 2017) Basisphenoid body similar in size to basioccipital (0) or greatly expanded with pneumatic cavities (1).

(30): (Leardi et al., 2017) Exoccipital does not contact distal end of quadrate (0) or contacts quadrate, enclosing internal carotid artery (1)

144

(31): (Leardi et al., 2017) Jugal straight below infratemporal fenestra (0) or strongly arched dorsally (1).

(32): (Leardi et al., 2017) Ventral edge of jugal flat or convex (0) or with longitudinal concavity (1).;

(33): (Leardi et al., 2017) Lateral end of paroccipital process convex (0) or concave (1).

(34): (Leardi et al., 2017) supratemporal portion of postorbital forms squared-off anterolateral corner to rectangular skull roof (0) or postorbital with oblique, anterolaterally facing edge so that skull roof and supratemporal fenestra narrow anteriorly

(1).

(35): (Leardi et al., 2017) Horizontal shelf in posterior part of supratemporal fenestra poorly developed or absent (0) or well developed, flooring posterior ⅓ of supratemporalfossa (1).

(36): (Leardi et al., 2017) Occipital portion of parietal narrow (0) or broad (1).

(37): (Leardi et al., 2017) Surangular foramen absent (0) or present (1).

(38): (Leardi et al., 2017) M. pterygoideus posterior insertion area on angular does not extend onto lateral surface (0) or extends well onto lateral surface (1).

(39): (Leardi et al., 2017) Anterior maxillary teeth similar in size to posterior teeth (0) or much larger than posterior teeth (1).

(40): (Leardi et al., 2017) Quadrate fenestra: absent (0); present but small (1); or present but large (2). (Ordered). [Modified from Clark et al. (2004) adding character state 2].

(41): (Pol et al., 2009: Character 125). Posterodorsal process of premaxilla: absent (0), or present extending posteriorly wedging between maxilla and nasals (1)

(42): (Clark, 1994: Character 13). Nasal contribution to narial border: yes (0), or no (1).

145

(43): (Clark, 1994: Character 66). External nares: divided by a septum (0), or confluent

(1).

(44): (modified from Clark, 1994: Character 6). External nares facing anterolaterally or anteriorly (0), dorsally not separated by premaxillary bar from anterior edge of rostrum

(1), or dorsally separated by premaxillary bar (2). (Ordered).

(45): (modified from Clark, 1994: Character 3). Rostrum proportions: narrow oreinirostral (0), broad oreinirostral (1), nearly tubular (2), or platyrostral (3). (Ordered).

(46): (modified from Clark, 1994: Character 79). Maxillary tooth size variation: absent or single wave of size variation (0), or enlarged maxillary teeth curved in two waves

(festooned) (1).

(47): (Pol et al., 2009: Character 139). External surface of maxilla and premaxilla: with a single plane facing laterally (0), or with ventral region facing laterally and dorsal region facing dorsolaterally (1).

(48): (Clark, 1994: Character 67). Antorbital fenestra: as large as orbit (0), about half the diameter of the orbit (1), much smaller than the orbit (2), or absent (3). (Ordered).

(49): (modified from Clark, 1994: Character 1). External surface of dorsal cranial bones: smooth (0), slightly grooved (1) and heavily ornamented with deep pits and grooves (2).

(Ordered).

(50): (Clark, 1994: Character 21). Frontals: paired (0), unpaired (1).

(51): (Clark, 1994: Character 24). Supratemporal roof dorsal surface: complex (0), or dorsally flat? Skull table? Developed, with postorbital and squamosal with flat shelves extending laterally beyond quadrate contact (1).

146

(52): (Clark, 1994: Character 37). Palatines: do not meet on palate below the narial passage (0), form palatal shelves that do not meet (1), or meet ventrally to the narial passage, forming part of secondary palate (2). (Ordered).

(53): (modified from Clark, 1994: Character 39). Choanal opening: continuous with pterygoid ventral surface except for anterior and anterolateral borders (0), or opens into palate through a deep midline depression (choanal groove) (1).

(54): (modified from Clark, 1994: Character 42 by Ortega et al., 2000: Character 139).

Depression on primary pterygoidien palate posterior to choana: absent or moderate in size being narrower than palatine bar (0), or wider than palatine bar (1).

(55): (Clark, 1994: Character 38). Pterygoid: restricted to palate and suspensorium, joints with quadrate and basisphenoid overlapping (0), or pterygoid extends dorsally to contact laterosphenoid and form ventrolateral edge of the trigeminal foramen, strongly sutured to quadrate and laterosphenoid (1).

(56): (Clark, 1994: Character 43). Pterygoids: form posterior half of the choanal opening

(0), or completely enclose choana (1).

(57): (Clark, 1994: Character 41). Pterygoids posterior to choanae: separated (0), or fused

(1).

(58): (Pol et al., 2009: Character 133). Acute anterior process of ectopterygoid projecting along medial surface of jugal: developed (0), or reduced or absent (1).

(59): (Leardi et al., 2017). Post-temporal fenestra: large and enclosed by parietal, squamosal, and exoccipital, well separated from the supraoccipital (0), or large and enclosed by the squamosal and exoccipital, with its medial end located close to the

147 lateral edge of the supraoccipital (1), or small and with supraoccipital participating from its medial margin (2). (Ordered).

(60): (Leardi et al., 2017) Subtriangular concavity located on the posterolateral surface of the squamosal, located posteriorly to the otic shelf recess and anterolaterally from the paroccipital process: absent (0), or present (1).

(61): (Leardi et al., 2017) Squamosal contact with the posterodorsal surface of the quadrate closing posteriorly the otic recess: absent (0), or present (1).

(62): (Clark, 1994: Character 49). Quadrate, squamosal, and otoccipital: do not meet to enclose cranioquadrate passage (0), enclose passage near lateral edge of skull (1), or meet broadly lateral to the passage (2). (Ordered).

(63): (Leardi et al., 2017) Lateral margin of squamosal-postorbital along the upper temporal bar, convex or straight (0), or laterally concave (1).

(64): (Ortega et al., 2000: Character 75). Anterior opening of temporo-orbital in dorsal view exposed (0), or hidden in dorsal view and overlapped by squamosal rim of supratemporal fossa (1).

(65): (Clark, 1994 Character 52). Eustachian tubes: not enclosed between basioccipital and basisphenoid (0), or entirely enclosed (1)

(66): (Clark, 1994: Character 56). Basisphenoid: exposed on ventral surface of braincase

(0), or virtually excluded from ventral surface by pterygoid and basioccipital (1).

(67): (modified from Wu & Sues, 1996: Character 24 and Wu et al., 1997: Character

124). Jaw joint: placed at level with basioccipital condyle (0), below basioccipital condyle about level of lower toothrow (1), or below level of toothrow (2). (Ordered);

148

(68): (modified from Clark, 1994: Character 77). Splenial involvement in symphysis in ventral view: not involved (0), involved slightly in symphysis (1), or extensively involved

(2). (Ordered).

(69): (Ortega et al., 1996: Character 9). Ventral exposure of splenials: absent (0), or present (1).

(70): (modified from Clark, 1994: Character 71). Retroarticular process: very short and robust projecting ventrally (0), with an extensive rounded, wide, and flat (or slightly concave) surface projected posteroventrally and facing dorsomedially (1), posteriorly elongated, triangular-shaped and facing dorsally (2).

(71): (Leardi et al., 2017) Basisphenoid-exoccipital suture: absent (0), or interdigitated suture lateral to the lateral Eustachian foramina (1).

(72): (Leardi et al., 2017) Supraoccipital: fused with the exoccipital (0); or, as a separate ossification (1).

(73): (Leardi et al., 2017) Supraoccipital shape: narrow, being dorsoventrally taller than lateromedially wide (0), or wide, being lateromedially wider than dorsally high (1).

(74): (modified from Clark, 1986) Quadrate-Laterosphenoid contact: absent (0); or, present (1).

(75): (Leardi et l., 2017) Basioccipital recesses: absent (0); or, present as paired foramina located in a median deep depression on the ventral surface of the bone (1).

(76): (Leardi et al., 2017) Length of the posterodorsal process of the postorbital: short, not reaching the midlength of the supratemporal fenestra (0); or, long, exceeding the midlength of the supratemporal fenestra (1).

149

(77): (Leardi et al., 2017) Quadrate fenestra: with participation of the quadratojugal (0); or, exclusively bounded by the quadrate (1).

(78): (Clark, 1994) Number of quadrate fenestrae: one (0), or more than one (1).

(79): (Leardi et al., 2017) Prootic: exposed in dorsal view, on the supratemporal fenestra

(0); or not exposed in dorsal view (1).

(80): (Leardi et al., 2017) Exit of the cranial nerves IX-XI: exit the braincase ventromedially (0); or, through a common foramen on the ventromedial region of the paroccipital process (vagus foramen) (1).

(81): (modified from Clark, 1986; Sereno & Wild, 1992; Nesbitt, 2011) Anterior process of the squamosal: elongated, less than one third of the lateromedial width of the supratemporal fossa (0), transversely broad more than one third the width of the supratemporal fenestra (1); or, very broad, as wide as the width of the supratemporal fossa.

(82): (modified from Nesbitt, 2011) Lateral extent of the paroccipital processes: ends lateral to lateral border of supratemporal fenestra (including fossa) (0); or, ends medial to or at the margin of the border of the supratemporal fenestra (1).

(83): (modified from Nesbitt, 2011) Anterior end of the dentaries: tapering to a point (0); or, dorsally expanded, forming a distinct step (1).

(84): (Parrish, 1994) Anterior part of the dentary: bears teeth (0); or, edentulous (1).

(85): (modified from Clark, 1994) Ventrally opened notch on ventral edge of rostrum at maxilla-premaxilla contact: absent (0), present as a notch (1), or present as a large notch

(2), or present as a notch that is closed ventrally (or largely constrained at its ventral edge) (3)

150

(86): (Nesbitt, 2011) Nasals-frontals contact: transverse (0); or, frontals taper to a point

(1).

(87): (Weinbaum & Hugerbühler, 2007) Ectopterygoid: single head (0); or, double headed (1).

(88): (modified from Dyrmala & Zanno, 2016) Antorbital fossa: (0) well defined anterodorsally, yet not well defined along entire length of posterior process of maxilla;

(1) well defined, forming complete circumference around the antorbital fenestra.

(89): NEW- Angular narrow in ventral and dorsal view, and long in both medial and lateral views (0) or expanding in size moving posteriorly, dominating posterior section of the mandible, becoming wide in both lateral and dorsal views. (1)

(90): NEW- anterior Pterygoid processes are elongate and narrow, allowing for long narrow openings (0) or the process are broad and flat. (1)

(91): NEW Majority of Articular process is exposed medially, with little exposure above the surangular (0) or major process of the articular is exposed laterally, posteriorly, and medially, posterior to the surangular and angular. (1)

(92): NEW- Prearticular present, ventral to Surangular (0) or absent (1).

(93): NEW- Dorso-posterior process of the Dentary in the lateral view extends as far (1), or is shorter than the ventral process (2) or makes up a large part of the mandibular foramen border, and is longer that the ventral process. (0) Ordered.

(94): NEW- Symphysis of the Dentary is weak and is 1/4th the length of the overall dentary or less (1) or it is stronger and over 1/4th length.

(95): NEW- 3rd or 4th tooth Mandibular caniform (twice height of adjacent teeth) (1) or identical to others (0)

151

(96): NEW- Ventral Premaxillary, Maxillary Notch, not present (0), shallow (1) or deep

(2). Ordered.

(97): NEW- Lateral opening of manibular fenestra is circular or square shaped and makes up less than 15% of mandible length (0) or the opening is ovular and above 18% of the mandibles length (1) or closed (2). ;

(98): NEW- Palatine teeth; 0- present, 1- absent

(99): (Young, 2009): Proportion of premaxilla length posterior to nares: >67% of length posterior to external nares (0), between 50 to 65% posterior to externnares (1), 36-45% of length posterior (2) 0r less than 28% (3).

(100): (Young, 2009: Nasal-premaxilla contact, present (0), absent (1), or nasals separates the external nares in 2 (2).

(101): (Young, 2009): Tabular rostrum-no (0), yes (1)

(102): (Young, 2009): Postorbital jugal contact: postorbital medial to jugal (0) postorbital lateral to jugal (1)

(103): (Young, 2009): relative length between Squamosal and postorbital: postorbital not longer (0) postorbital longer (1)

(104): (Young, 2009): Symphysis of dentary length: short- 29% of mandible length(0), moderate, 32-38% mandible length (1) and long over 40%.

(105): (Young, 2009): Mandibular groove on dentary and surangular: shallow, far longer on dentary than surangular (0), shallow and equally long on both (1), deep and strongly developed on both elements with large foramen on each end (2) or absent (3).

(106): (Young, 2009): Dorsal part of postorbital bar: constricted, distinct from dorsal part of postorbital(0) or boradens dorsally (1)

152

(107): (Young, 2009): jugal extends anteriorly in front of pre frontals. No (0), yes (1)

(108): (Young, 2009) Posterior skull table: nonplanar, squamosal ventral to horizontal level of postorbital and parietal (0), or planar, postorbital, squamosal and parietal on the the same plane (1).

(109): (Young, 2009): Dorsal primary head of the quadrate contacts: only squamosal (0), or squamosal and or near laterosphenoid (1)

(110): (Young, 2009): Dorsal and ventral rim of the squamosal groove for external ear flap muscles. absent (0), ventral, placed lateral to dorsal (1) or ventral directly beneath dorsal (2).

(111): (Young, 2009) Squamosal contribution to the supratemporal arch: 40% or less (0) or at least 50% (1).

(112): (Leardi et al., 2017) Coracoid subcircular in lpostateral view (0), with elongate, tapering postglenoid process posteromedially (1), with extremely elongate posteromedial process (2) or with elongate ventromedial process expanded ventrally (3).

(113): (Leardi et al., 2017): Proximal ends of metacarpals overlap (0) or abut one another while overlapping (1).

(114): (Leardi, 2017): Proximal head of femur confluent with shaft (0), or with distinct, medially directed head set off from shaft, representing less than 55% the total width (1), or medially directed head and projected more than 55% the total width (2). (Ordered)

(115): (Leardi 2017): Tibia/femur length ratio: less than 1 (0), greater than 1. (1)

(116): (Leardi 2017): Anterior edge of paramedian dorsal osteoderms straight (0) or with an anterior process (1).

153

(117): (Leardi 2017): Paramedian Dorsal Osteoderm flat (0) or with distinct longitudinal bend near lateral edge (1).

(118): (Leardi 2017): Vertebral centra shallow amphicoelous (0) or procoelous (1).

(119): (Leardi 2017): Well developed hypapophysis absent on cervical vertebrae (0) or present (1).

(120): (Leardi 2017): Zygapophyses of anterior dorsal vertebrae sub vertically oriented

(0) or horizontal (1).

(121): (Leardi 2017): Osteoderms present (0) or absent (1).

(122): (Leardi 2017): Anterior and Posterior borders of scapula similar, blade relatively narrow (0), or anterior edge more strongly concave than posterior, blade much broader

(1).

(123): (Leardi 2017): Glenoid Fossa on scapula ventrally or posteroventrally oriented (0) or posterolaterally oriented (1).

(124): (Leardi 2017): On distal end of ulna, medial articulation with elongate radiale not confluent with distal articulation with ulnare (0) or articulations confluent, distal end broadly arched (1).

(125): (Leardi 2017): First manus digit faces ventrally (0) or faces laterally, flexing towards digit 2 (1).

(126): (Leardi 2017): First metacarpal similar in thickness to second metacarpal or thicker (0) or more slender (1).

(127): (Leardi 2017): Coracoid length up to 2/3rds of the scapular length (0) or subequal in length to scapula (1).

154

(128): (Leardi 2017): Radiale: not elongated (0); elongated (1); or greatly elongated, being 30% the length of the humerus or femur (2) (ordered).

(129): (Leardi 2017): Pubis distal expansion: absent (0) or present, markedly expanded distal end (1).

(130): (Leardi 2017): Pubis: forms anterior half of ventral edge of acetabulum (0) or pubis partially or completely excluded from the acetabulum by the anterior process of the ischium (1).

(131): (Leardi 2017): Long axis of the femoral head and axis that joins the fibular and medial condyles at the distal femoral end: forming an angle of 10 degrees or more (0) or parallel to each other (1).

(132): (Leardi 2017): Lesser trochanter (trochanteric crest) in the anterolateral proximal femur: present as a long ridge (0) or trochanter absent with/without distinct scar for muscle attachment (1).

(133): (Leardi 2017): Pseudointernal trochanter in the posteroproximal femur for the insertion of the M.pubo-ischio-femoralis externis muscle: absent (0) or Present (1).

(134): (Leardi 2017): Development of fourth trochanter as a wide knob (0) as a sharp ridge (1).

(135): (Leardi 2017): Development of distal condyles: fibular condyle at the same level as the medial condyle (0), or fibular condyle slightly distal than the median condyle (1), or fibular condyle further distally than the medial condyle (2).

(136): (Leardi 2017): Rows of Dorsal Osteoderms: two parallel rows (0) or more than two rows (1).

155

(137): (Leardi 2017): Postzygodiapophyseal laminae on the posterior cervical and anterior dorsal vertebrae: absent or low (0) or present as a sharp lamina delimiting a pit posterior to them on the neural arch (1).

(138): Length of the radius: shorter than the humerus (0) or longer than the humerus (1).

(139): (Leardi 2017): Proximomedial process of the radiale: absent (0) or present (1).

(140): (Leardi 2017): Acromial process of scapula: in the same plane of the proximal surface of the scapula (0) or distinctly raised, forming an abrupt step between the scapular blade and the proximal end of the scapula (1)

(141): (Leardi 2017): Coracoid posteroventral edge: smooth (0) or with a groove (1).

(142): (Leardi 2017): Scapular contribution to the glenoid: lesser than the coracoid contribution (0) or equal or greater than the coracoid contribution (1).

(143): (Leardi 2017): Humeral proximal head: confined to the proximal surface (0) or posteriorly expanded and hooked (1).

(144): (Leardi 2017): Olecranon process of the ulna: present (0) or absent or very low (1).

(145): (Leardi 2017): Proximolateral process of the ulna: located at the mid point of the proximolateral surface of the ulna (0) or anteriorly displaced, at the level of the anterior process of the ulna (1).

(146): (Leardi 2017): Distal end of ulna: anteroposteriorly compressed or rounded (0) or with anterior expansion (1).

(147): (Leardi 2017): Metacarpals 2-5 configuration: spreading (0) or compact (1).

(148): (Leardi 2017): Dorsoventrally oriented crest dorsal to the supraacetabular crest: absent (0) or present (1).

156

(149): (Leardi 2017): Preacetabular process of the ilium: short and does not extend anteriorly to the acetabulum (0) or, elongated, but shorter that the post acetabular process

(1).

(150): (Leardi 2017): Ilium orientation: mainly vertical orientation (0-20 degrees) (0) or venterolaterally deflected (1).

(151): (Leardi 2017): Ventral margin of acetabulum: convex (0) or concave (1).

(152): (Leardi 2017): Ilium dorsal margin dorsal to the supracetabular rim: round or sharp

(0) or flat (1).

(153): (Leardi 2017): Obturator foramen on the pubis: present (0) or absent (1).

(154): (Leardi 2017): Ischium medial contact with its antimere: all along its medial margin, but excluding the proximal end (0) or restricted to the medial edge of the distal part (1).

(155): (Leardi 2017): Ischium distal end: plate like (0) or rounded (1).

(156): (Leardi 2017): Femur, proximal condylar fold: absent (0) or present (1).

(157): (Leardi 2017): Fibula, proximal end: rounded or ellipticle (0) or mediolaterally compressed (1).

(158): (Leardi 2017): Pedal digit 4, number of phalanges: five (0) or four (1).

(159): (Leardi 2017): Pedal Digit 5, number of phalanges, one or more (0) or none (1).

(160): (Leardi 2017): Ectepicondylar groove: absent (0) or present (1).

(161): (Leardi 2017): Radius, proximal end, medial process: absent, giving the radius a symmetrical aspect in anterior view (0) or present, giving the radius an asymmetrical aspect in anterior view (1).

157

(162): (Andrade 2009): Anterior and Posterior margins of Scapula in lateral view: symmetrically concave in lateral view (0) or anterior edge more strongly concave than posterior edge (1).

(163): (Andrade 2009): Fourth trochanter on femur: absent (0) or present but low (1).

(164): (Andrade 2009): Fore and Hindlimb lengths: hindlimb much longer than forelimb

(0) or subequal (1).

(165): (Andrade 2009): Basal tubera: reduced (0) or large and pendulous (1).

(166): (Wilberg, 2015) Sculpture of external surface of rostrum: absent (0) or present (1)

(167): Rostral proportions at orbits: broadening gradually at orbits (0) or broadening abruptly at orbits as in Gavialis gangeticus (1).

(168): Rostral length measured from anterior orbital edge to anterior contour of rostrum: equal to or longer than remainder of skull as measured to the posterior end of the quadrate (0) or shorter than remainder of skull (1).

(169): Rostral length measured from anterior orbital edge to anterior contour of rostrum: equal to or slightly longer than distance from anterior orbital edge to posterior parietal contour (0) or at least twice the distance from anterior orbital edge to posterior parietal contour (1).

(170): Lateral contour of snout in dorsal view: straight or gently convex (0) or sinusoidal

(1).

(171): Shape of external nares in dorsal view: wider than long (0), subequal (1) or longer than wide as in Cricosaurus suevicus (2).

(172): Notch in premaxilla on lateral edge of external nares: absent (0) or present on dorsal half (1).

158

(173): Premaxilla contribution to internarial bar: forming at least ventral half (0) or little, if any, contribution (1).

(174): Maximal width of premaxillae relative to maximal width of rostrum at level of fourth or fifth maxillary alveoli: premaxilla narrower (0) or premaxilla broader as in

Sarcosuchus imperator (1).

(175): Premaxillo-maxillary suture direction in lateral view: vertically directed (0) or posterodorsally directed (1).

(176): Premaxillo-maxillary suture shape in lateral view: straight (0) or zigzag shaped

(1).

(177): Premaxillo-maxillary suture direction in palatal view: anteriorly directed (0), sinusoidal, posteromedially directed on its lateral half and anteromedially directed along its medial region (1), posteriorly directed (2) or perpendicular to the longitudinal axis of the skull (3).

(178): Foramen at premaxillo-maxillary suture in lateral surface (not for large mandibular teeth): absent (0) or present as in Simosuchus clarki (1).

(179): Ventral edge of maxilla in lateral view: straight or convex (0) or sinusoidal (1).

(180): Maxillary fossa on posterolateral surface of maxilla: absent (0) or present as in

Sphagesaurus huenei and Terminonaris robusta (1).

(181): Groove along lateral margin of maxilla dorsal to toothrow: absent (0) or present

(1).

(182): Large and aligned neurovascular foramina on lateral maxillary surface: absent (0) or present (1).

159

(183): Position of anterior portion of maxillary tooth row in relation to dentary tooth row: adjacent to (0) or anterior to orbit (1).

(184): Posterior extent of posterior process of maxilla: terminating posterior to anterior margin or orbit (0) anterior to orbit (1).

(185): Lacrimal contribution to dorsal margin of antorbital fenestra: present (0) or absent- participates in the posterior margin only (1).

(186): Nasal lateral border near premaxilla/maxilla/nasal junction: anterior portion of nasal laterally concave posterior to external nares (nasals may send a small lateral process between maxilla and premaxilla as in Orthosuchus stormbergi) (0) or premaxilla-maxilla suture straight, continuous with the nasal-maxilla suture (1).

(187): Nasal contact with lacrimals: nasal extensively contacts lacrimal (0) lacrimo-nasal contact excluded (or very nearly) by anterior projection of prefrontal meeting posterior projection of maxilla as in Orthosuchus stormbergi (1).

(188): Lacrimal contact with nasal (inapplicable in taxa without extensive contact): contacting nasal along medial edge only (0) or along medial and anterior edges (1).

(189): Nasal orientation at posterior border: nasals converge at sagittal plane posteriorly

(0) or nasals separated posteriorly by and anterior sagittal projection of forntal (1).

(190): Distance between the posterior processes of nasals relative to the distance from the posterior process of the nasal to the anterior margin of the supratemporal fossa: much shorter (0) or nearly as long (1).

(191): Nasal bones: paired (0) or partially or completely fused as in Dyrosaurus phosphaticus (1).

160

(192): Maximal width of the nasals relative to the minimal width of the nsout in dorsal view: narrower than or nearly as wide (0), wider than (1) or more than twice as wide (2).

(193): Posterolateral regions of nasals: flat surface facing dorsally (0) or lateral region deflected ventrally, forming part of the lateral surface of the snout as in Metriorhynchus superciliosus (1).

(194): Midline depression at contact between anterior process of frontal and nasals extending as a groove between nasals: absent (0) or present as in Metriorhynchus supercilious (1).

(195): Anterior extent of lacrimal relative to anterior margin or antorbital fenestra: does not exceed (0) or exceeds (1).

(196): Anterior extent of anterior process of prefrontal relative to posterior margin of antorbital fenestra: reaches or exceeds (0) or remains posterior (1).

(197): Anterior extent of anterior process of jugal relative to anterior extent of lacrimal: does not exceed (0) or exceeds (1).

(198): Anterior process of jugal contacting nasal, separating maxilla from lacrimal: absent (0) or present as in Terminonaris robusta (1).

(199): Prefrontal contact with nasal: along medial edge only (0) or penetrates the nasal anteriorly, separating the nasal into posteromedial and a posterolateral (or posteroventrolateral) processes as in Metriorhynchus superciliosus (1).

(200): Anterior process of prefrontal relative to anterior process of lacrimal: absent (0) or exceeds anteriorly (1).

(201): Nasal-prefrontal suture with a pronounced, rectangular concavity: absent (0) or present as in Eoneustes gaudryi (1).

161

(202): Prefrontal medial extent: prefrontals do not meet at midline (0) or prefrontals meet

(or very nearly meet) at midline, excluding frontal contact with nasals as in

Pissarachampsa (1).

(203): Posterior extent of posterior process of prefrontal relative to the anterior margin of the supratemporal fossa: does not reach (0) or reaches or very nearly reaches (1).

(204): Prefrontal anterior process: two anterior processes, one anterodorsal and one anteroventral, separated by posterodorsal process of lacrimal (0), single short anterior process (shorter than or as long as the orbit) (1), or single long anterior process (longer than the orbit) (2).

(205): Anterior extent of anterior process of frontal relative to anterior process of lacrimal: much shorter than lacrimal (0) or reaches or exceeds anterior extent of lacrimal

(1).

(206): Sculpturing on postorbital and squamosal when parietal and frontal are ornamented (inapplicable in taxa lacking ornamentation on frontal and parietal): absent

(0) or present (1).)

(207): Supratemporal fenestra: present (0) or reduced to a thin slit or absent as in

Gobiosuchus (1).

(208): Supratemporal fenestra size relative to orbit: smaller or nearly the same size as orbit (0), larger than orbit, but less than twice as long than wide (1) or larger than orbit, but nearly twice as long than wide (2).

(209): Cranial table width relative to ventral portion of skull: nearly as wide as ventral portion of the skull (0) or narrower than ventral portion (1).

162

(210): Wide frontal plate in the anteromedial corner of the supratemporal fossa: absent

(0) or present (1).

(211): (ordered) Supratemporal fossa, anterior margin in dorsal view: anterior margin posterior to the postorbital (0), anterior margin reaches between the anterior and posterior points of the frontal-postorbital suture (1) or reaches at least as far anteriorly as the postorbital (2). Ordered

(212): supratemporal fossae, shape, anteroposterior and lateromedial axes: longitudinally ellipsoid/sub-rectangular (anteroposterior axis more than 10% longer than lateromedial axis (0), sub-square/sub-circular (anteroposterior and lateromedial axes subequal, plus or minus 5%) (1) or transversally ellipsoid/subrectangle (lateromedial axis more than 10% longOrdereder than anteroposterior axis (2).

(213): Supratemporal fossae, shape, parallelogram (lateral and medial margins, and anterior and posterior margins are sub-parallel-anterior and posterior margins swept back, being directed slightly posterolaterally): no (0) or yes, as in Machimosaurus hugii,

Steneosaurus leedsi (1)

(214): Supratemporal fossa, in dorsal view, posterior limit: terminates well before the posterior-most point of the parietal (0), either terminates near the posterior-most of the parietal or exceeds it, but never reaches the supraoccipital (1) or more posterior than intertemporal bar (2): Ordered.

(215): Posterior edge of Supratemporal fenestra: thin (with fossa extending to posterior limit-thin ridge) (0), thin, but not a narrow ridge (no posterior extension of fossa) (1) or thick (2).

163

(216): Frontal contribution to supratemporal fossa: excluded or nearly excluded from supratemporal fossa (0) or extends well into supratemporal fossa (1).

(217): Angle between posteromedial process (interfenestral bar) and lateral process of frontal (posterodorsal margin of orbit) in dorsal view: nearly 90 degrees (0) or much less than 90 degrees (1).

(218): Frontal- postorbital suture on the skull table (anterior to the supratemporal fenestra): straight or irregular (0) or V shaped, frontal taper laterally, sending a lateral process within the postorbital skull table (1).

(219): Anterolaterally directed ridges on frontal following postorbital/frontal suture, joining with midline frontal ridge posteriorly forming posteriorly pointing arrow sharp: absent (0) or present as in Pissarachampsa (1).

(220): Sagittal crest shape: narrow, but similar in height along entire length and dorsally flat as in Steneosaurus bollensis (0) narrow, but of uniform width with distinct medial groove as in Metriorhynchus superciliosus (1), narrows abruptly posteriorly at frontal parietal suture and is dorsoventrally expanded as in Suchodus brachyrhynchus (2) or broadens posteriorly- parietal portion is broader than frontal portion as in Dakosaurus andiniensis (3).

(221): Width of anterior and posterior portions of interfenestral bar: uniform- anterior and posterior portions approximately same width (0) or anterior portion (frontal) much wider than posterior portion (parietal) (1).

(222): Parieto-postorbital suture on dorsal skull roof: absent from skull roof and supratemporal fossa (0), absent from dorsal surface of skull roof, but broadly present

164 within supratemporal fossa (1) or present on dorsal surface of skull roof and within supratemporal fossa (2).

(223): Long, thin anterior process of parietal wedging between frontal and laterosphenoid in supratemporal fossa: absent (0) or present- participates to the anteroventral margin of the supratemporal fenestra, below the frontal within the fenestra (1).

(224): Posterior margin of the parietal in dorsal view: relatively straight or gently concave (0) or with pronounced, deep concavity opening posteriorly as in Cricosaurus elegans (1).

(225): Supraoccipital exposure on cranial roof: absent- parietals contact on occiput preventing dorsal exposure of supraoccipital (0) or present- clearly exposed on dorsal surface of cranial roof (1).

(226): Dorsal part of the postorbital: with anterior and lateral edges only (0), with anteriorly facing edge so that skull roof and supratemporal fenestrae narrow anteriorly- not for articulation with palpebral (1) or with anterolaterally facing edge for articulation with palpebral as in Notosuchus terrestris (2).

(227): Posterolateral projections of squamosal (squamosal prongs): absent- posterior edge of squamosal nearly flat (0) or present- posterolateral edge of squamosal extending posteriorly as an elongate process (1).

(228): Posterolateral edge of squamosal: without descending ornamented process (0) or with descending ornamented process as in Sichuanosuchus and Fruitachampsa (1).

(229): Posterior extent of squamosal relative to quadrate condyle in lateral view: squamosal terminates anterior to the quadrate condyle (0), reaches quadrate condyle (1) or extends far as posterior to the quadrate condyle (2).

165

(230): Posterolaterally directed facet on posterolateral margin of squamosal: absent (0) or present as in Metriorhynchus superciliosus (1).

(231): Dorsoventral height of squamosal portion of lateral rim of supratemporal fenestra with respect to interefenstral bar: at smale level as skull table (0), slightly deflected ventrally as in Steneosaurus bollensis (1) or strongly deflected vnetrally as in Cricosaurus suevicus (2): Ordered.

(232): Orbit orientation: more circular in lateral aspect (0) or more circular in dorsal aspect (1).

(233): Sclerotic ossicles: absent (0) or present (1).

(234): Lateral border of orbit relative to lateral border of supratemporal fenestra: lateral to (0) or medial to (1).

(235): Prefrontal and lacrimal around orbits: forming flat rims- flush with external surface of skull (0) or envaginated- forming elevated rims as in Gavialis gangeticus (1)

(236): Prefrontal pillars when integrated in palate (inapplicable in taxa lacking a descending process of prefrontal or lacking contact between prefrontal pillar and palate): pillars transversely expanded (0), pillars transversely expanded in their dorsal half and columnar ventrally (1) or pillars longitudinally expanded in their dorsal part and columnar ventrally (1).

(237): Lacrimal orbital contour: facing laterally (0) or facing laterodorsally (1).

(238): Lacrimal in dorsal view: visible (0) or not visible (1).

(239): Ventral portion of the lacrimal- contact with jugal: extending ventroposteriorly widely contacting the jugal (0) or tapering ventroposteriorly, does not contact or contacts the jugal only slightly as in Mariliasuchus amarali (1).

166

(240): Anterior extension of the jugal relative to the anterior margin of the orbit: does not exceed the anterior margin of the orbit (0), exceeds the anterior margin but length

(measured from anterior margin of orbit to anterior tip of jugal) is less than that the orbital length (1) or greatly exceeds anterior margin such that the anterior process of jugal is as long or longer than the orbital length (1).

(241): Dorsoventral height of antorbital region of the jugal with respect to infraorbital region: equal or narrower (0) antorbital region greatly expanded (150% or more than minimal height of the jugal below the orbit) and ins Sphagesaurus huenei (1).

(242): Lateral surface of anterior process of jugal: flat or convex (0) or with broad shelf ventral to the orbit with triangular depression beneath it as in Sphagesaurus huenei (1)

(243): Longitudinal ridge on lateral surface of jugal below infratemporal fenestra: absent

(0) or present as in Zaraasuchus shepardi (1).

(244): Elongate neurovascular groove on lateral surface of jugal beneath orbit: absent (0) or present as in Steneosaurus brevior (1).

(245): Jugal participation in ventral (or lateral) margin or orbit: jugal broadly participates in orbital margin (0) of jugal excluded or nearly excluded from the orbit by lacrimal- postorbital contact as in Platysuchus multiscrobiculatus (1).

(246): Prefrontal-maxilla contact in the inner anteromedial region of orbit: absent (0) or present as in Sphagesaurus (1).

(247): Lateral margin of prefrontal relative to dorsal margin of the orbit: continuous with, not laterally expanded (0) or laterally expanded, forming a "prefrontal overhang" over the orbit as in Metriorhynchus superciliosus (1)1

167

(248): Prefrontal "overhang" over orbit (inapplicable in taxa lacking overhang): very slight- approximately 5-10% of its width (0) or greatly enlarged >10% (1).

(249): Shape of lateral margin of prefrontal overhang in dorsal view (inapplicable in taxa lacking a prefrontal overhang): gently curved (obtuse angle) with posterior margin anterolaterally directed (0), gently curved with posterior margin directed laterally nearly perpendicular to sagittal plane as in Enaliosuchus macrospondylus (1), with distinct point formed with posterior margin directed anterolaterally as in Metriorhynchus casamiquelai (2) or with distinct point and posterior margin directed laterally nearly perpendicular to sagittal plane as in Dakosaurus andiniensis (3).

(250): Lateral extent or prefrontal overhang relative to the posterolateral corner of the supratemporal fossa in dorsal view: prefrontal does reach the same plane laterally as the posterolateral corner of the supratemporal fossa (0) prefrontal reaches or exceeds laterally than the posterolateral corner of the supratemporal fossa (1).

(251): Prefrontal/lacrimal suture raised, forming an anteroposteriorly directed ridge: absent (0) or present (1).

(252): Elements contributing to medial margin of the orbit: primarily frontal (0) or prefrontal contributes 50% or greater, reducing frontal contribution (1).

(253): Anterior process of the frontal relative to anterior margin of orbit: extending far anterior (0) or slightly anterior, subequal, or posterior (1).

(254): Palpebral elements: one small palpebral present (0), two large palpebral present

(1), one large palpebral present (2) or palpebrals absent (3).

(255): Palpebrals contact with frontal: separated from the lateral edge of the frontals (0) or extensively sutured to each other and to the lateral margin of the frontal (1).

168

(256): Postorbital orientation relative to jugal on postorbital bar: anterior to jugal (0), medial or posterior to jugal (1) or lateral to jugal (2).

(257): Sculpturing of postorbital portion of postorbital bar (when sculpturing present on skull): present (0) or absent (1).

(258): Postorbital bar shape when postorbital lies medial to jugal (inapplicable in taxa where postorbital lies anterior to or lateral to jugal): transversely flattened (0) or columnar (1).

(259): Vascular opening on lateral surface of the postorbital in the dorsal portion of postorbital bar: absent (0) or present (1).

(260): Ventral portion of postorbital bar: flush with lateral surface of jugal (0), slightly medially displaced (1) or medially displaced and a ridge separates postorbital bar from lateral surface of jugal (2).

(261): Orientation of the base of the postorbital bar: directed posterodorsally (0), directed dorsally (1) or directed anterodorsally (2).

(262): Postorbital bar: inset from the dorsolateral margin of the postorbital bar (0) or lateral surface of postorbital continuous with postorbital bar-- bar not inset (1).

(263): Postorbital bar orientation in dorsal view: ventrolaterally oriented- visible in dorsal view (0) or vertical- not visible in dorsal view (1).

(264): External surface of ascending process of jugal (inapplicable in taxa with postorbital forming external surface of postorbital bar as in Teleosaurus cadomensis): exposed laterally (0) or exposed posterolaterally as in Gobisuchus (1).

169

(265): Postorbital participation to posterodorsal (in lateral view) or posteromedial (in dorsal view) margin of orbit: postorbital in excluded from the orbit posterodorsal margin

(0) or postorbital reaches the orbit posterodorsal margin (1).

(266): Postorbital participation in posteroventral (in lateral view) or posterolateral (in dorsal view) margin of orbit: postorbital does not contribute to the posteroventral orbital margin (0) or postorbital reaches the orbits posteroventral margin, forming part of the ventral margin of the orbit (1).

(267): Anterolateral process of the postorbital: absent (0), small (1), extensive, contacting or nearly contacting the dorsal margin of the jugal as in Rhabdognathus aslerensis (2).

Ordered

(268): Ectopterygoid- postorbital contact (inapplicable for thalattosuchia):absent- ectopterygoid does not contact postorbital (0) or ectopterygoid contacts postorbital on medial side of postorbital bar (1).

(269): Infratemporal fenestra length: anteroposteriorly shorter than dorsoventral height or subequal (0) elongated, approximately twice as long as deep (1).

(270): Infratemporal fenestra orientation: facing laterally (0) or facing dorsolaterally (1).

(271): Jugal shape ventral to infratemporal fenestra: transversely flattened (0) or rod like

(1).

(272): Length of posterior process of jugal relative to anterior process: equal in length or longer (0), shorter, but greater than 50% of the length of the anterior process (1) or much shorter, less than 50% of the length of the anterior process (2).

170

(273): Posterior limit of posterior process of the jugal relative to infratemporal fenestra: exceeds the posterior border of the infratemporal fenestrae (0) or terminates anterior to or reaches posterior border of the infratemporal fenestra (1).

(274): Infratemporal fenestra length relative to supratemporal fenestra: much shorter than supratemporal fenestra: much shorter than supratemporal fenestra (0), subequal (1) or longer than supratemporal fenestra (2).

(275): Quadratojugal dorsal process contact with postorbital (inapplicable in taxa lacking quadratojugal/postorbital contact): narrow, contacting only small part of postorbital (0) or broad, extensively contacting postorbital and greatly reducing size of infratemporal fenestra as in Gobisuchus (1).

(276): Jugal-quadratojugal suture relative to posterior corner of the infratemporal fenestra in lateral view: jugal-quadratojugal suture lies at posteroventral corner (0) or quadratojugal extends anteriorly forming part of ventral edge of infratemporal bar (1).

(277): Posteroventral extent of quadratojugal: reaches the quadrate condyle (0) or terminates prior to reaching the quadrate condyle (1).

(278): Posterolateral end of quadratojugal: acute or rounded, tightly overlapping the quadrate (0) or with sinusoidal ventral edge and wide and rounded posterior edge slightly overhanging the lateral surface of the quadrate as in Zosuchus, Sichuanosuchus (1).

(279): Quadratojugal spine at posterior margin of infratemporal fenestrae: absent (0) or present (1).

(280): Palatal part of premaxillae: not in contact posterior to incisive foramen (0), in contact posterioly along contact with maxillae (1), in contact along entire length due to lack of incisive foramen as in Gobisuchus (2).

171

(281): Incisive foramen position: completely separated from premaxillary tooth row, at the level of the second or third alveolus (0) or abuts premaxillary tooth row (1).

(282): Palatal branches of maxillae: not in contact in palate (0), posterior portion not in contact on palate at sutures with palatines (1) or in contact for entire length (2). Ordered

(283): Paired foramina on palatal surface of the premaxilla-maxilla suture (not pits for dentary teeth): absent (0) or present as in Simosuchus clarki (1).

(284): Sculpturing of palatal surface of maxilla: maxillary palatal surface smooth (0), maxillary palatal surface ornamented with ridges as in Protosuchus richardsoni (1) or maxillary palatal surface ornamented with pits as in Kayentasuchus (2).

(285): Ornamentation of palatal surface of palatines: absent- smooth (0) or present- pitted as in Fruitachampsa callisoni (1).

(286): Palatine antermedial process anterior extent (inapplicable in taxa lacking midline contact between maxilla and palatine): exceeds the anterior margin of the suborbital fenestra (0) or terminates posterior to the anterior margin of the suborbital fenestrae as in

Striarosuchus (1).

(287): Shape of anterior process of palatine (maxilla-palatine suture) near midline on palatal surface (inapplicable in taxa lacking midline contact between maxilla and palatine): palatine tapers anteriorly (rounded or pointed) (0), palatine anteromedially straight, perpendicular to the longitudinal axis of the skull (1) or palatine invaginated (2).

(288): Palatine, anterior margin has two distinct non-midline anterior processes: absent

(0) or present as in Pelagosaurus typus (1).

172

(289): Palatomaxillary foramina (small anteriorly directed foramina on either side of midline at palatine/maxilla suture, opening into a ventral maxillary groove): absent (0) or

Present as in Pelagosaurus typus (1).

(290): Palatomaxillary groove extent (inapplicable in taxa lacking palatomaxillary foramina): restricted to maxillae (0) or extend anteriorly onto maxillae and caudally onto surface of palatines (1).

(291): Maxillo-palatal fenestrae (moderately sized fenestrae opening ventrally from maxilla/palatine suture on palate, connecting nasal cavity with oral cavity): absent (0) or present as in Notosuchus terrestris (1).

(292): Foramina on palatine ventral surface: absent (0) or present-foramina lying along grooves on either side of midline as in Pissarachampsa (1).

(293): Orientation of posterior region of palatines (inapplicable in taxa lacking midline contact between palatines): run parasagittaly along midline (0) or palatines diverge posterolaterally becoming rodlike caudally forming palatine bars as in Notosuchus terrestris (1).

(294): Elements contributing to anterior margin of choanae: vomers and maxillae (0), maxillae only (1), palatines only (2), pterygoids with small participation of palatine (3) or pterygoids only (4). Ordered.

(295): Palatine portion of anterior margin of the choanal opening (inapplicable in taxa lacking a palatine contribution to anterior border): gently rounded (0), tapers anteriorly between the palatines as in Pelagosaurus typus (1), w-shaped- with short midline posterior process of palatines forming middle portion of "w" as in Metriorhynchus leedsi

173

(2) or with short posterior processes (rugosities) on either side of midlines as in

Machimosaurus hugii (3).

(296): Internal choanal opening: opening posteriorly and continuous wither pterygoid surface (0) or closed posteriorly by an elevated wall formed by the pterygoids (1).

(297): Internal choanae groove: undivided (0), partially septated (1) or completely septated (2).

(298): Internal choanal septum shape (inapplicable in taxa lacking a choanal septum): narrow vertical bony sheet (0) or T-shaped bar expanded ventrally (1).

(299): Flat ventral surface of internal choanal septum (inapplicable in taxa lacking a choanal septum): uniform width (parallel sided) (0), tapering anteriorly as in

Araripesuchus gomesii (1) or expanding anteriorly as in Mahajangasuchus insignis (2).

(300): Anteroposterior position of posterior margin of choana relative to posterior edge of suborbital fenestra: anterior to (0) or posterior to (1).

(301): Anteroposterior position of the anterior margin of choanae relative to suborbital fenestra: anterior to (0) or posterior to (1).

(302): Palatine contribution to suborbital fenestra border: participates in suborbital fenestra (0) or entirely excluded from suborbital fenestra (1).

(303): Anterior extent of suborbital fenestrae relative to anterior border of orbit: end anteriorly at level of anterior border of orbit (0) extend further anteriorly (1) or terminates posterior to the anterior border of the orbit (2).

(304): Vomer contribution to secondary palate (does not include palatal exposure due to lack of palatine or maxillary palatal processes meeting at midline): vomer contributes

174 flattened plate to secondary palate as in Simosuchus clarki (0) or vomers forms no contribution to secondary palate (1).

(305): Edentulous portion of posterior process of ventral lamina of maxilla: short (0) or long (room for at least 2 additional posterior maxillary teeth) (1).

(306): Ectopterygoid-maxilla contact: absent-ectopterygoid does not contact palatal branch of maxilla (0) or present (1).

(307): Ectopterygoid contact with palatal branch of maxilla (inapplicable in taxa lacking ectopterygoid-maxilla contact): very slight contact (0), extensive contact- suture mediolaterally oriented (perpendicular to sagittal plane) (1), extensive contact- suture primarily oriented anteromedially (2) or contact- suture oriented anterolaterally as in

Teleosaurus cadomensis (3).

(308): Ectopterygoid medial process: single (0) or forked (1).

(309): Ectoperygoid pneumaticity: absent (0) or present (ectopterygoid foramen/foramina) as in Pissarrachampsa (1).

(310): Ectopterygoid projecting medially on ventral surface of pterygoid flanges: barely extended (0) or widely extended covering approximately the lateral half of the ventral surface of the pterygoid flanges as in Notosuchus terrestris, Dyrosaurus phosphaticus (1).

(311): Sculpturing of palatal surface of pterygoid: absent (0) or present as in Protosuchus richardsoni (1).;

(312): Pterygoids between basisphenoid and choana: separated- not in contact along midline on palatal surface as in Postosuchus kirkpatricki (0) or in contact along midline

(1).

175

(313): Pterygoid pneumatisation: absent- pterygoid thin, sheet-like (0) or present as in

Gobisuchus (1).

(314): Posterior extent of posteromedial process of pterygoid relative to medial

Eustachian foramen: terminates anterior to the level of the medial eustachian foramen (0) or reaches same level as the medial eustachian foramen (1).

(315): Anteroposterior position of posterolateral margin of the pterygoid (torus transiliens) relative to medial eustachian foramen (0), reaches approximately the same anteroposterior level as the medial eustachian foramen (1) or terminates far posterior to the medial eustachian foramen (2).

(316): Shape of quadrate ramus of pterygoid in ventral view: narrow and elongate (0), broad in ventral view (1) or narrow and very short in ventral view (2).

(317): Long anterior process of pterygoids that contact the maxillae anteromedial to primary choanae: absent (0) or present as in Calsoyasuchus (1).

(318): In ventral view, narrow flanges of pterygoid extend posterolaterally along lateral margin of basisphenoid forming small posterolateral pterygoid-basisphenoid "wings" separating basisphenoid from contact with quadrate laterally: absent (0) or present as in

Stenosaurus bollensis (1).

(319): Palatine-pterygoid contact on palate: palatines overlie pterygoids as in Protosuchus richardsoni (0) or palatines firmly sutured to pterygoids (1).

(320): Posteriorly facing notch between the base of the pterygoid wings: absent (0) or present (1).

176

(321): Pterygoid participation in posterior border of suborbital fenestrae: present (0) or absent- excluded by posterolateral processes of palatines or medial expansion of ectopterygoids (1).

(322): Anterior process of pterygoid ramus of quadrate contact with pterygoid: not sutured (0) or firmly sutured (1).

(323): Infratemporal fenestra in ventral view: largely hidden by the pterygoid flange (0) or largely visible lateral to the pterygoid flange (1).

(324): Parietal width on occipital surface relative to supraoccipital width: widely exposed- much wider than the supraoccipital (0) or minor occipital portion- similar in size to supraoccipital (1).

(325): Posterodorsal margin of the skull roof in occipital view: relatively flat surface or gently convex (0), sigmoidal, strongly W-shaped-dorsal margin of the supraoccipital is much higher than the dorsal margin of the squamosal as in Metriorhynchus superciliosus

(1), V-shaped- dorsal margin of supraoccipital at the midlines lies ventral to dorsal margin of squamosal as in Mahajangasuchus insignis (2) or strongly convex as in

Dyrosaurus phosphaticus (3).

(326): Supraoccipital shape: more or less triangular (0) or pentagonal as in Sphenosuchus acutus (1).

(327): Paroccipital process contact with squamosal: in loose contact with squamosal laterally (0), paroccipital process laterally narrow and sutured to squamosal (1) or paroccipital process very deep dorsoventrally, interlocked with squamosal as in

Dibothrosuchus elaphros (2).

177

(32): Unsculpted ventral projection of the squamosal enclosing the dorsal half of the paraoccipital process: absent (0) or present as in Rhabdoghathus aslerensis (1).

(329): Bilateral posterior prominences on posterior surface of exoccipitals: absent- relatively flat (0) or present as in Dyrosaurus phosphaticus (1).

(330): Ordered: Foramen for the internal carotid artery (inapplicable in taxa lacking contact between quadrate and exoccipital to enclose carotid artery): small, similar in size to the openings for cranial nerves IX-XI (0) or slightly enlarged, larger than other foramina on occipital surface as in Pelagosaurus typus (1) or extremely enlarged, more than twice as large as other foramina as in M.superciliosus (2).

(331): Medioventral projection of exoccipital in relation to ventral limit of basioccipital in occipital view: terminates well dorsal (0) or nearly reaches (1).

(332): Paroccipital process, orientation in occipital view: horizontal (0), dorsolaterally orientated, at a 45 degree angle as in Cricosaurus suevicus (1), ventral-edge horizontal, then terminal third sharply inclined dorsolaterally at a 45 degree angle as in Dakosaurus andiniensis (2) or prominently arched ventrally as in Dyrosaurus phosphaticus (3).

(333): Exoccipitals contribution to occipital condyle: slight contribution (0) or large contribution as in Dyrosaurus phosphaticus (1).

(334): Orientation of occipital condyle: posteriorly directed (0) or posteroventrally directed as in Notosuchus (1).

(335): Ventral portion of basioccipital in occipital view: thin, without well-developed bilateral tuberosities (0), ventral portion anteroposteriorly thick, rugous, with pendulous tubera formed primarily from basioccipital (1) or pendulous tubera with large contribution from exoccipitals as in Rhabdognathus aslerensis (2).

178

(336): Basioccipital with laterally directed knobs: absent or slightly developed (0) or strongly developed as in Lomasuchus palpebrosus (1).

(337): Ventral projection of the basioccipital in occipital view: ventrally indistinct from exoccipital (0) or distinct from exoccipital, ventrally offset (1).

(338): Posterior surface of basioccipital ventral to the occipital condyle: short and gently curved- dorsoventrally shorter than the occipital condyle (0) elongate, flat and nearly vertical- at least as high as occipital condyle (1).

(339): Contribution of squamosal to dorsal margin of the external otic aperture: absent- dorsal margin formed by the quadrate, the squamosal does not participate (0) or present- dorsal margin formed by squamosal (1).

(340): Deep groove along ventral edge of pterygoid ramus of quadrate: absent-ventral edge flat (0) or present as in Protosuchus richardsoni (1).

(341): Quadrate contact with basisphenoid in ventral view: absent (0) or present (1).

(342): Distal part of quadrate body: distinct (0) and indistinct due to ventromedial contact of quadrate body with otoccipital as in Gobisuchus (1).

(343): Ventral surface of quadrate: slightly concave (0) or strongly concave, with distinct, obliquely oriented crest as in Dibothrosuchus (1).

(344): Quadrate condyles: almost aligned (0) or medial condyle expanded ventrally as in

Notosuchus terrestris (1).

(345): Quadrate distal end: with only one plane facing posteriorly (0) or with two distinct faces in posterior view, a posterior one and a medial one bearing the foramen aereum (1).

(346): Quadrate major axis direction: posteroventrally (0) or ventrally or anteroventrally

(1).

179

(347): Orientation of quadrate body distal to otoccipital-quadrate contact in posterior view: ventrally (0) or ventrolaterally (1).

( 348): Cross section of distal end of quadrate: mediolaterally wide and anteroposteriorly thin (0) or sub-quadrangular (1).

(349): Posterior edge of quadrate: broad medial to tympanum, gently concave (0) or posterior edge narrow dorsal to otoccipital contact, strongly concave (1).

(350): Ridge along dorsal section of quadrate-quadratojugal contact: absent (0) or present as in Zaraasuchus shepardi (1).

(351): LArge depression on lateral surface of quadrate (quadrate depression) ventral to otic aperture: absent (0) or present as in Striatosuchus (1).

(352): Ventrolateral contact of otoccipital with quadrate: very narrow as in Protosuchus richardsoni (0) or broad (1).

(353): Dorsal lamina of exoccipital (anterior to the cranioquadrate canal) sutured to the quadrate or squamosal dorsally when the cranioquadrate canal is closed off anteriorly by a thin lamina (inapplicable in taxa with an open cranioquadrate canal and taxa in which cranioquadrate canal is closed off by a thick lamina): absent- dorsal lamina of exoccipital is not sutured to the quadrate or squamosal dorsally (0) or present- dorsal lamina of exoccipital is sutured to the quadrate or squamosal dorsally (1).

(354): Anterior opening of cranio-quadrate passage in otic area (inapplicable in taxa where cranio-quadrate canal is open laterally): not expanded (otic aperture oval in shape)

(0) or opening expanded forming a caudal notch as in Crocodylus niloticus (1).

(355): Laterally concave descending flange of otoccipital ventral to subcapsular process: absent (0) or present (1).

180

(356): Crista interfenestralis (between fenestra ovalis and fenestra pseudorotundal): nearly vertical (0) or horizontal (1).

(357): Basisphenoid rostrum (cultriform process) shape: slender- not dorsoventrally expanded (0) or dorsoventrally expanded (1).

(358): Length of basisphenoid rostrum: short (0) or extremely long anteriorly as in

Rhabdognathus (1).

(359): Basisphenoid rostrum sutured dorsally with laterosphenoid: absent- basisphenoid rostrum separate from laterosphenoid (0) or present as in Dyrosaurus phosphaticus (1).

(360): Anteroposterior crest on basisphenoid: absent- smooth (0), present- bears a single medial crest (1), present- bears two crests (2), or present bearing three crests (3).

(361): Basisphenoid pterygoid suture orientation: transverse- nearly straight (0) basisphenoid tapers anteriorly between the pterygoids (1).

(362): Shape of basisphenoid exposure in ventral view: wider than long (0) or longer than wide (1).;

(363): Basioccipital, midline crest on basioccipital plate below occipital condyle: absent

(0) or present (1).

(364): Prootic exposure on lateral surface of braincase: widely exposed (0) or very little exposure- obscured by expansion of quadrate and laterosphenoid (1).

(365): At maturity, prootic/laterosphenoid suture is raised, forming a dorsoventrally directed crest (directly above foramen for cranial nerve V) separating fossae for muscle attachment (M.pseudotemporalis profundus and M. adductor mandibulae externus profundus; Holliday and Witmer, 2009) this crest is greatly reduced or absent in small juveniles of taxa possessing the crest in larger specimens: absent (0), slightly raised as in

181

Metriorhynchus superciliosus (1), or present as a strong crest as in Pelagosaurus typus

(2). Ordered.

(366): Exit for trigeminal and middle cerebral vein in prootic: single circular or ovate opening (0), bilobate (or hourglass shaped) with anterior projection of prootic slightly dividing into dorsal (for middle cerebral vein) and ventral (for exit of trigeminal branches) portions as in Pelagosaurus typus, Metriorhynchus westermani (1) or fully divided into two openings by a bridge formed by prootic, dorsal opening interpreted as middle cerebral vein and ventral opening as exit for trigeminal branches as in

Steneosaurus bollensis (2).

(367): Otic capsule size: not enlarged, protruding slightly into endocranial cavity as in

Crocodylus niloticus (0) inflated but not meeting at midline with in endocranial cavity as in Chenanisuchus lateroculi (1) or highly inflated, meeting at midline within endocranial cavity, dividing cavity into dorsal and ventral chambers as in Rhabdognathis keinensis

(2).

(368): Quadratojugal participation in craniomandibular joint: absent (0) or present (1).

(369): Dorsal edge of dentary: straight (0), dorsally expanded at caniniform- showing a single concave arch posteriorly (1) or edge sinusoidal, with two concave waves (2).

(370): LArge occlusion pit on dentary lateral to seventh alveolus: absent (0) or present as in Mahajangasuchus (1).

(371): Groove along lateral margin of dentary ventral to toothrow separating ornamented region (ventrally) from unornamented region (dorsally): absent (0) or present as in

Terminonaris robusta (1).

182

(372): Shape of dentary symphysis in ventral view: tapering anteriorly forming an angle

(0), U-shaped, curving anteriorly (1) or very broad with extensive transversely oriented anterior edge as in Simosuchus (2)

(373): Dorsal surface of mandibular symphysis: flat or slightly concave (0), strongly concave and narrow, trough shaped as in Araripesuchus gomesii (1).

(374): Posteriorly directed peg at symphysis: absent (0) or present as in Notosuchus terrestris (1).

(375): Thickness of splenial posterior to symphysis: thin (0) or robust dorsally (1).

(376): Dorsal edge of surangular anterior to glenoid fossa: flat of concave (0) or arched dorsally (1).

(377): Longitudinal ridge along the dorsolateral surface of surangular: absent (0) or present in Zaraasuchus shepardi (1).

(378): Enlarged foramen at anterior end of surangulodentary groove (inapplicable in taxa lacking distinct surangulodentary groove): absent (0), present (1).

(379): Surangulodentary groove-

(380): Distinct coronoid process on surangular: absent (0) or present (1).

(381): Surangular in dorsal view: does not extended beyond the orbit along the dorsal surface of the mandible (0) exceeds orbit (1).

(382): Surangular contribution to glenoid fossa: forms lateral wall only (0) or surangular contributes approximately one third or more as in Dyrosaurus mahgribensis (1).

(383): Sharp ridge on the surface of the angular: absent (0), present on the ventral most margin as in Zaraasuchus shepardi (1) or present along the lateral surface as in

Shamosuchus djadochtaensis (2).

183

(384): Coronoid length: short (0) elongate, projecting further anteriorly than posterior- most dentary alveolus as in Metriorhynchus superciliosus (1) or absent as in Dyrosaurus mahgribensis (2).

(385): Coronoid, participates on the external face of the mandible: absent (0), present on lateral surface of coronoid process (1) or present anteriorly between coronoid process and tooth row (2).

(386): Mandible geometry, relative positions of the dentary tooth-row and coronoid process, and development of dorsal curvature of the caudal end of the mandible: gentle curvature in the dorsal margin of the mandible, from the coronoid process to the end of the tooth-row (0) or strong curvature, raising the coronoid process considerably above the tooth row (1).

(387): Size of glenoid fossa of articular relative to articular surface of quadrate: anteroposteriorly similar in length (0), slightly longer (1) or much longer- close to 300% of the length of the articular surface of the quadrate (2).

(388): Posterior ridge on glenoid fossa of articular: present (0) or absent as in Simosuchus

(1).;

(389): Retroarticular process dorsal extent: short, does not exceed to the articular glenoid

(0), slightly exceeds the articular glenoid (1) or extremely dorsally curved, greatly exceeds the articular glenoid cavity as in Dyrosaurus mahgribensis (2).

(390): Posteromedial process of the retroarticular process (inapplicable in taxa lacking a retroarticular process): absent (0) or present (1).

(391): Medial articular shelf of retroarticular process: (inapplicable in taxa lacking a retroarticular process): absent (0) or present (1).

184

(392): Medial shelf of retroarticular process orientation (inapplicable in taxa lacking a medial articular shelf on the retroarticular process): vertically and facing medially (0) facing dorsally (1).

(393): Medial articular shelf of retroarticular process dorsoventral position (inapplicable in taxa lacking a medial articular shelf on the retroarticular process): dorsal in position

(0), displaced ventrally to approximately mid-point of the retroarticular process as in

Terminonaris robusta (1) or extremely displaced ventrally to ventral portion of retroarticular process as in Chenanisuchus lateroculi (2).

(394): Anterior margin of mandible (dentary), in dorsal view: outer margin converging towards tip or parallel (0), distinct notched spatulate shape as in Steneosaurus bollensis

(1) or broadens anteriorly, but anterior margin straight as in Sarcosuchus imperator (2).

(395): Anterior of mandible (dentary), interalveolar space size: anterior interalveolar spaces are variable in size, ranging from being larger than the proceeding and preceding alveolus to being half the size (0) or all anterior interalveolar spaces are less than half the length of the adjacent alveoli (1).

(396): Posteroventral edge of mandibular ramus: straight or convex (0) or strongly deflected ventrally as in Simosuchus clarki (1).

(397): Orientation of premaxillary tooth row: curves posterolaterally from midline

"arched" (0) or angled posterolaterally at approximately 120 degree angle as in

Terminonaris robusta (1), or transverse as in Simosuchus clarki (2).

(398): Number of premaxillary teeth: five (0), four (1), three (2) or two (3). (Ordered);

(399): Position of first and second premaxillary alveoli: separated like adjacent teeth (0) or nearly confluent (1).

185

(400): Position of last premaxillary tooth relative to first maxillary tooth: anterior or slightly anteromedial (0) or anterolateral as in Sarcosuchus imperator (1).

(401): Premaxillary tooth row, dorsoventrally position relative to maxillary row: level (0) or ventrally offset as in Sarcosuchus imperator (1).

(402): Number of teeth partially supported by both premaxilla and maxilla: none (0) or one (1).

(403): Number of maxillary teeth: more than twenty (0), eight to twenty (1), seven (2), six (3), five (4) or 4 or fewer teeth (5). Ordered.

(404): Compression of maxillary tooth crowns: absent (0), present- obliquely disposed

(asymmetric), tear-drop shaped as in Notosuchus terrestris (1) or present- labiolingually compressed- oriented parallel to the longitudinal axis of skull (2).

(405): Position of first enlarged maxillary tooth: no prominent tooth (0), second or third alveoli enlarged (1) or fourth or fifth alveoli enlarged (2).

(406): Position of last maxillary tooth relative to anterior edge of suborbital fenestra: last maxillary tooth posterior to anterior edge of suborbital fenestra (0) or last maxillary tooth anterior to anterior edge of suborbital fenestra (1).

(407): Mid to posterior maxillary teeth, crown-root junction: unconstricted (0) and constricted (1).

(408): Cusps of posterior cheek of teeth: not multicusped (0) or multicusped (1).

(409): Morphology of cusps on multicusped teeth (inapplicable in taxa lacking multicusped teeth): one main cusp with smaller cusps arranged in one row as in

Simosuchus clarki (0), one main cusp with smaller cusps arranged in more than one row, forming lingual cingulum at base of middle and posterior teeth as in Malawisuchus

186 mwakasyungutiensis (1) or multiple small cusps along edges of occlusal surface as in

Edentosuchus (2).

(410): Posterior teeth with rings of undulating enamel: absent (0) or present as in

Dakosaurus andiniensis (1).

(411): Maxillary dental implantation: teeth in isolated alveoli (0) or teeth in a dental groove as in Simosuchus clarki (1).

(412): Edge of the Maxillary tooth alveoli relative to palate in lateral view: lower or at the same level as the palate (0) or higher than regions between tooth row (palate sits lower than alveoli in lateral view) (1).

(413): Maxillary teeth crown facets: either lacking or indistinct (0) or with facets in three distinct planes as in Geosaurus giganeus (1).

(414): Size of anterior dentary teeth opposite premaxilla-maxilla contact relative to other dentary teeth: no more that twice length (0) or more than twice the length (1).

(415): Size of dentary teeth, posterior tooth opposite premaxilla-maxilla contact: equal in size (0) or enlarged opposite smaller teeth on maxillary tooth row (1).

(416): Third and fourth dentary alveoli: third smaller than fourth- alveoli separated (0) or third and fourth dentary alveoli roughly equal in size- nearly confluent (1).

(417): Size of seventh mandibular tooth relative to adjacent teeth: similar in size to adjacent teeth (0) or small and set very close to the eight tooth as in Dyrosaurus mahgribensis (1).

(418): Tooth crown serrations: present (0) or absent (1).

(419): Tooth serrations (sensu Andrade et al. 2010- inapplicable in taxa lacking tooth crown serrations): macroziphodont (0) or microziphodont (1).

187

(420): Occlusion, relation between maxillary and dentary series: in-line or interlocked (0) or maxillary dentition overbites dentary dentition (1).

(421): Heterodonty of maxillary and dentary teeth: absent- homodonty (0) or present- different dental morphologies (1).

(422): Atlas intercentrum: broader than long (0) or as long as broad (1).

(423): Atlas intercentrum (hypocentrum) length relative to odontoid process: long-15% of odontoid process length (0) or short-subequal to odontoid process length (5% plus or minus) (1).

(424): Anteroposterior development of neural spine in axis: well developed- covering the length of the neural arch (0) or poorly developed- located over the posterior half of the neural arch as in Notosuchus terrestris (1).

(425): Prezygapophyses of axis: not exceeding anterior margin of neural arch (0) or exceeding the anterior margin of the neural arch (1).

(426): Axis neural arch diapophysis: absent (0) or present (1).

(427): Cervical vertebrae: amphicoelous or amphiplatyan (0), incipiently procoelous-

"shallow procoelous"(1) or procoelous (2). Ordered.

(428): Neural spine on posterior cervical vertebrae: as broad as those on anterior cervical vertebrae (0), posterior spines anteroposteriorly narrow, rod like (1) or all spines rod-like

(2). Ordered.

(429): Hyapophyses on cervicodorsal vertebrae: absent (0), well developed hypapophyses on cervical vertebrae only (1), present in cervicals and first two dorsals (2), present through third dorsal (3) or present through fourth dorsal (4). Ordered.

188

(430): Dorsal vertebrae: amphicoelous or amphiplatyan (0), incipiently procoelus-

"shallow procoelous" (1), or procoelous (2). Ordered.

(431): Thoracic vertebrae, shallow fossa on the anterior margin of the diapopohysis immediately lateral to the parapophysis: present (0) or absent (1).

(432): "Insertion" or sacral vertebrae between the first and second primordial sacral vertebrae: absent- two sacrals present (0) or present- three sacrals present with third

"inserted" between primordial 1st and 2nd sacral (1).

(433): Caudal vertebrae: all amphicoelous of amphiplatyan (0), first caudal vertebrae bioconvex, with other caudal vertebrae procoelous as in Alligator mississippiensis (1) first caudal vertebra biconvex, with other caudal vertebrae semiprocoelous, amphicoelous or amphiplatyan (2) or all caudal vertebrae procoelous (last sacral procoelous as in

Fruitachampsa) (3).

(434): Height of neural arch of caudal vertebrae relative to centrum length: less than two times the length of centrum (0) or more than two times then length of the centrum as in

Dyrosaurus maghribensis (1).

(435): Vertebral morphology near distal end of tail: distal vertebrae isomorphic to poorly heteromorphic, non hypocercal (0) or heteromorphic, bent, ventrally, defining lower lobe of tail fin (1).

(436): Axis rib: holocephalus (rib elongate with one articular head) (0) or dicephalous

(rib triradiate, with two articular heads) (1).

(437): Axis rib tuberculum: wide with broad dorsal tip (0) or narrow with acute dorsal tip

(1).

189

(438): Sacral ribs: short, robust, and slightly bent lateroventrally (0) or long, gracile, and strongly bent ventrally as in Metriorhynchus superciliosus (1).

(439): Orientation of sacral ribs: horizontal (0) or arched ventrally, at least in the first sacral (1).

(440): Shape of posterior chevrons (haemal arches) in anterior view: either 'V' or 'Y' shaped, no distinct anterodorsal process (0) or posterior chevrons have a 'W' shape when observed in anterior view, formed by an anterodorsal process rising between the 'Y' shape

(1).

(441): Scapular blade: scapular blade very large, more than 200% of the width of the scapular shaft, while the scapular shaft contributes less than 50% of total scapular length

(0) or scapular blade reduced, being narrower than the proximal region and less than

150% the width of the scapular shaft, while the scapular shaft contributes at least 50% of total scapula length (1).

(442): Glenoid surface of coracoid: extended on a subhorizontal plane as in Postosuchus

(0) extended on a vertical plane as in Protosuchus richarsoni (1) or extends on an oblique plane, and the glenoid lip facing outwards and posteroventrally as in Alligator mississippiensis (2). Ordered?

(443): Humeral shaft: straight (0) or sigmoidal, with a pronounced posterior curvature of shaft on proximal area of humerus (1).

(444): Distal portion of humeral shaft in cross section: rounded (0) or flattened as in

Metriorhynchus superciliosus (1).

(445): Length of the humerus relative to length of femur: more than two thirds (0), nearly_two-thirds (1), nearly one-third (2) or_much less than one third (3). Ordered.

190

(446): Humerus length relative to scapula: much longer than scapula (0) or shorter than or subequal to scapula as in Metriorhynchus superciliosus (1).

(447): Deltopectoral crest of humerus: present and robust (0) or very reduced, nearly continuous with proximal articulation surface as in Cricosaurus suevicus (1).

(448): Ulna length relative to humerus: subequal (0) or more than one quarter shorter (1).

(449): Ulnar shaft in cross section: ovate/round as in other long bones (0) or flattened (1).

(450): Radius and ulna length relative to width: elongate bones (much longer than wide)

(0) or length and width subequal forming plate like elements as in Cricosaurus suevicus

(1).

(451): Radiale length vs width (considering its proximal width as reference): longer tna wide as in Dibothrosuchus elaphros (0) or as long as wide as in Postosuchus (1).

(452): Proximal and distal ends of radiale: proximal end expanded symmetrically, similar to distal end (0) or proximal head wider than distal one, more expanded proximolaterally than proximomedially (1).

(453): Illium: large, anteroposteriorly longer than dorsoventral height (0) or small, dorsoventrally higher than anteroposterior length (1).

(454): Anterior process of ilium relative to posterior process: similar in length or slightly shorter (0) or one quarter or less (1).

(455): Iliac blade: with posterior and anterior laminae subequal in height (0) or with posterior lamin higher than anterior one (1).

(456): (Wilberg, 2015) Ilium posterior process: present (0) or basent as in

Metriorhynchus superciliosus (1).

191

(457): (Wilberg, 2015) Extent of ilium posterior process (inapplicable in taxa lacking a posterior process of the ilium): elongate and robust as in Alligator mississippiensis,

Stenosaurus bollensis (0) or reduced and fan shaped as in Steneosaurus leedsi (1).

(458): (Wilberg, 2015) Development and orientation of the rugose surface for the insertion of the M. iliotibialis that forms the supracetabular crest: reduced, barely present

(0), mediolaterally narrow and facing dorsally or slightly laterodorsally (1), mediolaterally broad, forming a wide and markedly rugose attachment surface facing laterodorsally (2) or mediolaterally broad and rugose that is highly deflected laterally forming a remarkably deep acetabulum (3). Ordered.

(459): Length from proximal articular facet of femur to distal end of fourth trochanter: more than one-third of total femoral length (0) or one-third or less of total femoral length

(1).

(460): Distal end of femur articular facet for fibula: large lateral facet (0) or very small facet (1).

(461): Flange for coccygeofemoralis musculature on anterior margin of femur: absent- femur linear (0) or present as in Mahajangasuchus insignis (1).

(462): Femur, medial distal condyle: tapers to a point on the medial portion in distal view

(0) smoothly rounded in distal view (1).

(463): Femur, distal surface between the lateral and medial condyles: nearly flat or flat

(0) or groove separating the medial condyle from the lateral condyle (1).

(464): Tibia, length: long (>45% of femur length) (0), reduced (31-45% of femur length)

(1) or very reduced (<30% of femur length) (2). Ordered.

192

(465): Calcaneum tuber: well developed- with long neck (subequal in length to main body of calcaneum plus or minus 5%), distal end wider than main body of calcaneum and projects inwards the body at >80 degrees (0) poorly developed- short neck (

(466): Phalanges of fifth pedal digit: present (0) or absent (1).

(467): Dorsal osteoderms: present (0) or absent (1).

(468): Shape of dorsal osteoderms in dorsal view (inapplicable with taxa lacking dorsal osteoderms): rounded, ovate (0), rectangular, wider than long (1), square (2) or rectangular, much wider than long (width >200% length) as in Sarcosuchus imperator (3).

(469): Anterolateral process on dorsal osteoderms (inapplicable in taxa lacking dorsal osteoderms): absent- osteoderms with straight anterior edge (0) or present (1).

(470): Longitudinal keels on dorsal surface of osteoderms (inapplicable in taxa lacking dorsal osteoderms): present (0) or absent (1).

(471): Continuity of dorsal armour: dorsal armour continues from neck to tail (0) or dorsal armour shows a distinct narrowing or gap at the cervico-thoracic junction (1).

(472): Dorsal paravertebral osteoderm curvature: flat, or weakly arched (0) or with a distinct ventral bend near lateral margin as in Postosuchus and Araripesuchus (1).

(473): Dorsal surface of osteoderms ornamented with anterolaterally and anteromedially directed ridges (fleur de lys pattern of Osmolska et al., 1997): absent (0) or present (1).

193

(474): Sacral osteoderms relative to dorsal osteoderms immediately preceding sacrum

(inapplicable in taxa lacking dorsal osteoderms): similar in size or smaller (0) or larger as in Platysuchus multiscrobiculatus_(1).

(475): Cervical region surrounded by lateral and ventral osteoderms sutured to the dorsal elements: absent (0) or present as in Gobisuchus (1).

(476): Tail osteoderms (inapplicable in_ taxa lacking dorsal osteoderms): dorsal osteoderms only (0) or completely surrounded by osteoderms (1).

(477): Ventral trunk osteoderms: absent (0) or present (1).

(478): Appendicular osteoderms: absent (0) or present (1).

(479): New- Supradentary dorsal to the splenial: absent (0) or present (1).

Added from Tennant et al., 2017:

(480): Sweetman et al., 2015: Snout, profile of the dorsal edge in lateral view (anterior to cranial table) : concave (0); convex (1); approximately straight (2).

481: (Gasparini et al., 2006): Antorbital fenestra, shape: rounded or dorsoventrally high

(0); dorsoventrally low and anteroposteriorly elongate, slit like (1).

482: (Andrade et al., 2011): External supratemporal fenestra, shape: square to subrectangular (0); circular to subcircular (1); mediolaterally narrow and slit like (2).

483: (Tennant et al., 2016): Intertemporal mediolateral width (minimum between supratemporal fenestrae), relative to interorbital mediolateral width (minimum between orbits): intertemporal region broader (0); intertemporal region equal or narrower (1).

194

484: (Andrade et al., 2011): Lateral temporal fenestra in lateral view, size proportional to orbit in dorsal view: small to orbit (0); more than 20 to 50% of the area of the orbit (1); area is larger than 50% of the area in the orbit (2).

485: (Ortega et al, 2000): Lateral temporal fenestra, shape: triangular (0); elliptical to subpolygonal (1).

486: (Andrade et al., 2011): Suborbital fenestra: small, 50% of orbital area (0); between

50% and the same size as the orbit (1); larger than the orbit (2).

487: (Andrade et al., 2011) Premaxilla, projection of the internarial bar relative to the main body of premaxilla and narial opening: does not project anterior to the main body of the premaxilla (0); strongly projected anteriorly from narial opening, extending anterior to main body of maxilla (1).

488: (Ortega et al., 2000): Premaxilla, ventral edge relative to maxilla: lower than ventral edge of maxilla, with dorsal contour of anterior part of dentary strongly concave to accomadate (0); at same height as ventral edge of maxilla (1); premaxilla ventral edge dorsal to maxilla (2).

489: (Andrade et al., 2011): Premaxilla, perinarial crests: absent (0); present as well defined and distinct ridges, cornering the lateral to posterior borders of the nares (1).

490: (Andrade et al., 2011): Premaxilla, perinarial fossa: absent (0); present (1).

491: (Andrade et al., 2011) Premaxilla, postnarial fossa: absent (0); present (1).

492: (Clark, 1994) Premaxilla-maxilla, suture: confluent ventrally (0); opened contact on ventral edge of rostrum (1).';

493: (Sereno et al., 2001) Premaxilla, orientation of anterior alveolar margin: vertical (0); out-turned (1).

195

494: (Pol, 1999) Maxilla premaxilla, suture in palatal view medial to alveolar region: sinusoidal, posteromedially directed on lateral half and anteromedially directed along medial region (0); posteromedially directed (1)

(495): Maxilla premaxilla, lateral fossa excavating alveolus of last premaxillary tooth: absent (0), present (1). (Larson and Sues, 2007);

496: (Wu et al., 1997) Maxilla, depression on posterolateral surface, laterally positioned: absent (0); present (1).

497: (Tennant et al., 2016) Maxilla, lateral surface of jugal process (posterior portion): heavily striated (0); ornamented, like rest of rostrum (1); smooth (2).

498: (Gasparini et al., 2006) Maxilla, evaginated alveolar edges: absent (0); present (1).

499: (Wu & Sues, 1996): Maxilla, lateral surface, unsculpted region along alveolar margin: absent (0); present (1).

500: (Tennant et al, 2016) Maxilla, foramen on palatal surface, dorsomedial to enlarged fifth tooth: absent (0); present (1); develops elongate groove (2) [ordered].

501: Nasal, lateral edges: subparallel (0); oblique to one another, converging anteriorly

(1) (Pol, 1999).

502: Nasal, participation in antorbital fenestra: present (0); absent (1) (Ortega et al.,

2000).

503: Nasals, posterior mediolateral widening adjacent to the maxilla (anterior to contact with periorbital elements): abrupt (0); gradual (constant) (1) (Lauprasert et al., 2011).

504: Lacrimal, total anteroposterior length relative to anteroposterior length of prefrontal: longer (0); shorter or equal to (1) (Brochu, 1999)

196

505: Lacrimal, shape: anteroposteriorly longer than mediolaterally broad (0); as anteroposteriorly long as mediolaterally broad (1) (Sereno & Larsson, 2009).

506: Lacrimal and jugal, anterior margins: confluent, with no notch at the anterior contact

(0); jugal edge convex, producing an anterior notch at contact (filled with maxilla) (1)

(Larsson & Sues, 2007).'

507: Jugal, foramen on the lateral surface near the anterior margin: absent (0); present (1)

(Zaher et al., 2006).

508: Jugal, orientation of base of postorbital process: directed posterodorsally (0); directed dorsally (1) (Pol, 1999).

509: Prefrontal, anterior morphology: tapers anteriorly to a point (0); anteriorly broad (1)

(Tennant et al, 2016).

510: Prefrontal?frontal sutures, form paired dorsal crests: absent (0); present (1) (Pol &

Powell, 2011).

511: Frontal, dorsal anteroposterior ridge(s): restricted to the posterior portion (0); restricted to median portion (1); restricted to anterior portion (2); occupy entire length of frontal (3).

512: Frontal, lateral margin relative to the skull surface: flush (0); elevated, forming ridged orbital margins (1) (Brochu, 1999).

513: Frontal, anterior process constriction with respect to main body of frontal, excluding sagittal projection into nasals anterior to orbits: absent, lateral edges parallel to subparallel (0) present, anterior portion mediolaterally constricted, with convergent lateral margins (1) (Montefeltro et al., 2013).

197

514: Parietal, posterior region dorsal surface: smooth (0); presenting a anteroposterior dorsal ridge (1); marked ventral deflection (?bevelled?) in posterior portion (2); sculpted as with the rest of the skull table (3) (Montefeltro et al., 2013).

515: Parietal/squamosal emargination (anterior concavity at suture contact), posterior margin in dorsal view: absent (0); present (1) (Wilkinson et al., 2008).

516: Supratemporal fenestra, relative contribution of frontal and parietal to medial margin: parietal with equal or greater contribution (0); frontal excluded from margin (1)

(Tennant et al, 2016).

517: Supratemporal fenestrae, minimum width between fenestrae, with respect to maximum width of cranial table: one-third or less of total width (0); more than one-third of total width (1) (Tennant et al, 2016).

518: Postorbital bar between orbit and supratemporal fossa, shape: broad and solid, as broad as dorsal surface of the cranial table lateral to the supratemporal fenestra (0); much narrower (1); much narrower and connected to orbit via a thin, superficial furrow in postorbital (2) (Clark, 1994) [ordered].

519: Postorbital bar, lateral surface formed by: postorbital and jugal (0); only by postorbital (1) (Gasparini et al., 2006).

520: Squamosal and postorbital, lateral margins, dorsal view excluding the squamosal posterolateral process: parallel (0); diverging posteriorly (1); medially concave (2); converging posteriorly (3) (Ortega et al., 2000).

521: Squamosal, posterolateral process, distal end: tapered and pointed (0); broad and rounded (1) (Larsson & Sues, 2007).

198

522: Squamosal, anterior process extending anteriorly to the orbital margin, overlapping the postorbital, in lateral view: absent (0); present (1)(Turner & Buckley, 2008).

523: Squamosal?parietal suture: flat, not elevated from the skull table (0); forms a well- developed anteroposterior groove (often bounded by elevated ridges) (1) (Tennant et al,

2016).

524: Quadratojugal, ornamentation at base (dorsolateral surface): absent (0); present (1)

(Pol, 1999).

525: Quadratojugal, length of anterior process relative to the lower temporal bar: absent or less than one-third of lower temporal bar (0); one third to one half the length of the lower temporal bar (1); long, greater than half of the lower temporal bar (2) (Larsson &

Sues, 2007) [ordered].

526: Quadratojugal, contribution to the lateral temporal fenestra, in dorsal view: extensive contact with the ventral and posterior margins (0); contributes to the posterior and dorsal margins (1); only contributes to the posterior margin (2) (Tennant et al, 2016).

527: Otic aperture (not including additional quadrate fenestrae): open posteriorly (0); closed posteriorly by quadrate and otoccipital (1) (Clark,1994).

528: Quadrate, ventral surface: smooth, with simple muscle scars (0); with multiple developed ridges (1) (Osi ? et al., 2007).

529: Ectopterygoid, main axis orientation: mediolaterally or slightly anterolaterally (0); anteroposteriorly, subparallel to anteroposterior axis of skull (1) (Pol et al., 2004).

530: Ectopterygoid, anterior process: developed (0); reduced or absent (1) (Pol, 1999).

531: Ectopterygoid, posterior process: developed (0); reduced or absent (1) (Pol, 1999).

199

532: Interfenestral bar, anterior half between suborbital fenestrae, lateral margins: parallel to subparallel (0); flared anteriorly (1) (Pol et al., 2009).

533: Interfenestral bar, posterior half between suborbital fenestrae, lateral margins: flared posteriorly (0); parallel to subparallel (1); converge posteriorly (2) (Pol et al., 2009)

[ordered].

534: Pterygoid, quadrate process: well developed, extending posterolaterally beyond anterior margin of basioccipital (0); poorly developed, only present as an incipient projection (1) (Pol, 1999).

535: Pterygoid flanges: mediolaterally expanded, laterally surpassing the quadrate medial condyle (0); relatively short, and do not reach laterally to the level of the quadrate medial condyle (1) (Osi et al., 2007).

536: Basisphenoid, lateral exposure on braincase: absent (0); present (1) (Pol, 1999).

537: Basisphenoid: ventral surface continuous with surrounding bones (0); body ventrally developed and separated from the remaining elements by a posteroventral step formed by a sulcus separating it from the main occipital plane, forming a postchoanal pterygoid?basisphenoid tuberosity (1) (Montefeltro et al.,2011).

538: Supraoccipital, posterodorsal exposure in skull roof: absent (0); present (1) (Ortega et al., 2000).

539: Supraoccipital, posterodorsal exposure: exposed in midline portion of posterior region of skull table (0); restricted to a thin surface attached to posterior-most portion of parietal and squamosal (1) (Montefeltro et al., 2011).

540: Mandible, outer surface sculpture, lateral surface: absent (0); present (1)

(Montefeltro et al., 2011).

200

541: External mandibular fenestra, orientation of main axis: horizontal to subhorizontal

(0); inclined, directed anteroventrally–posterodorsally (Andrade et al., 2011).

542: Dentary, lateral surface below alveolar margin, at middle to posterior region of tooth row: vertically orientated, continuous with rest of lateral surface of the dentaries (0); flat surface exposed dorsolaterally, divided by ridge from the rest of the lateral surface of the dentary (1); flat, unsculpted surface confluent with rest of the lateral surface (2) (Pol &

Apesteguia, 2005).

543: Dentary, mediolateral compression and ventrolateral surface anterior to mandibular fenestra (or of anterior portion posterior to symphysis if fenestra is absent): compressed and flat (0); uncompressed and convex (1) (Ortega et al.,1996).

544: Dentary, sculpted below the tooth row: lacking sculpting (0); present (1) (Pol, 1999).

545: Dentary alveoli: all independent of one another (0); some confluent (1); all confluent, within continuous alveolar groove (2) (Tennant et al, 2016) [ordered].

546: Dental alveoli, transitional shape morphology from circular to subcircular or oval: absent (0); present (1) (Tennant et al, 2016).

547: Dentary, distinct foramina on occlusal surface, lingual to dental arcade: absent (0); present (1) (Tennant et al, 2016).

548: Dentary, external alveolar margins, dorsal edge: vertically festooned, forming raised rims about each alveolus (0); flat (1) (Tennant et al, 2016).

549: Dentary, internal alveolar margins: forming raised rims (0); flat and confluent with dentary occlusal surface (1) (Tennant et al, 2016).

550: Dentary, diastema (gap) between D7 and D8: present (0); absent (1) (Tennant et al,

2016).

201

551: Dentary, pitted ornamentation of external surface: absent (0); present (1) (Tennant et al, 2016).

552: Dentary, grooved ornamentation of external surface: absent (0); present (1) (Tennant et al, 2016).

553: Dentary, symphysis and dentary arcade lateral to symphysis, in dorsoventral view: parallel (0); oblique (1) (Tennant et al, 2016).

554: Angular and posterior surangular, strong pitted pattern: absent (0); present (1); lateral surface with rugose pattern instead of pits (2) (Andrade et al., 2011).

555: Surangular, extension toward posterior end of retroarticular process: along entire length (0); pinched off anterior to posterior tip (Norell, 1988).

556: Articular, medial process articulating with otoccipital and basisphenoid: absent (0); present (1) (Clark, 1994).

557: Posterior premaxillary teeth, apicobasal length: <1.5 times the size of the anterior teeth (0); 1.5 times or greater than anterior teeth (1) (Clark, 1994).

558: Maxillary teeth, mesiodistal margin carinae: absent or with smooth and crenulated carinae (0); with denticulate carinae (ziphodont condition) (1) (Ortega et al., 1996).

559: Maxillary teeth, middle to posterior elements, ridged ornamentation on enamel surface: absent (0); present (1) (Andrade et al., 2011).

560: Maxillary teeth, enamel surface: smooth or slightly crenulated (0); with ridges at base of crown (often extending apically) (1) (Turner &Sertich, 2010).

561: Maxillary teeth, striations on labial and lingual faces: present (0); absent (1)

(Tennant et al, 2016).

202

562: Cheek teeth, base (i.e. immediately apical to root), with respect to remainder of tooth crown: not constricted (0); constricted (1) (Tennant et al, 2016).

563: Maxillary teeth, posterior teeth, mediolaterally compressed lanceolate-shaped morphotype (sometimes called ‘leaf-shaped’), visible in labial or lingual view, with wide crown tapering apically to a sharp point (note that the point can often be abraded): present (0); absent (1) (Tennant et al, 2016).

564: Tooth, present at premaxilla–maxilla contact with transitional size-based morphology: absent (0); present (1) (Turner & Sertich, 2010).

565: Enlarged maxillary teeth (at least 1.5 times the apicobasal size of remaining teeth): present at M2 and/or M3 (0); present at M4 and/or M5 (1) (Martin et al., 2014a,b). ';

566: Maxillary tooth 5, apicobasal size relative to adjacent maxillary teeth: subequal, or

<4.0 times the size of adjacent teeth (0); hypertrophied, at least 4.0 times the size of adjacent teeth (1) (Tennant et al, 2016).

567: Maxillary teeth, bulbous tooth morphotype (tribodont): present (0); absent (1)

(Sweetman et al., 2015).

568: Cervical vertebrae, number: six or fewer (0); seven (1); eight or more (2) (Tennant et al, 2016).

569: Cervical vertebrae, neural spine: absent, or extremely reduced (0); present, distinct from centrum body (1) (Tennant et al, 2016).

570: Dorsal vertebrae, number: 14 or fewer (0); 15–16 (1); 17 or more (2) (Tennant et al,

2016) [ordered].

571: Posterior dorsal vertebrae, transverse process shape: dorsoventrally low and laminar

(0); dorsoventrally high (1) (Buscalioni & Sanz, 1988).

203

572: Sacral vertebrae, orientation of transverse processes: project laterally (horizontally)

(0); deflected markedly ventrally (1) (Gasparini et al., 2006).

573: Caudal vertebrae, number: fewer than 50 (0); 50 or more (1) (Tennant et al, 2016).

574: Caudal vertebrae, anteroposterior ridge/lamina separating centrum and neural arch: present (0); absent (1) (Tennant et al, 2016).

575: Coracoid, medial process: elongate posteromedial process (0); distally expanded ventromedial process (1) (Wu & Sues, 1996).

576: Humerus, circular depression on the posterior surface of the proximal end, for the insertion of the M. scapulohumeralis caudalis: absent (0); present (1) (Pol et al., 2012).

577: Humerus, lateral and medial surfaces of distal end: flat and anteroposteriorly broad, similar in anteroposterior length to the transverse width of the distal end of the humerus

(0); convex and reduced in comparison with the transverse width of the distal humerus

(1) (Polet al., 2012).

578: Radius : tibia length, ratio: <0.6 (0); 0.6 to <0.7 (1); 0.7 or greater (2) (Tennant et al,

2016) [ordered].

579: Ilium, posterior end of the postacetabular process: tapering posteriorly to an acute tip (0); subrectangular with a vertically orientated posterior margin (1) (Pol et al., 2012).

580: Pubis, anterior process: absent (0); present (1) (Clark, 1994).

581: Femur, proximal development of greater trochanter: prominent, ridge-like lateral border that separates the lateral surface of the proximal femur from a flat posterior surface reaching down to the level of the fourth trochanter (0); proximodistally short trochanteric surface lacking a distinct ridge, terminating well above the fourth trochanter

(1) (Pol et al., 2012).

204

582: Tibia, distal projection of articular surfaces: medial region of distal articular surface extends further distally than the lateral region, forming a strongly oblique distal margin of the tibia (0); medial and lateral regions subequally extended, with distal margin subhorizontally orientated (1) (Pol et al., 2012).

583: Tibia, posterior surface of shaft: flattened and confluent with fibula (0); twists posteriorly, leaving a void between the tibia and fibula (1) (Tennant et al, 2016).

584: Astragalus, anterior margin of the tibial facet: forming a well-defined ridge that reaches medially the ball-shaped region for the articulation of metatarsals I–II and closes the proximomedial corner of the anterior hollow of the astragalus (0); forming a low ridge that is medially separated by a notch from the ball-shaped region for the articulation of the metatarsals I–II, failing to close the proximomedial corner of the anterior hollow

(1) (Pol et al., 2012).

585: Distal tarsals, digits 2–4, dorsal surface: longitudinally grooved (0); smooth and flat

(1) (Tenant et al, 2016).

586: Metatarsals I–IV: equidimensional (0); metatarsal I shorter than metatarsals II–IV

(1) (Tennant et al, 2016).

587: Osteoderms, dorsal surface: entirely sculpted (0); partially or completely unsculpted

(1) (Tennant et al, 2016).

588: Presacral armour: cervical and dorsal trunk shields undifferentiated, morphology grading continuously (0); cervical shields clearly differentiated from dorsal trunk shields by size and general morphology (regardless of contact between nuchal and trunk series)

(1); anterior most cervical osteoderms developed into distinct shield (2) (Andrade et al.,

2011).

205

589: Nuchal osteoderms: consistent morphology along series (0); vary substantially in size in a random fashion (1); systematically increase in size posteriorly (2) (Tennant et al,

2016).

590: Dorsal and cervical osteoderms: some or all imbricated (0); not in contact (1)

(Tennant et al, 2016).

591: Dorsal osteoderms, accessory osteoderms (sensu Frey, 1988; i.e. osteoderms not forming part of the dorsal shield): absent (0); present (1) (Turner & Sertich, 2010).

592: Dorsal osteoderms, anterior edge of dorsal surface (i.e. articular surface, if present): sculpted, undifferentiated from main osteoderm body (0); unsculpted (1) (Tennant et al,

2016).

593: Caudal osteoderms: ovate (0); subcircular (1); subrectangular (2) (Tennant et al,

2016).

594: Caudal osteoderms, bearing anteroposterior ridge: present (0); absent (1) (Tennant et al, 2016).

595: Caudal osteoderms, anteroposterior ridge: present medially (0); forms a distinct lateral step in posterior-most elements (1) (Tennant et al, 2016).

596: Caudal osteoderms, medial and lateral edges: serrated (0); smooth (1) (Tennant et al,

2016).

597: Caudal osteoderms, secondary osteoderms: present (0); absent (1) (Tennant et al,

2016).

598: Caudal osteoderms, anteroposterior ridges: same morphology along series (0); becoming more pronounced posteriorly, coincident with a decrease in osteoderm size (1)

(Tennant et al, 2016).

206

Appendix 2: Taxa and scorings

MATRIX Gracilisuchus._stipanicicorum 11110?00000101010000???0??000010000000?000000000000000000??0000??0101000? 000??00000000000?000110013210000110??10?00?100?001????00?0100?00???00????? 0000001000?0???10000000100010?000000001000-010000000000000-0000020100--0- 0200100-1010010-000100000?0---0013-0?-001110-00?000200-1100?1?????-??-????- 0????????????????0??????-????0?0?0-0??000???000?00000000--- 0???????0????0000??0???--?000??00?0----0000??00012???0- ?0??00?????0????10000?0?0???00?1?00100000??000002000?00000000?10?0??0?{1 2}0110000000000?00200?-11000001?000001000?00002000??????????00000??????00?- ?00000100000121000000?0?000000?000000???????0???? Turfanosuchus._dabenensis 1111100010000100000?0?100000000000000?01010000001000?????0001?10000?10??? ?10?0000?0000?00?011?100032100031000?1??2?11001000?????00?1012- ??????0????000001??????1??101000002000100?000000010?0-010001000000000- 0000000200001-02?0?00-?0?0010?0000?00000100?0013-0?- 002110000?000100?010?01?000????????0?0???????10???000?0??0?0???0?00?00?- ?00000?000000000?0?00????0???0?000?0?00001???00-- ?000?000?0?0??0?00000001???00- ?000000000?0?????00000000????0?00001?0?????01000?00?11?0?0200?1???????20001 0?01000?0?00201?-110000010000?01000?00????00???????0??000?00????100?- 000???????????1?00????0?000??-??1???????????0???? Yonghesuchus._sangbiensis 00?11?????1??0??????0???0010??00?0??0011?00000000??0???0?????????01111?????? 00????0001?0001001?0103200?0?10??????????0???????????????????????????????????? ???????????000?0100010?01000000?????00??0000?0?0??0???0?????????????????????? 0????-010000000????0??3-00??00100?00?00000?000??01???0??0?????0- 0?????0??0?????000??00????0????????????0?????000000?0?0??????00?010?????00??1 0?000??00??000??010??0?0000000122?00- ?00000?000?00????0????????????????????????????????????????????????????20??2010 1000?0?00201001000000?????????00????00????????0????00??000???100?- 1?0000?000101???????????????????????????????0???? Carnufex._carolinensis 01?1??????????????????????00??00?????10?1??00?001????????????????????0????????? ?????1??0??0?0?111?10??????1???????????01????????????????1?????0??????????????? ?1????????????0?10?00000?10?????????0?0??0???????????????????????????????0??0? 000110000??????????????????0?0????000??0???0?0??????????????????????0????????? ????????????????????????????????????????????????????????????????????2???0?010??? ??010000121?00- 0000????00?00?????0??0???????????00??0?????????????????????????????????1???0??0 000???0000?????00000????????????????????????????????????????????-??0?00?000- 01?1?0?????0?????????????????????0???? Postosuchus._kirkpatricki 01000??000110000110000000000000000111000?00000000000???001?00?0????0???0?

207

000??000010011?010011?10110000010000?10000??001?00?00001??100201000101001 010100?01110000110000001?00010?000?-0001000-0110000???00100- 0100010100?0000000000010?0?10-0001000?0???????1?10?-000110- 000000001101000?0?0?--?0-0??0-????0000-00-000000??000?-?01000000- 00000010?0000000000?0-????00?01????0?010?100000???0?000?0- 00???0?0010000121000- 0000000000?00?0?0000000000??0??000010?00?10001002000100000????10?0??00011 1200010000010010000011000000?00001000?00000200010--0100-000000000?1000- 100000?0011012110010000000?000010???????????0???? Erpetosuchus._granti 1110???1001100?2011?????0000?000?00?0??0?00000?0?000000?0??00?00??1??0?0??? 0??000?000??1010121111122000?31000?100??01000000?00?0?????????00??01??00??? ??????????011?0?000020001000000000010010000?000001000?0100000101000000020 0000?1011000??001101?0?0---1?13-100002110-100100001-0100111?00--0?-0?-0-- ???0000-10-0?0?0???0011-00000?00???00000?000?0?000?0??0??- ??00??1??0???000010000011?10???00??????0?00100005201?0-?00000?11- ?0??00000?0?????1????0000?1000011??????????????0101?10?0??0?01102100100000? 0020100110000000?00000000000000200100--01???00000?????1000- 0?00?0??00??120?0???0?0?????????0000????????0???? Dyoplax._arenaceus 00?10??11????????0???????????????0?10???10001001000????????????0??1??????????? ?????0?0?10??11???1?32000??110??1????00???0??????????????0??????1????????????? ???1?????00000200010?00000001?1- 00000000000100000?0100010020?000?0???0???0?1010?00010000000???0113- 1???021?0?10?0000?1??????????????????????????????0????????????????????????????? ??????????????????????????????????0?01?????????????0??01???0?0?????????????????? ?????????????????????10?????0???????????????????????010101000100??11011?01000 ?0?00001?0110000?0??100?1100???????????????????00??0?????1????0????????????2? 2????????????????000000201???????? Phyllodontosuchus._lufengensis ?0?1????0????????0?????????1??11?0??010?1???00011?0????????????0??1??????????? ????00?0?00???00??01??0?00???0????????????????????????????????????????????????? ????????00000?????0??0???000??????0??0?0????????- 000?????1????0???0???????00?0?0?0100??0?10?????3- ????????0????00011???????????????????????????????0???????????????????0????????0 ???????01000?0??0????????????????00?01?00?0??00??100???????0?0????00100?00??0 ?0??0000?1??????????????????????????????????????????????????????????0011010010 00?0?0??0????????0??0??000?10???00???????????0????0???????????????00100010?01 ???????????????????????????????????? Pseudhesperosuchus._jachaleri 01??11?101110??0?10????0000??0??0001?01??0?00000?00?????0??00?0????????100?? ??001100?1??000100??0112000031100?11?10??000?000?001???1001??111111111???? ?????????10?11000000100010?00000000?0?100000010000000?0- 0000100100000000?001002010000?000100??0?10000113- 1?10011?00000000110??0001110?0?00??00?1-02000000-00- ??0000?00011?0?1011?00???0000100??0?00100??0?????00?????????0000??0??0??000

208

??00????0??0?0020000122000- ?00000?0???0?????0000?000??????000010000000???????1?110??1????????????011111 ?01000?0?0020??1110000100?0?001000?10?0100????--00??-0?000??0???00?- ?00?00?00010?2110?0?1??20?0?0???????????????0???? Redondavenator._quayensis 0?1???????????????????????????????????0?10?00?0????????????????????????????????? ????0?????????11??02???????????0??????????0???0?????????????11????????????????? ??????1????2?0000000?0????1????00?????????????????????????????????????????????? ?????????????????????????????????????????111?2?????????????????????????????????? ??????????????????????????????????????????????????????????????????????????????00 0000?02??0- ?000????00?0????????????????????0????????????????????????????????????????????01 00000100?01?0??????????????????????????????????????????????????????00?01??0???? ??????1???????????????????????????? Macelognathus._vagans ?1????010?????1??1??112????11?????????11????????00???????????2??001????11?1??0 0??010?????????00????????0???01????21??000??????????10112????????????0??11???1 1?????1?0?????????????0?01???????????????????00???0?00????0???00?0000????????? ??????????????????????????????????????????00???????????????????????????????????? ?0????????01??0?1?0?001000101000000001?0000010?000?100000100??????????????? ?????????0??0?21?00-?0?0?0001- ?1?????0?200?????????????????????00000?1101100?????????????????????????????00? 01??????????0?00?0???????0???00???????001????0000???000????0?010????1?1?0???? ???0?010???????????????????? Trialestes._romeri 01?1??????????????????????0???0???????0??????000?????????1????????1??0????????? ???10????0??0111101?201?0???????01?0??0010000?001?0???0??1111?011?11?1??1?? 0?1??1?110???0?0????10??0001?0??????????0?0??????????????????????????????????0 ????0?0100000?????????????????????????001???????????????????????????????1?????? ?????????????????????????????????????????????????????????00010?000???100???0?01 0--0?0?1000?1?2?00- ?00001?0??00?????00000000???00?0000000000000000??0????00?0?????????????0??2 1??????????0201???????0???????????????????????????????00000??????00?00??00010? 0?01?1?00??100200010000????????????????? Almadasuchus._figarii 0??????10111111201111?101?0?11??0111???2????????2010??????11121000????01111 1100110???????00???????????1????1110??2?1????0??????2??10101???1?????????????? ??1??????1?0??????0??????????????????????????????0??10??0000101100010000100?0 0???????????????????????????0?????0???00????-???????????-??-????- ????????1???????????????-???00?20001200001000100??10???01- ??????01?000????????????????0??????010-- ?????????????????????????????????????????????????????????01??????0?0?1?????????? ????????11?1?????0???????????????????-??0010??11?0???{0 1}0???????00- ?????????????????0???????????????????????0?????????????????0???? Terrestrisuchus._gracilis ?111??1101100??0?0011??0??01?00000010002?00000000000000-

209

01?0000??0101??1001000002000?1?01000100001?20?00?10???11?21??0000000?001?0 ??1?000?11111011?1101100011101011?00000020001000000001010010000000000000 000-0000000110100?10?011001000000?000100000?0---0013-0?- 0021001000000111?10???11000--00??0-1--2010000-00- 000000?0000100?1001200?100000100000?00000000??000?0?01100?000000??0000-- 00000000?01?0?0?0???000121000- ?00000?00??0??00?0010000001000?00001000000100000?110110000010?00?????0212 10100100000?0020100110000{0 1}00?00?01010??0?0??0000?-- 00?0000?00??0??1000010??00??001??211??1?000100?10001????????????0???? Litargosuchus._leptohynchus 00?1??1101?00??1100????0??01????000?00??100000?0000????????00??????010?1?0?0 ??002100?1?0100010??11120000?1010013???????????????1????????011?0?1????????? ?10?1?1101?0?00000200010?00001010?00100000000000000?0-0000100100110- 01?0000?201?000?000100000?0---0013-0?0000?00000??000?10?????????0--0????-??0- ??????- 00?0?00???????1????00?20???0???????????????0???????????????0????000100000???0? 01000?01???0?0010000122?00- ?0?00??????00????00?0??000?????0?001000000????????????0?1????????????10120??0 01000?0?0020100111000000-000011100100?????010-- ?????000?000?00?00000?0?0??000?01211????1?010???0???????????????0???? Kayentasuchus._walkeri 0010????11111?1?100111?0???1?0??0?01?0001000000?0000??????1?1?0????11??100? ???0??0003??1?????0021?1?0??00??001???1?110??0?????????0011???????????????1?? ???1??????1??0??00????10?00000010?0???00000?00??????010????2011????- ???00??1?010?0??0001?0?000??????1?-????????????????????????01102??-0?-???1- ??????0?-0??????????????-???0012??- ?0??????????0??0?0??00???????????0????0001000????????10????????00?01000012??? 0-000000??1- ?0????????0?0?????00???????????????????1?011???011??10???1?1202????0100010?00 201001??0000??-0?001?00?000???0????--????- 00?000??111000???0000?100?01????0????????1?????0??00???????????? Hesperosuchus._agilis 00?1??11111101?0?101????0011000000010010100000?0000??0?????00?00????00??00 11??00111031111010001311020??0??0?0?1112?1100?0000?0?1???1001010101110111 1??010??11?0111??00000?10?000?0000?000?0010001000000000000- 00?000020101000000000????0??0?00?1?0??000--- 00?2?0??0011??000?0000??????????????-??-????--?-?????10??????????????- ???????????????????????????????????????????????0?0?1?????????1???0???????0??000 ?00121?00-0??0???????0?????0????????0????0?00?0?00000??????????1???0- 10010?0????011111?01000?0?00201?0100000000300?0000011000?0?000?-- 00???0?000??0???0000??0000?00???12???????00?????????000000??????0???? Dromicosuchus._grallator 00?1???111110??0?10????0??11??00000??01010000000000?????0??00?00????00???0? 0??001000?1?1001001120?020?00??000?11?101100?001???01??011110011?1?10?1?11 101????1??10?1000000020?0??000001010100100000010000000?0- 0000?10100000000?0?0001??0000?000100000?0---0013-

210

???0?1100000?000001?000??1???????????????2- ?????10????????????0??????0?20???0?????????01000?0??0????????????????000010000 ???00001000------0?0000000121?00- 00?0?0??00?00??000000000001000?00001?00000000000?1101100?010000000??11011 10??01000?0?00201?1110000?10300-01000?000???0?0??--????- 0??00000???01?0?00?00?000?01200?0?00002??100???000000??????0???? Dibothrosuchus._elaphros 0011110111110112111111000111000001110011100000000000000- 01110000001010010010?000100031?00000111301220000310000121??0000000000001 ?0?????00111011011101011???????10??100000010?01000000100011- 10000000000000000-0000110101?000000000002010?1000001000?000---011??1?- ?011001000000001-00?0011000--00-00-1--2000000-10-0000000000110001001200- 1000001000001000000?0---0000?0100000?0000100000??00001000-010-- 010000000121000- 0?00000000?0000000200?000?0?00??000?000000000000???0??????000000????012011 2110100010-0020100111000000300-01000100000?00001--0000- 000000001110000100000?10010121100?0100?0??0????000000??????0???? Sphenosuchus._acutus 001110111111011211111110010100000001001010?00?000000000- 0100000?00?0??010?111?0010002110000001121122000030000002?????00?001????1?? ?????????111101??????????????1?1??00000010?01000000?00010000000000000000000 -01002001010100100000001010000-00010000000---001??0?-0011101000000000- 00000?1000--?0-0??1--20?0000100-000000000011-001001200- 1000001000001000000?0---0000?0100001?0000100000--00101000-010-- 010000000121000- 0000000000?00000?00?0?????0????0000?00?0?0??0?????10???0?????0?0??0?0100111 100100010?0020100110000000300-01000110000200010--0000-0100000?111000- 000000?000?01?1?00?0100????1??1?0000?0??????0???? Junggarsuchus._sloani 0011101111100112111211100001111101011111100000010000000-00111200{0 1}0101001001{0 1}1100100001100000001211120100310010121????11011001102????????00110111111 ????????????101??00000010?000??000100011-10000000000000010- 0000000100?000010001001010100-00000000001000011110--0011101000000211- 1000011?00--0--?0-1--2000000110-0000?1??1001-01?001200{0 1}1000001000001000000000???0?0?010000000000100010--00001000-010-- 011000000121000- 00000000001000???1?41??0??10???0100?000000???????????????????????0?001001111 0?200010-0020101110000000300001000100000200001--0000- 010000001110000100?001000??1???????100?????????????????????1???? Protosuchus._richardsoni 1111101101101012101212211- 1111100010000100000001201000100020000?1010001111001111200031101110101301 02000001011113?101100101000001000100100011011111001011110111110111011100 0000000?000100001-0-00000000000000010010000021000-020000102000000- 00000000000---001101000010100000000001110000?1?10--?0-00?1-0???000?-?0- 000110?01000001110010001000000001110000000100--

211

1010?3100??00000000001000?0000000-000--0?0010000100100-000010001- 10000?000300000010000010010000000000002?101100101000000011000?0111?01000 10?00001?1111100100- 00001000100000210101????0???01010??1??10??1111001000?01201100?10010011001 001200021-1100???? Protosuchus._haughtoni 1111101101101012101212211- 1111100010000100000001201000100020000?1010001111001111200031101110101301 0200000101111?????????????????????????????????????????????????????01110000000 00?000100001-0-00000000000000010010000021000-020000102000000-0000?0?0000-- -001101000010100000000001110000?1??0--?0-0??1-0???000???0- 000110?0100000111001000?000000001110000000100-- 1010?1100??00000000001000?0000000-000--0?0010000100100-000010001- 10?????????????????????????????????????????????????????????0001111001000?0?000 01?0101000110-00001000?00000210111??000?- 101010??1??10?0111?001000?01???????1????0??????????????????0???? Gobiosuchus._kielanae 111111?1011?1012101212011- 01110000?00001000010111011001000?0?20?1010001111001111??0010?01101111121 120000010111?3?110000?0?????0?1?01???10?1??1???????????????????101010100000 0000000?0000100?00000000?0?00000?100?----??000-???000112?00??000000001?0?0-- -??1111?0001011-00000000?100002?1001--?0-0??1-010?0002- 010?0?01100100?0?11???100010000001?011000000010?-?0?1??010???0?000?1?0?01- ?0??10000-000--000020000100000-010010001- 10?????0??00?0?00??????00?00?0???00000???0??00?010000101111?20-- 011010001010020100111100000?0000--00010000210001??0000??- ?0?00?111??0?001000?010?011??0??????20???1???10001021-1??0???? Zosuchus._davidsoni 11111??101101012?012???11- 0?110000100?11101010122112101011?01?0110100?111100?111?1?02???01112012010 200003011111????????????????1????????????????????????????????????010100?0?0100 000000001000-000011000??0010?0010?????0?0?-0{0 1}0?10111??0?00?1000000???0- --?0110?10?010?0??0??1?001??11???1?0?-???- 1?02?01??11?2???????011001?000???1001?0?1?100001?011001001?10???0????2000?? 0?000?0000?0?????0?????0??????102?000401?00-010?1?001- ?0??????????????????????????????????????????????????????????001111??100010?001 01?011???01?0- 00001000?1000?210?????001001?00100?10?11010?0000?010?01???????????????????? ???????????0???? Zaraasuchus._shepardi ???????111?110?210????????0???00?10000??????1???201????????0??01??1??1?????0? ?1??????1??1100????2????00??1?1?11???????0?0????????????????0?????1??????????? ???????????0???????????????????????????????????????100???002?00?- ?????010?000?0????????1??????????11100001011100?000?0210??0???????-??- ?????????????????????????????????????0?????????????????????1???????????????????? ?????01?????1?????00????????????????????????????????????0?0?????????????00???00 0????????????????01000?1?1??1???2001??????????????????????1???00?01110?1000?2 ????????????1??????????01?100???????????21??????????????0??1?00????????0????

212

Nominosuchus._matutinus ?011??0101101012101?????1- ?00100??10001210?01012211210?010001?00111000????00111?2?10211?000??112010 20000?101111?????????????????????????????????????????????????????010100?0?000? 0000000001-1000000100000001010000100021000-00?0?1111000000?0000001?0?0--- 00110?0??01?1010???00011??100??1?100000-??02001001002?00- 0?011??010?1?0???0??0???????????011?0?0?1010?????????1??0??000?00?001100?001 ??0???????0??1100001?1000-??00????1- ?0???????????????????????????????????????????????????????????- 10010??????0?000?1??11?0??????1????????10?01?10?????00???10?01???????1?000??0 01?0??01???????????????????????????????0???? Fruitachampsa._callisoni 1?1?100111101012101?12?1??01010000010011100010132112101010?0?200?01001?11 1?0?11?1?00210?011?2012211200010101111??01?11010?1???????0????001???111?1? 0?01?????????0??0?10100????00?1000000-001?00000--0000?0010100100??020000- 0000101110000?0-?0?00?0?000---001100000010?00000??00????1000?1?01-?00- 10020????1002????00?1?????0?1?0???0?10??????0??????????0??0?01- ??????????10?000?0????10??00?0000??0????0?????00012??00-010010??1- 000??0020020030???0????000?0000??01?00??1??1??????1?00??????0?1100101000001 00201?01010??100?100?101?01000021?000??00??11010100?11110010?10001000??1? ??0????????1??????????????????0???? Sichuanosuchus.shunanensis 10111??1011010?21012???11- 0111000?10000110?010111?1110?01?201?0010100?1?11000111100021?0011121120?1 000?00?011113?????0??01?????11??????000110??1??0??????????????0?1011100?0?00 0?00001000?0?100000100?0?0001010000000010000-0?000?012??0?00?0000001?0?0-- -?0110110001010?00000020110010??1?1?--?0-0??1-0?0?0012?00- 0000110010000??1000100010?0000??01100?0010001????1??210?????000?100??1??0? ??0???-000--0?10100001?1000-00001?001- 11?????0?00??????????0??0?0?0??01???????????????1?0?????1???2010??10100010000 201?1111100100?00001000?1000??1?0????00?0110100??????00000?0????010000?0?0 1???00????0????????????????0???? Shantungosuchus._hangjinensis 1?1?0??1????101????{1 2}1???1- 1111000???0012????1?1{2 3}2?11101010?010??10100011?10??0?1??001?0?010?001101??????????????111??0010 1???002???1111?0011?111?0??101?????1???0111??110??0?0??0000??00??0?????????? ????????1???????????0?- ???????1????????0??00000???????????11?001??0???000020?10?100?1?11--?0- 0??1?010??01???0- 000?1100100?0001??????01?000001?01100100?0??1??0?1?0210?0?0?0000??0??0-- ??010???-000--??101000?1?0100-00?010001- ?00????0??0???????0?????00?????0?00????1?0?????????????????0?????10?1???10?001 01??1????01?????????00?1??0??100????01?0?1010100011111010???0???1????0?2???? 1??20???????00??????????0???? Steneosaurus._bollensis 101???0100100?12111100?01- ?001??0?00?000111120121?021?001??001?0101212?11?11??01??00?10?000??0??2111 111211100003100011010000?01111???1?0?000?111?01?0?111?01111??00010001010? 110??0000000?0010000011100000010001000001010000101000000111010?1002?0?11

213

?0---001?-21-00011-010- 1000110?000?12?0?0201?00020110?10021?1?00001010?0110001?0?2002?000101000? 00?00100001100?101?100022000?01?0??001010??0000?01101000??000000?00- 0010?00?1-00000?10?0000000000100?00100100000000001?0?100101{0 1}000010010100111?1-10000010?10101111000?00?100000010- 00?1?000??000?1001000?000001??0?0?00100010?012121?000?000?110?0000?00020?? ?10???? Metriorhynchus._superciliosus 101011?100100012?01??0?11- 00010000000000111{1 2}2012010210101?200100101212?11001??0111000???11102000211111121110001???? ??0011??????111??????00????01????001?1??????1?00010101010?01000000000- ?00100211--101100010001011001011101000000000110?100010100000011001110-21- 00111-100- 1010110100001200102111000211???10001110000010002011000101020021000101000 100000100001100010?1110011000000000000111001110000110000020000000000- 001000001-0001001000010?01101111201310111101001-0110011111------?- 0012111111-100000100?01??01100001??01-000003-1001000000??????-1- ?0000?11100000000000010?0111201010???????????????????????0???? Cricosaurus._arauanensis 101011?1001000121112?0?11- 0001000000000011022012010210101?200100?01212??1001??01110001??0100100021 3111121110000??0???00?1??????1??????????????01???????????????10???1000102000 1000000000-?00100211--101100110- 010130010111021?1100000?20110?0001000?0?1110?013-21-?1011-100- 10121001?00??2?0?02???0??20??????0?1????0001?00?0110?010102002110010100010 000010??01100?1?111100?10000?00??00??11?0121?0001?0000020000000100- 00?0?0001-00?10?10000?0001?01111201311111101001-0??0??2211------?- 00121111120100000??0?01?1?110?0?00?01-100013-0????0?????001?000- ???00?00?000-??00000010?0101101?00?1011110???????????????????? Geosaurus._suevicus (Cricosaurus._suevicus) 1010??010010???2?1?2?0?01- ?0010001011?0011022012110210?01???02??1012?2?110?1??????00?1??000??0??2131 11121110000??????00?????????????????0????????????????????????000100????0??10?? 000?00-?001?0?1?--101?0??11-01??200???1????????0?????0110?0?01000?0?110?1?1?- 2?-0110?-100- 0012?000??0????????????????????????????????????????????????????????????????????? ?????????????????10?0?0??0??00110?01?1???????0?0?????012??00- 00?100??0110?????0?00???????1??????????????1?01-??????????????????????2111112- 1000???00201?111?000?0??0?-1000??-000110????????????-?0000????00?- 0?000??010?010120100?0100?110???????????????????? Pelagosaurus._typus 1011??0100100012000200?01?0001000111??001111201221021?101?21020010121211 1011??011000010?010010101101111211000003?0100000100??00011???11000?0?101? 0??0011110?111?000011001010?010000000000?001001111010000001010200{0 1}00101000-001000000110110-00010000001000001?-20-00011-010- 10000101000??200002110000211???10011113000010002011000100020021000101000 100000100001000010011000210000010000001010010000001100000100000000?0-

214

001000001-00000?1000000000100101??01001?0000?11- 0?10?100?01?10000001012011112-1000011001{0 1}001110000000?10000101?- 00011000000000??010000000??110010000000010?012120100??1001??????0120002?? ???0???? Sarcosuchus._imperator 10111111011010?210?2???11?010110?0?00100?111201221121?1010?11?00111212111 110??????00?10??1?0201?01101?02?0011003?1?10001011???1?110??1?00??001?????? 0011110??????1???1001000?1102001000??100100000??00000001010010000020000- 000000100001001110010000000---011?0110021010001?010001010001020000000- 000201210100111120000101020010?10?0011001000101010100000101001- 0??10?000??0??0000100100--0??000000001102?0100110000100-000000001- 00??0000??000?0???00????0????0???01100???????0?031000000010?2?11102- 00000101020101?10000100?100010000100?10101101000???1101000?11100010?0010 0010?0121211??????01??????00200121-1110???? Pholidosaurus._purbeckensis 1?1???110110???210?2?1?11?0101?0??10??00111?301321121?1?10??1???1112121111 00????2010110?110??011?1??1002?001121????010??001???????????100??1?10?????00 ?0????1??1?0??010010????10?00??0?0-100100000--000?0001010010010021000- 000000100?01?0??1002?00?0?0---?10?011?121010001?01?00101001??2?0?0??0- 000201???100?001?0000101020010??00001100?00010101010??00101?01?0?????000?? ??1000??0000???0????000?0011120??????0000000-?00??0001- 00?????0??00??????????????0??????????????0??????310?00?001??2?01?0??1???01?10 00????10?00??0?0?30?00021?0?2?01?????0????????????????????1?010?01???1??????? ?????1??????????????????0???? Dyrosaurus._phosphaticus 1?1?11010110101200121??11?01011000010?00111120132112101011211?0011221211 1100??111000110?1101201101101?02?00112130101001000101010110??0?1?0?011011 1010010110?1??1001111001010?01000000000-100101000-- 00000001110210000011000-000000100011001010010000000--- 000001101220100020110002010001020000000- 00030120?10021112001010122001001013011101310201110100000001001- 00?111000110021000100000--0?102000020112000010000000000-000000011- 00??011004000010?0000-200000100000110111?0110010101?0010?10?2?11212- 10000010021101?1100?10000000000001000?01010000001100??100010011000??1010 0010?0121100??1?1?10???0100000????????0???? Crocodylus._niloticus 1?1?110101101012101212011?0101000010011010123113211210111020120111001211 1100??111000110?1101111101200000000112130100011100100011110?011100110101 110000101?0?1111001001000110?01010100000-100100000-- 00000001010010000010000-020010000001000210010000000--- 000101111200000001010001000011120000000-0004-10-- 11011012000010112001101010011000000001110100000101001- 10110000001?0002001000000000000000010110000000000102010-000001001- 0000000213200100?10000210000100100110011101100102001000010000-01112- 10000111010101?10000100?100010000000102101000010111110100110011111001000 011010111000001012010110000220102011010????

215

Alligator._mississippiensis 1?1?110101101012101212011?0101000010011010023113211210111020120111001111 1100??112000211?1101111211201000000112130100011100100011110?011100110101 1100001011011111001001000110101010100000-100100000-- 00010001010010000020000-020000000001000210010000000--- 000101111200000001010001010001120000100-0004- 120?11001012000010112001101010011000000001110100000101001- 10110000011?0002001000000000000000000110000000000102010-000001001- 1110000213200100?10000210000100100110011101100102001000010000- 01102010000101010101?11110100?100010000000112101001010111110101110011111 001000011010101100001012010110?00220102011010???? Simosuchus._clarki 1111110111101012101212011?0101000110010010003011111210101120120010211111 1100??11110000?0110100001130100010011113???0000001???0{0 1}11?0?0??100110101110?????11?11?110???010100?10010210001?001000- 00001000000001010000000020000-010012112000000?00010000000--- 00110110100010000000020110000102100??00- 00020121210020010000111021010??1?0?1000?00100?1101???01011001?000???00010 ?0000000201010--0?01?0001001?00112000001201110010000001- 1100?0002300?????????0???????0????????????1100?0200?0?00?11??01101101010001 00101011000000100000010000010102101101010011111?110000101121000100-10- 0121?0001110211001000000010200100????? Baurusuchus._salgadoensis 1?1?11?1011010121012?2011?0001000110011110100013111211101020120010211111 11001?11110031?01111001201001?0130?1111301000000010??011110001?0001?0101 1??00?1011011111?011010000110001?0000101010?0- 00000010000101010010000010000-01?012111000000?010211000?0--- 00110111100010000?00020110100?021?01100- 1002012121100?0??0011100?100?1111101100?0?000?111?100010110011????0???011 1??1000?200?00???00??000?101100?0010000401000- 000000?001010?01002000100011??00?100000000001101101?110010000000000???211 11120100010?10111?1111111010100001000?-000{1 2}21010001100111100000?11?01?- 101000???0000211000011001?01?00001???00001000???? 'Gomphosuchus. wellsi (UCMP-97638/125871) '111110?1?1????1??0?212211- 111100??0?00020?0000022?11111001???0??1011??1111001111?0003??011101102111 ?00?13?011???????????????????????????????????????????????????????010100?0?0001 00001000?0???000?0?00????0?0100???0002????-??0000?????0000- 0?10000000???????1?10?0?101-?0?000101111100??1?21-- 00??0?2101011002?0100000110010000011?0???0010000001?1100000000100??101000 00010000000000010--?000??00?000--0100100003200112000000101- 00??????????????????????????????????????????????????????????0011100?1000101001 010111?0001???0?0??010????00010001??011001021100011110000100001110- 00???????????????????????????????0????

216

'"Edentosuchus" (UCMP125358)' ??11???101?01?12???2???11- 0?0100001?0001????1002101??0?00?201?00?01010???1?011??10???1???1?0?0??11??0 000?101111??????00?0????????????????????????????????????????????????????0????00 0??0??????0000???0????0?0?00??000011000??1?0?01?1?00?00?1??000?00?0--- ?011011??0?010100?000?01100?0???????????????????????????????010??1?0???10???1 ???1000????1????000?0??00???????????????00?0???01?????00???0?000-- ?????????????0?????0????1- ??0????0???????????????????????????????????????????????0?????01000??????????02? ???11???00?0?0?00?00001000021??????00???0??????????00?00???????1????????????? ??????????????????????0???? Orthosuchus_stormbergi 101110?1011010121012???11- 0?11000010000100001001201000?00120000?1020??1111001111{1 2}0003111?100??12?1120000?1011113010110010100?0010101002000110111??00101 111011?11011101110010000000000000001-100000000001000?010010010021000-0- 0000101?00?00-000000?00?0---00110100000010000?0000011000001110?--?0-0??1- 01000002110-00001?00100?00111001000100000000?110000000?01-- 1?1??3?????0?000??00?00??0??00000-000--0100101105?01?0-000000001- ?00?000000000?0?00000010001000?0100000?010110010110000?0?000011121000000 00100201?0111000100- 002000000100?0110010??011011010100?111110?000?????10???20110??1?0200?100?? 0??000??????0???? Shamosuchus__djadochtaensis 1-?-1101111010121012???11-0101000010010100?{1 2}3?132112101010201?01110111111100101120???0??1101- ??????0000?30011213???000010?????????0??11?0?????01??????1???0?1???0?1?010?? 01??01000100000-11-100000--00000001010010000010000- 00000010000100021001?0000?0---000?01110210000000010101010011?20000000- 0?03010-- 10001?12000010112001001010010000000001110100000101001?10????000110?00?0?? 00000000?02???0000110??00000001010?0-0000??001- 110?0?021220???????????00??0?00???????????11?0??100?00?0?110?- 1110102???0??00011?1?10010000310100000010002111101?00011?1?211???0??10?00 0?00001?0001?1??????102??????100??0??????1?0???? Hsisosuchus_chungkingensis 10111??1011010121012????1- 0?010001100011?0??10?1201210?0102012001011101111001?1110?0?1?0010?20??11? 010003001?11????0000?01???001???????0?011010111?????????????101??0?000?????1 0?00000010?1??00000000?000001010010000011000-00?00{1 2}11100?000- 0001001?0?0---?002001-001000000?01000110000??2??0?--0-1?0?- 11?????01????000100000000??1?00100??00100?0?1000?10100?01- ?0?1??211?1???0?01??0?00??0?00???00001??0?0?????0121??0- 0?00?0??0010?????0?00??0???????0??0?000?0?????????1??????0100000?0110?211101 0?1???0???0201?0110000100?0100000000001011?1????00??11?100??0???00010??0?0 ?00?1?1?1???????0??0??????010010???????????

217

Figure 1: The holotype material of Junggarsuchus sloani: A, the skull of Junggarsuchus in left lateral view; B, the cervical vertebrae and shoulder of Junggarsuchus in left lateral view in the block B); C, the right forelimb of Junggarsuchus in right lateral view. Labels and details of these elements are shown in the following figures.

218

Figure 2: Generalized tree of crocodylomorph relationships. Thalattosuchia is placed in two positions within Crocodyliformes, and as the sister group to crocodyliforms. Sphenosuchia shown as a paraphyletic grade and Protosuchian as a monophyletic group. The 3 hypothesized placements of Thalattosuchia are in red. Crocodyliformes in blue, Crocodylomorpha in red.

219

Figure 3: Life reconstruction of Junggarsuchus sloani. Image produced by Portia Sloan Rollings in consultation with James Clark.

220

Figure 4. The skull of A, Junggarsuchus sloani and Dibothrosuchus elaphros: Junggarsuchus skull in lateral view; B, Dibothrosuchus rostrum in lateral view; C, Tooth serrations on the 4th maxillary tooth of Junggarsuchus. Abbreviations: anf antorbital fenestra; antf antorbital fossa; boc basioccipital condyle; bs basisphenoid; cc crista cranii; cpt capitate process; dpf descending process of the prefrontal; ect ectopterygoid f frontal; fm foramen magnum; itf infratemporal fenestra; j jugal; l lacrimal; lf lacrimal fenestra; ls laterosphenoid; m maxilla; n nasal; nf nutrient foramina; or orbit; oto otoccipital; parietal; pb palpebral; pdt pit for dentary tooth; pf prefrontal; 221 pl palatine; pm premaxilla; po postorbital; pop paroccipital process; pt pterygoid; q quadrate; qf quadrate fenestra; qj quadratojugal; s squamosal; so supraoccipital; v vomer; vps ventral process of the squamosal.

222

Figure 5: The skull of A, Junggarsuchus sloani; B, and Dibothrosuchus elaphros in dorsal view.

C, Alternative interpretation of the postorbtial, frontal, squamosal contact in dorsal view; D, alternative interpretations of squamosal - postorbital contact in left lateral view. Abbreviations: dect dorsal process of ectopterygoid; fr frontal ridges; ftoa fenestra for temporal orbital artery; pfo prefrontal overhang; po2 postorbital 2; po/f? Postorbital -frontal fragment?; pro prootic; r-q right displaced quadrate; s2 squamosal 2; so supraoccipital; stf supratemporal fenestra; stfo supratemporal fossa; vl ventral process of lacrimal.

223

Figure 6: The skull of A, Junggarsuchus; B, and Dibothrosuchus, in ventral view.

Abbreviations: apf additional palatine fenestra?; apl anterior process of the palatine; apt anterior process of the pterygoid; bo basioccipital; bpt basipterygoid process; ch choana; em edentulous portion of the maxilla; jg jugal ventral groove; plr palatine rod; ppl posteriir process of the palatine; prq pterygoid ramus of the quadrate; qrp quadrate ramus of the pterygoid; sof suborbital fenestra, tpt transverse process of the pterygoid, vd displaced vomer.

224

225

Figure 7: Occipital view of the skulls of A, Junggarsuchus; B, and Dibothrosuchus.

Abbreviations: crn cranioquadrate canal; fm foramen magnum; otr otic recess; oto/bo otoccipital or basioccipital ventral portion?; ptf: posterior temporal fenestra; XII exit for cranial nerve 12.

226

227

Figure 8: Junggarsuchus brain case in A, left lateral; B, right lateral; C, anterior view; D, Ventral view; E, basisphenoid in anterior view; F, dorsal view. Abbreviations hypf hypophyseal fossa;

IV cranial nerve 4; lsvf ventral fossa of the laterosphenoid

228

229

Figure 9: Ear region of Junggarsuchus. A, ear region in right ventral view, B, prootic in ventral view; C, alternative interpretation of prootic quadrate contact in anterior view; D, prootic in dorso medial view showing the medial openings of the semicircular canal through the prootic.

Abbreviations: adq suture for dorsal head of the quadrate on the prootic; aoto suture for otoccipital on prootic; apf anterior prootic foramen; as suture of prootic squamosal; aso suture of prootic with supraoccipital; ci crista interfenestralis; cpprof crista prootica foramin; dmq dorsomedial process of quadrate on prootic; dq dorsal head of the quadrate; dvt dorsal vestibule: fo fenestra ovalis; op opisthotic; oscc3 opening for the third semicircular canal; scc medial wall enclosing semicircular canal in prootic; ssp subscapular process; tri trigeminal nerve exit; trir trigeminal recess; vg+cn exit for cranial nerves and vagus nerve; VII exit for cranial nerve 7

230

Figure 10: Dibothrosuchus brain case and ear region in A, right ventrolateral view; B, anterior view; C, ventral view; D, dorsal view. Abbreviations: crista interfenestralis; cppro crista prootica; dq dorsal head of quadrate; fa facial antra; feu foramina for the eustachian tubes; fo foramen ovalis; ic inner carotid; IX-XI cranial nerves 9-11; ma mastoid antrum; mpq medial process of the quadrate; pc pituitary canal; ptr posterior tympanic recess; str superior tympanic recess; tri-V exit for trigeminal nerve; VII exit for cranial nerve 7

231

Figure 11: Mandible of Junggarsuchus in left lateral A) dorsal B) ventral view C). Alternative interpretations of surangular in left lateral view D) and alternative prearticular in left medial view

E), coronoid and splenial F) Abbreviations: ang angular; ar articular; atc atlas centrum; atna atlas neural arch; atns atlas neural spine; cor: coronoid; d dentary; ect ectopterygoid; mnf mandibular

232 fenestra; rap retroarticular process; r rib (displaced): rmp ridge for M.pterygoideus ventralis; sa surangular; sp splenial; srf surangular fenestra.

233

Figure 12: Mandible of Dibothrosuchus in A, left lateral; B, dorsal, C, ventral view. D, quadrate distal end contact with articular in dorsal view; E, coronoid and splenial in left dorsolateral view;

234

F. surangular/ quadrate fragment on medial side of left mandible in right lateral view.

Abbreviations: dmrap dorsomedial process of retroarticular process; sur/q surangular/ quadrate fragment.

Figure 13: Close up of the left scapula and coracoid of Junggarsuchus in A) lateral view and B) ventral view of the glenoid fossa and coracoid body. C) Right scapula and coracoid of

Dibothrosuchus Abbreviations: corc corocoid; gf glenoid fossa; pp postglenoid process; sc scapula

235

Figure 14. The left forelimb of Junggarsuchus sloani showing the elements as preserved in A) anterior and B) posterior view; C) proximal head of the humerus and D) articular ends of ulna and radius E) the right radius and ulna of Dibothrosuchus. Abbreviations: dh distal end of humerus; dr distal end of radius; du distal end of ulna; h humerus; hh head of humerus; ho humerus oval depression on head; mc metacarpals; prl proximal end of radiale; pul proximal end of ulnare; r radius; rl radiale; u ulna; ul ulnare

236

Figure 15. Close up of the left wrist bones of Junggarsuchus sloani. A) close up of right wrist bones in Dibothrosuchus B) Abbreviations: dc distal carpals; pi pistiform; rl radiale; ur ulnare.

237

Figure 16. A, the left manus of Junggarsuchus with the metacarpals labeled; B, right manus of

Dibothrosuchus. Note the phalange (pI) still attached to metacarpal I in Junggarsuchus.

238

Figure 17. The axis of Junggarsuchus in A) lateral B) anterior and C) posterior view.

Abbreviations: atin atlas intercentrum; atr atlas rib; ax axis; ct centrum; na neural arch; ns neural spine; od odontoid process; poz postzygapophysss; prz prezygapophysis;

239

Figure 18. A, the posterior cervical vertebrae of Junggarsuchus. Note the parapophysis on the complex neural arch and the hypophyses present on the ventral surface of each centrum B, and the full cervical series of Dibothrosuchus with articulated osteoderms. Cervical vertebrae numbered on specimen Abbreviations: at atlas; dia diapophyses; hya hyapophyses; ost osteoderms;

240

Figure 19. Close up view of the dorsal vertebrae of Junggarsuchus. Note the simple centrum and neural arch and the anteroposteriorly broad neural spines. The double-headed rib retains its natural contact with the parapophyses and diapophyses. Abbreviations: hri heads of rib; ri rib;

241

Figure 20. Strict consensus of 3 most parsimonious trees of 1799 steps with CI =.39 and RI =

.68 generated by TNT and FigTree. This analysis made use of Tennant’s characters, Postosuchus as the outgroup and equal weights. Scores given on nodes. Score supports under 20 are collapsed into polytomies.

242

Figure 21: Results of implied weight analysis using Tennant’s characters with Gracilisuchus as the outgroup and rouge taxa omitted. Scores given on nodes. Score supports under 20 are collapsed into polytomies.

243

Figure 22: Results of implied weight analysis without Tennant’s added characters. Postosuchus as the outgroup with rouge taxa omitted. Scores given on nodes. Score supports under 20 are collapsed into polytomies.

244

G-All- G- All- P- all-E P- omit- G- all-I G-omit-I P-all-I P- omit- I E omit-E E

CI .33 .34 .36 .37 Na Na Na Na

RI .588 .597 .641 .65 Na Na Na NA

Steps 2005 1975 1832 1799 Na Na Na Na

Table 1: CI, RI, step for analyses without the inclusion of Tennant’s characters. Maximum number of step 3907, minimum number 673. Abbreviations: G-All-E (Gracilisuchus outgroup, with all taxa with equal weights); G-Omit-E (Gracilisuchus outgroup with rouge taxa omitted and equal weights); P-All-E (Postosuchus outgroup with all taxa and equal weights): P-Omit-E (Postosuchus outgroup with rouge taxa omitted and equal weights). G-All-Im (Gracilisuchus outgroup with all taxa and implied weight (K=12)). G-Omit-I (Gracilisuchus outgroup with rouge taxa omitted and implied wieghts weights); P-All-Im (Postosuchus outgroup with all taxa and implied weights); P-omit-Im (Postosuchus outgroup with rouge taxa omitted and implied weights); CI Consistency index; RI Retention index

G-All- G- All- P- all- P- omit- G- all- G-omit- P-all-I P- omit- I E omit-E E E I I

CI .35 .36 .39 .39 Na Na Na Na

RI .619 .627 .672 .68 Na .Na .Na .Na

Steps 1570 1548 1429 1407 Na Na Na Na

Table 2: CI and RI numbers and steps for analyses found with the inclusion of Tennant’s characters. Maximum number of steps 3223 and minimum steps 553. Abbrevations: same as above.

245

Gra/all Gra/om Po/all/e Po/omi Gra/all/ Gra/omi Po/all/i Po/omi/im

eq i/eq q t/eq im /im m

Erpeto + Dyo (1) 72 0-poly Omit Omit 1 (80) 1 (74) Omit Omit

Lit + Terr 0-poly 0-poly 0-poly 0 0 0 0 0

Dib + Sph 1 (39) 1 (47) 1 (45) 0 1 (56) 1 (60) 1 (54) 1

Dib/Sphen+S 0-poly 0-poly 0 0 1 (32) 1 (46) 1 (30) 1 (43)

Sph+Dib+Sol 0-poly 0-poly 0-poly 0 0 0 0 0

Dib+ Solido 0-poly 0-poly 0-poly 0 0 0 0 0

Solidocrania 0-poly 1 (64) 0 1 (63) 1 (59) 1 (83) 1 (62) 1 (84)

Almad+Mac 0-poly 0 0 0 0 0 0 0

Mac/Al+Cro 0-poly 0 0 0 0 0 0 0

Mac+Al+Cro 0-poly 1 (59) 0 1 (55) 1 (64) 1 (77) 1 (63) 1 (77)

Almad+Croc 0-poly 1 (57) 0 1 (51) 1 (58) 1 (58) 1 (55) 1 (59)

Proto-mono 1 (40) 1 (28) 1 (48) 1 (38) 1 (29) 0 1 (34) 1 (29)

Proto-para 0 0 0 0 0 1 - 0 0

Protos 1 (56) 1 (46) 1 (29) 1 (50) 1 (59) 1 (54) 1 (57) 1 (56)

Gobi 1 (91) 1 (94) 1 (94) 1 (91) 1 (98) 1 (94) 1 (98) 1 (98)

Shart 1 (45) 1 (51) 1 (51 1 (52) 1 (54) 1 (56) 1 (56) 1 (55)

Th out Croc 0 0 1 (79) 0 0 0 0 0

Th in Croc 1 (89) 1 (87) 0 1 (86) 1 (92) 1 (93) 1 (90) 1 (89)

Meso+Th 0-poly 1 (49) 0 0-poly 1 (64) 0 1 (55) 1 (61)

Longirostrine 0-poly 0 0 0-poly 0 1 (36) 0 0

246

Table 3: Groups found in analyses without Tennant’s Characters. 0 indicates absence and 1 presence of the group. When group is recovered, node support is given in parenthesis (xx). Abbreviations: Erpet +

Dyo (Erpetosuchus + Dyoplax); Lit + Terr (Litargosuchus + Terrestrisuchus ); Dib + Sph

(Dibothrosuchus + Sphenosuchus); Dib + Sphen+S (Dibothrosuchus + Sphenosuchus)+Solidorcania);

Dib + Solido (Dibothrosuchus + Solidocrania); Almad + Mac (Almadasuchus + Macelognathus);

Almad/Mac+Cro (Almadasuchus + Macelognathus )+ Crocodyliformes); Mac+Alam+ Cro

(Macelognathus + (Almadaschus + Crocodyliformes)); Almad + Croc (Almadasuchus +

Crocodyliformes); Proto-mono (monophyletic ‘Protosuchia’); Proto-para (paraphyletic ‘Protosuchia’)- support can’t be given; Protos Protosuchidae (Orthosuchus+ Protosuchus); Gobi Gobiosuchidae; Shart

Shartegosuchidae; Th out Croc (Crocodyliformes where Thalattosuchia is sister group of

Crocodyliformes); Th in Croc (Crocodyliformes where Thalattosuchia is in Crocodyliformes); Meso + Th

(Thalattosuchia as the sister group to mesoeucrocodylia); poly polytomy.

247

Gra/all Gra/om Po/all/ Po/omit/ Gra/all/ Gra/o Po/all/i Po/omi/i

eq i/eq eq eq im mi/im m m

Erpeto + Dyo 1 (88) 0-poly Omit Omit 1 (88) 1 (84) omit Omit

Lit + Terr 1 (42) 1 (39) 0-poly 0 1 (51) 1 (50) 0 1 (29)

Dib + Sph 0-Poly 0-poly 0-poly 0-poly 1 (31) 0 1 (33) 0

Dib 0-poly 0-poly 0-poly 1 (40) 1 (23) 0 1 (28) 0

/Sphen+S

Sph+Dib+Sol 0-poly 0-poly 0-poly 0-poly 0 1 (39) 0 1 (60)

Dib + Solido 0-poly 0-poly 0-poly 0-poly 0 1 (28) 0 1 (33)

Solidocrania 0-poly 1 (55) 0-poly 1 (59) 1 (44) 1 (64) 1 (55) 1 (81)

Almad+Mac 0-poly 0-poly 0-poly 0 1 (37) 1 (42) 1 (36) 1 (40)

Mac/Alm+Cr 0-poly 0-poly 0-poly 0 1 (49) 1 (66) 1 (54) 1 (78) o

Mac+Alm+C 0-poly 0-poly 0-poly 1 (56) 0 0 0 0 ro

Almad+Croc 0-poly 1 (60) 0-poly 1 (44) 0 0 0 0

Proto-mono 0-poly 0-poly 0-poly 1 (26) 0 0 0 0

Proto-para 0-poly 0-poly 0-poly 0 1- 1- 1- 1-

Protos- 0-poly 0-poly 0-poly 0 0 0 0 0

Gobi- 1 (81) 1 (82) 1 (81) 1 (78) 1 (96) 1 (97) 1 (97) 1 (97)

Shart- 1 (36) 1 (35) 1 (35) 1 (38) 1 (50) 1 (49) 1 (55) 1 (49)

Th in Croc 1 (81) 1 (86) 1 (81) 1 (86) 1 (90) 1 (88) 1 (92) 1 (90)

Th + Neo 0-poly 0-poly 1 (64) 1 (69) 1 (63) 1 (67) 1 (69) 1 (77)

248

Table 4: Groups found in analyses with Tennant’s characters. 0 codes for absence, 1. codes for presence.

Abbreviations: Th + Neo (Thalattosuchia as the sister group to Neosuchia)

Group Synapomorphies

Litargosuchus + 9 (0), 18 (0), 81 (2), 93 (1), 171 (2), 217 (1), 482 (2), 518 (1) Terrestrisuchus: 4/16 Dibothrosuchus + 26 (1), 94 (1), 99 (2), 318 (1), 436(0)

Sphenosuchus: 8/16

Dibothrosuchus + 34 (1), 35 (1), 40 (1), 60 (1), 141(0), 148 (0), 256 (1), 305 (1),

Solidocrania: 2/18 531 (1)

Dibothrosuchus 6 (1), 27 (1), 234 (1), 263 (0), 484 (2) autapomorphies

Junggarsuchus 31 (1), 32 (1), 37 (1), 38 (1), 85 (0), 102 (1), 118 (1), 119 (1), autapomorphies 121 (1), 125 (1), 126 (1), 128 (1), 222 (1), 233 (1), 247 (1),

272 (2), 273 (1), 276 (1), 427 (1), 430 (1), 477 (0), 488 (2), 598

(1)

‘Solidocraniua’: 12/16 20 (2), 29(1), 30(1), 48(1), 58 (0), 61(1), 62 (2), 78 (1); 109(1), 122 (1), 138(0) 144(1), 204(1), 211 (0), 254 (1), 313 (1), 316 (1), 322 (1), 376 (1), 442 (1)

Macelognathus + 114 (1), 131 (1), 132 (0), 580 (0)

Almadasuchus: 4/16

Maceloganthus/Almadasuchus 13 (1), 25 (1), 49 (2), 51 (1), 73 (1), 74 (1), 80 (1), 108 (1), 110 (1), 231 (0), 341 (1), 343 (0), 352 (1), 418 (1), + Crocodyliformes: 6/16

Thalattosuchia wo 13(1), 22(2), 24(2), 51 (0), 74 (0), 75(0), 79(1), 110 (0), 214(0),

Crocodyliformes: 1/16 215(2), 227 (0), 262(0), 324(1), 327(1), 338(1), 479 (1), 527(0),

249

Crocodyliformes + 1(1), 12(0), 14(0), 59(2), 65(1),390(0) 404(0),

Thalattosuchia: 1/16

Protosuchia monophyly 22(1), 69(0), 105 (1), 168(1), 175(0), 240(0), 285 (1), 303 (2) 337(0), 338 (1), 342(1), 350(1) 384 (0), 414(1), 479 (0), 492(1), 504 (1), 508(1), 543(1),

Crocodyliformes with 1(1), 18(0), 22(2), 24(2), 59(2), 65(1),79(1), 90(1), 214(0),

Thalattosuchia 326(0),390(0)

Mesoeucrocodylia: 2(0), 50(1), 52(1-2), 53(1), 57(1), 580 (0)

Thalattosuchia + 29(0), 70(2), 101(1), 149(0), 261(0), 282(2), 296(1), 347(1),

Mesoeucrocodylia 391(1), 486 (2), 497(1), 501(1), 525(1), 548(0), 549(0), 561(0)

Thalattosuchia: 16/16 10(0), 21(0) 28(0), 104(1), 140(0), 146(0), 163(0), 169(1),

197(1), 208(1), 230(1), 234(1), 269(1), 287(2), 289(1), 330(1),

335(1), 365(1-2), 381(1), 403(0)*, 412(1), 426(1), 439(1),

516(0), 519(1), 527(0), 538 (0), 569(2), 571(1), 574(0), 577(0)

Thalattosuchia + Neosuchia : 40 (0), 226 (0), 228 (0), 229 (0), 275 (0), 276 (1), 316 (2), 319

(1), 335 (1), 395 (0), 448 (1), 512 (1), 547 (0), 548 (0), 576 (1),

588 (1), 597 (1)

Longirostrine: 1/16 42 (1), 104 (1), 165(1), 169 (1), 269 (1), 335 (1), 403 (0), 405

(0), 426 (1)

Table 5: Synapomorphies for groups found in all or at least all but one analyses. Recovered

Synapomorphies of major groups--The characters listed as synapomorphies above are found in at least all but one analyses, or in each analysis. The number next to the name is the number of times that the group was found out of 16 analyses. Unambiguous characters found in all analyses in bold.

250