This is the accepted version of the journal article: Parra, Sara; Sellés, Albert G. «New cranial remains of the broad-nosed crocodile Elosuchus (Pholidosauridae; Mesoeucocodrylia) and its palaeoecological implica- tions». Historical Biology, (October 2022). DOI 10.1080/08912963.2022.2130791 This version is available at https://ddd.uab.cat/record/266787 under the terms of the license 1 New cranial remains of the broad-nosed crocodile Elosuchus (Pholidosauridae; 2 Mesoeucocodrylia) and its paleoecological implications 3 4 Parra, Sara1 & Sellés, Albert1,2,* 5 6 1 Institut Català de Paleontologia Miquel Crusafont – ICTA-ICP, Edifici Z, C/ de les 7 Columnes s/n. Campus Universitat Autònoma de Barcelona, 08193 Cerdanyola del Vallès 8 2 Museu de la Conca Dellà, c/Museu 4, E-25650 Isona, Lleida, Spain 9 10 *To whom correspondence should be addressed. E-mail: albert.garcia@icp.cat 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 1 26 Abstract 27 Sometimes Natural History Museums unknowingly treasure singular specimens in their 28 collections for decades. The re-discovery of such elements allows the description of 29 previously unknown features of well-known taxa. Here, we describe a nearly complete left 30 premaxilla attributed to the pholidosaurid Elosuchus. Specimen IPS3303 shows some 31 remarkable differences with previously known premaxillae of Elosuchus, especially 32 concerning the relative position of the fifth premaxillary tooth and the premaxilla-maxilla 33 sutural surface morphology. Because of the scarcity of current data, it is difficult to evaluate 34 the significance of such differences. The described element is about 40% larger than the 35 largest premaxilla of Elosuchus known so far, suggesting that this taxon could achieve 36 remarkable body size proportions. The preliminary analyses of the dentition and the 37 neurovascular system of the premaxilla provide new insight into the paleoecology of this 38 riverine crocodylomorph from the Cretaceous of Africa. 39 40 Keywords: Pholidosauridae, Elosuchus, paleoecology, Cretaceous, Museum Collection. 41 42 43 INTRODUCTION 44 Aside from performing public outreach and educational missions, one of the pivotal points of 45 any Natural History Museum is to curate unique specimens that hold critical information for 46 evolutionary sciences. Due to the large number of specimens that usually manage NHM 47 collections, sometimes accounting for thousands of elements, it is not surprising that some 48 specimens “get lost” or remain hidden for decades. There are several examples of that, but 49 once re-discovery and re-evaluation, these hidden treasures either come out as new species or 50 additional specimens of previously described taxa (e.g. Boersma and Pyenson 2016; Chapelle 2 51 et al. 2019; Borths and Stevens 2019; Smitha et al. 2021; Spiekman et al. 2021), while other 52 times they represent new remains of an already known but rare taxon (i.e. Prieto-Márquez 53 and Wargne 2013; Prieto-Márquez and Gutarra 2016). 54 Funded in 1965 as the Institut de Paleontologia de Sabadell (IPS) and re-named as 55 Institut Català de Paleontologia Miquel Crusafont (ICP) in 2006, the ICP has acted as a pool 56 of attraction and a state reference in the preservation of vertebrate fossils from all eras. 57 Consequently, it is not rare that time to time enthusiastic, amateur, and private collectors end 58 up giving their collections to the institution. To date, the ICP collection accounts for up to 59 126,000 specimens. In 1992, the ICP Museum received the private collection of an amateur 60 paleontologist from Barcelona (Sr. Andreu), who donated two specimens –a left premaxilla 61 of a crocodylomorph (IPS3303) and a tooth of Spinosaurus sp. (IPS3304)– from an unknown 62 locality of the southern Er-Rachdia Province, close to the Morocco-Algeria boundary (J. 63 Agustí per. com.). The two specimens remained stored in the museum collections for more 64 than 30 years, without anyone knowing of its existence since a review made by one of the 65 authors (A.S) brought the remains to light. 66 Here, we describe a new nearly complete premaxilla (IPS3303) attributed to the 67 pholidosaurid Elosuchus. Despite the fragmentary nature of the finding, it is the largest 68 premaxilla known for this taxon so far and provides new clues about the paleoecology of this 69 pholidosaurid species. 70 Institutional abbreviations. –– ICP-Museum of Institut Català de Paleontologia - Miquel 71 Crusafont, Sabadell, Spain; MNHN-Muséum national d'Histoire Naturelle, Paris, France. 72 73 MATERIAL AND METHODS 74 The element of interest of the present study is a nearly complete left premaxilla of a putative 75 large crocodylomorph (IPS3303), which is housed in the paleontological collections of the 3 76 Museum of the Institut Català de Paleontologia Miquel Crusafont. All anatomical measures 77 were taken using a digital caliper Mitutoyo 15CPX with a range of error of 0.005 mm. 78 Anatomical descriptions follows the morphological nomenclature of previous studies on 79 pholidosaurid crocodylomorphs (Jouve 2005; Young et al. 2016; Meunier and Larsson 2017). 80 A three-dimensional model of the studied fossil specimen was produced following the 81 photogrammetric protocol of Mallison and Noble (2016), and using Agisoft Photoscan Pro (v. 82 1.2.4, www.agisoft.com), to perform scaling and alignment. 83 Alanyses of the topographic relieve and neurobascular densty were perfomerd using 84 ImageJ software (Schneider et al. 2012) 85 86 SYSTEMATIC PALAEONTOLOGY 87 Superorder: Crocodylomorpha Hay 1930 88 Clade: Neosuchia Benton & Clark 1988 89 Family: Pholidosauridae von Zittel & Eastman 1902 90 Genus: Elosuchus de Lapparent de Broin 2002 91 Elosuchus sp. 92 93 Specimen 94 IPS3303, a nearly complete left premaxilla. 95 96 Locality and age 97 Unknown fossil site near the southern border of Er Rachida Province, Morocco. Kem Kem 98 Group; late Albian-early Cenomanian (see Ibrahim et al. 2020 for further details). 99 100 4 101 Description and Comparison 102 IPS3303 is a nearly complete left premaxilla, only lacking the most rostromedial part of the 103 bone, including most of the first and half of the second premaxillary alveolar region, and the 104 ventral palatine process. 105 In dorsal view, the main body of the premaxilla is strongly sculptured with numerous 106 ridges and pits that vary in size and shape. The anterior margin of the premaxilla is rounded 107 in shape, and transversally broad (Fig 1). The maximal width of the element is at the level of 108 the fourth premaxillary alveolus, as in Elosuchus (de Lapparent de Broin 2002; Young et al. 109 2016), Meridiosuchus (Fortier et al. 2011), Pholidosaurus (Martin et al. 2016), and 110 Oceanosuchus (Hua et al. 2007). Except from the embayment of the external nares, the 111 medial side of the bone is nearlt streaigh. The resulting “half pan-like” shape of IPS3303 112 resembles that of Pholidosaurus (Martin et al. 2016), Chalawan (Martin et al. 2013), 113 Meridiosaurus (Fortier et al. 2011), and Elosuchus (de Lapparent de Broin 2002; Young et al. 114 2016). Despite the absence of nasal bone, given that the premaxilla-nasal sutural surface 115 extends until the posterior edge of the external nares (Fig 1), it is likely that the premaxilla 116 did not contact each other medially and therefore nasals contributed to the posterior margin of 117 the opening of the nares. Such anatomical configuration is, so far, only reported in Elosuchus 118 (de Lapparent de Broin 2002; Young et al. 2016). 119 The dorsally facing external naris is more expanded lateromedially than 120 anteroposteriorly, occupying about 33% of the maximum transverse width of the premaxilla 121 (Fig 1). This feature is shared by other broad-nose pholidosaurids such as Oceanosuchus 122 (Hua et al. 2007), Terminonaris (Shimada and Parris 2007), and Elosuchus (Young et al. 123 2016), but not as developed as in Sarcosuchus (Sereno et al. 2001) and Chalawan (Martin et 124 al. 2013). The anterior margin of the external nares exhibits a small posteriorly directed bump 125 (Fig 1) like in Terminonaris (Shimada and Parris 2007) and Elosuchus (Young et al. 2016). 5 126 A set of neurovascular foramina arranged in a semicircular distribution surrounds the 127 lateral margin of the external nares. A prominent rim surrounds the anterior margin of the 128 external nares (Fig 1), as in the most well-preserved specimen of Elosuchus (MNHN.F SAM 129 129 and MNHN.F MRS 334; Young et al. 2016). An enlarged, laterodorsally semi-rounded 130 fossa for the accommodation of the first dentary teeth, named “elosuchid fossa” by Mourier 131 and Larsson (2016), is located in the anterior half of the premaxilla, near the lateral edge of 132 the bone (Fig 1), a feature only reported in the African pholidosaurid Elosuchus (de 133 Lapparent de Broin 2002; Young et al. 2016). 134 In palatal view, the premaxilla displays five alveoli oriented lateroventrally (Fig 1), as 135 Pholidosaurus, Terminonaris, and Elosuchus (Shimada and Parris 2007; Young et al. 2016; 136 Martin et al. 2016). The morphology of the premaxilla and alveoli form a broad rosette 137 dentition. Despite being partially broken, the second alveolus (P2) seems to be the smallest of 138 the premaxillary series, followed by the P5, P4 and P3 being the largest one as in 139 Pholidosaurus (Martin et al. 2016) and Elosuchus (Young et al. 2016). The first two alveoli 140 (P1 and P2) are located anteromedially to the large “elosuchid fossa” (Fig 1), as in Elosuchus 141 (Young et al. 2016). This fossa, intended to the reception of the enlarged first dentary tooth 142 (D1), separates the P2 from the P3 alveolus (Fig 1). Although similar notches are present in 143 other pholidosaurids (e.g. Mook 1933; Sereno et al. 2001; Lepage et al. 2008; Fortier et al. 144 2011), the position of this opening that splits the premaxillary tooth series into distinct units 145 is considered autapomorphic of Elosuchus (Young et al. 2016). 146 The third and fourth premaxillary alveoli are widely separated from each other by a 147 deep concave interalveolar space. P5 is located posteromedially relative to the P4 alveolus, as 148 in Elosuchus (de Lapparent de Broin 2002; Fortier et al. 2011; Young et al. 2016), and 149 Chalawan thailandicus (Martin et al. 2013), while P5 is posterolateral to the P4 alveoli in 150 most pholidosaurids (e.g. Mook 1934; Sereno et al. 2001; Wu et al. 2001; Hua et al. 2007; 6 151 Lepage et al. 2008; Fortier et al. 2011). The posterolateral position of the fifth premaxillary 152 tooth regarding the tooth row was proposed as a synapomorphy of Pholidosauridae by Fortier 153 et al. (2011), but the medial ‘migration’ of the last premaxillary alveolus could be 154 characteristic of a more inclusive clade within pholidosaurids (Young et al. 2016). The gap 155 between the P4 and P5 alveoli is proportionally smaller than the P3–P4 interalveolar space. 156 This condition is similar to that of MNHN.F INA 30 (Young et al. 2016), while the P4-P5 157 interalveolar space seems much larger in E. cherifiensis (MNHN.F MRS 334; de Lapparent 158 de Broin 2002). The interalveolar space between P3-P4, and P4-P5 alveoli would have 159 accommodated the second and third dentary teeth respectively. 160 Only the third replacement tooth is preserved in place in IPS3303 (Fig 1), although 161 fragments of the fifth premaxillary tooth are observed within the alveolus. The third 162 premaxillary tooth is conical in shape, and circular in cross-section (Fig 1). While most of the 163 crown is covered by smooth enamel with fine apicobasal striation, the apex is ornamented 164 with anastomosing enamel (Fig 1). There is no evidence of carinae. This ornamental pattern 165 resembles that of Elosuchus and Chalawan (de Lapparent de Broin 2002; Martin et al. 2013) 166 and differs from the curved cone-like tooth with carinae and apicobasal ridges of 167 Terminonaris, Pholidosaurus, and Sarcosuchus (Sereno et al. 2001; Shimada and Parris 168 2007; Martin et al. 2016). It is worth noting that the dental features of IPS3303 and Elosuchus 169 strongly resemble that of the thalattosuchid Machimosaurus (Young et al. 2014). 170 The ventral palatine process of the premaxilla is broken, and its extinction can only be 171 assessed from the scar left on the medial edge of the palatine, which expands from the 172 posterior margin of the “elosuchid fossa” to the level of the posterior margin of the P4. A 173 series of neurovascular foramina are present on the palatal shelves, mainly surrounding the 174 medial margins of the alveoli. The posterior process tapers posteriorly to the fifth 175 premaxillary alveolus. Three large foramina are located at the anterior premaxilla-maxilla 7 176 sutural surface, which seems to be related to the trigeminal nerve (CN V; Fig 1). The 177 trigeminal nerve seems to extend along all the anteroposterior length of the premaxilla. 178 In lateral view, the premaxilla is beak-shaped, with the anterior margin oriented 179 anteroventrally (Fig 1), a characteristic feature of pholidosaurids. Several neurovascular 180 foramina cover most of the lateral surface of the bone. The premaxilla-maxilla sutural 181 surface, which is curved posterodorsally, is restricted to the most posterior part of the lateral 182 edge of the premaxilla (Fig 1). 183 Finally, the straight premaxilla-nasal sutural surface is the most salient structure in the 184 medial side of the premaxilla, occupying more than half of the total length of the element and 185 reaching at the level of P4 (Fig 1).This sutural surface is depicted by a strongly sculptured 186 rectangular surface consisting of anteroventrally ridges (Fig 1). The inner margin of the 187 external naris displays five neurovascular foramina, being the largest one located in the 188 posterior border. 189 The combination of the above-described character in IPS3303–transversally broad 190 beak-like premaxilla, strong posterior constriction of the posterior process, rosette-like 191 alveolar region, a large “elosuchid fossa” and nasal contributing to the posterior margin of 192 external nares–allow us to tentatively assign the studied specimen to the pholidosaurid genus 193 Elosuchus. 194 195 [FIGURE 1 SHOULD BE PLACED HERE] 196 197 DISCUSSION 198 Taxonomic attribution and morphological variability 199 Pholidosauridae is a group of specialized longirostrine crocodylomorphs adapted to aquatic 200 environments (Hua et al. 2007). Throughout their evolutionary history, pholidosaurid 8 201 achieved a wide geographic distribution, being present in Europe, Africa, North America, and 202 South America. It is also worth noting that the family expanded several millions of years, 203 from 167Ma to 61Ma, proving that the clades survived to the K-Pg biotic mass extinction 204 (Jouve and Jalil 2020). 205 Elosuchus is perhaps one of the most controversial pholidosaurid taxa described so far 206 due to its complex history. The original material was collected in 1951 by the French 207 paleontologist René Lavocat to the north of Gara Sbaa (SE of Morocco). Initially identified 208 as ‘Thoracosaurus’ cherifiensis (Lavocat 1955), the type material was re-described years 209 later by de Lapparent de Broin (2002), who concluded that the taxon discovered by Lavocat 210 should be attributed to a new genus: Elosuchus. In the same study, de Lapparent de Broin 211 described a new elosuchid species –‘Elosuchus’ felixi– and erected the family ‘Elosuchidae’, 212 which initially included E. cherifiensis, ‘E’. felixi (currently Fortignathus felixi, Young et al. 213 2016) and Stolokrosuchus lapparenti (Larsson and Gado 2000). However, Jouve (2005) was 214 the first in noting that such grouping was inconsistent. He observed that Stolokrosuchus 215 lacked the ‘elosuchid fossa’, that several alleged premaxillary ‘elosuchid’ characters were 216 also apomorphie of Peirosauridae, and that E. cherifiensis –as defined by de Lapparent de 217 Broin (2002)– shared several characters with basal dyrosaurids. 218 In a posterior re-evaluation of the genus, Meunier and Larsson (2017) proposed to 219 restrict E. cherifiensis to Morocco, defined E. broinae from Argelia, and questioned the 220 validity of ‘E’. felixi. By the same time, the latter taxon was re-assessed by Young et al. 221 (2016) to the genus Fortignatus. In consequence, the purported family ’Elosuchidae’ sensu de 222 Lapparent de Broin (2002) is currently questioned by several authors (Jouve 2005; Young et 223 al. 2016; Meunier and Larsson 2017). 224 Therefore, the genus Elosuchus currently includes two species: E. cherifiensis and E. 225 broinae. Ibrahim et al. (2020) noted that they only differ in minor anatomical characters, but 9 226 none of them pertaining to the morphology of the premaxilla (Young et al. 2016). As a result, 227 the attribution of IPS3303 to a lower rank than the genus level is impossible here. 228 Nevertheless, it is worth noting that there are some anatomical differences between 229 IPS3303 and other premaxillae attributed to Elosuchus (Fig 2). First, the intra-alveolar space 230 between P4 and P5 of IPS3303 is similar in size to MNHN.F INA 30, but they both are 231 smaller than in MNHN.F MRS 334. This feature can be quantified as the distance between 232 the posterior margin of the P4 alveolus and the anterior edge of the alveolus P5 (Fig 2). In 233 turn, the difference in the spacing is also reflected in the relative position of the fifth 234 premaxillary alveolus, which is located more medially in MNHN.F MRS 334 and FSAC-KK 235 10 than in MNHN.F INA 30 and IPS 3303 (Fig. 2) 236 In addition, in dorsal view, the premaxilla-maxilla sutural surface in IPS3303 differs 237 from the other known Elosuchus premaxillae (Fig 2). In that it extends anterolaterally in the 238 IPS specimen, while in other specimens seems limited to the axial plane of the snout. With 239 the current data at hand, it is difficult to establish if such differences may correspond to 240 interspecific variations (even sexual dimorphism), ontogenetic changes, or taxonomic 241 differences. Whatsoever, the unexpected discovery of a large premaxilla attributed to the 242 pholidosaurid Elosuchus in the IPS collections provides new ecological information to this 243 enigmatic taxon. 244 245 [FIGURE 2 SHOULD BE PLACED HERE] 246 247 Body size 248 Body size is one of the most fundamental biological traits because it affects virtually all 249 aspects of an organism’s physiology and ecology. To date, with a basal skull length of 94 cm 250 (measured from the tip of snout to the posterior side of the occipital condyle in MNHN SAM 10 251 129; Meunier and Larsson 2017), Elosuchus is considered a mid-length pholidosaurid, almost 252 a half of its relative Sarcosuchus (skull length of 178 cm; Sereno et al. 2001). 253 Judging by the graphic information available of the most complete premaxillae (see 254 de Lapparent de Broin 2002; Meunier and Larsson 2007), MNHN SAM 129 has a total 255 anteroposterior length of about 16.3 cm, being that of MNHN SAM 334 just slightly smaller. 256 In this regard, IPS3303 is the largest premaxilla attributed to the genus Elosuchus so far, 257 being 143.7% larger than MNHN SAM 129 (see Supplementary Table 1). 258 By assuming a similar cranial morphological ontogenetic trajectory of Elousuchus 259 than the extant long-snout Gavialis gangeticus (Piras et al. 2010; Hone et al. 2020) for 260 Elosuchus, the total skull length of IPS3303 would likely measure about 135 cm in length, 261 almost twice the skull size of the current Nile crocodiles. If so, Elosuchus could be as large as 262 Chalawan (Martin et al. 2013) but smaller than the gigantic skull of Sarcosuchus (Sereno et 263 al. 2001). Consequently, Elosuchus could be recognized as the third-largest species of 264 pholidosaurid known so far. 265 266 Paleoecological implications 267 The occurrence of a premaxillary tooth in the IPS3303 provides new insights into the 268 potential dental variability and feeding strategy in Elosuchus. Overall, the premaxillary tooth 269 of IPS3003 resembles the maxillary teeth of MNHN MRS 1112 and SAM 129. All of them 270 show a similar conical-shaped crown with apical ornamented enamel consisting of 271 anastomosing ridges, but the premaxillary tooth does not have carinae as the maxillary ones, 272 it is less curved vocally, and the anastomosing texture covers more surface of the tooth crown 273 (Fig 3). 274 Similar morphology differences are recognized along the jaw in extant crocodiles, 275 putative semiaquatic dinosaurs (Ibrahim et al. 2014), and mosasaurs (Street et al. 2021). In 11 276 fact, it is noteworthy that the false denticulations and anastomosing enamel texture are 277 produced by the same amelogenesis process (Sander 1999, 2000) –in which the underlying 278 dentine does not contribute to the relieve development, but as a differential thickening of the 279 enamel–, and because of that it is not surprising to find them coexisting in the same tooth. 280 While the anastomosing enamel texture covering the apical region of the crowns favor 281 the increase of the grip during biting, the conical morphology of the premaxilla teeth are ideal 282 for gripping, crushing, and pulling. The combination of these features suggests that the 283 premaxillary teeth of Elosuchus were able to deal with hard-biting elements (i.e. bones or 284 shells). In contrast, the ziphodont maxillary teeth seem primarily intended to slice the meat 285 (Hornung and Reich 2015). Such heterodonty is consistent with a bite-and-swallow feeding 286 style (O’Connor et al. 2010), similar to that of the extant Crocodylus niloticus. 287 [FIGURE 3 SHOULD BE PLACED HERE] 288 289 The external surface of IPS3303 is strongly sculpured with several sinuous grooves 290 (Fig 4a,c), some of them allocating neurovascular foramina of different sizes and shapes (Fig 291 4b,d). The distribution of these neurovascular foramina is uneven and they are predominantly 292 concentrated around the labial margin of the mouth (n=44; Fig 4b,d) and surrounding the 293 external nares (n=14; Fig 4d). This topographic distribution resembles that of extant 294 crocodiles, whose foramina are associated with a high-resolution mechanosensory system 295 linked to the terminal innervations of the ophthalmic premaxillary branch (CN V1) of the 296 trigeminal nerve (Leitch and Catania 2012; George and Holliday 2013). Pending further 297 analyses, in IPS3303 the preserved ophthalmic division of trigeminal nerve (CN V1) cavities 298 extend rostrally through all the craniocaudal length of the premaxilla (Fig 4) and suggest the 299 presence of a complex neurovascular system intender for detecting waterborne vibrations and 12 300 optimal temperature of nesting site (Brazaitis and Watanabe 2011; George and Holliday 301 2013), as in extant crocodylians. 302 The combination of relatively large body size, a highly sensitive sensorial system, and 303 a specialized dentition leads suggest that Elosuchus could likely be a remarkable ambusher 304 predator of its time. Hidden underwater, Elosuchus could feel vibrations and pressure 305 changes produced by its potential prey underwater. Once attacking, it could grasp its prey 306 with the bulbous massive conic premaxillary teeth, until it pull part of the prey and sliced 307 with the posterior ziphodont dentition. Although its primary food source could be fishes, it is 308 also likely that Elosuchus could easily prey upon mid-sized vertebrates, such as turtles –the 309 Kem Kem Group has plenty of turtles remains of Dirqadim Hamadachelys, or Galianemys– 310 or even small dinosaurs –most likely young Rebbachisaurus. 311 [FIGURE 4 SHOULD BE PLACED HERE] 312 313 ACKNOWLEDGEMENTS 314 The present work has been supported by the Spanish Agencia Estatal de Investigación 315 (CGL2016-76431-P, CGL2016-73230-P, and CGL2017-82654-P) and the Generalitat de 316 Catalunya (CERCA Program). This research is part of the project I+D+i/PID2020- 317 119811GB-I00 funded by MCIN/ AEI/10.13039/501100011033/. ASG research is supported 318 by the project VIGEOCULT (PLEC2021-00793) founded by 319 MCIN/AEI/10.13039/501100011033 and by the European Union NextGeneration EU/ PRTR. 320 The authors further acknowledge the collaboration Dr. Bernat Vila, from the ICP institute, for 321 his assistance in the photogrammetric virtual reconstruction of the studied specimen. 322 323 324 13 325 REFERENCES 326 Boersma A, Pyenson ND. 2016. Arktocara yakataga, a new fossil odontocete (Mammalia, 327 Cetacea) from the Oligocene of Alaska and the antiquity of Platanistoidea. PeerJ. 4: 328 e2321. doi:10.7717/peerj.2321. 329 Borths MR, Stevens NJ. 2019. Simbakubwa kutokaafrika, gen. et sp. nov. (Hyainailourinae, 330 Hyaenodonta, 'Creodonta,' Mammalia), a gigantic carnivore from the earliest Miocene of 331 Kenya. Journal of Vertebrate Paleontology: e1570222. 332 doi:10.1080/02724634.2019.1570222. 333 Brazaitis P, Watanabe ME. 2011. Crocodilian behaviour: a window to dinosaur behaviour?. 334 Historical Biology. 23 (1): 73–90. doi:10.1080/08912963.2011.560723. 335 Chapelle KEJ, Barrett PM, Botha J, Choiniere JN. 2019. Ngwevu intloko: a new early 336 sauropodomorph dinosaur from the Lower Jurassic Elliot Formation of South Africa and 337 comments on cranial ontogeny in Massospondylus carinatus. PeerJ. 7: e7240. doi: 338 10.7717/peerj.7240. 339 George ID, Holliday CM. 2013. Trigeminal nerve morphology in Alligator mississippiensis 340 and its significance for crocodyliform facial sensation and evolution. Anatomical 341 Records. 296 (4): 670–680. doi: 10.1002/ar.22666. 342 Hone D, Mallon JC, Hennessey P, Witmer LM. 2020. Ontogeny of a sexually selected 343 structure in an extant archosaur Gavialis gangeticus (Pseudosuchia: Crocodylia) with 344 implications for sexual dimorphism in dinosaurs. PeerJ. 8: e9134. doi: 345 10.7717/peerj.9134. 346 Hornung JJ, Reich M. 2015. Tylosaurine mosasaurs (Squamata) from the Late Cretaceous of 347 northern Germany. Netherlands Journal of Geosciences Geologie en Mijnbouw. 94 (1): 348 55-71. doi: 10.1017/njg.2014.31. 14 349 Hua S, Buffetaut E, Legall C, Rogron P. 2007. Oceanosuchus boecensis n. gen, n. sp., a 350 marine pholidosaurid (Crocodylia, Mesosuchia) from the Lower Cenomanian of 351 Normandy (western France). Bulletin de la société géologique de France. 178 (6): 503– 352 513. doi:10.2113/gssgfbull.178.6.503. 353 Ibrahim N, Sereno PC, Varricchio DJ, Martill DM, Dutheil DB, Unwin DM, Baidder L, 354 Larsson HCE, Zouhri S, Kaoukaya A. 2020. Geology and paleontology of the Upper 355 Cretaceous Kem Kem Group of eastern Morocco. ZooKeys. 928: 1–216. doi: 356 10.3897/zookeys.928.47517. 357 Lapparent de Broin F. 2002. Elosuchus, a new genus of crocodile from the Lower Cretaceous 358 of the North of Africa. Comptes Rendus Palevol. 1 (5): 275–285. doi: 10.1016/S1631- 359 0683(02)00049-0. 360 Lavocat R. 1955. Découverte d'un Crocodilien du genre Thoracosaurus dans le Crétacé 361 Supérieur d'Afrique. Bulletin du Muséum National d'Histoire Naturelle, Paris, Serie 2. 27 362 (4): 338-340. 363 Leitch DB, Catania KC. 2012. Structure, innervation, and response properties of 364 integumentary sensory organs in crocodilians. The Journal of Experimental Biology. 215 365 (3): 4217-4230. doi:10.1242/jeb.076836. 366 Mallison H, Wings O. 2014. Photogrammetry in Paleontology – a Practical Guide. Journal of 367 Paleontological Techniques. 12: 1-31. 368 Martin JE, Lauprasert K, Buffetaut E, Liard R, Suteethorn V. 2013. A large pholidosaurid in 369 the Phu Kradung Formation of north-eastern Thailand. Palaeontology. 57 (4): 757-769. 370 doi:10.1111/pala.12086. 371 Meunier LMV, Larsson HCE. 2017. Revision and phylogenetic affinities of Elosuchus 372 (Crocodyliformes). Zoological Journal of the Linnean Society. 179: 169–200. doi: 373 10.1111/zoj.12448. 15 374 O’Connor PM, Sertich JW, Stevens NJ, Roberts EM, Gottfried MD, Hieronymus TL, Jinnah 375 ZA, Ridgely R, Ngasala SE, Temba J. 2010. The evolution of mammal-like 376 crocodyliforms in the Cretaceous Period of Gondwana. Nature. 466 (7307): 748-751. 377 doi:10.1038/nature09061. 378 Piras P, Colangelo P, Adams DC, Buscalioni A, Cubo J, Kotsakis T, Meloro C, Raia P. 2010. 379 The Gavialis–Tomistoma debate: the contribution of skull ontogenetic allometry and 380 growth trajectories to the study of crocodylian relationships. Evolution & Development. 381 12 (6): 568-579. doi: 10.1111/j.1525-142X.2010.00442.x. 382 Prieto-Márquez A, Gutarra S. 2016. The ‘duck-billed’ dinosaurs of Careless Creek (Upper 383 Cretaceous of Montana, USA), with comments on hadrosaurid ontogeny. Journal of 384 Paleontology. 90 (1): 133–146. doi: 10.1017/jpa.2016.42. 385 Prieto-Márquez A, Wagner JR. 2013. The ‘Unicorn’ Dinosaur That Wasn’t: A New 386 Reconstruction of the Crest of Tsintaosaurus and the Early Evolution of the 387 Lambeosaurine Crest and Rostrum. PLoS ONE. 8 (11): e82268. 388 doi:10.1371/journal.pone.0082268. 389 Sander PM. 1999. The microstructure of reptilian tooth enamel: terminology, function, and 390 phylogeny. Münchner Geowissenschaftliche Abhandlungen (A). 38:1–102. 391 Sander PM. 2000. Prismless enamel in amniotes: Terminology, function and evolution. in: 392 Teaford MF, Smith MM, Ferguson MWJ (eds.), Development, Function and Evolution 393 of Teeth. Cambridge University Press, Cambridge. pp. 92–106. 394 Schneider CA, Rasband WS, Eliceiri KW. 2012. NIH Image to ImageJ: 25 years of image 395 analysis. Nature Methods. 9: 671-675. doi:10.1038/nmeth.2089. 396 Sereno PC, Larson HCE, Sidor CA, Gado B. 2001. The Giant Crocodyliform Sarcosuchus 397 from the Cretaceous of Africa. Science. 294 (5546): 1516–9. 398 doi:10.1126/science.106652. 16 399 Smitha RE, Martill DM, Unwin DM, Steel L. 2021. Edentulous pterosaurs from the 400 Cambridge Greensand (Cretaceous) of eastern England with a review of Ornithostoma 401 Seeley, 1871. Proceedings of the Geologists' Association. 132 (1): 110-126. 402 doi:10.1016/j.pgeola.2020.10.004. 403 Spiekman SN, Ezcurra MD, Butler RJ, Fraser NC, Maidment SC. 2021. Pendraig milnerae, a 404 new small-sized coelophysoid theropod from the Late Triassic of Wales. Royal Society 405 Open Science. 8 (10): 210915. doi:10.1098/rsos.210915. 406 Street HP, LeBlanc ARH, Caldwell MW. 2021. A histological investigation of dental crown 407 characters used in mosasaur phylogenetic analyses. 83Vertebrate Anatomy Morphology 408 Palaeontology. 9:83–94. doi:10.18435/vamp29372. 409 Young MT, Steel L, Brusatte SL, Foffa D, Lepage Y. 2014. Tooth serration morphologies in 410 the genus Machimosaurus (Crocodylomorpha, Thalattosuchia) from the Late Jurassic of 411 Europe. Royal Society Open Science. 1 (3): 1140269140269. doi: 10.1098/rsos.140269. 412 Young MT, Hastings AK, Allain R, Smith TJ. 2016. Revision of the enigmatic crocodyliform 413 Elosuchus felixi de Lapparent de Broin, 2002 from the Lower-Upper Cretaceous 414 boundary of Niger: potential evidence for an early origin of the clade Dyrosauridae. 415 Zoological Journal of the Linnean Society. 179: 377-403. doi: 10.1111/zoj.12452. 416 417 418 FIGURE CAPTIONS 419 420 Figure 1. IPS3303. Left premaxilla of Elosuchus sp. in a) dorsal, b) palatal, d) medial, and e) 421 lateral view. c) Detail of the posterior process of the premaxilla showing the anatomic 422 configuration of the maxilla and nasal sutural surface and details of the neurovascular 423 foramina. f) Close-up view of the third premaxillary replacement tooth showing the 17 424 characteristic anastomosing wrinkle of the enamel at the apex of the tooth (white arrow) 425 running until the mid-high of the tooth crown (black arrow). 426 The black triangle in “a” indicates the possible location of an incipient anterior narial bar. 427 Abbreviations: dD1,dD2,dD3-diastema for the reception of the dentary teeth; en-external 428 nares; if- inferior fossa ;nvf-neurovascular foramen; pp-palatine process; px-mx- premaxilla- 429 maxila sutural surface; px-n- premaxilla-nasal sutural surface; P1 to P5 -premaxillary tooth 430 position; r-rimtgn-V- trigeminal nerve or cranial nerve V. [double-column; color only in 431 electronic version] 432 433 Figure 2. Features comparison between premaxillae attributed to the pholidosaurid 434 Elosuchus in palatal (upper row) and dorsal (lower row) views. Dashed black lines and 435 triangles illustrate the lateromedial displacement of P5 regarding P4. Dashed grey lines and 436 triangles show the craniocaudal spacing between P4 and P5. Dashed black lines and triangles 437 in the lower row show the layout of the premaxilla-maxilla sutural surface in dorsal view. 438 Schemes are not at scale. Abbreviations: mx-maxilla; n-nasal; px-premaxilla. [double- 439 column] 440 441 Figure 3. Dental features of Elosuchus. Schematic illustration showing the dental 442 morphology variations on the upper jaw of Elosuchus, in both having marginal carinae (fc) 443 and relative contribution of enamel types within each tooth. Note that anastomosing enamel 444 has a relative more contribution on the apex of the premaxillary tooth than in maxillary one. 445 Tooth schemes are not at scale. [one-column] 446 447 Figure 4. Neurovascular system distribution in IPS3303. Colored topographic relieve 448 showing the distribution of the sculpturing grooves in (a) dorsal and (c) lateral views. Red 18 449 points illustrate the presence of a neurovascular foramina (NVF). Density map distribution of 450 neurovascular foramina upon a tridimensional model in the (b) dorsal surface and (d) the 451 lateral surface of the premaxilla. Note the higher concentration of foramina around the 452 external nares and the labial margin of the bone, just above the alveolar region. [double- 453 column; color only in electronic version] 19 Supplementary Table 1. Measures of premaxillae attributed to the genus Elosuchus. Width measurements were taken considering just one premaxillary bone (half of the total rostrum width). All measures are in cm. Measures are taken from: de Lapparent de Broin, 2002; Meunier and Larsson, 2016; Young et al. 2016; Ibrahim et al., 2017. *Values extrapolated by using ImageJ software. **Value calculated using linear regression. IPS3303 MNHN SAM MNHN MRS MNHN FINA FSAC-KK 129 334 30 10 max. length 23.43 16.45* 16.98* - - max. width 11.25 8.93* 11.22* 6.87* - min. width 5.28 3.99* 4.6* 3.06* 4.28* (posterior to P5) max. nares length 5.07 5.27* 5.31* 5.26* - max. nares width 3.2 2.7* 3.01* 2.40* - Skull length 135.12* 94* - - - * 1