Theropoda: Spinosauridae) from the Middle Cretaceous of Morocco and Implications for Spinosaur Ecology

Cretaceous Research 93 (2019) 129e142

Contents lists available at ScienceDirect

Cretaceous Research

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

Juvenile spinosaurs (Theropoda: Spinosauridae) from the middle Cretaceous of Morocco and implications for spinosaur ecology

* Rebecca J. Lakin , Nicholas R. Longrich

Department of Biology and Biochemistry, Milner Centre for Evolution, University of Bath, Bath, BA2 7AY, UK article info abstract

Article history: The Spinosauridae is a specialised clade of theropod dinosaurs known from the Berriasian to the Cen- Received 29 May 2018 omanian of Africa, South America, Europe and Asia. Spinosaurs were unusual among non-avian di- Received in revised form nosaurs in exploiting a piscivorous niche within riverine and estuarine habitats, and they include the 24 August 2018 largest known theropod. Although fossils of giant spinosaurs are increasingly well-represented in the Accepted in revised form 18 September fossil record, little juvenile material has been described. Here, we describe new examples of juvenile 2018 Available online 19 September 2018 spinosaurines from the middle Cretaceous (Cenomanian) Kem Kem beds of Morocco. The fossils include material from a range of sizes and are relatively common within the Kem Kem deposits, suggesting that Keywords: juveniles exploited the same semiaquatic niche as the adults throughout ontogeny. This implies that the Dinosauria Cenomanian delta habitats supported an age-inclusive population of spinosaurs that was neither Spinosauridae geographically or environmentally separated, though some ecological separation between juveniles and Juvenile adults is likely based on the large variation in size. Bones or teeth of very small (<2 m) spinosaurs have Ontogeny not been found, however. This could represent a taphonomic bias, or potentially an ecological signal that Cretaceous the earliest ontogenetic stages inhabited distinct environments. Skeletal remains include individuals referable to Sigilmassasaurus brevicollis and a second spinosaurine taxon. Consistent with this, we also identify two distinct cranial morphs of Spinosauridae present within the Kem Kem, supporting previous recognition of distinct taxa in the assemblage. © 2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

1. Introduction cervical vertebrae (Evers et al., 2015), but also recently from portions of the skull (Hendrickx et al., 2016). These are Sigilmassasaurus The Spinosauridae is a diverse family of theropod dinosaurs brevicollis (¼Spinosaurus maroccanus; Evers et al., 2015) and a including ten described genera (Taquet & Russell, 1998; Buffetaut & second spinosaurid taxon, possibly Spinosaurus (Hendrickx et al., Ouaja, 2002; Kellner et al., 2011; Allain et al., 2012; Buffetaut, 2012; 2016) or a relative (Evers et al., 2015). Malafaia et al., 2013; Evers et al., 2015) found across ﬁve continents Spinosaurs are thought to have been piscivorous on the basis of (Ruiz-Omenaca~ et al., 2005; Keller et al., 2011; Barrett et al., 2011; their elongated, crocodile-like skulls and their conical, spearing Buffetaut, 2012; Allain et al., 2012; Candeiro et al. 2017; Gasca teeth (Sues et al., 2002; Buffetaut, 2007; Hasegawa et al., 2010; et al., 2018). Spinosaurs are known from the Berriasian (Sales Richter et al., 2013). Furthermore, an intricate series of canals et al., 2017) to the Cenomanian (Buffetaut & Ouaja, 2002), and are across the distal parts of the upper and lower jaws, interpreted as a particularly common in the middle Cretaceous (Cenomanian) aged system of pressure receptors, have been cited as an adaptation for deposits of North Africa, including Egypt (Stromer, 1915), Algeria hunting aquatic prey (Ibrahim et al., 2014b; Vullo et al., 2016; Arden (Taquet & Russell, 1998; Benyoucef et al., 2015), and Morocco et al., 2018), though this is disputed (Leich & Catania, 2012; Barker (Buffetaut, 1989; Ibrahim et al., 2014a, 2014b). Within the Kem Kem et al., 2017; Henderson, 2018). The skeletal morphology of spino- region are two distinct morphologies, originally assigned only from saurines suggest a semiaquatic lifestyle, similar to extant crocodilians (Amiot et al., 2010; Ibrahim et al., 2014b; Goedert et al., 2016; Aureliano et al., 2018). Adaptations for aquatic locomotion include retracted nares (Dal Sasso et al., 2005), a reduced pelvis and * Corresponding author. E-mail addresses: [email protected] (R.J. Lakin), [email protected] hindlimbs, and pachyostotic long bones (Ibrahim et al., 2014b; (N.R. Longrich). Aureliano et al., 2018). https://doi.org/10.1016/j.cretres.2018.09.012 0195-6671/© 2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). 130 R.J. Lakin, N.R. Longrich / Cretaceous Research 93 (2019) 129e142

Spinosaurs are the largest known theropod dinosaurs (Dal Sasso 3. Materials & methods et al., 2005; Therrien & Henderson, 2007; Gimsa et al., 2016). The largest specimen currently known is MSNM V4047 referred to The material presented here consists of three cervical vertebrae Spinosaurus aegyptiacus (Dal Sasso et al., 2005), which has an (FSAC-KK-7280, FSAC-KK-18119, FSAC-KK-18122), two dorsal verte- estimated skull length of 175 cm and an estimated body length of brae (FSAC-KK-18118, FSAC-KK-18121), a right premaxilla (FSAC-KK- 12e15 m (Therrien & Henderson, 2007). Other specimens, 7281) and a right quadrate (FSAC-KK-18120). We describe our including the holotype of S. aegyptiacus, have estimated total body vertebrae using the laminar nomenclature laid out in Wilson (1999). lengths of 10e14 m (Therrien & Henderson, 2007; Ibrahim et al., Specimens were acquired from commercial sources in the region of 2014b). Like all dinosaurs, however, giant spinosaurs would have Erfoud and Begaa (Fig. 1C) between October 2017 and March 2018 hatched from relatively small eggs before ultimately growing to (Fig.1B) Of the numerous Kem Kem outcrops in the region (Fig.1C), it adult size. Fossils of juvenile dinosaurs are generally very rare is difficult to know from which locality any of our isolated specimens (Lockley, 1994; Hone & Rauhut, 2010), and juvenile spinosaurs have originated, and given that the material is likely to have been only recently been described for the first time (Maganuco & Dal reworked, it is also difficult to know where in the environment these Sasso, 2018). However, little is still known about the growth and specimens may have originally been deposited post-mortem. How- life histories of giant spinosaurs. Small spinosaurine teeth are ever, the horizon can be verified from the taphonomic differences common from Kem Kem deposits (Richter et al., 2013), however, between the Ifezouane and Aoufous Formations. In the Ifezouane these cannot be reliably used for size or age estimations because of Formation, preservation is generally good, with bones preserved as the variation in spinosaur teeth within regions of the jaw (Monfroy, white, red fossils, or dark colored fossils in sandy matrix, often with 2017), and the polyphyodonts nature of all theropods (Erickson, bright red clays (Krassilov & Bacchia, 2013). In the Aoufous Forma- 1996; Therrien et al., 2005). tion, preservation is generally poorer, with specimens preserved in Here, we describe new material of juvenile spinosaurs. These more powdery yellow to tan coloured matrix (Krassilov & Bacchia, include a quadrate, cervical and dorsal vertebrae, and a premaxilla, 2013). These differences provide an independent means of veri- all found in the Kem Kem beds of eastern Morocco near the towns fying provenance and mean that fossils can typically be assigned to of Begaa and Erfoud. one of the two Kem Kem subunits with some confidence. We estimated body length by isometrically scaling from the proportions of the spinosaur reconstruction published in Ibrahim 2. Geological & palaeoecological setting et al. (2014b). Due to the controversial nature of this reconstruction at the time of publication (Scott & Currie, 2017), and ongoing The Kem Kem beds represent the upper part of the informally- research into the locomotory and lifestyle adaptations of Spinosau- named ‘Continental Intercalaire’ (Rodriguez et al., 2011; rus, we also reconstruct body length estimations from the bipedal Benyoucef et al., 2015). The Kem Kem beds were originally reconstruction by Dal Sasso et al. (2005). This was done by cali- thought to be early Cretaceous in age, but the current consensus is brating the length, width and height of our juvenile specimens that this geological unit is Cenomanian in age based on both the (Table 1) with the corresponding known adult specimens from the shark fauna and faunal similarities with the Cenomanian aged literature (Stromer,1915; Dal Sasso et al., 2005; Ibrahim et al., 2014b) Bahariya Formation of Egypt (Sereno et al., 1996; Cavin et al., 2010). and scaling up by dividing the length, width and height of a juvenile The Kem Kem beds sit unconformably on marine strata of specimen with those of its corresponding adult specimen to calcu- Cambrian-Devonian age (Cavin et al., 2010; Ibrahim et al., 2014a). late a multiplication factor, then multiplying the estimated length of Above, they are overlain by cliff-forming Cenomanian-Turonian the adult model with the multiplicative inverse of the multiplication marine limestones (Martill et al., 2011), reflecting the pattern of factor. Measurements for all specimens are given in Table 1. increasing marine encroachment represented by the fining- Institutional Abbreviations: Names have been abbreviated as upwards sequence seen in the Kem Kem. follows throughout (FSAC) Faculte des Sciences Aïn Chock, Uni- Two stratigraphic subunits have been identified within the Kem versity Hassan II, Casablanca. Kem beds, the lower Ifezouane Formation, which outcrops exten- sively near the towns of Begaa and Taouz, and the upper Aoufous 4. Systematic palaeontology Formation, which is well exposed to the north near the towns of Erfoud, Aoufous, and Jorf (Cavin et al., 2010). The specimens we 4.1. Sigilmassasaurus brevicollis describe primarily come from the Ifezouane Formation, with the remainder from the overlying Aoufous Formation (Cavin et al., Dinosauria Owen, 1842 2010). In the Ifezouane Formation, coarse, vertebrate-rich sand- Saurischia Seeley, 1887 stone red beds make up the majority of the formation (Cavin et al., Theropoda Marsh, 1881 2015), with rare conglomerates. These patterns of deposition sug- Spinosauridae Stromer, 1915 gest a habitat dominated by episodic flooding and drought (Russell Sigilmassasaurus brevicollis Russell, 1996 & Paesler, 2003). Oxygen isotope data from the lower unit vertebrates seems to affirm this, with seasonal changes of habitat Referred specimens. FSAC-KK-18122. (aquatic to terrestrial, and vice versa) being reflected in the remains Horizon. Lower Member (Ifezouane Formation) of the ?Cenomanian of Spinosauridae (Amiot et al., 2010; Goedert et al., 2016). The fauna Kem Kem beds; specimen FSAC-KK-18122. itself has been the subject of recent interest, with many sources Locality. Near Erfoud, Draa-Ta^ filalet region, Morocco; specimen finding an ‘unbalanced’ food web, with an apparent glut of FSAC-KK-18122. terrestrial carnivores (McGowan & Dyke, 2009; Lang€ et al., 2013). Description. FSAC-KK-18122 is a posterior cervical vertebra (C9), Between the vertebrate-rich deposits of the lower Ifezouane complete except for the hypapophysis and dorsalmost point of the Formation and the upper Aoufous Formation is a pattern of current neural spine (Fig. 2C). The proportions of the centrum (Table 1)are ripples (Cavin et al., 2010) and a section of sandy clay/marlstone. similar to those of BSPG 2011 I 115, referred to Sigilmassasaurus The Aoufous Formation is made up of more powdery sandstones, (Evers et al., 2015). The parapophyses are pointed anteriorly, giving a claystones and marlstones, with a fauna comprised of a diverse slightly triangular shape in lateral view (Fig. 2A), and are set towards range of vertebrates as well as gastropods (Cavin et al., 2010). the anterior end of the centrum, below the ventral surface of the R.J. Lakin, N.R. Longrich / Cretaceous Research 93 (2019) 129e142 131

Fig. 1. Localities of the Kem Kem beds surrounding the region of Erfoud and Begaa, Errachidia Province, Southeastern Morocco. Map (B) provided by Google Maps. Geological outcrops adapted from Richter et al., (2013). Scale bar represents 20 km. centrum in lateral view (Fig. 2A & 2D). These meet approximately transverse processes with pneumatic openings at the base. Short, halfway down the ventral surface at the keel, which goes on to ventrally-inclined postzygopophyses meeting centrally at a very flatten into the ventral posterior rugose triangular platform char- short neural spine at the dorsal midline. Epipophyses absent acteristic of Sigilmassasaurus. This specimen has very large trans- (McFeeters et al., 2013; Evers et al., 2015). verse processes with deep pneumatic openings at the base, most notably on the posterior surface (Fig. 2F & 2L). The neural spine is incomplete, though even the most conservative extrapolation 4.2. cf. Spinosaurus aegyptiacus would indicate that the complete neural spine would be very short. In addition to these, the specimen also carries anteriorly-pointing Horizon. Ifezouane Formation (FSAC-KK-18118 and FSAC-KK-7280) prezygopophyses with dorsally-facing articular facies (Fig. 2B & and Aoufous Formation (FSAC-KK-18120 and FSAC-KK-7281), ? 2C). The interpostzygopophyseal laminae are strongly reduced, Cenomanian Kem Kem beds. and the vertebral foramen is very large compared to comparable Locality. Near Begaa, Er Rachidia, Morocco (FSAC-KK-18118 and specimens described in Evers et al., (2015), approximately 20% as FSAC-KK-7280), and near Erfoud, Draa-Ta^ filalet region, Morocco wide as the distance between the transverse processes. (FSAC-KK-18120 and FSAC-KK-7281). Comments. Evers identified two spinosaurs in the Kem Kem beds, Description. FSAC-KK-7280 is a cervical vertebra (C5 or C6; Sigilmassasaurus brevicollis, and a second taxon, which may be Fig. 3AeF). The neural arch is absent. The centrum is elongated referable to the same species as the Spinosaurus neotype skeleton. (length:height ratio ¼ 3.29), indicative of the Spinosauridae (Evers Based on the following synapomorphies we assign FSAC-KK-18122 et al., 2015), and the anterior articular facet of the centrum is almost to Sigilmassasaurus: longitudinally compressed posterior cervical round, with a height:width ratio of 1:1.2. The ventral midline nar- vertebra with condyle width almost as long at the centrum, and rows to a smooth keel and a ventral triangular plateau is absent. twice the height of the condyle. Very broad, ventrally-angled The ventral triangular plateau is diagnostic of Sigilmassasaurus (Evers et al., 2015). Each of the lateral surfaces carries a large central Table 1 pneumatic foramen. In lateral view (Fig. 3D), the pneumatic fora- Measurements for Kem Kem specimens, with standardised measurements, men is not obscured by the parapophysis, which, in FSAC-KK-7280, including: length (shortest distance from anteriormost point to posteriormost point is approximately the same size as the lateral foramen, and posi- of the specimen); height (shortest distance from the ventralmost point to the dor- tioned towards the anterior articular facet. Close examination of the salmost point of the specimen) and width (shortest distance from the leftmost lateral point to the rightmost lateral point, except for vertebrae, where this is taken external periosteum reveals a pattern of striations across the as the width of the cotyle, following Evers et al., 2015). ventral side of the vertebra (Fig. 3E). Due to the poor preservation of the specimen, other diagnostic features such as the broadened Specimen no. Length (mm) Height (mm) Width (mm) centroprezygapophyseal laminae, spinopre-postzygapophyseal FSAC-KK-7280 52 20 23 fossae and details of the neural spine cannot be interpreted. FSAC-KK-7281 110 85 36 FSAC-KK-18118 42 25 29 FSAC-KK-18119 70 65 21 FSAC-KK-18118 is an anterior dorsal vertebra. The neural arch is FSAC-KK-18120 57 78 42 missing and as in FSAC-KK-7280, the neurocentral sutures are open. FSAC-KK-18122 39 44 62 The centrum is relatively short (<1.5 as long as wide) and shal- FSAC-KK-18121 50 32 23 lowly opisthocoelous, with approximately circular articular faces. 132 R.J. Lakin, N.R. Longrich / Cretaceous Research 93 (2019) 129e142 R.J. Lakin, N.R. Longrich / Cretaceous Research 93 (2019) 129e142 133

The entire centrum is elongated. The lateral fossa is visible in lateral plane. Ventrally to this point, the flange recurves at a 70 angle view and is not obscured by the parapophysis, the lateral surface of (relative to the first ventral slope) towards the facet with which which, in FSAC-KK-18118, is approximately 1.4 larger than the the quadrate articulates with the mandible. On the lateral surface lateral pneumatic foramen. Like FSAC-KK-7280, there are striations of this facet is a concave articular surface for the quadratojugal. across the periosteum on the lateral and ventral sides. The ventral Ventral to this is the ecto- and endocondyle. The endocondyle is keel is sharply defined and lacks a posterior triangular plateau, but ventrally convex, and followed laterally by a shallow inter- towards the posterior end of the ventral surface, the sharp keel condylar sulcus and the ectocondyle, which is narrow and smooths out to a more rounded hull. tapering, giving the ventral profile of the quadrate a tadpole- shaped outline, with its thick and bulbous endocondyle tapering 4.3. Spinosauridae indet. to the ectocondyle ‘tail’. Dorsal to the quadratojugal articulation, the quadrate recurves proximally into the quadrate foramen. This Referred specimens. FSAC-KK-7281, FSAC-KK-18120, FSAC-KK-18119, feature is largely reduced, as in the juvenile material described by FSAC-KK-18121. Hendrickx et al., (2016). It has a shallow semilunar shape, and is Horizon. Upper Member (Aoufous Formation) of the ?Cenomanian positioned approximately 60% of the distance from the dorsal Kem Kem beds. edge of the quadrate. From the quadrate foramen the edge of the Locality. Near Erfoud, Draa-Ta^ filalet region, Morocco. foramen rises to the dorsal quadratojugal articulation before Description. FSAC-KK-7281 is a right premaxilla 8.31 cm in length, ascending almost vertically back to the squamosal articular facet. identified by the sharply curved shape of the anterior rosette in As well as the anteriormost portion of the pterygoid flange, the lateral view (Fig. 4A & 4B), which hooks proximally from along the ventral projection of the dorsal quadratojugal contact is also ventral-dorsal plane. In the corresponding portion of the dentary, missing. this is more bulbous (Stromer, 1915). The specimen bears numerous FSAC-KK-18119 is an incomplete mid-cervical vertebra of an neurovascular foramina along the distal and lateral portions of the indeterminate spinosaur. The preservation quality is very poor. premaxilla (Fig. 4F), and the distalmost part of the premaxilla ends The neural arch is partially missing, and most of the neural spine, ventrally in a sharp point (Fig. 4A, 7E), curving upwards towards the as well as the transverse process, postzygopophyses and part of posterior portion and narrowing caudally to a point just 70% the the prezygopophyses are absent (Fig. 6A & 6D). At the base of the height of the tallest point on the specimen. This creates a strongly incomplete neural spine, the attachment site is anterioposteriorly ventrally-recurved upper jaw (Vullo et al., 2016). In FSAC-KK-7281, elongated, similar to that of specimen BSPG 2006 I 57 (Evers et al., the anterior portion of the specimen has been reconstructed post- 2015), assigned to Spinosauridae indet. The periosteum of the excavation by fossil dealers. The anteriormost portion can be seen lateral surfaces is also poorly defined and does not appear to to have been reattached (Fig. 4B & 4E), presumably by commercial carry the striations seen on the other vertebrae (Fig. 6E). The dealers. articular surfaces of the centra are strongly opisthocoelous, and in the case of the anterior articular surface, the distance between Within the tooth layout, the first anterior pair of alveoli are the the body of the centrum and the anteriormost point of the smallest, and are separated by a space 23% the width of the largest articular surface is half the total diameter of the centrum (Fig. 6A alveolus. These are followed posteriorly by the largest pair of alveoli & 6B). The entire vertebra is mediolaterally crushed, with in the premaxilla, and within the smaller of these two alveoli is a distortion visible in both articular surfaces of the centrum and in partially erupted tooth 6e8 mm in diameter (Fig. 4G). It is elongate the neural arch (Fig. 6B & 6C). The ventral surface is noticeably and needle-shaped, with numerous narrow longitudinal striations flat, with the lateral curves around the centrum coming abruptly and is finely serrated along the posterior ridge. This second pair of to a more horizontal plane along the ventral surface, without the alveoli are followed by a third, smaller pair, approximately 60% the sharp keel present in the other vertebrae. This is a feature diameter of the larger pair, and the space separating the third pair potentially referable to S. brevicollis, however due to the preser- of alveoli is just 17% the diameter of each alveolus. Following this is vation quality of the specimen this is equivocal. In lateral view, a large diastema that separates the premaxillary teeth from the the ventral surface has a sigmoid-shaped profile (Fig. 6A & 6D), maxillary ones. This diastema corresponds to a dorsal incline in the with the ventral platform descending ventrally towards the roof of the palate (Fig. 4B), forming an arc in the distal part of the posterior end. rostrum (Dal Sasso et al., 2005). On the ventral side of the upper FSAC-KK-18121 is a partial anterior thoracic vertebra. The neural rostrum, there is a well-developed ossified ‘secondary palate’, arch is completely absent and the articular facet of the centrum is believed to brace the skull during predation (Holtz, 1998) and poorly defined. There is no central pneumatic foramen, nor is the formed from the maxilla. parapophysis present on the centrum (Fig. 6G & 6J), supporting the FSAC-KK-18120 is a right quadrate with well-defined articular diagnosis of an anterior thoracic vertebra. The centrum is highly facets on the ventral and lateral sides (Fig. 5). From the quadrate elongate, twice as long as wide. It is also tall, and the height of the head (sensu Hendrickx et al., 2016), the dorsal edge of the quad- posterior articular cotyle measures 50% more than the width rate descends anteriorly at a slope of approximately 42 (relative (Table 1). There is a highly rugose area on the posterior surface of to the posterior edge) towards the anteriormost point of the the ventral keel, approximately 1 cm from the posterior articular pterygoid flange (Fig. 5A). The pterygoid flange itself is incom- condyle (Fig. 6K). This could be used as a diagnosis for S. brevicollis, plete. In dorsal view, the pterygoid flange is distally convex, and at however, the proportions of the centrum, as in FSAC-KK-18119, are the widest point, the quadrate measures 5.30 cm along the sagittal contradictory and render this feature equivocal.

Fig. 2. Posterior cervical from Sigilmassasaurus brevicollis (FSAC-KK-18122) in lateral right (A), dorsal (B), anterior (C), lateral left (D), ventral (E) and posterior (F). ns, neural spine; pozp, postzygopophysis; przp, prezygopophysis; tp, transverse process; pps, parapophysis; apf, anterior pneumatic foramen; ppf, posterior pneumatic foramen. Laminae high- lighted following Wilson (1999): Red, prespinal lamina; yellow, spinopostzygopophyseal lamina; purple, spinoprezygopophyseal lamina; blue, prezygodiapophyseal lamina; light green, postzygodiapophyseal lamina; orange, centroprezygopophyseal lamina; dark green, anterior centrodiapophyseal lamina; dark pink, intrapostzygopophyseal lamina; light pink, osterior centrodiapophyseal lamina. Scale bar represents 3 cm. (For interpretation of the references to color/colour in this ﬁgure legend, the reader is referred to the Web version of this article.) 134 R.J. Lakin, N.R. Longrich / Cretaceous Research 93 (2019) 129e142

Fig. 3. Vertebrae from cf. Spinosaurus aegyptiacus from the Cenomanian of Morocco. Mid-cervical (FSAC-KK-7280) in lateral right (A), dorsal (B), anterior (C), lateral left (D), ventral (E) and posterior (F) views. Anterior dorsal (FSAC-KK-18118) in lateral right (G), dorsal (H), anterior (I), lateral left (J), ventral (K) and posterior (L) views. Scale bar represents 3 cm.

5. Discussion transverse processes with deep pneumatic openings at the base; (6) small neural spine. 5.1. Identiﬁcation of spinosaur material FSAC-KK-7280: We refer this cervical vertebra to cf. Spinosaurus aegyptiacus based on the following features: (1) a highly elongated FSAC-KK-18122: We refer this cervical vertebra to Sigilmassa- centrum, with a width to length ratio of 1:1.96 (Based on the width saurus brevicollis based on autapomorphies laid out by Evers et al. of the anterior articular facet, length not including condyle; >1:1.6 (2015). This specimen is almost identical in its proportions to considered diagnostic for Spinosauridae; Evers et al., 2015); (2) BSPG 2011 I 115, referred to Sigilmassasaurus by Evers et al. (2015), absence of a ventral rugose triangular plateau, considered diag- and exhibits the following features: (1) anterodorsally angled nostic of Sigilmassasaurus brevicollis (Evers et al., 2015); (3) anterior centrum when posterior articular surface is positioned vertically; articular facet with a height to width ratio of 1:1.23 (>1:1.5 (2) anterior articular surface that exceeds the total width; (3) well- considered diagnostic for Sigilmassasaurus; Evers et al., 2015); (4) developed ventral keel that expands posteriorly into a rugose wide lateral central pneumatic foramen, with a width to height triangular platform; (4) concave parapophyseal facets (5) massive ratio of 1:1.4 (3:1 considered diagnostic for Sigilmassasaurus; Evers R.J. Lakin, N.R. Longrich / Cretaceous Research 93 (2019) 129e142 135

Fig. 4. Right premaxilla (FSAC-KK-7281) of a juvenile Spinosauridae indet. In lateral right (A), lateral left (B), ventral (C), dorsal (D) and anterior (D) view. Close-ups show neurovascular foramina and striation pattern of bone (F) and partially erupted tooth with anterior microstructures (G). p1ep5; premaxillary alveoli 1e5; pg, premaxillary groove; nvf, neurovascular foramina. Scale bars represent 5 cm (A-E) and 5 mm (F, G). et al., 2015); (5) parapophysis larger than the lateral foramen, as in carinae (Hasegawa et al., 2010); (3) teeth arranged in a terminal Spinosaurus aegyptiacus (Stromer, 1915). rosette (Fig. 7; Dal Sasso et al., 2005). FSAC-KK-18118: We refer this dorsal vertebra to cf. Spinosaurus FSAC-KK-18120: we refer this quadrate to Spinosauridae indet. aegyptiacus based on: (1) A highly elongated centrum, with a width based on the taxonomy of spinosaur quadrates laid out by to length ratio of 1:1.96 (Based on the width of the anterior articular Hendrickx et al. (2016). This study divided Moroccan spinosaurine facet, length not including condyle; >1:1.6 considered diagnostic quadrates into ‘Morphotype A’ (identified as Spinosaurus aegyp- for Spinosauridae; Evers et al., 2015); (2) the absence of a ventral tiacus), and ‘Morphotype B’ (?Sigilmassasaurus brevicollis). These rugose triangular plateau, considered diagnostic of Sigilmassasau- morphotypes are distinguished by the highly posteriorly recurved rus brevicollis (Evers et al., 2015); (3) anterior articular facet with a ectocondyle present in Morphotype A, whereas that of Morphotype height to width ratio of 1:1.3 (>1:1.5 considered diagnostic for B is more flattened and more deeply depressed. Additionally, the Sigilmassasaurus; Evers et al., 2015); (4) parapophysis larger than intercondylar sulcus is more pronounced in Morphotype B, the the lateral foramen, as in Spinosaurus aegyptiacus (Stromer, 1915). pterygoid flange is larger and distinctly sail-shaped in Morphotype FSAC-KK-7281: we refer this premaxilla to Spinosauridae indet. A and the quadrate foramen is smaller and more sharply defined in based on the following: (1) ventrally curved upper jaw with paired Morphotype B. Based on their interpretations of spinosaur quadrate dental alveoli (Dal Sasso et al., 2005); (2) tooth with longitudinal morphology, we assign FSAC-KK-18120 to Morphotype A based on flutes (Hendrickx et al., 2016) and fine microstructures on the the greatly enlarged quadrate facet, strongly curved mandibular 136 R.J. Lakin, N.R. Longrich / Cretaceous Research 93 (2019) 129e142

Fig. 5. Right quadrate (FSAC-KK-18120) from juvenile Spinosauridae indet. in lateral right (A), posterior (B), lateral left (C), anterior (D), ventral (E) and dorsal (F) view. ptf, pterygoid flange; qr, quadrate ridge; qf, quadrate foramen; vqjc, ventral quadratojugal contact; enc, entocondyle; qh, quadrate head; inc, intercondylar sulcus; ecc, ectocondyle. Abbreviations after Hendrickx et al. (2016). Scale bars represent 3 cm. articular facet, shallow intercondylar sulcus and broad pterygoid Highly elongated mid-cervical vertebra with total length 2 greater flange. These features are missing in the specimens assigned to than width (Evers et al., 2015); (2) nearly round centrum, not more Morphotype B by Hendrickx et al., (2016). We do not ascribe Mor- than twice as wide as high (Stromer, 1915; Evers et al., 2015); (3) in photypes A and B to either S. aegyptiacus or S. brevicollis due to the spinosaurs, the parapophysis becomes more dorsally-positioned lack of cranial material in either holotype and the composite nature further back in the column (Evers et al., 2015), until the anterior- of the neotype. dorsal vertebrae, where the parapophysis moves onto the neural FSAC-KK18119: We refer this cervical vertebra to Spinosauridae arch and is absent from the centrum. This provides a reliable po- indet. based on the following characteristics: (1) highly elongated sitional diagnosis for disarticulated vertebrae; (4) smooth keel and mid-cervical vertebra with total length 2.8 greater than width ventral rugose triangular plateau, indicative of Sigilmassasaurus (Evers et al., 2015); (2) nearly round centrum, not more than twice (Evers et al., 2015); (6) Height:width ratio of <1:1.5, indicating a as wide as high (Stromer, 1915; Evers et al., 2015); (3) laterally diagnosis of Spinosaurus aegyptiacus (Stromer, 1915). narrow neural spine that extends from anteriorly broad centroprezygopophyseal laminae that extend laterally to a rounded 5.2. Spinosaurus or Sigilmassasaurus? corner (Fig. 2); (4) in spinosaurs, the parapophysis becomes more dorsally-positioned further back in the column (Evers et al., 2015), Rostra did not form part of the holotypes of either Spinosaurus until the anterior-dorsal vertebrae, where the parapophysis moves aegyptiacus or Sigilmassasaurus brevicollis. Because of this, the onto the neural arch and is absent from the centrum. This provides a assignation of MSNM 4047 to Spinosaurus aegyptiacus by Dal reliable positional diagnosis for disarticulated vertebrae; (5) smooth Sasso et al., (2005) is difficult to corroborate. Milner (2003) keel and ventral rugose triangular plateau, indicative of Sigilmas- described a second rostrum from the Moroccan Kem Kem, spec- sasaurus (Evers et al., 2015); (6) height:width ratio of <1:1.5, indi- imen NHMUK 16665. Here, we identify a number of diagnostic cating a diagnosis of Spinosaurus aegyptiacus (Stromer, 1915). morphological features that mark this specimens apart from the FSAC-KK-18121: We refer this anterior thoracic vertebra to Milan rostrum, and assign them to Morphotype A and Morpho- Spinosauridae indet. based on the following characteristics: (1) type B, respectively. R.J. Lakin, N.R. Longrich / Cretaceous Research 93 (2019) 129e142 137

Fig. 6. Mid-cervical from Spinosauridae indet. (FSAC-KK-18119) in lateral right (A), dorsal (B), anterior (C), lateral left (D), ventral (E) and posterior (F) views. Anterior dorsal from Spinosauridae indet. (FSAC-KK-18121) in lateral right (G), dorsal (H), anterior (I), lateral left (J), ventral (K) and posterior (L) views. Scale bar represents 3 cm. 138 R.J. Lakin, N.R. Longrich / Cretaceous Research 93 (2019) 129e142

Fig. 7. Comparisons of the rostral morphologies of MSNM V447 (in ventral (A) and lateral left view (D)), NHMUK 16665 (in ventral (B) and lateral left view (E)) and FSAC-KK-7281 (in ventral (C) and lateral left view (F)). Red guides indicate differences in rostral morphology between Morph A (A, B, D, E) and Morph B (C, F). Scale bars represent 20 cm. R.J. Lakin, N.R. Longrich / Cretaceous Research 93 (2019) 129e142 139

Evidence for the distinct status of these morphotypes is found Another possibility is these morphotypes represent two distinct primarily in the lateral part of the rostrum, where, in lateral view, a taxa of Sigilmassasaurus, although evidence for a second Sigilmas- deeply concave dorsal profile strongly marks out NHMUK 16665 sasaurus species has never been reported from the Kem Kem. A from MSNM V4047 (Fig. 7D & 7F), and a curved premaxillary third option is that each morphotype could represent cranial ma- ventral profile in NHMUK 16665 which, in MSNM V4047, is flatter. terial from the only two spinosaurs known from the Kem Kem, cf. Additionally, the maxillary toothrow is relatively straight in Spinosaurus aegyptiacus and Sigilmassasaurus brevicollis (Russell, NHMUK 16665, and in MSNM V4047, this is more straight. MSNM 1996), as suggested by Hendrickx et al. (2016). Despite this, we V4047 carries uniquely shaped nares which are anteriorly tapered, contend there is not sufficient evidence to assign either morpho- while in NHMUK16665 (Fig. 7D & 7F) the nares taper posteriorly type to either cf. S. aegyptiacus or S. brevicollis. We do, however, and are proportionally larger than in MSNM V4047. In ventral view, suggest that of the two rostral morphotypes, our specimen (FSAC- the anterior outline of the premaxilla is more U-shaped in MSNM KK-7281) more closely resembles Morphotype A (MSNM V4047) V4047, whereas in NHMUK 16665, this is more V-shaped, even (Fig. 7D & 7E). accounting for taphonomic damage. Despite these observations, they could be considered equivocal 5.3. Juvenile status of spinosaur material due to the possibility that they arose as a result of taphonomy and crushing, as opposed to actual morphological differences. How- The vertebrae described here are identified as juvenile by ever, in the lateral profile of the premaxilla of MSNM V4047, it can several features, not least their small size (Table 1). As well as this, be seen that the terminal rosette is more sharply pointed ventrally there is a pattern of microscopic striations across the ventral side of (Fig. 7D). Meanwhile, in NHMUK 16665 (Fig. 7F), the rosette flat- the vertebra (Fig. 3E & 3K) that extend into the cortex. This texture tens into a ventral plateau that holds the first five teeth, before represents a longitudinally arranged vessel structure, and is a recurving dorsally further back in the rostrum towards the distal characteristic of growing bone in other dinosaurs (Brown et al., diastema. This feature is clear in both specimens and is unlikely to 2009; Xu et al., 2010) including birds (Tumarkin-Deratzian et al., be a result of post-mortem crushing. Based on this, as well as the 2006; Watanabe & Matsuoka, 2013). Larger striations are present many other (equivocal) differences, we argue that NHMUK 16665 too, and likely represent muscle scars of the hypaxial musculature, represent a distinct morphology so different from either MSNM as described by Evers et al. (2015). On the posterior cotyle, dense V4047 or FSAC-KK-7281 that they cannot be classified within the patterns of lamellar bone can be observed on the peripheral margin same taxon. of the cotyle alongside haphazard trabecular supports at the sur- However, this disparity between Morphotypes A and B does not face, suggesting high vascularity just beneath the periosteum necessarily support the hypothesis that these belong to distinct (Fig. 3F & 3L). This is a feature linked to adolescence in living ar- species. One possibility is that they represent sexual dimorphism chosaurs (Myers & Fiorillo, 2009) and has been inferred as such for within a single species. Although poorly documented, sexual di- dinosaurs too (Horner et al., 2001; Rauhut et al. 2012). In several of morphisms in the skull of theropod dinosaurs has been discussed the vertebrae the neural arch is missing, and the neurocentral su- by Larson (2008) as an explanation for the dental dimorphism seen tures along the dorsal side of the vertebra are open (Figs. 3B, 3H & in Tyrannosaurus rex. The concept of sexual differentiation in the 6B). In other taxa, this feature is used to infer infancy (Brochu, 1996) crania of theropods was explored by Molnar (2005), which argues but in spinosaurs, it is present even in adult-sized vertebrae (Evers that the major form of sexual differentiation in theropod crania is in et al., 2015). the presence or absence of skull ornamentation, including horns FSAC-KK-7281 correlates well with the morphology of the and crests. However, none of the ‘bizarre structures’ usually asso- ‘Spinosaurus aegyptiacus’ rostrum described by Dal Sasso et al., ciated with mating displays or ornamentation are present on either (2005), yet it too represents just a small percentage of the spinosaurine rostral morphotype, and the differences in maximum size when scaled isometrically. The corresponding morphology seem to be purely mechanical. This is not to say that portion of the premaxilla of NHMUK 16665 (Milner, 2003) mea- they were purely functional, as variances in the shape of the skull sures approximately 20 cm, and similar patterns of striations are can be attributed to species recognition (Padian & Horner, 2011). present in FSAC-KK-7281, as in FSAC-KK-7280 and FSAC-KK-18118. However, it does suggest that a sex-based explanation for the dif- These run longitudinally across the entire outer surface of the ferences in morphology between these specimens seems unlikely, premaxilla and along the palate. especially considering that there are few examples of modern ar- FSAC-KK-18120 is morphologically very similar to the ‘State 1’ chosaurs which show this kind of sexual differentiation in the skull Morphotype 1 spinosaurine quadrates described by Hendrickx or rostrum. et al., (2016). Based on the diagnostic characters described It is also possible that these changes are assignable to ontoge- therein, we feel confident in ascribing FSAC-KK-18120 to this netic development or intraspecific variation. There is little evidence ontogenetic stage and morphotype. in the literature for such huge differences between adults of the same species in extant archosaurs (Hospitaleche & Tambussi, 2006; 5.4. Estimations of body size Foth et al., 2015). Also, growth in extant crocodilians has been shown to be fairly consistent, with little change in shape occurring The disarticulated and fragmentary nature of this material throughout most of the life history (Monteiro et al., 1997). In other makes it difficult to arrive at robust estimates of body length, and non-avian theropods, evidence is less readily available. Carr (1999) even more difficult to estimate mass given the lack of reliable in- found few differences in the shape of the facial region of tyranno- dicators such as skull length (Therrien & Henderson, 2007), or limb saurs, with much of the development occurring in the positioning bone dimensions (Campione & Evans, 2011; Benson et al., 2014). of facial features (nares, fenestrae, orbits, etc.) and in size. Addi- However, the proportions and shapes of the vertebrae are very tionally, a comparison of allosaurus found little in the way of shape similar to adult material described in previous work (Stromer, 1915; change (Rauhut & Fechner, 2005). It is likely, therefore, that the Dal Sasso et al., 2005; Ibrahim et al., 2014b), suggesting a broadly extent of ontogenetic changes of this sort is limited throughout isometric growth pattern for Spinosaurus that could allow us to development, and that we would be unlikely to observe the estimate total body lengths from the specimens described here. changes seen between NHMUK 16665 and MSNM V4047 if the This approach has obvious drawbacks, but serves to provide differences were due to ontogeny or intraspecific differences alone. approximate estimations of the size of juvenile spinosaurs. Due to 140 R.J. Lakin, N.R. Longrich / Cretaceous Research 93 (2019) 129e142

Table 2 spinosaurs. The concept of dietary, habitat and behavioural segre- Isometrically scaled estimations of total body length for spinosaur material. gation between adult and juvenile theropods living in the same Specimen no. Total body length Total body length habitat has previously been proposed (Carbone et al., 2011), but this (bipedal model) (m) (quadrupedal model) (m) effect is likely magnified in the aquatic habitat that spinosaurs is FSAC-KK-7280 4.5 4.5 known to have occupied due to the greater predation pressures FSAC-KK-7281 7.6 5.2 aquatic habitats are able to support (Shurin et al., 2006). FSAC-KK-18118 3.7 4.0 An idiosyncrasy of this collection that is only made clear by FSAC-KK-18119 4.5 3.9 visual representation of the individuals (Fig. 8) is the fact that the FSAC-KK-18120 3.7 4.8 FSAC-KK-18122 3.4 3.4 juveniles presented here are all of a broadly similar size, and FSAC-KK-18121 3.0 3.3 possibly represent individuals of a similar age group. This can also be seen in other collections that have mentioned juvenile material incidentally (Hendrickx et al., 2016). So far, there appears to be a the ongoing controversy over spinosaur locomotion, isometric distinct lack of hatchling-sized specimens. This phenomenon could scaling estimated for both the bipedal model (Dal Sasso et al., 2005) be assigned to a multitude of explanations, none of which can be and quadrupedal model (Ibrahim et al., 2014b) are given. These tested without further sampling. Possible explanations include (1) body lengths are given in Table 2. taphonomic biases against the preservation of hatchling material, (2) collection bias against the recovery of juvenile material, or (3) 5.5. Ecological implications habitat segregation, with hatchlings occupying a distinct habitat. Taphonomic bias or collector bias is a plausible explanation. This material represents some of the smallest spinosaurs yet Hatchling bones would likely be delicate and poorly ossified, and described (Fig. 8B) and suggests co-occurrence of adult spinosaurs might be unlikely to survive transport and burial. Hatchlings might and their offspring. This implies that juvenile spinosaurs occupied also be more vulnerable to predation, and this may have been a the same habitat as their parents and may have occupied a similar more common cause of death than natural death and burial ecological niche. If the adults were semiaquatic, then it appears (Farlow, 1976; Hone & Rauhut, 2010). Collection bias against small likely that the juveniles were as well. Similar habits are seen in bones is also common. However, collecting of Kem Kem fossils extant crocodiles (Pooley, 1977) and have been inferred for the produces abundant small and even microscopic teeth of theropods, small number of dinosaurs for which juveniles and adults have pterosaurs, and fish, which are even smaller than the smallest been found in association (Varricchio et al., 1997; Norell et al., spinosaur teeth (Kellner & Mader, 1996; Cavin & Brito, 2001). The 1995). However, despite this association, it is unlikely that adult absence of very small (>20 mm) spinosaurid shed teeth is therefore and juvenile spinosaurs occupied the same ecological niche given difficult to explain, and raises the possibility that the smallest the large disparity in size between the known largest and smallest spinosaurs may have lived in different habitats.

Fig. 8. Artistic representation of the relative sizes of our juvenile spinosaur material. FSAC-KK-7280 (A), FSAC-KK-18122 & FSAC-KK-18121 (B), FSAC-KK-7281 (C), FSAC-KK-18118 & FSAC-KK-18119 (D), FSAC-KK-18120 (E) and MSNM V4047 (F), with human for scale. Specimens measuring within 10 cm of each other are illustrated together. Scale bar represents 2 m. R.J. Lakin, N.R. Longrich / Cretaceous Research 93 (2019) 129e142 141

6. Conclusion Years of Sustained Ecological Innovation on the Avian Stem Lineage. PLOS Biology 12 (5), e1001853. Benyoucef, M., Lang,€ E., Cavin, L., Mebarki, K., Adaci, M., Bensalah, B., 2015. Over- Non-hatchling juvenile spinosaurids are relatively common abundance of piscivorous dinosaurs (Theropoda: Spinosauridae) in the mid- members of the Kem Kem assemblage. This suggests that juveniles Cretaceous of North Africa: The Algerian dilemma. Cretaceous Research 55, e of these species shared their habitat with the adults, and that the 44 55. Brochu, C., 1996. Closure of neurocentral sutures during crocodilian ontogeny: semiaquatic habits hypothesized for adult spinosaurs also charac- Implications for maturity assessment in fossil archosaurs. Journal of Vertebrate terized the juveniles. It is conceivable if not likely that juveniles of a Palaeontology 16 (1), 49e62. certain age may have associated with adults, given that this Brown, C.M., Russell, A.P., Ryan, M.J., 2009. Pattern and transition of surficial bone texture of the centrosaurine frill and their ontogenetic and taxonomic impli- behaviour is seen in extant relatives such as crocodilians and birds. cations. Journal of Vertebrate Palaeontology 29, 132e141. Very small (~1e2 m) juveniles are not yet known from the Kem Buffetaut, E., 1989. New remains of the enigmatic dinosaur Spinosaurus from Kem beds despite the fact that other small vertebrates, with similar the Cretaceous of Morocco and the affinities between Spinosaurus and Baryonyx. Neues Jahrbuch für Geologie und Palaontologie,€ Monatshefte preservation potential, are known from these same deposits. This 1989, 79e87. suggests that the earliest ontogenetic stages had a distinct habitat Buffetaut, E., 2007. The spinosaurid dinosaur Baryonyx (Saurischia, Theropoda) in and ecology. Testing these hypotheses requires improved sampling the Early Cretaceous of Portugal. Geological Magazine 144 (6), 1021e1025. of adult and juvenile spinosaurs in Morocco. Additionally, we have Buffetaut, E., 2012. An early spinosaurid dinosaur from the Late Jurassic of Tenda- guru (Tanzania) and the evolution of the spinosaurid dentition. Oryctos 10, 1e8. identified two distinct morphotypes of spinosaurine rostra present Buffetaut, E., Ouaja, M., 2002. A new specimen of Spinosaurus from the Lower in the Moroccan Kem Kem which supports the identification of at Cretaceous of Tunisia, with remarks on the evolutionary history of Spinosaurus. e least two distinct species in the fauna. At least some of the juvenile Bulletin de la SocieteGeologique de France 173 (5), 415 421. Campione, N.E., Evans, C., 2011. Cranial growth and variation n edmontosaurs material exhibits the same diagnostic characters that separate the (Dinosauria: Hadrosauridae): implications for latest Cretaceous megaherbivore adults. Although ontogenetic changes must be taken into account diversity in North America. PLoS ONE 6 (9), e25186. in diagnosing species, this suggests that some skeletal elements of Carbone, C., Turvey, S.T., Bielby, J., 2011. Intra-guild competition and its implications for one of the biggest terrestrial predators, Tyrannosaurus rex. Proceedings of the juveniles can be diagnostic. the Royal Society B 278 (1718), 2682e2690. Carr, T.D., 1999. Craniofacial ontogeny in Tyrannosauroidea (Theropods, Coelur- e Acknowledgements osauria). Journal of Vertebrate Palaeontology 19 (3), 497 520. Cavin, L., Brito, P.M., 2001. A New Lepisosteidae (Actinopterygii: Ginglymodi) from the Cretaceous of the Kem Kem beds, southern Morocco. We thank the fossil collectors and dealers of Begaa, whose hard Cavin, L., Buffetaut, E., Tong, H., Lang,€ E., Tabouelle, J., Vullo, R., 2015. Taxonomic work and enterprising collection methods have resulted in Composition and Trophic Structure of the Continental Bony Fish Assemblage fi from the Early Late Cretaceous of Southeastern Morocco. PLoS ONE 10 (5), acquiring material previously unknown to the scienti c commu- e0125786. nity. We thank Samir Zouhri and the University Hassan II in Casa- Cavin, L., Tong, H., Boudad, L., Meister, C., Piuz, A., Tabouelle, J., Aarab, M., Amiot, R., blanca for their cooperation in cataloguing and housing our Buffetaut, E., Dyke, G., Hua, S., Le Loeuff, J., 2010. Vertebrate assemblages from the early Late Cretaceous of southeastern Morocco: An overview. Journal of material. We thank Emma Randle and Conrad van den Ende for African Earth Sciences 57, 391e412. their support and advice regarding methods, as well as editing. We Candeiro, C.R.A., Brusatte, S.L., de Souza, A.L., 2017. Spinosaurid Dinosaurs from the thank Aubrey Roberts for providing training in the preparation of Early Cretaceous of North Africa and Europe: Fossil Record, Biogeography and Extinction. Anuario do Instituto de Geociencias^ 40 (3), 294e302. specimens, and for manuscript revisions. We thank Catherine Klein Dal Sasso, C., Maganuco, S., Buffetaut, E., Mendez, M.A., 2005. New information on for her encouragement and for providing training in work man- the skull of the enigmatic theropod Spinosaurus, with remarks on its size and agement and manuscript editing. We thank our reviewers, Chris- affinities. Journal of Vertebrate Paleontology 25 (4), 888e896. Erickson, G.M., 1996. Incremental lines of von Ebner in dinosaurs and the assess- tophe Hendrickx and Serjoscha Evers, for their excellent and ment of tooth replacement rates using growth line counts. Proceedings of the thorough critique of our work. The authors declare that every National Academy of Sciences 93 (25), 14623e14627. precaution has been taken to protect the rights of the Moroccan Evers, S.W., Rauhut, O.W.M., Milner, A.C., McFeeters, B., Allain, R., 2015. A reappraisal people to the material described in this study. The specimens have of the morphology and systematic position of the theropod dinosaur Sigilmas- sasaurus from the "middle" Cretaceous of Morocco. PeerJ 3, e1323. been accessioned at a Moroccan institute (University Hassan II, Farlow, J.O., 1976. A consideration of the trophic dynamics of a Late Cretaceous Casablanca) and upon completion of our research, they will be large-dinosaur community (Oldman Formation). Ecological Society of America e returned to this institute. This work was supported by the Natural 57 (5), 841 857. Foth, C., Bona, P., Desojo, B., 2015. Intraspecific variation in skull morphology of the Environment Research Council [NE/L002434/1] and the Lev- black caiman Melanosuchus niger (Alligatoridae, Caimaninae). Acta Zoologica erhulme Foundation (RL-2016-036). 96, 1e13. Gasca, J.M., Díaz-Martínez, I., Moreno-Azanza, M., Canudo, J.I., Alonso, A., 2018. A hypertrophied ungual phalanx from the lower Barremian of Spain: Implica- References tions for the diversity and palaeoecology of Spinosauridae (Theropoda) in Iberia. Cretaceous Research 84, 41e152. Allain, R., Xaisanavong, T., Richir, P., Khentavong, B., 2012. The first definitive Asian Gimsa, J., Sleigh, R., Gimsa, ., 2016. The riddle of Spinosaurus aegyptiacus' sail. spinosaurid (Dinosauria: Theropoda) from the Early Cretaceous of Laos. Geological Magazine 153 (3), 544e547. Naturwissenschaften 99 (5), 369e377. Goedert, J., Amiot, R., Boudad, L., Buffetaut, E., Fourel, F., Godefroit, P., Kusuhashi, N., Amiot, R., Wang, X., Lecuyer, C., Buffetaut, E., Boudad, L., Cavin, L., Ding, Z., Suteethorn, V., Tong, H., Watabe, M., Lecuyer, C., 2016. Preliminary Investigation Fluteau, F., Kellner, A.W.A., Tong, H., Zhang, F., 2010. Oxygen and carbon isotope of Seasonal Patterns Recorded in the Oxygen Isotope Compositions of Theropod compositions of middle Cretaceous vertebrates from North Africa and Brazil: Dinosaur Tooth Enamel. PALAIOS 31, 10e19. Ecological and environmental significance. Palaeogeography, Palae- Hasegawa, Y., Tanaka, G., Takakuwa, Y., Koike, S., 2010. Fine sculptures on a tooth of oclimatology, Palaeoecology 297 (2), 439e451. Spinosaurus (Dinosauria, Theropoda) from Morocco. Bulletin of Gunma Museum Arden, T.M.S., Klein, C.G., Zouhri, S., Longrich, N.R., 2018. Aquatic adaptation in the of Natural History 14, 11e20. skull of carnivorous dinosaurs (Theropida: Spinosauridae) and the evolution of Henderson, D.M., 2018. A buoyancy, balance and stability challenge to the hy- aquatic habits in spinosaurs. Cretaceous Research (in press). pothesis of a semi-aquatic Spinosaurus Stromer, 1915 (Dinosauria: Theropoda). Aureliano, T., Ghilardi, A.M., Buck, P.V., Fabbri, M., Samathi, A., Delcourt, R., PeerJ 6, e5409. Fernandes, M.A., Sander, M., 2018. Semi-aquatic adaptations of a spinosaur from Hendrickx, C., Mateus, O., Buffetaut, E., 2016. Morphofunctional Analysis of the the Lower Cretaceous of Brazil. Cretaceous Research 90, 283e295. Quadrate of Spinosauridae (Dinosauria: Theropoda) and the Presence of Spi- Barker, C.T., Naish, D., Newham, E., Katsamenis, O.L., Dyke, G., 2017. Complex nosaurus and a Second Spinosaurine Taxon in the Cenomanian of North Africa. neuroanatomy in the rostrum of the Isle of Wight theropod Neovenator salerii. PLoS ONE 11 (1), e0144695. Scientific Reports 7 (3749). Holtz, T.R., 1998. Spinosaurs as crocodile mimics. Science 282, 1276e1277. Barrett, P.M., Benson, R.B., Rich, T.H., Vickers-Rich, P., 2011. First spinosaurid dino- Hone, D.W.E., Rauhut, O.W.M., 2010. Feeding behaviour and bone utilization by saur from Australia and the cosmopolitanism of Cretaceous dinosaur faunas. theropod dinosaurs. Lethaia 43 (2), 232e244. Biology Letters 7 (6), 933e936. Horner, J.R., Padian, K., de Ricqles, A., 2001. Comparative Osteohistology of Some Benson, R., Campione, N.E., Carrano, M.T., Mannion, P.D., Sullivan, C., Upchurch, P., Embryonic and Perinatal Archosaurs: Developmental and Behavioral Implica- Evans, D.C., 2014. Rates of Dinosaur Body Mass Evolution Indicate 170 Million tions for Dinosaurs. Palaeobiology 27 (1), 39e58. 142 R.J. Lakin, N.R. Longrich / Cretaceous Research 93 (2019) 129e142

Hospitaleche, C.A., Tambussi, C., 2006. Skull morphology of Pygoscelis (Sphenisci- Rauhut, O.W.M., Foth, C., Tischlinger, H., Norell, M.A., 2012. Exceptionally preserved formes): inter and intraspecific variation. Polar Biology 29, 728e734. juvenile megalosauroid theropod dinosaur with filamentous integument from Ibrahim, N., Varricchio, D.J., Sereno, P.C., Wilson, J.A., Dutheil, D.B., Martill, D.M., the Late Jurassic of Germany. Proceedings of the National Academy of Sciences Baidder, L., Zouhri, S., 2014a. Dinosaur Footprints and Other Ichnofauna from 109 (29), 11746e11751. the Cretaceous Kem Kem Beds of Morocco. PLoS ONE 9 (3), e90751. Richter, U., Mudroch, A., Buckley, L.G., 2013. Isolated theropod teeth from the Kem Ibrahim, N., Sereno, P.C., Dal Sasso, C., Maganuco, S., Fabbri, M., Martill, D.M., Kem Beds (Early Cenomanian) near Taouz, Morocco. Palaontologische€ Zeits- Zouhri, S., Myhrvold, N., Iurino, D.A., 2014b. Semiaquatic adaptations in a giant chrift 87 (2), 291e309. predatory dinosaur. Science 345 (6204), 1613e1616. Rodriguez, T., Kellner, A.W., Mader, B.J., Russell, D.A., 2011. New pterosaur specimens Keller, A.W.A., Azevado, S.A.K., Machado, E.B., de Carvalho, L.B., Henriques, D.D.R., from the Kem Kem beds (Upper Cretaceous, Cenomanian) of Morocco. Rivista 2011. A new dinosaur (Theropoda, Spinosauridae) from the Cretaceous (Cen- Italiana di Paleontologia e Stratigrafia 117 (1), 49e160. omanian) Alcantara^ Formation, Cajual Island, Brazil. Anais da Academia Brasi- Ruiz-Omenaca,~ J.I., Canudo, J.I., Cruzado-Caballero, P., Infante, P., Moreno- leira de Ciencias^ 83 (1), 99e108. Azanza, M., 2005. Baryonychine teeth (Theropoda: Spinosauridae) from the Kellner, A.W., Mader, B.J., 1996. Archosaur teeth from the Cretaceous of Morocco. Lower Cretaceous of La Cantalera (Josa, NE Spain). In: Current Research in Journal of Palaeontology 71 (3), 525e527. Vertebrate Palaeontolgy 3rd Annual Meeting of the European Association of Krassilov, V., Bacchia, F., 2013. New Cenomanian florule and a leaf mine from Vertebrate Palaeontologists (EAVP) in the Hessisches Landesmuseum Darm- southeastern Morocco: Palaeoecological and climatological inferences. Creta- stadt 18th to 23th July 2005. ceous Research 40, 218e226. Russell, D.A., 1996. Isolated dinosaur bones from the middle Cretaceous of the Lang,€ E., Boudad, L., Maio, L., Samankassou, E., Tabouelle, J., Tong, H., Cavin, L., 2013. Tafilalt, Morocco. Bulletin du Museum National d'Histoire Naturelle, Paris, 4e Unbalanced food web in a Late Cretaceous dinosaur assemblage. Palae- serie, section C 18 (2e3), 349e402. ogeography, Palaeoclimatology, Palaeoecology 381e382, 26e32. Russell, D.A., Paesler, M.A., 2003. Environments of Mid-Cretaceous Saharan di- Larson, P., 2008. Variation and Sexual Dimorphism in Tyrannosaurus rex. In: nosaurs. Cretaceous Research 24, 569e588. Larson, P., Carpenter, K. (Eds.), Tyrannosaurus Rex, the Tyrant King. Indiana Sales, M.A.F., Liparini, A., de Andrade, M.B., Aragao,~ P.R.L., Schutlz, C.L., 2017. The University Press, Bloomington, pp. 103e106. oldest South American occurrence of Spinosauridae (Dinosauria, Theropoda). Leich, D.B., Catania, K.C., 2012. Structure, innervation and response properties of Journal of South American Earth Sciences 74, 83e88. integumentary sensory organs in crocodilians. Journal of Experimental Biology Scott, P.W., Currie, P.J., 2017. The functional origin of dinosaur bipedalism: Cumu- 215, 4217e4230. lative evidence from bipedally inclined reptiles and disinclined mammals. Lockley, M.G., 1994. Dinosaur ontogeny and population structure: Interpretations Journal of Theoretical Biology 420, 1e7. and speculations based on fossil footprints. In: Carpenter, K., Hirsche, K.F., Sereno, P.C., Dutheil, D.B., Iarochene, M., Larsson, H.C.E., Lyon, G.H., Magwene, P.M., Horner, Dinosaur Eggs, J.R., Babies (Eds.). Cambridge University Press, New York, Sidor, C.E., Varricchio, D.J., Wilson, J.A., 1996. Predatory Dinosaurs from the NY, p. 349. Sahara and Late Cretaceous Faunal Differentiation. Science 272 (5264), Maganuco, S., Dal Sasso, C., 2018. The smallest biggest theropod dinosaur: a tiny 986e991. pedal ungual of a juvenile Spinosaurus from the Cretaceous of Morocco. PeerJ 6, Shurin, J.B., Gruner, D.S., Hillebrand, H., 2006. All wet or dried up? Real differences e4785. between aquatic and terrestrial food webs. Proceedings of the Royal Society B Malafaia, E., Ortega, F., Escaso, F., Mocho, P., 2013. Rediscovery of a lost portion of 273 (1582), 1e9. the holotype of Suchosaurus girardi (Sauvage, 1897-98), now related to the Stromer, E., 1915. Ergebnisse der Forschungsreisen Prof. E. Stromers in den Wüsten € spinosaurid theropod Baryonyx. In: VI Jornadas Internacionales sobre Paleon- Agyptens. II. Wirbeltier-Reste der Baharije-Stufe (unterstes Cenoman). 3. Das tología de Dinosaurios y su Entorno. Salas de los Infantes, Burgos. Original des Theropoden Spinosaurus aegyptiacus nov. gen., nov. spec. Abhand- Martill, D.M., Ibrahim, N., Brito, P.M., Baider, L., Zhouri, S., Loveridge, R., Naish, D., lungen der Koniglich€ Bayerischen Akademie der Wissenschaften. Mathema- Hing, R., 2011. A new Plattenkalk Konservat Lagerstatte€ in the Upper Cretaceous tisch-physikalische Klasse 28 (3), 1e32. of Gara Sbaa, south-eastern Morocco. Cretaceous Research 32, 433e446. Sues, H., Frey, E., Martill, D.M., Scott, D.M., 2002. Irritator challenger, a spinosaurid McFeeters, B., Ryan, M.J., Hinic-Frlog, S., Schroder-Adams,€ C., 2013. A reevaluation of (Dinosauria: heropods) from e Lower Cretaceous of Brazil. Journal of Vertebrate Sigilmassasaurus brevicollis (Dinosauria) from the Cretaceous of Morocco. Ca- Palaeontology 22 (3), 535e547. nadian Journal of Earth Sciences 50 (6), 636e649. Taquet, P., Russell, D.A., 1998. New data on spinosaurid dinosaurs from the Early McGowan, A.J., Dyke, G.J., 2009. A surfeit of theropods in the Moroccan Late Cretaceous of the Sahara. Comptes Rendus de l'Academie des Sciences - Series Cretaceous? Comparing diversity estimates from field data and fossil shops. IIA - Earth and Planetary Science 327 (5), 347e353. Geology 37 (9), 843e846. Therrien, F., Henderson, D.M., 2007. My theropod is bigger than yours... or not: Milner, A.C., 2003. Fish-eating theropods: A short review of the systematics, biology estimating body size from skull length in theropods. Journal of Vertebrate and palaeobiogeography of spinosaurids. In: Huerta Hurtado, P., Torcida Palaeontology 27 (1), 108e115. Fernandez-Baldor, F. (Eds.), 2003. e Actas de las ii Jornadas Internacionales Therrien, F., Henderson, D.M., Ruff, C.B., 2005. Bite me: biomechanical models of Sobre Paleontologíıa de Dinosaurios y Su Entorno, 2001. Salas de los Infantes, theropod mandibles and implications for feeding behavior. The Carnivorous Colectivo Arqueologico ePaleontologico. de Salas, pp. 129e138. Dinosaurs 179e237. Monfroy, Q.T., 2017. Correlation between the size, shape and position of the teeth Tumarkin-Deratzian, A.R., Vann, D.R., Dodson, P., 2006. Bone surface texture as an on the jaws and the bite force in Theropoda. Historical Biology 29 (8), ontogenetic indicator in long bones of the Canada goose Branta canadensis 1089e1105. (Anseriformes: Anatidae). Zoological Journal of the Linnean Society 148 (2), Monteiro, L.R., Cavalcanti, M.J., Sommer, H.S.J., 1997. Comparative ontogenetic shape 133e168. changes in the skull of Caiman species (Crocodylia, Alligatoridae). Journal of Varricchio, D.J., Jackson, F., Borkowski, J.J., Horner, J.R., 1997. Nest and egg clutches of Morphology 231, 53e62. the dinosaur Troodon formosus and the evolution of avian reproductive traits. Myers, T.S., Fiorillo, A.R., 2009. Evidence for gregarious behavior and age segrega- Nature 385 (6613), 247. tion in sauropod dinosaurs. Palaeogeography, Palaeoclimatology, Palaeoecology Vullo, R., Allain, R., Cavin, L., 2016. Convergent evolution of jaws between spino- 274, 96e104. saurid dinosaurs and pike conger eels. Acta Palaeontologia Polonica 61 (4), Norell, M.A., Clark, J.M., Chiappe, L.M., Dashzeveg, D., 1995. A nesting dinosaur. 825e828. Nature 378, 774e776. Watanabe, J., Matsuoka, H., 2013. Ontogenetic change of morphology and surface Padian, K., Horner, J.R., 2011. The evolution of ‘bizarre structures’ in dinosaurs: texture of long bones in the Gray Heron (Ardea cinerea, Ardeidae). Palae- biomechanics, sexual selection, social selection or species recognition? Journal ornithological Research 279e306. of Zoology 283 (1), 3e17. Wilson, J.A., 1999. A nomenclature for vertebral laminae in sauropods and other Pooley, A.C., 1977. Nest opening response of the Nile crocodile Crocodylus niloticus. saurischian dinosaurs. Journal of Vertebrate Paleontology 19 (4), 639e653. Journal of Zoology 182 (1), 17e26. Xu, X., Zheng, X., You, H., 2010. Exceptional dinosaur fossils show ontogenetic Rauhut, O.W.M., Fechner, R., 2005. Early development of the facial region in a non- development of early feathers. Nature 464 (7293), 1338e1341. avian theropod dinosaur. Proceedings of the Royal Society B: Biological Sciences 272 (1568), 1179e1183.