identifier	taxonID	type	CVterm	format	language	title	description	additionalInformationURL	UsageTerms	rights	Owner	contributor	creator	bibliographicCitation
03A39E0B9644FFFAFEBAFB99FF2E1C15.text	03A39E0B9644FFFAFEBAFB99FF2E1C15.taxon	http://purl.org/dc/dcmitype/Text	http://rs.tdwg.org/ontology/voc/SPMInfoItems#GeneralDescription	text/html	en	Spinosaurinae	<div><p>Spinosaurinae gen. et sp. indet.</p><p>(Figs 2–12)</p><p>Referred specimens</p><p>An almost complete rostrum preserving both premaxillae, maxillae, and the anterior portion of the nasals (NHMUK PV R 16420); a paired premaxillae with a fragment of attached maxilla (NHMUK PV R 16422); and a fragment of an isolated right premaxilla (NHMUK PV R 16424); conjoined nasals (NHMUK PV R 16426); a partial skull roof comprising two fused frontals, left prefrontal, and the posterior portion of the left nasal (NHMUK PV R 16423); a mid-dorsal neural arch (NHMUK PV R 16430); a fragmentary mid-dorsal neural spine (NHMUK PV R 16431); two partial ilia (NHMUK PV R 16391 and NHMUK PV R 16438); and a nearly complete left femur (NHMUK PV R 16433).</p><p>Morphological description (rostrum)</p><p>NHMUK PV R 16420 consists of the paired premaxillae, mainly the anterior body of the right and left maxillae, and the anterior portions of the conjoined nasals (Fig. 2). The anterior end of the snout is damaged and seems to be slightly taphonomically deformed (e.g. Fig. 2C). The rostrum is anteroposteriorly long (Milner 2003), being c. 600 mm in length, lateromedially compressed and dorsoventrally rather low (Fig. 2A–C). In lateral view (Fig. 2A–C), the dorsal surface of the rostrum is straight anteriorly, posteriorly becoming concave. The ventral surface is heavily concave anteriorly and straightens posteriorly at the level of the fourth maxillary alveolus [elongate 'S-curve' of Milner (2003)]. The anterior tip of the premaxillae does not project ventral to the maxillary tooth row (Fig. 2A, B). The preserved rostral roof is formed by the premaxillae and the nasals, with the skull roof being narrow mediolaterally. Ventrally the rostrum thickens because the lateral walls of the maxillae are ventrolaterally directed—both the premaxillae and maxillae exhibit a sinusoidal groove (Isasmendi et al. 2023), which separates the lateral dentigerous section from the medial part of the rostral bones and creates a secondary palate visible along the entire lateral view (Fig. 2A, B). The external nares are significantly retracted posteriorly, at about 560 mm from the anterior tip of the snout, and they are proportionately small, being located between the eighth and ninth maxillary alveoli (Fig. 2A, B). The premaxilla does not participate in the external nares. Instead, the latter is delimited anteriorly, ventrally, and posteriorly by the maxillae, and dorsally by the nasals (Fig. 2A, B). Although the entire shape of the external naris cannot be determined, the anterior and dorsal margins are relatively straight and the posteroventral margin seems to be concave in shape (Fig. 2A).</p><p>Premaxillae</p><p>The paired premaxillae of the NHMUK PV R 16420 specimen suggest that the dorsal rim was partially fused, potentially indicating that this individual was nearly a somatically mature animal. In the other specimens, NHMUK PV R 16422 (Fig. 3) and NHMUK PV R 16424 (Fig. 4), no bony fusion can be noticed, suggesting that these were somatically immature individuals. The premaxillary body of the NHMUK PV R 16420 rostrum is c. 200 mm in length anteroposteriorly and c. 120 mm in height, with a c. 90 mm broad premaxillary 'rosette' (sensu Charig and Milner 1997). If the posterodorsal process of the premaxilla is considered, this would be c. 570 mm long anteroposteriorly (Fig. 2). The premaxillae of NHMUK PV R 16420 contact the nasals posterodorsally and the maxillae posteroventrally (Fig. 2A, B). The anteriormost portion of all the studied premaxillae is damaged and its shape cannot be determined; however, the anterior premaxillary 'rosette' is present, at least partially, in all of them [see Milner (2003) and this work for NHMUK PV R 16420]. Posterior to the 'rosette', the premaxillary body is constricted both dorsoventrally and mediolaterally (Figs 2–4). In lateral view (Fig. 2A, B, H–K), the anteroventral margin of the NHMUK PV R 16420 premaxillae projects anteroventrally— the anterior surface of the premaxillae is convex, straightening dorsally, while the ventral surface is concave throughout its entire length. The lateral wall of the premaxillae is thoroughly rugose and bears several neurovascular foramina that pierce the premaxilla (e.g. Fig. 3E–H). Generally, the foramina are oval to subcircular in shape and exhibit a ventrally oriented shallow groove, which can be straight or parabolic. The ventral foramina form a row that is parallel to the ventral surface of the lateral wall. The dorsal foramina seem to be scattered. Also in lateral view, the alveolar row can be distinguished (see Figs 2–4).</p><p>As aforementioned, the external naris is considerably retracted posteriorly (Fig. 2A, B), the typical nasal and subnarial processes of the premaxilla that delimit the external naris dorsally and ventrally, as seen in other theropods, is not present. Nevertheless, two equivalent processes (i.e. the posterodorsal and posteroventral processes of the premaxillae; Fig. 2A, B) are present in the NHMUK PV R 16420 individual, which delimit the subnarial foramen (better preserved on the left lateral side; Fig. 2A). The posterodorsal process of the premaxillae mainly projects posteriorly, being long and forming the skull roof between its base and the nasals, and gradually tapering posteriorly (Fig. 2A, B). This process is straight and anteriorly horizontal; however, it slightly projects posterodorsally, as seen in lateral view (Fig. 2B). The ventral surface of the posterodorsal process contacts the maxilla ventrally from its base to slightly anterior to the external naris; posteriorly it contacts the nasals. The contact between the premaxillae and nasals has a complex morphology. In dorsal view, the premaxillae and nasals contact each other dorsally and interlock via a 'peg-like' process that extends from each premaxillae penetrating the nasals, giving it a 'W-shape' morphology at the contact region. The posteroventral process of the premaxillae is considerably less-developed; however, it is taphonomically damaged (mainly on the right side; Fig. 2B). This tapering process is posteriorly directed and laterally overlaps the maxillae. The bases of the posterodorsal and posteroventral processes are separated by the subnarial foramen (Fig. 2A). In ventral view, each premaxilla (i.e. NHMUK PV R 16420, NHMUK PV R 16422, and NHMUK PV R 16424) exhibits a medially placed robust bone element, forming a convex secondary palate. That secondary palate is also visible in lateral view due to the lateral walls of the premaxillae, which do not project ventrally (Figs 2–4). These elements are divided from the lateral dentigerous portion of the premaxillae by a sinusoidal groove. The robust bone elements meet anteromedially at the level of the second and third premaxillary alveoli (e.g. Fig. 2C, D). This medial contact consists of the interdigitation of both elements, whereas they are posteriorly separated by a gap and, more posteriorly, by the anteromedial processes of the maxillae (Milner 2003).</p><p>The dentigerous portions of the premaxillae are concave (Figs 2H–J, 3E–H, 4B, C). Each of the NHMUK PV R 16420 premaxillae bear seven alveoli, which are subcircular in shape (Milner 2003, Lacerda et al. 2022). The anteriormost alveolus (premaxillary alveolus 1; pm1) is small, and the second (pm2) and third ones (pm3) are considerably larger, being the largest alveoli of the premaxillary row; towards the rear of the skull, the premaxillary alveoli decrease in size. The fourth (pm4) and fifth alveoli (pm5) are separated from their anterior and posterior ones by a diastema; and the sixth (pm6) and seventh alveoli (pm7) are 'coupled' in the studied premaxillae (Figs 2C, D, 3A, B, 4B, C), similar to other spinosaurines (Lacerda et al. 2022). Between the last premaxillary alveolus (pm6 or pm7) and the first maxillary tooth (mx1), another diastema is present. In NHMUK PV R 16422 and NHMUK PV R 16424, the posterior four alveoli are preserved, which gradually decrease in size posteriorly (Figs 3A, B, 4B, C). If the pattern observed in the spinosaurine specimens, such as NHMUK PV R 16420, MNHN SAM 124, and MSNM V 4047, is considered (Taquet and Russell 1998, Milner 2003, Dal Sasso et al. 2005, Lacerda et al. 2022), the NHMUK PV R 16422 individual would bear six premaxillary teeth and thus the preserved alveoli are interpreted as pertaining to pm3 to pm6 (Fig. 3A–D). On the other hand, the specimen NHMUK PV R 16424 would exhibit seven alveoli on its right premaxilla and the preserved ones would correspond to pm4 to pm7 (Fig. 4). Alveolar metrics are in Table 2.</p><p>In the left premaxilla of NHMUK PV R 16422, lpm3 is separated from the rest of the premaxillary alveoli by a diastema (Fig. 3A, B). In the right premaxilla of NHMUK PV R 16422 and the right premaxilla NHMUK PV R 16424 (Fig. 4), the two anterior preserved alveoli (pm3–4 of NHMUK PV R 16422 and pm4–5 of NHMUK PV R 16424), as well as the two most posterior alveoli (pm5–6 of NHMUK PV R 16422 and pm6–7 of NHMUK PV R 16420), are paired and separated from each other by a diastema (Figs 2–4). Another diastema is present between the last alveolus of the premaxilla (pm6/7) and the maxilla (m1). In the regions of the premaxillae where the diastemata are located, the lateral walls of the premaxillae have ventrally located concavities (Fig. 2A, B), where the laterocumbent teeth of the dentary would probably have inserted, as indicated by Dal Sasso et al. (2005). The overall outline of the preserved alveoli is subcircular in all three specimens (Figs 2–4).</p><p>Both NHMUK PV R 16420 and NHMUK PV R 16422 still retain some teeth preserved in situ. In NHMUK PV R 16420, these are present in the lpm1 on the left side, and the rpm 3 in the right premaxilla, and comprise the bases and roots of the teeth (Fig. 2C–H). In the NHMUK PV R 16422 specimen, teeth are preserved in the alveoli lpm4 and lpm6 of the left premaxilla and rpm3 and rpm5 of the right premaxilla (Fig. 3). However, these teeth are not completely erupted, being replacement teeth (e.g. rt in Fig. 3E). The general dentition is conidont, being not distally recurved, with a centrally located apex, and laterocumbent (sensu Hendrickx et al. 2019). The basal cross-section of the teeth is subcircular and lenticular in the middle of the crown. The mesial and distal carinae are not denticulated and their basal portion reaches the cervix. The mesial carinae of the teeth of NHMUK PV R 16422 are oriented mesiolingually and the distal carinae are oriented distolabially, mainly based on rpm5 and lpm6 (Fig. 3A, B). The enamel texture of the crowns is veined/ anastomosed (sensu Hendrickx et al. 2015). The largest tooth in NHMUK PV R 16420 (rpm3) exhibits flutes, at least on the labial surface of the crown. However, the teeth of NHMUK PV R 16422 are devoid of flutes, with the labial and lingual surfaces of the crown being smooth.</p><p>Maxillae</p><p>The NHMUK PV R 16422 specimen only preserves the anteromedial processes of a maxilla and the most anteroventral point of the left maxilla, including the mesial border of the left maxillary alveolus 1 (Fig. 3A–D). Therefore, our description focuses mainly on the well-preserved maxillae of NHMUK PV R 16420.</p><p>Each maxilla of the NHMUK PV R 16420 snout mainly preserves the anterior portion of the maxillary body and the base of the ascending process of the maxilla (Fig. 3), thus, there is no sign of the jugal processes. The maxilla is elongated (Milner 2003), with a preserved length of c. 420 mm. The anterior body of the maxilla is continuous in depth (c. 100 mm), becoming deeper at the level of the ascending process (Fig. 2A, B). The anterior process of the maxilla is hypertrophied and contacts the premaxilla anteriorly in ventral view (Fig. 2C). The anteroventral contact between the premaxilla and the maxilla is not preserved, but just adjacent to the subnarial foramen, the posteroventral process of the premaxilla interlocks with the maxilla (Milner 2003). Posterodorsally, the premaxillary–maxillary contact becomes more horizontal and straighter, slightly curving dorsally anterior to the premaxillary–nasal contact (Fig. 2A, B). The intermaxillary suture extends from the most anterior portion of the maxillae to the region of the eighth maxillary alveoli (m8) (Milner 2003). Anteriorly, the intermaxillary suture, between alveoli m1 and m2, presents an interdigitated pattern (Fig. 2C). In lateral view, the ventral margin of the maxilla faces anteroventrally and subsequently this margin gradually faces ventrally proximally, with a medial shelf being evident in lateral view (Fig. 2B, C). In ventral view, the maxilla tapers anteriorly, widening posteriorly. The maxilla is widest at the level of the m4, narrowing slightly posteriorly and widening towards the back of the skull. Therefore, in ventral view, the lateral wall is 'S-shaped' and directed lateroventrally (Fig. 2C).</p><p>The lateral surface of the maxilla is smooth, being perforated by numerous neurovascular foramina with a subcircular to oval outline (Fig. 2A). Basally, these foramina are arranged in an anteroposteriorly oriented row that is parallel to the ventral margin of the lateral wall of the maxilla and interpreted as maxillary alveolar foramina. In addition to these, other scattered foramina are present on the lateral wall of the maxilla and are interpreted as median maxillary foramina. In ventral view, the lateral wall of the maxilla is sinusoidal, being convex due to the projection of the lateral wall along the alveoli, and concave between the alveoli (Fig. 2C).</p><p>The maxillae bound the anterior, ventral, and posterior edges of the external nares. In NHMUK PV R 16420, the external naris is retracted posteriorly and reduced, being positioned between alveoli m7 and m9 (Fig. 2A, B). Adjacent to the external naris, the maxilla exhibits a fossa that deepens posterodorsally so that a thin bone lamina (or nasal process— Dal Sasso et al. 2005) projects dorsally into the ventral margin of the external naris. Posteriorly, the base of the ascending process of the maxilla is present, which is overlapped by the maxillary process of the nasal (Fig. 2B). However, the ascending process is poorly preserved.</p><p>In ventral view, the anteromedial processes are basally located in the maxillae and project anteriorly from the level of alveoli m3 reaching the pm4. The anteromedial processes medially divide the conjoined premaxillae ventrally, so that the premaxillary– maxillary contact is 'peg-like' or 'finger-like' (Milner 2003) in ventral view (Fig. 2C). Both processes contact each other medially almost throughout their entire extent, but they diverge at their anteriormost end in the premaxilla.</p><p>The medial shelf in NHMUK PV R 16420 is located on the medial surface of the maxilla and ventrally displaced, being a smooth cylinder-like bony bar (Fig. 2A–C). The medial shelf projects more ventrally than the lateral wall of the maxilla along its entire extension (Fig. 2A, B). The ventromedial surface of the medial shelf is quite flattened anteriorly and this surface becomes rounded posteriorly. Both medial shelves contact each other from the m3 to the m8 alveoli (Milner 2003), forming a secondary palate visible in lateral view (Fig. 2A–C). This contact gradually increases posteriorly and from m8 towards the rear of the skull, the maxillae would come into contact with the nasals in ventral view. In the same view, both maxillae present an anteroposteriorly oriented groove at the posteriormost end that is interpreted as the contact for the palatine (Fig. 2C). The medial shelf dorsomedially limits the interdental plates, which are somewhat rough and oriented ventrolaterally. These are more dorsoventrally oriented anteriorly and become more horizontal posteriorly (Fig. 2C). Regarding their shape, these are anteriorly 'V-shaped' and become subquadrangular from m5 onwards. Furthermore, in ventral view, the interdental plates with the continuous paradental bone (sensu Currie 1987) are 'hourglassshaped' (Fig. 2C). The interdental plates are fused together to form an interdental wall. The paradental groove (nutrient groove sensu Hendrickx and Mateus 2014) is located between the interdental wall and the medial shelf, dividing the lateral and medial half of the maxilla (Fig. 2C). The paradental groove is sinusoidal in ventral view due to its adjustment with the maxillary alveoli (Isasmendi et al. 2023).</p><p>Because the NHMUK PV R 16420 maxillae are not posteriorly complete, the total number of maxillary alveoli cannot be accurately determined. Nevertheless, 10 subcircular-shaped alveoli are preserved in the right maxilla and eight in the left (Fig. 2C). All the alveoli measurements are provided in Table 3.</p><p>The alveoli increase in size from m1 to m4 and then gradually become smaller thereafter (Fig. 2C; Table 3). The three anterior alveoli are oriented anteroventrally, and from the fourth toward the rear of the skull, the alveoli are oriented slightly ventrolaterally rather than ventrally. Hence, an inflection point can be noticed between the alveoli m3 and m4. The anterior alveoli are closer packed (i.e. m1 to m4) and become more spaced posteriorly in the skull, so that more posteriorly located alveoli have the anteroposterior distance between adjacent alveoli similar to the anteroposterior length of the alveoli themselves (Fig. 2C).</p><p>Most of the maxillary alveoli of NHMUK PV R 16420 do not have teeth in situ; however, they are present in alveoli lm2, lm4, and lm6 (Fig. 2A, C). Since the preserved teeth are not fully erupted (e.g. lm4 and lm 6 in Fig. 2A) and are located lingually in the alveoli, these are interpreted as replacement teeth. Similar to the premaxillary dentition, the teeth of the left maxilla have a conidont morphology, the crowns are straight with a centrally located apex. The enamel texture is veined/anastomosed and several flutes are present, at least on the lingual surface of the crown. The basal cross-section of the teeth appears to be subcircular and lenticular in the middle of the crown. The mesial and distal carinae are present and do not have denticles/serrations. Notably, the lm2 tooth is procumbent, while the other two, lm4 and lm6, are laterocumbent (Fig. 2C).</p><p>Nasals</p><p>Only the anterior portion of the conjoined nasals is preserved in the NHMUK PV R 16420 specimen. Additionally, a posterior portion of an isolated conjoined nasal (NHMUK PV R 16426) is also presented (Fig. 5). The preserved portion of the nasals contacts the premaxilla anteriorly and the maxilla anteroventrally, so that the nasals form part of the skull roof, slightly anterior to the external nares towards the posterior part of the skull. Furthermore, the maxillary and premaxillary processes of the nasals are projected laterally on the rostrum, participating in the lateral wall of the skull (Fig. 2A, B).</p><p>The premaxillary and maxillary processes of the nasal bifurcate laterally anteriorly at the posterior margin of the external nares. The premaxillary process of the nasals delimits the straight dorsal margin of the external nares (Fig. 2A, B). The maxillary processes may have delimited the posterior margin of the external nares, but this area is not preserved. The maxillary process of the external nares does not bound its posteroventral margin.</p><p>The premaxillary processes diverge from the dorsal margin and extend on to the lateral surface of the skull above the external nares, more anteriorly than the external nares and overlapping the maxilla laterally. The premaxillary process is projected anteriorly and tapers anteriorly. In lateral view, this is almost horizontal, with its dorsal margin sloping anteroventrally and the ventral margin nearly horizontal (Fig. 2A, B).</p><p>The maxillary process is projected anteroventrally, tapers anteroventrally, and contacts the maxilla laterally. The ventral margin of this process overlaps the maxilla and, posteriorly, this overlap gradually decreases, as the ventral margin of the nasal process ascends posteriorly (Fig. 2A, B). In posterior view, the nasal shaft has an arched morphology, with a convex dorsal margin and a concave ventral surface. The suture line of the nasals is still visible on the preserved portion and no sagittal crest is present on the dorsal surface.</p><p>The posterior portion of the conjoined nasals (NHMUK PV R 16426) is raised into a sagittal crest giving it an inverted-' V shape', which is more intact on the left nasal (Fig. 5A–D). The lateral surface of this crest is smooth, but inside it is pneumatic, as seen in the right nasal, which is abraded and has large foramina in its upper section (Fig. 5G, H). The dorsal edge of the crest has a crenulated profile when viewed from the side. The junction with the lacrimal has a 'wing-like' shape when viewed from above, featuring a deep concavity at the front and a posteroventral concavity at the back (Fig. 5E–H). The posterior portion of the nasal, where it contacts the frontals, is convex when viewed from the side, though the very end of this portion is abraded. In posterior view, this section has a longitudinal groove, with a chambered structure, indicating it was highly pneumatized. The internal (ventral) surface of the nasals is slightly concave and divided by a low sagittal ridge running from front to back. This internal area is mostly smooth, except for a few small foramina and parallel striations. The conjoined nasals narrow slightly at the midpoint of the preserved bones, widen towards the posterior end where they meet the lacrimal, and then narrow again at the very back where they meet the frontal (Fig. 5A–D). However, this last section is broken, missing the tip of the bone.</p><p>Morphological comparisons</p><p>As the skull of spinosaurids is highly derived and specialized, many synapomorphies can be found within their cranium. Owing to this, specimens NHMUK PV R 16420, NHMUK PV R 16422, and NHMUK PV R 16424 share many derived traits with Spinosauridae, as follows.</p><p>The premaxillae of the specimens studied here are hypertrophied and fused (except the poorly preserved NHMUK PV R 16424), with a concave ventral margin in lateral view, posteriorly tapered, and displaying more than five alveoli as in Spinosauridae (Carrano et al. 2012, Barker et al. 2021, Lacerda et al. 2022, Schade et al. 2023). Furthermore, the premaxillary–maxillary articulation is interlocked as in Spinosauridae, rather than flat as in the other early-diverging theropods; additionally, the external nares are posteriorly retracted (Sereno et al. 1998, Taquet and Russell 1998, Dal Sasso et al. 2005, Sales and Schultz 2017, Rauhut and Pol 2019, Barker et al. 2021, Schade et al. 2023).</p><p>Regarding the maxillae, NHMUK PV R 16420 shares the following anatomical features with Spinosauridae: (i) the hypertrophied anterior process of the maxillae; (ii) a plate-shaped anteromedial process of the maxillae, long and projecting far anteriorly; (iii) a proportionally anteroposterior and mediolaterally large, and cylinder-shaped medial shelf; (iv) the sinusoidal lateral wall of the maxilla in ventral view; (v) the size pattern of the alveoli that increases from m1 to m4 and then gradually decreases towards the rear of the skull; (vi) anterior alveoli that are angled anteriorly due to a convex anterior maxillary border; (vii) procumbent anterior maxillary teeth; (viii) conidont dentition; and (ix) veined/anastomosed enamel texture of the crowns (Charig and Milner 1997, Sereno et al. 1998, Taquet and Russell 1998, Sues et al. 2002, Dal Sasso et al. 2005, Canudo et al. 2008, Benson 2010, Kellner et al. 2011, Carrano et al. 2012, Hendrickx and Mateus 2014, Hendrickx et al. 2015, 2019, Alonso et al. 2017, 2018, Sales and Schultz 2017, Rauhut and Pol 2019, Isasmendi et al. 2023, Schade et al. 2023, Souza et al. 2023). Furthermore, the nasals are also at least partially fused in NHMUK PV R 16420, as observed in Spinosauridae (Carrano et al. 2012) .</p><p>Within Spinosauridae, the premaxillary alveoli paired with the presence of diastemata of the specimens described here, as well as the small pm1 tooth/alveolus compared to the other premaxillary teeth/alveoli, are also features shared with Spinosaurinae species, and the latter feature is also present in the baryonychine Riparovenator (Dal Sasso et al. 2005, Carrano et al. 2012, Barker et al. 2021, Lacerda et al. 2022). Furthermore, the alveolar row is interrupted by the premaxillary–maxillary contact in NHMUK PV R 16420, a feature recovered as a synapomorphy of Spinosaurinae by Barker et al. (2021).</p><p>Other features found in the maxillae of NHMUK PV R 16420 snout that are worth highlighting are: (i) lateral wall not projected too far ventrally, so that the lateral maxillary alveoli are visible in lateral view; (ii) conidont teeth with a very subcircular cross-sectional base, without denticles in the distal and mesial carinae; (iii) laterocumbent lateral teeth; and (iv) presence of diastema in the maxillary teeth (Charig and Milner 1997, Carrano et al. 2012, Hendrickx et al. 2015, 2019, Alonso and Canudo 2016). Moreover, the medial shelves of the maxillae in NHMUK PV R 16420 are strongly medially directed, the paradental groove is sinusoidal in ventral view, and the external nares are strongly retracted posteriorly as in other North African spinosaurines (e.g. MNHN SAM 124 and MSNM V 4047) (Dal Sasso et al. 2005, Sales and Schultz 2017, Barker et al. 2021, Isasmendi et al. 2023, Souza et al. 2023).</p><p>In NHMUK PV R 16420, the maxillae delimit the external nares ventrally and posteroventrally, also as in MSNM V 4047 and Irritator (Dal Sasso et al. 2005, Schade et al. 2023). However, in Irritator the narial margin is comparatively more anteriorly placed (Sales and Schultz 2017, Isasmendi et al. 2023) and the latter has a straighter paradental groove. The sinusoidal paradental groove is also present in Oxalaia, but the medial platform is not as medially projected (Isasmendi et al. 2023). Therefore, based on the above, the NHMUK PV R 16420, NHMUK PV R 16422, and NHMUK PV R 16424 specimens can be safely assigned to Spinosauridae, and furthermore, based on the stricter comparisons they can be assigned to Spinosaurinae .</p><p>Lakin and Longrich (2019) compared the NHMUK PV R 16420 snout [NHMUK 16665 in Lakin and Longrich (2019)] with the MSNM V 4047 rostrum and proposed different morphotypes based on some morphological differences in the premaxilla, maxilla, and morphology of external nares, attributing these differences to ontogenetic changes, sexual dimorphism, or distinct taxa. The possibility that they belonged to two different taxa was later ruled out by (Smyth et al. 2020b), who assigned the NHMUK PV R 16420 and MSNM V 4047 to the same taxon. Lacerda et al. (2022) also noted that in the premaxillae of Spinosaurinae specimens from North Africa, there are specific differences such as the degree of constriction of the posterior region of the premaxilla and the degree of expansion of the anterodorsal edge of the premaxilla, as well as the size and location of alveoli pm3 and pm4, which may be due to taphonomic alterations. Based on the reassessment of NHMUK PV R 16420 provided here, we highlight that the only significant difference between the two rostra is the number of premaxillary teeth and the pattern of the intramaxillary suture anteriorly. Six premaxillary alveoli are present in MSNM V 4047 and seven in NHMUK PV R 16420 (Dal Sasso et al. 2005), as in MNHN SAM 124 (Taquet and Russell 1998). However, this feature seems to have no systematic significance, because the right premaxilla of Baryonyx has six alveoli, whereas the left has seven (Charig and Milner 1997, Hendrickx et al. 2016, Lacerda et al. 2022).</p><p>In ventral view, at the level of alveoli m3 and m4, the maxillae of NHMUK PV R 16420 have a curvature similar to that observed in MSNM V 4047, and differing from MNHN SAM 124, in which the curvature is not as pronounced. Furthermore, the dorsoventral constriction of the posterior portion of the premaxillae of NHMUK PV R 16420 is more similar to that observed in MSNM V 4047 and less pronounced than the constriction of the snout of MNHN SAM 124 (thus, MNHN SAM 124 is distant in the premaxilla morphospace from NHMUK PV R 16420 and MSNM V 4047 based on analysis in lateral view— Lacerda et al. 2022). Besides that, the anteriormost alveolar foramina are much smaller in NHMUK PV R 16420 than in MNHN SAM 124, and similar to those found in MSNM V 4047. Therefore, due to these similarities, NHMUK PV R 16420 is considered the same taxon as MSNM V 4047. NHMUK PV R 16420 was previously assigned to cf. Spinosaurus aegyptiacus by Milner (2003) and MSNM V 4047 assigned to Spinosaurus cf . aegyptiacus by Dal Sasso et al. (2005). Other studies (e.g. Ibrahim et al. 2014a, Smyth et al. 2020b) considered NHMUK PV R 16420, MSNM V 4047, and MNHN SAM 124 to belong to Spinosaurus aegyptiacus. Nevertheless, as previous work has suggested (e.g. Evers et al. 2015, Lacerda et al. 2022), because the holotype of Spinosaurus aegyptiacus lacks these elements for direct comparison and correlation, here we consider NHMUK PV R 16420, NHMUK PV R 16422, and NHMUK PV R 16424 as indeterminate Spinosaurinae .</p><p>Posterior portions of spinosaurid nasals are rarely preserved, with only a few species offering points of comparison. The specimen NHMUK PV R 16426 bears resemblance to the spinosaurid nasal UCPC-2 (Dal Sasso et al. 2005), although the latter specimen is located more anteriorly in the snout compared to the former one. The lateral profile of the sagittal crest in NHMUK PV R 16426 is similar to those of UCPC-2 and Baryonyx, featuring a crenulated lateral profile (Charig and Milner 1997, Dal Sasso et al. 2005). However, NHMUK PV R 16426 lacks the longitudinal wrinkles seen in UCPC-2, which are also absent in Baryonyx, probably due to its more posterior position in the skull. The high degree of pneumatization in the crest of NHMUK PV R 16426 is also observed in UCPC-2 (Dal Sasso et al. 2005), in contrast to the solid nasal crest of Irritator (Schade et al. 2023), and probably in Baryonyx and Riparovenator . When viewed from underneath, NHMUK PV R 16426 is broader than the nasal of Baryonyx, which narrows towards the front, giving it an 'arrow-like' shape (Charig and Milner 1997). Despite its incompleteness, NHMUK PV R 16426 shows a convex articulation with the frontals, while in Riparovenator, the lateral profile of the nasals has a concave shape, and in Baryonyx, it is straight (Barker et al. 2021).</p><p>Morphological description (skull roof)</p><p>The specimen NHMUK PV R 16423 is a partial skull roof comprising two frontals, the left prefrontal, and the posterior portion of the left nasal (Fig. 6). The frontals are strongly fused without any signal of suture, measuring 140 mm across the postorbital processes and 160 mm long giving it a width/length ratio of c. 87%, being slightly longer than wide. The frontals taper anteriorly, and the width across the contact with the prefrontals is almost 50% of the width at the postorbital processes (Fig. 6A–D). The medial keel is well projected dorsally in the posteriormost portion of the frontals (for contact with parietal) and is flat and wide, dividing the frontals in two deep lateral trapezoidal fossae at middle-length (Fig. 6A, B). In the posteriormost portion of the frontals, the medial keel is tripartite with flat lateral edges and projected 60° through the postorbital processes disappearing laterally. The medial keel follows the concavity of the lateral fossae and rises again in the middle of the frontals where it becomes tripartite again and high for articulation with the nasals (Fig. 6). The posterior edge of the frontals is straight in dorsal view, suggesting a straight profile for the supratemporal fossa that is shallow in posterior view.</p><p>The contacts for the postorbitals are 'wing-like' shaped and bipartite, with two large foramina in the anterior portion of each process in dorsal view. The right process is more vascularized than the left, especially in the posterior portion of the process. The dorsal surface of the processes is flat laterally to the lateral edge of the medial keel and concave in the anteriormost portion (Fig. 6). In lateral view, the postorbital contact is dorsoventrally elongated and deeply excavated. The nasal contact extends more than 50% of the frontal length and narrows anteriorly from the level of the posterior prefrontal contact (Fig. 6). It is bifurcated in the anteriormost portion and bears small foramina, longitudinal striae, and a keel in the midline between the frontals, originating from the medial keel in dorsal view (Fig. 6A, B).</p><p>The frontals are arched at 40° in lateral view (Fig. 6), making the orbit elevated in the articulated skull. The dorsal orbital margin is circular, with the frontal forming a large part of the orbit. The orbit has a trapezoidal outline in ventral view, with the posterior facet wider than the anterior (Fig. 6C, D).</p><p>Most of the ventral bone surface is smooth. A well-defined ridge of the medial orbitonasal region of the lacrimal/prefrontal complex projects posteriorly to form the paired cristae cranii (Fig. 6C, D) that border the sulcus olfactorius (Fig. 6). Both ridges taper posteriorly, narrowing the sulcus in the middle of the orbit, and widen again to define the lateral border of the cerebral fossae — which are elongated anteroposteriorly and divided by a short, rough septum (Fig. 6C–H). Anterior to this septum, the surface of the bone is crenulated, bearing several small foramina. The contacts of the laterosphenoid are well projected ventrally. They are 'pillar-like' shaped and the ventral surface is very rough and concave (Fig. 6).</p><p>The prefrontal has a large contribution to the margin of the orbit, being visible in lateral view (Fig. 6). The contact between the other bones is rough, well-marked with several foramina (suggesting that the specimen was not a mature individual). In the dorsal portion, the prefrontal has a rough crest forming a cornual process in the form of a protrusion that rises from a concave surface (Fig. 6). The posterior process contacts the frontal by a peg-and-socket suture in lateral view, while in ventral view the suture between the bones is tightly interdigitating. Ventrally, the ventral border of the anteroventral process is confluent with the crista cranii of the frontals. The anterior process extends more anteriorly than the anteroventral process, giving the prefrontal a slight 'T-shape' in lateral view (Fig. 6). The anteroventral process is smaller than the anterior and posterior ones, and inclined almost 45° relative to the roof of the skull.</p><p>Morphological comparisons</p><p>The frontals of the specimen NHMUK PV R 16423 share the following features with Spinosauridae: (i) participation of the frontals in the orbits; (ii) nasals extending posteriorly over the frontals; (iii) elevated orbital margins; (iv) presence of a well-developed medial keel on the dorsal surfaces of the frontals indicating a well-developed parietal crest; and (v) ventrally deflected rostrum (Arden et al. 2019).</p><p>However, the participation of the frontal in the orbit varies within Spinosauridae, being absent in Ceratosuchops and to a lesser extent in Riparovenator (Barker et al. 2021), as well as Baryonyx and Suchomimus (i.e. Baryonychinae). In Spinosaurinae, the frontals are more conspicuous and curved in lateral view (e.g. Arden et al. 2019). The frontal excluded from the orbital margin due to the lacrimal/postorbital contact is observed in carcharodontosaurids (e.g. Meraxes — Canale et al. 2022, Eocarcharia and Carcharodontosaurus — Sereno and Brusatte 2008), and abelisaurids (e.g. Majungasaurus — Sampson and Witmer 2007 and Carnotaurus, Cerroni et al. 2021), the other two large-bodied theropod groups that are also found in the Kem Kem Group (e.g. Ibrahim et al. 2020b).</p><p>The overall morphology of NHMUK PV R 16423 resembles those of FSAC-KK-3209, FSAC-KK-3210 ('Morphotype A'), and FSAC-KK-7715 ('Morphotype B') of Arden et al. (2019). These specimens were initially proposed as cf. Spinosaurus aegyptiacus ('Morph A') and? Sigilmassasaurus brevicollis ('Morph B') based on different proportions. 'Morph B' is shorter and wider, with weaker concave orbital margins, less concave orbital edges, a more deeply excavated postorbital process, a higher sagittal crest, and a wider overlapping contact of the frontals with the prefrontals (Arden et al. 2019). Later, Ibrahim et al. (2020b:160) commented that Arden et al (2019) 'Morph A' resembles the morphology of marine crocodyliforms, considering the condition of the posterior distance of the braincase in relation to the frontals; however, no detailed morphological redescription/comparison was provided, therefore, we consider here the spinosaurid association of the 'Morph A' braincase provided by Arden et al. (2019).</p><p>The specimen studied here, NHMUK PV R 16423, has the anterior portion of the frontal tapering to come into contact with the nasals, as in 'Morph A' of Arden et al. (2019). NHMUK PV R 16423 is more similar to FSAC-KK-3209, which has the contact for the nasals occupying around 50% of the total frontal length. Also, it resembles 'Morph A' by having a similar curvature and dorsal projection of the orbit in lateral view. Conversely, the general proportion is more similar to 'Morph B' having a width/ length ratio of c. 87% in NHMUK PV R 16423, compared to the c. 93% of 'Morph B' (Arden et al. 2019). NHMUK PV R 16423 also shares with 'Morph B' a deep-notched postorbital process, a broad overlapping contact between the frontal and prefrontal, and a high sagittal crest. These features of NHMUK PV R 16423 overlap with the morphology of both morphotypes from Arden et al. (2019), which may reinforce the suggestion of Smyth et al. (2020b) that the morphotypes represent variations (individual, ontogenetic, or sexual) within a unique species.</p><p>In ventral view, NHMUK PV R 16423 presents a closed interfrontal suture, differing from those of Ceratosuchops, Riparovenator (Barker et al. 2021), and possibly FSAC-KK-3209, as well as FSAC-KK-7715 (Figs 2B and 3B respectively— Arden et al. 2019). In NHMUK PV R 16423 and 'Morphs A and B' (Arden et al. 2019), the nasal overlaps the frontal contact, a different condition observed in Irritator (Schade et al. 2023), Ceratosuchops, and Riparovenator (Barker et al. 2021) . Besides that, in NHMUK PV R 16423 and 'Morphs A and B' (Arden et al. 2019), the contact with frontal and nasal is strongly interdigitated, having an 'M-shape' in dorsal view, unlike the condition of Irritator, Ceratosuchops, and Riparovenator, which have a rough and semicircular surface for nasal contact (Barker et al. 2021, Schade et al. 2023).</p><p>The parietal contact in NHMUK PV R 16423 differs from those other spinosaurids in having a straight transverse posterior margin of the postorbital process, which is posteriorly projected in Irritator, Ceratosuchops, Riparovenator, 'Morph A', and 'Morph B' (Arden et al. 2019, Barker et al. 2021, Schade et al. 2023). However, in the latter 'Morph' the postorbital process is less projected than in other spinosaurids, approaching the condition seen in NHMUK PV R 16423.</p><p>Finally, the prefrontal of NHMUK PV R 16423 differs from other spinosaurids in having a 'T-shape' rather than a 'hookshape', as seen in Irritator (Schade et al. 2023), Ceratosuchops, Riparovenator (Barker et al. 2021), and Suchomimus (Sereno et al. 1998) . The prefrontal of Baryonyx differs from other spinosaurids because it lacks a ventral process (Charig and Milner 1997). The prefrontal cornual process of NHMUK PV R 16423 is also seen in Ceratosuchops, Riparovenator, and Suchomimus, but not in Irritator and Baryonyx (Charig and Milner 1997, Sereno et al. 1998, Barker et al. 2021, Schade et al. 2023). The angle of the anteroventral process resembles that of Irritator (Schade et al. 2023), whereas in Baryonychinae the angle of this process is more acute (Charig and Milner 1997, Sereno et al. 1998, Barker et al. 2021). Based on the aforementioned six synapomorphies shared between NHMUK PV R 16423 and Spinosauridae, we can safely consider this new specimen as belonging to this clade. Furthermore, considering the similarities with Irritator and the contemporaneous spinosaurines described by Arden et al. (2019), we identify NHMUK PV R 16423 as an indeterminate Spinosaurinae .</p><p>Morphological description (axial skeleton)</p><p>The dorsal series is represented by a mid-dorsal vertebra that preserves the neural arch and part of the centrum (NHMUK PV R 16430) and a neural dorsal spine lacking its dorsalmost portion (NHMUK PV R 16431). Specimen NHMUK PV R 16430 is interpreted as a middle dorsal vertebra, since the parapophysis is between the neural arch and the centrum, as in the mid-dorsal vertebrae of the spinosaurids Baryonyx and Vallibonavenatrix (Charig and Milner 1997, Malafaia et al. 2020).</p><p>The neural spines in both specimens, NHMUK PV R 16430 and NHMUK PV R 16431, are hypertrophied, being large and tall (Figs 7, 8). In lateral view, they are narrower anteroposteriorly expanding dorsally. Dorsal to this expansion, the neural arches become narrower anteroposteriorly and gradually expand dorsally. The neural spines are laterally compressed (Figs 7A–D, 8C–F), projected dorsally or even slightly anteriorly, and their cross-section is quite tabular, being slightly transversely wider anteriorly and at the level of the transverse processes in dorsal view (Fig. 8E, F). Interspinous ligament scars are present and developed similarly on the anterior and posterior surfaces, but the posterior scar extends further dorsally than the anterior (Figs 7A–D, 8C–F). The scars of the interspinous ligament are limited laterally by the spinoprezygapophyseal (e.g. Fig. 8A, B, G, H) and postzygodiapophyseal laminae. The bases of the neural arch in NHMUK PV R 16430 exhibit webs at the level of the spinodiapophyseal fossa (Fig. 7G, H).</p><p>The spinoprezygapophyseal laminae extend from the neural spine to the dorsal surface of the bases of the prezygapophyses, delimiting the spinoprezygapophyseal fossa (Figs 7C, D, 8E–H), which is dorsoventrally elongated and deep (e.g. Fig. 8E, F). Spinopostzygapophyseal laminae are also present in both partial vertebrae. These structures are narrower than the spinoprezygapophyseal laminae and extend from the posterior surface of the neural spine to the dorsal margin of the postzygapophyses, reaching the most posterior point of the postzygapophyses (Fig. 7C, D). In both NHMUK PV R 16430 and NHMUK PV R 16431, these laminae delimit the spinopostzygapophyseal fossa, which is similar to the spinoprezygapophyseal fossa but much deeper at the base in NHMUK PV R 16430 (Fig. 7A–D). Only the spinopostzygapophyseal fossa is preserved in NHMUK PV R 16431 (Fig. 8E–H). In NHMUK PV R 16430, the postzygodiapophyseal lamina runs anteriorly from the postzygapophyses to the posterior margin of the transverse processes.</p><p>The bases of the transverse processes in NHMUK PV R 16430 are projected laterally and close to the horizontal. These are also dorsoventrally compressed, 'sheet-like' and appear to narrow slightly anteroposteriorly towards the diapophyses (Fig. 7A–F). Under the transverse processes, the neural arch is strongly pneumatized. NHMUK PV R 16430 has both anterior and posterior centrodiapophyseal laminae (Fig. 7). The anterior centrodiapophyseal lamina extends from the ventral surface of the transverse process and reaches the posterodorsal margin of the parapophysis. The posterior centrodiapophyseal lamina runs from the ventral surface of the transverse process and extends posteroventrally to the centrum (Fig. 7I–J). The anterior centrodiapophyseal lamina is narrower and sharper compared to the more robust and rounded posterior centrodiapophyseal lamina. Adjacent to the transverse processes, three triangular fossae in lateral view are also present. These are the prezygocentrodiapophyseal fossa, the centrodiapophyseal fossa, and the postzygocentrodiapophyseal fossa (Fig. 7G–J). The three fossae are equally deep; however, the centrodiapophyseal fossa is thelargestofthem.Theprezygocentrodiapophysealfossaisdelimited by the prezygodiapophyseal and anterior centrodiapophyseal laminae (Fig. 7G–J). The centrodiapophyseal fossa is delimited by the anterior and posterior centrodiapophyseal laminae, and the postzygocentrodiapophyseal fossa is delimited by the posterior centrodiapophyseal and the postzygodiapophyseal laminae (Fig. 7G–J).</p><p>The postzygapophyses project dorsolaterally and their articular surfaces face ventrolaterally. The articular surface of the postzygapophyses is elliptical and flat (Fig. 7C–J). In posterior view, the postzygapophyses are medially fused, but there is no apparent hyposphene.</p><p>Morphological comparisons</p><p>The base of the NHMUK PV R 16430 vertebra displays the webbing pattern in its spinodiapophyseal fossae, a feature that has been recovered as a synapomorphy of Spinosauridae (Carrano et al. 2012, Evers et al. 2015, Malafaia et al. 2020, Barker et al. 2021, Mateus and Estraviz-López 2022, Isasmendi et al. 2024). The NHMUK PV R 16430 vertebra does not have the accessory lamina that projects anteroventrally from the posterior centrodiapophyseal lamina, differing from the middle caudal vertebrae of Baryonychinae species (Sereno et al. 1998, 2022, Benson 2010, Allain et al. 2012, Carrano et al. 2012, Barker et al. 2021) or other spinosaurids (e.g. Evers et al. 2015, Malafaia et al. 2020, Isasmendi et al. 2024), and as in Spinosaurus aegyptiacus, including the neotype FSAC-KK 11888 (e.g. Carrano et al. 2012, Ibrahim et al. 2014 a, Schade et al. 2023).</p><p>Furthermore, the hypertrophied dorsal neural spine is much taller than that of non-spinosaurid tetanuran theropods such as Acrocanthosaurus (Stovall and Langston 1950), and other spinosaurids such as Baryonyx, Ichthyovenator, and Suchomimus (Charig and Milner 1997, Sereno et al. 1998, 2022, Allain et al. 2012). Similarly, hypertrophied neural spines occur in the spinosaurines Spinosaurus aegyptiacus and FSAC-KK 11888, which are also directed anterodorsally (Stromer 1915, Smith et al. 2006, Ibrahim et al. 2014a) as the condition noted in NHMUK PV R 16431. The neural spines of the dorsal vertebrae described here further resemble the African spinosaurines, e.g. in NHMUK PV R 16431, the neural spine widens anteroposteriorly near the base of the neural spine, then narrows dorsally, and after this constriction it widens again, but gradually in the same way as occurs in Spinosaurus aegyptiacus (Stromer 1915, Smith et al. 2006, Ibrahim et al. 2014a). Therefore, these specimens closely resemble other contemporary African spinosaurines in morphology [an autapomorphic feature of Spinosaurus aegyptiacus; the hypertrophied dorsal neural spine — Stromer (1915), and the webbing (striated) pattern at the base of the neural spine, being a synapomorphic feature of Spinosauridae — Malafaia et al. (2020) and Isasmendi et al. (2024)]. Thus, both NHMUK PV R 16430 and NHMUK PV R 16431 are referred here as indeterminate Spinosaurinae .</p><p>Morphological description (ilia)</p><p>The two partial right ilia present different degrees of preservation. The structures that are preserved in both specimens, and therefore overlap, are virtually identical. Specimen NHMUK PV R 16391 preserves the postacetabular blade, the iliac peduncle, the entire brevis fossa, and part of the supra-acetabular crest (Fig. 9); whereas NHMUK PV R 16438 preserves part of the iliac blade, both peduncles, the brevis fossa, and the supra-acetabular crest (Fig. 10).</p><p>Specimen NHMUK PV R 16438 has an anteroposterior total length of c. 447 mm (Fig. 10). Only the base of the pre-acetabular process is preserved in NHMUK PV R 16438, expanding anteroventrolaterally, providing a smooth 'U-shaped' curvature between the ventral margin of the pre-acetabular blade and the anterior surface of the pubic peduncle (Fig. 10A, B). The portion between the anteroventral process of the ilium and pubic peduncle forms the pre-acetabular notch (or 'cuppedicus' fossa), but does not form a true fossa, as this region is shallow in NHMUK PV R 16438 (Fig. 10A–D). There is no sign of a shelf medial to the pre-acetabular notch. The pubic peduncle of NHMUK PV R 16438 is relatively small and directed ventrally, suggesting a propubic pelvis. Two foramina are on the pubic peduncle. The articulations of the ilium with the pubis and ischium appear to have the same or similar proportions, so both the pubic and ischiatic peduncles are relatively similar in size (Fig. 10A, B). The articular surface of the pubis is triangular in ventral view, being wider in the acetabular region and narrower anteriorly (Fig. 10C, D). Furthermore, the distal outline of the pubic peduncle has its anterior part more ventrally positioned in lateral view, whereas its posterior portion near to the acetabulum is more dorsally positioned, being the anterior limit of the supra-acetabular crest (Fig. 10A, B).</p><p>The dorsal surface of the acetabulum is smooth, wide, and slightly straight, presenting a dorsally smooth concave shape in NHMUK PV R 16438. In ventral view, its supra-acetabular crest is completely preserved, projecting from the posterolateromedial portion of the ischiatic peduncle to form the posterolateral acetabular edge of the pubic peduncle (Fig. 10A–D). The supra-acetabular crest is directed ventrolaterally, being well developed but not a large/pendant 'hood' (sensu Carrano et al. 2012), with an ovoid outline, and occluding the anteroventral region of the acetabulum in lateral view (Fig. 10C, D). In the specimen NHMUK PV R 16391 (total length of c. 430 mm), the supra-acetabular crest projects from the posterolateral portion of the ischiatic peduncle (slightly more posterior than in NHMUK PV R 16438), also being a well-developed and short crest directed anterolaterally, but only the posterior part is preserved (Fig. 9A–F).</p><p>On the posterolateral surface of both ilia, there is a gap that separates the supra-acetabular crest and the most anteroventral part of the brevis shelf in the postacetabular blade, posterior to the ischiatic peduncle (Figs 9E, F, 10C, D). The brevis shelf is a cylinder-like bony bar; more pronounced posteriorly; giving it a 'lobular' appearance in posterior view (Fig. 9I, J). The ischiatic peduncle is directed anteroposteriorly; its anterior acetabular surface is flat (particularly in the better-preserved specimen NHMUK PV R 16391), while the posterior surface is concave with the distal end tapering posteroventrally (Figs 9A, B, 10A, B). Thus, the outline of the ventral border of the brevis shelf and the posterior edge of the ischiatic peduncle is 'hook-shaped' (Fig. 9A–D). The distal facet of the ischiatic peduncle in NHMUK PV R 16391 is rounded and NHMUK PV R 16438 has its distal ischiatic peduncle eroded medially; however, it still seems to have a rounded morphology.</p><p>The postacetabular length in relation to the length of the ischiatic peduncle is greater than 1 in both specimens, being a deep process projected posteroventrolaterally with a straight outline (Figs 9A–D, 10A, B). In the most ventral postacetabular blade, the lateral wall of the brevis fossa is high throughout its entire length, thus, the medial wall of the fossa is concealed in lateral view (Figs 9A–D, 10A, B). The brevis fossa is located ventrally in the postacetabular blade, posterior to the ischiatic peduncle and restricted to the lateral and medial wall of the brevis fossa. This is relatively deep and wide, becoming posteriorly wider. The lateral wall of the fossa is expanded laterally in anterior view, contributing to the posterior widening of the brevis fossa (Figs 9, 10C, D). In posterior view, the lateral wall of the brevis fossa is oblique relative to the main axis of the ischiatic peduncle, while the medial wall is almost horizontal and directed ventromedially (Figs 9I, J, 10E, F).</p><p>At least two foramina can be seen on the lateral surface of the iliac blade of NHMUK PV R 16391, one positioned anteriorly at the same level as the anterior surface of the ischiatic peduncle and the other positioned posteriorly at the medial level of the postacetabular process (Fig. 9A, B).</p><p>Dorsal to the supra-acetabular crest, some radial scars are present on the iliac blade, reminiscent of muscular origins. The preserved part of the iliac blade is flat and smooth in NHMUK PV R 16391 (Fig. 9A, B) and NHMUK PV R 16438; thus, there is no sign of a vertical ridge on the ilium. In lateral view, the shape of the dorsal margin of the ilium is convex. In dorsal view, the outline of the dorsal edge of the ilium has a sigmoid shape; more laterally expanded posteriorly, and directed medially becoming anteriorly projected (or straight) in the most anterior region of the dorsal border. This straighter dorsal part can also be seen in lateral view, where the anterior portion is flatter dorsoventrally, softening the curve of the dorsal outline of the ilium anteriorly (Fig. 9A, B). In this way, the dorsal edge of the postacetabular process and its ventral edge become parallel (Fig. 9A, B).</p><p>In the dorsoposterior border of the postacetabular process, NHMUK PV R 16391 presents the crista dorsolateralis ilii (sensu Baumel and Witmer 1993) of the ilium well preserved, in addition to the muscle–bone contact area, which is rough, presenting some undulations of the bone surface (Fig. 11). Such characteristics are the osteological correlates of the most superficial and posterior muscles of the thigh M. iliotibialis 3 (IT3; level I inference of Witmer 1995).</p><p>The shape of the posterior surface of the postacetabular process in NHMUK PV R 16391 is convex; however, its posterior border is asymmetrical, with its posterodorsal margin flattening posteriorly ventrally and projecting posteriorly (but not as much as in Megalosaurus; Benson 2010, Lacerda et al. 2023). This posterior projection occurs at the level of contact with the medial wall of the brevis fossa. The morphology of the posteroventral edge of the postacetabular process is straight (Fig. 9A–D). Although specimen NHMUK PV R 16438 has its postacetabular process more eroded, its posteroventral edge is also straight (Fig. 10A, B).</p><p>The medial surface of NHMUK PV R 16391 is well preserved. The medial wall of the brevis shelf, as previously noted, is quite horizontal, being an arched structure; its posterior limit is in the region of greatest posterior expansion of the ilium, becoming more dorsal anteriorly, and thus participating in the deepening of the brevis fossa (Fig. 9C, D). The medial wall occludes the brevis fossa anteriorly, as this structure becomes more ventrally projected, mainly in the probable contact with the sacral vertebra 1 (in an irregular and striated region), and anteriorly in the contact with the sacral vertebra 2, located adjacent to the ischiatic peduncle (Fig. 9C, D).</p><p>Morphological comparisons</p><p>In a number of features, both NHMUK PV R 16391 and NHMUK PV R 16438 resemble Tetanurae in: (i) the presence of a reduced ventrolateral supra-acetabular crest, occluding the anterodorsal portion of the acetabulum; (ii) ventrally oriented pubic peduncle (also seen in ceratosaurs); (iii) convex shape of the dorsal margin of the ilium; and (iv) the morphology between the supra-acetabular crest and the brevis shelf, on the lateral surface, as a gap (differing from the continuous ridge of ceratosaurs) (Hutchinson 2001b, Benson 2010, Carrano et al. 2012, Cuesta et al. 2018, Malafaia et al. 2020, Lacerda et al. 2023). Moreover, the presence of a shallow pre-acetabular notch, not forming a true fossa, as well as the lack of a ridge on the ilium, just medial to the pre-acetabular notch, which can be considered based on the NHMUK PV R 16438 specimen, indicates differences between this individual and taxa of Avetheropoda (sensu Hutchinson 2001b, Benson 2010, Carrano et al. 2012, Malafaia et al. 2020). The presence of a brevis fossa that widens posteriorly differentiates both ilia described here from metriacanthosaurid avetheropods (Carrano et al. 2012, Lacerda et al. 2023).</p><p>Based on their combination of morphological features, both ilia described here more closely resemble the general morphology of spinosaurids than other theropods. Similar to the spinosaurid Ichthyovenator (Allain et al. 2012), both ilia related to North African spinosaurines (FSAC-KK 11888 and MSNM V 6900; Ibrahim et al. 2014a), as well as the indeterminate specimen MN 4819- V (Machado 2010), both NHMUK PV R 16391 and NHMUK PV R 16438 do not present any sign of a vertical ridge on the lateral surface of the ilium. This differs from the spinosaurids Vallibonavenatrix (Malafaia et al. 2020) and Suchomimus, which present a low ridge, and later-diverging metriacanthosaurids, which have a double ridge (Carrano et al. 2012, Lacerda et al. 2023).</p><p>The moderately ventromedial extension of the supra-acetabular crest and its outline morphology in ventral view in NHMUK PV R 16438 resemble spinosaurids such as Vallibonavenatrix and the lateral morphology of this crest is also similar to, for instance, Vallibonavenatrix, Ichthyovenator, and Suchomimus (Allain et al. 2012, Malafaia et al. 2020, Lacerda et al. 2023).</p><p>The brevis shelf as a cylinder-like bony bar ventrally is also seen in Ichthyovenator (Allain et al. 2012), Vallibonavenatrix (less robust) (Malafaia et al. 2020), MSNM V 6900, and probably Suchomimus; besides this, the shape of the brevis fossa— widening posteriorly—is also shared among spinosaurids (except Ichthyovenator) and both NHMUK PV R 16391 and NHMUK PV R 16438. However, the widening of the brevis fossa is homoplastic in theropods (Lacerda et al. 2023). Several theropods have the lateral wall of the brevis fossa shorter than the medial wall anteriorly, exposing the medial wall in lateral view. In both ilia, the lateral wall is taller along its entire length, hiding the brevis fossa in lateral view, as occurs in the spinosaurids Suchomimus, FSAC-KK 11888 and MSNM V 6900 (Ibrahim et al. 2014a). The shape of the brevis fossa in posterior view and its medial and lateral walls in NHMUK PV R 16391 is identical to the morphology of the FSAC-KK 11888 specimen by having a horizontal medial wall, and the posterolateroventrally directed distal portion forming the 'lobular-like' lateral wall of the brevis fossa.</p><p>NHMUK PV R 16391 also shares with Suchomimus, FSAC-KK 11888, and probably also MSNM V 6900 a flat and dorsally convex shape of the dorsal rim of the postacetabular process of the ilium. Although the posterior outline of the postacetabular is convex in NHMUK PV R 16391 and other spinosaurids (but also avetheropods). The shape of the ilia described here resembles more the morphology of FSAC-KK 11888 and MSNM V 6900 than other specimens, by having an asymmetrical posterior surface of the postacetabular process, with its posterodorsal margin flattening ventroposteriorly and projecting posteriorly (but not as much as in Megalosaurus; Benson 2010, Lacerda et al. 2023) and the straight posteroventral edge of the postacetabular process. Nevertheless, the shape of the posterior facet of the ischiatic peduncle, as well as the posteroventrolaterally directed lateral wall of the brevis fossa, giving it a 'hook-shaped' posteroventral outline. This shape is similar among NHMUK PV R 16391, NHMUK PV R 16438, and the spinosaurines FSAC-KK 11888 and MSNM V 6900. In lateral view, the ventral margin of the postacetabular process is straight in spinosaurids, such as Ichthyovenator, Suchomimus, and FSAC-KK 11888. Moreover, this margin is ventrally directed posteriorly; the same morphology noted in NHMUK PV R 16391 and NHMUK PV R 16438.</p><p>Interestingly,among tetanurans a feature noted only in NHMUK PV R 16438 and spinosaurines, such as FSAC-KK 11888, MSNM V 6900, and the Brazilian specimen MN 4819- V (Machado 2010), is the morphology of the pubic peduncle and its relative size with the ischiatic peduncle. Generally, tetanurans have the pubic peduncle larger than the ischiatic peduncle. This can be 130% larger or more, and most of these theropods have a distal pubic expansion (Hutchinson 2001b, Benson 2010, Allain et al. 2012, Carrano et al. 2012, Malafaia et al. 2020, Lacerda et al. 2023, 2024). As previously mentioned, the exceptions noted by Malafaia et al. (2020) are the spinosaurines, which have both iliac peduncles similar in proportion. Concerning the ilium FSAC-KK 11888, Lacerda et al. (2023) considered the pubic peduncle slightly larger than the ischiatic, whereas Malafaia et al. (2020) considered both at the same proportions. In spite of that, a small pubic peduncle is only found in the Spinosaurinae clade among Tetanurae, this being the same feature noted in NHMUK PV R 16438.</p><p>Finally, the (M. iliotibialis 3) muscle scar on the dorsal rim of the postacetabular blade, noted in NHMUK PV R 16391, is similar in topology with other theropods (e.g. Carrano and Hutchinson 2002), and the rough pattern is similar to other megalosauroids (Lacerda et al. 2024).</p><p>Thus, based on the several features described and the similarities with spinosaurine theropods, we refer both NHMUK PV R 16391 and NHMUK PV R 16438 as indeterminate Spinosaurinae individuals.</p><p>Morphological description (femur)</p><p>Specimen NHMUK PV R 16433 is a well-preserved left femur, missing only the distal end (Fig. 12). The femur is c. 371 mm in length. The femoral shaft, in anterior and posterior views, has a straight shape (Fig. 12A–C); in medial and lateral views, it curves markedly posteriorly (Fig. 12E–H), being especially arched in the most distal portion of the diaphysis. Thus it is sigmoid in medial and lateral views (Fig. 12E–H). The transverse section of the diaphysis (based on a distal break) is elliptical in shape.</p><p>The femoral head is positioned c. 15° anteromedially (in dorsal view; Fig. 12I, J) relative to the posterior edge of the proximal portion of the distal condyles (not shown in Fig. 12I, J), and oriented medially (based on the main axis of the diaphysis), with the dorsal surface positioned horizontally in anterior view (Fig. 12A, B). The femoral neck is well developed. The medial articular surface of the femoral head is straight, tapering distally, having an irregular and almost oval shape (Fig. 12G, H). On the posterior surface of the caput there is a longitudinal groove that slopes laterally in the distal portion. The distolateral delimitation of this groove appears to be the base of the flange of the caput; however, this region may have suffered abrasion (Fig. 12C, D). There is a shallow portion on the posterior surface of the femoral head in NHMUK PV R 16433, which seems to be homologous with the oblique ligament groove in theropods (e.g. Carrano et al. 2012, Malafaia et al. 2018). In dorsal view, the head has a straight posterior surface and an arched anterior surface, becoming mediolaterally narrow towards the greater trochanter (Fig. 12I, J).</p><p>The greater trochanter is continuous with the femoral head. It is wide and posterolaterally expanded in its distal portion, forming a smooth and rounded junction with the posteroproximal femoral shaft (Fig. 12C, D). The junction of the greater trochanter in the anteroproximal region of the diaphysis becomes even smoother with the distal narrowing of the greater trochanter (Fig. 12E, F).</p><p>The lesser trochanter is robust and relatively thick, especially in the distal region at its junction with the femoral shaft, being more tapered posteroproximally near the cleft that separates this structure from the greater trochanter (Fig. 12A, B I, J). The proximal contact of the lesser trochanter with the greater trochanter is smooth, and the cleft that divides these structures is shallow. The lesser trochanter forms a wide semi-oval flange. Its dorsal edge is 'U-shaped' and its proximal end slightly exceeds the ventral margin of the femoral head (Fig. 12A, B). The anterolateral surface of the lesser trochanter is flat, while its anteromedial surface is irregular. The accessory trochanter projects slightly anterodistally from the lesser trochanter, but this is a poorly developed structure represented by a thickening of the distal margin, giving an irregularly arched shape to the anterior edge of the lesser trochanter (Fig. 12A, B E, F). At the laterodistal base of the lesser trochanter, a distinct trochanteric shelf is present, formed by a smooth protuberance, located on the lateral surface of the diaphysis, proximal to the fourth trochanter (Fig. 12C–F). These structures border a shallow concavity between the distolateral greater trochanter, the posterior surface of the lesser trochanter, and the proximal edge of the trochanteric shelf (Fig. 12E, F).</p><p>The fourth trochanter is hypertrophied and appears as a prominent elliptical-shaped flange, being longer proximodistally than anteroposteriorly and occupying one-third of the femur (Fig. 12G, H). Proximally, the fourth trochanter originates smoothly near the level of the distal base of the lesser trochanter and expands proximodistally to the medial region of the femoral shaft, ending in a gentle curve (Fig. 12G, H). In lateral view, the fourth trochanter protrudes slightly, having a straight outline (Fig. 12E, F). A protrusion is present on the most posterior rim of the fourth trochanter, being slightly asymmetrical and positioned distally (Fig. 12G, H).</p><p>The posterolateral surface of the fourth trochanter is smooth, while the posteromedial surface is irregular with a well-marked keel separating two depressions: (i) a deeper and narrower sulcus that extends longitudinally and parallel to the fourth trochanter at its base, originating in the proximal portion and occupying at least two-thirds of the fourth trochanter; and (ii) a shallower and wider depression, located medial to the fourth trochanter and the sulcus previously mentioned. It originates approximately at the medial height of the proximal fourth trochanter (more distal than the longitudinal sulcus), widening distally and becoming shallower at the distal base of the fourth trochanter (Fig. 12G, H). Both depressions represent osteological correlates related to the attachment of strong hip extensor muscles (e.g. Gatesy 1990, Hutchinson 2001a, Carrano and Hutchinson 2002). Based on the femur NHMUK PV R 16433, we infer that the entire lateral surface of the fourth trochanter was the region of insertion of the caudofemoralis brevis muscle (CFB); whereas the medial surface of the fourth trochanter, the medial sulcus, and the large elliptical groove on the medial side was a broad insertion for a large caudofemoralis longus muscle (Fig. 12G, H). Both Mm. caudofemorales are a level II inference by Witmer's (1995) systematization, being in positions topologically similar to other theropods (e.g. ceratosaurs— Cerroni et al. 2024, early tetanurans— Lacerda et al. 2024, and late-diverging coelurosaurs— Carrano and Hutchinson 2002).</p><p>Distally, the posterior region of the femoral shaft is straight and expands slightly mediolaterally at the distal end, proximally to where the condyles would be (Fig. 12A–D). On the medial and distal surfaces of the femur, the anteromedial (or 'craniomedial distal') crest is rounded and poorly developed. NHMUK PV R 16433 also has a poorly developed, rounded medial epicondyle (or medial distal crest) (Fig. 12C, D). Although the distal condyles are not preserved, distally the femoral shaft, in posterior view, forms a gap that separates two posterior elevations, which probably represent the proximal border of the linea aspera (which is visible in anterior view; Fig. 12A, B) on the medial side of the tibiofibular crest; therefore, being homologous to the flexor groove of other theropods (e.g. Hutchinson 2001a, Benson 2010).</p><p>Morphological comparisons</p><p>Specimen NHMUK PV R 16433 shares several features with tetanuran theropods, such as: (i) the femoral head positioned anteromedially and oriented medially; (ii) the dorsal edge of the lesser trochanter slightly surpassing the ventral margin of the femoral head; (iii) the reduced trochanteric shelf; and (iv) the lesser trochanter resembling a wide flange (Gauthier 1986, Hutchinson 2001a, Benson 2010, Carrano et al. 2012, Cuesta et al. 2018, Lacerda et al. 2023).</p><p>In the specimen studied here, the position and orientation of the femoral head differs from most allosauroids, such as the metriacanthosaurid Neovenator and the carcharodontosaurids Acrocanthosaurus and Carcharodontosaurus, and Giganotosaurus, being oriented medially and at a dorsomedial angle in the latter (Stromer 1931, Carrano et al. 2012, Cuesta et al. 2018, Lacerda et al. 2023). A femoral head positioned anteromedially and oriented medially is shared between NHMUK PV R 16433 and both megalosauroid and metriacanthosaurid tetanurans (Charig and Milner 1997, Benson 2010, Machado 2010, Carrano et al. 2012, Malafaia et al. 2018, Lacerda et al. 2023).</p><p>Interestingly, there is only a very shallow portion on the posterior surface of the femoral head in NHMUK PV R 16433, which may be homologous with the groove of the oblique ligament. In theropods, it is generally represented by a deep and bounded structure, except in the megalosauroids Aþovenator, Megalosaurus, Torvosaurus, and FSAC-KK 11888, which only have a shallow groove (Benson 2010, Carrano et al. 2012, Lacerda et al. 2023).</p><p>In non-Avetheropoda theropods, including African spinosaurine FSAC-KK 11888 (Lacerda et al. 2023) and NHMUK PV R 16433, the accessory trochanter (which derives from the lesser trochanter), is represented by a weak structure that forms the slightly thickened margin of the lesser trochanter. This is the condition observed in NHMUK PV R 16433, which differs from the spinosaurid Suchomimus and also Avetheropoda taxa in general, which have the accessory trochanter as a triangular flange (Benson 2010, Carrano et al. 2012, Sereno et al. 2022, Lacerda et al. 2023).</p><p>NHMUK PV R 16433 also shares with spinosaurids (and other megalosauroids) the rounded and poorly developed morphology of the medial epicondyle, which differs from the ridge-shaped structure present in ceratosaurs and allosauroids (Allain et al. 2007, Benson 2010, Carrano et al. 2012, Evans et al. 2015, Malafaia et al. 2018, Sereno et al. 2022, Lacerda et al. 2023, 2024, Cerroni et al. 2024, Isasmendi et al. 2024).</p><p>Based on the four main characteristics of the proximal femur described previously, along with the presence of a low and rounded medial epicondyle and the probably shallow groove of the oblique ligament, it is possible to assign the NHMUK PV R 16433 femur to the Megalosauroidea clade. Within this clade, NHMUK PV R 16433 more closely resembles the morphology of spinosaurids than that of other megalosauroids.</p><p>The overall shape of the femoral head in NHMUK PV R 16433 resembles the spinosaurids Baryonyx, FSAC-KK 11888, Suchomimus (Ibrahim et al. 2014 a, Sereno et al. 2022, Lacerda et al. 2023), and specimens MN 4819- V (Machado 2010) and CMP-MS-0/22 (Malafaia et al. 2018). The medial surface of the femoral head in NHMUK PV R 16433 resembles the shape of Suchomimus more than in FSAC-KK 11888, as it has a horizontal and continuous dorsal surface of the femoral head, which is less arched on its anterior edge, as seen in dorsal view. Nevertheless, the remainder of the proximal femoral morphology described here is similar to that observed in FSAC-KK 11888. Similarities between FSAC-KK 11888 and NHMUK PV R 16433 include: (i) greater trochanter morphology; (ii) position and morphology of the lesser trochanter, which has a weak accessory trochanter forming only a thick distal margin (differing from MN 4819- V — Machado 2010 and Suchomimus — Lacerda et al. 2023); (iii) position and shape of the trochanteric shelf; (iv) the overall shape and position of the fourth trochanter; and (v) degree of anteroposterior curvature of the femoral shaft, which is markedly curved in FSAC-KK 11888, NHMUK PV R 16433, and probably Riojavenatrix, than in Suchomimus and CMP-MS-0/22 (Ibrahim et al. 2014a, Malafaia et al. 2018, Sereno et al. 2022, Isasmendi et al. 2024). Furthermore, based on the muscles that were inserted on to the fourth trochanter of NHMUK PV R 16433, it is possible to infer that the Mm. caudofemorales represented relatively robust muscles (or the tendons thereof) in this specimen. The osteological correlates are much more evident than in other megalosauroids (e.g. piatnitzkysaurids— Lacerda et al. 2024), which potentially is consistent with the proposal of a robust and strong tail in the African spinosaurine described by Ibrahim et al. (2020a), and topologically similar to the femur FSAC-KK 11888. Thus, the morphology of the specimen studied here is more similar to that of FSAC-KK 11888 and other spinosaurids than that of other theropods; on this basis, NHMUK PV R 16433 is interpreted as representing an indeterminate Spinosaurinae .</p></div>	https://treatment.plazi.org/id/03A39E0B9644FFFAFEBAFB99FF2E1C15	Public Domain	No known copyright restrictions apply. See Agosti, D., Egloff, W., 2009. Taxonomic information exchange and copyright: the Plazi approach. BMC Research Notes 2009, 2:53 for further explanation.		Plazi	Mauro;Lacerda;Isasmendi, Erik;Delcourt, Rafael;Marcelo;Fernandes;John;Hutchinson	Mauro, Lacerda, Isasmendi, Erik, Delcourt, Rafael, Marcelo, Fernandes, John, Hutchinson (2024): New theropod dinosaur remains from the ºpper Cretaceous of the Kem Kem Group (Eastern Morocco) clarify spinosaurid morphology. Zoological Journal of the Linnean Society 202 (2): 1-31, DOI: 10.1093/zoolinnean/zlae109, URL: https://doi.org/10.1093/zoolinnean/zlae109
03A39E0B9652FFFAFEB8FCDCFF9B1C8F.text	03A39E0B9652FFFAFEB8FCDCFF9B1C8F.taxon	http://purl.org/dc/dcmitype/Text	http://rs.tdwg.org/ontology/voc/SPMInfoItems#GeneralDescription	text/html	en	Sigilmassasaurus Russell 1996	<div><p>Sigilmassasaurus Russell, 1996</p><p>Type-species: Sigilmassasaurus brevicollis</p></div>	https://treatment.plazi.org/id/03A39E0B9652FFFAFEB8FCDCFF9B1C8F	Public Domain	No known copyright restrictions apply. See Agosti, D., Egloff, W., 2009. Taxonomic information exchange and copyright: the Plazi approach. BMC Research Notes 2009, 2:53 for further explanation.		Plazi	Mauro;Lacerda;Isasmendi, Erik;Delcourt, Rafael;Marcelo;Fernandes;John;Hutchinson	Mauro, Lacerda, Isasmendi, Erik, Delcourt, Rafael, Marcelo, Fernandes, John, Hutchinson (2024): New theropod dinosaur remains from the ºpper Cretaceous of the Kem Kem Group (Eastern Morocco) clarify spinosaurid morphology. Zoological Journal of the Linnean Society 202 (2): 1-31, DOI: 10.1093/zoolinnean/zlae109, URL: https://doi.org/10.1093/zoolinnean/zlae109
03A39E0B9652FFF8FE82FC73F95A1D73.text	03A39E0B9652FFF8FE82FC73F95A1D73.taxon	http://purl.org/dc/dcmitype/Text	http://rs.tdwg.org/ontology/voc/SPMInfoItems#GeneralDescription	text/html	en	Sigilmassasaurus brevicollis Russell 1996	<div><p>Sigilmassasaurus brevicollis</p><p>(Fig. 13)</p><p>Referred specimen</p><p>A completely preserved posterior cervical vertebra (NHMUK PV R 38358).</p><p>Morphological description</p><p>NHMUK PV R 38358 is an almost complete cervical vertebra, with only part of the neural spine missing. Based on the axial sequence proposed by Evers et al. (2015), NHMUK PV R 38358 can be identified as a C9 (Fig. 13).</p><p>The vertebral centrum is c. 125 mm long anteroposteriorly, c. 140 mm wide, and c. 95 mm high. It is strongly opisthocoelous and the anterior articular surface is surrounded by a rim (Fig. 13A, B). NHMUK PV R 38358 has a vertebral centrum that is shorter than it is wide, being 12% wider than it is long, and its dorsal surface is slightly anteroposteriorly shorter than ventral surface. The articular facets are wide, with the anterior and the posterior articular surfaces being 1.4 and 1.6 times wider than they are tall, respectively (Fig. 13). The anterior articular surface is strongly convex and the posterior one is strongly concave. In anterior view, the anterior articular surface is pronouncedly elliptical (Fig. 13A, B) and, in posterior view, the posterior one has a reniform outline (Fig. 13C, D). The ventral surface exhibits a well-developed ventral keel that extends to the anterior and posterior margins of the vertebral centrum, widening slightly transversely towards these edges (Fig. 13G, H). In lateral view, the anterior and posterior parts of the keel project further ventrally, so that it is concave at midlength of the centrum (Fig. 13I–L). Anteriorly, the ventral keel gradually becomes lower and merges into a small, rough, triangular area located immediately anterior to the parapophyses (Fig. 13G, H). Lateral to the ventral keel, a fossa is present on each side of the centrum. These fossae are delimited anterolaterally by the parapophyses and the lamina that connects the parapophyses to a small anterior triangular area, and posterolaterally by a lamina that extends from the parapophyses posteriorly to the posterior articular facet (Fig. 13G–L). No hypapophysis seems to be present in NHMUK PV R 38358.</p><p>The parapophyses are robust, 'button-shaped' structures that are long and project ventrolaterally; located anteroventrally on the lateral surfaces of the centrum (Fig. 13A–D, G–L). The articular surfaces of the parapophyses are concave and oval in outline. From the posterior margin of each of the parapophyses, a rounded ridge extends posteriorly; these are the ridges that laterally delimit the ventral fossae (Fig. 13G, H). Above the parapophyses, a single and large central pneumatic foramen is present on each side of the NHMUK PV R 38358 vertebral centrum and it penetrates the bone anteroventrally (Fig. 13I–L). Each foramen has a different shape: the right one is triangular and anteroposteriorly larger than tall (Fig. 13), whereas the left is oval and taller than its length anteroposteriorly (Fig. 13K, L).</p><p>The neurocentral suture in NHMUK PV R 38358 is still visible. The neural canal is large, subrectangular to oval in shape, and transversely wider than tall (Fig. 13A–D). The pleurocentral depressions are located ventral to this suture, anterodorsally located on the centrum.</p><p>The transverse processes are very large and project ventrolaterally at an angle of c. 40° to the lateral surface (Fig. 13A–D). In anterior view, the transverse processes are quite straight dorsoventrally (their dorsal surfaces are distinctly flat), but towards the diapophyses they curve slightly more ventrally. In dorsal view, the posterior surfaces of the transverse processes are straight and laterally directed (Fig. 13E, F). The anterior margin is gently concave in dorsal view and expands anteriorly near the diapophyses. The diapophyses face ventrolaterally and are convex and triangular in outline. Both prezygodiapophyseal and postzygodiapophyseal laminae are present. These laminae are robust, rounded, developed in a similar way, and are less marked near the diapophysis. The prezygodiapophyseal lamina extends laterally from the anterolateral margin of the prezygapophyses along the anterodorsal margin of the transverse process, reaching the anterodorsal margin of the diapophyses. The postzygodiapophyseal lamina runs from the anterolateral margin of the postzygapophyses toward the diapophyses, forming the posterodorsal edge of the transverse process. The centroprezygapophyseal lamina is directed anterolaterally and merges with the prezygodiapophyseal lamina on the anterior surface of the base of the prezygapophyses, with the lamina being less marked, but more robust and rounded than the prezygodiapophyseal lamina.</p><p>The centrodiapophyseal lamina is present on the ventral surface of the transverse process. This is very prominent and it seems not to be bifurcated into the anterior centrodiapophyseal and posterior centrodiapophyseal laminae. However, there is another lamina that runs from the ventral surface of the transverse process toward the posterodorsal margin of the centrum [probably the posterior centrodiapophyseal lamina according to Evers et al. (2015)]. The centrodiapophyseal lamina runs from the diapophysis, extending from the ventral surface of the transverse process toward the lateral side of the centrum, resulting in a 'T-shaped' cross-section of the transverse process.</p><p>Regarding the pneumaticity of the vertebra, the prezygocentrodiapophyseal fossa is open ventrolaterally and extends throughout the anteroventral part of the transverse process (Fig.13A,B).Thisfossaisdelimitedbythecentroprezygapophyseal lamina anteriorly, the centrodiapophyseal lamina posteriorly, and the prezygodiapophyseal lamina dorsally. The prezygocentrodiapophyseal fossa presents a 'slit-shaped' foramen, quite high, narrow, and oval, which penetrates the pedicle of the prezygapophysis. The postzygocentrodiapophyseal fossa is located on the posterior surface of the transverse process, being larger than the prezygocentrodiapophyseal fossa (Fig. 13C, D). This fossa is delimited anteriorly by the centrodiapophyseal lamina, posteriorly by the centropostzygapophyseal lamina and dorsally by the postzygodiapophyseal lamina. Moreover, it exhibits another pneumatic foramen, which penetrates posterolaterally into the transverse process, and is also more circular than that of the prezygocentrodiapophyseal fossa (Fig. 13C, D).</p><p>The prezygapophyses are large and considerably separated, projecting more laterally than the vertebral centrum itself, being projected mainly dorsolaterally and slightly anteriorly (Fig. 13A–F). The prezygapophyses are located dorsal to the transverse process in the anterior half of the transverse processes and their anterior margin is placed slightly anterior to the anterior margin of the transverse process (Fig. 13I–L). The prezygapophyseal articular facets of the prezygapophyses are oval to subcircular in dorsal view, being wider lateromedially. They are dorsomedially facing, with an angle of 125° between both facets. These facets are slightly convex but almost flat, and slightly posteriorly inclined. There is no intraprezygapophyseal lamina between the prezygapophyses.</p><p>The postzygapophyses are also large, but more compact compared to the prezygapophyses and without an epipophysis. They extend posterodorsally beyond the posterior margin of the centrum, with half of the postzygapophyses posterior to the posterior margin of the centrum (Fig. 13I–L). The postzygapophyseal facets are concave, with an inverted 'teardropshape', posteroventrally facing, and laterally oriented. They are connected anteromedially by the spinopostzygapophyseal laminae, but they are not connected ventromedially because there is no intrapostzygapophyseal lamina. Nevertheless, from the posteromedial margin of the postzygapophyses, two narrow spinopostzygapophyseal laminae extend ventrally, reaching the dorsolateraledgeoftheneuralspine.Thespinopostzygapophyseal laminae are well developed, robust, and rounded, being delimited dorsolaterally by the spinopostzygapophyseal fossa, which is quite triangular and open ventrally. This fossa is laterally delimited by postzygapophyses and ventromedially oriented laminae.</p><p>The neural spine is not completely preserved, but its base suggests that it was a 'spike-like' process that projected posterodorsally (Fig. 13A–F). The cross-section of its base is subcircular in dorsal view, and is connected to the prezygapophyses by the spinoprezygapophyseal laminae. These laminae are poorly developed in comparison to the other laminae present in the vertebra; besides that, they are rounded and extend from the anterior surface of the neural spine to the posteromedial margin of the prezygapophyses, forming an inverted 'V-shaped' structure (Fig. 13A–F). A prespinal lamina (sensu Evers et al. 2015) is also present on the anterior surface of the neural arch, situated between the spinoprezygapophyseal laminae. This lamina is low, dorsoventrally oriented, and projects slightly ventrally into the neural canal. This lamina, together with the spinoprezygapophyseal laminae, delimits a shallow triangular depression on the anterior surface of the neural arch.</p><p>Morphological comparisons</p><p>Specimen NHMUK PV R 38358 shares with many earlybranching tetanuran theropods the single pneumatic foramen on the lateral surface of the vertebral centrum (Carrano et al. 2012). Furthermore, this posterior cervical vertebra shares with Megalosauroidea taxa the bordered (or rimmed) anterior articular surface of the centrum (Carrano et al. 2012, Evers et al. 2015, Malafaia et al. 2020, Barker et al. 2021). However, this trait may be also present in Allosaurus (Rauhut and Pol 2019) . In the specimen studied here, the parapophyses are enlarged and exhibit a strongly concave facet, similar with the spinosaurids Baryonyx, Ichthyovenator, Sigilmassasaurus, and Suchomimus (Allain et al. 2012, Evers et al. 2015). Another feature shared between NHMUK PV R 38358 and spinosaurids is the presence of a ventral keel on the vertebral centrum, with the anterior end projecting anteriorly (a synapomorphy for Spinosauridae — Schade et al. 2023). The neural arch of NHMUK PV R 38358 lacks epipophyses, similar to the posterior cervicals of Baryonyx [considering the arrangement proposed by Evers et al. (2015)], and Sigilmassasaurus (Russell 1996, McFeeters et al. 2013, Evers et al. 2015). The neural spine is inferred to be 'spike-like' in NHMUK PV R 38358, a feature also shared with Baryonyx and Sigilmassasaurus (Evers et al. 2015) .</p><p>The vertebral centrum of NHMUK PV R 38358 is very wide, with the anterior articular surface being 1.4 times wider than it is tall, similar to the condition observed in Ichthyovenator and Sigilmassasaurus, in the latter being more than 1.5 times larger than high (Russell 1996, McFeeters et al. 2013, Evers et al. 2015). Specimen NHMUK PV R 38358 also shares with Ichthyovenator and Sigilmassasaurus the lack of intraprezygapophyseal and intrapostzygapophyseal laminae (Allain et al. 2012, Evers et al. 2015). The posterior cervical vertebra (C9) studied here further resembles the posterior cervicals of Sigilmassasaurus brevicollis by having large, transverse processes that exhibit pneumatic foramina deep beneath their base (McFeeters et al. 2013, Evers et al. 2015).</p><p>If the emended diagnosis of Sigilmassasaurus brevicollis proposed by Evers et al. (2015) is considered, specimen NHMUK PV R 38358 shares with the former the reduced lamination of the neural arch with the centrodiapophyseal laminae not being divided into anterior and posterior centrodiapophyseal laminae [autapomorphy proposed by Evers et al. (2015)]. Nevertheless, the specimen described here lacks the anterior tubercle that is present on the anterior articular surfaces of the posterior cervical and anterior dorsal vertebrae of Sigilmassasaurus brevicollis (McFeeters et al. 2013, Evers et al. 2015), but this feature is more subtle in the BSPG 2011 I 115 cervical vertebra (Evers et al. 2015) and also absent in ROM 65537 (McFeeters et al. 2013). The posterior vertebra NHMUK PV R 38358 can be safely assigned to Spinosauridae, furthermore, due to the above-mentioned features shared with Sigilmassasaurus brevicollis, we here assign this specimen to the same taxon.</p></div>	https://treatment.plazi.org/id/03A39E0B9652FFF8FE82FC73F95A1D73	Public Domain	No known copyright restrictions apply. See Agosti, D., Egloff, W., 2009. Taxonomic information exchange and copyright: the Plazi approach. BMC Research Notes 2009, 2:53 for further explanation.		Plazi	Mauro;Lacerda;Isasmendi, Erik;Delcourt, Rafael;Marcelo;Fernandes;John;Hutchinson	Mauro, Lacerda, Isasmendi, Erik, Delcourt, Rafael, Marcelo, Fernandes, John, Hutchinson (2024): New theropod dinosaur remains from the ºpper Cretaceous of the Kem Kem Group (Eastern Morocco) clarify spinosaurid morphology. Zoological Journal of the Linnean Society 202 (2): 1-31, DOI: 10.1093/zoolinnean/zlae109, URL: https://doi.org/10.1093/zoolinnean/zlae109
03A39E0B9650FFF4FC34FD57FF0F1AA7.text	03A39E0B9650FFF4FC34FD57FF0F1AA7.taxon	http://purl.org/dc/dcmitype/Text	http://rs.tdwg.org/ontology/voc/SPMInfoItems#GeneralDescription	text/html	en	Carcharodontosauridae Stromer 1931	<div><p>Carcharodontosauridae gen. et sp. indet.</p><p>(Fig. 14)</p><p>Referred specimen</p><p>A partial left ischium (NHMUK PV R 16437).</p><p>Morphological description</p><p>The left ischium NHMUK PV R 16437 has its proximal portion preserved, missing the distalmost portion of the shaft and the ventralmost portion of the obturator process (Fig. 14). The iliac and pubic peduncles are separated by a concave acetabular rim that is shallow, wide in anterior view with a middle constriction, giving it an 'hourglass-shape' in proximal view (Fig. 14). The peduncles are subequal in size. In the proximal region, the iliac articular surface is triangular and deeply concave suggesting a peg-and-socket articulation (Fig. 14). In the anteriormost region of the iliac peduncle, anterior to the articular surface with the ilium, the ischium is thick mediolaterally, forming the ischial antitrochanter, which is parallelogram-shaped in anterior view. Although the antitrochanter is thick, it is not well projected, being a reduced ridge. In the posterior part of the iliac joint, there is a posterior flange that rises, and it is broken in the dorsalmost portion. The pubic peduncle is subtriangular and medially concave with the articular surface being laterally oriented.</p><p>The lateral and medial surfaces of the ischium are concave between the peduncles, with the concavity displaced dorsally in the lateral side and ventrally in the medial side. Posteroventral to the flange in the iliac peduncle, a deep and rugose sulcus, homologous to the ischial tuberosity (e.g. Hutchinson 2001b, Brusatte et al. 2008, Cuesta et al. 2018), runs in the lateral surface becoming shallower posteriorly (Fig. 14). The most proximal part of this sulcus, somewhat elliptical in shape, represents the osteological correlate of the origin of the muscle flexor tibialis internus 3 (FTI3), which is delimited slightly more distally. The most distal part of the sulcus in the posterodorsal region of the ischial shaft, which is more linear, less elliptical than the proximal part, becomes shallower distally, and represents the osteological correlate of the origin of the muscle adductor femoris 2 (ADD2) (Fig. 14A, B). Both of these muscle origins are level II inferences by Witmer's (1995) systematization and are topologically compatible with other theropods (e.g. ceratosaurs— Cerroni et al. 2024, early tetanurans— Lacerda et al. 2024, and derived coelurosaurs— Carrano and Hutchinson 2002).</p><p>In the preserved portion of the ischium, the shaft is dorsoventrally flattened, giving it a subrectangular shape, lacking the distalmost part (Fig. 14). Ventral to the shaft and posterior to the pubic peduncle, the obturator process is separated from the pubic peduncle by a shallow and anteroposteriorly notch. Posterior to this notch, the obturator process is twisted medially from the pubic peduncle and broken in its ventralmost portion (however, a notch ventral to obturator process can be noted) and seems to be confluent with the shaft (Fig. 14). Although the ischial shaft is not completely preserved, the preserved part is straight, suggesting that the orientation of the main axis of the ischium was straight in NHMUK PV R 16437.</p><p>Morphological comparisons</p><p>The overall shape of the NHMUK PV R 16437 partial ischium resembles that of carcharodontosaurid theropods rather than of other dinosaurs. The acetabular rim is shallow in lateral view with a weak 'U-shape' due to the ventral position of the pubic peduncle, as seen in allosauroids (sensu Rauhut and Pol 2019), including the carcharodontosaurids Acrocanthosaurus, Concavenator, Giganotosaurus, Mapusaurus, and Neovenator . In some neovenatorids, such as Siats, the acetabular rim is shallow, not forming a 'U-shaped' border, being straighter than in other allosauroids (Zanno and Makovicky 2013). The 'U-shape' of the acetabular rim is more pronounced in megalosauroids such as Piatnitzkysaurus and also in the spinosaurids Baryonyx, Ichthyovenator, Vallibonavenatrix, FSAC-KK 11888, and possibly in Suchomimus due to a dorsal projection of the pubic peduncle (Allain et al. 2012, Malafaia et al. 2020, Sereno et al. 2022, Lacerda et al. 2024). This condition differs from that observed in NHMUK PV R 16437.</p><p>The ilioischiatic articulation of NHMUK PV R 16437 has a deep peg-and-socket (or ball-and-socket) configuration, as seen in Acrocanthosaurus, Concavenator, Giganotosaurus, Mapusaurus, and Siats (Stovall and Langston 1950, Coria and Currie 2006, Carrano et al. 2012, Zanno and Makovicky 2013, Cuesta et al. 2018, Rauhut and Pol 2019, Lacerda et al. 2023). In other theropods, this articulation has a concave plane configuration (e.g. Carrano et al. 2012, Lacerda et al. 2023, Isasmendi et al. 2024).</p><p>The ischial antitrochanter is a well-developed, notch-shaped structure in coelophysoids, some ceratosaurs, the early tetanuran Sinosaurus, the spinosaurid Ichthyovenator, and the neovenatorid Siats; it is a reduced notch in the ischium of other theropods (e.g. Allain et al. 2012, Carrano et al. 2012, Zanno and Makovicky 2013, Cuesta et al. 2018, Lacerda et al. 2023). In NHMUK PV R 16437, although the ischial antitrochanter is a reduced ridge, it represents a thick structure at its base, and this is more comparable with forms such as Acrocanthosaurus, Giganotosaurus, and Sinraptor, than Siats and other contemporary theropods such as spinosaurids.</p><p>The presence of the sulcus in the dorsolateral shaft of the ischium, which is posterior to the flange, is similar in NHMUK PV R 16437 and other theropods (e.g. Zanno and Makovicky 2013, Cuesta et al. 2018); however, in the ischium described here it is deeper, similar to carcharodontosaurids and neovenatorids rather than spinosaurids. Consequently, the osteological correlates of the origins of the muscles FTI3 (proximal) and ADD2 (distal) in NHMUK PV R 16437 are deeper than those noted in other non-carcharodontosaurid theropods (e.g. Carrano and Hutchinson 2002, Cerroni et al. 2024, Lacerda et al. 2024).</p><p>The obturator process of NHMUK PV R 16437 presents a notch that is shared with other theropods such as the ceratosaur Ceratosaurus, piatnitzkysaurids, spinosaurids (except Ichthyovenator), and allosauroids such as Acrocanthosaurus, Allosaurus, Giganotosaurus, and Sinraptor (Stovall and Langston 1950, Carrano et al. 2012, Lacerda et al. 2023). However, NHMUK PV R 16437 has a lamina, part of the dorsalmost portion of the obturator process, which is immediately ventral to the pubic peduncle. This feature is also shared with the metriacanthosaurid Sinraptor and the carcharodontosaurid Giganotosaurus .</p><p>Furthermore, if the ischial shaft in NHMUK PV R 16437 is indeed straight, as the preserved portion suggests, this would be another feature shared between this individual and carcharodontosaurids (also seem in some allosauroids— Madsen 1976), but not with some early diverging tetanurans and also metriacanthosaurids. Thus, based on the set of characteristics shared between NHMUK PV R 16437 and carcharodontosaurid theropods, as well as the differences between this ischium and those of spinosaurids, we assign this material to an indeterminate Carcharodontosauridae .</p><p>Brief survey of theropod dinosaurs from the Kem Kem Group</p><p>Several records of theropod dinosaurs are known from the North Africa, especially those that derive from Middle Cretaceous rocks from the Kem Kem Group region, therefore, suggesting a high theropod diversity. However, much of the diversity known for the region is represented by shed teeth, which are more likely to be fossilized as they are more resistant to weathering and taphonomic alterations (Benyoucef et al. 2015, Hendrickx et al. 2024). There is also a good and broad ichnological record that helps confirm faunal occurrences and serves as a proxy for more reliable palaeoenvironmental reconstructions (Belvedere et al. 2013, Ibrahim et al. 2014b). Regarding skeletal remains, records are scarcer (Hendrickx et al. 2024); however, they are still of broad relevance for understanding both biogeographical and evolutionary issues in different clades. Below, a brief nonexhaustive survey of the body fossil occurrences is presented, as well as integration with our findings and their relevance to current knowledge.</p><p>Abelisauridae</p><p>The fossil record of abelisaurids is one of the most abundant in the Kem Kem Group, being less abundant only than spinosaurids. Russell (1996) described several bone fragments, including two partial right dentaries, in addition to two partial cervical vertebrae that were attributed to an undetermined theropod, but recently were assigned to abelisaurids (Souza-Júnior et al. 2023). Although the locality from which these materials were derived is unknown, they were probably recovered from the Cenomanian of southern Morocco (Souza-Júnior et al. 2023). A partially preserved maxilla that probably comes from Erfoud was described by Mahler (2005). Novas et al. (2005) related a ungual pedal to Abelisauroidea from the Tafilalt region. A partially preserved left maxilla that derives from a region near Taouz was described by Porchetti et al. (2011). The proximal part of a femur was described by Chiarenza and Cau (2016). In addition to these, an axis vertebra described by Smyth et al. (2020a) was also attributed to Abelisauridae .</p><p>Noasauridae</p><p>A postcranial skeleton that is relatively well preserved, including hindlimbs and partial forelimbs as well as part of the tail, was erected as Deltradomeus agilis by Sereno et al. (1996). However, the assignment of this taxon to noasaurid theropods occurred only in later phylogenetic analyses (e.g. Sereno et al. 2004) and the phylogenetic position of this taxon remains under debate. Evans et al. (2015) described a well-preserved femur from south-east Taouz, which could possibly belong to Deltradomeus. Another noasaurid occurrence was presented by Smyth et al. (2020a), based on an isolated cervical vertebra.</p><p>Spinosauridae</p><p>This group is one of the most representative in the Kem Kem Group fossil record; however, the nature of the material is isolated or semi-articulated fossils. Two partial dentaries from south-eastern Morocco were described by Buffetaut (1989), being both referred to Spinosaurus cf. Sp. aegyptiacus . Based on a set of isolated bones, including cervical vertebrae, dorsal neural arch, and a fragmentary dentary recovered from Morocco, Russell (1996) erected the name Spinosaurus maroccanus; later on, Taquet and Russell (1998) referred a rostrum and some axial elements recovered from the Algerian portion of the Sahara Desert to this species. However, this species is frequently regarded as a nomen dubium (e.g. Carrano et al. 2012), a junior synonym of Spinosaurus aegyptiacus (e.g. Ibrahim et al. 2014a), or Spinosaurinae indet. (e.g. Lacerda et al. 2022). Milner (2003) presented a rostrum and a relatively well-preserved dentary from Morocco in a brief note; both of them were referred to Sp. aegyptiacus . This rostrum is redescribed in detail in this work (NHMUK PV R 16420) and considered as an indeterminate Spinosaurinae, agreeing with the current debate (e.g. Sales and Schultz 2017, Lacerda et al. 2022). Another rostrum, which represents the most complete and well-preserved known to date, and a pair of nasals from Morocco were described by Dal Sasso et al. (2005), with both referred to Spinosauru s cf. Sp. aegyptiacus . Lately, some studies (e.g. Sales and Schultz 2017, Lakin and Longrich 2019, Lacerda et al. 2022) have argued that the most reliable identification of that rostrum is Spinosaurinae indet. due to a lack of overlap between these specimens and the Sp. aegyptiacus holotype. Based on several skeletal elements of a subadult individual, Ibrahim et al. (2014a) proposed a neotype to Sp. aegyptiacus, and later on, new skeletal remains possibly of the same individual represented by a robust tail were also recovered from Morocco (Ibrahim et al. 2020a). Two morphotypes of spinosaurids, Sp. aegyptiacus and cf. Sigilmassasaurus brevicollis, were identified by Hendrickx et al. (2016) based on six quadrates that probably derived from Morocco. A tiny pedal ungual recovered from between the villages of Taouz and Begaa was presented by Maganuco and Dal Sasso (2018). Arden et al. (2019) also described two morphotypes of spinosaurids from the Kem Kem Group: one referred to Sp. cf. Sp. aegyptiacus based on frontals and a frontoparietal, and a skull roof referred to Sigilmassasaurus cf. Si. brevicollis . Based on cranial remains and isolated axial elements, Lakin and Longrich (2019) also described fossils referred to Si. brevicollis and Sp. cf. Sp. aegyptiacus . Besides the previous mentions, other studies also described axial elements of Sigilmassasaurus brevicollis recovered from the same region (e.g. Russell 1996, McFeeters et al. 2013, Evers et al. 2015).</p><p>In this contribution we have described several specimens that expand the knowledge of occurrences of spinosaurids in the Kem Kem Group. Among these, we have presented a cervical vertebra referred to Si. brevicollis (NHMUK PV R 38358), in addition to nine specimens (NHMUK PV R 16391, 16422, 16423, 16424, 16426, 16430, 16431, 16433, 16438) that we conservatively classified as indeterminate Spinosaurinae . Our findings, combined with materials mentioned above, contribute to general aspects of the occurrence of spinosaurids in the region, making them one of the most abundant groups of theropod dinosaurs in the Kem Kem Group.</p><p>Carcharodontosauridae</p><p>At least two species are known from the Kem Kem Group (although some studies consider only one species— Ibrahim et al. 2020a). A nearly complete skull and some vertebral elements of Carcharodontosaurus saharicus were described by Sereno et al. (1996), in which a neotype was designated, recovered from the Douira Formation, south-eastern Morocco (Sereno et al. 1996, Ibrahim et al. 2020a). An isolated and fragmentary portion of a dentary [originally referred to an abelisaurid by Russell (1996) —see Ibrahim et al. (2020a)] can also be assigned to C. saharicus . A second carcharodontosaurid species, Sauroniops pachytholus, was erected by Cau et al. (2013) based on an almost complete frontal. Later on, Paterna and Cau (2023) also referred additional materials (a partial maxilla and a jugal) to Carcharodontosauridae, discussing the status of both— Carcharodontosaurus and Sauroniops . A probably indeterminate carcharodontosaurids manual ungual [Ibrahim et al. (2020a), originally described as Theropoda indet. by Russell (1996)] adds to the fossil record of this clade.</p><p>Here we also provided the description of an isolated ischium we identify as an indeterminate carcharodontosaurid (NHMUK PV R 16437), adding to the fossil record of this clade from the Kem Kem Group.</p><p>Taxonomic attributions of NHMUK PV R 16420 and MSNM V4047 snouts</p><p>There is a consensus in the literature considering both well-preserved rostra from the Kem Kem Group—NHMUK PV R 16420 and MSNM V4047—as Spinosaurinae theropods (e.g. Milner 2003, Dal Sasso et al. 2005, Lakin and Longrich 2019, Lacerda et al. 2022). However, the referral of both to Spinosaurus aegyptiacus (e.g. Milner 2003, Dal Sasso et al. 2005, Ibrahim et al. 2014a) is not possible to corroborate, at least yet, due to the lack of overlap among these and the Sp. aegyptiacus holotype (Evers et al. 2015, Sales and Schultz 2017, Lakin and Longrich 2019, Lacerda et al. 2022).</p><p>Some studies (Lakin and Longrich 2019, Lacerda et al. 2022) noted some differences between both specimens. Lakin and Longrich (2019) considered NHMUK PV R 16420 as having a deeply concave dorsal profile and a curved premaxillary ventral profile, a straighter maxillary tooth row, larger external nares, and distinct outline of the premaxilla when compared with MSNM V4047. Although Lakin and Longrich (2019) did not consider these, necessarily, as two distinct taxa, they considered this set of features enough to designate two distinct morphotypes. However, Lacerda et al. (2022) quantitatively showed a high degree of compression/erosion in NHMUK PV R 16420, preventing any morphological differentiation from MSNM V4047.</p><p>As previous noted, the only major differences between the two rostra are the number of premaxillary teeth and the pattern of the intramaxillary suture anteriorly. However, these features seem to have no systematic significance. Based on our detailed redescription, we reject the possibility of two distinct taxa based on the NHMUK PV R 16420 and MSNM V4047 rostra. Excluding some features that are taphonomic artefacts, both snouts have virtually the same shape and probably represent the same taxon.</p><p>One or two Spinosaurinae taxa in the Kem Kem Group?</p><p>There is a prolific debate regarding the presence of one or more spinosaurid species in the Cenomanian of the Kem Kem Group. Several studies consider that Spinosaurus aegyptiacus is the only well-established species from this region and Sigilmassasaurus brevicollis is not supported due to the lack of autapomorphies (e.g. Ibrahim et al. 2014a, 2020a, b, Maganuco and Dal Sasso 2018, Smyth et al. 2020b). Meanwhile, other studies have argued the plausibility of two (contemporary?) species, considering Si. brevicollis a valid taxon (McFeeters et al. 2013, Evers et al. 2015, Hendrickx et al. 2016, Hone and Holtz 2017, Arden et al. 2019). However, there is a degree of plausibility to both propositions, as we explain here. On one hand, those studies that consider Sp. aegyptiacus the only species argue that any morphological variation noted in multiple specimens is due to ontogeny, individual variations or sexual dimorphism (e.g. Ibrahim et al. 2014b, Smyth et al. 2020b). Nevertheless, there are no studies showing the main ontogenetic stages of Sp. aegyptiacus, for example, leaving these propositions as speculative. On the other hand, several 'diagnostic' or 'autapomorphic' features are described in the literature, and several studies discuss distinct morphotypes of spinosaurids from the Kem Kem Group (e.g. Chiarenza and Cau 2016, Hendrickx et al. 2016, Arden et al. 2019; McFeeters 2021).</p><p>This study also supports different morphotypes from this geological unit, for example, based on the proximal portion of the femur described here (NHMUK PV R 16433) when compared with the proximal femur attributed to Sp. aegyptiacus (FSAC-KK 11888; Ibrahim et al. 2014 a, Sereno et al. 2022). Besides that, several fossil specimens that have been synonymized with, and referred to, Sp. aegyptiacus do not even overlap with the lost holotype specimen (Evers et al. 2015, Sales and Schultz 2017, Lakin and Longrich 2019, Lacerda et al. 2022), and thus remain difficult to corroborate. Although this study does not intend to resolve these issues, we considered the extended diagnosis provided by Evers et al. (2015), and thus the potential validity of Si. brevicollis . Meanwhile, we did not relate other materials to Sp. aegyptiacus due to lack of overlap and the possibility of a proper morphological comparison. Our study highlights the complexity of the analysis of spinosaurid diversity from the Kem Kem Group, and describes some fossil remains in detail, instead of only mentioning the occurrences of fossils, lacking proper descriptions and anatomical comparisons. We urge a detailed review of all spinosaurid materials from the Kem Kem Group, and also suggest caution in considering either a single taxon of spinosaurine with a large amount of morphological variation, or two taxa lacking proper diagnosis; thus, encouraging more detailed/rigorous studies. Resolution of this issue probably depends upon new discoveries of specimens.</p></div>	https://treatment.plazi.org/id/03A39E0B9650FFF4FC34FD57FF0F1AA7	Public Domain	No known copyright restrictions apply. See Agosti, D., Egloff, W., 2009. Taxonomic information exchange and copyright: the Plazi approach. BMC Research Notes 2009, 2:53 for further explanation.		Plazi	Mauro;Lacerda;Isasmendi, Erik;Delcourt, Rafael;Marcelo;Fernandes;John;Hutchinson	Mauro, Lacerda, Isasmendi, Erik, Delcourt, Rafael, Marcelo, Fernandes, John, Hutchinson (2024): New theropod dinosaur remains from the ºpper Cretaceous of the Kem Kem Group (Eastern Morocco) clarify spinosaurid morphology. Zoological Journal of the Linnean Society 202 (2): 1-31, DOI: 10.1093/zoolinnean/zlae109, URL: https://doi.org/10.1093/zoolinnean/zlae109
