An exceptional neurovascular system in abelisaurid theropod skull: New evidence from Skorpiovenator bustingorryi

Abstract

Abelisaurids were one of the most successful theropod dinosaurs during Cretaceous times. They are featured by numerous derived skull traits, such as heavily ornamented bones, short and tall snout, and a strongly thickened cranial roof. Furthermore, nasals are distinctive on having two distinct nasal patterns: strongly transversely convex and heavily sculptured (e.g., Carnotaurus), and transversely concave, with marked bilateral crests and poorly sculptured surfaces (e.g., Rugops). Independently of the pattern, some abelisaurid nasals (e.g., Rugops) show a distinctive row of large foramina on the dorsal surface, which were in general associated to skin structures (scales). Skorpiovenator bustingorryi is a derived abelisaurid coming from the upper Cretaceous beds of northwestern Patagonia, represented by an almost complete skeleton including a well‐preserved skull. Particularly, the skull of Skorpiovenator shows nasal bones characterized by being transversely concave, rimmed by lateral crests and with a conspicuous row of foramina on the dorsal surface. But more interesting is that the skull roof also exhibits a row of large foramina that seem to be continuous with the previous nasal foramina. CT scans made on the skull corroborates a novel feature within theropods: the nasal foramina on the external surface are linked to an internal canal that runs across the nasal bones. We compared this feature with CT scans of Carnotaurus and revealed that it also possess an internal system as in Skorpiovenator, but being notably smaller. The symmetry and disposition of the foramina in the nasal and skull roof bones of Skorpiovenator would indicate a neurovascular correlate (i.e., blood vessels and nerves), probably to the lateral nasal and supraorbital vessels and the trigeminal nerve. The biological significance of such neurovascular system can be conjectured from several hypotheses. A possible one involves an enhanced blood volume in these bones linked to a zone of thermal exchange, which may help avoid overheat of encephalic tissues. Another plausible hypothesis takes into account the presence of display skin structures in which blood volume nourished the mineralized skin, which would have a role in intraspecific communication. However, other more speculative explanations should not be discarded such as a correlation with integumentary sensory organs.

Skorpiovenator bustingorry is a abelisaurid theropod featured by a row of neurovascular foramina in the nasal bones. Further, it has distinctive foramina on the skull roof. The nasal foramina show an internal canal that links each one, which probably represent correlates of blood vessels and nerves. The possible implications of this neurovascular system include thermoregulation and nourishing of display structures, but other speculative hypotheses such as sensory functions are not discarded.

1. INTRODUCTION

Abelisaurid ceratosaurians were a successful theropod group that inhabited Gondwanan landmasses, as well as southern Europe (Bonaparte, 1991; Novas et al., 2013; Tortosa et al., 2013). Among its main features, the skull of abelisaurids is characterized by having a short and deep cranium at the level of the snout, antorbital fenestra with reduced antorbital fossa, a strongly encircled orbital cavity due the well‐developed lacrimal and postorbital bones, frontals strongly thickened and ornamented conforming well‐developed cornual structures (e.g., bulges, Abelisaurus comahuensis; horns, Carnotaurus sastrei and Majungasaurus crenatissimus), expanded parietal crest with a tall parietal eminence, as well as an enhanced mineralization of the cranial bones that produces complex patterns of sculpturing (Bonaparte et al., 1990; Bonaparte, 1991; Sampson et al., 1998; Delcourt, 2018).

The snout of abelisaurids such as Abelisaurus, Carnotaurus, and Majungasaurus bear transversally convex nasals with a dorsal surface highly sculptured. A different ornamented pattern is observed on the external surface of the premaxillae and maxillae, consisting of strong vertical wrinkles (Bonaparte et al., 1990; Novas, 1997; Sampson and Witmer, 2007). This sculptured pattern, originally recognized in Abelisaurus (Bonaparte and Novas, 1985), was later documented in all the remaining abelisaurids; thus, isolated maxillae are readily referred as belonging to this theropod clade.

Nonetheless, nasal bones are preserved in a handful of abelisaurids reflecting an interesting grade of nasal variability not only from its shape, but also from other features, such sculpturing and presence of foramina. Abelisaurus, Carnotaurus, and Majungasaurus have nasals transversely convex and extensively sculptured by highly projected rugosities (Bonaparte et al., 1990; Sampson and Witmer, 2007). The presence of foramina piercing the surface of the nasals is another variable trait: Abelisaurus lacks any sort of openings (Bonaparte and Novas, 1985), Carnotaurus shows a dorsal median row of relatively large foramina on each nasal (Carrano and Sampson, 2008), and Majungasaurus has autapomorphical foramina lateroventrally placed (Sampson et al., 1998). The nasals of Rugops primus constituted an exception among abelisaurids in having the surface of the snout transversely concave, poorly ornamented, surrounded by lateral thick crests, and with a lateral row of large foramina. This combination of features in Rugops was considered as autapomorphic of this taxon (Sereno et al., 2004).

Skorpiovenator bustingorryi is a monospecific abelisaurid known from a complete and articulated skeleton (Canale et al., 2009), discovered near Villa El Chocón town (Neuquén Province) from rocks of the Huincul Formation (Late Cenomanian–Early Turonian). Since its original description, most of the skull and post‐cranial skeleton has been entirely prepared. Skorpiovenator preserves both nasal bones, whose show interesting features on the dorsal surface including a lateral row of large foramina, so far unknown in other Patagonian abelisaurids. Although more astonishing is the presence of several large and laterally placed foramina over the skull roof, which seem to be topographically continuous with the lateral foramina of the nasals.

Several studies focused on the skull of living archosaurs (e.g., crocodiles and birds) comprise descriptions of the neurovascular vessels (Midtgärd, 1984; Baumel, 1993; Sedlmayr, 2002; Porter and Witmer, 2016; Porter et al., 2016). But regarding fossil archosaurs, the inference and reconstruction of cranial soft tissues related to the neurovasculature has been increasing recently, although mostly focused to the skull roof and braincase (e.g., Holliday and Gardner, 2012; Bona et al., 2013; Holliday et al., 2019). Within theropods, there are some previous works regarding the neurovascular system of the snout (e.g., Spinosaurus and Neovenator, Ibrahim et al., 2014; Barker et al., 2017) but stands out the extensive study made by Porter and Witmer (2019) about the overall cephalic vasculature in non‐avian theropod dinosaurs. Although little is known about the internal anatomy of theropod nasals and their soft tissue significance, there are some examples of inferences regarding the possible soft tissues filling these bones (e.g., Monolophosaurus, Zhao and Currie, 1993; Guanlong, Xu et al., 2006; Majungasaurus, Sampson and Witmer, 2007; Corythoraptor, Lü et al., 2017).

Here, we carried out an analysis of the morphological variation of the nasals of abelisaurids based on Skorpiovenator, describing its external ornamentation and foramina and their internal structures related with soft tissues. The aim of this study is to discuss the anatomical and functional significance of this variation. Furthermore, a remarkable set of large foramina over the skull roof is described and constitute a novelty among theropods.

Institutional abbreviations: MACN‐Pv, Museo Argentino de Ciencias Naturales “Bernardino Rivadavia”, Paleontología de Vertebrados, Buenos Aires, Argentina; MCF PVPH, Museo Carmen Funes, Plaza Huincul, Neuquén Argentina; MMCh‐PV, Museo Municipal “Ernesto Bachmann”, Villa El Chocón, Neuquén, Argentina; MPCA‐Pv, Colección Paleontología de Vertebrados, Museo Provincial “Carlos Ameghino”, Cipolletti, Río Negro, Argentina; MUCPv, Museo de la Universidad Nacional del Comahue, Neuquén, Neuquén, Argentina; NCSM, North Carolina Museum of Natural Sciences, Raleigh, North Carolina, United States of America. USNM, National Museum of Natural History, Smithsonian Institution, Washington, United States of America.

2. MATERIAL AND METHODS

The skull of Skorpiovenator bustingorryi (MMCh‐PV 48) was CT‐scanned at the Moguillansky Clinic (Neuquén, Argentina) using a medical tomographer GE BrightSpeed Elite and an energy of 140 kv and 153 mAs. The CT scan resulted in a number of 2236 slices of 0.62 mm thickness. Additionally, we used CT scans of Carnotaurus sastrei (MACN‐CH 894) for anatomical comparisons, which was CT‐scanned at the TCba Salguero Diagnostic Center (Buenos Aires, Argentina) in 2010, using a CT 64 Ingenuity Core medical tomographer. The slice thickness was of 0.62 mm and the scan energy parameters were of 119 mA and 120 kV.

In the case of Skorpiovenator, the presence of numerous cracks (caused by roots and erosion) that strongly affected the complete skull, hindered recognition of the total extent of the internal cavities. The virtual three‐dimensional nasals and its internal structure were obtained and visualized using the software 3D Slicer version 4.10 (Fedorov et al., 2012). The resulting files (.stl) were imported to Design Spark Mechanical 2.0 and exported as a 3D.pdf format (Figures S1 and S2). Final illustrations were made using Adobe Photoshop (CS6).

3. RESULTS

3.1. Description and comparisons of nasal bones of Skorpiovenator

Skorpiovenator shows unfused nasals although the internasal suture is somewhat obscured due deformation (Figure 1). Each nasal has a mostly flat surface rimmed laterally by a relatively tall and markedly thick crest that extends anteroposteriorly along most of the dorsolateral margin, being the crest continuous as a bony shelf in the lacrimal and postorbital. Therefore, the transversal profile of the dorsal aspect of the skull is transversally concave throughout most of its extent, a condition clearly accentuated by the bilateral crests. However, the most conspicuous feature over each nasal is a lateral row of large foramina piercing the dorsal surface. These foramina extend anteroposteriorly over the nasals spreading onto the most posterior sector reaching the level of the lacrimal. Each foramen is separated from the others by wide bony struts, and the foramina have a variable diameter between 8 and 12 mm. A surprising fact is that nasal foramina row is continuous with the row of large foramina on the skull roof (Figure 1a,b); these latter foramina are bounded by several bones such frontal, lacrimal, and postorbital bones (see below).

Figure 1.

Skull of Skorpiovenator bustingorryi (MMCh‐PV 48) in dorsal view (a); Close up of the of the skull dorsal aspect showing the dorsal foramina (b); skull of Skorpiovenator in right lateral view (c); and, detail of a foramen with associated neurovascular grooves (d). Interpretive draw in (a) shows the position of the nasals in red. Arrows in (d) indicates anatomical orientations. Abbreviations: af, anterior foramen; aof, antorbital fenestra; d, dentary; exc, excavations; fr, frontals; hum, hummocky rugosities; ins, internasal suture; itf, infratemporal fenestra; lac, lacrimal; mf, middle foramen; mx, maxilla; n, nasal; nfo, nasal foramina; nlc, nasal lateral crest; pa, parietal; pf, posterior foramen; po, postorbital; sq, squamosal; srfo, skull roof foramina; stf, supratemporal fossa. Scale bar equals 10 cm (a‐c) and 1 cm (d).

In order to compare, the size of some selected foramina was measured considering its diameter and the area that occupy relative to the mediolateral width of the nasal (Table 1). Thus, it can be seen that the average diameter of the nasal foramina of Skorpiovenator is similarly sized to those present in Rugops, but notably larger than those foramina in the nasals of Carnotaurus (Table 1). As for example, in Skorpiovenator and Rugops a foramen at the mid‐length region occupies about 15‐20% of the mediolateral width of the nasal, respectively, whereas in Carnotaurus (MACN‐CH 894), the foramen width at the mid‐length of the bone roughly reaches 10% of the nasal mediolateral width (Mauricio Cerroni, pers. obs.).

Table 1.

Measurements of nasal foramina of Skorpiovenator bustingorryi (mediolateral diameter)

Taxa	Skorpiovenator		Rugops		Carnotaurus
Element (side)	Nasal (l)	Nasal (r)	Nasal (l)	Nasal (r)	Nasal (l)	Nasal (r)
AF/NWL	10.5/32	10.5/30.2	5.5/27	4.5/26	4/38	3/42
MF/NWL	8.3/42.8	8.3/52	4.5/30	4.5/28	3.3/40	2.7/39
PF/NWL	12/69.6	9.5/57.9	4/40	4/41	3.2/52	2.3/50

Measurements of Carnotaurus sastrei were made at first hand observation on MACN‐CH 894; measurements of Rugops primus were taken from Sereno et al. (2004: Figure 3).

Measurements in millimeters (mm).

Abbreviations: AF, foramen located in the anterior zone; l, left; MF, foramen located at mid‐length; NWL, nasal width at this level; PF, foramen at the posterior end; r, right.

Regarding nasal sculpturing, Skorpiovenator is not strongly ornamented, it shows hummocky‐like rugosities over the surface, which constitutes a bonycorrelate for epidermal cover by scales (Hieronymus et al., 2009). In this way, Skorpiovenator shows almost the same ornamentation pattern observed in Rugops (Sereno et al., 2004; also see Delcourt, 2018). The ornamentation described above notably differs with the strongly projected rugosities of other abelisaurids, such as Carnotaurus and Abelisaurus. Although the nasal surface of Skorpiovenator was affected by erosion, some foramina have been preserved with associated vascular grooves on the dorsal surface (Figure 1c). Similarly, Rugops have conspicuous grooves leading into each foramen; structures that were argued by Sereno et al. (2004) as correlated to some kind of soft tissue. Vascular grooves associated to sculpturing of the skull roof have been observed in extinct crocodyliforms as well (e.g. Holliday and Gardner, 2012; Bona et al., 2013).

Another outstanding feature of the nasals of Skorpiovenator, revealed by CT scans, is the presence of a large internal longitudinal canal, connected externally through the row of foramina (Figure 2). This canal consists in a solely and anteroposteriorly long duct that spreads along each nasal from near the mid‐length to the posterior end, close to the frontal contact (Figure 2). Each canal is connected to the large foramina on the dorsal surface of the nasals through short but noticeable canals. The internal canal does not form a uniform space and seems to be somewhat partitioned by internal struts that are projected into the cavity. Although the CT scan resolution is not optimal, the internal canal seems to approach with the antorbital cavity through some ventral and narrow projections (Figure 2: “vp”); however, the best preserved zones of the ventral surface of the nasals do not show any sort of large apertures that could indicate a connection between the inner canal and the antorbital cavity, as it is certainly present in the ventral aspect on the nasals of Rugops that show large foramina (Porter, 2015).

Figure 2.

Digital reconstruction of the skull of Skorpiovenator bustingorryi (MMCh‐PV 48) in dorsal view (a); bone is rendered solid. Rendered and isolated nasals (and partially the frontals) in dorsal view (b) showing the exit branches through foramina (red); bone rendered solid. Isolated nasals (and partially frontals) in dorsal (c) and anterolateral (d) views, showing the internal canals linked to the foramina (red); bone is rendered semitransparent. Arrows in (b‐d) indicates anatomical orientations. Abbreviations: a, anterior; af, anterior foramen of skull roof; aof, antorbital fenestra; icln, intercanal of left nasal; icrn, intercanal in right nasal; ins, internasal suture; l, lateral; mf, middle foramen of skull roof; n, nasals; nlc, nasal lateral crest; p, posterior; sc, short canals exiting through dorsal foramina; sn, snout; so, supraoccipital; stf, supratemporal fenestra, vp: ventral projection. Scale bars equal 10 cm.

3.2. Skull roof foramina of Skorpiovenator

Notably, the skull roof of Skorpiovenator shows an outstanding series of three large foramina that appear to be an extension of the foramina row from the nasals (Figure 1a,d). These openings are laterally placed over the dorsal aspect of the skull being bounded by several osseous contacts of the skull roof and circumorbital bones. The most anterior of these foramina is a foramen delimited laterally by the lacrimal, posteriorly by the frontal, and anteriorly by the nasal. It is interesting to note that such anterior foramen is physically connected to the main internal canal coming from the nasal through a short passage. Posterior to the anterior foramen is another much larger foramen, termed here as middle foramen, bounded exclusively by lacrimal and frontal laterally and medially, respectively. Finally, the most posterior foramen is bounded laterally by postorbital and lacrimal and the frontal medially, as occurs in certain abelisaurids (Sereno et al., 2004; Tortosa et al., 2013). These large foramina vary in diameter, as the nasal foramina, but the middle foramen is by far the largest (Table 2). Regrettably, CT scans do not allowed major details about the internal pathway of the middle and posterior foramina due the ventral aspect of the skull roof is severely damaged by erosion. Finally, it is interesting to note some deep excavations located between the each skull roof foramen; these excavations are deep but do not pierce the skull.

Table 2.

Measurements of the skull roof foramina of Skorpiovenator bustingorryi (mediolateral diameter)

Skull roof (side)	Left	Right
AF	8.8	5.2
MF	12	12.4
PF	8.4	7.5

Measurements in millimeters (mm).

Abbreviations: AF, anterior foramen (located between nasal, lacrimal and frontal); MF, middle foramen (bounded by lacrimal and frontal); PF, posterior foramen (bounded by lacrimal, postorbital and frontal; =”frontal fenestra”).

It is interesting to note that Rugops exhibits a large foramen between the lacrimal, postorbital, and frontal; such foramen is continuous with the lateral row of nasal foramina (Sereno et al., 2004). Furthermore, a very similar “triosseal” foramen has been reported in the skull roof of some abelisaurids for which nasals remain unknown (Figure 3); further, such foramen was named as “frontal fenestra” by Tortosa et al. (2013). Abelisaurids with this foramen are Ekrixinatosaurus (Calvo et al., 2004), Arcovenator (Tortosa et al., 2013), and a yet unnamed abelisaurid from Bajo Barreal Formation of Patagonia (Lamanna et al., 2011). However, neither Ekrixinatosaurus nor Arcovenator holotypes preserve nasals, so it is unknown if these abelisaurids had similar foramina pattern on the dorsal aspect as the otherwise evidenced by the much more complete skulls of Skorpiovenator and Rugops.

Figure 3.

Abelisaurid theropods in dorsal view showing skull roof foramina: Skorpiovenator bustingorryi (a) (MMCh‐PV 48), Rugops primus (b) (modified after Sereno et al., 2004), and braincase of Arcovenator escotae (c) (modified after Tortosa et al., 2013). Braincase of Majungasaurus (MACN‐PV 19776, cast) in dorsal view, lacking of foramina (d). Gray areas indicate the presence of “frontal fenestra”. Abbreviations: af, anterior foramen; “frontal fenestra”, frontal fenestra; fr, frontals; mf, middle foramen; n, nasal; nc, nasal contact; po, postorbital; poc, postorbital contact; sn, snout; srfo, skull roof foramina; stf, supratemporal fenestra. Not to scale.

The most exceptional feature of Skorpiovenator, however, is the entire set of foramina, not only the “frontal fenestra” also present in some other abelisaurids (Figure 1b). The array of foramina over the dorsal surface of the skull roof is not reported in any theropod skull up to date. Rugops, Arcovenator, and Ekrixinatosaurus do not show any sort of extra foramina on their skull roofs that could represent homologues to that present in Skorpiovenator (Figure 3). Thereby, these foramina on the skull roof likely represent an autapomorphy of Skorpiovenator and most likely housed neurovascular tissues in life.

4. DISCUSSION

4.1. Character analysis

With the aim to discuss the nasal disparity (Figures 4 and 5) present among ceratosaurs, we discretized the observed morphological variation into the following traits, comparing in detail Skorpiovenator and those ceratosaurs with preserved nasals, except for noasaurids in which preserved nasals are currently unknown (Carrano et al., 2011) and the basal ceratosaur Limusaurus whose nasals are not fully described (Xu et al., 2009).

Figure 4.

Nasal pattern “A”. Skulls of Carnotaurus sastrei (MACN‐CH 894), Abelisaurus comahuensis (modified after Bonaparte and Novas, 1985), and Majungasaurus crenatissimus (modified after Sampson et al., 1998) in left lateral (a, c, e) and dorsal (b, d, f) views. Nasals are highlighted for better observation (a,b). Dotted lines indicate missing bone (c,d). Gray areas indicate limits of nasal bones. Abbreviations: bc, braincase; fn, fused nasals; fr, frontals; ins, internasal suture; nfo, nasal foramina; ntc, nasal lateral crest; o, orbit; prg, projecting rugosities; sn, snout; spe, subequal posterior end; stf, supratemporal fenestra. Not to scale.

Figure 5.

Nasal pattern “B”. Skulls of Skorpiovenator bustingorryi (MMCh‐PV 48) in right lateral (a) and dorsal (b) views, and Rugops primus (modified and reversed after Sereno et al., 2004) in right lateral (c) and dorsal (d) views. Nasals are highlighted for better observation (a,b). Dotted lines indicate missing bones. Gray areas indicate limit of nasal bones. Abbreviations: aof, antorbital fenestra; epe, expanded posterior end; fr, frontal; ins, internasal suture; nfo, nasal foramina; nlc, nasal lateral crest; o, orbit; sn, snout; stf, supratemporal fenestra. Not to scale.

4.1.1. Transverse section

The plesiomorphic condition of nasals present in most theropods is being transversally convex or rounded throughout the entire length, giving the snout a vaulted roof (Rauhut, 2003). Among ceratosaurians (Figure 4), this condition is moderately present in Ceratosaurus (USNM 4735; Madsen and Welles, 2000), but well‐developed in Abelisaurus, Carnotaurus, and Majungasaurus (MACN‐CH 894; MPCA‐Pv 11908; Sampson and Witmer, 2007), although the nasals in Majungasaurus are extremely inflated probably as a consequence of the extensive nasal sinus that invaded internally (Sampson and Witmer, 2007). By contrast, the nasals of Skorpiovenator and Rugops (Sereno et al., 2004) are transversally concave due the raised crests that delimit the dorsolateral aspect of each bone (Figure 5); a condition similarly developed in basal tetanurans and extremely developed in some coelophysids (Rauhut, 2003). Although the transversal profile of the nasals of Limusaurus is not described, it is known that these bones bear a lateral shelf (Xu et al., 2009).

4.1.2. Overall shape

Viewed dorsally, the plesiomorphic condition of nasals in most archosaurs as well in basal theropods is to be posteriorly expanded (Rauhut, 2003). However, Ceratosaurus (USNM 4735), Carnotaurus (MACN‐CH 894), Majungasaurus (Sampson and Witmer, 2007), and seemingly Abelisaurus (MPCA‐Pv 11908) show nasals with a subequal width throughout their length (Figure 4), even though being relatively short and wide. Conversely, Skorpiovenator and Rugops (Sereno et al., 2004) have nasals with a narrow anterior end and a posterior end posteriorly expanded (Figure 5), resembling the plesiomorphic condition in theropods. In the lateral aspect, almost all abelisaurids show nasals with a nearly straight profile, with the exception of Majungasaurus which have a distinct dorsal hump resulting in a faintly curvature of the dorsal margin (Sampson and Witmer, 2007).

4.1.3. Nasal fusion

Most theropods retain the condition of unfused nasals (Carrano and Sampson, 2008). Within ceratosaurs, this is true for Ceratosaurus (USNM 4735; Madsen and Welles, 2000) and Carnotaurus (MACN‐CH 894; Bonaparte et al., 1990); however, nasals are fused into a single element in Abelisaurus (MPCA‐Pv 11098; Bonaparte, 1985, 1985) and Majungasaurus (Sampson and Witmer, 2007). Rugops have both nasals fused on its most anterior portion; however, on the distal two‐thirds the internasal suture is observable (Sereno et al., 2004: Figure 3), whereas in Skorpiovenator both nasals are completely unfused (Figures 1 and 5).

4.1.4. Sculpturing

Theropods retain the primitive condition of having nasals with smooth or slightly textured dorsal surfaces (Carrano and Sampson, 2008). Among ceratosaurs, the nasals of Ceratosaurus are mainly smooth, except for the nasal horn (USNM 4735; Madsen and Welles, 2000). However, the nasal sculpturing of Abelisaurus, Carnotaurus, and Majungasaurus is featured on being highly rugose (Figure 5) (Sampson and Witmer, 2007), as well as having coarse pitting and grooving surfaces (e.g., Carnotaurus, Majungasaurus) even developing prominent bone lobules as in Abelisaurus (Delcourt, 2018). The nasals of Skorpiovenator and Rugops are certainly ornamented with hummocky rugosities (Figure 5) and clearly differ with the highly sculptured nasals of Abelisaurus or Carnotaurus (see Hieronymus, 2009 and Delcourt, 2018 for detailed discussions of cranial sculpturing).

4.1.5. Dorsal foramina

Most theropods including ceratosaurs lack nasal bones bearing apertures or large foramina over the dorsal surface (Carrano and Sampson, 2008). The known specimens of Ceratosaurus preserve nasals with the characteristic medial eminence, although devoid of dorsal openings except for a medial long vacuity posterior to the horn core (UNSM 4735; Gilmore, 1920; Madsen and Welles, 2000). Within abelisaurids, Carnotaurus (MACN‐CH 894) has a dorsal anteroposterior row of eight foramina on each nasal (Figures 4a and 6). Each foramen has a diameter of approximately 5 mm and is partially obscured by the rugosities. Conversely, Abelisaurus (MPCA‐Pv 11098) and Majungasaurus (Sampson and Witmer, 2007) do not have foramina over the dorsal surfaces, although the nasals of Majungasaurus show ventrolaterally placed foramina (Figure 4c), an autapomorphy of this taxon (Sampson et al., 1998). In Rugops, each nasal possesses a row of dorsal foramina of large diameter, close to the external margin but just medial to the bony crests (Sereno et al., 2004), a feature considered as an autapomorphy of Rugops (Carrano and Sampson, 2008) or even an apomorphy of Abelisauridae (Sampson and Witmer, 2007). Nevertheless, Skorpiovenator also shows a row of large foramina over the dorsal surface (Figure 1a,b), although this row is uniformly continuous toward the frontal contact. This condition is slightly different from the pattern seen in Rugops (Sereno et al., 2004), in which the nasal foramina are positioned mostly over the nasal mid‐length (Figure 5c).

Figure 6.

Digital reconstruction of the skull of Carnotaurus sastrei (MACN‐CH 894) and internal canals in nasal bones. Complete skull in right lateral (a) and (b) dorsal views; bone is rendered semitransparent. Rendered and isolated nasals in dorsal view (c, d); bone is rendered (c) solid and (d) semitransparent, showing the dorsal foramina and the internal canals (red). Nasals in anterolateral view (e); bone rendered semitransparent. Arrows in c‐e indicates anatomical orientations. Abbreviations: a, anterior; ac?, accessory canal?; antf, antorbital fenestra; dfo, dorsal foramina; fh, frontal horns; l, lateral; mc, main canal; p, posterior; o, orbit; sc, short canals linking with the main canal; sn, snout; stf, supratemporal fenestra; q, quadrate. Scale bars equal 20 cm (a, b) and 5 cm (c‐e).

4.1.6. Presence of nasal recesses

The recent amount of evidence indicates that certain theropods developed some sort of internal pneumaticity on their nasals, a possibly apomorphic trait (e.g., Monolophosaurus, Zhao and Currie, 1993; Guanlong, Xu et al., 2006). The first documented abelisaurid with a conspicuous nasal recess is Majungasaurus in which their fused nasals are extensively pneumatic, evidenced by an anteroposteriorly long and smooth walled cavity, probably originated by a nasal sinus (Sampson and Witmer, 2007). The novel information described here shows that Skorpiovenator has a large and elongate canal inside each nasal that in turn is connected externally trough the dorsal foramina (Figure 2), but the pneumatic origin for this canal is doubtful (see Biological implications, below). Furthermore, the CT scans of Carnotaurus revealed that both nasals are predominantly compact, lacking any evidence of pneumaticity or at least large recesses. Instead, the nasals of Carnotaurus show a system of slender canals in connection with the dorsal foramina (Figure 6). The pattern consists in a main canal that runs anteroposteriorly very close to the median plane of the nasal and links all the canals coming from each dorsal foramen; this pattern is comparable to that of Skorpiovenator. The diameter of the main canal on the nasals of Carnotaurus is similar to that of the individual canals coming from the dorsal foramina. However, the openings on the dorsal surface of each individual canal are wider than the canal itself. Further, the main canal seems to have some narrow canals leading to the antorbital cavity; however, the ventral aspect of the holotype nasals does not evidence distinct ventral foramina. These narrow canals in the nasals of Carnotaurus do not constitute pneumatic recesses, which undoubtedly characterized the nasals of Majungasaurus (Sampson and Witmer, 2007).

CT scans of Abelisaurus (MPCA‐Pv 11908) do not have enough resolution to determine if numerous foramina are present, except for two small foramina over the ventral surface facing the antorbital cavity (Ariana Paulina‐Carabajal, pers. obs.); thus, it is not clear whether they are connected to some internal canals. But regarding the presence of large recesses, these are definitely absent in Abelisaurus. On the other hand, the internal structure of the nasals of Rugops is currently unknown, but the external morphology of these bones (Sereno et al., 2004) is broadly similar from that of Skorpiovenator, especially sharing a closely similar foramina pattern. Moreover, Porter (2015) reported a series of conspicuous foramina over the ventral surface in Rugops, which were likely connected to some internal system and consequently to the dorsal foramina. Thereby, the presence of some kind of internal canals, or even recesses, within the nasals of Rugops would not be unexpected.

4.2. Nasal patterns and comments on the presence of foramina in Abelisauridae and other theropods

In order to describe the nasal disparity within abelisaurids, we putatively classified it into two general patterns that do not necessarily reflect particular phylogenetical relationships. These patterns were based on the set of traits previously discussed with emphasis on some of them such as the nasal transversal profile, sculpturing, and presence of foramina.

4.2.1. Pattern A

Nasals characterized by subequal width along the entire length, lack of prominent lateral crests, and therefore, a transversal profile that becomes notably convex. Also, the ornamentation in general consists of a strongly rugose surface, with coarse pitting and grooving areas, and even bony lobes. This pattern is present in Abelisaurus, Carnotaurus, and Majungasaurus (Figure 4). However, within these abelisaurids there are other features that are variable, such as the degree of internasal fusion (e.g., fused nasals in Abelisaurus and Majungasaurus, unfused in Carnotaurus), the presence of dorsal foramina obscured by rugosities (present in Carnotaurus, but absent in Abelisaurus and Majungasaurus) and the presence of large internal cavities (nasal recess in Majungasaurus; narrow canals in Carnotaurus; absent cavities in Abelisaurus).

4.2.2. Pattern B

Nasals posteriorly expanded, generally unfused, characterized by having lateral and tall bony crests, which give the nasals a transversally concave profile, conspicuous foramina over the dorsal surface, and sculpturing relatively developed composed by hummocky rugosities. This pattern is present in the nasals of Skorpiovenator and Rugops (Figure 5). The nasal pattern of these abelisaurids shows some differences concerning the distribution of the dorsal foramina. In Rugops, the foramina seem to be located mostly over the mid‐portion of the bone, whereas in Skorpiovenator, the foramina are disposed as a homogeneous line along the complete dorsal surface. Regarding size, the foramina of Skorpiovenator are similarly sized than those of Rugops when comparing its diameter relative to the nasal mediolateral width.

Although relatively few abelisaurids preserve nasals, it is interesting to note that the most outstanding trait, asides from the transversal concave profile, is the presence of a lateral row of dorsal foramina which are distinctly large in Skorpiovenator and Rugops (Figure 5). As we stated above, though present in Carnotaurus, these foramina are proportionally smaller (Figure 4), whereas in other well‐known ceratosaurs, the nasals lack any kind of dorsal and large foramina (e.g., Ceratosaurus, Abelisaurus, Majungasaurus). Furthermore, the presence of such peculiar row of dorsal foramina in Skorpiovenator and Rugops contrasts with most theropod groups in which the nasal dorsal surface generally devoid of large neurovascular foramina (e.g., coelophysoids, allosauroids, megaraptorans, coelurosaurs). Nonetheless, there are some non‐ceratosaur theropods that exhibit rows of large foramina on the dorsal nasal surface. Nasals pierced by a row of dorsal foramina are documented in some tyrannosaurids such as Albertosaurus, Gorgosaurus, and Daspletosaurus (Currie, 2003; Carr et al., 2017; Figure 1), which have numerous foramina (about ten foramina); whereas some gigantic tyrannosaurids (Tyrannosaurus and Tarbosaurus; Hurum and Sabath, 2003) tend to have a row of fewer and relatively smaller foramina. Likewise, the basal carcharodontosaurid Acrocanthosaurus (NCSM 14345) also shows a distinct row of dorsal foramina on each nasal, although those kind of openings are not present in derived carcharodontosaurids such as Giganotosaurus (MUCPv‐CH‐1) and Mapusaurus (MCF PVPH 108.1; Coria and Currie, 2006) where the dorsal surface of the nasals is covered by large rugosities and bumps. However, the nasals of the above‐cited tyrannosaurids and Acrocanthosaurus are perhaps more similar to those of Carnotaurus, rather than the transversally concave and less‐ornamented nasals of Skorpiovenator and Rugops.

In sum, it is striking that some taxa of different theropod clades such as Abelisauridae (e.g., Skorpiovenator, Carnotaurus), Tyrannosauridae (Gorgosaurus, Tyrannosaurus), and basal Carcharodontosauridae (e.g., Acrocanthosaurus) developed a dorsal row of foramina on each nasal) revealing that perhaps most theropods shared a similar innervation or vascularization within the nasals but probably the vast majority does not have distinct bony correlates on the dorsal surface, as certainly occurs in some abelisaurids.

4.3. Biological implications

Extant archosaurs lack any sort of dorsal foramina on their nasal bones as certainly occur in some abelisaurid theropods. However, the nasals of crocodiles and birds are vascularized by several blood vessels that run over the ventral aspect of the nasal bones leaving bony grooves and, at least in some crocodiles, these vessels send small branches within the nasals through ventral foramina (Caiman yacare, Mauricio Cerroni, pers. obs.; Alligator mississippiensis, Porter and Witmer, 2016).These blood vessels correspond to the lateral and medial nasal arteries and veins (Sedlmayr, 2002; Porter and Witmer, 2016; Porter et al., 2016). Regarding innervation of this zone, it is mostly composed by sensory branches of the ophthalmic division of the trigeminal nerve (CN V₁), which conveys to the brain sensory information gathered from the skin (Dubbeldam, 1993; Witmer, 1995; Leitch and Catania, 2012; George and Holliday, 2013). Considering these soft tissues present in extant archosaurs, we interpret the nasal foramina of Skorpiovenator and Carnotaurus as osteological correlates of a neurovascular system. Despite crocodiles and birds do not show homologous dorsal foramina on their nasals, the evidence for a neurovascular hypothesis of these foramina in abelisaurids is based on the continuous nature of the foramina row, the symmetrical arrangement and lateralized position of the foramina on each nasal, and the presence of grooves associated with the foramina. This occurs in other bones perforated by foramina (e.g., premaxilla, maxilla, dentary, lacrimal), which have been shown to be related to a neurovascular system in living archosaurs such as crocodiles and birds (Leitch and Catania, 2012; George and Holliday, 2013; Barker et al., 2017). Furthermore, the neurovascular assumption of the abelisaurid nasal foramina is also demonstrated by the inner main canal associated to each dorsal foramen through short passages. Thus, the canals seen in abelisaurids may represent a correlate for a branch of the lateral nasal vessels (Porter et al., 2016; Porter and Witmer, 2019), running from the anterior region to the posterior end of the nasals, and emerging over the dorsal surface through the large foramina.

It should be noted that bone correlates for the lateral nasal blood vessels were firstly proposed by Porter (2015). This author reported grooves associated with foramina on the ventral aspect of the nasals of Rugops and the tyrannosaurid Albertosaurus, leading the author hypothesize that such ventral system of foramina was probably linked internally to the dorsal foramina; finally concluded that such foramina are correlated to the lateral nasal vessels (Porter, 2015). As bony canals usually not only house blood vessels but nerves as well, conforming “neurovascular bundles” (see Porter and Witmer, 2019), the innervation of the neurovascular structures in abelisaurid nasals was probably given by some subdivision of the ophthalmic branch of the trigeminal nerve as in living archosaurs (Witmer, 1995).

In spite of both Skorpiovenator and Carnotaurus share a common neurovascular canal that links with all dorsal foramina, they exhibit differences in the inner canal width and foramina diameter. Skorpiovenator preserves a main internal canal notably larger in transversal section than in Carnotaurus, in which is markedly slenderer (Figures 2 and 6). Moreover, the dorsal foramina of Skorpiovenator are proportionally wider and are well exposed in dorsal view, whereas in Carnotaurus they smaller and partially obscured by the projecting rugosities. The presence of a notably larger canal in the nasals of Skorpiovenator would lead to the hypothesis that these bones have some pneumatic component. Majungasaurus and some non‐ceratosaurian theropods (e.g., Sinraptor, Monolophosaurus, Guanlong) show highly pneumatic nasals, with distinct nasal recesses, which were presumably originated by the invasion of a diverticulum from the main antorbital sinus (Witmer, 1997; Sampson and Witmer, 2007). Although it is not unusual that air‐filled sinuses invade opportunistically bone taking vascular paths, sometimes resulting in the loss of vessel boundaries (Witmer, 1995, 1997), the preserved zones of the ventral aspect of the nasals of Skorpiovenator does not exhibit any sort of ventral or ventrolateral apertures that could indicate an entrance path for some sinus derived from the antorbital cavity, as well as lack of large and notorious cavities (i.e., pneumatic recesses), as it certainly does in Majungasaurus (Sampson and Witmer, 2007). It is worth mentioning that the ventral surface of the nasals of Skorpiovenator is poorly preserved due erosion, and if there was some ventral opening (as the reported in Rugops; Porter, 2015) it is currently lost; whereas the nasals of Carnotaurus are well preserved but does not have any ventral large foramina. Thereby, taking into account that a neurovascular correlate of the dorsal foramina and the internal canals seen in Skorpiovenator and Carnotaurus is the most probable, then the markedly size variation on both foramina and canals of these abelisaurids may respond to a difference on the amount of vascularization in the nasal bones, and even may reflect differences in the number of nerve fibers, as occurs in fossil and extant crocodyliforms (George and Holliday, 2013).

On the other hand, regarding the large foramina on the skull roof of Skorpiovenator, we can envision that such foramina were also related to neurovasculature as evidenced by the most anterior foramen, which is connected with the main internal canal of the nasal. As we stated in the description section, it is unknown if the middle and posterior foramina on the skull roof of Skorpiovenator have been associated to internal canals or grooves, due to preservational artifact. However, in extant birds and crocodiles the supraorbital blood vessels (artery and vein) send branches that exit on the dorsal surface of the frontal, supplying the skin (Porter and Witmer, 2016; Porter et al., 2016). Hence, it would be plausible that the most posteriorly located foramina of Skorpiovenator would be vascularized by some branch from the supraorbital vessels; however, the markedly large size of these foramina is intriguing considering the tiny foramina that pierce the frontal in extant archosaurs. With regard to the innervation, it is possible that some ramification of the trigeminal nerve reached these foramina (Sedlmayr, 2002; Leitch and Catania, 2012). In sum, Skorpiovenator shows a unique skull in which the pattern of dorsal foramina of the skull roof and nasals represents an exceptional novelty among theropods.

The most substantial question here is what possible functions these canals and foramina played during life of these abelisaurids. Although speculative, some biological inferences (all with several levels of speculation) can be done.

4.3.1. Thermoregulation

In living amniotes, it is currently known that the nasal region is highly vascularized and acts as a main site of thermal exchange as in squamates (Crawford et al., 1977), birds (Midtgård, 1984; Porter and Witmer, 2016) and possibly crocodiles (Porter et al., 2016). More precisely this thermal exchange intervenes through evaporative cooling of dural sinuses and brain, avoiding overheat of the encephalic tissues. It was recently hypothesized that several structures such as grooves, canals and foramina on the dinosaur skull, including theropods (e.g., Majungasaurus), served as sites where blood vessels in life were acting as thermoregulatory zones of the antorbital cavity that helped brain to avoid overheating (Porter and Witmer, 2019). These authors pointed out that several correlates on the nasal bones indicate that a large volume of blood supplied the antorbital sinus, allowing this sinus to act as a thermoregulatory zone, in addition to the orbital and oral zones. In this way, if the canals and the large dorsal foramina in the nasals of Skorpiovenator, Carnotaurus and Rugops correspond to sites that housed large volumes of blood (probably correlates of the lateral nasal vessels) indicating a highly specialized vascularization, we can speculate a significant zone of thermal exchange. However, due to the fact that the nasals of Skorpiovenator show markedly larger canals and foramina, compared with those structures present in Carnotaurus, it could mean that Skorpiovenator had a major vascularization resulting perhaps in an enhanced thermoregulatory zone.

Nonetheless, the neurovascular features on the nasals of these abelisaurids do not preclude that other zones of the skull might have played a thermoregulatory rol, such as the oral and orbital regions, as occurs in extant archosaurs (Porter et al., 2016). Likewise, the notably enlarged foramina of the skull roof in Skorpiovenator also would allow a large blood volume acting as a thermoregulatory site, in addition to the described nasal foramina. This is not rare among fossil archosaurs; it is known that several crocodyliforms bear correlates on their skull roof related to superficial vascularization, which has been interpreted as a physiological mechanism of thermoregulation of the cephalic tissues (Holliday and Gardner, 2012; Bona et al., 2013; Souza‐Filho et al., 2019). However, the size and amount of foramina in Skorpiovenator is far beyond the observed in such extinct crocodyliforms. Therefore, the neurovascular correlates seen in the nasals and skull roof of Skorpiovenator (as possibly in the nasals of Carnotaurus and Rugops as well) would represent an enhanced site of thermoregulation, likely related to the paranasal sinuses (at least the nasal foramina), and perhaps denoting a focused thermoregulatory strategy in the sense proposed by Porter and Witmer (2019).

4.3.2. Display structures

Following the idea that the foramina of abelisaurids would be structures highly supplied by blood, another potential role is the vascularization of dermal tissues. A greater role for blood vessels within the nasal bones would provide a large amount of vascularization by nourishing the mineralized skin. This large volume of blood may have incidence on the development of display structures, which may be useful in social interactions between co‐specific individuals (Horner, 2000; Hieronymus, 2009; Padian and Horner, 2010; Delcourt, 2018). The highly projected rugosities in the nasals of Carnotaurus may have a considerable amount of blood coming from the foramina, when compared with the lesser sculptured nasals of Skorpiovenator and Rugops in which the rugosities are conformed by hummocky rugosities within the anastomosed web of minute furrows. Although in Skorpiovenator and Rugops the sculpturing is not as strongly developed as in Carnotaurus, the large foramina exhibit grooves leading into each foramen. As previously identified Sereno et al. (2004) in Rugops, these grooves likely correspond to vascular correlates and the authors proposed that they served as anchorage structures for soft tissues involved in social displays. Thereby, we cannot rule out the possibility that the dorsal foramina (and the internal canals as well) corresponded to a site of vascularization that acted supplying display structures with large amounts of blood, an idea previously suggested for the cranial bones of other dinosaurs (Barsbold, 1988; Xu et al., 2006; Padian and Horner, 2010; Porter, 2015). Nevertheless, it should be noted that vascular traits associated with thermoregulatory and nourishing functions are not mutually exclusive, being possible both physiological roles (Farlow et al., 2010; Holliday and Garner, 2012; Porter, 2015).

4.3.3. Sensory structures

Further conjectures concerning the biological role of the dorsal foramina can be made; however, these involve a higher level of speculation when compared to the thermoregulatory and display functions. On one hand, a possible sensory structure could have been present, as was proposed previously for Rugops (Sereno et al., 2004). This assumption is based on the sensory structures currently present in living crocodiles. These animals have a complex neurosensorial system distributed over the epidermal cover of the skull and part of the neck in alligatoroid crocodiles, whereas in other groups (crocodylids and gavialids) the mechanoreceptors are more widespread all over the body (Leitch and Catania, 2012). These structures called ISO (Integumentary Sensory Organs) are present over the snout and the dorsal surface of the skull and it has been shown that it has mechano‐, thermo‐ and chemosensory functions (Soares, 2002; Di‐Poï and Milinkovitch, 2013). The ISOs that cover the dorsal surface of the skull are mostly innervated by the ophthalmic ramus of the trigeminal nerve (CN V₁), whereas the ISOs present in snout bones are innervated by the maxillo‐mandibular ramus (CN V_2,3), leaving numerous foramina in the bone surface of several bones (e.g., premaxilla, maxilla, lacrimal; Barker et al., 2017). Despite nasal and lacrimal bones of crocodylians are largely free of receptors, the ophthalmic ramus innervates these bones as well as the dorsum of the head (Leitch and Catania, 2012). Further, it has been proven that the trigeminal nerve innerves somatosensory structures in the facial bones in a variety of tetrapods such as snakes, birds and monotremes (Soares, 2002; Schneider et al., 2016). We do not mean that abelisaurid theropods developed ISOs on their skin, because these structures are apparently related to semiaquatic habits. However, the dorsal foramina of Skorpiovenator, Carnotaurus and Rugops might suggest a pathway for the innervation by the trigeminal nerve leading to the hypothesis that these abelisaurids (as well as some large theropods) would have developed some type of sensory structures on the dorsum of the head.

Nevertheless, we take into account that many theropods including abelisaurids (Sampson and Witmer, 2007) and non‐ceratosaur theropods (e.g., Spinosaurus, Ibrahim et al., 2014; Neovenator, Barker et al., 2017; Daspletosaurus, Carr et al., 2017) exhibit snout bones (premaxilla, maxilla, dentary) pierced by a large amount of neurovascular foramina, which transmitted fibers of the trigeminal nerve. This led some authors to argue in favor of a highly sensitive integument in the snout, as occurs in extant crocodilians, through the specialization of mechanoreceptors (Barker et al., 2017; Carr et al., 2017). Further, Barker et al. (2017) noted that a specialized sensory snout would have played a major role in facial contact behaviors (e.g., intraspecific combat, prey capture). In this context, the nasals of Skorpiovenator, Rugops and Carnotaurus displays enlarged foramina facing dorsally, a situation that clearly contrasts with the numerous foramina scattered and concentrated on the most anterior portion of the snout of many theropods (Ibrahim et al., 2014). Thus, the hypothesis of a very sensitive area in abelisaurid nasals has weaker support, although this need to be tested with further studies focused on the dorsal region of theropod skulls and it would be interesting to see in detail more similitudes and differences with the snout region.

4.3.4. Glandular hypothesis

Finally, a farther speculative hypothesis involves a glandular function for the dorsal foramina of both nasals and skull roof. Although the dorsal foramina of abelisaurids might have involved glandular tissues, like a salt gland (as in living birds), it would be unlikely. In crocodilians the nasal gland is ventromedially placed to the nasals (within the context of the antorbital cavity) (Witmer, 1995). On the other hand, a salt gland is present in some living birds occupying different locations including the supraorbital region and the antorbital cavity (Witmer, 1995, 1997). Even though the skull roof foramina of Skorpiovenator, especially middle and posterior foramina, might suggest a potential site for a salt gland due its similarly supraorbital position (just dorsal to the orbit), it is unlikely in light of the previous arguments, mostly due the neurovascular inference (i.e., correlate for the artery blood vessels and trigeminal nerve).

5. CONCLUSIONS

The novel information of the skull anatomy of Skorpiovenator presented here increases the disparity regarding the nasal morphology within abelisaurid theropods. Furthermore, it provides evidence that some structures, such as the dorsal foramina of the nasals, which were previously known for Carnotaurus and Rugops, are more widespread than previously known for other abelisaurids. Further, the series of large foramina on the skull roof is a trait not recorded in other theropod dinosaurs constituting an intriguing novel feature, and being a possible new autapomorphy of Skorpiovenator. But more interestingly, is the information afforded by the CT scans. These show that both Skorpiovenator and Carnotaurus have an internal system of canals linked to the dorsal nasal foramina, which likely represent a correlate for a neurovascular complex. The neurovascular system, probably related to the lateral nasal vessels and perhaps innervated by the trigeminal nerve as in extant archosaurs, indicate an enhanced supply of blood to these bones that might be related to some possible biological mechanisms. The most likely implies a site of thermal exchange, in close association with the antorbital sinus, which helped in the thermoregulation of dural sinuses. However, other explanations, such as the vascularization of dermal tissues for display interactions and the presence of some sensory structures as seen in extant archosaurs, cannot be ruled out and deserves more emphasis in future studies on the skull of these theropods.

AUTHOR CONTRIBUTIONS

All the authors were responsible in the editing of the manuscript. M.A.C., J.I.C. and A.P‐C acquired the CT scans. J.I.C. and F.E.N. advised on the design of the research. M.A.C. segmented tomograms and build the 3D models. All authors conducted and analyzed the results and discussion.

Supporting information

Figure S1

Figure S2

Cerroni MA, Canale JI, Novas FE, Paulina‐Carabajal A. An exceptional neurovascular system in abelisaurid theropod skull: New evidence from Skorpiovenator bustingorryi . J. Anat..2022;240:612–626. 10.1111/joa.13258

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available in the supplementary material of this article.