Spinosaur taxonomy and evolution of craniodental features: Evidence from Brazil (original) (raw)

Loading metrics

Open Access

Peer-reviewed

Research Article

Cesar L. Schultz

Spinosaur taxonomy and evolution of craniodental features: Evidence from Brazil

Marcos A. F. Sales,
Cesar L. Schultz

Published: November 6, 2017
https://doi.org/10.1371/journal.pone.0187070

Figures

Abstract

Fossil sites from Brazil have yielded specimens of spinosaurid theropods, among which the most informative include the cranial remains of Irritator, Angaturama, and Oxalaia. In this work some of their craniodental features are reinterpreted, providing new data for taxonomic and evolutionary issues concerning this particular clade of dinosaurs. The mesial-most tooth of the left maxilla of the holotype of Irritator is regarded as representing the third tooth position, which is also preserved in the holotype of Angaturama. Thus, both specimens cannot belong to the same individual, contrary to previous assumptions, although they could have been the same taxon. In addition, the position of the external nares of Irritator is more comparable to those of Baryonyx and Suchomimus instead of other spinosaurine spinosaurids. In fact, with regards to some craniodental features, Brazilian taxa represent intermediate conditions between Baryonychinae and Spinosaurinae. Such a scenario is corroborated by our cladistic results, which also leave open the possibility of the former subfamily being non-monophyletic. Furthermore, the differences between spinosaurids regarding the position and size of the external nares might be related to distinct feeding habits and degrees of reliance on olfaction. Other issues concerning the evolution and taxonomy of Spinosauridae require descriptions of additional material for their clarification.

Citation: Sales MAF, Schultz CL (2017) Spinosaur taxonomy and evolution of craniodental features: Evidence from Brazil. PLoS ONE 12(11): e0187070. https://doi.org/10.1371/journal.pone.0187070

Editor: Ulrich Joger, State Museum of Natural History, GERMANY

Received: March 31, 2017; Accepted: October 12, 2017; Published: November 6, 2017

Copyright: © 2017 Sales, Schultz. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: All relevant data are within the paper and its Supporting Information file.

Funding: The first author received a Ph.D. grant from the Conselho Nacional de Desenvolvimento Científico e Tecnológico (Grant 141268/2013-1) and a short-term research grant from the Deutscher Akademischer Austauschdienst (Grant 57130097). The informed funders provided support for data collection throught visits to scientific paleontological collections in Brazil and Europe.

Competing interests: The authors have declared that no competing interests exist.

Introduction

Spinosauridae are among the most iconic dinosaur groups of all time [1]. This status is strengthened every time a new study on them is published (e.g., [2–7]). The history of the knowledge on spinosaurids began in 1912, when the fossil collector Richard Markgraf unearthed a partial skeleton with neural spines up to 165 cm in height from the Bahariya Oasis, western Egypt [8,9]. These remains, which represented the holotype of Spinosaurus aegyptiacus Stromer, 1915 and were housed at the Paläontologische Staatssammlung München, were destroyed by a British air raid on Munich during World War II, along with other specimens also referred to that genus [1,8–11].

Consequently, for a long time spinosaurid material was restricted to isolated bones and teeth, although many records were probably misinterpreted as crocodylomorph remains (even prior to the discovery of Spinosaurus) [12–14]. However, since the 1980s, new remains have been reported from various parts of the world, providing insights on the anatomy of these theropods (e.g., [1,2,4,5,7,11,15–18]). These records extended the geographic range of Spinosauridae to Europe, Asia, and Gondwanan regions other than Africa, such as South America and perhaps Australia, and its temporal distribution back into the Jurassic [2,16,19–23]. Some authors have also suggested their possible presence in the Late Jurassic of North America [16,24,25].

Spinosaurid remains from Brazil have played an important role in discussions regarding this clade. Firstly, they comprised the first unequivocal record out of Africa and Europe [19,26], despite putative reports from Asia by that time [27]. Secondly, the holotype of Irritator challengeri Martill et al., 1996, from the Romualdo Formation (Albian, Araripe Basin), is the spinosaurid specimen with the most complete preserved skull [15]. On the other hand, the holotype of Oxalaia quilombensis Kellner et al., 2011, from the Alcântara Formation (Cenomanian, São Luís-Grajaú Basin), represents the largest theropod hitherto known from Brazil, likely reaching c. 12–14 m in length [28]. Despite being smaller than the largest spinosaurid skull reported [5], it provides additional evidence for the wide distribution of large-bodied spinosaurs during the mid-Cretaceous [1,29]. Finally, the Brazilian spinosaurid record also includes isolated teeth and postcranial remains [30–37]. Among them, the most outstanding specimen is a partial postcranial skeleton firstly mentioned in scientific meeting abstracts [38–40] and later described in an unpublished Master’s thesis [41].

Current research on spinosaurids is largely focused on African material, whereas most Brazilian specimens have already been described and redescribed. However, the re-examination of the latter revealed the need for further considerations on a set of their traits. This is particularly true for cranial remains, which overlap other relevant specimens from overseas fossil sites with respect to the known portions of the skull. Moreover, many diagnostic features both at or below family rank for spinosaurids reside in the skull [1,4,5,29]. Here, we provide a reappraisal of spinosaurid skull materials from Brazil, focusing mainly on attributes that were either overlooked or previously interpreted. The new data also provide taxonomic and evolutionary implications for Spinosauridae as a whole, regarding craniodental features and phylogenetic ingroup relationships.

Geological settings

Brazilian spinosaurids are Cretaceous in age and were found in the northeastern region of Brazil (Fig 1). Cranial remains correspond mainly to the holotypes of all formally described species, namely Irritator challengeri, Angaturama limai, and Oxalaia quilombensis. The first two taxa are from the concretion-bearing beds of the Araripe Basin, one of the most famous Fossil Lagerstätten in the world [42,43]. For nomenclatural purposes, we will follow here Valença et al. [44], referring to the lithostratigraphic unit of the concretion-bearing beds as the Romualdo Formation. Besides its remarkable and abundant carbonate concretions, usually containing exquisitely preserved fossils, this formation is also characterized by the presence of shales, marls, and limestones. Along with its fossil content, geological data indicate a transitional depositional setting for the majority of the stratigraphic sequence of the Romualdo Formation and a late Early Cretaceous (Albian) age. However, in the western portion of the formation and above the concretion-bearing beds, there are strata containing echinoids and mollusks, thereby evidencing local marine conditions [44–47]. In fact, the holotypes of Irritator and Angaturama have been found inside concretions, which is common for most Romualdo Formation fossils. Again, as many Romualdo fossils, both spinosaurs lack accurate geographic and stratigraphic information [43,48]. The concretion containing Irritator was briefly described as including the ostracod Pattersoncypris and scales of the ichthyodectiform fish Cladocyclus. It may have come from Buxexé, a locality near Santana do Cariri Municipality, Ceará State [15,49]. On the other hand, the concretion enclosing Angaturama was considered as typically resembling those from the Romualdo Formation (Romualdo Member of the Santana Formation in Kellner and Campos [19]). Both cranial remains are distinct yet complementary to each other. Due to coming from the same stratigraphic unit, some authors suggested that they could have pertained to the same specimen (e.g., [4,50]).

Fig 1. Map of northeastern Brazil showing the location of the Araripe and São Luís-Grajaú basins.

The (likely) provenance of the holotypes of local spinosaurid taxa is indicated. Modified from Bittencourt and Langer [51].

https://doi.org/10.1371/journal.pone.0187070.g001

On the other hand, the holotypic premaxillae of Oxalaia, along with a fragmentary left maxilla referred to the same taxon, came from a fossil site called Laje do Coringa, in Cajual Island, Alcântara Municipality, Maranhão State (Fig 1) [28]. This locality corresponds to the most important fossiliferous outcrop of the Alcântara Formation, which is characterized by sandstones and mudstones intercalated with conglomeratic beds where most fossil remains are found. The majority of these fossils are released from the matrix by the continuous action of tides and waves. Due to its macrofossil and palynological content, the Alcântara Formation is usually regarded as the Cenomanian strata of the São Luís-Grajaú Basin, deposited in a transitional setting related to the opening of the equatorial portion of the South Atlantic Ocean [52–54]. Fossil remains are frequently fragmentary and show evidence of reworking; therefore, some degree of time-averaging is expected for the Alcântara fossil assemblage, which is composed of continental, coastal, and marine taxa. Despite the time-averaging, most of the identified taxa might have been coeval throughout part of the Cenomanian and ecologically related to each other [30,31,54,55]. The holotype of Oxalaia was reportedly found in situ, while the referred maxilla was collected as “surface float” [28].

Materials and methods

Descriptive basis

As already stated, this work focuses on craniodental features either overlooked or interpreted previously in a different manner. We follow Carrano et al. [29] and Hendrickx and Mateus [56] wherever possible when referring to cranial structures. Regarding teeth and alveoli, we adopt the terminology proposed by Smith and Dodson [57], which was further developed by Hendrickx et al. [58].

Below, we provide a revised diagnosis for each Brazilian taxon in light of the current knowledge on spinosaurids. The revised diagnoses are differential in essence and do not focus solely on autapomorphies, given that distinct features are, in general, independently distributed among all spinosaurids. Therefore, they constitute a unique set of characters for each taxon. However, wherever possible, autapomorphies are temporarily recognized with the caution of each of them being possibly later figured out to be an intraspecific or ontogenetic variation, or even a pathologic condition. This possibility is real as most taxa are represented only by a single individual with the skull preserved, considering formally published specimens [1].

Cladistic analyses

The cladistic analyses presented here are based on the data matrix of Carrano et al. [29], which was designed for evaluating ingroup relationships of Tetanurae. However, some significant changes were performed. First of all, we incorporated all the changes made by Evers et al. [17], regarding the vertebral character coding of Carcharodontosaurus. We then added three character statements and modified others as the result of our new perspective on relative spinosaurid cranial features (see S1 File).

With respect to operational taxonomic units (OTUs), we also performed additional modifications. For example, we added Oxalaia as a new OTU in order to investigate the phylogenetic relationships of all Brazilian taxa. In addition, we excluded the OTUs previously regarded by Carrano et al. [29] as wild cards, Megaraptor (whose coding did not include recent data [59]), and the spinosaurids whose holotypes correspond solely to non-cranial remains (i.e., Ichthyovenator and Sigilmassasaurus [16,17,60,61]). Spinosaurus is another spinosaurid whose holotype comprises mainly postcranial bones, but normally this taxon has been coded (e.g., [4,29]) considering also cranial remains later assigned to it [5,60,62]. Nevertheless, we consider their attribution as tentative because some of those remains were found either isolated or lacking accurate field data and they only just overlap the holotypic remains described by Stromer [8]. Furthermore, keeping in mind that many mid-Cretaceous African sites are only roughly coeval and the possible evidence indicating the presence of more than one spinosaurid morphotype (and taxon?) therein [17,18,63], it is difficult to accurately attribute isolated skulls to any of the taxa recognized locally. This is the case of the Kem Kem Beds of Morocco, from which comes material also assigned to Sigilmassasaurus [17,60,61]. As Spinosaurus and Sigilmassasaurus may not be synonymous (see Ibrahim et al. [7,64] for a different opinion), it is not possible yet to assign a huge Moroccan partial skull (MSNM V4047, Museo Civico di Storia Naturale di Milano, Milan, Italy) [5] to any of them. The same is true for another skull found in Algeria and tentatively attributed to S. maroccanus (MNHN SAM 124, Muséum National d’Histoire Naturelle, Paris, France) [62], which is a nomen dubium according to Sereno et al. [4] and Carrano et al. [29]. Thus, we prefer not to refer these two non-Egyptian skulls to S. aegyptiacus. Recently, a partial skeleton, including cranial bones, was proposed as the neotype of S. aegyptiacus [7]. However, some authors [17] have challenged its attribution specifically to that species. Considering that we could not handle the proposed neotype, which awaits a thorough description, we opted for not including Spinosaurus in our analyses. Instead, we coded MSNM V4047 as a distinct OTU in place of the latter.

Another taxonomic issue pervading the choice of spinosaurid OTUs is the possible synonymy between Cristatusaurus and Suchomimus, both from the Elrhaz Formation (horizon GAD 5; Aptian-Albian?) of the Tegama Group of Niger [4,28,29,62]. The former taxon was the first described, but its remains and proposed diagnosis are thought to be too general to differentiate it from other serrated-toothed spinosaurids [3,4,29]. Shortly after the description of Cristatusaurus, Suchomimus was described and the provided diagnosis for this taxon is enough to set it apart from Baryonyx [2–4,29]. Despite claims of Cristatusaurus as a nomen dubium, as well as some similarities between it and Suchomimus or even Baryonyx, we recognize some differences between them with possible taxonomic significance (see Discussion). However, as specifically addressing this issue is not the main goal of this work, for practical purposes we regard Cristatusaurus and Suchomomimus as distinct OTUs also coded for this analysis.

We also performed one analysis excluding only Irritator and another removing only the other Brazilian taxa. We did so because, due to the large amount of missing data, each OTU from Brazil acted as a wild card within Spinosauridae. This procedure recovered more refined hypothetical phylogenetic relationships for Irritator and both Angaturama and Oxalaia in each analysis, which contributed to the discussions regarding the evolution of the craniodental features considered in this work.

Finally, we performed the analysis techniques using the free software TNT v. 1.5 [65], following the same parameters set by Carrano et al. [29].

Additional institutional abbreviations

MN, Museu Nacional/Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil; MNN, Musée National du Niger, Niamey, Republic of Niger; SMNS, Staatliches Museum für Naturkunde Stuttgart, Stuttgart, Germany; UCPC, University of Chicago Paleontological Collection, Chicago, United States; USP, Universidade de São Paulo, São Paulo, Brazil.

Results

Systematic paleontology

Dinosauria Owen, 1842 [66]

Theropoda Marsh, 1881 [67]

Tetanurae Gauthier, 1986 [68]

Spinosauridae Stromer, 1915 [8]

Irritator challengeri Martill et al., 1996 [49]

Holotype and only known specimen.

SMNS 58022, an almost complete and articulated skull associated to the mandibular rami, lacking the anterior tip of the rostrum and the dentigerous portions of both mandibulae (Figs 2 and 3).

Fig 2. Specimen SMNS 58022, holotype of Irritator challengeri.

A, Left lateral view. B, Right lateral view. The abbreviation for the third tooth of the left maxilla follows Hendrickx et al. [58]. Additional abbreviations: ao.f, antorbital fenestra; c.c, crista cranii; e.n, external naris; m, maxilla; m.r, mandibular ramus; n, nasal; n.c, nasal sagittal crest; n.p, nasal process; o, orbit; pm, premaxilla; s, stapes.

https://doi.org/10.1371/journal.pone.0187070.g002

Fig 3. Dentition of Irritator challengeri.

A, Detail of the left maxillary tooth row. B, Detail of the first tooth crown preserved in situ of the left maxilla. C, Detail of the artificially detached fragment of the right maxilla bearing the right maxillary tooth row. D, Detail of the sixth tooth crown preserved in situ of the right maxilla in mesial view. Abbreviations for teeth follow Hendrickx et al. [58]. Additional abbreviations: e.n, external naris; m, maxilla; m.r, mandibular ramus; m.u, marginal undulations; n, nasal; pm, premaxilla; r.m, rock matrix.

https://doi.org/10.1371/journal.pone.0187070.g003

Revised diagnosis.

Spinosaurid that differs from Baryonyx, Suchomimus, and Cristatusaurus by the unserrated condition of its teeth and nearly half the number of maxillary alveoli (teeth). Differs from MSNM V4047 and MNHN SAM 124 by the relatively larger and more anteriorly placed external nares and the participation of the premaxillae in the anterior border of the external nares. Finally, an autapomorphy of Irritator is the presence of a sagittal crest formed by the conjoined nasals that ends in a knob-like projection over the frontals.

Occurrence.

Exact geographical provenance unknown, but likely near Buxexé, 5 km south to Santana do Cariri Municipality, Ceará State, Araripe Basin, northeastern Brazil. Romualdo Formation; Albian, Lower Cretaceous.

Remarks and comparisons

Prior to its assignment to Spinosauridae, the unprepared skull of Irritator was thought to be a pterosaur [15]. It was afterwards described as a theropod dinosaur, more specifically as a member of Bullatosauria [49]. Much of this taxonomic confusion was due to the modifications and damages in the specimen, which were likely caused by its collectors or additional personnel that dealt with it before its study. Actually, the etymology of the generic name is an allusion to such problems. After a more careful preparation of SMNS 58022, Sues et al. [15] were able to confirm the spinosaurid identity of Irritator, which was firstly proposed by Kellner [26].

Despite these damages, the holotype of Irritator is considered to be the most complete spinosaurid skull, at least in relation to the posterior portion (Fig 2) [15,29]. In fact, it is one of the few non-avian theropod specimens with one of its stapes preserved, in this case the right one (Fig 2B). Accordingly, the right side of the skull is best preserved [15]. The amount of data recovered from Irritator must have been possible due to its fossilization within an early diagenetic carbonate concretion. This becomes even more reasonable when taking into account the scarcity of (articulated) spinosaurid postnarial bones [2–5,19,62]. In this sense, Irritator reveals important aspects of the cranial morphology of Spinosauridae. The skull becomes progressively deeper towards its posterior end so that its dorsal outline projects posterodorsally (Fig 2), resembling the condition inferred for Baryonyx [3]. This morphology is followed by other reconstructions of spinosaurid skulls [5,7] but contrasts with that of Suchomimus [4].

Although Irritator was described and redescribed in detail, some of its craniodental features were not addressed further. For example, this taxon is remarkable because it is the spinosaurid skull with the highest number of in situ preserved teeth. However, due to the missing portion of its rostrum, including the anterior end of both maxillae, neither Martill et al. [49] nor Sues et al. [15] properly identified the preserved maxillary teeth in relation to their position along the tooth row.

The maxillary dentition of Irritator is composed of nearly straight conical crowns, which are labiolingually fluted but bear no serration on their carinae (Fig 3) [15]. These features are common in other spinosaurids (e.g., [3–5,8,36]) and, particularly, the absence of serrations is characteristic of Spinosaurinae after Sereno et al. [4]. Some teeth also exhibit short marginal undulations close to the distal carina (Fig 3B) [15]. Regarding tooth size variation, in the left maxilla the first (mesial-most) crown preserved in situ is the second largest, being slightly smaller than the second crown, whereas the seven remaining teeth become gradually smaller (Figs 2A and 3A) [15].

In spinosaurid specimens with the maxillary tooth row preserved, the largest crowns (and alveoli) are the third and the fourth. In addition, there is typically an increase in size from m1 to m4, which are followed by progressively smaller teeth. Thus, the spinosaurid maxillary dentition comprises a tooth size variation that begins with a tendency of increasing, reaching its maximum at m4, regarding the crown base length and crown height, and continues with an opposite trend. This is the case both for Baryonychinae, i.e., Baryonyx and Suchomimus, and Spinosaurinae, such as MNHN SAM 124 and MSNM V4047 [3–5]. This pattern is widely distributed among spinosaurids and no current evidence indicates that this condition is not also present in Irritator. The tooth size variation seen in the left maxilla of the Brazilian taxon matches the described pattern, so that the first two preserved teeth correspond to Lm3 and Lm4. This implies a total number of eleven left maxillary teeth (Figs 2A and 3A). Accordingly, Sues et al. [15] inferred that the left maxilla would have comprised at least 11 teeth, although they neither justified their claim nor attributed a specific location within the tooth row for any of the preserved crowns. Furthermore, this number is nearly identical to that of the spinosaurine MSNM V4047, which displays 12 maxillary teeth (Fig 4C and 4D) [5]; this dissimilarity could be explained by the difference in size, interspecific variation, different ontogenetic stages, alveolar remodeling after the loss of teeth in vivo, other dental abnormalities, etc. [28,69].

Fig 4. Schematic illustration of spinosaurid cranial remains.

A, Premaxillae and maxillae of Baryonyx walkeri. B, Partial rostrum (MNN GDF501) referred to Suchomimus tenerensis. C, Specimen MSNM V4047. D, Reconstructed skull of Irritator challengeri. All specimens are aligned based on the position of the fourth maxillary tooth, whereas the last maxillary tooth (or alveolus) in each is indicated by a black star. Abbreviations for teeth follow Hendrickx et al. [58]. Additional abbreviations: n.c, nasal contact; ao.f, antorbital fenestra; e.n, external naris. A, B, C, and D are modified from Charig and Milner [3], Sereno et al. [4], Dal Sasso et al. [5], and Sues et al. [15], respectively.

https://doi.org/10.1371/journal.pone.0187070.g004

Regarding the artificially detached fragment of the right maxilla of Irritator, nine teeth are preserved in situ as well as the broken base and partial impression in the matrix of the tenth (distal-most) crown. The second and the fourth are replacement teeth (Fig 3B). In general, the first four crowns are more damaged than those from the left side. This precludes an accurate estimation of their dimensions, although the mesial-most crown seems to have been the largest one. If the latter was indeed the largest, the preserved portion of the right maxilla would also comprise teeth from m4 to m11.

The identification of the maxillary teeth of Irritator implies the reinterpretation of the location of its external nares. Dal Sasso et al. [5] considered that these were restricted to the mid-part of the maxillary tooth row. However, their anterior-most outline must have been placed somewhere between m3 and m4 (Figs 2A, 3A and 4D). Therefore, the external nares of Irritator begin in a comparable position to those of Baryonyx and Suchomimus, in which the external nares start between m2 and m3 and between m3 and m4, respectively (Fig 4A and 4B) [3,4]. On the other hand, the external nares of MSNM V4047 start only at the level of m9 (Fig 4C) [5].

Another difference between Irritator and African material attributed to Spinosaurus is the relationship between the bones that bind the external nares. In the former genus, the premaxillae form a minor part of the ventral outline of the nares, whereas their dorsal and ventral outlines are surrounded by the nasals and maxillae (Figs 2A, 3A and 4D). In MSNM V4047, the premaxillae are completely excluded from the external nares (Fig 4C) [5]. Both spinosaurids differ from Baryonyx and Suchomimus; in these, the premaxillae are major components of the external nares, but they differ from each other in relation to the contribution of the maxillae. In Baryonyx, the maxillae form the posterior half of the preserved ventral border of the left naris (Fig 4A) [3]. In Suchomimus, the maxillae are minor components of the ventral outline of the nares and are constricted by the premaxillae and descending rami (= maxillary processes) of the nasals (Fig 4B) [4]. Furthermore, Dal Sasso et al. [5] considered the exclusion of premaxillae from the nares of MSNM V4047 as an autapomorphy of Spinosaurus. Among spinosaurids, MSNM V4047 is also unique in presenting premaxillae, maxillae, and nasals meeting at a single point, whereas Irritator, Baryonyx, and Suchomimus do not present such a triple joint (Fig 4).

Another important issue related to spinosaurid external nares is their size. The external nares of Irritator are both absolutely and comparatively (in relation to the antorbital fenestrae) smaller than those of the baryonychine Suchomimus and, possibly, Baryonyx (Fig 4A, 4B and 4D). On the other hand, they are larger than those of the spinosaurine MSNM V4047 in absolute length (Fig 4C and 4D), despite the Brazilian taxon being inferred to have had a shorter skull (c. 60 cm) [15] than the African specimen (c. 175 cm) [5,15].

Irritator is also unique in displaying a sagittal crest formed by the conjoined nasals that ends in a knob-like projection overhanging the frontals. In fact, where observable, the unfluted condition of the nasal crest of Irritator might also distinguish it from an isolated fragmentary pair of fused nasals from Morocco (UCPC-2) [5] and those reported by Ibrahim et al. [7] (and, consequently, the taxon to which they belonged). Finally, additional differences between Irritator and African material assigned to Spinosaurus might include features of the quadrates, which were well addressed by Hendrickx et al. [18] and, as such, are not further considered here.