ELECTRONIC SUPPLEMENTARY MATERIAL
First spinosaurid dinosaur from Australia and the cosmopolitanism of Cretaceous dinosaur faunas
Paul M. Barrett1,*, Roger B. J. Benson2, Thomas H. Rich3,4 and Patricia Vickers-Rich3,4
1Department of Palaeontology, Natural History Museum, Cromwell Road, London SW7 5BD, UK; 2Department of Earth Sciences, University of Cambridge, Downing Street, Cambridge CB3 3EQ, UK; 3Palaeontology Department, Museum Victoria, Melbourne, Victoria 3000, Australia; 4School of Geosciences, Monash University, Clayton, Victoria 3800, Australia.
*Corresponding author ()
Description of NMV 221081
NMV P221081 is a partial cervical vertebra comprising the centrum and left side of the neural arch (Figs 1, S1). The suture between the neural arch and centrum is clearly visible, suggesting that the vertebra pertains to a juvenile individual [1]. The neural spine and both postzygapophyses are missing, but the preserved part of the vertebra is generally well-preserved and only slightly distorted.
Figure S1. Spinosaurid vertebra NMV P221081 in anterior (a, cast), left lateral (b, cast), dorsal (c, cast), posterior (d) right lateral (e, g), and ventral (f) views, including magnification of the right pneumatic foramen (‘pleurocoel’) (g). Images of casts are included to highlight particular features that are otherwise difficult to image from the original specimen due to poor contrast.
The centrum is strongly opisthocoelous, with a ball-shaped anterior articular surface and a deeply excavated posterior surface. This condition is present in most allosauroids and megalosauroids, the small coelurosaur Compsognathus, and some alvarezsaurids [2–5]. NMV P221081 differs from most theropods, in which the anterior surface of the centrum is weakly concave [3]. It also differs from many ceratosaurs, tyrannosauroids, Piatnitzkysaurus and Condorraptor, in which the anterior surface is flat [3, 6–7]. Compsognathus and alvarezsaurids that possess similar cervical vertebrae to NMV P221081 are small-bodied taxa, with cervical vertebrae that are much smaller than NMV P221081 [2, 8]. This is especially significant given the juvenile status of NMV P221081. Perhaps more importantly, however, alvarezsaurids and Compsognathus show detailed anatomical differences from NMV P221081, which preclude a close relationship. Alvarezsaurid cervical vertebrae possess several derived features that are not present in NMV P221081 (e.g. carotid processes: ref. 2). Compsognathus cervical vertebrae are considerably more anteroposteriorly elongated (>3·0 times the dorsoventral height) than those of NMV P221081, and lack the detailed similarities shared by NMV P221081 and spinosaurids [8]. NMV P221081 has a maximum anteroposterior length (excluding the anterior articular ball) that is approximately 2·0 times the dorsoventral height of the posterior articular surface. Although this is short when compared to Compsognathus, it is long in comparison to most allosauroids and megalosauroids, among which the longest cervical centra are only slightly longer anteroposteriorly than high dorsoventrally [3, 7, 9–10]. The only exception to this among allosauroids and megalosauroids is in spinosaurid megalosauroids (e.g. Fig. 1). Among allosauroids, Carcharodontosaurus has the proportionally longest cervical centra. These are 1·5 times as long anteroposteriorly as they are high dorsoventrally (ref. 4: anteroposterior measurement excluding the convex anterior surface), substantially less than the ratio in NMV P221081 (2·0). Furthermore, this 'elongate' ratio in Carcharodontosaurus arises from a proportional decrease in height rather than genuine elongation: the centrum has dorsoventrally compressed proportions and is 1·8 times as wide mediolaterally as it is high dorsoventrally. This does not occur in NMV P221081. Furthermore, derived carcharodontosaurids such as Carcharodontosaurus lack substantial dorsoventral offset between the articular surfaces of the centrum [4], also unlike NMV P221081 and spinosaurids (Figs 1, S1b).
In NMV P221081 the lateral surfaces of the centrum are mildly concave, both anteroposteriorly and dorsoventrally, and a strong ridge separates the lateral and ventral surfaces. A small pneumatic foramen (‘pleurocoel’) is positioned in the anterior third of the lateral surface. The presence of a single, anterior pleurocoel is a tetanuran synapomorphy [3, 5]. Most non-tetanuran theropods also possess a posterior pleurocoel [11], although the possible coelophysoid Liliensternus is an exception [3]. The pleurocoel of NMV P221081 has an elliptical outline, with the long axis of the ellipse oriented anterodorsally. This foramen is partially obscured by the diapophysis in lateral view and is therefore only fully visible in lateral view on the right side (which is missing the diapophysis), or in ventrolateral or posterolateral views on the left side. It is subdivided into two unequal portions by a vertical lamina (Fig. S1g). The anterior portion is extremely small on the left side. This condition is also present in some vertebrae of the spinosaurid Baryonyx walkeri (Fig. 1h–i), but has not been seen in any other theropod by any of the authors. Although the anterior pleurocoel is also subdivided in some of the cervical vertebrae of carcharodontosaurian allosauroids [9–10, 12], this is achieved by a massive, buttress that effectively results in the formation of two anterior pleurocoels and is substantially different to the situation in NMV P221081. The prominent parapophysis of NMV P221081 is located anteroventrally on the lateral surface of the centrum, immediately anteroventral to the pleurocoel. The articular surface of the parapophysis is concave and the long axis of this surface is oriented anteroposteriorly. The degree of elongation of the centrum suggests that this is a middle or posterior cervical vertebra, potentially cervical 7 or 8. Indeed, the vertebra is almost identical to cervical 8 of Baryonyx walkeri (Fig. 1a–g) and shares the same character states if scored for a phylogenetic analysis [5].
In ventral view, the centrum is widest anteriorly, at a point level with the parapophyses. It constricts in width posteriorly before expanding again to form the posterior articular surface. The ventral surface is concave anteriorly, but flattens posteriorly. A short longitudinal ridge occupies the centre of the anteroventral concavity. The neural canal is visible in dorsal view, due to the disarticulation and partial breakage of the neural arch. A narrow, longitudinal groove extends along the centre of the neural canal incising the dorsal surface of the centrum. The articular surface for the right side of the neural arch is also exposed in dorsal view. It has a roughened texture, as would be expected if the neural arch were unfused to the centrum (see above). Broken areas on the ventral surface of the centrum reveal that the interior is composed of relatively large internal spaces that may be marrow cavities or large pneumatic chambers (camerae). The interior of the neural arch is composed of compact bone. A highly divided, ‘honey-comb like’ network of small pneumatic spaces (camellae) is absent. This absence precludes referral to Ceratosauria, Carcharodontosauria and most coelurosaurian clades, which have a camellate internal pneumatic structure that is often clearly visible in broken vertebrae [10, 13–14]. The absence of camellae in NMV P221081 is corroborated by the relatively thick external walls of the bone; camellate bones have extremely thin external walls [13]. The presence/absence of this feature cannot be linked to ontogenetic status of the animal (RBJB, pers. obs. on numerous theropod specimens).
The preserved left prezygapophysis is distant from the midline, indicating that the prezygapophyses were widely-spaced, as in tetanurans [3] and some abelisauroids [15], but unlike most non-tetanuran theropods, in which the space between the prezygapophyses is narrower than the neural canal [3]. In lateral view, the prezygapophysis is elongate and possesses a convex dorsolateral surface that flattens posteriorly as it merges with the base of the diapophysis. A well-developed prezygodiapophyseal lamina (prdl: terminology following ref. 16) links these two structures. The articular surface of the prezygapophysis is flat and subovate in medial view: this surface is steeply inclined so that it is oriented at an angle of approximately 80 degrees to the horizontal in anterior view. The diapophysis extends ventrolaterally and forms a triangular sheet in dorsal view: the tip of the diapophysis is missing. A short, but robust posterior centrodiapophyseal lamina (pcdl) is present. The anterior centrodiapophyseal lamina (acdl) is reduced to a very short thin ridge that is only visible in ventral view. Although the postzygapophysis is missing, the base of an extensive postzygodiapophyseal lamina (podl) is present. The pcdl and podl define a deep triangular infrapostzygapophyseal fossa on the lateral surface of the neural arch. This extends dorsomedially into the body of the neural arch and is exposed where the arch is broken dorsally (Fig. S1c). This morphology is consistent with the camerate pneumatic architecture described above, as taxa with camellate internal architecture often have shallower fossae that produce distinct foramina of small or large size that enter into the bone. These foramina are absent in NMV P221081. A stout lamina extends dorsally from the anterodorsolateral corner of the centrum and contacts the prdl at a point approximately halfway between the prezygapophysis and diapophysis, and the remains of a second lamina extends from the same point of origin to a position on the medial surface of the prezygapophysis closes to its base. Taken together, these laminae may represent parts of a bifurcated centroprezygapophyseal lamina (cpzl).
Measurements of NMV P221081
Maximum anteroposterior length of centrum = 42 mm
Maximum mediolateral width of anterior articular surface (including parapophyses) = 18 mm
Maximum mediolateral width of posterior articular surface = 12 mm
Maximum dorsoventral height of posterior articular surface = 17 mm
Similarity Indices
Agnolin et al. [17] used Sorensen’s Similarity Index (SSI) to assess the similarity of Cretaceous dinosaur faunas in terms of the presence/absence of shared taxa. This index can be expressed as:
SSI = 2a/(2a + b + c)
where ‘a’ equals the number of taxa that are common to two faunas, ‘b’ equals the number of taxa present in fauna A, but absent from fauna B, and ‘c’ equals the number of taxa present in fauna B, but which are absent from fauna A. This equation produces a number between 0 and 1, where 1 indicates perfect similarity. This measure is widely used in community ecology [18], but is not particularly appropriate for palaeobiogeographical analyses (see main text: some of these points are expanded upon below). Nevertheless, we used this method to investigate the effects of including/excluding taxa based on the addition of new material and a reinterpretation of some of the data presented in the analyses of Agnolin et al. [17] in order to explore their results. The original results clearly favour close similarity between South America and Australia to the definite exclusion of biogeographical relationships with Laurasian continents [17].
Following the work presented herein, spinosaurids were added to the faunal list for the Australian Cretaceous provided in ref. 17, and tyrannosauroids were also added [19]. Several other changes were made to the faunal lists presented in ref. 17 to bring them into line with more specialist taxonomic reviews, correcting several errors in the original analysis. In our analysis (presented below), stegosaurs are coded as absent from both North and South America [20], ankylosaurs are coded as absent from Africa and India [21], pachycephalosaurs are coded as absent from Europe [22], abelisauroids are coded as absent from Australia [5, 23], and diplodocids are coded as absent from the Cretaceous [24]. In addition, we added the presence of spinosauroids to Asia (on the basis of unnamed material from Thailand and China: ref. 25) and the presence of styracosternans to Africa (on the basis of Ouranosaurus and Lurdusaurus, which are both members of this clade: phylogeny in ref. 26). This produces the following distribution of taxa (Table S1) and the resultant SSIs (Table S2).
Table S1. Modified presence/absence data for Cretaceous dinosaur faunas, based on ref. 17 with the additions and deletions mentioned in the text. Additions of taxa to the Agnolin et al. dataset are shown as a bold red cross; deletions as bold red zeros. Abbreviations: Af, Africa; An, Antarctica; As, Asia; Au, Australia; Eu, Europe; In, India; Ma, Madagascar; NA, North America; SA, South America.
NAAsEuSAAnAfMaInAu
TyrannosauroideaXXX-----X
OrnithomimosauriaXXX------
TherizinosauroideaXX------
OviraptorosauriaXX------
AlvarezsauridaeXXXX-----
TroodontidaeXXX------
DromaeosauridaeXXXXXXX-X
Abelisauroidea--XX-XXX0
Spinosauroidea-XXX-X--X
Neovenatoridae-XXX----X
CarcharodontosauridaeXX-X-X---
Rebbachisauridae--XX-X---
Dicraeosauridae---X-X---
TitanosauriaXXXX-XXXX
Stegosauria0XX0-X---
AnkylosauriaXXXXX0-0X
Basal OrnithopodaXXXXXX--X
StyracosternaXXXXXX--X
PachycephalosauriaXX0------
CeratopsiaXXX------
Table S2. SSIs for Cretaceous dinosaur faunas calculated on the basis of the modified presence/absence data provided in ref. 17 (see Table S1). ‘a’ equals taxa shared between the two continents; ‘b’ equals taxa unique to Australia; ‘c’ equals taxa unique to the other continent.
Continent / 2a / b / c / SSINorth America / 12 / 2 / 9 / 0·55
Asia / 16 / 0 / 9 / 0·67
Europe / 16 / 0 / 8 / 0·70
South America / 14 / 1 / 5 / 0·70
Antarctica / 8 / 4 / 0 / 0·67
Africa / 10 / 3 / 6 / 0·52
Madagascar / 4 / 6 / 1 / 0·36
India / 2 / 7 / 1 / 0·20
Although South American faunas are recovered as similar to those of Australia (as concluded in ref. 17), it is noteworthy that the SSI for Europe is indistinguishable from that for South America, and that obtained for Asia is also similar. All SSIs for comparisons between Australia and Laurasian continents increase relative to the values obtained in ref. 17 following these minor changes (also helping to demonstrate the sensitivity of this index to issues of taxon sampling: see main text), whereas values for most Gondwanan continents (India, Africa, Antarctica and Madagascar) are reduced.
Following this analysis, we restricted comparisons to taxa of Early–‘middle’ Cretaceous age (Neocomian and Gallic in old stratigraphic nomenclature, comprising the Berriasian to Turonian stages), in order to exclude taxa known only from the Late Cretaceous and thereby make comparisons between the late Early Cretaceous Australian fauna and those from elsewhere more relevant. Unfortunately, this precludes comparisons between Australia and several Gondwanan continents (Antarctica, India and Madagascar), as these lack dinosaur faunas of appropriate age [27]. Similarly, this led to the exclusion of pachycephalosaurs from the analysis, as their entire record post-dates the Turonian stage [28]. The restricted version of the dataset is presented in Table S3 and the results of the new analyses are in Table S4.
Table S3. Presence/absence data for Early to ‘middle’ Cretaceous dinosaur faunas, based the data presented in Table S1, pruned to exclude areas and taxa not present in these time slices (see text). Abbreviations: Af, Africa; As, Asia; Au, Australia; Eu, Europe; NA, North America; SA, South America. Changes to the original matrix in ref. 17 shown in bold red type.
NAAsEuSAAfAu
TyrannosauroideaXXX--X
OrnithomimosauriaXXX---
TherizinosauroideaXX----
OviraptorosauriaXX----
TroodontidaeXXX---
DromaeosauridaeXXXXXX
Abelisauroidea--XXX0
Spinosauroidea-XXXXX
Neovenatoridae-XXX-X
CarcharodontosauridaeXX-XX-
Rebbachisauridae--XXX-
Dicraeosauridae---XX-
TitanosauriaXXXXXX
Stegosauria0XX0X-
AnkylosauriaXXX--X
Basal OrnithopodaXXXXXX
StyracosternaXXXXXX
CeratopsiaXX----
Table S4. SSIs for Early to ‘middle’ Cretaceous dinosaur faunas calculated on the basis of the pruned dataset presented in Table S3. ‘a’ equals taxa shared between the two continents; ‘b’ equals taxa unique to Australia; ‘c’ equals taxa unique to the other continent.
Continent / 2a / b / c / SSINorth America / 12 / 2 / 6 / 0·60
Asia / 16 / 0 / 6 / 0·73
Europe / 16 / 0 / 5 / 0·76
South America / 12 / 2 / 4 / 0·67
Africa / 10 / 3 / 5 / 0·56
Analysis of this restricted dataset underscores the strong similarities between the Early–‘middle’ Cretaceous faunas of Australia and those of Laurasian continents (see main text). In addition, several other dinosaur clades have been proposed as members of the Early Cretaceous Australian dinosaur fauna, including oviraptorosaurian and ornithomimosaurian theropods [29–30] and the possible ceratopsian Serendipaceratops [29, 31]. The systematics of these specimens were disputed in ref. 17, and they were not included in either the SSI analyses presented by these authors, or in the modified versions of these analyses that are presented above. Discussion of the affinities of these specimens lies outside of the scope of this paper and will be addressed elsewhere (RBJB, THV, PV-R and others, in preparation), but it should be noted that the Australian oviraptorosaur, ornithomimosaur and ceratopsian were identified as a dromaeosaurid, ?unenlagiinine dromaesaurid and an indeterminate ornithischian, respectively, in ref. 17. However, if the presence of either an oviraptorosaur, ornithomimosaur or ceratopsian is confirmed in Australia, each of them would lead to an increase in the SSI values obtained from comparisons between the Australian biota and those from North America, Europe and Asia: the converse would be true for the SSI values obtained from comparisons with Gondwanan continents.
References
1 Varricchio, D. J. 1997 Growth and embryology. In Encyclopedia of Dinosaurs (eds P. J. Currie & K. Padian), pp. 282–288. San Diego, United States: Academic Press.
2 Chiappe, L. M., Norell, M. A. & Clark, J. M. 2002 The Cretaceous, short-armed Alvarezsauridae. In Mesozoic Birds (eds L. M. Chiappe & L. M. Witmer), pp 87–120. Berkeley, United States: University of California Press.
3 Rauhut, O. W. M. 2003. The interrelationships and evolution of basal theropod dinosaurs. Spec. Pap. Palaeontol. 69, 1–213.
4 Brusatte, S. L. & Sereno P. C. 2007 A new species of Carcharodontosaurus (Dinosauria: Theropoda) from the Cenomanian of Niger and a revision of the genus. J. Vert. Paleontol.27, 902–916.
5 Benson, R. B. J., Carrano, M. T. & Brusatte, S. L. 2010 A new clade of archaic large-bodied predatory dinosaurs (Theropoda: Allosauroidea) that survived to the latest Mesozoic. Naturwissenschaften97, 71–78 (doi: 10.1007/s00114-009-0614-x).
6 Brochu, C. A. 2002 Osteology of Tyrannosaurus rex: insights from a nearly complete skeleton and high-resolution computed tomographic analysis of the skull. Mem. Soc. Vert. Paleontol. 7, 1–138.
7 Rauhut, O. W. M. 2005. Osteology and relationships of a new theropod dinosaur from the Middle Jurassic of Patagonia. Palaeontology48, 87–110.
8 Peyer, K. 2006 A reconsideration of Compsognathus from the Upper Tithonian of Canjeurs, southeastern France. J. Vert. Paleontol.26, 879–896.
9 Harris, J. D. 1998 Reanalysis of Acrocanthosaurus atokensis, its phylogenetic status and implications, based on a new specimen. Bull. New Mexico Mus. Nat. Hist. Sci. 13, 1–75.
10 Brusatte, S. L., Benson, R. B. J. & Hutt, S. 2008 The osteology of Neovenator salerii (Dinosauria: Theropoda) from the Wealden Group (Barremian) of the Isle of Wight. Palaeontol. Soc. Monographs162(631), 1–75.
11 Tykoski, R. S. & Rowe, T. 2004 Ceratosauria. In The Dinosauria (2nd Edition) (eds D. B. Weishampel, P. Dodson & H. Osmólska), pp. 47–70. Berkeley, United States: University of California Press.
12 Smith, N. D., Makovicky, P. J., Hammer, W. R. & Currie, P. J. 2007 Osteology of Cryolophosaurus ellioti (Dinosauria: Theropoda) from the Early Jurassic of Antarctica and implications for early theropod evolution. Zool. J. Linn. Soc.151, 377–421.
13 Britt, B. B. 1993 Pneumatic postcranial bones in dinosaurs and other archosaurs. PhD thesis, University of Calgary.
14 Sues, H.-D. 1997 On Chirostenotes, a Late Cretaceous oviraptorosaur (Dinosauria: Theropoda) from western North America. J. Vert. Paleontol.17, 698–716.