Revision of the Sauropod Dinosaur Diamantinasaurus Matildae Hocknull Et Al

GR-01247; No of Pages 39 Gondwana Research xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Gondwana Research

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

Revision of the sauropod dinosaur Diamantinasaurus matildae Hocknull et al. 2009 from the mid-Cretaceous of Australia: Implications for Gondwanan titanosauriform dispersal☆

Stephen F. Poropat a,b,⁎, Paul Upchurch c, Philip D. Mannion d,ScottA.Hocknulle,BenjaminP.Keara, Trish Sloan b, George H.K. Sinapius b, David A. Elliott b a Department of Earth Sciences, Uppsala University, Uppsala, Sweden b Australian Age of Dinosaurs Museum of Natural History, The Jump-Up, Winton, Queensland, Australia c Department of Earth Sciences, University College London, Gower Street, London WC1E 6BT, United Kingdom d Department of Earth Science and Engineering, Imperial College London, South Kensington Campus, London SW7 2AZ, United Kingdom e Geosciences, Queensland Museum, Hendra, Queensland, Australia article info abstract

Article history: The osteology of Diamantinasaurus matildae, the most complete Cretaceous sauropod described from Australia to Received 28 October 2013 date, is comprehensively reassessed. The preparation of additional material from the type locality, pertaining to Received in revised form 17 March 2014 the same individual as the holotype, sheds light on the morphology of the axial skeleton and provides additional Accepted 17 March 2014 information on the appendicular skeleton. The new material comprises two dorsal vertebrae, an incomplete Available online xxxx sacrum (including four partial coalesced vertebrae), the right coracoid, the right radius, an additional manual ﬁ Keywords: phalanx, and a previously missing portion of the right bula. In this study we identify thirteen autapomorphic ﬁ Lithostrotia characters of Diamantinasaurus, and an additional ve characters that are locally autapomorphic within Titanosauria Titanosauriformes. This work provided an opportunity to revisit the phylogenetic placement of Diamantinasaurus. Cretaceous In two independent data matrices, Diamantinasaurus was recovered within Lithostrotia. One analysis resolved Australia Diamantinasaurus as the sister taxon to the approximately coeval Tapuiasaurus from Brazil, whereas the second Palaeobiogeography analysis recovered Diamantinasaurus as the sister taxon to Opisthocoelicaudia from the latest Cretaceous of Mongolia. The characters supporting the recovered relationships are analysed, and the palaeobiogeographical implications of the lithostrotian status of Diamantinasaurus are explored. A brief review of the body fossil record of Australian Cretaceous terrestrial vertebrates suggests close ties to South America in particular, and to Gondwana more generally. © 2014 The Authors. Published by Elsevier B.V. on behalf of International Association for Gondwana Research. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

1. Introduction directly overlying the Diamantinasaurus type specimen, indicating a maximum age for this layer of 95 Ma; therefore, the underlying Diamantinasaurus matildae is the most complete Cretaceous sauro- bonebed has a minimum age of 95 Ma. AODL 85 sits on the western pod described from Australia to date (Hocknull et al., 2009); the only side of the Cork Fault, east of which the sediments of the Winton Forma- non-Cretaceous sauropod skeleton known from Australia, the Jurassic tion have been upthrust by at least 100 m (Vine and Jauncey, 1964). Rhoetosaurus brownei Longman 1926, is slightly more complete This would appear to place AODL 85 higher within the Winton Forma- (Longman, 1927; Nair and Salisbury, 2012). The type locality for tion than the dinosaur-bearing fossil sites considered by Tucker et al. Diamantinasaurus is AODL 85 (the “Matilda Site”), which is located (2013) to be uppermost Cenomanian to lowermost Turonian. Tucker west-northwest of Winton, central Queensland, in the northeast of et al. (2013) provided detrital zircon dates for a range of fossil-bearing Australia (Fig. 1). The sediments of AODL 85 pertain to the Winton For- sites, including those at Bladensburg National Park and Lark Quarry, mation (Fig. 2), and were dated as Cenomanian by Bryan et al. (2012). which occur on the eastern side of the Cork Fault. However, the sites These authors provided detrital zircon dates for the sandstone unit considered by Tucker et al. (2013) are topographically higher than AODL 85, and represent a slightly stratigraphically younger succession within the Winton Formation than the beds at the Diamantinasaurus ☆ This article belongs to the Special Issue on Gondwanan Mesozoic biotas and bioevents. type locality. ⁎ Corresponding author at: Department of Earth Sciences, Uppsala University, Uppsala, ﬁ – Sweden. Excavations at AODL 85 took place over ve dig seasons (2006 2010) E-mail address: [email protected] (S.F. Poropat). and led to the discovery of two dinosaurs, isolated crocodylomorph bones http://dx.doi.org/10.1016/j.gr.2014.03.014 1342-937X/© 2014 The Authors. Published by Elsevier B.V. on behalf of International Association for Gondwana Research. This is an open access articleundertheCCBY-NC-NDlicense (http://creativecommons.org/licenses/by-nc-nd/3.0/).

Fig. 1. Map of Queensland, Australia showing the location of the town of Winton and the distribution of Cretaceous sedimentary strata surficial exposures in Queensland. and teeth, and numerous freshwater molluscs (Fig. 3; Hocknull et al., the identification of the glenoid fossa and coracoid foramen, have demon- 2009). Deposition of the bones most likely occurred within an abandoned stratedthatitispartoftherightcoracoidofDiamantinasaurus. The ele- channel system (Hocknull et al., 2009). Several large siltstone and iron- ment previously identified as the left sternal plate of Diamantinasaurus stone concretions were extracted from within the deposit, and additional seems to represent the missing dorsal margin of this element. This reas- remains of the two dinosaurs were preserved within. Preparation of the sessment, if correct, would mean that sternal plates are currently not rep- concretions, and material in the remaining plaster jackets, has yielded resented in the holotype of Diamantinasaurus. The right coracoid was additional sauropod and theropod elements to those originally described found deep within the accumulation, but close to the right scapula by Hocknull et al. (2009). These new elements are assignable to (Fig. 3); in contrast, the ‘sternal plate’ section was found towards the the holotype individuals of Diamantinasaurus (AODF 603) and upper-most part of the accumulation and approximately 3 m to the Australovenator wintonensis Hocknull et al. 2009 (AODF 604). The sauro- south. Both the right scapula and the right coracoid show signs of pod remains from the site derive from a similarly-sized sauropod and crushing and dislocation. The mechanism for the distortion and displace- do not duplicate any of the elements previously catalogued as AODF ment of these elements in a low-energy depositional environment is sug- 603; therefore, it is clear that these bones derive from the same individual gested to be dinoturbation of the sediment, probably by other sauropods. as the holotype of Diamantinasaurus. The new elements pertaining to When first described, Diamantinasaurus was known mostly from Australovenator have been described elsewhere (White et al., 2012, appendicular elements, the only axial elements present being ribs. 2013). Excavations in 2010 recovered a single bone unusually deep with- Most elements in the pectoral girdle and forelimb were represented in the accumulation. It was preserved within a fossil-poor grey-blue basal from at least one side of the body, with the notable exception of the cor- siltstone layer that directly underlies the main bone accumulation. When acoid and the radius (Hocknull et al., 2009). Following the preparation of found, the bone's position and preservation suggested that it may have new material, the right coracoid, right radius and manual phalanx IV-1 come from a third dinosaur; however, reanalysis of this element, and are now assigned to Diamantinasaurus as paratypes (following the

Fig. 2. Stratigraphy of the Eromanga Basin, central Queensland, with silhouettes representing known tetrapod taxa.

International Commission on Zoological Nomenclature, 1999); as stated redescription of the previously reported but only briefly described sacral above, the element originally described as the ‘sternal plate’ appears to processes, greatly increase our understanding of the axial skeleton of represent the dorsal portion of the right coracoid. The fibula was missing Diamantinasaurus. a portion of the shaft when described and figured by Hocknull et al. In this paper, we provide a comprehensive redescription of the (2009); this has since been identified and re-attached. Perhaps most im- Diamantinasaurus holotype material, describe new paratype material, portantly, two dorsal vertebrae and four coalesced partial sacral verte- reassess the phylogenetic position of the genus, and examine the impli- brae have also been recovered from AODL 85. These vertebrae, and the cations of our improved knowledge of middle Cretaceous Australian

Fig. 3. Map of AODL 85 (the “Matilda Site”), showing the approximate relative positions of the holotype and paratype specimens of Diamantinasaurus matildae. Non-shaded specimens mostly pertain to Australovenator wintonensis. Abbreviations: ast, astragalus; c, coracoid; dor A, dorsal vertebra A; dor B, dorsal vertebra B; f, ﬁbula; fem, femur; h, humerus; il, ilium; is, ischium; m, metacarpal(s); p, pubis; r, dorsal rib; rad, radius; s, scapula; sac, sacral vertebrae; tib, tibia; u, ungual; ul, ulna. Scale bar = 1 m.

Please cite this article as: Poropat, S.F., et al., Revision of the sauropod dinosaur Diamantinasaurus matildae Hocknull et al. 2009 from the mid-Cretaceous of Australia: Implications for Gondwanan..., Gondwana Research (2014), http://dx.doi.org/10.1016/j.gr.2014.03.014 4 S.F. Poropat et al. / Gondwana Research xxx (2014) xxx–xxx sauropods for our understanding of titanosauriform palaeobiogeography central Queensland, the protostegid turtles Cratochelone Longman during the later stages of the break-up of Gondwana. 1915, Notochelone (Owen 1882) Lydekker 1889,andBouliachelys Kear and Lee 2006, the huge pliosaur Kronosaurus Longman 1924,and 2. Palaeogeographic and palaeobiogeographic setting the elasmosaur Eromangasaurus Kear 2007,areallfoundinthe lower upper Albian Toolebuc Formation. The ubiquitous ichthyosaur The middle Cretaceous was a tumultuous time during Earth's history, Platypterygius australis (McCoy 1867) McGowan 1972 has been identi- characterised by the continued break-up of Gondwana (Jokat et al., fied in both the Toolebuc Formation and the overlying Allaru Mudstone, 2003; König and Jokat, 2006). Rifting between South America and and an undescribed polycotylid derives from the upper Albian Mackunda Africa, initiated in the Late Jurassic (Mohriak et al., 2008), facilitated Formation (Kear, 2003). brief, intermittent marine incursions into the proto-South Atlantic The aforementioned Albian formations (Toolebuc, Allaru, Mackunda) from the North from as early as the late Barremian (Chaboureau et al., of central Queensland, though deposited mostly in marine settings, have 2013)orearlyAptian(Koutsoukos et al., 1991; Koutsoukos, 1992)and all produced dinosaurs: fragmentary sauropod remains (Molnar, 2001; led to the establishment of a permanent marine connection between Molnar and Salisbury, 2005) and enantiornithine birds (Kurochkin and the closing Tethys Ocean to the North and the Weddell Sea in the late Molnar, 1997), including Nanantius Molnar 1986, are known from Albian (Azevedo, 2004). The shift from non-marine to marine conditions the Toolebuc Formation; the sauropod Austrosaurus mckillopi Longman in the proto-South Atlantic, with an initial Tethyan-only and later 1933, ornithopod material assigned to Muttaburrasaurus (Molnar, Tethyan-mixed-with-south-polar signature, is reflected in the echinoid 1996a), and an almost complete ankylosaur referred to Minmi (Molnar, (Néraudeau and Mathey, 2000), ostracod (Dingle, 1999), foraminiferan 1996b), were found in the Allaru Mudstone; and the holotype of (Koutsoukos et al., 1991; Koutsoukos, 1992) and ammonite (Kennedy Muttaburrasaurus langdoni Bartholomai and Molnar 1981 derives from and Cooper, 1975) fossil records. Magnetic data show that Madagascar the Mackunda Formation. The uppermost Albian-lower Turonian-aged had separated from eastern Africa (specifically Kenya and Somalia) by Winton Formation, deposited during and after the final retreat of the in- the mid-Cretaceous (Rabinowitz et al., 1983)butremainedattachedto land sea (Bryan et al., 2012; Tucker et al., 2013), has most famously India (with Mauritius and other modern day Indian Ocean islands) yielded numerous dinosaur footprints, including those that make up the well into the Late Cretaceous (Raval and Veeraswamy, 2003; Torsvik world famous Lark Quarry Dinosaur Stampede (Thulborn and Wade, et al., 2013). The effects of the breakup of Gondwana on the palaeoflora 1979, 1984), which has been the subject of two recent papers (Romilio were explored by McLoughlin (2001). and Salisbury, 2011; Romilio et al., 2013) presenting an alternative Throughout the Cretaceous, Australia was connected to Antarctica interpretation of the site, the first of which has been refuted by (Veevers et al., 1991; Betts et al., 2002), though by the end of the Thulborn (2013). A broad range of invertebrates and plants dominate Maastrichtian the land bridge between the two appears to have been macrofossil assemblages from the Winton Formation (Dettmann et al., limited to a relatively narrow strip corresponding to modern day 1992; McLoughlin et al., 1995, 2010). Scattered vertebrate remains Tasmania (Veevers, 2006). During the late Early Cretaceous, much of include the derived neosuchian crocodylomorph Isisfordia Salisbury southern Australia was located within the Antarctic Circle and would et al. 2006,adolichosaur(Scanlon and Hocknull, 2008), a small ornitho- have experienced winters characterised by long periods of darkness pod (Hocknull and Cook, 2008), an ankylosaur (Leahey and Salisbury, and, quite possibly, very low to freezing temperatures (Constantine 2013), Australia's most complete non-avian theropod Australovenator et al., 1998; Vickers-Rich et al., 1999). In spite of these challenging (Hocknull et al., 2009; White et al., 2012, 2013), and abundant sauropods conditions, the Victorian sediments of the Otway and Strzelecki groups (Coombs and Molnar, 1981; Molnar, 2001, 2011; Molnar and Salisbury, (Rich and Vickers-Rich, 2012a) are dominated by small ornithopods 2005) including Wintonotitan wattsi Hocknull et al. 2009 and (Rich et al., 1989) including Leaellynasaura Rich and Rich 1989, Diamantinasaurus, the focus of this paper. A revision of Wintonotitan Atlascopcosaurus Rich and Rich 1989 and Qantassaurus Rich and Vickers- will shortly be published (Poropat et al., in press), wherein it is com- Rich 1999, alongside which lived ankylosaurs (Barrettetal.,2010), the pared in detail with Diamantinasaurus; rather than duplicating this ef- possible neoceratopsian Serendipaceratops Rich and Vickers-Rich 2003 fort in this paper, we direct the reader to this forthcoming publication. (Rich et al., in press), and a range of theropods including ceratosaurs (Fitzgerald et al., 2012), spinosaurids (Barrett et al., 2011), neovenatorid 2.1. Institutional abbreviations allosauroids (Smith et al., 2008), tyrannosauroids (Benson et al., 2010a, 2010b;thoughseeHerne et al., 2010), and maniraptorans including pos- AAOD: Australian Age of Dinosaurs Natural History Museum, sible oviraptorosaurs (Currie et al., 1996), dromaeosaurs (Benson et al., Winton, Queensland, Australia; AODF: Australian Age of Dinosaurs 2012) and enantiornithine birds (Close et al., 2009). The chigutisaurid Fossil; AODL: Australian Age of Dinosaurs Locality; LHV: Long Hao temnospondyl Koolasuchus Warren et al. 1997, the cryptodiran turtle Geologic and Paleontological Research Center, Nei Mongol, China; Otwayemys Gaffney et al. 1998, plesiosaurs (Benson et al., 2013), NHMUK: Natural History Museum, London, United Kingdom; NMV: crocodylomorphs (Vickers-Rich, 1996) and a wide variety of mammals National Museum of Victoria, Melbourne, Victoria, Australia; QM F: (including Ausktribosphenos Rich et al. 1997, Teinolophos Rich et al. Queensland Museum (Brisbane, Australia) Fossil. 1999a, Bishops Rich et al. 2001, Kryoryctes Pridmore et al. 2005 and Corriebaatar Rich et al. 2009) are also represented in the fauna. 3. Systematic palaeontology Before, during, and after the deposition of the Victorian dinosaur- bearing sediments, a significant percentage of the Australian continent Dinosauria Owen, 1842 was covered by an inland sea (Frakes et al., 1987). The extent and Saurischia Seeley, 1887 depth of this sea varied over the span of its 30 million year existence Sauropodomorpha von Huene, 1932 (Gallagher and Lambeck, 1989; Campbell and Haig, 1999), lasting Sauropoda Marsh, 1878 from the Valanginian until the late Albian (Cook and McKenzie, 1997). Eusauropoda Upchurch, 1995 During this time, a range of marine reptiles inhabited the centre of Neosauropoda Bonaparte, 1986 Australia (Kear, 2003). The lower Aptian to lower Albian Bulldog Shale Macronaria Wilson and Sereno, 1998 of South Australia, interpreted to have been deposited in cool marine Titanosauriformes Salgado et al., 1997a conditions based on the presence of glendonites (Sheard, 1990; De Somphospondyli Wilson and Sereno, 1998 Lurio and Frakes, 1999) and lonestones (Frakes et al., 1995), has yielded Titanosauria Bonaparte and Coria, 1993 plesiosaurs, including the rhomaleosaurid Umoonasaurus Kear et al. Lithostrotia Upchurch et al., 2004 2006 and the aristonectid Opallionectes Kear 2006. Further north, in Diamantinasaurus matildae Hocknull et al., 2009

Holotype: AODF 603: Three partial cervical ribs, fragmentary visible in posterior, lateral and ventral views. Within Titanosauriformes, gastralia, dorsal ribs, two isolated sacral processes, right scapula, right local autapomorphies of Diamantinasaurus are: (1) scapular glenoid ar- and left humeri, right ulna, left metacarpal I, right metacarpals II–V, ticular face laterally bevelled; (2) metacarpal I distal condyle medially four manual phalanges (including manual ungual I-2), left ilium, right bevelled; (3) manual ungual phalanx I-2 proximal height: total length and left pubes, right and left ischia, right femur, right tibia, right fibula, ratio = 0.4; (4) ilium with prominent lateral protuberance near base right astragalus. of ischiadic articular region (if Diamantinasaurus and Opisthocoelicaudia Paratypes: AODF 603: two incomplete dorsal vertebrae, four are sister taxa, this is instead a synapomorphy of that clade); and (5) fib- coalesced sacral vertebrae with bases of two sacral processes, right ula with medial triangular scar at proximal end. coracoid (including the holotype element formerly identified as a sternal plate), right radius, one manual phalanx. 4. Description Type locality: AODL 85 (the “Matilda Site”), Elderslie Sheep Station, approximately 60 km west-northwest of Winton, west-central Queens- Descriptions of the cervical ribs, dorsal and sacral vertebrae, sacral land, Australia (Fig. 1). processes, right coracoid, right radius and right manual phalanx IV-1 Formation and age: Winton Formation (Fig. 2), latest Albian– are provided for the first time. The original descriptions of the previous- Turonian, Cretaceous; AODL 85 specifically is of Cenomanian age ly reported skeletal elements (Hocknull et al., 2009), though generally (Bryan et al., 2012). accurate, are thoroughly revised and additional details on each element Comments on original diagnosis: Of the 40 features listed in are provided. Major emendations are only required for a few bones: the the original diagnosis, Hocknull et al. (2009) identified five as bone originally described as the left sternal plate is reinterpreted herein autapomorphic: 1) humerus with intermediate robusticity; 2) manual to represent the dorsal margin of the right coracoid; manual phalan- phalanx III-1 heavily reduced; 3) femur with intermediate robusticity; gesIII-1andV-1werepreviouslymisidentified as IV-1 and III-1 4) tibia cnemial crest projects anteriorly rather than laterally; and respectively; and the right fibula was found to be missing a portion 5) fibula with intermediate robusticity. Of these, three relate to the of the shaft, which has since been added to the specimen. robusticity of limb bones and, in the light of the large dataset on continuous characters for Titanosauriformes presented by Mannion et al. 4.1. Axial skeleton (2013) such features no longer uniquely diagnose Diamantinasaurus. We suggest that Hocknull et al. (2009) erred in their identification of 4.1.1. Cervical ribs (Supplementary Table 1) the manual phalanges: another manual phalanx was present in the Three cervical ribs from Diamantinasaurus have been identified to same block as the right metacarpals and phalanges, intermediate in date. None is complete; however, potentially autapomorphic character- size between those identified as II-1 and IV-1 by Hocknull et al. istics can be identified in the least incomplete specimen. (2009). This suggests that the phalanx that Hocknull et al. (2009) identified as III-1 is actually V-1, and that their IV-1 is in fact III-1. The 4.1.1.1. Cervical rib A (Fig. 4A–F). The best-preserved cervical rib, from remaining character identified as a potential autapomorphy only pro- the left side, retains the broken base of the tuberculum, a large part of vides weak support for the validity of Diamantinasaurus, since the tibiae the capitulum, the anterior process and the base of the distal shaft. In of several other sauropods also have anteriorly projecting cnemial the following description, it is presumed that the capitulum projected crests (see Mannion et al., 2013); furthermore, there is always the dorsomedially towards the cervical parapophysis. The anterior process possibility that the orientations of structures (in this case, the cnemial is relatively short and bluntly rounded in dorsal view. This process is crest) have been affected by post-mortem distortion. Nevertheless, formed from a transversely oriented sheet of bone that curves upward Diamantinasaurus is clearly a distinct sauropod taxon and we support along its lateral and medial margins so that its dorsal surface is concave this view below by providing a revised diagnosis based on newly iden- and its ventral surface is correspondingly convex. Cervical ribs with tified autapomorphies. similarly robust, but more anteriorly extended anterior projections Revised diagnosis: Diamantinasaurus can be diagnosed on the basis are evident in Trigonosaurus Campos et al. 2005, Phuwiangosaurus of the following autapomorphies: (1) cervical rib distal shafts lack a dor- Martin et al. 1994 (Martin et al., 1999; Suteethorn et al., 2009)and sal midline trough and instead possess a laterodistally directed ridge on Malawisaurus Jacobs et al. 1993 (Gomani, 2005). The medial edge the dorsal surface; (2) posterior centroparapophyseal lamina (PCPL) bi- of the anterior process is linked to the anterior base of the capitulum furcated dorsally in the middle–posterior dorsal vertebrae; (3) accesso- by a stout ridge. The capitulum is well-developed and dorsoventrally ry longitudinal ridge and fossa at mid-length on the lateral surface of the compressed. A stout sheet of bone extends from the lateral face of the scapular blade, lying dorsal to the main lateral ridge; (4) the posterolat- capitulum to the medial face of the main rib at the base of the eral margin of the proximal humeral shaft is formed by a stout vertical tuberculum. In medial view, this transverse sheet slopes anterodorsally, ridge that increases the depth of the lateral triceps fossa; (5) a ridge ex- creating a deep fossa at the posterior end of the ‘U’-shaped concavity on tends medially from the humeral deltopectoral crest and then turns to the dorsal surface of the anterior process. It also defines a medially fac- extend proximally, creating a fossa lying medial to the dorsal part of ing ovate fossa (tapering posteriorly) that occupies the posterior two the deltopectoral crest on the anterior face of the humerus; (6) pubis thirds of the medial surface of the capitulum; this is a different structure obturator foramen region with grooves and ridges on lateral surface; to the pneumatic fossa identified in the cervical ribs of Rapetosaurus (7) femur with shelf linking posterior ridges of fibular condyle; Curry Rogers and Forster 2001 (Curry Rogers, 2009). The lateral margin (8) proximal lateral face of tibia with double ridge extending distally of the main shaft is transversely thick as it extends from the tubercular from lateral projection of proximal articular area; (9) tibia with postero- region, but thins distally. The base of the main shaft is a transversely lateral fossa posterior to the double ridge, containing a lower tuberosity compressed sheet of bone, and does not seem to form a trough along and an upper deep pit; (10) anterolateral margin of tibia shaft, its dorsal surface. This condition contrasts with that seen in many distal to cnemial crest, forms a thin flange-like projection extending other sauropods, including Mamenchisaurus hochuanensis Young and proximodistally along the central region of the element; (11) medial Zhao 1972 and Mamenchisaurus sinocanadorum Russell and Zheng surface of fibula shaft, between proximal triangular scar and 1993 (PU pers. obs.), Euhelopus zdanskyi (Wiman 1929) Romer 1956 mid-length, with vertical ridge separating anterior and posterior (SFP pers. obs.), and the titanosaurs Trigonosaurus (Campos et al., grooves; (12) lateral fossa of astragalus divided into upper and 2005) and Rapetosaurus (Curry Rogers, 2009), in which the proximal lower portions by anteroposteriorly directed ridge; and (13) postero- part of the distal shaft possesses a longitudinal groove on its dorsal ventral margin of astragalus, below and medial to the ascending pro- surface bounded by lateral and medial walls. The absence of this groove cess, with well-developed, ventrally projecting rounded process in Diamantinasaurus is thus potentially autapomorphic. The posterior

Fig. 4. Holotype cervical ribs of Diamantinasaurus matildae: cervical rib A (left) in A, anterior; B, dorsal; C, left lateral; D, ventral; E, posterior; and F, medial views; and cervical rib B (right) in G, dorsal; H, right lateral; I, anterior; J, medial; K, ventral; and L, posterior views.

edge of the capitulum gives rise to a ridge that does not merge into the preservation of the neural arches in both specimens makes determina- medial margin of the main distal shaft; instead it extends onto tion of their serial position within the vertebral column difﬁcult; that the dorsal surface and slants a little laterally over a distance of said, the location of the parapophyses with respect to the centrum and approximately 80 mm. This ridge is also potentially autapomorphic the diapophyses suggests that these vertebrae are either middle or pos- for Diamantinasaurus. In ventromedial view, the base of the distal terior dorsal vertebrae. Nomenclature for vertebral laminae and fossae shafthasaslightsigmoidcurvelikethatinApatosaurus louisae follows that of Wilson (1999) and Wilson et al. (2011). Holland 1915 (Gilmore, 1936), ﬁrst extending slightly dorsally and mainly distally, then slightly ventrally and mainly distally. 4.1.2.1. Dorsal vertebra A (Fig. 5A–I). Dorsal vertebra A of Diamantinasaurus is reasonably complete and well-preserved, although 4.1.1.2. Cervical rib B (Fig. 4G–L). This specimen is part of a right cervical the posterior surface has been subjected to some weathering; this and rib, preserving much of the distal shaft, but lacking most of the anterior other broken surfaces show that the internal bone of both centrum process, tuberculum and capitulum. It displays a similar morphology to and neural arch are camellate. The right prezygapophysis has been bro- the left cervical rib described above in that it is ventrally convex (in the ken into medial and lateral portions, resulting in the ventral displace- transverse plane), dorsally concave between the bases of the capitulum ment of the lateral part of this region, and meaning that the right and tuberculum, and preserves a ridge on the dorsal surface of the distal parapophysis is currently positioned more ventrally than it would shaft. The capitulum is very poorly preserved but was clearly a stout have been in life. process. The base of the distal shaft is sub-triangular in transverse The anteroposteriorly elongate centrum is strongly opisthocoelous. cross-section, with wider ventrolateral and ventromedial faces, and a In anterior view, the condyle is heart-shaped: a medial depression narrower dorsal surface. Thus the rib is moderately compressed trans- aligned with the neural canal is present on the dorsal margin, and the versely. The shaft curves slightly dorsally in lateral view and becomes condyle narrows towards its ventral margin. However, the ventral even more compressed transversely towards its distal end. narrowing may be exaggerated as a result of missing bone along the ventrolateral surfaces. The posterior cotyle is also heart-shaped, with a 4.1.1.3. Cervical rib C. This portion of a cervical rib is fragmentary, repre- prominent notch aligned with the neural canal present on the dorsal sented only by a part of the distal process. Little additional information surface of the cotyle, similar to that seen in Europasaurus Sander et al. on the cervical rib morphology of Diamantinasaurus can be obtained 2006 (Carballido and Sander, 2014)andEuhelopus (Wiman, 1929). from this specimen. The centrum is slightly wider transversely than dorsoventrally tall (Supplementary Table 2), lacking the very strong dorsoventral 4.1.2. Dorsal vertebrae (Supplementary Table 2) compression of posterior dorsal centra seen in Giraffatitan Paul Two dorsal vertebrae have been recovered from Diamantinasaurus. 1988, and the derived titanosaurs Epachthosaurus Powell 1990 and Both vertebrae, though incompletely preserved, provide an abundance Opisthocoelicaudia Borsuk-Białynicka 1977 (Martínez et al., 2004; of anatomical information. They will be referred to in the following de- Taylor, 2009; Mannion et al., 2013). The ventral face is concave both scription as dorsal vertebra A and dorsal vertebra B, since the poor anteroposteriorly and transversely, with thin ridges of bone extending

Fig. 5. Paratype dorsal vertebra (dorsal vertebra A) of Diamantinasaurus matildae in A, anterior; B, ventral; C; left lateral; D, dorsal; E, posterior; and F, right lateral views; and restored schematic of dorsal vertebra A in G, anterior; H, left lateral; and I, posterior views.

along the entire length of the ventrolateral margins. This central extended laterally and that these laminae effectively formed the antero- concavity is divided longitudinally by a sagittal ridge. Ventral keels are lateral margins of the neural spine (Fig. 5G–H). SPRLs are absent in the knowninthedorsalvertebraeofseveral diplodocoids and brachiosaurids, dorsal vertebrae of Opisthocoelicaudia and the majority of the middle- and the somphospondylans Barrosasaurus Salgado and Coria 2009, posterior dorsal vertebrae of Rapetosaurus (Curry Rogers, 2009). Opisthocoelicaudia (Borsuk-Białynicka, 1977), and Phuwiangosaurus Only the right postzygapophysis is preserved and, based on its (Upchurch et al., 2004; Curry Rogers, 2005; Mannion and Otero, 2012; location relative to the neural canal, the postzygapophyses appear Mannion et al., 2013). However, only Opisthocoelicaudia shares to have been situated further apart from one another than the with Diamantinasaurus the presence of a ventral keel set within a ridge- prezygapophyses. The left centropostzygapophyseal lamina (CPOL) is bound concavity (Borsuk-Białynicka, 1977). also well defined. This lamina creates a longitudinal postzygapophyseal The dorsal half of the lateral face of the centrum is excavated by an centrodiapophyseal fossa (POCDF), trending ventromedially between anteroposteriorly elongate pneumatic fossa, which is more acutely itself and the posterior centrodiapophyseal lamina (PCDL), immediately pointed posteriorly than anteriorly, and houses a deeply-ramifying below the postzygapophysis. There is no indication that a hyposphene– pneumatic foramen. Ventral to this fossa, the lateral face of the centrum hypantrum system was present, and we suggest that this was genuinely is essentially flat. absent in Diamantinasaurus as in many derived titanosaurs (Salgado The neural arch (inclusive of the neural spine) is slightly taller et al., 1997a). The base of the postzygodiapophyseal lamina (PODL) is dorsoventrally than the centrum (Supplementary Table 2). The observable on the posterior face of the diapophysis. This lamina can be dorsomedially facing prezygapophyses are supported ventrally by traced posteriorly as a thin plate extending to the postzygapophysis. stout, undivided centroprezygapophyseal laminae (CPRL), which form Its presence contrasts with the condition in many derived titanosaurs, the lateral margins of the anterior neural canal opening. The bases of wherein the PODL is absent (Salgado et al., 1997a). the prezygapophyses are connected to one another along their midline As noted above, the right parapophysis has been deflected so that it by dorsolaterally angled intraprezygapophyseal laminae (TPRL), which now projects slightly ventrolaterally; in life, it would have projected form a widened ‘V’-shape in anterior view and roof the neural canal. laterally and would have been located just anteroventral to the These laminae thicken laterally towards the prezygapophyses. Robust diapophysis. Thus, the parapophysis was apparently connected to the prezygoparapophyseal laminae (PRPL) connect the prezygapophyses anteroventral margin of the diapophysis by a very short and stout to the parapophyses. The prezygapophyses are linked to the neural parapodiapophyseal lamina (PPDL). The left parapophysis appears to spine by single spinoprezygapophyseal laminae (SPRL), which decrease be preserved almost in situ, and is level with the prezygapophysis, but in thickness as they ascend the anterolateral margins of the neural the associated diapophysis has been lost. Each parapophysis lies at the spine. Both SPRLs have been deflected laterally to the left — the right end of a ‘peduncle’ and has a dorsoventrally elongate, elliptical articular SPRL is now aligned with the base of the median prespinal lamina surface. The parapophysis is connected to the centrum via ashort (PRSL), whereas the left SPRL now extends towards the left anterior centroparapophyseal lamina (ACPL), which shares a base parapophysis. Nevertheless, it seems likely that the base of each SPRL with the CPRL. These two laminae create a wide, anterolaterally opening

Please cite this article as: Poropat, S.F., et al., Revision of the sauropod dinosaur Diamantinasaurus matildae Hocknull et al. 2009 from the mid-Cretaceous of Australia: Implications for Gondwanan..., Gondwana Research (2014), http://dx.doi.org/10.1016/j.gr.2014.03.014 8 S.F. Poropat et al. / Gondwana Research xxx (2014) xxx–xxx parapophyseal centroprezygapophyseal fossa (PACPRF) on the lower would have been visible throughout its length in anterior view (if the part of the neural arch, below the level of the prezygapophyseal articu- vertebra was undistorted). The region between the SPRL and SPDL lar surface. A ridge extends anterodorsally across the lateral surface of forms a dorsoventrally tall prezygapophyseal spinodiapophyseal fossa the neural arch, from the dorsal part of the centrum towards the (PRSDF). posteroventral part of the parapophysis. This is probably the posterior The morphology of the spinopostzygapophyseal laminae (SPOL) centroparapophyseal lamina (PCPL). This PCPL defines the dorsal mar- cannot be fully ascertained; nevertheless, their presence is clearly indi- gin of the lateral pneumatic fossa on the centrum. Towards its cated on the vertebra. The broad and transversely rounded base of the anterodorsal tip, the PCPL bifurcates: the upper branch extends to the SPOL can be observed above the right postzygapophysis, and there is parapophysis itself, whereas the lower one merges into the posterior no indication that this lamina was bifurcated. Bifurcation of the SPOL margin of the ACPL. This bifurcation, therefore, creates a small but in middle–posterior dorsal neural spines is the derived condition for deep fossa between the upper and lower branches of the PCPL and the most eusauropods, but many somphospondylans reverted to a singular posterior margin of the ACPL. This fossa appears to have been further di- structure (Wilson, 1999, 2002), with the derived titanosaurs Elaltitan vided by small thin laminae, but this might merely reflect the loss of the Mannion and Otero 2012 and Opisthocoelicaudia notable exceptions surface bone that has revealed internal chambers. The dorsally bifurcat- (Mannion and Otero, 2012). A lamina appears to curl upwards and out- ed PCPL is provisionally considered to be autapomorphic for wards near the top of the spine; this could be part of the SPOL, perhaps a Diamantinasaurus. damaged sub-triangular aliform process. A postspinal lamina (POSL) is The right diapophysis is formed from a stout, dorsoventrally com- present on the posterior face of the neural spine, though its extent can- pressed plate of bone. This projects laterally and only slightly dorsally, not be fully determined. but it might have projected slightly more dorsally in life and then been ventrally displaced into its current position. However, it seems 4.1.2.2. Dorsal vertebra B (Fig. 6A–F). The second, larger vertebra recov- unlikely that it could have been directed strongly dorsally (i.e. approx- ered from Diamantinasaurus is not as well preserved as dorsal vertebra imately 45° to the horizontal), as is the case in some sauropods A. The anterior, ventral, left lateral and the left half of the posterior faces (Upchurch, 1998). The anterior centrodiapophyseal lamina (ACDL) is of the centrum are in good condition, as are the left prezygapophysis apparently absent, though this is commonly a consequence of the dorsal and parapophysis. The remainder of the neural arch and the right side migration of the parapophysis along the dorsal series, which interrupts of the centrum are missing. The vertebra has been distorted and a itscourseandseparatesitintotwolaminae,theACPLandPPDL(Wilson, 250 mm long rib fragment remains attached via matrix to the left dorso- 1999). The steeply anterodorsally oriented PCDL, though poorly lateral margin of the posterior cotyle. Matrix also remains within the preserved, is clearly present. Unlike many derived somphospondylans, pneumatic fossa, the posterior cotyle, and on parts of the neural arch. such as Andesaurus Calvo and Bonaparte 1991, Epachthosaurus, Euhelopus Nevertheless, this specimen reveals some aspects of posterior dorsal (Wilson and Upchurch, 2009) and Saltasaurus Bonaparte and Powell morphology that are not preserved in dorsal vertebra A, and also con- 1980 (Salgado et al., 1997a), the ventral end of the PCDL does not expand firms the presence of several notable features described in the latter. anteroposteriorly or bifurcate (Salgado et al., 1997a; Wilson and The centrum is strongly opisthocoelous, although the posterior Upchurch, 2009; D'Emic, 2012; Mannion et al., 2013). The plesiomorphic cotyle remains partially infilled with matrix. The external bone on the condition of a ventrally unbifurcated PCDL is also displayed in several condyle has been lightly eroded, revealing the camellate nature of the titanosaurs, including Alamosaurus Gilmore 1922 (Lehman and Coulson, interior of the vertebra. The ventral face is transversely and longitudi- 2002), Lirainosaurus Sanz et al. 1999 (Díez Díaz et al., 2013a) and nally concave, bound laterally by ridges extending ventrally from the Opisthocoelicaudia (Borsuk-Białynicka, 1977). Posteriorly, the PCPL termi- lateral faces of the centrum; however, unlike the other vertebra of nates close to the base of the PCDL. On the left side of the neural arch, a Diamantinasaurus, there is no midline ridge in the ventral concavity. second lamina runs anteroventral to the PCDL and extends parallel to it. The absence of this midline ridge may reflect damage to the ventral sur- This latter lamina is potentially the ‘upper PCPL’ noted by D'Emic face of the centrum, since the surface bone has been removed in places, (2012), which also occurs in several brachiosaurids, and several revealing the internal camellae; however, it is possible that the absence somphospondylans, including Andesaurus (Mannion and Calvo, 2011), of the keel merely represents variation along the series, as observed Jiangshanosaurus Tang et al. 2001, Ligabuesaurus Bonaparte et al. 2006 in Lirainosaurus (Díez Díaz et al., 2013a). Several anteroposteriorly and Opisthocoelicaudia (D'Emic, 2012; Mannion et al., 2013). Poor preser- elongate, elliptical vascular foramina pierce the ventral surface of the vation means that it is not clear where the upper PCPL terminates at its centrum. anterodorsal end on the left lateral side, although it probably met the The area ventral to the pneumatic fossa on the lateral surface of the underside of the diapophysis or posterior face of the PPDL. The ventral centrum is essentially flat. The pneumatic fossa itself occupies the dorsal surface of the upper PCPL gives rise to a short, anteroventrally directed region of the centrum, effectively extending from the anterior to the ridge that extends towards the lower PCPL. posterior margin. Despite still being infilled with matrix, it is clear that As preserved, the neural spine is dorsoventrally short, though it the pneumatic fossa housed a pneumatic foramen that extended into seems to be slightly incomplete, missing its very dorsal-most portion. the body of the centrum. On both sides, there are additional thin lami- Consequently, the possibility that the dorsal margin of the neural nae forming two deep fossae located anterodorsal to the pneumatic spine was slightly bifid (as seen in Dongyangosaurus Lü et al. 2008 and foramen. As in dorsal vertebra A, these additional fossae lie below the Opisthocoelicaudia (Borsuk-Białynicka, 1977)) cannot be completely anterior end of an anterodorsally directed ridge that forms the dorsal ruled out, though is not inferred here. The neural spine does not appear margin of the lateral pneumatic fossa. Thus the ‘lower PCPL’ can also to project strongly posterodorsally as in many derived titanosaurs be tentatively identified in dorsal vertebra B. (Wilson, 2002), but this might be the result of post-mortem distortion. The prezygapophyses are supported ventrally by the lateral margins Despite being affected by distortion, the spine was evidently transverse- of the neural canal, which are formed by CPRLs manifested as ly wide and relatively compressed anteroposteriorly. The extremely mediolaterally wide, wall-like sheets of bone. The prezygapophyses thin PRSL, which is not bifid ventrally, extends from the dorsal margin are connected to each other by TPRLs. The intersection of these laminae of the neural canal (intersection of the TPRLs) to around one-third the is ‘V’-shaped, as in dorsal vertebra A, although in this case the presence height of the neural spine and forms a distinct ridge, as in diplodocoids of a PRSL cannot be ascertained because of the absence of the neural and most somphospondylans (Mannion et al., 2013). The diapophysis spine. A thin vertical lamina projects ventrally from the TPRL intersec- appears to be connected to the lateral margin of the neural spine via a tion, terminating slightly above the dorsal margin of the neural canal. spinodiapophyseal lamina (SPDL). This lamina apparently projects Near their bases, the CPRLs form the dorsolateral margins of the anterior further laterally than the SPRL, so that the lateral margin of the SPDL neural canal opening. These laminae help to define the margins of fossae

Fig. 6. Paratype dorsal vertebra (dorsal vertebra B) of Diamantinasaurus matildae in A, anterior; B, dorsal; C, left lateral; D, posterior; E, right lateral; and F, ventral views.

that lie dorsolateral to the neural canal, below the TPRLs and ventrome- reflects torsion (of around 45°) between the main rib shaft and the dial to the prezygapophyses. The postzygapophyses are not preserved, proximal plate. Between the posterior margin and the anteromedial and the bases of the laminae by which they were supported cannot be flange of the proximal end, there is a posteromedially facing fossa. Bro- observed. ken surfaces at the proximal end indicate that the rib was pneumatic The parapophyses are supported by an ACPL and probably by a lower and had a camellate internal tissue structure (Hocknull et al., 2009), as PCPL; the identification of these laminae is complicated by the presence in other titanosauriforms (Wilson and Sereno, 1998). The main rib of matrix on the lateral margin of the neural arch. shaft is moderately curved throughout its length with the concave On the left side, a steeply inclined PCDL can be seen, and this again side facing medially. Its anterior margin is thicker transversely than its indicates that the ventral end of this lamina is not expanded or bifurcated. posterior one. The distal end is transversely compressed and somewhat Another ridge lies anteroventral to the PCDL and extends anterodorsally expanded anteroposteriorly. parallel to it. This appears to be equivalent to the upper PCPL in dorsal vertebra A. The upper PCPL meets the posterior surface of a ridge that 4.1.4. Sacral vertebrae (Fig. 7A; Supplementary Table 3) extends steeply anteroventrally from the region of the base of the The remains of four incomplete, fused sacral centra, to which the diapophysis. The latter ridge bifurcates into two parallel ridges towards bases of two pairs of acetabular processes are attached, have been its ventral end. At present, however, it is not certain whether this bifurcat- extracted from a large nodule recovered from AODL 85. The dorsal por- ed ridge represents parts of the ACPL, the lateral part of the CPRL, or some tions of these are incompletely exposed, though the ventral portion has other lamina. been completely prepared, revealing two essentially complete centra, two further partial centra (one attached anteriorly, one posteriorly), 4.1.3. Dorsal ribs and the bases of two pairs of acetabular processes. The remainder of Hocknull et al. (2009) reported the presence of ten dorsal rib the sacrum awaits preparation; until this is complete, it will not be pos- portions in the holotype of Diamantinasaurus. Below, we briefly summa- sible to determine the position of the preserved centra within the sa- rise and supplement their description based on the largest and most crum. However, assuming that the sacrum of Diamantinasaurus complete rib available. This rib preserves the region from approximately consisted of a least five vertebrae, these vertebrae represent sacrals the distal end to just below the proximal plate (capitulum and 1–4or2–5; if the sacrum consisted of six vertebrae, they could also con- tuberculum are not preserved). Near the proximal end, the posterior ceivably be sacrals 3–6. margin of the shaft thickens and becomes a ridge that extends proxi- The preserved centra are transversely convex and anteroposteriorly mally and anteriorly to form the anterolateral edge of the proximal very slightly concave on their ventral surfaces. The apex of the ventral plate. A second ridge starts on the medial face of the shaft a short way convexity manifests as a subtle ridge running down the mid-line; it below the proximal plate. This first appears as a low, anteroposteriorly flares slightly as it approaches the intercentral sutures. The anterior- rounded bulge, but becomes more acute and prominent as it extends most suture preserved is little more than a subtle linear depression. proximally. This ridge forms the posterior margin of the proximal The middle suture is much more clearly marked as a deep but narrow plate. The shift in the relative positions of these ridges and margins linear depression, whereas the posterior-most suture preserved is

Fig. 7. Type sacral elements of Diamantinasaurus matildae: coalesced sacral vertebrae in A, ventral view; sacral acetabular process A (left) in B, dorsal; C, anterior; D, ventral; E, lateral; F, posterior; and G, medial views; and sacral acetabular process B (left) in H, dorsal; I, anterior; J, ventral; K, lateral; L, posterior; and M, medial views.

more distinct still. This suture could represent the junction between a to be mediolaterally narrow relative to the anterior or posterior sacral posteriorly concave vertebra and an anteriorly convex vertebra, as is vertebrae, unlike Neuquensaurus which shows distinct mediolateral evident in Neuquensaurus australis (von Huene 1929) Powell 1992 and narrowing of the middle sacral centra (D'Emic and Wilson, 2011). Trigonosaurus (Campos and Kellner, 1999; Campos et al., 2005; D'Emic and Wilson, 2011). The combination of these features suggests that 4.1.5. Sacral processes (Supplementary Table 3) these vertebrae are the posterior-most four sacral vertebrae of Two isolated partial sacral processes are preserved in the holotype of Diamantinasaurus. Broken surfaces indicate that the internal bone was Diamantinasaurus, but were only brieﬂy mentioned by Hocknull et al. camellate, and neither of the complete centra possessed external pneu- (2009). Based on their relative sizes, it is likely that the ﬁrst sacral pro- matic fossae; this combination of features has also been observed in cess described below (A) was attached to one of the middle sacral ver- Wintonotitan and Saltasaurus, and, in general, pneumatisation of the tebrae, whereas the second (B) was attached either to an anterior or sacrum appears to characterise somphospondylans (Mannion et al., posterior sacral vertebra. Both represent acetabular processes (sensu 2013). The middle sacral vertebrae of Diamantinasaurus do not appear Wilson, 2011) and probably come from the left side of the sacrum.

4.1.5.1. Sacral process A (Fig. 7B–G). Both the medial (sacral) and lateral 4.2. Appendicular skeleton (iliac) articular facets of this sacral process are incomplete; this is unsur- prising, since these would have been fused to one of the sacral vertebrae 4.2.1. Right scapula (Fig. 8A–B; Supplementary Table 4) and the ilium respectively in life. The medial end is much taller dorso- The right scapula preserves most of the blade, but there is a badly ventrally than anteroposteriorly long, whereas the dimensions of the damaged section between the proximal end of the blade and the lateral end are approximately equal. Between these expanded ends, remains of the acromion. The surface bone is clearly crushed from the element is anteroposteriorly compressed. Its ventral margin is impact forces exerted from above the blade. This occurred post- strongly rounded anteroposteriorly and concave (arched upwards) mortem and probably post-burial where the bone was damaged and transversely. Towards the lateral end, the dorsal surface gives rise to a moved, although it remained proximal to the undistorted bone. Direct- tall and anteroposteriorly thin plate. The anterior face of this plate is ed crushing damage of this sort is likely to have occurred as a result of generally flat or slightly convex, whereas the posterior face is shallowly dinoturbation of the surrounding sediments. The acromion itself only concave both transversely and dorsoventrally. The lateral and dorsal preserves the glenoid and the region just posterior to the glenoid, margins of this plate are broken, but there is a concave finished margin such that anatomical information is not available for the acromial at the medial end where the plate diminishes in height. This concave ridge, coracoid articulation, or dorsal two-thirds of the acromion. The margin suggests that there was a transverse foramen (sensu Wilson, scapula will be described as if the blade were oriented horizontally. 2011) extending parasagittally between the sacral rib and the sacral The proximal end is robust, with a roughened glenoid region orient- vertebra. Transverse foramina have been noted in several other ed at approximately 45° with respect to the scapular blade in lateral sauropods, including Opisthocoelicaudia (Borsuk-Białynicka, 1977)and view. The glenoid articular face (which is larger than that of the cora- Camarasaurus lewisi (Jensen 1988) McIntosh et al. 1996.Damaged coid) is broadly triangular, expanded dorsally and bevelled slightly later- surfaces, especially at the articular ends, demonstrate that the acetabular ally (contra Hocknull et al., 2009). In this way, the glenoid differs from sacral process was highly pneumatised, as noted by Hocknull et al. the medially deflected condition that characterises somphospondylans (2009). (Wilson, 2002), and is thus regarded as a local autapomorphy of Diamantinasaurus within this clade (based on the phylogenetic position of this taxon resolved in the analyses run in this study; see below). Pos- 4.1.5.2. Sacral process B (Fig. 7H–M). The larger left sacral process terior to the glenoid, a small area on the ventrolateral margin is flattened assigned to Diamantinasaurus is incomplete at both its medial and later- and smooth; this may have been a point of muscle attachment (possibly al ends, and on the posterior face. The long axis of the medial articular the scapular head of the triceps (Borsuk-Białynicka, 1977)). On the later- facet is dorsoventrally oriented, whereas that of the lateral articular al surface, dorsal to this attachment area, there is a longitudinal fossa facet extends anteroposteriorly. Thus, in anterior and posterior views, that is defined dorsally by an acute ridge. Although partially obscured the medial end appears much more expanded than the lateral, with by damage, it can be seen that the region posterior to the glenoid gives the reverse being true in dorsal and ventral views. Despite their perpen- rise to a convexity on the ventromedial margin (Hocknull et al., 2009). dicular long axes, the expansion of the ends is subequal, though both are This low, rounded process lies approximately below the most posterior incomplete. The iliac attachment area has a flattened articular facet that part of the acromial plate and, therefore, probably represents the more faces ventrolaterally. This posterior portion tapers dorsoventrally to- anterior of the two ventral processes noted in the scapula of Alamosaurus wards its terminus, but there is no clear facet or broken area where it (D'Emic et al., 2011). Such a process has been observed in various other would attach to the next part of the sacricostal yoke (possibly the result somphospondylan titanosauriforms, including Wintonotitan (Hocknull of damage). As in sacral process A, the dorsal surface of the acetabular et al., 2009), Chubutisaurus Del Corro 1975 (Carballido et al., 2011), process is convex in sacral process B. There is no sign of a foramen Angolatitan Mateus et al. 2011, and Ligabuesaurus (Bonaparte et al., between ilium and rib, but this is probably the result of damage. The 2006), and probably represents another point for muscle attachment. ventral surface of the process is convex anteroposteriorly and rounds There is no second posterior process on the ventral margin of the smoothly into the anterior surface. Posteriorly, however, there is a blade. The scapular blade is dorsoventrally tall, transversely thin, flat- more abrupt junction between the ventral and posterior surfaces to-convex laterally and flat-to-concave medially. The medial concavity because the latter is more concave dorsoventrally. is deepest slightly ventral of the mid-height of the blade. Both above

Fig. 8. Type pectoral girdle elements of Diamantinasaurus matildae: right scapula in A, lateral; and B, medial views; and right coracoid in C, lateral; and D, glenoid views.

Please cite this article as: Poropat, S.F., et al., Revision of the sauropod dinosaur Diamantinasaurus matildae Hocknull et al. 2009 from the mid-Cretaceous of Australia: Implications for Gondwanan..., Gondwana Research (2014), http://dx.doi.org/10.1016/j.gr.2014.03.014 12 S.F. Poropat et al. / Gondwana Research xxx (2014) xxx–xxx and below the lateral convexity, the blade flattens out. Thus, the base of other advanced titanosaurs, such as Opisthocoelicaudia (Borsuk- the blade has a ‘D’-shaped cross-section (contra Hocknull et al., 2009), Białynicka, 1977) and Saltasaurus (Powell, 1992, 2003). The proximal contrasting with the rectangular cross-section expressed in many and lateral surfaces of the humerus meet at approximately 90°, with somphospondylans (Wilson, 2002). The convexity of the lateral face is the proximal-most point of the lateral margin approximately level located closer to the ventral margin so that the blade thins dorsally with the remaining lateral half of the proximal surface: this is a derived and has a stouter ventral margin. This convexity creates a lateral ridge, state that characterises the humeri of somphospondylans (Upchurch, at around one-third of the blade height from the ventral margin, which 1999; Wilson, 2002; Mannion et al., 2013). The proximal articular sur- extends almost to the distal end of the scapular blade. At about mid- face is rugose and moderately convex anteroposteriorly. It has a shallow length on the lateral side of the blade there is a low, rounded longitudi- ‘C’-shaped profile in proximal view, with an anterior concavity created nal ridge located a short distance below the dorsal margin. This creates a largely by the lateral portion curving anterolaterally. The humeral shallow longitudinal fossa between itself and the dorsal margin of the head is located slightly medial to the mid-width of the proximal end. main lateral ridge. It also creates a flattened area that faces dorsolaterally This gives rise to a vertical ridge on the posterior surface that extends and leads to the true dorsal margin of the blade. This additional ridge and distally to the level of the distal end of the deltopectoral crest. The shallow fossa on the lateral surface of the scapula blade has not been posterolateral margin of the proximal end is thickened to form a stout observed in other sauropods and is, therefore, regarded as an vertical ridge that defines and deepens the lateral triceps fossa. This autapomorphy of Diamantinasaurus. The blade expands dorsoventrally ridge fades out level with the point where the deltopectoral crest is towards its distal end, with this expansion mainly restricted to the dorsal most prominent. A similar structure has not been observed in any margin. In lateral view, the distal end has a mildly convex profile, meet- other sauropod and we, therefore, provisionally regard this feature as ing the dorsal and ventral margins at distinct ‘corners’. Consequently, the an autapomorphy of Diamantinasaurus. Because of this posterolateral distal end is the widest part of the blade dorsoventrally. The lateral sur- ridge, the lateral surface of the deltopectoral crest meets the posterior face of the distal end has a mildly sigmoid curvature in transverse cross- surface at a rounded but acute angle, rather than merging smoothly, section, with a shallow fossa located dorsally and a more ventral convex- but this might reflect crushing. ity (the latter being the terminus of the main lateral ridge). The extent to which the deltopectoral crest extends across the anterior face of the humerus varies markedly between the two humeri — in 4.2.2. Right coracoid (Fig. 8C–D; Supplementary Table 5) the right humerus, the deltopectoral crest projects anteromedially and As seen in the right scapula, the right coracoid was broken in several extends medially across one-third of the humeral width, whereas in places and deformed prior to burial. A portion of the coracoid, previously the left humerus the deltopectoral crest is oriented anteriorly. More- identified as a sternal plate by Hocknull et al. (2009),ismorelikelythe over, on the right humerus the deltopectoral crest appears to extend dorsal margin of the coracoid which, as a result of dinoturbation, has distally for more than half the length of the humerus, and to be ex- been detached and moved away from the rest of the scapulocoracoid panded distally. In contrast, the deltopectoral crest extends no fur- complex. Our description below, therefore, follows this re-interpretation ther than the mid-length and shows less distal expansion in the left of the ‘sternal plate’ fragment as the dorsal portion of a now near- humerus. The latter element appears to be less distorted and is, complete right coracoid. The coracoid is broadly quadrangular, the therefore, regarded as providing more accurate data on the morphology preserved margins being relatively straight. Both laterally and medially, of the deltopectoral crest. A medially deflected deltopectoral crest the coracoid has an essentially flat surface; however, it is likely that the characterises most somphospondylans (Wilson, 2002; Mannion et al., original curvature of the element has been distorted by crushing. The 2013) and several brachiosaurids (Mannion et al., 2013), but several coracoid is thickest transversely at the glenoid fossa and thinnest at the somphospondylan taxa revert to the plesiomorphic state of an anterior- incompletely preserved dorsal margin. The glenoid articular surface has ly projecting deltopectoral crest, including Euhelopus (Young, 1935)and a ‘D’-shaped outline in posteroventral view, with a relatively flat medial Rapetosaurus (Curry Rogers, 2009). There are no clear ridges or grooves margin, and it curves up onto the lateral margin of the coracoid. There on the lateral side of the deltopectoral crest. is a remnant of the notch anterior to the glenoid on the ventral margin, A distinct anterior fossa is present on the proximal half of the humer- but its depth and length cannot be fully assessed; nor can we determine us,boundedlaterallybythedeltopectoralcrestandmediallybyaridge whether an infraglenoid lip was present. No ridges or tuberosities can that initially projects laterodistally, becoming more horizontal along its be detected on the lateral surface of the coracoid, contrasting with the lateral half, and meeting the medial margin of the deltopectoral crest. As anterodorsally placed tuberosity present in Huabeisaurus Pang and such, this ridge forms the ventromedial margin of the anterior fossa. Cheng 2000 (D'Emic et al., 2013), and the anteroventrally positioned According to Borsuk-Białynicka (1977), this medial ridge would have turberosity observed in Lirainosaurus (Díez Díaz et al., 2013b) and been the attachment site for the M. coracobrachialis brevis. In other ad- Opisthocoelicaudia (Borsuk-Białynicka, 1977). The coracoid foramen, vanced titanosaurs where this ridge is present, like Opisthocoelicaudia which is completely encircled by bone, is separated at its posterior margin (Borsuk-Białynicka, 1977), it typically extends medially and creates from the scapular articulation on the lateral face by a very narrow a much larger and shallower fossa. This strongly curving ridge and (17 mm) portion of bone. This foramen has been crushed so that its out- deeper fossa in Diamantinasaurus is present in both humeri, and is line in lateral view cannot be determined accurately. Based on our recon- regarded as autapomorphic. The anteriorly most prominent point of struction of the coracoid, the foramen would be situated at approximately the deltopectoral crest lies relatively close to the proximal end of the mid-height. The scapular articular facet of the coracoid is flat and marked- humerus (proximal-most 15–20%), and this margin of the deltopectoral ly thinner (62 mm) than the glenoid fossa (129 mm). As noted above, the crest is slightly expanded transversely. coracoid contributes less to the glenoid fossa than the scapula. Based on The point of attachment of the M. latissimus dorsi is marked by a low the preservation of this margin and the corresponding margin on the ridge on the posterolateral margin, slightly above the mid-length of scapula, these bones were not fused when the animal was alive, possibly the humerus. However, this region is not as well developed as that in indicating that this individual was not fully grown at its time of death. Neuquensaurus (Otero, 2010)orMagyarosaurus von Huene 1932 (NHMUK R3857, PU pers. obs. 2012). 4.2.3. Right and left humeri (Figs. 9A–F, 10A–F; Supplementary Table 6) In anterior view, the lateral margin of the humerus is concave Both left and right humeri are well preserved, with the left showing (Hocknull et al., 2009). At mid-length, the shaft has an anteroposteriorly less evidence of post-mortem distortion. The left is better preserved at compressed ‘D’-shaped cross-section (mediolateral width to the proximal end, which has a sigmoid transverse profile in anterior anteroposterior length ratio of approximately 1.5), with a mildly convex view, created by a laterally placed concave area adjacent to a proximally anterior and more rounded posterior surface. The transverse width of projecting medial convexity. A similar morphology occurs in several the shaft at mid-length to humerus length ratio is 0.2 (Hocknull et al.,

Fig. 9. Holotype left humerus of Diamantinasaurus matildae in A, anterior; B, distal; C, lateral; D, proximal; E, posterior; and F, medial views.

2009). The distal end of the medial surface is flat and faces mainly me- titanosaurs, including Opisthocoelicaudia and Saltasaurus (Wilson, dially. The M. coracobrachialis longus would have attached to the distal 2002). The condyle for the radius is more expanded transversely than end of the humerus along the anteromedial margin — a low ridge of the ulnar condyle (Hocknull et al., 2009). In anterior view, the ulnar con- bone appears to demarcate this muscle attachment point, somewhat dyle projects further distally than the radial one (Hocknull et al., 2009), more proximally than interpreted in Opisthocoelicaudia (Borsuk- such that the distal end surface is not strictly perpendicular to the long Białynicka, 1977). There is a vertical fossa on the anterolateral part of axis of the humeral shaft. the distal end, bounded posteriorly by a stout prominence. Along the lateral half of the distal end, the anterior surface forms a divided con- 4.2.4. Right ulna (Fig. 11A–H; Supplementary Table 7) dyle. The two ridges comprising this condyle are quite widely separated The right ulna is well preserved and very robust (Hocknull et al., from one another, and there are no additional intervening ridges. The 2009). The maximum width of the proximal end divided by ulna length more lateral of the anterodistal processes is reasonably prominent, is 0.44; thus, Diamantinasaurus has a robust ulna as represented in projecting about 40−50 mm from the anterior surface. As such, several derived titanosaurs (e.g. Opisthocoelicaudia and Saltasaurus), Diamantinasaurus differs from titanosaurs (and some derived non- where this ratio is typically 0.4 or higher (Wilson, 2002; Curry Rogers, titanosaurian somphospondylans, e.g. Chubutisaurus; Carballido et al., 2005; Mannion et al., 2013). In proximal view, the ulna is triradiate. 2011), in which this condyle is undivided (D'Emic, 2012). Two relatively The olecranon process is pronounced, clearly exceeding the height prominent ridges define a moderately deep anconeal fossa at the distal of the anteromedial and anterolateral processes in anterior view. A end of the posterior surface. The distal end of the humerus is wider well-developed olecranon is a titanosaur feature (McIntosh, 1990; transversely than anteroposteriorly. Its articular surface is rugose and Upchurch, 1995; Wilson and Sereno, 1998; D'Emic, 2012; Mannion mildly convex anteroposteriorly, but only extends slightly onto the et al., 2013), although it is also present in the basal macronarian anterior surface. This distal articular surface is not strongly divided Janenschia Wild 1991 (Bonaparte et al., 2000). The anteromedial and an- into lateral and medial condyles: Diamantinasaurus,therefore,lacks terolateral processes are of a similar length, though the anteromedial the concave distal end profile in anterior view seen in some derived process is slightly longer and 1.5 times broader in dorsal aspect. In

Fig. 10. Holotype right humerus of Diamantinasaurus matildae in A, anterior; B, distal; C, medial; D, proximal; E, posterior; and F, lateral views. medial and lateral views, the anteromedial process is distinctly concave, damaged. The shaft becomes more trapezoidal in cross-section near sloping towards the olecranon process. This concave profile is present in the broadly ‘D’-shaped distal end, though with the straight margin of the ulnae of several titanosauriforms, and Janenschia (Upchurch, 1995, the ‘D’ being slightly concave. This concave margin forms the anterior 1998; D'Emic, 2012; Mannion et al., 2013). The anterolateral process is or anteromedial fossa that received the radius. The distal surface lacks essentially flat, sloping dorsally to merge smoothly with the olecranon the comma-shaped outline and well-developed slender anteromedial process. The posterior surface, distal to the olecranon process, forms a process present in many other sauropods. It is slightly wider transverse- prominent ridge, which separates the flattened posterolateral face of ly than anteroposteriorly, and there is a slight projection from its other- the ulna from the more concave posteromedial face. The proximal wise transversely rounded posterior margin. This projection marks the ends of both the posteromedial and posterolateral faces are capped by distal end of the ridge descending from the posterior process of the prominent ridges of bone running along the bases of the anteromedial proximal end. The distal articular surface is rugose and moderately and anterolateral processes respectively. The rim of bone lining the convex transversely. It is slightly more convex anteroposteriorly, largely anteromedial process is more prominent due to the concave nature of because it curves upwards onto the posterior projection. the posteromedial face below. The anteromedial process gives rise to a distally extended ridge that 4.2.5. Right radius (Fig. 12A–F; Supplementary Table 8) is sharper than those extending from the posterior and anterolateral The radius of Diamantinasaurus is complete and very well- processes. The mid-shaft has a sub-triangular cross-section, with sur- preserved. The ratio of proximal end width to overall length for faces that face approximately posteromedially, posterolaterally and an- Diamantinasaurus is 0.36, a figure exceeded among titanosaurs by teriorly. A clear interosseous ridge cannot be seen, but this region is Opisthocoelicaudia (0.44) and approached by Saltasaurus (0.34), with

Fig. 11. Holotype right ulna of Diamantinasaurus matildae in A, lateral; B, anterolateral; C, anterior; D, anteromedial; E, medial; F, proximal; G, posterior; and H, distal views.

most other taxa recording ratios of ~0.25 or less (Mannion et al., 2013). end is more than double that of the mid-shaft, comparable to the condi- The proximal articular surface is oval, with a rounded lateral margin and tion in many somphospondylans (Wilson, 2002; Mannion et al., 2013). more acute medial tip. This articular surface is flat, rugose and approxi- The surface of the distal end is rugose and moderately convex mately twice as wide transversely as anteroposteriorly (Supplementary transversely. The lateral half of this surface slopes upwards in Table 4). In most sauropods, such as Saltasaurus (Powell, 1992, 2003), anterior view: such bevelling of the distal radius occurs in several the proximal articular surface is slightly concave, although somphospondylans (Wilson, 2002; Mannion et al., 2013), and in various Opisthocoelicaudia is apparently unusual in having a mildly convex sur- more basal eusauropods (Mannion et al., 2013; Mateus et al., 2014). The face (Borsuk-Białynicka, 1977). The anteroposteriorly widest part of the distal end has a broad elliptical outline but is slightly flattened proximal end is located close to the midline, and is partly created by the posterolaterally where it articulates with the ulna. This means that the posterior margin at this point producing a moderately developed lip lateral portion is slightly narrower than the medial part, but this is sub- that overhangs the posterior surface. There is a small protuberance on tle. The degree of transverse expansion of both the proximal and distal the anterolateral corner of the bone, just below the proximal end. The ends of the radius relative to the width of the shaft at mid-length is var- long axes of the proximal and distal ends are sub-parallel so that there iable in sauropods. Diamantinasaurus possesses an intermediate level of is little torsion of the radial shaft. In anterior view, the lateral margin transverse expansion of its articular surfaces, with somewhat greater of the radius is mainly concave, whereas the medial margin is sinuous. expansion occurring in Epachthosaurus (Martínez et al., 2004), and On the anterior surface of the shaft, there is a low transversely rounded less expansion in Aeolosaurus rionegrinus Powell 1987 (Powell, 2003), ridge that extends distally and medially, becomes slightly more promi- Argyrosaurus Lydekker 1893 and Elaltitan (Mannion and Otero, 2012). nent towards mid-length, and then fades out towards the distal end. The The radius of Uberabatitan Salgado and Carvalho 2008 has a much lateral face of the shaft, near the proximal end is rounded more anteroposteriorly expanded distal end than in Diamantinasaurus. anteroposteriorly rather than flat as in Wintonotitan (Hocknull et al., 2009). At about one-quarter of the way from the proximal end there is 4.2.6. Manus (Figs. 13–14) a vertical ridge that originates on the anterior surface close to the lateral No carpal elements are known for Diamantinasaurus and, given the margin of the bone. This ridge extends distally and laterally until it relatively complete nature of the forelimb material, it is possible forms the lateral margin of the shaft just above the distal end. Conse- that there were no ossified carpals in the living animal. If correct, quently, the lateral surface of the shaft shifts from facing Diamantinasaurus possesses the derived absence of ossified carpal posterolaterally near the proximal end to anterolaterally near the distal elements that also characterises several advanced titanosaurs such as end. Another ridge originates on the posterior surface at about the same Opisthocoelicaudia and Alamosaurus (Gilmore, 1946; Borsuk-Białynicka, level as the lateral one. This extends distally and laterally and becomes 1977; Upchurch, 1998). more prominent, reaching its greatest projection just above the distal As reported by Hocknull et al. (2009),allfive metacarpals of end, about one-third of shaft width from the lateral margin. This proba- Diamantinasaurus are represented in the holotype specimen, though bly represents the interosseous ridge for ligamentous attachment to the the first metacarpal is from the left side and the remaining four are ulna. Between this interosseous ridge and the lateral ridge, there is a from the right. The metacarpals of Diamantinasaurus, like those of vertical striated rugosity near the distal end. Just lateral to this slight other sauropods, vary in length; they are, from longest to shortest, III– bulge, on the posterior surface, there is a small vertical ridge that was II–I–IV–V(Hocknull et al., 2009). The fact that metacarpals I and II are probably a site for muscle attachment. The transverse width of the distal shorter than metacarpal III means that Diamantinasaurus retains the

Fig. 12. Holotype right radius of Diamantinasaurus matildae in A, anterior; B, distal; C, medial; D, proximal; E, posterior; and F, lateral views. plesiomorphic state seen in most sauropods rather than the derived laid side-by-side on a flat surface with the long axis of each distal artic- condition (metacarpal I longest) that occurs in Opisthocoelicaudia and ular end oriented transversely. This means that each element has its Alamosaurus (Upchurch, 1998). Metacarpal V is 80 mm shorter than palmar surface facing ventrally, its external surface facing dorsally, metacarpal III, though there is only a 10 mm difference between meta- and lateral and medial surfaces (the latter two usually being the places carpals II and III. Metacarpal III is ~0.6 times the length of the radius, of contact between adjacent elements; see Mannion and Otero, 2012 for equivalent to the otherwise highest known longest metacarpal to radius a similar treatment of the manus). Thus, the external surface will be ratio of any sauropod, that of the titanosaur Argyrosaurus (Mannion and termed dorsal, and the palmar, ventral. Otero, 2012). The metacarpals of Diamantinasaurus, like those of other titanosaurs, The eusauropod manus typically has a columnar ‘U’-shaped arrange- were arranged in a ~360° columnar horseshoe-shaped arrangement ment (first recognised by Hatcher, 1902), in which the five metacarpals (Fig. 13B) as interpreted by Hocknull et al. (2009). Metacarpal I would are oriented vertically with their proximal ends forming an arc of have been located in the posteromedial portion of the articulated between 180° and nearly 360° (Janensch, 1961; Upchurch, 1994, manus when viewed proximally. In this position, the dorsal surface of 1998; Wilson and Sereno, 1998). This morphology creates difficulties the metacarpal would have faced posteriorly and towards the midline for the accurate and consistent description of the positions and orienta- of the animal, and the ventral surface would have faced anteriorly and tions of structures on the metacarpals and the relationships between away from the midline. In life, metacarpal II would have occupied the elements. For example, the surface that is generally termed the ‘palmar’ anteromedial part of the manus in proximal view, the dorsal face surface in most dinosaurian manüs faces in different directions in the would have faced anteromedially, and the rounded ventral process sauropod manus depending on which metacarpal is being considered would have projected posterolaterally. Metacarpal III would have occu- (Upchurch, 1994). Here, therefore, we describe the metacarpals as if pied the central anterior portion of the articulated manus, with its

Fig. 13. Holotype manus elements of Diamantinasaurus matildae: A, right metacarpals II–V and manual phalanges II-1 and III-1 as discovered; B, right manus in proximal view (left metacarpal I reversed); left metacarpal I in C, dorsal; D, distal; E, lateral; F, proximal; G, ventral; and H, medial views; right metacarpal II in I, dorsal; J, distal; K, medial; L, proximal; M, ventral; and N, lateral views; right metacarpal III in O, dorsal; P, distal; Q, medial; R, proximal; S, ventral; and T, lateral views; right metacarpal IV in U, dorsal; V, distal; W, medial; X, proximal; Y, ventral; and Z, lateral views; and right metacarpal V in AA, dorsal; AB,proximal; AC, medial; AD, ventral; AE, distal; and AF, lateral views.

Fig. 14. Type manual phalanges Diamantinasaurus matildae: right manual phalanx I-2 in A, dorsal; B, lateral; C, ventral; D, distal; E, medial; and F, proximal views; right manual phalanx II-1 in G, proximal; H, dorsal; I, lateral; J, ventral; K, distal; and L, medial views; right manual phalanx III-1 in M, proximal; N, dorsal; O, lateral; P, ventral; Q, distal; and R, medial views; right manual phalanx IV-1 in S, proximal; T, dorsal; U, lateral; V, ventral; W, distal; and X, medial views; and right manual phalanx V-1 in Y, proximal; Z, dorsal; AA, lateral; AB, ventral; AC, distal; and AD, medial views. dorsal surface facing anteriorly. In life, metacarpal IV would have occu- settings. The absence of manual phalanges in other derived titanosaurs pied the anterolateral portion of the manus in proximal view. Its dorsal is based upon relatively weak and, of necessity, almost exclusively neg- surface would have faced anteriorly and away from the midline of the ative evidence. The most complete skeleton of Rapetosaurus (FMNH PR animal, and the curved tapering ventral process of the proximal end 2209) was found almost entirely disarticulated (Curry Rogers, 2009: would have been directed approximately towards the midline. In the fig. 4E), and yet, as stated by Curry Rogers (2009, p. 1073), “… the pres- articulated manus, metacarpal V would have been placed in the ence of articular surfaces at the distal ends of metacarpals implies the posterolateral portion in proximal view. The dorsal surface of the shaft presence of phalanges”, despite the fact that manual phalanges were would have faced posteriorly and perhaps away from the midline of not found. In the case of Saltasaurus, the disarticulation of the remains the animal. (shown in the site map provided by Bonaparte et al., 1977), the relative The identifications of the metacarpals made by Hocknull et al. rarity of metapodials (five metacarpals and seven metatarsals recovered (2009) are all accepted here; however, some corrections to the figure from a site with a minimum of five individuals; Powell, 2003) and the depicting the metacarpals (Hocknull et al., 2009: fig. 8) are required: total absence of any phalanges (manual or pedal) do not necessarily G (McII), labelled as external, is ventral; H (McII), labelled as internal, support the interpretation of a genuine absence of manual phalanges is medial; I (McII), labelled as medial, is lateral; O (McIII), labelled as in this taxon. Similarly, Neuquensaurus is known from multiple individ- medial, is lateral; P (McIII), labelled as lateral, is medial and upside uals (von Huene, 1929), yet only three metacarpals have yet been iden- down; U (McIV), labelled as medial, is ventrolateral; V (McIV), correctly tified (Otero, 2010); it is unfortunate that the specimen reported by labelled as lateral, is upside down; Y (McV), labelled as external, is later- Salgado et al. (2005) lacks any forelimb elements distal to the humerus. al and upside-down; Z (McV) is upside-down; AA (McV), labelled as Finally, in describing a mostly articulated skeleton of Epachthosaurus medial, is actually dorsomedial and upside-down; CC (McV), labelled Martínez et al. (2004, p. 114) stated that, “With the exception of a ves- as proximal, is distal; and DD (McV), labelled as distal, is proximal. tigial element fused to the distal surface of metacarpal IV, no phalanges Five manual phalanges are preserved in Diamantinasaurus, and an are present. A nearly identical condition is present in the Mongolian additional one would have been present between metacarpal I and titanosaurian Opisthocoelicaudia (Borsuk-Białynicka, 1977). This corrob- the manual ungual on this digit; this gives an estimated phalangeal orates the hypothesis that the manual phalanges of titanosaurian sauro- formula of 2–1–1–1–1(contra Hocknull et al., 2009). Thus, pods were strongly reduced, unossified, or absent (Giménez, 1992; Diamantinasaurus retains the plesiomorphic state relative to the more Salgado et al., 1997a)”. Contrary to this latter statement, we would derived Opisthocoelicaudia (Borsuk-Białynicka, 1977)andAlamosaurus argue that the presence of a manual phalanx attached to metacarpal (Gilmore, 1946). However, the interpretation that Opisthocoelicaudia IV is probably more likely to indicate that Metacarpals I, II and III also had no manual phalanges is potentially contradicted by Borsuk- bore phalanges that were lost as a result of taphonomic processes in Białynicka's (1977, p. 31) observation of a “small rounded body fused Epachthosaurus and Opisthocoelicaudia, and most likely in Alamosaurus, with the distal surface of the metacarpal IV, which is probably a rudi- Neuquensaurus, Saltasaurus and Rapetosaurus. In sum, we do not think mentary phalanx”. Moreover, the absence of phalanges in Gilmore's there is a strong case for the total absence of manual phalanges in (1946) North Horn Alamosaurus specimen (despite the presence of derived titanosaurs based on the available skeletal evidence. the rest of the forelimb) might not necessarily indicate that these structures were absent — much of the rest of the skeleton was also missing, 4.2.6.1. Metacarpals (Fig. 13; Supplementary Table 9) and it can be presumed that the manual phalanges were not strongly 4.2.6.1.1. Left metacarpal I (Fig. 13C–H). Metacarpal I is complete but attached to the metacarpals and, being small, may have been prone to is crushed at mid-shaft. It is straight with only slightly expanded proxi- being transported away from carcasses even in relatively low energy mal and distal ends relative to the width of the shaft at mid-length. As

Please cite this article as: Poropat, S.F., et al., Revision of the sauropod dinosaur Diamantinasaurus matildae Hocknull et al. 2009 from the mid-Cretaceous of Australia: Implications for Gondwanan..., Gondwana Research (2014), http://dx.doi.org/10.1016/j.gr.2014.03.014 S.F. Poropat et al. / Gondwana Research xxx (2014) xxx–xxx 19 such, it lacks the bowed shape present in the titanosaurs Andesaurus and medial distal condyle. The lateral surface is mildly convex dorsoventral- Argyrosaurus (Apesteguía, 2005; Mannion and Calvo, 2011), though it is ly near the proximal end and bears two proximodistally elongate pits or not impossible that crushing has artificially straightened it. The proxi- foramina, one above the other, separated by around 10 mm of bone. The mal articular end is comma-shaped with a strongly rounded medial middle and distal part of the lateral surface is flatter and faces laterally margin and tapered lateral process. It lacks the strongly compressed and also ventrally. At mid-length, the shaft has a sub-triangular trans- morphology of some titanosauriforms, e.g. Andesaurus, Chubutisaurus, verse cross-section (Hocknull et al., 2009). At about one-fifth of the Rapetosaurus and Venenosaurus Tidwell et al. 2001 (Apesteguía, 2005; way from the distal end, there is a low rounded tubercle on the lateral Mannion and Calvo, 2011). The tapered lateral process has a convex surface, at the point where this surface rounds into the ventral surface. proximal surface in dorsal view because it curves outwards and a little The distal articular surface is expanded transversely and dorsoventrally, distally towards its tip. In proximal view, the dorsal margin is convex, forming a transversely elongate sub-rectangular shape in distal end whereas the ventrolateral margin is concave and creates the tapered view. As a result, the distal end is wider transversely than the proximal form of the lateral process. This concavity lies at the proximal end of a end in dorsal view. This rugose distal surface is flat transversely and is shallow striated fossa, which, in life, probably articulated with metacar- most strongly divided into condyles ventrally. The distal articular area pal II. Just distal to the proximal end, the shaft rapidly transforms its is strongly convex dorsoventrally and curves onto both the ventral cross-section, becoming transversely widened and dorsoventrally com- and dorsal surfaces (especially the latter). The medial surface of the pressed, though this might have been exaggerated by crushing. The medial distal condyle is flat and faces medially and moderately proxi- dorsal surface is transversely convex over the proximal third of the mally. Its ventral margin gives rise to a small medial projection located element. There is no longitudinal ridge on the ventral surface, but a short distance proximal to the distal articular surface itself. The lateral there is a bulge on the ventral margin of the ventrolaterally facing surface of the lateral distal condyle is slightly more convex dorsoven- fossa, located around one-quarter of the element's length from the trally and lacks a ventral lateral process. proximal end. The ventral surface was probably transversely flat over 4.2.6.1.3. Right metacarpal III (Fig. 13O–T). Metacarpal III is complete much of the middle and distal part of the shaft. The distal end is moder- but slightly crushed in the mid-shaft region. In dorsal view, this element ately expanded both transversely and dorsoventrally (especially is straight and has strongly expanded proximal and distal ends, relative ventrally). The distal articular surface is rugose and fairly flat trans- to the shaft (Hocknull et al., 2009). The proximal articular surface is sub- versely, and is strongly convex dorsoventrally. Thus the articular surface triangular in outline, with mildly convex dorsal and lateral margins and curves a little on to the ventral surface and quite strongly on to the a slightly concave ventromedial margin. This surface is rugose and dorsal surface of the element. In this way, Diamantinasaurus differs mildly convex dorsoventrally and even more strongly convex trans- from most titanosauriforms, in which the distal articular surface does versely because the dorsolateral and dorsomedial corners extend distal- not extend onto the dorsal surface of the metacarpal (D'Emic, 2012), ly. This is particularly noticeable on the dorsomedial side, where, as in although this plesiomorphic state also appears to be present in the metacarpal II, there is a ‘D’-shaped facet that faces medially. Just distal derived titanosaur Alamosaurus (Gilmore, 1946: fig. 10). The distal con- to this ‘D’-shaped facet, there is a ventromedial facing fossa for articula- dyles are most clearly divided from each other on their ventral surface, tion with metacarpal II. The ventral margin of this fossa is defined by a forming a wide ventral notch in distal end view, although this does not longitudinal ridge from the ventrolateral corner of the proximal end. continue proximally as a ventral concavity. The lateral condyle is slight- This ridge and the ventromedial fossa extend further along the element ly taller than the medial one, but the difference is subtle. In dorsal view, than in metacarpal II, disappearing at approximately two-thirds of the the lateral distal condyle projects further distally than the medial one, way from the proximal end. The lateral surface is flat or mildly concave such that the metacarpal is longer along its lateral side. This distal bevel- dorsoventrally and faces mainly laterally. The dorsal face of the meta- ling represents a local autapomorphy of Diamantinasaurus:inother carpal is flattened proximally, becomes convex above mid-shaft, and titanosauriforms, the distal end is perpendicular to the long axis of the remains convex all the way to the distal margin, ultimately extending shaft (Wilson, 2002; D'Emic, 2012). The lateral surface of the lateral onto the medial face of the element at the distal end. Just proximal to condyle is flat, whereas the surface of the medial condyle is flat but the distal end, the shaft is thickest laterally and thins medially, giving faces slightly proximally and medially. it a sub-triangular or oval transverse cross-section. The distal end is 4.2.6.1.2. Right metacarpal II (Fig. 13I–N). Metacarpal II is complete, flat transversely and sub-rectangular in outline, but the lateral distal but the element is fractured in places. The proximal end is sub- condyle is transversely thicker and projects further distally than the triangular with a short rounded ventral process. This profile is formed medial one. The distal end expands more strongly along its medial from a long, mildly convex lateral margin, an intermediate length than lateral margin. The articular surface might have expanded on to straight dorsal margin that faces dorsally and a little medially, and a the dorsal surface of the distal end, but this is less clearly marked than long, concave ventromedial margin that faces more medially than ven- in metacarpals I and II. The ventromedial margin also produces a trally. The proximal end articular surface is mildly convex and rugose. It flange-like projection in an equivalent position to the ventrolateral slopes a little distally towards the dorsolateral corner. A ‘D’-shaped tubercle described above. articular area extends directly distally from the dorsomedial corner of 4.2.6.1.4. Right metacarpal IV (Fig. 13U–Z). Metacarpal IV is complete the proximal articular surface and encroaches upon the dorsal part of and generally well preserved. In dorsal view, this element displays little the ventromedial fossa described below. This ‘D’ shape has a rounded transverse expansion of its proximal end, but its distal end is strongly distal margin and straight proximal margin. The rounded ventral expanded relative to the width of the shaft at mid-length. The metacar- process lies at the proximal end of a thin ridge that extends along the pal has a sub-triangular proximal end that is comma-shaped because of slightly more lateral part of the ventral surface up to approximately a slender and slightly curved ventromedially directed process that curls mid-length where it dissipates. The distal half of the ventral surface is, around the proximal end of metacarpal III. As such, the metacarpal lacks therefore, flat transversely. The aforementioned ridge defines the ven- the chevron-shaped proximal end seen in several titanosauriforms tral margin of the ventromedial facing fossa that articulates with meta- (D'Emic, 2012; Mannion et al., 2013). The outline of the proximal end carpal I. The dorsal edge of this fossa is defined by a sharp ridge that is created by a transversely convex dorsal margin, longer and nearly extends to form the dorsomedial margin of the shaft and fades out at straight lateral margin, and moderately concave medial margin. The around mid-length. The dorsal surface of the metacarpal is flat and proximal articular surface is rugose and moderately convex transversely faces dorsally and slightly medially over the proximal and middle part, and dorsoventrally. A rugosity is located at the proximal end of the and faces more dorsally on the distal part of the shaft. This change in ori- medial surface. The ventral process of the proximal end gives rise to a entation of the dorsal surface is produced by a low rounded ridge that long ridge that extends distally and laterally along the ventral surface. extends from the dorsolateral corner of the proximal end towards the As a result, the ventral surface faces more ventromedially towards the

Please cite this article as: Poropat, S.F., et al., Revision of the sauropod dinosaur Diamantinasaurus matildae Hocknull et al. 2009 from the mid-Cretaceous of Australia: Implications for Gondwanan..., Gondwana Research (2014), http://dx.doi.org/10.1016/j.gr.2014.03.014 20 S.F. Poropat et al. / Gondwana Research xxx (2014) xxx–xxx distal end of the element. This thin ridge widens a little transversely to- claw is long relative to its height at the proximal end, with a proximal wards the middle of the element, and then dissipates. There is a moder- height to length ratio of approximately 0.4. This is close to the ratio of ately deep medial fossa on the proximal part of the shaft. This is defined 0.43 seen in Shunosaurus Dong et al. 1983 (Zhang, 1988), but signifi- ventrally by the ventral ridge just described, and dorsally by a ridge that cantly lower than the respective ratios of 0.6 in Apatosaurus louisae is quite prominent and acute that starts some distance distal to the (Gilmore, 1936), 0.66 in Janenschia (Janensch, 1922) and 0.52 in proximal end. A similar fossa and ridge occurs on the lateral side of “Bothriospondylus madagascariensis” Lydekker 1895 (Läng and the shaft. The ridges that define the dorsal margin of the lateral and me- Goussard, 2007); this feature is, therefore, potentially autapomorphic dial fossae near the proximal end extend distally and ventrally towards within Titanosauriformes for Diamantinasaurus. However, this feature the distal condyles. The dorsomedial ridge extends all of the way to the should be treated with caution given that the ungual was not found in projecting part of the medial distal condyle (see below), whereas the articulation with any other manual elements, and the locality that dorsolateral one dissipates before reaching the condyle. The distal end yielded the Diamantinasaurus holotype has also produced remains of of the dorsolateral ridge projects slightly laterally, possibly representing other vertebrates including the theropod Australovenator (Hocknull a structure equivalent to the ventrolateral tuberosities on metacarpals II et al., 2009). The claw is nearly straight in dorsal view, with a very slight and III. The ventral surface just proximal to the distal end is flat trans- lateral curve towards its tip. In lateral view, the distal tip curves a little versely. The shaft in this region has a convex dorsal surface and it ap- ventrally. The medial and lateral surfaces are convex dorsoventrally, pears that it is a little thicker on the medial side, making the medial with the lateral face being more strongly curved. A longitudinal nail surface more rounded and the lateral one more tapered. The distal groove can only be seen on the lateral side, defined ventrally by a end is rugose and not clearly divided into two condyles, even on ridge. This ridge creates a convex ventrolaterally facing surface. A tuber- the ventral surface. It has a dorsoventrally compressed hexagonal osity appears to be present on the ventral margin of the claw, though outline (Hocknull et al., 2009), with longer dorsal and ventral mar- this might represent a deformational artefact (note the crack in the ele- gins and short dorsomedial, dorsolateral, ventromedial and ventro- ment running immediately distal of this region). lateral margins. This is because the distal condyles produce lateral 4.2.6.2.2. Right manual phalanx II-1 (Fig. 14G–L). This phalanx is com- and medial projections, most strongly on the medial side. The medial plete but damaged on its dorsal surface. It is relatively large and sub- process is more pointed in distal view, whereas the lateral one is circular in dorsal view. As in most sauropods, the phalanx is wider trans- more rounded. The distal articulation curves up on to the dorsal sur- versely than long proximodistally (Wilson, 2002; Upchurch et al., 2004, face of the shaft. As in metacarpal III, the distal end surface is slightly 2007; Yates, 2007). This element also seems to be a little longer on its concave transversely and more flat dorsoventrally than in metacar- lateral margin than on its medial one. In dorsal view, the lateral margin pals I and II. is concave, whereas the medial margin seems to be gently convex. The 4.2.6.1.5. Right metacarpal V (Fig. 13AA–AF). This metacarpal is nearly proximal surface is ovate, with a pronounced dorsal lip for articulation complete, apart from some damage to the medial part of the proximal with metacarpal II. In proximal end view, the dorsal margin of this end. The proximal articular surface has a sub-triangular outline formed surface is strongly rounded, whereas the ventral one is straighter. The from a short, mildly convex ventral margin that faces slightly laterally, a articular surface is not divided into lateral and medial fossae by a low longer mildly convex lateral margin that faces a little dorsally, and a vertical midline ridge. Distal to the proximal end, the very short ‘shaft’ damaged medial margin that would have been longer than the two and distal end of the phalanx are compressed dorsoventrally. The lateral other edges. The dorsal tip of the proximal end extends into a longitudi- surface of the phalanx is concave towards the proximal end. The dorsal nal ridge. This ridge extends distally and slants medially, eventually surface is flat just distal to the lip region, whereas the ventral surface is meeting the dorsomedial margin of the shaft. Near the proximal end, mildly convex transversely. The distal end appears to be about equally the lateral surface is mildly convex dorsoventrally and does not bear a thick dorsoventrally on its lateral and medial sides, though the lateral fossa. At about one third of the way from the proximal end, there is a half is perhaps slightly thicker than the medial one. ventrolaterally located tuberosity that gives rise to a low ridge that ex- 4.2.6.2.3. Right manual phalanx III-1 (Fig. 14M–R). This manual pha- tends distally and dorsally to the lateral distal condyle. As a result, the lanx is poorly-preserved, missing the entire proximal surface. As can surface of the shaft, which faces laterally close to the proximal end, ro- be seen in Fig. 13A, this element was found in close association with tates to face dorsally near the distal end. The central third of the shaft the right metacarpals and right manual phalanx II-1. Based on the rela- has an acute ridge along its ventromedial margin, roofing a shallow me- tive size of this element to the other preserved manual phalanges, it is dial fossa. At about one-fifth of the way from the distal end, on the me- most likely that this element represents manual phalanx III-1. This reas- dial surface, there is a low rounded projection not seen in the other sessment means that a putative autapomorphy of Diamantinasaurus metacarpals. Towards the distal end, the shaft becomes more com- identified by Hocknull et al. (2009), i.e., “manual phalanx III-1 heavily pressed dorsoventrally and widens transversely. The distal end is rugose reduced”, is no longer valid. and square in outline, with rounded corners (especially the ventrolater- 4.2.6.2.4. Right manual phalanx IV-1 (Fig. 14S–X). This element was al corner). The articular surface is rugose and moderately concave trans- not included in the original description of Diamantinasaurus by versely. The distal end is convex dorsoventrally and curves onto the Hocknull et al. (2009). It is complete apart from damage to its proximal ventral surface and, to a lesser extent, the dorsal one. In ventral view, articular surface, which has a roughened texture. It is dorsoventrally the medial distal condyle develops a vertical ridge along its outer mar- compressed and wider transversely than long proximodistally. The dor- gin that projects further distally than the rest of the articular surface. sal surface is moderately convex transversely and it is not clear that the Neither of the distal condyles has a strong lateral or medial projection. medial margin is thicker than the lateral one. The distal end is convex The medial condyle has a flat medial surface, whereas the lateral surface dorsoventrally and subtly divided into condyles (this is clearest of the lateral condyle is more convex dorsoventrally and faces moder- distoventrally), with a slight groove on the midline between the con- ately ventrally. dyles. The lateral condyle is a little thicker dorsoventrally than the medial one. 4.2.6.2. Manual phalanges (Supplementary Tables 10, 11) 4.2.6.2.5. Right manual phalanx V-1 (Fig. 14Y–AD). This terminal pha- 4.2.6.2.1. Left manual phalanx I-2 (Fig. 14A–F). The ungual is nearly lanx is smaller than the first phalanges on digits II, III and IV. In proximal complete (Hocknull et al., 2009). It is moderately compressed trans- and distal views, it has a slightly dorsoventrally compressed, sub- versely, but is relatively wide compared to its dorsoventral height. The circular to elliptical outline. The element is wider transversely than proximal articular end is not well defined but apparently faced proxi- long proximodistally. The proximal end is mildly concave dorsoventral- mally and moderately laterally. This surface is sub-triangular in outline, ly and slightly convex transversely. The remaining surfaces are rugose; with apices located medially, dorsolaterally and ventrolaterally. The they are perhaps damaged or pathological.

above or in front of the base of the pubic process (Hocknull et al., 2009), as occurs in most titanosauriforms (Upchurch, 1998; Mannion et al., 2013). The pubic process has a comma-shaped horizontal cross- section, with a concave posterior face, transversely rounded anterior face, a convex lateral margin and sharper medial edge. The distal end of this process is expanded so that it is at least twice as wide transversely as anteroposteriorly. Its lateral margin is narrow near the main body of the ilium, but distally it broadens and becomes more strongly convex anteroposteriorly. The medial margin of the pubic process curves posteromedially so that it is visible in lateral view behind the rest of the process. There is no ridge and triangular hollow morphology above the base of the pubic process on the lateral surface, unlike the condition in Cetiosaurus Owen 1841 (Upchurch and Martin, 2003), and the titanosaurs Rocasaurus Salgado and Azpilicueta 2000 (García and Salgado, 2013: ﬁg. 8b) and Lirainosaurus (Díez Díaz et al., 2013b). The ischiadic articulation is greatly reduced as in other sauropods (Upchurch, 1998). There is a prominent protuberance on the lateral part of the base of the ischiadic articulation, where it meets the main body of the ilium, as in Camarasaurus grandis (Marsh 1877) Osborn and Mook 1921 (YPM 1901 [holotype] and YPM 1902 [referred specimen; holotype of “Morosaurus robustus” Marsh 1878]asﬁgured by Ostrom and McIntosh, 1966: pl. 65, 66); within Titanosauriformes, this feature appears to be a possible synapomorphy uniting Diamantinasaurus and Opisthocoelicaudia (Borsuk-Białynicka, 1977).

4.2.8. Right and left pubes (Figs. 15, 17A–F; Supplementary Table 13) The right pubis is complete, except for some small portions missing from the margin of the distal expansion; the left pubis is completely preserved. The pubis in Diamantinasaurus is a flattened plate of bone that generally resembles those of other advanced sauropods in most re- spects. Unlike derived diplodocoids (McIntosh, 1990; Upchurch, 1998), there is no hook-like ambiens process lying just anterior to the iliac articulation in Diamantinasaurus; instead, this process is low and ridge-like. The articular surface of the ilium is longer anteroposteriorly Fig. 15. Holotype left pelvis (ilium, pubis and ischium) of Diamantinasaurus matildae in lateral view. than mediolaterally (though note that this long axis would have been directed anterolaterally in life). This surface is mildly rugose and slightly concave, with this concavity facing dorsally and partly medially. The acetabular region is not well preserved in the right pubis. In the left pubis, 4.2.7. Left ilium (Figs. 15, 16A–B; Supplementary Table 12) however, this region forms a well-developed concave margin in lateral The left ilium is displayed lying on its medial side, is fragile, and can- view and is strongly thickened transversely. The acetabular surface not be turned over easily; therefore it is difficult to obtain information faces posterodorsally and slightly laterally. In the left pubis, there is a on the morphology of its medial surface. This element is nearly notch on the lateral margin of the acetabulum, just posterior to the complete apart from some damage to its dorsal margin and the loss of iliac articular surface. This notch extends distally for a short distance, much of the postacetabular process (Hocknull et al., 2009). As noted with its anterior and posterior margins defined by low ridges (the by Hocknull et al. (2009), broken surfaces demonstrate that the internal more posterior of these ridges lying just anterior to the obturator tissue structure of the ilium comprises numerous large camerate opening); these grooves and ridges on the lateral surface of the chambers. Pneumatisation of the ilium is a derived state that occurs pubis near the obturator foramen might represent an autapomorphy in several other titanosaurs, including Alamosaurus (Poropat, 2013), of Diamantinasaurus. The elliptical obturator foramen passes Epachthosaurus (Martínez et al., 2004), Lirainosaurus (Díez Díaz et al., dorsomedially through the bone and, in the right pubis, is continuous 2013b), Saltasaurus (Powell, 2003; Cerda et al., 2012)andSonidosaurus with a vertical groove on the medial face extending towards the acetab- Xu et al. 2006a and in the basal somphospondylan Euhelopus (Wilson ular margin. and Upchurch, 2009); see also Mannion et al. (2013). The main shaft of the pubis is broad anterolaterally to posteromedially In lateral view, the preacetabular process is rounded, largely because and has a flat anteromedial surface. The posterolateral surface of the shaft its blunt anteroventral tip does not project as far anteriorly as the most is more convex transversely than the anteromedial one, but remains anterior part of the process. The preacetabular process is transversely relatively flat. There is a weakly developed ridge extending vertically thin near its base, but becomes strongly thickened at its anteroventral along the anterior part of this posterolateral surface, but this is not prom- tip. This tip forms a lateral expansion located just posteroventral to inent enough to create the comma-shaped horizontal cross-section the true anterior tip of the blade. In lateral view, this anteroventral through the shaft that occurs in many other sauropods (Upchurch et al., expansion projects a short way below the rest of the ventral margin of 2004). This low rounded ridge bifurcates below the obturator foramen, the preacetabular process, giving the anterior part an almost bifid producing branches that extend anterior and posterior to this foramen. profile. In dorsal view, the preacetabular process is deflected anterolat- The more anterior branch is the ridge that forms the posterior margin erally, with its anteroventral portion projecting laterally to form a hori- of the vertical groove leading to the lateral acetabular notch described zontal shelf as is also seen in titanosaurs such as Opisthocoelicaudia above. The more posterior branch of the lateral ridge forms the anterior (Borsuk-Białynicka, 1977) and Saltasaurus (Powell, 2003). The dorsal margin of the concave fossa lying below the ischiadic articular region margin of the main body of the ilium is strongly rounded antero- and anterior to the symphyseal junction on the midline. The ischiadic posteriorly. It appears that the highest part of this margin was located articulation is long, approximately one-half (0.46) of pubis length

Fig. 16. Holotype left ilium of Diamantinasaurus matildae in A, dorsal; and B, lateral views.

(Hocknull et al., 2009), comparable to the ratio in most macronarians in most somphospondylans, including Opisthocoelicaudia, Qiaowanlong (Salgado et al., 1997a; Wilson and Sereno, 1998), although this is reduced You and Li 2009 and Tastavinsaurus Canudo et al. 2008 (Salgado et al., in some derived titanosaurs (e.g. Alamosaurus and Opisthocoelicaudia; 1997a; Upchurch, 1998; D'Emic, 2012; Mannion et al., 2013). The artic- Mannion et al., 2013). The ischiadic articular surface extends distally, ulation for the pubis is very long and has a vertically concave medial then merges into the pubic symphyseal region. The symphyseal surface face. The length of the articulation between pubis and ischium suggests is deflected first dorsomedially and then runs more directly distally, as that the opening in the ventral floor of the pelvis was probably closed in occurs in other sauropod pubes. Unlike the pubes of many sauropods Diamantinasaurus (i.e. there is no emargination distal to pubic articula- (PU pers. obs.), there is no ‘D’-shaped or elliptical expansion on this tion), as also occurs in all titanosaurs (McIntosh, 1990; Upchurch, surface just anteroventral to the ischiadic articulation, but this might be 1998; Wilson, 2002; Upchurch et al., 2004; Mannion and Calvo, 2011; because of damage. Distally, the symphysis terminates in a small broad- D'Emic, 2012). In lateral view, the acetabulum forms a strongly concave ened sub-triangular region for attachment with its partner on the mid- margin, such that the pubic articular surface forms an anterodorsal line. The distal end is only slightly expanded relative to the main shaft, projection, as in most other titanosaurs (D'Emic, 2012; Mannion et al., and the rugose surface rises up on to the anterolateral margin: however, 2013). Its articular surface faces slightly outwards and is widest trans- this region does not form a distinct sub-triangular process or ‘boot’. versely at the iliac peduncle, narrowing somewhat as it extends anteroventrally. However, this surface does not display the strong 4.2.9. Right and left ischia (Figs. 15, 18A–F; Supplementary Table 14) transverse narrowing in its central region that occurs in several The left and right ischia are nearly complete, but have been broken rebbachisaurids (Mannion et al., 2012a). The medial margin of the ace- into four main sections (Hocknull et al., 2009). They are plate-like tabulum is marked by a sharp ridge, as in other sauropods: however, elements with well-developed proximal ends and relatively short distal there is no prominent medial flange projecting from the iliac articulation, shafts. As a result, the ischium is only 0.68 of the pubis length (Hocknull unlike Sonidosaurus (LHV 0010, PDM pers. obs. 2007), Huabeisaurus et al., 2009). Such short ischia represent a derived state that also occurs (D'Emic et al., 2013) and “Ornithopsis” eucamerotus Hulke 1882

Fig. 17. Holotype right pubis of Diamantinasaurus matildae in A, anteromedial; B, distal; C, anterior; D, proximal; E, lateral; and F, posterolateral views.

(Upchurch et al., 2011). In lateral view, the iliac peduncle does not form a transversely than anteroposteriorly, tapering a little medially. This sur- distinct ‘neck’, unlike those of several rebbachisaurids (Sereno et al., face consists of two portions: a medially located area that faces 2007; Whitlock, 2011) and some titanosaurs (Mannion and Calvo, proximomedially, and a distinct, laterally placed area that faces 2011). In horizontal cross-section, the iliac articular surface is wider proximolaterally and slightly anteriorly. Posterior to the base of the

Fig. 18. Holotype right ischium of Diamantinasaurus matildae in A, anterior; B, lateral; C, posterior; D, proximal; E, medial; and F, distal views.

Fig. 19. Holotype right femur of Diamantinasaurus matildae in A, anterolateral, B, anteromedial, C, medial; D, proximal; E, posterior; F, distal; and G, lateral views.

Fig. 20. Holotype right tibia of Diamantinasaurus matildae in A, anterior; B, proximal; C, medial; D, posterior; E, lateral; and F, distal views.

Please cite this article as: Poropat, S.F., et al., Revision of the sauropod dinosaur Diamantinasaurus matildae Hocknull et al. 2009 from the mid-Cretaceous of Australia: Implications for Gondwanan..., Gondwana Research (2014), http://dx.doi.org/10.1016/j.gr.2014.03.014 26 S.F. Poropat et al. / Gondwana Research xxx (2014) xxx–xxx other derived titanosaurs, including Magyarosaurus (NHMUK R3856, PU the medial one, but the latter projects further distally. The lateral pers. obs.). The posterolateral fossa forms a small deep groove between malleolus terminates in a blunt, posteriorly projecting process. the main posterior part of the fibular condyle and the additional posterolateral ridge, which terminates distally at a shelf that links the two posterior ridges. Although the additional ridge is more widespread 4.2.12. Right fibula (Fig. 21A–F; Supplementary Table 17) among titanosaurs, the shelf-like structure is potentially autapomorphic The holotype fibula of Diamantinasaurus was described by Hocknull for Diamantinasaurus.Thefibular condyle projects further distally than et al. (2009); however, a small portion (90 mm long) of this bone had the tibial one, indicating that the distal end was proximomedially not been attached to this element at the time of publication. The total bevelled relative to the long axis of the shaft. This medial bevelling of length of the element was originally given as 710 mm+; this can now the distal femur has previously been optimised as a saltasaurid feature be amended to 769 mm (overlap of surface bone accounts for the seem- (Wilson and Carrano, 1999; Wilson, 2002), but also occurs in other de- ing discrepancy between the size of the added piece and the previously rived titanosaurs, e.g. Rapetosaurus and the basal somphospondylan measured length of the element). The addition of this section to the Dongbeititan Wang et al. 2007 (Mannion et al., 2013). fibula also means that the point at which mid-shaft breadth was measured originally is shifted distally, and is reduced from 150 mm to 4.2.11. Right tibia (Fig. 20A–F; Supplementary Table 16) 126 mm. The robusticity of the element is 0.16; therefore, it is still The tibia is complete and very well preserved (Hocknull et al., 2009). classed as being of “intermediate robusticity” (i.e. between 0.15 and Tibia length is 59% of femur length, which lies in the typical range for 0.25), but falls very close to the value (≤0.15) considered to represent sauropods (Upchurch et al., 2004; D'Emic et al., 2013). The proximal a gracile element (Curry Rogers, 2005). articular end is approximately square in outline, with rounded corners. The fibula is nearly complete, but cracked and damaged, with some These corners project approximately anteriorly, laterally, medially and crushing. The proximal end is transversely compressed and, in lateral posteriorly. The proximal articular surface is rugose and gently convex view, is directed slightly posteriorly. The rugose proximal articular sur- in its anterior part and along its anterolateral and anteromedial margins, face is convex both anteroposteriorly and transversely. In proximal whereas the central and posterior regions are mildly concave. As in view, the anterior corner of the articular surface tapers and projects most neosauropods (Wilson and Sereno, 1998), the ratio between the anteromedially as a triangular process, as in most somphospondylans transverse and anteroposterior diameters of the proximal articular sur- (Wilson and Upchurch, 2009; D'Emic, 2012; Mannion et al., 2013). face is greater than 1.0 (1.2). As noted by Hocknull et al. (2009), the However, the region distal to this process is broken and the preserved cnemial crest is a stout plate that is directed anteriorly and then curves section appears to be a broad rounded ridge, rather than a thin plate. laterally towards its most prominent point, creating a deep fossa poste- The posterior part of the proximal end is relatively stout and is trans- rior to it that would have received the anterior flange of the proximal versely rounded. There is a clear triangular scar on the medial side of fibula. There is no ‘second cnemial crest’ (Bonaparte et al., 2000; the proximal end, marked by a ridge extending anterodistally from the Mannion et al., 2013) posterior to the cnemial crest. The lateral corner proximal posterior corner to the anterior margin. This scar is wider of the proximal end gives rise to a prominent but anteroposteriorly nar- anteroposteriorly at its proximal end than it is long proximodistally at row double ridge that descends distally and fades out around the level its anterior end. D'Emic (2012) regarded the absence of a triangular of the distal end of the cnemial crest. The more anterior ridge, which striated area as a synapomorphy of Titanosauriformes; thus, its presence is longer than the more posteriorly located ridge, becomes quite sharp in Diamantinasaurus might represent a local autapomorphy within the towards its distal end. These ridges mark the posterior border of the group. There is no bulge or anterior trochanter at the distal end of the deep anterolateral fossa located behind, and extending proximally ridge defining the triangular medial scar. The horizontal cross-section from the base of, the cnemial crest; at its apex, this fossa forms a through the upper shaft is approximately ‘D’-shaped, with a flat or smooth-walled notch, located between the cnemial crest and a small slightly convex medial surface and strongly rounded lateral surface. anterior projection arising from the lateral corner of the proximal There is a faint vertical ridge on the medial surface extending from a articular surface. An additional, shallow, vertical fossa is present poste- point just below the margin of the triangular proximal scar to around rior to the double ridge. This fossa, which is bounded posteriorly by a mid-length. This ridge lies at approximately one-third of shaft width ridge on the posterolateral margin of the shaft, houses a proximodistally from the posterior margin. Consequently, there are shallow vertical elongated bulge or tubercle, lying immediately behind the distal end grooves anterior and posterior to this medial ridge, the more anterior of the more posterior of the double ridges. Above this bulge, just one being wider and deeper than the posterior one. This medial ridge below the proximal end, there is a small but deep fossa leading into and associated grooves have not been observed in other sauropods and the bone. These various ridges, fossae and tuberosities on the proximal are provisionally regarded as an autapomorphy of Diamantinasaurus. part of the lateral face of the tibia have not been observed in other The lateral trochanter is a relatively prominent bulge on the lateral sauropods and collectively represent a suite of autapomorphies for surface located at around one-third of fibula length from the proximal Diamantinasaurus. end. This bulge lies just posterior to a low ridge that extends a short The shaft has the typical horizontal cross-section at mid-length, with distance proximally and as far distally as the mid-length of the fibula. a flattened posterolateral surface and more rounded anteromedial face. This lateral ridge is located midway between the anterior and posterior Below the cnemial crest, the anterolateral margin of the shaft develops margins of the shaft and is directed slightly posterodistally in lateral into a thin acute ridge that extends distally as a flange-like projection view. At mid-length, the cross-section through the shaft has an acute before fading out into the rest of the anterolateral margin at around anterior margin and broader rounded posterior one. This cross-section- one-third of the way above the distal end. This thin flange is not contin- persiststothedistalend,butthethinanteriormarginseeninmostsauro- uous with the cnemial crest; it is separated from it by a narrower but pod fibulae just above the distal end cannot be detected, possibly because rounded margin, facing anterolaterally. This anterolateral flange ex- of damage. The transverse width of the distal end is greater than twice tending distally from mid-shaft is regarded as a potential autapomorphy that of the mid-shaft (ratio = 2.58). The distal articular surface is rugose of Diamantinasaurus. The transverse width of the distal end of the tibia is and mildly convex transversely. This surface has a distally facing posterior more than twice the diameter of the mid-shaft long axis, a derived state section and an anterodistally facing anterior portion. The medial margin seen in many titanosaurs (Wilson, 2002). The distal end of the anterior of the distal end forms a shelf-like projection: this might have been exag- face forms a mildly concave sub-triangular region as represented in gerated by distortion, but was probably a genuine feature. In distal view, most sauropods (Upchurch et al., 2004). There is a vertical groove locat- the articular surface is sub-triangular in outline, with medial, anterolater- ed on the posterolateral face of the distal end between the lateral and al and posterior margins. This surface is slightly wider anteroposteriorly medial malleoli. The lateral malleolus projects further posteriorly than than transversely.

Fig. 21. Holotype right ﬁbula of Diamantinasaurus matildae in A, anterior; B, medial; C, distal; D, posterior; E, proximal; and F, lateral views.

4.2.13. Right astragalus (Fig. 22A–F; Supplementary Table 18) the anteroposterior length. The plesiomorphic condition observed The right astragalus (Hocknull et al., 2009) is well preserved and in most sauropods is a ratio of 1.5 or higher (D'Emic, 2012); thus, broadly resembles those of basal titanosauriforms such as Giraffatitan Diamantinasaurus possesses the derived state, as in many other (Janensch, 1961) and more derived titanosaurs such as Aeolosaurus titanosauriforms including Giraffatitan, Euhelopus and Opisthocoelicaudia (Powell, 2003). In Diamantinasaurus, the astragalus has a reduced medial (Mannion et al., 2013). The proximal surface is divided into two main section so that the element covers approximately 80% of the width of the portions: a block-like ascending process situated laterally, and a medially distal end of the tibia. In most sauropods, the astragalus completely caps placed fossa for reception of the lateral part of the distal end of the tibia. the distal tibia, but in several somphospondylans (e.g., Euhelopus)this The lateral face of the ascending process forms the shallow fossa for surface is partially exposed because of the reduction of the medial articulation with the distal end of the ﬁbula.Inbothproximalandposte- part of the astragalus (Wilson and Upchurch, 2009; Ksepka and Norell, rior views, the astragalus tapers in width towards its medial apex, with a 2010). The transverse width of the astragalus is 1.47 times greater than nearly straight anterior margin and curved posterior one. The proximal

Fig. 22. Holotype right astragalus of Diamantinasaurus matildae in A, anterior; B, medial; C, proximal; D, posterior; E, distal; and F, lateral views.

surface of the ascending process is mildly concave because it turns more strongly convex anteroposteriorly. The posterior surface of the as- upwards to form a mild proximal lip along its anterior margin. The prox- cending process continues distally into a prominent rounded rugose pro- imal and medial surfaces of the ascending process meet each other at jection that extends below the rest of the distal articular surface. In approximately 90° along a well-defined margin: thus, this process does posterior view, this projection divides the posterior part of the distal sur- not simply merge into the medial fossa in a gradual fashion. The ascend- face into two regions, the lateral one being narrower transversely than ing process has a sub-rectangular proximal outline, widening slightly the medial one. In lateral view, this produces an acute angle to the transversely towards its anterior end. In lateral view, with its proximal posterodistal margin. This process is also visible in proximal and distal surface lying horizontally, the ascending process extends to the posterior views, forming the most posterior part of the astragalus. Such a margin of the element: this is a derived state that occurs in posterodistal process has not been observed in any other sauropod and Neosauropoda (Wilson and Sereno, 1998). As a result, the proximal sur- is, therefore, regarded as an autapomorphy of Diamantinasaurus. face of the ascending process meets the anterior and posterior surfaces of the astragalus at approximately 90°. There is no fossa or set of foramina 5. Phylogenetic analyses at the anterior base of the ascending process, a derived condition that is typical for eusauropods (Wilson and Sereno, 1998). The posterior surface In order to explore the phylogenetic relationships of Diamantinasaurus, of the ascending process does not display smooth bone in a shallow fossa this taxon was scored for the two most up-to-date and comprehensive between two vertical ridges as would typically occur in most sauropod titanosauriform phylogenetic datasets currently available: Carballido astragali: instead, this surface is as rugose as the proximal and anterior and Sander (2014) and Mannion et al. (2013). ones. In most sauropods there is a ‘tongue’-like projection located on the posterior margin of the astragalus just medial to the posterior termi- 5.1. Carballido and Sander (2014) nation of the ascending process. In Diamantinasaurus, however, this process is greatly reduced, as is the condition in most titanosauriforms The data matrix used by Carballido and Sander (2014) to test the (D'Emic, 2012; Mannion et al., 2013). Two foramina are located on the phylogenetic relationships of Europasaurus within Sauropoda was medial face of the ascending process, where it meets the floor of the me- utilised. This data matrix, which is essentially the same as that used by dial fossa. The smaller foramen lies immediately anterior to the larger Carballido et al. (2012) to test the placement of Comahuesaurus one. The lateral fossa, for articulation with the fibula, faces mainly later- Carballido et al. 2012, was based upon Carballido et al. (2011), which ally and only slightly proximally. It is divided into an upper fossa that was in turn largely derived from Wilson (2002). Aside from the inser- faces laterally, and a lower one that faces proximolaterally, by a broad tion of Diamantinasaurus into the dataset, no other changes were rounded ridge extending anteriorly and a little proximally across the lat- made. Diamantinasaurus was scored for 117 of the 341 characters eral surface of the ascending process. At its posterior end, this ridge sep- (34.3%; Supplementary Table 19). arates from the lateral wall of the ascending process to form a deep cleft Analyses were run in TNT v.1.1 (Goloboff et al., 2008) following the that opens proximally and posteriorly. This ridge and cleft structure methodology employed by Carballido and Sander (2014). Forty-nine of might have been caused by the presence of a piece of adhering bone the characters were multistate characters, and of these 24 were ordered from another element, possiblytheratherbadlydamagedfibula. Never- (12, 58, 95, 96, 102, 106, 108, 115, 116, 119, 120, 154, 164, 213, 216, 232, theless, the anteroposteriorly directed ridge that divides the lateral fossa 233, 234, 235, 256, 267, 298, 299, 301), and an equally weighted parsi- into upper and lower portions appears to be a genuine feature and is po- mony analysis was performed using the heuristic tree search. One tentially autapomorphic for Diamantinasaurus. The laterodistal edge of thousand Wagner tree replicates were utilised, with random addition the lateral fossa is broad rather than thin or shelf-like. The distal surface sequence of taxa followed by tree bisection–reconstruction branch is mildly convex transversely, slopes proximally and medially, and is swapping and ten trees saved per replicate. Twenty most parsimonious

Fig. 23. Phylogenetic tree derived from the Carballido and Sander (2014) matrix, placing Diamantinasaurus as the sister taxon to Tapuiasaurus macedoi, and therefore as a lithostrotian titanosaur.

Fig. 24. Phylogenetic tree derived from the Mannion et al. (2013) matrix, placing Diamantinasaurus as the sister taxon to Opisthocoelicaudia skarzynskii, and therefore as a saltasaurid lithostrotian titanosaur.

6. Discussion recovered as the sister taxon to Alamosaurus + Saltasauridae. Rescoring of Diamantinasaurus for the Mannion et al. (2013) LSDM matrix 6.1. Phylogenetic placement of Diamantinasaurus resolved it in a more derived position, as a saltasaurid lithostrotian closely related to Opisthocoelicaudia fromtheLateCretaceousof Diamantinasaurus has been resolved in many analyses as a Mongolia, and Dongyangosaurus from the early Late Cretaceous of lithostrotian titanosaur, closely related to the derived Late Creta- China. This latter result, if valid, could potentially have serious impli- ceous taxa Opisthocoelicaudia, Alamosaurus and the South American cations for titanosaur palaeobiogeography, since Diamantinasaurus saltasaurines (Saltasaurus, Neuquensaurus). It was first included in phylo- would have been separated geographically by thousands of kilometres genetic analyses by Hocknull et al. (2009), based on the data matrices of and temporally by up to 25 million years from the taxa with which González Riga et al. (2009) and Canudo et al. (2008, derived from Wilson this matrix suggests it was most closely related. Mannion et al. (2013) (2002)). Subsequently, it was included in an analysis by Zaher et al. cautioned that the focus of their dataset was basal titanosauriforms (2011, again derived from Wilson (2002)) and resolved as a lithostrotian (and Tapuiasaurus was not included in this analysis), meaning that draw- titanosaur. However, Mannion et al. (2013) resolved Diamantinasaurus ing palaeobiogeographical conclusions about the timing and route of dis- as a somphospondylan outside Titanosauria in their Lusotitan standard persal of derived titanosaurs on the basis of their analysis alone is discrete matrix (LSDM), and as a titanosaurian outside Lithostrotia in inadvisable. Superficially, the result from the Carballido and Sander their Lusotitan continuous and discrete matrix (LCDM). The thorough (2014) analysis appears to fit better temporally and geographically revision of the holotype, and the description of new material, have with the available data, and suggests a radiation of lithostrotians within strengthened the case for the inclusion of Diamantinasaurus as a member Gondwana prior to the Cenomanian. However, given that Australia's of Titanosauria (it is here reconstructed as such: Fig. 25), and also indi- Cretaceous sauropods occupy a very narrow timespan (Albian– cates that it is more derived than the positions recovered in the original Cenomanian), it would not be advisable to make palaeobiogeographic analyses of Mannion et al. (2013). hypotheses on the basis of them alone at this point. Further work on The insertion of Diamantinasaurus into the matrix of Carballido and the sauropods of central Queensland will allow us to more rigorously Sander (2014) positioned this taxon within Lithostrotia as the sister assess the diversity of the Australian Cretaceous fauna and make more taxon of the Brazilian taxon Tapuiasaurus from slightly older (Aptian) informed palaebiogeographic hypotheses than can be undertaken in Gondwanan strata. This clade (Diamantinasaurus + Tapuiasaurus)is this work.

Fig. 25. Life restoration of Diamantinasaurus matildae as a lithostrotian titanosaur by Travis R. Tischler.

6.2. Sauropod absence in Cretaceous Victoria: genuine, or the result of 6.3. Palaeobiogeographic implications of Australia's Cretaceous taphonomic bias? terrestrial fauna

Despite the fact that over 1500 individual vertebrate elements have The identification of all of Queensland's Cretaceous sauropods as been recovered from the Aptian–Albian sediments along the Victorian Titanosauriformes (Molnar, 2001; Molnar and Salisbury, 2005) has coast of southeast Australia, not a single sauropod element has been never caused doubts about a connection between Australia and South recovered (Benson et al., 2012). The conspicuous absence of sauropod America via Antarctica during the Cretaceous. However, other Australian remains in these sediments could be a result of taphonomic bias. Depo- Cretaceous non-volant terrestrial vertebrates have been utilised as sitional conditions appear to have favoured the preservation of relative- a basis for palaeobiogeographical interpretations or to highlight ly small bones: mammal jaws less than 10 mm long represent the small Australia's potential as a refugium for relict taxa. These are summarised end of the spectrum, whereas the holotype femur of the theropod briefly below. Timimus Rich and Vickers-Rich 1994, one of the largest bones recovered from the Lower Cretaceous sediments of Victoria to date, is only Temnospondyls: The discovery in Victoria of the youngest known 195 mm long. However, it is also possible that sauropods were genuine- chigutisaurid temnospondyl, Koolasuchus (Jupp and Warren, 1986; ly absent in Victoria during the Aptian–Albian, during which time it was Rich et al., 1992; Warren et al., 1997), was one of many lines of evi- located within the Antarctic Circle (Veevers, 2006), enduring sufficiently dence which led to the idea that Australia was a refuge for relict taxa, cool temperatures to have periodically frozen the ground (Constantine since Koolasuchus is the only unquestionable temnospondyl known et al., 1998; Vickers-Rich et al., 1999); furthermore, the rift valley from post-Jurassic sediments (Rich et al., 1988; Warren et al., 1997; would have had quite pronounced topography and slightly different Vickers-Rich et al., 1999). The temnospondyl status of the remains vegetation, and these factors may not have been conducive to the ascribed to Koolasuchus has never been questioned (Schoch, 2013). presence of sauropods in Victoria at this time (McLoughlin et al., Turtles: Three Cretaceous non-marine turtles have been named from 2002). Sea surface temperatures across Australia have been interpreted Australian Cretaceous sediments. Otwayemys from the late Early Cre- to have been low during the Aptian on the basis of the presence of taceous of Victoria has been shown to belong to Meiolaniformes glendonites (De Lurio and Frakes, 1999) and dropstones (Frakes et al., (Gaffney et al., 1998; Sterli and De la Fuente, 2011, 2013; Sterli 1995) in the Bulldog Shale. However, sea surface temperatures appear et al., 2014), a group which was globally distributed by at least the to have risen during the Aptian–Cenomanian, based on isotopic data end of the Early Cretaceous (Sterli and De la Fuente, 2013) and has from molluscs (Henderson and Price, 2012), including belemnites been hypothesised to have originated before the break-up of Pangaea (Price et al., 2012), and carbonates (Clarke and Jenkyns, 1999), despite (Sterli et al., 2014). The fragmentary Chelycarapookus Warren 1969 the fact that Victoria has been interpreted to have become cooler and has proven difficult to place phylogenetically (Gaffney, 1981), though dryer during the Albian on the basis of palynological analysis the interpretation of this turtle as a cryptodiran has been widely ac- (Wagstaff et al., 2013). Sauropods appear to have preferred warmer cepted (Gaffney et al., 1998). Spoochelys Smith and Kear 2013,aturtle climates, especially given that during the Late Cretaceous at least, sauro- from the upper Lower Cretaceous Griman Creek Formation of Light- pods were more diverse at lower palaeolatitudes than ornithischians, ning Ridge, was interpreted as a meiolaniid-like form (Smith and which tended to occupy higher latitudes (Mannion et al., 2012b), Kear, 2013), though this has been questioned and a stem- closely matching Australia's Cretaceous dinosaur faunas. A tempera- testudinate position has been proposed (Sterli et al., 2014). Unnamed ture rise in Victoria during the Albian–Cenomanian transition might chelid pleurodires are also present in the Griman Creek Formation. have facilitated the dispersal of lithostrotian titanosaurs across These are the only pre-Eocene chelids known from outside of Patago- Antarctica from South America into Australia at this time, and allowed nia, and as such they have been used to support the hypothesis of the ancestors of Diamantinasaurus to reach at least as far north as east–west dispersal across Antarctica between South America and central Queensland. Australia during the Early Cretaceous (Smith, 2010).

Crocodylomorphs: The best known crocodylomorph from the Creta- for a close connection between the Australian Cretaceous fauna and ceous of Australia is Isisfordia from the Winton Formation of central that of South America, since megaraptoran theropods (be they Queensland (Salisbury et al., 2006). The holotype specimen was allosauroids (Benson et al., 2010c) or coelurosaurs (Novas et al., initially suggested to be similar to the Early Cretaceous Brazilian 2013)) are most abundant in South America. This identification Susisuchus Salisbury et al. 2003, and it was later interpreted as was strengthened by the discovery of the most complete theropod being slightly more derived than this taxon: a basal member of known from Australia, Australovenator (Hocknull et al., 2009; Eusuchia, the group which includes modern crocodylians (Salisbury White et al., 2012, 2013), which has been resolved as a close relative et al., 2006). This prompted the hypothesis that eusuchians originated of Megaraptor Novas, 1998 and, more often than not, as a in Gondwana (Salisbury et al., 2006). The position of Isisfordia as a neovenatorid (Benson et al., 2010c), though it should be noted that eusuchian less derived than modern crocodylians has been supported the position of Megaraptora within Neovenatoridae (and, therefore, in subsequent analyses (Delfino et al., 2008; Martin, 2010; Holliday Allosauroidea) has recently been contested and a position within and Gardner, 2012). Coelurosauria proposed (Novas et al., 2013). Other megaraptoran/ Non-avian theropods:AsidefromAustralovenator,allAustralian neovenatorid elements have also been recovered from Victoria non-avian theropod remains comprise individual isolated (Benson et al., 2012), including claws and teeth which are directly elements. The identifications of the isolated bones constituting the comparable to the Australovenator holotype (AODF 604). Victorian theropod fauna have been extensively debated. That all The Victorian theropod fauna has also been interpreted to include of the specimens are isolated renders most of their identifications spinosaurs (NMV P221081: Barrett et al., 2011), tyrannosauroids contentious, and necessitates that palaeobiogeographic hypotheses (NMV P186046: Benson et al., 2010a), ceratosaurs (NMV P221202: based upon them be viewed as tentative. Fitzgerald et al., 2012), and dromaeosaurs (Currie et al., 1996; The first non-avian theropod fossil identified from Australia, the Benson et al., 2012). The interpretation of NMV P221081 as a “Cape Paterson Claw” (NMV P10058), was discovered in Victoria spinosaur (Barrett et al., 2011; Benson et al., 2012) has been over a century ago (Smith Woodward, 1906). Given that NMV questioned, and a basal tetanuran or averostran identification pro- P10058 cannot be positively identified as a manual or pedal ungual, posed (Novas et al., 2013). The identification of NMV P186046 as a it was most recently regarded as an indeterminate theropod tyrannosauroid has been debated, with a less precise identification (Benson et al., 2012). as a tetanuran preferred by some authors (Herne et al., 2010), the An astragalus (NMV P150070) recovered from Victoria was originally tyrannosauroid status upheld by others (Benson et al., 2010b, identified as pertaining to Allosaurus Marsh 1877 (Molnar et al., 2012), and a megaraptoran identification proposed most recently, 1981). Later, it was questioned as an allosaurid, let alone a specimen despite the authors acknowledgment that the morphology of of Allosaurus (Welles, 1983); reaffirmed as Allosaurus (Molnar et al., NMV P186046 was indeed consistent with identification as a ty- 1985); removed from Allosaurus but retained as an allosauroid rannosaurid (Novas et al., 2013). Finally, the interpretation of (Chure, 1998); and subsequently aligned with Fukuiraptor Azuma NMV P221202 as a ceratosaur has also been questioned, with a and Currie 2000 and Australovenator (Hocknull et al., 2009), both basal averostran position posited (Novas et al., 2013). identified as megaraptoran neovenatorid allosauroids (Benson et al., The arguments over the identifications of Australia's theropods have 2010c). More recently, the allosauroid affinities of NMV P150070 unsurprisingly led to different views of the implications of the fauna. have been questioned, and abelisauroid affinities suggested On the one hand, those who accept that ceratosaurs, spinosaurs, (Agnolin et al., 2010); however, the allosauroid and neovenatorid neovenatorids, tyrannosauroids and dromaeosaurs were present in status of this specimen has since been defended (Benson et al., Victoria (which would constitute a rather different fauna to that of 2012; Fitzgerald et al., 2012), and a megaraptoran identification ac- South America), have argued for pre-Cretaceous cosmopolitanism cepted (Novas et al., 2013). of theropod groups as a result of broader distributions earlier in Timimus (NMV P186303) was originally identified as an their history. This was followed by local extinctions (Barrett et al., ornithomimosaur, and this identification prompted the hypothesis 2011; Fitzgerald et al., 2012), possibly caused by climatic changes that Australia was the site of origin for this group, which subsequent- (Benson et al., 2012). On the other hand, others favour broader, ly became common in the Late Cretaceous of North America and Asia less specific, affinities for Victoria's theropods and have used this to (Rich and Vickers-Rich, 1994), or that Australia might have had a land argue for a Gondwanan signal in the fauna (Agnolin et al., 2010; connection to East Asia during the Cretaceous (Vickers-Rich, 1996). Herne et al., 2010; Novas et al., 2013).Thefactthatmostof More recently, NMV P186303 has been interpreted as a paravian Australia's theropods are based on isolated remains will ensure close to unenlagiines (Agnolin et al., 2010; Novas et al., 2013), or as that debates over their identifications and palaeobiogeographic im- a tyrannosauroid (Benson et al., 2012). Unenlagiines are otherwise plications continue. For the moment, Australovenator is the most restricted to South America (Gianechini and Apesteguía, 2011; reliable theropod taxon upon which palaeobiogeographic recon- Novas et al., 2013), whereas tyrannosauroids are more common in structions can be based, and its close relationship with Megaraptor the Northern Hemisphere but almost certainly would have had a superficially suggests a closer connection to the South American global distribution given that basal tyrannosauroids have been recov- fauna than to elsewhere. However, this is again complicated by the ered from the Middle Jurassic of England (Rauhut et al., 2010) and the presence of the closely related Fukuiraptor in the Cretaceous of Late Jurassic of China (Xu et al., 2006b). Japan. Oviraptorosaurs were identified in Australia on the basis of a Ornithopods: Australia's ornithopod record is fairly good, with repre- surangular (NMV P186386) and a dorsal vertebra (NMV P186302) sentatives in Queensland, New South Wales and Victoria. The first from Victoria (Currie et al., 1996). As was the case with Timimus, ornithopod discovered in Australia, Fulgurotherium von Huene the identification of oviraptorosaurs in Victoria was seen as evidence 1932, was originally interpreted as a theropod but correctly re- for a possible connection with China (Vickers-Rich, 1996). However, identified as an ornithopod by Molnar (1980a). Additional material NMV P186386 has been considered to be of indeterminate position from both New South Wales and Victoria has been referred to this within Theropoda (Agnolin et al., 2010), whereas a reassessment of genus (Molnar and Galton, 1986; Rich and Rich, 1989), though NMV P186302 supported its identification as a non-dromaeosaurid more recently these referrals have been questioned (Agnolin et al., maniraptoran, though not necessarily an oviraptorosaur (Benson 2010). et al., 2012). Australia's most complete ornithopod to date, Muttaburrasaurus, The identification of a theropod ulna (NMV P186076) from Victoria was originally interpreted as an iguanodontid (Bartholomai and as being megaraptoran-like (Smith et al., 2008) was used as support Molnar, 1981; Molnar and Galton, 1986) or an iguanodontian of

uncertain position (Norman and Weishampel, 1990), though the Non-mammalian synapsids; The presence of a possible dicynodont in discovery of an additional specimen and a reassessment of the holo- the upper Lower Cretaceous Allaru Mudstone of Queensland was type led to the idea that it was more basal within Ornithopoda reported only a decade ago, despite the fact that the specimen (Molnar, 1996a). Later workers retained it within Iguanodontoidea (QM F15990) was discovered almost a century ago (Thulborn and (Norman, 2004), assigned it to Styracosterna based on character as- Turner, 2003). Like the late-surviving temnospondyls of Victoria, sessment without phylogenetic analysis (Agnolin et al., 2010), or re- this possible dicynodont has been cited as evidence for Australia as solved it as the earliest (and thus far only Australian) rhabdodontid a refuge for relict taxa during the Cretaceous, since this occurrence on the basis of a phylogenetic analysis (McDonald et al., 2010). It post-dates the other most recent dicynodonts (which were globally should be noted that McDonald et al. (2010) were only able to distributed) by over 100 million years (Thulborn and Turner, 2003). code from the literature and from casts, and that the holotype and However, the interpretation of QM F15990 as a dicynodont has not referred specimens of Muttaburrasaurus remain incompletely pre- been universally accepted: Agnolin et al. (2010) suggested that it pared from their surrounding matrix. This taxon requires much may represent a baurusuchian crocodyliform, although these au- more work before it can be reliably used to test palaeobiogeographic thors did not fully explain the lines of reasoning that led to this hypotheses. interpretation. Within Australia, ornithopods are most commonly found in Victoria. Mammals: Ausktribosphenos, from the Early Cretaceous of Victoria, The discovery of a diverse assemblage of small ornithopods in was originally identified as a placental mammal; this led to the pro- Victoria (Rich and Rich, 1988, 1989; Rich and Vickers-Rich, 1999) posal of several palaeobiogeographic hypotheses, including two was one line of evidence that led to the idea that Australia was a wherein contact between Australia and Asia was made possible Cretaceous haven, in which taxa that otherwise showed low global during the Cretaceous, the first by island stepping stones, and the diversity were able to become speciose and abundant (Rich et al., second by northward-rafting “Noah's Ark”-like terranes (Rich et al., 1988; Vickers-Rich et al., 1999). On the basis of this evidence, the 1997, 1999a). The discovery of Bishops was used to further advance Australian fauna was also hypothesised to have strong ties to the the hypothesis of a southern origin for placentals and for a direct Antarctic fauna (Rich et al., 1989), matching palaeogeographical connection between Asia and Australia during the Cretaceous (Rich reconstructions of the time (Zinsmeister, 1987). The discovery of et al., 2001). However, the placental status of these mammals has small- to medium-sized ornithopods in the Late Cretaceous of been debated (Kielan-Jaworowska et al., 1998; Rich et al., 1998), Antarctica (Coria et al., 2013) and South America (Coria and and an alternative interpretation proposed: that Ausktribosphenos Salgado, 1996; Salgado et al., 1997b; Martínez, 1998; Coria and and its kin are actually related to monotremes (Luo et al., 2001; Calvo, 2002; Novas et al., 2004; Calvo et al., 2007; Cerda, 2008; Sigogneau-Russell et al., 2001), and to Jurassic forms (grouped as Ibiricu et al., 2010; Cerda and Chinsamy, 2012; Canudo et al., 2013; australosphenids) from Madagascar (Flynn et al., 1999)and Egerton et al., 2013) supports this hypothesis, though taphonomic Argentina (Rauhut et al., 2002). Following this interpretation, the processes might have favoured the preservation of ornithopods, mammal fauna would appear to show a clear Gondwanan signal and the high-latitude location of Victoria during the late Early (Luo et al., 2002). The affinities of these fossils have generated Cretaceous might have promoted higher diversity. much debate, though the relationship of ausktribosphenids to Ceratopsians: The identification of an ulna from the Early Cretaceous monotremes has garnered wider acceptance (Woodburne et al., of Victoria as a neoceratopsian (Rich and Vickers-Rich, 1994), and its 2003; Kielan-Jaworowska et al., 2004; Rich and Vickers-Rich, 2004, subsequent description as the holotype of Serendipaceratops (Rich 2012b; Martin and Rauhut, 2005; Rich, 2008; Phillips et al., 2009). and Vickers-Rich, 2003), has been met with scepticism (Agnolin The acceptance of australosphenids as monotreme relatives, coupled et al., 2010). Ceratopsians are currently known almost exclusively with the presence of the monotremes Teinolophos (originally identi- from the Northern Hemisphere, with the only putative Southern fied as a placental) and Kryoryctes in the late Early Cretaceous of Hemisphere remains, designated Notoceratops Tapia 1918, compris- Victoria (Rich et al., 1999a; Pridmore et al., 2005), and Steropodon ing an incomplete dentary (von Huene, 1929). The discovery of Archer et al., 1985 and Kollikodon Flannery et al., 1995 in the late Yinlong Xu et al. 2006c and Xuanhuaceratops Zhao et al. 2006 in Ju- Early Cretaceous of New South Wales, is strongly suggestive of rassic strata in China suggests that ceratopsians evolved prior to Gondwanan relationships: monotremes are otherwise only known the breakup of Gondwana and thus may have been able to spread from Argentina (Pascual et al., 1992a, 1992b, 2002). Until the discov- throughout the world prior to the Cretaceous; however, more com- ery of Corriebaatar, multituberculates were unknown from Australia plete remains from Southern Hemisphere strata will be required to (Rich et al., 2009). This specimen helped to demonstrate that determine the validity of this hypothesis. Following this line of re- multituberculates were globally distributed (Wilson et al., 2013), search, a recent study by Rich et al. (2014) lends renewed support and the recent discovery of an Early–Middle Jurassic multitubercu- to the identification of Serendipaceratops as a genuine Australian late from India (the oldest yet known) suggests that this clade may ceratopsian. even have originated in Gondwana (Parmar et al., 2013). Ankylosaurs: The presence of ankylosaurs in the Late Cretaceous of South America (Salgado and Coria, 1996), exclusively in deposits 6.4. Summary that have also yielded hadrosaurs, has been used as evidence for a Late Cretaceous connection between South America and North This brief review has highlighted that the majority of Australia's America (Coria and Salgado, 2001). That ankylosaurs (Gasparini Cretaceous fossil terrestrial vertebrates are known from very incom- et al., 1987, 1996; Salgado and Gasparini, 2006) and hadrosaurs plete material. Consequently, the affinities of these specimens are highly (Case et al., 2000) are both present in the Late Cretaceous of debatable, and their incompleteness means that they are only included Antarctica has been used as further evidence of a dispersal event infrequently in phylogenetic analyses. This increases the difficulty of from the north. However, the presence of ankylosaurs in the Lower performing cladistic biogeographic analyses, meaning that they have and lower Upper Cretaceous sediments of Australia (Molnar, only rarely been generally applied; they have never been performed 1980b, 1996b; Molnar and Frey, 1987; Barrett et al., 2010; Leahey specifically with the Cretaceous of Australia as the focus. A factor and Salisbury, 2013), and very rarely in Lower Cretaceous African compounding this difficulty is that Australian Cretaceous terrestrial sediments (Lapparent, 1960; Weishampel et al., 2004), could be vertebrates are almost entirely known from a narrow time interval: utilised as evidence counter to this interpretation, since they dem- the Aptian to Cenomanian. onstrate that ankylosaurs were already present in the Southern This review shows that most of Australia's Cretaceous terrestrial Hemisphere during the Cretaceous (Agnolin et al., 2010). vertebrates have been resolved as the sister taxa of, or within clades

Please cite this article as: Poropat, S.F., et al., Revision of the sauropod dinosaur Diamantinasaurus matildae Hocknull et al. 2009 from the mid-Cretaceous of Australia: Implications for Gondwanan..., Gondwana Research (2014), http://dx.doi.org/10.1016/j.gr.2014.03.014 34 S.F. Poropat et al. / Gondwana Research xxx (2014) xxx–xxx comprised dominantly of, other Gondwanan taxa. This could be was facilitated by a Leverhulme Trust Research Grant (RPG-129), interpreted as being a genuine reflection of close ties to Gondwana and which also enabled him to take part in this collaborative work. South America in particular (which would make palaeobiogeographic sense), but inadequacies in data matrices, taxon sampling, and patchi- Appendix A. Supplementary data ness of the fossil record again render such hypotheses tentative (Upchurch, 2008). Supplementary data to this article consisting of tables of measure- Based on Cretaceous palaeogeographic reconstructions, the hypoth- ments and phylogenetic character scores can be found online at esis that Australia's fauna should show close ties to that of eastern Asia http://dx.doi.org/10.1016/j.gr.2014.03.014. does not appear to be supported. However, it is interesting to note that one of the phylogenetic analyses performed here to evaluate the relationships of Diamantinasaurus resolved it as closely related to a References Late Cretaceous Asian taxon (Opisthocoelicaudia in Mannion et al., 2013). Superficially, this could be viewed as evidence for a connection Agnolin, F.L., Ezcurra, M.D., Pais, D.F., Salisbury, S.W., 2010. A reappraisal of the Cretaceous non-avian dinosaur faunas from Australia and New Zealand: evidence for their between Australia and Asia during the latest Cretaceous. However, Gondwanan affinities. Journal of Systematic Palaeontology 8, 257–300. this interpretation does not match palaeobiogeographic reconstruc- Apesteguía, S., 2005. Evolution of the titanosaur metacarpus. In: Tidwell, V., Carpenter, K. tions; furthermore, a close relationship to Opisthocoelicaudia may repre- (Eds.), Thunder-lizards: The Sauropodomorph Dinosaurs. Indiana University Press, Bloomington & Indianapolis, pp. 321–345. sent discrepancies in coding (this taxon was not coded for the Mannion Archer, M., Flannery, T.F., Ritchie, A., Molnar, R.E., 1985. First Mesozoic mammal from et al. (2013) matrix from personal observation, but only from the liter- Australia—an Early Cretaceous monotreme. Nature 318, 363–366. ature), or problems with taxon sampling (few derived titanosaurs were Azevedo, R.L.M., 2004. Paleoceanografia e a evolução do Atlântico Sul no Albiano (South Atlantic paleoceanography and evolution during the Albian). Boletim de Geociências included in the Mannion et al. (2013) matrix). da Petrobrás 12, 231–249. Azuma, Y., Currie, P.J., 2000. A new carnosaur (Dinosauria: Theropoda) from the Lower 7. Conclusion Cretaceous of Japan. Canadian Journal of Earth Sciences 37, 1735–1753. Barrett, P.M., Rich, T.H., Vickers-Rich, P., Tumanova, T.y.A., Inglis, M., Pickering, D., Kool, L., Kear, B.P., 2010. Ankylosaurian dinosaur remains from the Lower Cretaceous of We have presented a full redescription of the holotype material of southeastern Australia. Alcheringa 34, 205–217. Diamantinasaurus and described paratypic material pertaining to the Barrett, P.M., Benson, R.B.J., Rich, T.H., Vickers-Rich, P., 2011. First spinosaurid dinosaur same individual. This redescription has led to the identification of thirteen from Australia and the cosmopolitanism of Cretaceous dinosaur faunas. Biology Letters 7, 933–936. autapomorphies of this taxon, and five local autapomorphies within Bartholomai, A., Molnar, R.E., 1981. Muttaburrasaurus, a new iguanodontid (Ornithischia: Titanosauriformes. On the basis of our reassessment of Diamantinasaurus, Ornithopoda) dinosaur from the Lower Cretaceous of Queensland. Memoirs of the – we have provided strong support for the interpretation that this species is Queensland Museum 20, 319 349. fi Benson, R.B.J., Barrett, P.M., Rich, T.H., Vickers-Rich, P., 2010a. A southern tyrant reptile. a lithostrotian titanosaur based on ve synapomorphies in the Carballido Science 327, 1613. and Sander (2014) matrix. The inclusion of the dorsal vertebrae and the Benson, R.B.J., Barrett, P.M., Rich, T.H., Vickers-Rich, P., Pickering, D., Holland, T., 2010b. radius, in particular, has shifted the position of Diamantinasaurus from Response to comment on “Asoutherntyrantreptile”. Science 329, 1013-d. Benson, R.B.J., Carrano, M.T., Brusatte, S.L., 2010c. A new clade of archaic large-bodied outside Titanosauria to within Lithostrotia in the Mannion et al. (2013) predatory dinosaurs (Theropoda: Allosauroidea) that survived to the latest Mesozoic. LSDM tree, though this position is only weakly supported. Naturwissenschaften 97, 71–78. Diamantinasaurus is the most derived titanosauriform sauropod Benson, R.B.J., Rich, T.H., Vickers-Rich, P., Hall, M., 2012. Theropod fauna from southern – Australia indicates high polar diversity and climate-driven dinosaur provinciality. known from Australia to date. Its presence in the upper Lower lower PloS One 7, e37122. Upper Cretaceous Winton Formation approximately coincides with Benson, R.B.J., Fitzgerald, E.M.G., Rich, T.H., Vickers-Rich, P., 2013. Large freshwater the appearance of closely related lithostrotian titanosaurs in eastern plesiosaurian from the Cretaceous of Australia. Alcheringa 37, 456–461. Betts, P.G., Giles, D., Lister, G.S., Frick, L.R., 2002. Evolution of the Australian lithosphere. Africa (Malawisaurus) and Brazil (Tapuiasaurus), suggesting that a Australian Journal of Earth Sciences 49, 661–695. lithostrotian radiation took place in Gondwana during or before the Bonaparte, J.F., 1986. Les dinosaures (carnosaures, allosauridés, sauropodes, cétiosauridés) end of the Early Cretaceous, although this is complicated by a potentially du Jurassique Moyen de Cerro Cóndor (Chubut, Argentine) (Part 2). Annales de – close relationship with Late Cretaceous East Asian sauropods. Paléontologie (Vert.-Invert.) 72, 325 386. Bonaparte, J.F., Coria, R.A., 1993. Un nuevo y gigantesco saurópodo titanosaurio de la A brief review of the Cretaceous terrestrial vertebrate body fossil Formación Rio Limay (Albiano-Cenomaniano) de la Provincia del Neuquén, record of Australia highlights the various hypotheses that have been Argentina. Ameghiniana 30, 271–282. made on the basis of the individual components of the fauna. Much of Bonaparte, J.F., Powell, J.E., 1980. A continental assemblage of tetrapods from the Upper Cretaceous beds of El Brete, northwestern Argentina (Sauropoda–Coelurosauria– the fauna seems to show ties to South America in particular, and to Carnosauria–Aves). Mémoires de la Société Géologique de France (Nouvelle Série) Gondwana as a whole in a broader sense. However, Gondwanan 139, 19–28. palaeobiogeography is proving to be more complex than previously Bonaparte, J.F., Salfity, J.A., Bossi, G., Powell, J.E., 1977. Hallazgo de dinosaurios y aves Cretácicas en la Formación Lecho de El Brete (Salta), proximo al limite con Tucumán. realised (Upchurch, 2008), and it is hoped that further work on Acta Geologica Lilloana 14, 5–17. Australia's sauropod fauna will inform future hypotheses of Cretaceous Bonaparte, J.F., Heinrich, W.-D., Wild, R., 2000. Review of Janenschia Wild, with the palaeobiogeography. description of a new sauropod from the Tendaguru beds of Tanzania and a discussion on the systematic value of procoelous caudal vertebrae in sauropods. Palaeontographica Abteilung A 256, 25–76. Acknowledgments Bonaparte, J.F., González Riga, B.J., Apesteguía, S., 2006. Ligabuesaurus leanzai gen. et sp. nov. (Dinosauria, Sauropoda), a new titanosaur from the Lohan Cura Formation (Aptian, Lower Cretaceous) of Neuquén, Patagonia, Argentina. Cretaceous Research The authors would like to thank: I. and S. Muir for allowing AAOD to 27, 364–376. conduct excavations on their property; J. Elliot, B. Elliot and H. Elliot Borsuk-Białynicka, M., 1977. A new camarasaurid sauropod Opisthocoelicaudia skarzynskii for their tireless efforts in support of AAOD; the staff and volunteers gen. n., sp. n. from the Upper Cretaceous of Mongolia. Palaeontologia Polonica 37, – at AAOD for pouring vast amounts of time and energy into pre- 5 64. Bryan, S.E., Cook, A.G., Allen, C.M., Siegel, C., Purdy, D.J., Greentree, J.S., Uysal, I.T., 2012. paring the specimens; T. R. Tischler for providing the restoration of Early-mid Cretaceous tectonic evolution of eastern Gondwana: from silicic LIP Diamantinasaurus matildae; A. G. Cook for providing the Cretaceous sed- magmatism to continental rupture. Episodes 35, 142–152. – imentary extents used in Fig. 1; and the Willi Hennig Society. The au- Calvo, J.O., Bonaparte, J.F., 1991. Andesaurus delgadoi gen. et sp. nov. (Saurischia Sauropoda), dinosaurio Titanosauridae de la Formación Río Limay (Albiano– thors would also like to warmly thank J. L. Carballido and V. Díez Díaz Cenomaniano), Neuquén, Argentina (Andesaurus delgadoi gen. et sp. nov. for their helpful reviews, and to S. McLoughlin for his corrections. SFP (Saurischia–Sauropoda), a titanosaurid dinosaur from Río Limay Formation and BPK's research was funded by an Australian Research Council link- (Albian–Cenomanian), Neuquén, Argentina). Ameghiniana 28, 303–310. Calvo, J.O., Salgado, L., 1995. Rebbachisaurus tessonei sp. nov. a new Sauropoda from the age grant (LP100100339); PDM's research was funded by an Imperial Albian–Cenomanian of Argentina: new evidence on the origin of the Diplodocidae. College London Junior Research Fellowship; and PU's visit to AAOD Gaia 11, 13–33.

Calvo, J.O., Porfiri, J.D., Novas, F.E., 2007. Discovery of a new ornithopod dinosaur from the (Chubutiano), Chubut, Argentina. Actas del I Congreso Argentino de Portezuelo Formation (Upper Cretaceous), Neuquén, Patagonia, Argentina. Arquivos Paleontología y Bioestratigrafía. Asociación Paleontológica Argentina, Tucumán, do Museu Nacional, Rio de Janeiro 65, 471–483. pp. 229–240. Campbell, R.J., Haig, D.W., 1999. Bathymetric change during Early Cretaceous Delfino, M., Codrea, V., Folie, A., Dica, P., Godefroit, P., Smith, T., 2008. A complete skull of intracratonic marine transgression across the northeastern Eromanga Basin, Allodaposuchus precedens Nopcsa, 1928 (Eusuchia) and a reassessment of the Australia. Cretaceous Research 20, 403–446. morphology of the taxon based on the Romanian remains. Journal of Vertebrate Campos, D.d.A., Kellner, A.W.A., 1999. On some sauropod (Titanosauridae) pelves from the Paleontology 28, 111–122. continental Cretaceous of Brazil. In: Tomida, Y., Rich, T.H., Vickers-Rich, P. (Eds.), D'Emic, M.D., 2012. Early evolution of titanosauriform sauropod dinosaurs. Zoological Second Gondwanan Dinosaur Symposium. National Science Museum Monograph, Journal of the Linnean Society 166, 624–671. 15. National Science Museum, Tokyo, Japan, pp. 143–166. D'Emic, M.D., Wilson, J.A., 2011. New remains attributable to the holotype of the sauropod Campos, D.d.A., Kellner, A.W.A., Bertini, R.J., Santucci, R.M., 2005. On a titanosaurid dinosaur Neuquensaurus australis, with implications for saltasaurine systematics. Acta (Dinosauria, Sauropoda) vertebral column from the Bauru Group, Late Cretaceous Palaeontologica Polonica 56, 61–73. of Brazil. Arquivos do Museu Nacional, Rio de Janeiro 63, 565–593. D'Emic, M.D., Wilson, J.A., Williamson, T.E., 2011. A sauropod dinosaur pes from the latest Canudo, J.I., Royo-Torres, R., Cuenca-Bescós, G., 2008. A new sauropod: Tastavinsaurus Cretaceous of North America and the validity of Alamosaurus sanjuanensis sanzi gen. et sp. nov. from the Early Cretaceous (Aptian) of Spain. Journal of Vertebrate (Sauropoda, Titanosauria). Journal of Vertebrate Paleontology 31, 1072–1079. Paleontology 28, 712–731. D'Emic, M.D., Mannion, P.D., Upchurch, P., Benson, R.B.J., Pang, Q., Cheng, Z., 2013. Osteology Canudo, J.I., Salgado, L., Garrido, A., Carballido, J., 2013. Primera evidencia de dinosaurios of Huabeisaurus allocotus (Sauropoda: Titanosauriformes)fromtheUpperCretaceousof ornitopodos en la base de la Formacion Huincul (Cenomaniense Superior- China. PloS One 8, e69375. Turoniense, Cuenca Neuquina, Argentina). Geogaceta 53, 9–12. Dettmann, M.E., Molnar, R.E., Douglas, J.G., Burger, D., Fielding, C., Clifford, H.T., Francis, J., Carballido, J.L., Sander, P.M., 2014. Postcranial axial skeleton of Europasaurus holgeri, Jell, P., Rich, T., Wade, M., Rich, P.V., Pledge, N., Kemp, A., Rozefelds, A., 1992. (Dinosauria, Sauropoda) from the Upper Jurassic of Germany: implications for sauro- Australian Cretaceous terrestrial faunas and floras: biostratigraphic and biogeographic pod ontogeny and phylogenetic relationships of basal Macronaria. Journal of Systematic implications. Cretaceous Research 13, 207–262. Palaeontology 12, 335–387. Díez Díaz, V., Pereda Suberbiola, X., Sanz, J.L., 2013a. The axial skeleton of the titanosaur Carballido, J.L., Pol, D., Cerda, I., Salgado, L., 2011. The osteology of Chubutisaurus insignis Lirainosaurus astibiae (Dinosauria: Sauropoda) from the latest Cretaceous of Spain. Del Corro, 1975 (Dinosauria: Neosauropoda) from the ‘Middle’ Cretaceous of Central Cretaceous Research 43, 145–160. Patagonia, Argentina. Journal of Vertebrate Paleontology 31, 93–110. Díez Díaz, V., Pereda Suberbiola, X., Sanz, J.L., 2013b. Appendicular skeleton and dermal Carballido, J.L., Salgado, L., Pol, D., Canudo, J.I., Garrido, A., 2012. A new basal armour of the Late Cretaceous titanosaur Lirainosaurus astibiae (Dinosauria: rebbachisaurid (Sauropoda, Diplodocoidea) from the Early Cretaceous of the Neuquén Sauropoda) from Spain. Palaeontologia Electronica 16, 19A. Basin; evolution and biogeography of the group. Historical Biology 24, 631–654. Dingle, R.V., 1999. Walvis Ridge Barrier: its influence on palaeoenvironments and source Carrano, M.T., Hutchinson, J.R., 2002. Pelvic and hindlimb musculature of Tyrannosaurus rock generation deduced from ostracod distributions in the early South Atlantic rex (Dinosauria: Theropoda). Journal of Morphology 253, 207–228. Ocean. In: Cameron, N.R., Bate, R.H., Clure, V.S. (Eds.), The Oil and Gas Habitats of Case, J.A., Martin, J.E., Chaney, D.S., Reguero, M., Marenssi, S.A., Santillana, S.M., the South Atlantic. Geological Society Special Publication, No. 153. Geological Society, Woodburne, M.O., 2000. The first duck-billed dinosaur (family Hadrosauridae) from London, pp. 293–302. Antarctica. Journal of Vertebrate Paleontology 20, 612–614. Dong, Z., Zhou, S., Zhang, Z., 1983. The dinosaur remains from Sichuan Basin, China. Cerda, I.A., 2008. Gastroliths in an ornithopod dinosaur. Acta Palaeontologica Polonica 53, Palaeontologia Sinica 162, 1–145. 351–355. Egerton, V.M., Novas, F.E., Dodson, P., Lacovara, K., 2013. The first record of a neonatal Cerda, I.A., Chinsamy, A., 2012. Biological implications of the bone microstructure of the ornithopod dinosaur from Gondwana. Gondwana Research 23, 268–271. Late Cretaceous ornithopod dinosaur Gasparinisaura cincosaltensis. Journal of Fitzgerald, E.M.G., Carrano, M.T., Holland,T.,Wagstaff,B.E.,Pickering,D.,Rich,T.H.,Vickers- Vertebrate Paleontology 32, 355–368. Rich, P., 2012. First ceratosaurian dinosaur from Australia. Naturwissenschaften 99, Cerda, I.A., Salgado, L., Powell, J.E., 2012. Extreme postcranial pneumaticity in sauropod 397–405. dinosaurs from South America. Paläontologische Zeitschrift 86, 441–449. Flannery, T.F., Archer, M., Rich, T.H., Jones, R., 1995. A new family of monotremes from the Chaboureau, A.-C., Guillocheau, F., Robin, C., Rohais, S., Moulin, M., Aslanian, D., 2013. Cretaceous of Australia. Nature 377, 418–420. Paleogeographic evolution of the central segment of the South Atlantic during Early Flynn, J.J., Parrish, J.M., Rakotosamimanana, B., Simpson, W.F., Wyss, A.R., 1999. A Middle Cretaceous times: paleotopographic and geodynamic implications. Tectonophysics Jurassic mammal from Madagascar. Nature 401, 57–60. 604, 191–223. Frakes, L.A., Burger, D., Apthorpe, M., Wiseman, J., Dettmann, M., Alley, N., Flint, R., Chure, D.J., 1998. A reassessment of the Australian Allosaurus and its implications for Gravestock, D., Ludbrook, N., Backhouse, J., Skwarko, S., Scheibnerova, V., McMinn, the Australian refugium concept. Journal of Vertebrate Paleontology 18, 34A A., Moore, P.S., Bolton, B.R., Douglas, J.G., Christ, R., Wade, M., Molnar, R.E., (supplement). McGowran, B., Balme, B., Day, R.A., 1987. Australian Cretaceous shorelines, stage by Chure, D., Britt, B.B., Whitlock, J.A., Wilson, J.A., 2010. First complete sauropod dinosaur stage. Palaeogeography, Palaeoclimatology, Palaeoecology 59, 31–48. skull from the Cretaceous of the Americas and the evolution of sauropod dentition. Frakes, L.A., Alley, N.F., Deynoux, M., 1995. Early Cretaceous ice rafting and climatic zona- Naturwissenschaften 97, 379–391. tion in Australia. International Geology Review 37, 567–583. Clarke, L.J., Jenkyns, H.C., 1999. New oxygen-isotope evidence for long-term Cretaceous Gaffney, E.S., 1981. A review of the fossil turtles of Australia. American Museum Novitates climatic change in the Southern Hemisphere. Geology 27, 699–702. 2720, 1–38. Close, R.A., Vickers-Rich, P., Trusler, P., Chiappe, L.M., O'Connor, J., Rich, T.H., Kool, L., Gaffney, E.S., Kool, L., Brinkman, D.B., Rich, T.H., Vickers-Rich, P., 1998. Otwayemys, a new Komarower, P., 2009. Earliest Gondwanan bird from the Cretaceous of southeastern cryptodiran turtle from the Early Cretaceous of Australia. American Museum Australia. Journal of Vertebrate Paleontology 29, 616–619. Novitates 3233, 1–28. Constantine, A., Chinsamy, A., Vickers-Rich, P., Rich, T.H., 1998. Periglacial environments Gallagher, K., Lambeck, K., 1989. Subsidence, sedimentation and sea-level changes in the and polar dinosaurs. South African Journal of Science 94, 137–141. Eromanga Basin, Australia. Basin Research 2, 115–131. Cook, A.G., McKenzie, E.D., 1997. The Great Artesian Basin. M'Choinneach Press, García, R.A., Salgado, L., 2013. The titanosaur sauropods from the Allen Formation (late Ilfracombe, Queensland 54 pp. Campanian–early Maastrichtian) of Salitral Moreno (Patagonia, Río Negro, Coombs Jr., W.P., Molnar, R.E., 1981. Sauropoda (Reptilia, Saurischia) from the Cretaceous Argentina). Acta Palaeontologica Polonica 58, 269–284. of Queensland. Memoirs of the Queensland Museum 20, 351–373. Gasparini, Z., Olivero, E., Scasso, R., Rinaldi, C., 1987. Un Ankylosaurio (Reptilia, Coria, R.A., Calvo, J.O., 2002. A new iguanodontian ornithopod from Neuquen Basin, Ornithischia) Campanico en el Continente Antarctico. Anais do X Congresso Brasil Patagonia, Argentina. Journal of Vertebrate Paleontology 22, 503–509. de Paleontologia, Rio De Janeiro, 19–25 Julho, 1987, pp. 131–141. Coria, R.A., Salgado, L., 1996. A basal iguanodontian (Ornithischia: Ornithopoda) from the Gasparini, Z., Pereda-Suberbiola, X., Molnar, R.E., 1996. New data on the ankylosaurian Late Cretaceous of South America. Journal of Vertebrate Paleontology 16, 445–457. dinosaur from the Late Cretaceous of the Antarctic Peninsula. Memoirs of the Coria, R.A., Salgado, L., 2001. South American ankylosaurs. In: Carpenter, K. (Ed.), The Queensland Museum 39, 583–594. Armored Dinosaurs. Indiana University Press, Bloomington, pp. 159–168. Gianechini, F.A., Apesteguía, S., 2011. Unenlagiinae revisited: dromaeosaurid theropods Coria, R.A., Moly, J.J., Reguero, M., Santillana, S., Marenssi, S., 2013. A new ornithopod from South América. Anais da Academia Brasileira de Ciências 83, 163–195. (Dinosauria; Ornithischia) from Antarctica. Cretaceous Research 41, 186–193. Gilmore, C.W., 1922. A new sauropod dinosaur from the Ojo Alamo Formation of New Currie, P.J., Vickers-Rich, P., Rich, T.H., 1996. Possible oviraptorosaur (Theropoda, Mexico. Smithsonian Miscellaneous Collections 72, 1–9. Dinosauria) specimens from the Early Cretaceous Otway Group of Dinosaur Cove, Gilmore, C.W., 1936. Osteology of Apatosaurus with special reference to specimens in the Australia. Alcheringa 20, 73–79. Carnegie Museum. Memoirs of the Carnegie Museum 11, 175–300. Curry Rogers, K.A., 2005. Titanosauria: a phylogenetic overview. In: Curry Rogers, K.A., Gilmore, C.W., 1946. Reptilian fauna of the North Horn Formation of central Utah. United Wilson, J.A. (Eds.), The Sauropods: Evolution and Paleobiology. University of California States Geological Survey Professional Paper 210-C, 29–53. Press, Berkeley, pp. 50–103. Giménez, O., 1992. Estudio preliminar del miembro anterior de los sauropodos Curry Rogers, K., 2009. The postcranial osteology of Rapetosaurus krausei (Sauropoda: titanosauridos. Ameghiniana 30, 154. Titanosauria) from the Late Cretaceous of Madagascar. Journal of Vertebrate Paleon- Goloboff, P.A., Farris, J.S., Nixon, K.C., 2008. TNT: a free program for phylogenetic analysis. tology 29, 1046–1086. Cladistics 24, 774–786. Curry Rogers, K., Forster, C.A., 2001. The last of the dinosaur titans: a new sauropod from Gomani, E.M., 2005. Sauropod dinosaurs from the Early Cretaceous of Malawi, Africa. Madagascar. Nature 412, 530–534. Palaeontologia Electronica 8, 27A. De Lurio, J.L., Frakes, L.A., 1999. Glendonites as a paleoenvironmental tool: implications González Riga, B.J., Previtera, E., Pirrone, C.A., 2009. Malarguesaurus florenciae gen. et sp. for early Cretaceous high latitude climates in Australia. Geochimica et Cosmochimica nov., a new titanosauriform (Dinosauria, Sauropoda) from the Upper Cretaceous of Acta 63, 1039–1048. Mendoza, Argentina. Cretaceous Research 30, 135–148. Del Corro, G., 1975. Un nuevo saurópodo del Cretácico Superior, Chubutisaurus insignis Hatcher, J.B., 1902. Structure of the forelimb and manus of Brontosaurus. Annals of the gen. et sp. nov. (Saurischia, Chubutisauridae nov.) del Cretácico Superior Carnegie Museum 1, 356–376.

Henderson, R.A., Price, G.D., 2012. Paleoenvironment and paleoecology inferred from Longman, H.A., 1924. A new gigantic marine reptile from the Queensland Cretaceous, oxygen and carbon isotopes of tropical mollusks from the Late Cretaceous Kronosaurus queenslandicus new genus and species. Memoirs of the Queensland (Cenomanian) of Bathurst Island, Australia. PALAIOS 27, 617–626. Museum 8, 26–28. Herne, M.C., Nair, J.P., Salisbury, S.W., 2010. Comment on “A southern tyrant reptile”. Longman, H.A., 1926. A giant dinosaur from Durham Downs, Queensland. Memoirs of the Science 329, 1013-c. Queensland Museum 8, 183–194. Hocknull, S.A., Cook, A.G., 2008. Hypsilophodontid (Dinosauria, Ornithischia) from latest Longman, H.A., 1927. The giant dinosaur, Rhoetosaurus brownei. Memoirs of the Queensland Albian, Winton Formation, central Queensland. Memoirs of the Queensland Museum Museum 9, 1–18. 52, 212. Longman, H.A., 1933. A new dinosaur from the Queensland Cretaceous. Memoirs of the Hocknull, S.A., White, M.A., Tischler, T.R., Cook, A.G., Calleja, N.D., Sloan, T., Elliott, D.A., Queensland Museum 10, 131–144. 2009. New mid-Cretaceous (latest Albian) dinosaurs from Winton, Queensland, Lü, J., Azuma, Y., Chen, R., Zheng, W., Jin, X., 2008. A new titanosauriform sauropod from Australia. PloS One 4, e6190. the early Late Cretaceous of Dongyang, Zhejiang Province. Acta Geologica Sinica 82, Holland, W.J., 1915. A new species of Apatosaurus. Annals of the Carnegie Museum 10, 225–235. 143–145. Luo, Z.-X., Cifelli, R.L., Kielan-Jaworowska, Z., 2001. Dual origin of tribosphenic mammals. Holliday, C.M., Gardner, N.M., 2012. A new eusuchian crocodyliform with novel cranial Nature 409, 53–57. integument and its significance for the origin and evolution of Crocodylia. PloS One 7, Luo, Z.-X., Kielan-Jaworowska, Z., Cifelli, R.L., 2002. In quest for a phylogeny of Mesozoic e30471. mammals. Acta Palaeontologica Polonica 47, 1–78. Hulke, J.W., 1882. Note on the os pubis and ischium of Ornithopsis eucamerotus. Quarterly Lydekker, R., 1889. Note on some points in the nomenclature of fossil reptiles and Journal of the Geological Society of London 38, 372–376. amphibians, with preliminary notices of two new species. Geological Magazine 6 Ibiricu, L.M., Martínez, R.D., Lamanna, M.C., Casel, G.A., Luna, M., Harris, J.D., Lacovara, K.J., (Decade III), 325–326. 2010. A medium-sized ornithopod (Dinosauria: Ornithischia) from the Upper Creta- Lydekker, R., 1893. Contributions to the study of the fossil vertebrates of Argentina. I. The ceous Bajo Barreal Formation of Lago Colhué Huapi, southern Chubut Province, dinosaurs of Patagonia. Anales del Museo de la Plata, Seccion de Paleontologia 2, Argentina. Annals of the Carnegie Museum 79, 39–50. 1–14. International Commission on Zoological Nomenclature, 1999. Article 73.1.3, Fourth Mannion, P.D., Calvo, J.O., 2011. Anatomy of the basal titanosaur (Dinosauria, Sauropoda) edition. International Code of Zoological Nomenclature. Andesaurus delgadoi from the mid-Cretaceous (Albian–early Cenomanian) Río Limay Jacobs, L.L., Winkler, D.A., Downs, W.R., Gomani, E.M., 1993. New material of an Early Formation, Neuquén Province, Argentina: implications for titanosaur systematics. Cretaceous titanosaurid sauropod dinosaur from Malawi. Palaeontology 36, Zoological Journal of the Linnean Society 163, 155–181. 523–534. Mannion, P.D., Otero, A., 2012. A reappraisal of the Late Cretaceous Argentinean sauropod Janensch, W., 1922. Das Handskelett von Gigantosaurus und Brachiosaurus brancai aus dinosaur Argyrosaurus superbus, with a description of a new titanosaur genus. Journal den Tendaguruschichten Deutsch-Ostafrika. Centralblatt für Mineralogie, Geologie of Vertebrate Paleontology 32, 614–638. und Paläontologie 1922, 464–480. Mannion, P.D., Benson, R.B.J., Upchurch, P., Butler, R.J., Carrano, M.T., Barrett, P.M., 2012b. Janensch, W., 1961. Die Gliedmaszen und Gliedmaszengürtel der Sauropoden der A temperate palaeodiversity peak in Mesozoic dinosaurs and evidence for Late Creta- Tendaguruschichten. Palaeontographica, Supplement VII 3, 177–235. ceous geographical partitioning. Global Ecology and Biogeography 21, 898–908. Jensen, J.A., 1988. A fourth new sauropod dinosaur from the Upper Jurassic of the Colorado Mannion, P.D., Upchurch, P., Mateus, O., Barnes, R.N., Jones, M.E.H., 2012a. New informa- Plateau and sauropod bipedalism. Great Basin Naturalist 48, 121–145. tion on the anatomy and systematic position of Dinheirosaurus lourinhanensis Jokat, W., Boebel, T., König, M., Meyer, U., 2003. Timing and geometry of early Gondwana (Sauropoda: Diplodocoidea) from the Late Jurassic of Portugal, with a review of breakup. Journal of Geophysical Research 108, 2428. European diplodocoids. Journal of Systematic Palaeontology 10, 521–551. Jupp, R., Warren, A.A., 1986. The mandibles of the Triassic temnospondyl amphibians. Mannion, P.D., Upchurch, P., Barnes, R.N., Mateus, O., 2013. Osteology of the Late Jurassic Alcheringa 10, 99–124. Portuguese sauropod dinosaur Lusotitan atalaiensis (Macronaria) and the evolution- Kear, B.P., 2003. Cretaceous marine reptiles of Australia: a review of taxonomy and distri- ary history of basal titanosauriforms. Zoological Journal of the Linnean Society 168, bution. Cretaceous Research 24, 277–303. 98–206. Kear, B.P., 2006. Marine reptiles from the Lower Cretaceous of South Australia: elements Mantell, G.A., 1852. On the structure of the Iguanodon, and on the fauna and flora of the of a high-latitude cold-water assemblage. Palaeontology 49, 837–856. Wealden Formation. Notices of the Proceedings at the Meetings of the Members of Kear, B.P., 2007. Taxonomic clarification of the Australian elasmosaurid genus the Royal Institution of Great Britain, 1, pp. 141–146. Eromangasaurus, with reference to other austral elasmosaur taxa. Journal of Vertebrate Marsh, O.C., 1877. Notice of new dinosaurian reptiles from the Jurassic Formation. Paleontology 27, 241–246. American Journal of Science 14 (series 3), 514–516. Kear, B.P., Lee, M.S.Y., 2006. A primitive protostegid from Australia and early sea turtle Marsh, O.C., 1878. Principal characters of American Jurassic dinosaurs: part I. American evolution. Biology Letters 2, 116–119. Journal of Science 16 (series 3), 411–416. Kear, B.P., Schroeder, N.I., Lee, M.S.Y., 2006. An archaic crested plesiosaur in opal from the Martin, J.E., 2010. Allodaposuchus Nopsca, 1928 (Crocodylia, Eusuchia), from the Late Lower Cretaceous high latitude deposits of Australia. Biology Letters 2, 615–619. Cretaceous of southern France and its relationships to Alligatoroidea. Journal of Kennedy, W.J., Cooper, M., 1975. Cretaceous ammonite distributions and the opening of Vertebrate Paleontology 30, 756–767. the South Atlantic. Journal of the Geological Society 131, 283–288. Martin, T., Rauhut, O.W.M., 2005. Mandible and dentition of Asfaltomylos patagonicus Kielan-Jaworowska, Z., Cifelli, R.L., Luo, Z., 1998. Alleged Cretaceous placental from down (Australosphenida, Mammalia) and the evolution of tribosphenic teeth. Journal of under. Lethaia 31, 267–268. Vertebrate Paleontology 25, 414–425. Kielan-Jaworowska, Z., Cifelli, R.L., Luo, Z.-X., 2004. Mammals From the Age of Dinosaurs: Martin, V., Buffetaut, E., Suteethorn, V., 1994. A new genus of sauropod dinosaur from the Origins, Evolution, and Structure. Columbia University Press, New York pp. 630. Sao Khua Formation (Late Jurassic or Early Cretaceous) of northeastern Thailand. König, M., Jokat, W., 2006. The Mesozoic breakup of the Weddell Sea. Journal of Geophysical Comptes Rendus des Seánces de l'Académie des Sciences 319, 1085–1092. Research 111, 102. Martin, V., Suteethorn, V., Buffetaut, E., 1999. Description of the type and referred material Koutsoukos, E.A.M., 1992. Late Aptian to Maastrichtian foraminiferal biogeography and of Phuwiangosaurus sirindhornae Martin, Buffetaut and Suteethorn, 1994, a sauropod palaeoceanography of the Sergipe Basin, Brazil. Palaeogeography, Palaeoclimatology, from the Lower Cretaceous of Thailand. Oryctos 2, 39–91. Palaeoecology 92, 295–324. Martinelli, A.G., Forasiepi, A.M., 2004. Late Cretaceous vertebrates from Bajo de Santa Rosa Koutsoukos, E.A.M., Mello, M.R., De Azambuja Filho, N.C., Hart, M.B., Maxwell, J.R., 1991. (Allen Formation), Río Negro province, Argentina, with the description of a new The upper Aptian–Albian succession of the Sergipe Basin, Brazil: an integrated sauropod dinosaur (Titanosauridae). Revista del Museo Argentino de Ciencias paleoenvironmental assessment. AAPG Bulletin 75, 479–498. Naturales 6, 257–305. Ksepka, D.T., Norell, M.A., 2006. Erketu ellisoni, a long-necked sauropod from Bor Guvé Martínez, R.D., 1998. Notohypsilophodon comodorensis gen.etsp.nov.unHypsilophodontidae (Dornogov Aimag, Mongolia). American Museum Novitates 3508, 1–16. (Ornithischia: Ornithopoda) del Cretaceo Superior de Chubut, Patagonia Central, Ksepka, D.T., Norell, M.A., 2010. The illusory evidence for Asian Brachiosauridae: new Argentina. Acta Geologica Leopoldensia 21, 119–135. material of Erketu ellisoni and a phylogenetic reappraisal of basal Titanosauriformes. Martínez, R., Giménez, O., Rodríguez, J., Luna, M., Lamanna, M.C., 2004. An articulated American Museum Novitates 3700, 1–27. specimen of the basal titanosaurian (Dinosauria: Sauropoda) Epachthosaurus sciuttoi Kurochkin, E.N., Molnar, R.E., 1997. New material of enantiornithine birds from the Early from the early Late Cretaceous Bajo Barreal Formation of Chubut Province, Argentina. Cretaceous of Australia. Alcheringa 21, 291–297. Journal of Vertebrate Paleontology 24, 107–120. Läng, É., Goussard, F., 2007. Redescription of the wrist and manus of ?Bothriospondylus Mateus, O., Jacobs, L.L., Schulp, A.S., Polcyn, M.J., Tavares, T.S., Neto, A.B., Morais, M.L., madagascariensis: new data on carpus morphology in Sauropoda. Geodiversitas 29, Antunes, M.T., 2011. Angolatitan adamastor, a new sauropod dinosaur and the first 549–560. record from Angola. Anais da Academia Brasileira de Ciências 83, 221–233. Lapparent, A.F.d., 1960. Les dinosauriens du “Continental Intercalaire” du Sahara Central. Mateus, O., Mannion, P.D., Upchurch, P., 2014. Zby atlanticus, a new turiasaurian sauropod Mémoires de la Société Géologique de France (Nouvelle Série) 39, 1–56. (Dinosauria, Eusauropoda) from the Late Jurassic of Portugal. Journal of Vertebrate Lavocat, R.J.M., 1954. Sur les Dinosauriens du continental intercalaire des Kem-Kemm Paleontology (in press). de la Douara. Comptes Rendus 19th International Geological Congress, 1952, 15, McCoy, F., 1867. On the occurrence of Ichthyosaurus and Plesiosaurus in Australia. Annals pp. 65–68. and Magazine of Natural History 19 (series 3), 355–356. Leahey, L.G., Salisbury, S.W., 2013. First evidence of ankylosaurian dinosaurs McDonald, A.T., Kirkland, J.I., DeBlieux, D.D., Madsen, S.K., Cavin, J., Milner, A.R.C., (Ornithischia: Thyreophora) from the mid-Cretaceous (late Albian–Cenomanian) Panzarin, L., 2010. New basal iguanodonts from the Cedar Mountain Formation of Winton Formation of Queensland, Australia. Alcheringa 37, 249–257. Utah and the evolution of thumb-spiked dinosaurs. PloS One 5, e14075. Lehman, T.M., Coulson, A.B., 2002. A juvenile specimen of the sauropod dinosaur McGowan, C., 1972. The systematics of Cretaceous ichthyosaurs with particular reference Alamosaurus sanjuanensis from the Upper Cretaceous of Big Bend National Park, to the material from North America. University of Wyoming Contributions to Geology Texas. Journal of Paleontology 76, 156–172. 11, 9–29. Longman, H.A., 1915. On a giant turtle from the Queensland Lower Cretaceous. Memoirs McIntosh, J.S., 1990. Sauropoda. In: Weishampel, D.B., Dodson, P., Osmólska, H. (Eds.), The of the Queensland Museum 3, 24–29. Dinosauria. University of California Press, Berkeley, pp. 345–401.

McIntosh, J.S., Miller, W.E., Stadtman, K.L., Gillette, D.D., 1996. The osteology of Pang, Q., Cheng, Z., 2000. A new family of sauropod dinosaur from the Upper Cretaceous Camarasaurus lewisi (Jensen, 1988). Brigham Young University Geology Studies 41, of Tianzhen, Shanxi Province, China. Acta Geologica Sinica 74, 117–125. 73–115. Parmar, V., Prasad, G.V.R., Kumar, D., 2013. The first multituberculate mammal from India. McLoughlin, S., 2001. The breakup of Gondwana and its impact on pre-Cenozoic floristic Naturwissenschaften 100, 515–523. provincialism. Australian Journal of Botany 49, 271–300. Pascual, R., Archer, M., Ortiz Jaureguizar, E., Prado, J.L., Godthelp, H., Hand, S.J., 1992a. First McLoughlin, S., Drinnan, A.N., Rozefelds, A.C., 1995. A Cenomanian flora from the Winton discovery of monotremes in South America. Nature 356, 704–706. Formation, Eromanga Basin, Queensland, Australia. Memoirs of the Queensland Pascual, R., Archer, M., Ortiz Jaureguizar, E., Prado, J.L., Godthelp, H., Hand, S.J., 1992b. The Museum 38, 273–313. first non-Australian monotreme: an Early Paleocene South American platypus McLoughlin, S., Tosolini, A.-M.P., Nagalingum, N.S., Drinnan, A.N., 2002. The Early Creta- (Monotremata, Ornithorhynchidae). In: Augee, M.L. (Ed.), Platypus and Echidnas. ceous (Neocomian) flora and fauna of the lower Strzelecki Group, Gippsland Basin, Royal Zoological Society of New South Wales, Sydney, Australia, pp. 1–14. Victoria, Australia. Association of Australasian Palaeontologists Memoir 26, 1–144. Pascual, R., Goin, F.J., Balarino, L., Udrizar Sauthier, D.E., 2002. New data on the Paleocene McLoughlin, S., Pott, C., Elliott, D., 2010. The Winton Formation flora (Albian– monotreme Monotrematum sudamericanum, and the convergent evolution of trian- Cenomanian, Eromanga Basin): implications for vascular plant diversification and gulate molars. Acta Palaeontologica Polonica 47, 487–492. decline in the Australian Cretaceous. Alcheringa 34, 303–323. Paul, G.S., 1988. The brachiosaur giants of the Morrison and Tendaguru with a description Mo, J., Wang, W., Huang, Z., Huang, X., Xu, X., 2006. A basal titanosauriform from the Early of a new subgenus, Giraffatitan, and a comparison of the world's largest dinosaurs. Cretaceous of Guangxi, China. Acta Geologica Sinica 80, 486–489. Hunteria 3, 1–14. Mohriak, W., Nemčok, M., Enciso, G., 2008. South Atlantic divergent margin evolution: Phillips, M.J., Bennett, T.H., Lee, M.S.Y., 2009. Molecules, morphology, and ecology indicate rift-border uplift and salt tectonics in the basins of SE Brazil. In: Pankhurst, R.J., a recent, amphibious ancestry for echidnas. Proceedings of the National Academy of Trouw, R.A.J., Brito Neves, B.B., De Wit, M.J. (Eds.), West Gondwana: Pre-Cenozoic Sciences of the United States of America 106, 17089–17094. Correlations Across the South Atlantic Region. Geological Society Special Publication, Poropat, S.F., 2013. Carl Wiman's sauropods: the Uppsala Museum of Evolution's 294. Geological Society, London, pp. 365–398. collection. GFF 135, 104–119. Molnar, R., 1980a. Australian late Mesozoic terrestrial tetrapods: some implications. Poropat, S.F., Mannion, P.D., Upchurch, P., Hocknull, S.A., Kear, B.P., Elliott, D.A., 2014. Mémoires de la Société Géologique de France (Nouvelle Série) 139, 131–143. Reassessment of the non-titanosaurian somphospondylan Wintonotitan wattsi Molnar, R.E., 1980b. An ankylosaur (Ornithischia: Reptilia) from the Lower Cretaceous of (Dinosauria: Sauropoda: Titanosauriformes) from the mid-Cretaceous Winton southern Queensland. Memoirs of the Queensland Museum 20, 77–87. Formation, Queensland, Australia. Special Papers in Palaeontology (in press). Molnar, R.E., 1986. An enantiornithine bird from the Lower Cretaceous of Queensland, Powell, J.E., 1987. The Late Cretaceous fauna of Los Alamitos, Patagonia, Argentina. Part VI. Australia. Nature 322, 736–738. The titanosaurids. Revista del Museo Argentino de Ciencias Naturales 3, 147–153. Molnar, R.E., 1996a. Observations on the Australian ornithopod Muttaburrasaurus. Powell, J.E., 1990. Epachthosaurus sciuttoi gen. et sp. nov, un nuevo dinosaurio Memoirs of the Queensland Museum 39, 639–652. titanosáurido del Cretácico de Patagonia (Provincia de Chubut, Argentina). Actas del Molnar, R.E., 1996b. Preliminary report on a new ankylosaur from the Early Cretaceous of V Congreso Argentino de Paleontología y Bioestratigrafía, Tucumán, 1, pp. 123–128. Queensland, Australia. Memoirs of the Queensland Museum 39, 653–668. Powell, J.E., 1992. Osteologia de Saltasaurus loricatus (Sauropoda–Titanosauridae) del Molnar, R.E., 2001. A reassessment of the phylogenetic position of Cretaceous sauropod Cretácico Superior del noroeste Argentino. In: Sanz, J.L., Buscalioni, A.D. (Eds.), Los dinosaurs from Queensland, Australia. In: Leanza, H.A. (Ed.), VII International Sympo- Dinosaurios y Su Entorno Biótico: Actas del Segundo Curso de Paleontologia in Cuenca. sium on Mesozoic Terrestrial Ecosystems. Asociacíon Paleontológica Argentina Institutio Juan de Valdes, pp. 165–230. Publicación Especial, No. 7. Asociacion Paleontologica Argentina, Buenos Aires, Powell, J.E., 2003. Revision of South American titanosaurid dinosaurs: palaeobiological, pp. 139–144. palaeobiogeographical and phylogenetic aspects. Records of the Queen Victoria Molnar, R.E., 2011. New morphological information about Cretaceous sauropod dinosaurs Museum 111, 1–173. from the Eromanga Basin, Queensland, Australia. Alcheringa 35, 329–339. Price, G.D., Williamson, T., Henderson, R.A., Gagan, M.K., 2012. Barremian–Cenomanian Molnar, R.E., Frey, E., 1987. The paravertebral elements of the Australian ankylosaur palaeotemperatures for Australian seas based on new oxygen-isotope data from bel- Minmi (Reptilia: Ornithischia, Cretaceous). Neues Jahrbuch für Geologie und emnite rostra. Palaeogeography, Palaeoclimatology, Palaeoecology 358–360, 27–39. Paläontologie, Abhandlungen 175, 19–37. Pridmore, P.A., Rich, T.H., Vickers-Rich, P., Gambaryan, P.P., 2005. A tachyglossid-like Molnar, R.E., Galton, P.M., 1986. Hypsilophodontid dinosaurs from Lightning Ridge, New humerus from the Early Cretaceous of south-eastern Australia. Journal of Mammalian South Wales, Australia. Geobios 19, 231–243. Evolution 12, 359–378. Molnar, R.E., Salisbury, S.W., 2005. Observations on Cretaceous sauropods from Australia. Rabinowitz, P.D., Coffin, M.F., Falvey, D., 1983. The separation of Madagascar and Africa. In: Tidwell, V., Carpenter, K. (Eds.), Thunder-lizards: The Sauropodomorph Dinosaurs. Science 220, 67–69. Indiana University Press, Bloomington & Indianapolis, pp. 454–465. Rauhut, O.W.M., Martin, T., Ortiz-Jaureguizar, E., Puerta, P., 2002. A Jurassic mammal from Molnar, R.E., Flannery, T.F., Rich, T.H.V., 1981. An allosaurid theropod dinosaur from the South America. Nature 416, 165–168. Early Cretaceous of Victoria, Australia. Alcheringa 5, 141–146. Rauhut, O.W.M., Milner, A.C., Moore-Fay, S., 2010. Cranial osteology and phylogenetic Molnar, R.E., Flannery, T.F., Rich, T.H.V., 1985. Aussie Allosaurus after all. Journal of position of the theropod dinosaur Proceratosaurus bradleyi (Woodward, 1910) from Paleontology 59, 1511–1513. the Middle Jurassic of England. Zoological Journal of the Linnean Society 158, 155–195. Nair, J.P., Salisbury, S.W., 2012. New anatomical information on Rhoetosaurus Raval, U., Veeraswamy, K., 2003. India–Madagascar separation: breakup along a pre- brownei Longman, 1926, a gravisaurian sauropodomorph dinosaur from the existing mobile belt and chipping of the craton. Gondwana Research 6, 467–485. Middle Jurassic of Queensland, Australia. Journal of Vertebrate Paleontology Remes, K., 2007. A second Gondwanan diplodocid dinosaur from the Upper Jurassic 32, 369–394. Tendaguru Beds of Tanzania, East Africa. Palaeontology 50, 653–667. Néraudeau, D., Mathey, B., 2000. Biogeography and diversity of South Atlantic Cretaceous Rich, T.H., 2008. The palaeobiogeography of Mesozoic mammals: a review, 66. Arquivos echinoids: implications for circulation patterns. Palaeogeography, Palaeoclimatology, do Museu Nacional, Rio de Janeiro pp. 231–249. Palaeoecology 156, 71–88. Rich, T.H., Rich, P.V., 1988. A juvenile dinosaur brain from Australia. National Geographic Norman, D.B., 2004. Basal Iguanodontia, In: Weishampel, D.B., Dodson, P., Osmólska, Research 4, 148. H. (Eds.), The Dinosauria, Second edition. University of California Press, Berkeley, Rich, T.H., Rich, P.V., 1989. Polar dinosaurs and biotas of the Early Cretaceous of southeastern pp. 413–437. Australia. National Geographic Research 5, 15–53. Norman, D.B., Weishampel, D.B., 1990. Iguanodontidae and related ornithopods. In: Rich, T.H., Vickers-Rich, P., 1994. Neoceratopsians and ornithomimosaurs: dinosaurs of Weishampel, D.B., Dodson, P., Osmólska, H. (Eds.), The Dinosauria. University of Gondwana origin? National Geographic Research and Exploration 10, 129–131. California Press, Berkeley, pp. 510–533. Rich, T.H., Vickers-Rich, P., 1999. The Hypsilophodontidae from southeastern Australia. In: Novas, F.E., 1998. Megaraptor namunhuaiquii, gen. et sp. nov., a large-clawed, Late Tomida, Y., Rich, T.H., Vickers-Rich, P. (Eds.), Second Gondwanan Dinosaur Sympo- Cretaceous theropod from Patagonia. Journal of Vertebrate Paleontology 18, sium. National Science Museum Monograph, 15. National Science Museum, Tokyo, 4–9. Japan, pp. 167–180. Novas, F.E., Cambiaso, A.V., Ambrosio, A., 2004. A new basal iguanodontian (Dinosauria, Rich, T.H., Vickers-Rich, P., 2003. Protoceratopsian? Ulnae from Australia. Records of the Ornithischia) from the Upper Cretaceous of Patagonia. Ameghiniana 41, 75–82. Queen Victoria Museum 113, 1–12. Novas, F.E., Agnolín, F.L., Ezcurra, M.D., Porfiri, J., Canale, J.I., 2013. Evolution of the carniv- Rich, T.H., Vickers-Rich, P., 2004. Diversity of Early Cretaceous mammals from Victoria, orous dinosaurs during the Cretaceous: the evidence from Patagonia. Cretaceous Australia. Bulletin of the American Museum of Natural History 285, 36–53. Research 45, 174–215. Rich, T.H., Vickers-Rich, P., 2012a. Australia's polar Early Cretaceous dinosaurs. In: Osborn, H.F., Mook, C.C., 1921. Camarasaurus, Amphicoelias, and other sauropods of Cope. Godefroit, P. (Ed.), Bernissart Dinosaurs and Early Cretaceous Terrestrial Ecosystems. Memoirs of the American Museum of Natural History (New Series) 3, 247–387. Indiana University Press, Bloomington & Indianapolis, pp. 488–503. Ostrom, J.H., McIntosh, J.S., 1966. Marsh's Dinosaurs: The Collections From Como Bluff. Rich, T.H., Vickers-Rich, P., 2012b. Palaeobiogeography of Mesozoic mammals — revisited. Yale University Press, New Haven 338 pp. In: Talent, J.A. (Ed.), Earth and Life. Springer, Netherlands, pp. 913–934. Otero, A., 2010. The appendicular skeleton of Neuquensaurus, a Late Cretaceous Rich, P.V., Rich, T.H., Wagstaff, B.E., McEwen Mason, J., Douthitt, C.B., Gregory, R.T., Felton, saltasaurine sauropod from Patagonia, Argentina. Acta Palaeontologica Polonica 55, E.A., 1988. Evidence for low temperatures and biologic diversity in Cretaceous high 399–426. latitudes of Australia. Science 242, 1403–1406. Owen, R., 1841. A description of a portion of the skeleton of the Cetiosaurus, a gigantic Rich, T.H., Rich, P.V., Wagstaff, B., McEwen-Mason, J., Douthitt, C.B., Gregory, R.T., 1989. extinct saurian reptile occurring in the Oolitic formations of different portions of Early Cretaceous biota from the northern side of the Australo-Antarctic rift valley. England. Proceedings of the Geological Society of London 3, 457–462. In: Crame, J.A. (Ed.), Origins and Evolution of the Antarctic Biota. Geological Society Owen, R., 1842. Report on British fossil reptiles. Part II. Report of the Eleventh Meeting of Special Publication, 47. Geological Society, London, pp. 121–130. the British Association for the Advancement of Science, Held at Plymouth, July 1841, Rich, T.H., Rich, P.V., Wagstaff, B., McEwen Mason, J., Flannery, T.F., Archer, M., Molnar, R.E., pp. 60–204. Long, J.A., 1992. Two possible chronological anomalies in the Early Cretaceous tetra- Owen, R., 1882. On an extinct chelonian reptile (Notochelys costata, Owen), from pod assemblages of southeastern Australia. In: Mateer, N.J., Chen, P.-J. (Eds.), Aspects Australia. Quarterly Journal of the Geological Society of London 38, 178–183. of Nonmarine Cretaceous Geology. China Ocean Press, Beijing, pp. 165–176.

Rich, T.H., Vickers-Rich, P., Constantine, A., Flannery, T.F., Kool, L., Van Klaveren, N., 1997. A Seeley, H.G., 1887. On the classiﬁcation of the fossil animals commonly named Dinosauria. tribosphenic mammal from the Mesozoic of Australia. Science 278, 1438–1442. Proceedings of the Royal Society of London 43, 165–171. Rich, T.H., Flannery, T.F., Vickers-Rich, P., 1998. Alleged Cretaceous placental from down Sereno, P.C., Beck, A.L., Dutheil, D.B., Larsson, H.C.E., Lyon, G.H., Moussa, B., Sadleir, R.W., under: reply. Lethaia 31, 346–348. Sidor, C.A., Varricchio, D.J., Wilson, G.P., Wilson, J.A., 1999. Cretaceous sauropods Rich, T.H., Vickers-Rich, P., Constantine, A., Flannery, T.F., Kool, L., Van Klaveren, N., 1999a. from the Sahara and the uneven rate of skeletal evolution among dinosaurs. Science Early Cretaceous mammals from Flat Rocks, Victoria, Australia. Records of the Queen 286, 1342–1347. Victoria Museum 106, 1–34. Sereno, P.C., Wilson, J.A., Witmer, L.M., Whitlock, J.A., Maga, A., Ide, O., Rowe, T.A., 2007. Rich, T.H., Vickers-Rich, P., Giménez, O., Cúneo, R., Puerta, P., Vacca, R., 1999b. Anew Structural extremes in a Cretaceous dinosaur. PloS One 2, e1230. sauropod dinosaur from Chubut Province, Argentina. In: Tomida, Y., Rich, T.H., Sheard, M.J., 1990. Glendonites from southern Eromanga Basin in South Australia: Vickers-Rich, P. (Eds.), Second Gondwanan Dinosaur Symposium. National Science palaeoclimatic indicators for Cretaceous ice. Quarterly Geological Notes 114, Museum Monograph, 15. National Science Museum, Tokyo, Japan, pp. 61–84. 17–23. Rich, T.H., Flannery, T.F., Trusler, P., Kool, L., Van Klaveren, N.A., Vickers-Rich, P., 2001. A Sigogneau-Russell, D., Hooker, J.J., Ensom, P.C., 2001. The oldest tribosphenic mammal second tribosphenic mammal from the Mesozoic of Australia. Records of the Queen from Laurasia (Purbeck Limestone Group, Berriasian, Cretaceous, UK) and its bearing Victoria Museum 110, 1–9. on the ‘dual origin’ of Tribosphenida. Comptes Rendus de l'Académie des Sciences — Rich, T.H., Vickers-Rich, P., Flannery, T.F., Kear, B.P., Cantrill, D.J., Komarower, P., Kool, L., Series IIA — Earth and Planetary Science 333, 141–147. Pickering, D., Trusler, P., Morton, S., Van Klaveren, N., Fitzgerald, E.M.G., 2009. An Smith, E.T., 2010. Early Cretaceous chelids from Lightning Ridge, New South Wales. Australian multituberculate and its palaeobiogeographic implications. Acta Alcheringa 34, 375–384. Palaeontologica Polonica 54, 1–6. Smith Woodward, A., 1906. A tooth of Ceratodus and a dinosaurian claw from the Lower Rich, T.H., Kear, B.P., Sinclair, R., Chinnery, B., McHugh, M.L., Vickers-Rich, P., 2014. Jurassic of Victoria, Australia. Annals and Magazine of Natural History 18 (series 7), Serendipaceratops arthurcclarkei Rich & Vickers-Rich, 2003 is an Australian Early 1–3. Cretaceous ceratopsian. Alcheringa (in press). Smith, E.T., Kear, B.P., 2013. Spoochelys ormondea gen. et sp. nov., an archaic meiolaniid-like Riggs, E.S., 1903. Brachiosaurus altithorax, the largest known dinosaur. American Journal of turtle from the Early Cretaceous of Lightning Ridge, Australia. In: Brinkman, D.B., Science 15 (Series 4), 299–306. Holroyd, P.A., Gardner, J.D. (Eds.), Morphology and Evolution of Turtles. Springer, Romer, A.S., 1923. The pelvic musculature of saurischian dinosaurs. Bulletin of the Dordrecht, pp. 121–146. American Museum of Natural History 48, 605–617. Smith, N.D., Makovicky, P.J., Agnolin, F.L., Ezcurra, M.D., Pais, D.F., Salisbury, S.W., 2008. A Romer, A.S., 1956. Osteology of the Reptiles. University of Chicago Press, Chicago 772 pp. Megaraptor-like theropod (Dinosauria: Tetanurae) in Australia: support for faunal Romilio, A., Salisbury, S.W., 2011. A reassessment of large theropod dinosaur tracks from exchange across eastern and western Gondwana in the Mid-Cretaceous. Proceedings the mid-Cretaceous (late Albian–Cenomanian) Winton Formation of Lark Quarry, of the Royal Society B 275, 2085–2093. central-western Queensland, Australia: a case for mistaken identity. Cretaceous Sterli, J., De la Fuente, M.S., 2011. A new turtle from the La Colonia Formation (Campanian– Research 32, 135–142. Maastrichtian), Patagonia, Argentina, with remarks on the evolution of the vertebral Romilio, A., Tucker, R.T., Salisbury, S.W., 2013. Reevaluation of the Lark Quarry dino- column in turtles. Palaeontology 54, 63–78. saur Tracksite (late Albian–Cenomanian Winton Formation, central-western Sterli, J., De la Fuente, M.S., 2013. New evidence from the Palaeocene of Patagonia Queensland, Australia): no longer a stampede? Journal of Vertebrate Paleontology (Argentina) on the evolution and palaeo-biogeography of Meiolaniformes 33, 102–120. (Testudinata, new taxon name). Journal of Systematic Palaeontology 11, Royo-Torres, R., Alcalá, L., Cobos, A., 2012. A new specimen of the Cretaceous sauropod 835–852. Tastavinsaurus sanzi from El Castellar (Teruel, Spain), and a phylogenetic analysis of Sterli, J., De la Fuente, M.S., Umazano, A.M., 2014. New remains and new insights on the the Laurasiformes. Cretaceous Research 34, 61–83. Gondwanan meiolaniform turtle Chubutemys copelloi from the Lower Cretaceous of Russell, D.A., Zheng, Z., 1993. A large mamenchisaurid from the Junggar Basin, Patagonia, Argentina. Gondwana Research (in press, available online 9 September Xinjiang, People's Republic of China. Canadian Journal of Earth Sciences 30, 2013). 2082–2095. Suteethorn, S., Le Loeuff, J., Buffetaut, E., Suteethorn, V., Talumbook, C., Chonglakmani, C., Salgado, L., Azpilicueta, C., 2000. Un nuevo saltasaurino (Sauropoda, Titanosauridae) de la 2009. A new skeleton of Phuwiangosaurus sirindhornae (Dinosauria, Sauropoda) from provincia de Río Negro (Formación Allen, Cretácico Superior), Patagonia, Argentina. NE Thailand. In: Buffetaut, E., Cuny, G., Le Loeuff, J., Suteethorn, V. (Eds.), Late Ameghiniana 37, 259–264. Palaeozoic and Mesozoic Ecosystems in SE Asia. The Geological Society, London, Salgado, L., Bonaparte, J.F., 1991. Un nuevo sauropodo Dicraeosauridae Amargasaurus Special Publications, 315. Geological Society, London, pp. 189–215. cazaui gen. et sp. nov., de la Formación la Amarga, Neocomiano de la Provincia del Tang, F., Kang, X.-M., Jin, X.-S., Wei, F., Wu, W.-T., 2001. A new sauropod dinosaur of Neuquén, Argentina. Ameghiniana 28, 333–346. Cretaceous from Jiangshan, Zhejiang Province. Vertebrata PalAsiatica 39, 272–281. Salgado, L., Carvalho, I.d.S., 2008. Uberabatitan ribeiroi, a new titanosaur from the Marília Tapia, A., 1918. Uña mandibula de dinosaurio procedente de Patagonia. Physis 4, Formation (Bauru Group, Upper Cretacous), Minas Gerais, Brazil. Palaeontology 51, 369–370. 881–901. Taylor, M.P., 2009. A re-evaluation of Brachiosaurus altithorax Riggs 1903 (Dinosauria, Salgado, L., Coria, R.A., 1996. First evidence of an ankylosaur (Dinosauria: Ornithischia) in Sauropoda) and its generic separation from Giraffatitan brancai (Janensch 1914). South America. Ameghiniana 33, 367–371. Journal of Vertebrate Paleontology 29, 787–806. Salgado, L., Coria, R.A., 2009. Barrosasaurus casamiquelai gen. et sp. nov., a new titanosaur Thulborn, R.A., 2013. Lark Quarry revisited: a critique of methods used to identify a large (Dinosauria, Sauropoda) from the Anacleto Formation (Late Cretaceous: early dinosaurian track-maker in the Winton Formation (Albian–Cenomanian), western Campanian) of Sierra Barrosa (Neuquén, Argentina). Zootaxa 2222, 1–16. Queensland, Australia. Alcheringa 37, 312–330. Salgado, L., Gasparini, Z., 2006. Reappraisal of an ankylosaurian dinosaur from the Upper Thulborn, T., Turner, S., 2003. The last dicynodont: an Australian relict. Proceedings of the Cretaceous of James Ross Island (Antarctica). Geodiversitas 28, 119–135. Royal Society B 270, 985–993. Salgado, L., Coria, R.A., Calvo, J.O., 1997a. Evolution of titanosaurid sauropods. I: Phylogenetic Thulborn, R.A., Wade, M., 1979. Dinosaur stampede in the Cretaceous of Queensland. analysis based on the postcranial evidence. Ameghiniana 34, 3–32. Lethaia 12, 275–279. Salgado, L., Coria, R.A., Heredia, S.E., 1997b. New materials of Gasparinisaura cincosaltensis Thulborn, R.A., Wade, M., 1984. Dinosaur trackways in the Winton Formation (Ornithischia, Ornithopoda) from the Upper Cretaceous of Argentina. Journal of (mid-Cretaceous) of Queensland. Memoirs of the Queensland Museum 21, Paleontology 71, 933–940. 413–517. Salgado,L.,Garrido,A.,Cocca,S.E.,Cocca,J.R.,2004.Lower Cretaceous rebbachisaurid Tidwell, V., Carpenter, K., Meyer, S., 2001. New titanosauriform (Sauropoda) from the sauropods from Cerro Aguada del León(LohanCuraFormation),Neuquén Poison Strip Member of the Cedar Mountain Formation (Lower Cretaceous), Utah. Province, northwestern Patagonia, Argentina. Journal of Vertebrate Paleontology In: Tanke, D.H., Carpenter, K. (Eds.), Mesozoic Vertebrate Life: New Research Inspired 24, 903–912. by the Paleontology of Philip J. Currie. Indiana University Press, Bloomington & Salgado, L., Apesteguía, S., Heredia, S.E., 2005. A new specimen of Neuquensaurus australis, Indianapolis, pp. 139–165. a Late Cretaceous saltasaurine titanosaur from north Patagonia. Journal of Vertebrate Torsvik, T.H., Amundsen, H., Hartz, E.H., Corfu, F., Kusznir, N., Gaina, C., Doubrovine, P.V., Paleontology 25, 623–634. Steinberger, B., Ashwal, L.D., Jamtveit, B., 2013. A Precambrian microcontinent in Salisbury, S.W., Frey, E., Martill, D.M., Buchy, M.-C., 2003. A new crocodilian from the the Indian Ocean. Nature Geoscience 6, 223–227. Lower Cretaceous Crato Formation of north-eastern Brazil. Palaeontographica Tucker, R.T., Roberts, E.M., Hu, Y., Kemp, A.I.S., Salisbury, S.W., 2013. Detrital zircon age Abteilung A 270, 3–47. constraints for the Winton Formation, Queensland: contextualizing Australia's Late Salisbury, S., Molnar, R.E., Frey, E., Willis, P.M.A., 2006. The origin of modern Cretaceous dinosaur faunas. Gondwana Research 24, 767–779. crocodyliforms: new evidence from the Cretaceous of Australia. Proceedings of the Upchurch, P., 1994. Manus claw function in sauropod dinosaurs. In: Lockley, M.G., Santos, Royal Society B 273, 2439–2448. V.F.,Meyer,C.A.,Hunt,A.P.(Eds.),AspectsofSauropodPaleobiology.Gaia,10. Sander, P.M., Mateus, O., Laven, T., Knötschke, N., 2006. Bone histology indicates insular Museu Nacional de Historia Natural, Lisbon, Portugal, pp. 161–172. dwarﬁsm in a new Late Jurassic sauropod dinosaur. Nature 441, 739–741. Upchurch, P., 1995. The evolutionary history of sauropod dinosaurs. Philosophical Transac- Sanz, J.L., Powell, J.E., Le Loeuff, J., Martínez, R., Pereda Suberbiola, X., 1999. Sauropod tions: Biological Sciences 349, 365–390. remains from the Upper Cretaceous of Laño (northcentral Spain), titanosaur Upchurch, P., 1998. The phylogenetic relationships of sauropod dinosaurs. Zoological phylogenetic relationships. Estudios del Museo de Ciencias Naturales de Alava Journal of the Linnean Society 124, 43–103. 14, 135–155. Upchurch, P., 1999. The phylogenetic relationships of the Nemegtosauridae (Saurischia, Scanlon, J.D., Hocknull, S.A., 2008. A dolichosaurid lizard from the latest Albian (mid- Sauropoda). Journal of Vertebrate Paleontology 19, 106–125. Cretaceous) Winton Formation, Queensland, Australia. In: Everhart, M.J. (Ed.), Upchurch, P., 2008. Gondwana break-up: legacies of a lost world? Trends in Ecology & Proceedings of the Second Mosasaur Meeting. Fort Hays Studies, Special Issue, 3. Evolution 23, 229–236. Fort Hays State University, Hays, Kansas, pp. 131–136. Upchurch, P., Martin, J., 2003. The anatomy and taxonomy of Cetiosaurus (Saurischia, Schoch, R.R., 2013. The evolution of major temnospondyl clades: an inclusive phylogenetic Sauropod) from the Middle Jurassic of England. Journal of Vertebrate Paleontology analysis. Journal of Systematic Palaeontology 11, 673–705. 23, 208–231.

Upchurch, P., Barrett, P.M., Dodson, P., 2004. Sauropoda, In: Weishampel, D.B., Dodson, P., Wilson, J.A., 1999. A nomenclature for vertebral laminae in sauropods and other saurischian Osmólska, H. (Eds.), The Dinosauria, Second edition. University of California Press, dinosaurs. Journal of Vertebrate Paleontology 19, 639–653. Berkeley, pp. 259–322. Wilson, J.A., 2002. Sauropod dinosaur phylogeny: critique and cladistic analysis. Zoological Upchurch, P., Barrett, P.M., Galton, P.M., 2007. A phylogenetic analysis of basal Journal of the Linnean Society 136, 217–276. sauropodomorph relationships: implications for the origin of sauropod dinosaurs. Wilson, J.A., 2011. Anatomical terminology for the sacrum of sauropod dinosaurs. Contri- In: Barrett, P.M., Batten, D.J. (Eds.), Evolution and Palaeobiology of Early butions from the Museum of Paleontology, University of Michigan 32, 59–69. Sauropodomorph Dinosaurs. Special Papers in Palaeontology, 77, pp. 57–90. Wilson, J.A., Carrano, M.T., 1999. Titanosaurs and the origin of “wide-gauge” trackways: a Upchurch, P., Mannion, P.D., Barrett, P.M., 2011. Sauropod dinosaurs. In: Batten, D.J. (Ed.), biomechanical and systematic perspective on sauropod locomotion. Paleobiology 25, English Wealden Fossils; Field Guides to Fossils 14. The Palaeontological Association, 252–267. London, pp. 476–525. Wilson, J.A., Sereno, P.C., 1998. Early evolution and higher-level phylogeny of sauropod Veevers, J.J., 2006. Updated Gondwana (Permian–Cretaceous) earth history of Australia. dinosaurs. Society of Vertebrate Paleontology Memoir 5, 1–68. Gondwana Research 9, 231–260. Wilson, J.A., Upchurch, P., 2009. Redescription and reassessment of Euhelopus zdanskyi Veevers, J.J., Powell, C.M., Roots, S.R., 1991. Review of seafloor spreading around Australia. (Dinosauria: Sauropoda) from the Early Cretaceous of China. Journal of Systematic I. synthesis of the patterns of spreading. Australian Journal of Earth Sciences 38, Palaeontology 7, 199–239. 373–389. Wilson, J.A., D'Emic, M.D., Ikejiri, T., Moacdieh, E.M., Whitlock, J.A., 2011. A nomenclature Vickers-Rich, P., 1996. Early Cretaceous polar tetrapods from the Great Southern for vertebral fossae in sauropods and other saurischian dinosaurs. PloS One 6, Rift Valley, southeastern Australia. Memoirs of the Queensland Museum 39, 719–723. e17114. Vickers-Rich, P., Rich, T.H., Constantine, A., 1999. Environmental setting of the polar Wilson, G.P., Evans, A.R., Corfe, I.J., Smits, P.D., Fortelius, M., Jernvall, J., 2013. Adaptive faunas of southeastern Australia and adaptive strategies of dinosaurs. In: Tomida, Y., radiation of multituberculate mammals before the extinction of dinosaurs. Nature Rich, T.H., Vickers-Rich, P. (Eds.), Second Gondwanan Dinosaur Symposium. National 483, 457–460. Science Museum Monograph, 15. National Science Museum, Tokyo, Japan, pp. Wiman, C., 1929. Die Kreide-Dinosaurier aus Shantung. Palaeontologia Sinica (Series C) 6, 181–195. 1–67. Vine, R.R., Jauncey, W., 1964. Mackunda 1:250 000 Geological Series Sheet SF 54-11. Woodburne, M.O., Rich, T.H., Springer, M.S., 2003. The evolution of tribospheny and the Bureau of Mineral Resources, Canberra. antiquity of mammalian clades. Molecular Phylogenetics and Evolution 28, 360–385. von Huene, F., 1929. Los Saurisquios y Ornitisquios del Cretáceo Argentina. Anales Museo Xu, X., Zhang, X., Tan, Q., Zhao, X., Tan, L., 2006a. A new titanosaurian sauropod from Late de La Plata 3, 1–196. Cretaceous of Nei Mongol, China. Acta Geologica Sinica 80, 20–26. von Huene, F., 1932. Die fossile Reptil-Ordnung Saurischia, ihre Entwicklung und Xu, X., Clark, J.M., Forster, C.A., Norell, M.A., Erickson, G.M., Eberth, D.A., Jia, C., Zhao, Q., Geschichte. Monographien zur Geologie und Paläontologie 4 (Series 1), 1–361. 2006b. A basal tyrannosauroid dinosaur from the Late Jurassic of China. Nature 439, Wagstaff, B.E., Gallagher, S.J., Norvick, M.S., Cantrill, D.J., Wallace, M.W., 2013. High 715–718. latitude Albian climate variability: palynological evidence for long-term drying in a Xu, X., Forster, C.A., Clark, J.M., Mo, J., 2006c. A basal ceratopsian with transitional features greenhouse world. Palaeogeography, Palaeoclimatology, Palaeoecology 386, 501–511. from the Late Jurassic of northwestern China. Proceedings of the Royal Society B 273, Wang, X., You, H., Meng, Q., Gao, C., Chang, X., Liu, J., 2007. Dongbeititan dongi, the first 2135–2140. sauropod dinosaur from the Lower Cretaceous Jehol Group of western Liaoning Yates, A.M., 2007. The first complete skull of Melanorosaurus Haughton Province, China. Acta Geologica Sinica 81, 911–916. (Sauropodomorpha: Anchisauria). In: Barrett, P.M., Batten, D.J. (Eds.), Evolution and Warren, J.W., 1969. A fossil chelonian of probably Lower Cretaceous age from Victoria, Palaeobiology of Early Sauropodomorph Dinosaurs. Special Papers in Palaeontology, Australia. Memoirs of the National Museum of Victoria 29, 23–28. 77, pp. 9–55. Warren, A., Rich, T.H., Vickers-Rich, P., 1997. The last labyrinthodonts? Palaeontographica You, H.-L., Li, D.-Q., 2009. The first well-preserved Early Cretaceous brachiosaurid Abteilung A 247, 1–24. dinosaur in Asia. Proceedings of the Royal Society B 276, 4077–4082. Weishampel, D.B., Barrett, P.M., Coria, R.A., Le Loeuff, J., Xu, X., Zhao, X., Sahni, A., Gomani, Young, C.-C., 1935. Dinosaurian remains from Mengyin, Shantung. Bulletin of the Geological E.M.P., Noto, C.R., 2004. Dinosaur distribution, In: Weishampel, D.B., Dodson, P., Society of China 14, 519–534. Osmólska, H. (Eds.), The Dinosauria, Second edition. University of California Press, Young, C.-C., Zhao, X.-J., 1972. Mamenchisaurus hochuanensis sp. nov. Monograph Series I, Berkeley, pp. 517–606. 8. Institute of Vertebrate Paleontology and Paleoanthropology pp. 1–30. Welles, S.P., 1983. Allosaurus (Saurischia: Theropoda) not yet in Australia. Journal of Zaher, H., Pol, D., Carvalho, A.B., Nascimento, P.M., Riccomini, C., Larson, P., Juarez-Valieri, Paleontology 57, 196. R., Pires-Domingues, R., Silva Jr., N.J.d., Campos, D.d.A., 2011. A complete skull of an White, M.A., Cook, A.G., Hocknull, S.A., Sloan, T., Sinapius, G.H.K., Elliott, D.A., 2012. New Early Cretaceous sauropod and the evolution of advanced titanosaurians. PloS One forearm elements discovered of holotype specimen Australovenator wintonensis 6, e16663. from Winton, Queensland, Australia. PloS One 7, e39364. Zhang, Y., 1988. Sauropod dinosaur (1) Shunosaurus. The Middle Jurassic Dinosaurian White,M.A.,Benson,R.B.J.,Tischler,T.R.,Hocknull,S.A.,Cook,A.G.,Barnes,D.G.,Poropat,S.F., Fauna from Dashanpu, Zigong, Sichuan. Sichuan Publishing House of Science and Wooldridge,S.J.,Sloan,T.,Sinapius,G.H.K.,Elliott,D.A.,2013.New Australovenator hind Technology, Chengdu, Sichuan pp. 1–89. limb elements pertaining to the holotype reveal the most complete neovenatorid leg. Zhang, X., Lü, J., Xu, L., Li, J., Yang, L., Hu, W., Jia, S., Ji, Q., Zhang, C., 2009. Anewsauropod PloS One 8, e68649. dinosaur from the Late Cretaceous Gaogou Formation of Nanyang, Henan Province. Whitlock, J.A., 2011. A phylogenetic analysis of Diplodocoidea (Saurischia: Sauropoda). Acta Geologica Sinica 83, 212–221. Zoological Journal of the Linnean Society 161, 872–915. Zhao, X., Cheng, Z., Xu, X., Makovicky, P.J., 2006. A new ceratopsian from the Upper Jurassic Wild, R., 1991. Janenschia n. g. robusta (E. Fraas 1908) pro Tornieria robusta (E. Fraas 1908) Houcheng Formation of Hebei, China. Acta Geologica Sinica 80, 467–473. (Reptilia, Saurischia, Sauropodomorpha). Stuttgarter Beiträge zur Naturkunde, Serie Zinsmeister, W.J., 1987. Cretaceous paleogeography of Antarctica. Palaeogeography, B: Geologie und Paläontologie 173, 1–4. Palaeoclimatology, Palaeoecology 59, 197–206.