A Reassessment of the Phylogenetic Position of Cretaceous Sauropod Dinosaurs from Queensland, Australia
Total Page:16
File Type:pdf, Size:1020Kb
Asociación Paleontológica Argentina. Publicación Especial 7 ISSN 0328-347X VII International Symposium on Mesozoic Terrestrial Ecosystems: 139-144.Buenos Aires, 30-6-2001 A ReasSessmenT of the phylogenetic position of CretaceoUS SaUROPOd dinosaURS from Queensland, Australia Ralph E. MOLNAR1 Abstract. The Cretaceous sauropod material from Queensland, Australia, has been regarded as pertaining to a persistently primitive sauropod lineage (e.g., Coombs and Molnar). The specimens derive from the Toolebuc and Allaru (Albian marine) and Winton (Cenomanian continental) Formations. Recent phyloge- netic analyses carried out by workers in Argentina, the USA and England permit a reassessment of this fragmentary material. As far as can be ascertained from the material, there is no indication from the char- acter states that more than a single taxon is represented. Character states diagnostic of the Titanosauriforrnes, the Titanosauria, the Somphospondyli and the Titanosauridae are present. Thus the Queensland material does not pertain to cetiosaurids but belongs to titanosaurs, extending their range in- to Australia Key words. Sauropods. Austrosaurus. Titanosaurs. Cretaceous. Paleozoogeography. IntroductioN (1998),has made it possible to reassess the phyloge- netic affinities of the Australian Cretaceous sauropod By the 1950's titanosaurs were widely recognized material and address the anomalous absence of ti- both as the latest sauropod group to diversify and as tanosaurs. This paper looks specifically at pre-eminently the sauropods of Gondwanaland. Austrosaurus mckillopi Longman (1933)and the mate- Both conclusions had been foreshadowed by von rial from the Winton Formation, much of which has Huene (1929).With titanosaurs prominent in South been referred to Austrosaurus sp. América, Africa and India - as well as a few occur- The first sauropod material recorded from rences in North America and Europe - it seemed Queensland (Queensland Museum {= QM} F311) anomalous that there was no evidence whatever that was collected in 1913 from Blackall, east-central they had inhabited Australia. Titanosaurs were Queensland, probably from the Winton Formation. chiefly recognized by their possession of procoelous Identified as a femur of Diprotodon, it was not cor- caudal vertebrae. Although Australian sauropod ma- rectly recognized as from a sauropod until 1980. In terial was sparse and fragmentary, the tail was ar- 1932,large but incomplete presacral vertebrae were guably the best known part of the anatomy, and pro- discovered in the Allaru Mudstone near Maxwelton, coelous caudals were clearly lacking. Salgado (1993) north-central Queensland. Studied by Longman departed from the usual interpretation of (1933),they were made the holotype of Austrosaurus Austrosaurus as a plesiomorphic sauropod and sug- mckillopi (QM F2316).Vertebrae (QM F2470)acquired gested titanosaurian affinities, specifically that it was in 1935 were identified as a cervical series of the the member of a clade also including the ti- pliosaur Kronosaurus. Only in 1980 was it realized tanosaurids. Study of new titanosaur material from they were sauropod caudals. Sauropod vertebrae, Argentina, supplemented by recently discovered correctly identified, were acquired in 1952 from specimens from Malawi, has shown that procoelous Chorregon, central Queensland (QM F10916) also caudals are neither necessary nor sufficient to recog- probably from the Winton Formation. Thus, frag- nize titanosaurs (Calvo and Bonaparte, 1991;Salgado mentary Cretaceous sauropod material had slowly and Coria, 1993;Jacobs et al., 1993).The recent inter- accumulated at the Queensland Museum from sev- est in the morphology of titanosaurs, and the phylo- eral sites in central Queensland. But, with the excep- genetic relationships of sauropods, exemplified by tion of Longman's work on Austrosaurus, it was not Giménez (1992), Upchurch (1995), Salgado et al. until the early 1970's - with the collection of several (1997), Wilson and Sereno (1998) and Upchurch specimens from the Winton Formation in the vicini- ty of Winton, central Queensland - that research in- terest was shown in this material, resulting in 'Museum of Northern Arizona, 3101 North Fort Valley Road, Flagstaff, Arizona, 86001, USA. Coombs and Molnar (1981).All of this material de- ©Asociación Paleontológica Argentina 0328-347Xi01$00.00+50 140 R.E. Molnar rives from the "míddle Cretaceous" Albían (marine ever posteriorly-positioned caudal neural arches Allaru or Toolebuc Formations) and Cenomanian have never been reported in sauropods, so the arches stages (continental Winton Formation). The few oth- must have been anteriorly-positioned. One character, er Cretaceous sauropod specimens from Queensland 19 (node 6), presence of prominent lateral bulge in will be treated elsewhere. femur below greater trochanter, is present in QM Longman (1933) tentatively attributed A. mckillopi F3390 (Coombs and Molnar, 1981). to the Cetiosauridae. Likewise Coombs and Molnar Four of the characters used for vertebrae by (1981) concluded that the material from near Winton Wilson and Sereno (1998) are present in both QM probably derived from a cetiosaurid, then considered F2316 and QM F6737. These are characters: 59, ante- a clade rather than agrade. Because of the fragmen- rior dorsal centra opisthocoelous; 68, presacral pleu- tary nature of these specimens, resulting in a large rocoels deep and invaginated; 92, posterior dorsal number of unknown character states, analyzing the centra opisthocoelous, with convexity present on an- sauropods as a whole with the inclusion of the char- terior face of centrum; and 102, presacral vertebrae acter states for the Queensland material is less likely composed of spongy bone. However characters 59 to illuminate the position of that material than to and 92 are uncertain for QM F6737 in that the incom- destabilize the results of the analysis. Thus the results pleteness of the vertebrae makes the identification of of the analyses of Salgado et al. (1997), Wilson and their position uncertain: they may be anterior rather Sereno (1998) and Upchurch (1998) (which are basi- than posterior dorsals. At least one of these charac- cally similar) are here accepted, and the likely phylo- ters is present, but if only one is present it is unclear genetic position of the Queensland material estab- which one. Two derived features are present in the fe- lished in relation to them. mur of QM F3390: lO, femoral shaft with elliptical Coombs and Molnar (1981) described five indi- cross-section, long axis oriented mediolaterally; and vidual specimens, but were uncertain whether all of 100, femur with proximal one-third of shaft deflected them represented the same taxon. Here nine speci- medially (which is taken to be equivalent to state 19 mens (including the previously studied five) are ex- of Salgado et al., 1997). Character 3, deltopectoral amined, and will be treated individually to avoid as- crest of humerus low, is present in QM F311. sumptions regarding their taxonomic uniformity. The following apomorphic vertebral states of Upchurch (1998) are present in the Queensland ma- Character states Present terial: B5 (C97), pleurocoels in dorsal centra; C27 (C132), centrum length divided by centrum height The vertebrae of these specimens are more abun- (in most craniad caudals) is approximately 0.5-0.6; dant and generally better preserved than the appen- H7 (C145) middle and distal chevrons are 'open' at dicular material, thus they are emphasized here. their proximal ends; K4 (C138), neural arches of mid- Seven vertebral apomorphies used by Salgado et al. dle caudals are situated on the cranial half of the cen- (1997) are present in the Queensland material (the trum; K9 (C205), osseous tissue structure of presacral nodes are those of Salgado et al., 1997, fig. 19): 1 (node vertebrae is composed of a small number of very 2) opisthocoelous cervical and anterior trunk verte- large 'cancellar' spaces separated by thin bony lamel- brae; 8 (node 4) presence of pleurocoels on presacral lae; MI (C96), pleurocoels in cranial dorsal centra vertebrae; 9 (node 5) opisthocoelous posterior trunk have tapering acute, caudal margins; P1 (C137), cen- and sacral centra; 15 (node 6) neural arches posi- tra of middle caudals display a dorsoventrally com- tioned anteriorly in mid- and posterior caudal centra; pressed transverse cross-section, and; Q5 (C136), 20 (node 8) eye-shaped pleurocoels in trunk verte- ventral surfaces of cranial caudal centra are mildly or brae, and 22 (node 8) presence of centro-parapophy- deeply excavated, with the excavation bounded by a se al lamina in posterior trunk vertebrae. Character 34 ventrolateral ridge on each side (convergent in (node 14) depressed mid- and posterior caudal centra Barosaurus and Diplodocus). One state, 12 (node 7), of with dorsoventrally convex lateral faces is present in Upchurch (1995) is present in QM F3390 and QM the mid-caudal of QM F2470, but absent in the other F7291, femoral distal condyles extend prominently caudals and other specimens. Characters 1 and 9 are anteriorly as well as posteriorly. assumed to be present because opisthocoelous ante- rior and posterior trunk vertebrae are clearly present. Character assessment by specimen Cervicals are possibly, but not certainly, known in the Winton material and Austrosaurus mckillopi, and QM F311: This incomplete humerus presents no sacral centra have not been recovered. Character 15 is useful characters beyond one - deltopectoral crest also assumed to be present (in