Society of Vertebrate Paleontology Memoir 5 Journal of Vertebrate Paleontology Volume 18. Supplement to Number 2 0 1998 by the Society of Vertebrate Paleontology DINOSAUR SOC l ETYa EARLY EVOLUTION AND HIGHER-LEVEL PHYLOGENY OF SAUROPOD DINOSAURS JEFFREY A. WILSON and PAUL C. SERENO Department of Organismal Biology and Anatomy, University of Chicago, 1027 East 57th Street, Chicago, Illinois 60637 ABSTRACT-Although sauropods played a major role in terrestrial ecosystems during much of the Mesozoic Era, little effort has been directed toward diagnosing Sauropoda and establishing higher-level interrelationships among sauropods. As a consequence, the origin and evolution of major skeletal adaptations in sauropods has remained largely speculative. The cladistic analysis presented here focuses on higher-level relationships among sauropods. Based on 109 characters (32 cranial, 24 axial, 53 appendicular) for 10 sauropod taxa, the most parsimonious arrangement places four genera (Vulcanorlon, Shunosaurus, Barapasaurus, and Omeisaurus) as a sequence of sister-taxa to a group of advanced sau- ropods. defined here as Neosauropoda. Neosauropoda, in turn, is composed of the sister-clades Diplodocoidea and Macronaria; the latter is a new taxon that includes Haplocanrhosaunrs, Catnarasaurus, and Titanosauriformes. Titano- sauriformes includes Brachiosauridae and Somphospondyli. a new taxon uniting Euhelopus and Titanosauria. Among macronarians, the position of Haplocunthosaurus is the least stable as a result of the absence of cranial remains. The basic structure of the phylogeny is resilient to various tests and establishes the evolutionary sequence of many functionally significant sauropod adaptations, such as the digitigrade posture of the manus in neosauropods. Other characteristic sauropod adaptations, such as narrow tooth crowns, increases in length and number of cervical vertebrae, and bifid neural spines, are shown to have evolved more than once. As these results underscore, the higher-level phylogeny of sauropods must be based on a broad sampling of character data. The fossil record of sauropods, although relatively limited during the early phase of the radiation (Late Triassic through Early Jurassic). nonetheless indicates that all major clades were established prior to the Late Jurassic. when substantial faunal interchange among major continental regions was still possible. The functional, temporal. and bio- geographic implications of the higher-level phylogeny of sauropods are explored. RESUMEN-Los saurdpodos habrian sido un componente importante de las comunidades terrestes durante gran parte del Mesosoico; se ha dirigido poco esfuerzo al estudio de la diagnosis de Sauropoda o a establecer las relaciones entre los distintos subgrupos. Por esto. el origen y la evoluci6n de sus adaptaciones esqueletales han sido especulativos. El analisis cladistico aqui presentado enfoca en las relaciones de alto-nivel de 10s saur6podos, sobre la base de 109 caracteres (31, craneanos, 24 axiales y 53 apendiculares) de diez grupos terminales. Se establece a 10s cuatro generos (Vulcarzodotz, Shunosaunts, Barapasaurus y Omeisa~trus) como una sucesi6n de grupos-hermanos a un grupo de sau- r6podos avanzados, aqui definido como Neosauropoda. Neosauropoda es compuesto por los clados-hermanos Diplo- docoidea y Macronaria; el ultimo es un nuevo grupo que incluye Haplocanthosa~tr~ts, Cumarasaurus y Titanosauri- formcs. Titanosauriformes incluye Brachiosauridae y Somphospondyli, un nuevo grupo unificando Euhelopus y Titano- sauria. La ausencia de restos crineanos hace menos segura la posici6n de Haplocanthosaurus entre macronarios. La estructura de la filogenia es confirmada por varios analisis, y pone de manifesto la evolution de adaptaciones impo~rantes, como por ejemplo la mano digitigrada en 10s neosauropodos. Otras adaptaciones tradicionalemente men- cionadas-tales como las coronas angostas. aumentos del largo y numero de vCrtebras cervicales y las espinas neurales dobles en 10s vertebras presacras-fueron adquiridas mas de una vez. Esto subraya la importancia de una prueha con amplio muestreo de caracteres para analisis filogeneticos de saur6podos. El registro fosil de los saur6podos, aunque relativamente limitado en 10s primeros tiempos de su radiation (Triasico tardio hasta Jur6sico medio), indica que todos 10s clados mayores habrian estado establecidos antes del Jurasico tardio, cuando aun era posible el intercambio faunistico entre las mayores regiones continentales. Las implicaciones funcion- ales, temporales y biogeograficos de la filogenia de alto-nivel de 10s saur6podos son analizadas. INTRODUCTION define and diagnose the ingroups supported by this phylogenetic analysis. Finally, we assess the reliability of these phylogenetic In 1841 Richard Owen described the scattered remains of the results and their evolutionary implications. first sauropod known to science as Cetiosaurus, or "whale liz- ard." Since that time, abundant sauropod remains have been HISTORICAL BACKGROUND discovered on every continent except Antarctica. Sauropods constitute a major proportion of the large herbivores in most Traditional Classification continental faunas of Jurassic and Cretaceous age. Despite their Traditional (pre-cladistic) classifications divided sauropods rich and long fossil record, higher-level relationships among into families and subfamilies (Fig. 1). When Marsh (1878) sauropods have received little attention. coined the suborder Sauropoda, it included only one family (At- We provide a phylogenetic framework for more focused work lantosauridae) and was based principally on skeletal material of within Sauropoda by analyzing a broad range of character data Camarasaurus (=Morosaurus; Riggs, 1901; Mook, 1913). Lat- across the most completely known taxa. First, we summarize er Marsh (1 882) erected a new family for Camarasaurus (Mo- historical and more recent efforts to resolve natural (monophy- rosauridae), leaving other sauropods in Atlantosauridae (prin- letic) groups among sauropods. Second, we define and diagnose cipally the diplodocids Apatosaurus and Diplodocus; Fig. 1A). the sauropod taxa that we have selected for analysis. Third, we Although Marsh's (1882) familial diagnoses differed only in an