A new basal sauropodomorph dinosaur from the Early Jurassic of South Africa

This article was downloaded by: [Bonnan, Matthew F.] On: 11 May 2011 Access details: Access Details: [subscription number 937434023] Publisher Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 3741 Mortimer Street, London W1T 3JH, UK Journal of Vertebrate Paleontology Publication details, including instructions for authors and subscription information: http://www.informaworld.com/smpp/title~content=t917000010 A new basal sauropodomorph dinosaur from the Early Jurassic of South Africa Adam M. Yatesa; Matthew F. Bonnanb; Johann Nevelingc a Bernard Price Institute for Palaeontological Research, University of the Witwatersrand, South Africa b Department of Biological Sciences, Western Illinois University, Macomb, Illinois, U.S.A. c Council for Geoscience, Pretoria, South Africa Online publication date: 09 May 2011 To cite this Article Yates, Adam M. , Bonnan, Matthew F. and Neveling, Johann(2011) 'A new basal sauropodomorph dinosaur from the Early Jurassic of South Africa', Journal of Vertebrate Paleontology, 31: 3, 610 — 625 To link to this Article: DOI: 10.1080/02724634.2011.560626 URL: http://dx.doi.org/10.1080/02724634.2011.560626 PLEASE SCROLL DOWN FOR ARTICLE Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf This article may be used for research, teaching and private study purposes. Any substantial or systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material. Journal of Vertebrate Paleontology 31(3):610–625, May 2011 © 2011 by the Society of Vertebrate Paleontology ARTICLE A NEW BASAL SAUROPODOMORPH DINOSAUR FROM THE EARLY JURASSIC OF SOUTH AFRICA ADAM M. YATES,*,1 MATTHEW F. BONNAN,2 and JOHANN NEVELING3 Bernard Price Institute for Palaeontological Research, University of the Witwatersrand, Johannesburg, Private Bag 3, WITS, 2050, South Africa, yatesam@gmail.com; 2 Department of Biological Sciences, Western Illinois University, Macomb, Illinois 61455, U.S.A.; 3 Council for Geoscience, Pretoria 0001, South Africa 1 Downloaded By: [Bonnan, Matthew F.] At: 17:42 11 May 2011 ABSTRACT—A new basal sauropodomorph dinosaur, Arcusaurus pereirabdalorum, sp. nov., is named and described on the basis of a partial, disarticulated but associated skull and dispersed cranial and postcranial elements from at least two individuals. Arcusaurus is part of a distinctive local fauna from the upper Elliot Formation (Lower Jurassic) in the Senekal District, Free State, South Africa. It can be diagnosed by various details of the premaxilla, nasal, and dentary in the skull and the shape of the distal caudal vertebrae. The taxon displays an unusual mix of characteristics. It lacks several synapomorphies of Plateosauria (Plateosaurus + Massospondylus and all descendants of their most recent common ancestor) but does display other derived characteristics that are otherwise known only from less inclusive clades within Plateosauria. In a cladistic analysis a position outside Plateosauria as the sister group of the clade of Efraasia + more-derived sauropodomorphs is supported; however, this position was not found to be a significantly better explanation of the data as a relatively derived position within Plateosauria. If the basal position for Arcusaurus is accepted, then a divergence from other sauropodomorphs in the middle Norian and a ghost lineage up to 35 Ma is implied. No other non-plateosaurian sauropodomorphs are known from the Jurassic, making Arcusaurus a potentially relictual taxon in the Early Jurassic. INTRODUCTION Lower Jurassic dinosaurs from South Africa have been found in the upper member of the Elliot Formation and the overlying Clarens Formation of the Karoo Supergroup. Together these units form the biostratigraphic zone known as the Massospondylus Range Zone (Kitching and Raath, 1984). Numerous sauropodomorph dinosaur taxa have been named from the Massospondylus RZ. However, almost all of those named prior to 2009 are now regarded as nomina dubia or junior synonyms of Massospondylus carinatus (Yates and Barrett, 2011). Nevertheless M. carinatus was clearly not the sole sauropodomorph in the Early Jurassic of southern Africa. Yates et al. (2004) described an isolated caudal vertebra from the upper Elliot Formation that belongs to a true sauropod, with a body size several times greater than M. carinatus. Barrett (2004) also briefly discussed the remains of a small sauropodomorph from the upper Elliot Formation, which has since been named M. kaalae (Barrett, 2009). Gryponyx africanus, a taxon that has been sunk into synonymy with Massospondylus carinatus, was resurrected by Vasconcelos and Yates (2004), although doubts about its distinctiveness still linger (Yates, unpubl. observ.). Lastly, Yates et al. (2007) briefly reported the discovery of a new assemblage of upper Elliot Formation dinosaurs that includes three new sauropodomorph taxa. One of these taxa, Aardonyx celestae, has since been named (Yates et al., 2010). The three taxa are part of an unusual and highly localized assemblage found in a distinctive package of stacked channel sands on the farm Spion Kop 932, in the Senekal District, northern Free State (Fig. 1). This sedimentary package appears to represent a discrete microenvironment on the upper Elliot floodplain with a distinctive local fauna. Here we describe and name the smallest of these new sauropodomorph taxa. *Corresponding author Institutional Abbreviations—BP, Bernard Price Institute for Palaeontological Research, University of the Witwatersrand, Johannesburg, South Africa; BRSUG, Department of Earth Sciences, University of Bristol, Bristol, U.K.; NM, National Museum, Bloemfontein, South Africa; SAM, South African Museum, Izikio Museums, Capetown, South Africa; YPM, Peabody Museum of Natural History, Yale University, New Haven, Connecticut, U.S.A. Anatomical Abbreviations—acet, acetabulum; al, alveolus; aof, antorbital fenestra; c, coronoid; ca, caudal vertebra; cf, cubital fossa; clpn, caudolateral process of the nasal; clpp, caudolateral process of the premaxilla; cmpp, caudomedial process of the premaxilla; cp, capitate process; d, dentary; dip, dorsal intercondylar process; dp, semilunate dorsal process of the postorbital; dt, dentary tooth; ect, ectepicondyle; emf, external mandibular fenestra; en, external naris; ent, entepicondyle; ff, facet for articulation with the frontal; fj, facet for articulation with the jugal; fmx, facet for articulation with the rostromedial process of the maxilla; fp, facet for articulation with the parietal; fpro, facet for articulation with the prootic; fsq, facet for articulation with squamosal; fsr, facet for articulation with sacral rib; idp, interdental plate; inf, internarial fenestra; ip, internarial process of the premaxilla; ipf, interpremaxillary fenestra; isp, ischiadic peduncle; itf, infratemporal fenestra; ldc, lateral distal condyle; lp, ligament pit; ls, lingual sulcus; mc, meckelian canal; mcv, notch for passage of the mid-cerebral vein; mdc, medial distal condyle; ms, medial shelf of the premaxilla; mxt, maxillary tooth; n, nasal; nc, neural canal; nvf, neurovascular foramen; nvg, neurovascular groove; pal, palatine; po, postorbital; poz, postzygapophysis; prz, prezygapophysis; ps, palatal shelf of the premaxilla; r, autapomorphic ridge; rc, radial condyle; sna, sutural surface for neural arch; rdf; rostral dentary foramen; rmp, rostromedial process of the nasal; rt, replacement tooth; rvp, rostroventral process of the nasal; snf, subnarial foramen; ssr, sutural surface for sacral rib; stf, 610 YATES ET AL.—NEW BASAL SAUROPODOMORPH 611 Downloaded By: [Bonnan, Matthew F.] At: 17:42 11 May 2011 FIGURE 1. A, Map of South Africa showing the location of the Free State and the farm Spion Kop 932. B, Map of Spion Kop 932 showing the location of the sauropod quarry (SQ). C, Quarry map for the sauropod quarry showing the bones of an undescribed sauropod (grey), the holotype of Arcusaurus pereirabdalorum (black), and the area over which the referred specimens of Arcusaurus were recovered (dashed line). supratemporal fenestra; uc, ulnar condyle; III, notch for passage of the oculomotor nerve; IV, notch for the passage of the trochlear nerve; V1 , sulcus for the passage of the ophthalmic branch of the trigeminal nerve; ?, fragment of an unidentified bone. SYSTEMATIC PALEONTOLOGY SAURISCHIA Seeley, 1887 SAUROPODOMORPHA von Huene, 1932 ARCUSAURUS PEREIRABDALORUM, gen. et sp. nov. Holotype—A disarticulated right side of the skull, BP/1/6235, including postorbital, nasal, coronoid (supradentary), dentary, palatine, and loose maxillary teeth (Figs. 2, 3). The specimen also includes a distal caudal vertebra but the association with the same individual as the skull is not certain. Referred Specimens—Several small sauropodomorph bones were found within a radius of 50 cm of the holotype skull (Fig. 1). These specimens belong to individuals that are comparable in size to the holotype and are much smaller than the basal sauropod that is found in the same quarry. Referral of these small bones to Arcusaurus is based on their proximity to the holotype within the quarry and their consistent size. In the case of the dentary, nasal, and isolated teeth, the referral to Arcusaurus is further sup- ported by anatomical characteristics that are consistent with the holotype and the lack of any characteristics to distinguish them from the holotype. The presence of a second right nasal does indicate the presence of more than one individual, consequently each of the isolated elements have been designated their own accession numbers: BP/1/6842, 6843 isolated teeth; 6853 incomplete left dentary; 6844 right nasal; 6845 left premaxilla; 6846 right non-terminal pedal phalanx, probably IV-1, 6847 sectioned nonterminal phalanx; 6848, ungual of right pedal digit I; 6849 ungual of left pedal digit III; 6850 distal end of a right humerus; 6851 sacral centrum; 6925 right laterosphenoid; 6926 a fragment from the ischiadic peduncle of a left ilium. Most of the isolated postcranial remains can be differentiated from Massospondylus carinatus (see below), further supporting their referral to the new taxon. Horizon and Locality—Sauropod Quarry, Spion Kop 932, (S28◦ 28.002 E027◦ 49.523 ), Senekal District, Free State, South Africa (Fig. 1). Upper Elliot Formation, Early Jurassic, possibly Pliensbachian (Yates et al., 2004). Etymology—From arcus (Latin for ‘rainbow’) and sauros (Greek for ‘lizard’), gender is masculine; pereirabdalorum for Lucille Pereira and Fernando Abdala, who discovered most of the bones of this taxon. Genus name honors the people of South Africa, the ‘rainbow nation.’ Diagnosis—Basal sauropodomorph with following autapomorphies: presence of horizontal medial shelf projecting from dorsal margin of the premaxillary body and offset caudally by a 612 JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 31, NO. 3, 2011 Downloaded By: [Bonnan, Matthew F.] At: 17:42 11 May 2011 FIGURE 2. Holotype (BP/1/6235) of Arcusaurus pereirabdalorum, gen. et sp. nov., partial right side of the skull in lateral view with distal caudal vertebra. A, photograph. Inset shows caudal process of the postorbital during preparation, before accidental loss of its dorsal prong. B, interpretive drawing. Scale bar equals 50 mm. notch; base of caudomedial process expanded into a horizontal palatal shelf that protrudes medially to contact its antimere; postorbital with an articulation surface for the parietal borne on dorsally projecting, semilunate process; fine ridge on nasal extending from medial edge of prefrontal embayment to a point near center of bone; tongue-shaped rostroventral process of nasal; dentary with a single enlarged neurovascular foramen opening near the rostral tip of the lateral surface; distal caudal vertebra with dorsoventrally compressed centrum (cranial centrum face 1.43 times wider than high). Differential Comparisons—In addition to its autapomorphies, Arcusaurus can be distinguished from the contemporary Massospondylus carinatus by the following traits: base of the rostrodorsal process of the postorbital with barely any medial curvature; a pointed rostrodorsally oriented tip to the dentary; lack of a dorsolateral ridge at the caudal end of the dentary; a relatively deep dentary, with a maximum depth that is 23.5% of its length; maxillary teeth with denticles that extend well below the apical half of the crown; poorly developed cubital fossa of the humerus; absence of a distal caudal heel on the ischiadic peduncle of the ilium; proximal breadth of pedal phalanx IV-1 is equal to its length; length of the ungual of pedal digit III is at least 75% of that of pedal digit I. It differs from M. kaalae by the pointed rostrodorsally oriented tip of the dentary and the absence of a dorsolateral ridge at the caudal end of the dentary. It also differs from the contemporary Aardonyx by lacking lateral plates on the dentary and premaxilla; possessing only four premaxillary teeth (vs. five); possessing an acutely pointed rostral tip of the dentary bearing just one enlarged neurovascular foramen; lacking a ventrally curved rostral end of the ventral margin of the dentary; lacking a dense band of fine pits along the lateral alveolar margins of tooth-bearing bones; lacking a strong medial curvature of the rostrodorsal ramus of the postorbital; possessing well-developed denticulations of the marginal carinae of the YATES ET AL.—NEW BASAL SAUROPODOMORPH 613 Downloaded By: [Bonnan, Matthew F.] At: 17:42 11 May 2011 FIGURE 3. Holotype (BP/1/6235) of Arcusaurus pereirabdalorum, gen. et sp. nov., partial right side of the skull in medial view with distal caudal vertebra. A, photograph and B, interpretive drawing. Scale bar equals 50 mm. teeth that extend well below the apical half of the crowns; possessing smooth unornamented tooth enamel. Arcusaurus most closely resembles Thecodontosaurus and Pantydraco from the Late Triassic of Great Britain. It can distinguished from Thecodontosaurus by the lack of a dorsolateral ridge at the caudal end of the dentary, poorly developed cubital fossa, and the stout proportions of its proximal pedal phalanges. It differs from Pantydraco by its larger premaxilla and consequently deeper snout; greater number of dentary teeth (18 vs. 12); pointed rostrodorsally directed tip of the dentary; and stout proximal pedal phalanges. Recently Knoll (2010) has erected Ignavusaurus rachelis a supposed non-plateosaurian sauropodomorph from the upper Elliot Formation. We consider it to be a likely synonym of either Massospondylus carinatus or M. kaalae (see below). Regardless of its status as a distinct taxon, Ignavusaurus can be distinguished from Arcusaurus by its everted orbital margin of the postorbital, transversely expanded ventral process of the postorbital and coarser, apically restricted denticles of the dentary teeth. DESCRIPTION Skull The incomplete holotype skull and referred skull bones would have come from skulls measuring no more than 150 mm in length (Table 1; Fig. 4). However, they probably represent juveniles because the neurocranium had open sutures, judging from the isolated laterosphenoid. The sacrum also has open sutures and is consistent with the hypothesis that the preserved Arcusaurus individuals were juveniles. Premaxilla—The premaxilla (BP/1/6845) is missing most of the dorsal process and the rostral margin due to damage incurred when the block containing it was split (Fig. 5). The premaxillary body is nearly as deep in the dorsoventral dimension as it is rostrocaudally long. This, together with the steeply angled base of the dorsal process, suggests that the snout was relatively deep, as it is in plateosaurian sauropodomorphs (e.g., Plateosaurus, Huene, 1926:pl. 1, figs. 1, 2; Massospondylus, Sues et al., 2004:figs. 3, 5; Melanorosaurus, Yates, 2007:figs. 1, 2), rather 614 JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 31, NO. 3, 2011 that the premaxillary portion of the snout was distinctly ‘pinched’ relative to the maxillary portion. Nasal—The nasal is a distinctive tetraradiate plate-like bone, with caudomedial, caudolateral, rostromedial, and rostroventral processes. Between the processes there are four embayments in the margin of the bone: one caudal, rostral, medial, and lateral embayment each. Starting at the rostral end there is a deep U-shaped embayment between the rostromedial and rostroventral processes that forms the caudodorsal rim of the external naris. Whereas the rostroventral process of most basal sauropodomorphs forms a triangular flange (e.g., Plateosaurus, Galton, 1984:pl. 5, fig. 9; Massospondylus, Sues et al., 2004:fig. 5), the rostroventral process of Arcusaurus is tongue-shaped (Fig. 2A). Furthermore, the basal width of the process is less than that of the rostromedial process, the opposite condition to that seen in basal plateosaurian sauropodomorphs. The rostromedial process, which is the process that would have contacted the dorsal process of the premaxilla to complete the internarial bar, is short and triangular with a pointed rostral tip. The medial margin is not straight and instead includes a centrally located semilunate embayment. When the two nasals were in articulation, these embayments would have produced an elliptical median fenestra on the dorsal surface of the snout (Figs. 4A, 6B), as has been found and restored in the skull of Melanorosaurus (Yates, 2007:figs. 3,7). This fenestra cannot be the result of damage to the thin medial edge of the nasals because its margin forms a wellpreserved natural edge in the isolated referred nasal (BP/1/6844; Fig. 6A). Caudally there is a broad, flat medial process. The termination of this process is damaged in both preserved nasals; however, there does not appear to be much missing bone in the referred nasal, indicating that the process had a blunt termination that produced a rather straight fronto-nasal suture. Lateral to the caudomedial process is a caudal U-shaped embayment that would have received the rostral end of the prefrontal when the skull was complete. Projecting caudolaterally and slightly ventrally from the lateral side of the prefrontal embayment is the subtriangular caudolateral process. Its tip is noticeably pitted and probably represents the rostral extention of a dorsolateral rugosity of the lacrimal. Similar rugosities can be observed in many basal sauropodomorphs including Massospondylus (Sues et al., 2004:fig. 3), Mussaurus (Pol and Powell, 2007a:fig. 5), and Plateosaurus (pers. observ., GPIT skelett 1). A tiny, curved, caudally directed, thorn-like process arises from the lateral margin of the main caudolateral process. This would have overlapped the TABLE 1. Dimensions (mm) of the skull bones of Arcusaurus pereirabdalorum, gen. et sp. nov. Maximum length Premaxilla (BP/1/6845) Nasal (BP/1/6235) Postorbital (BP/1/6235) Laterosphenoid (BP/1/6925) Dentary (BP/1/6235) 23 38 39 17 71 ∗ Maximum width 10 21 8 12 — ∗ Maximum depth 20∗ , 14 (corpus) — 38 16 16 Asterisk indicates that the measurement is of an incomplete bone as it is preserved. than the pointed, Eoraptor-like snout that has been restored for non-plateosaurians such as Pantydraco (Yates, 2003:fig. 2). The base of the dorsal process is placed at the rostral end of the bone, indicating that the external nares were terminally positioned. BP/1/6845 has four preserved alveoli. Judging from the rostral convergence of the line of dental foramina with the lateral alveolar margin, there are no missing alveoli (Fig. 5C, D). The interdental plates differ from those of most other basal saurischians (e.g., Plateosaurus, Galton, 1985:fig. 6; Dracovenator, Yates, 2006:figs. 3, 4; Massospondylus: SAM PK K398) in that they are higher than they are wide. The lateral alveolar margin only protrudes below the ends of the interdental plates for a distance that is less than the width of an alveolus, thus lateral plates cannot be said to be present. Two horizontal shelves protrude medially from the caudal half of the premaxillary body and between them define a U-shaped slot to accept the rostromedial process of the maxilla. The subrectangular ventral shelf extends from the base of caudomedial process and would have made contact with its antimere, formed a broad flat premaxillary palate. The rostral margin of the premaxillary palate defines a large elliptical median foramen. The dorsal medial shelf is less extensive and would not have contacted its antimere. A distinct notch separates the dorsal medial shelf from the caudolateral process in dorsal view. The caudolateral process extends caudally for half the length of the premaxillary body. It is a slender horizontal prong in lateral view but is noticeably curved laterally in dorsal view. The much shorter caudomedial process is also directed laterally, indicating Downloaded By: [Bonnan, Matthew F.] At: 17:42 11 May 2011 FIGURE 4. Reconstruction of the skull of Arcusaurus pereirabdalorum, gen. et sp. nov., in A, dorsal and B, lateral views based on BP/1/6235 and BP/1/6845. Scale bar equals 50 mm. YATES ET AL.—NEW BASAL SAUROPODOMORPH 615 Downloaded By: [Bonnan, Matthew F.] At: 17:42 11 May 2011 FIGURE 5. Left premaxilla (BP/1/6845) of Arcusaurus pereirabdalorum, gen. et sp. nov., in A, B, lateral, C, D, medial, E, F, ventral, and G, H, dorsal views. I, reconstruction of premaxillary pair in ventral view. Grey areas represent regions of matrix, hatched areas represent broken bone surfaces in B, D, F, H. Scale bar equals 20 mm. lacrimal and is probably homologous with the caudolateral spur of other eusaurischians such as Coelophysis (Downs, 2000:fig. 2b) and Melanorosaurus (Yates, 2007:fig. 3). The lateral margin of the nasal is gently concave between the caudolateral process and the rostroventral process. Currently it is not possible to determine if this embayment formed part of the dorsal margin of the antorbital fenestra or whether it was excluded by the maxilla and lacrimal. A low but sharp, scarp-like ridge extends rostrally from the medial side of the prefrontal embayment to a point in the center of the bone. As in other basal sauropodomorphs there is a cluster of three neurovascular foramina on the dorsal surface of the bone, just caudal to the caudodorsal corner of the external naris. Postorbital—The dorsoventrally flattened rostral ramus of the triradiate postorbital is relatively elongate, being almost equal to the ventral ramus in length (Fig. 3). In dorsal view, the rostral ramus is straight, lacking the medial curvature that is usually present in sauropodomorph postorbitals. Two large articulation facets on the dorsal and ventral surfaces of the rostral ramus indicate that the postorbital inserted into a horizontal slot in the frontal (Fig. 2). About midway along the length of the rostral ramus, the medial edge expands into a dorsally directed, semicircular process. The medial surface of this process is concave and striated, indicating that it served as the articulation facet for the parietal. A dorsally expanded parietal-postorbital articulation is apparently unique amongst dinosaurs. Two short, deep sulci excavate the ventral surface of the rostral ramus, directly below the semicircular flange. These may represent additional articulation scars for the parietal but more likely represent the articulation site of the laterosphenoid. The dorsolateral rim of the rostral ramus is not raised, indicating that there was no distinct supratemporal fossa impressed upon the dorsal surface of the postorbital. The ventral ramus curves rostrally along its length, indicating a relatively large, circular orbit similar to that of Eoraptor (Sereno et al., 1993:fig. 1). The degree of rostral curvature of this process is quite variable amongst basal sauropodomorphs, even within the same species (e.g., Massospondylus carinatus, BP/1/5241, BP/1/4934, Sues et al., 2004:figs. 5, 7), so is of little systematic value. The transverse section of the ventral ramus is triangular and mediolaterally compressed, with an articular facet for the dorsal process of the jugal developed along the ventral half of the caudal margin. The caudal ramus of the postorbital is strongly mediolaterally compressed and arises from the dorsal end of the postorbital. It is smoothly flush with the lateral surface of the rest of the bone. The dorsal margin of the postorbital is only weakly concave in lateral view, with a gentle embayment developed between the rostral and caudal rami. Unlike most other basal sauropodomorphs, the caudal ramus of Arcusaurus is not distally tapered, instead it maintains its depth along its length and has a bifurcate terminus (Fig. 2). The dorsal prong of the terminal bifurcation was lost soon after preparation but photographs taken during preparation record its existence (Fig. 2). Palatine—The right palatine is only partially exposed (Fig. 2). Three of the usual four processes are visible. The caudolateral process, which would articulate against the medial side of the maxilla and form part of the rostrolateral margin of the suborbital fenestra, is apparently buried and therefore must have been short and poorly developed. A long-slender caudomedial process protrudes from the matrix on the medial side of the holotype block (Fig. 3). The lateral side of the block exposes the medial vomeropterygoid process, which appears as a short, rectangular tab. However, this process is most likely damaged and was probably larger in life. A slender craniolateral process extends cranially from the base vomeropterygoid process and curves ventrolaterally towards its tip. An articulation surface for the medial side of the maxilla extends along the lateral side of this process. The dorsal fossa for the dorsal pterygoideus muscle and ventral surface of the palatine are obscured by matrix. Laterosphenoid—The laterosphenoid (Fig. 7) is a relatively short, deep bone in comparison to the laterosphenoids of Massospondylus and Plateosaurus where the rostrocaudal length of 616 JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 31, NO. 3, 2011 shaped process formed part of the rostral rim of the trigeminal foramen. Unlike Plateosaurus (Galton, 1985:pl. 6, fig. 12), there is no pit for the reception of the dorsal tip of the epipterygoid. A broad, rounded notch for the passage of the mid-cerebral vein excavates the caudal margin of the bone, dorsal to the tongueshaped process (Fig. 7A–C). The gap between this notch and sulcus of the ophthalmic nerve indicates that, like Massospondylus, Plateosaurus, and Melanorosaurus (Galton, 1985:fig. 1b; Gow, 1990:fig. 3; Yates, 2007:fig. 12), the foramen for the mid-cerebral vein was separated from the trigeminal foramen. The rostromedial edge of the laterosphenoid slopes caudoventrally and is excavated by two distinct notches (Fig. 7C). The more ventral of these forms part of the border of the oculomotor nerve (III). It is a large semicircular notch that is as deep as the notch for the trochlear nerve (IV), unlike the weak embayment seen in Massospondylus (BP/1/5276; Gow, 1990:fig. 3a), or the shallow notch seen in Melanorosaurus (NM QR 3314; Yates, 2007:fig. 12). A tab-like rostromedial projection separates the oculomotor notch from the trochlear notch, which is narrower than the oculomotor notch and V-shaped rather than broadly rounded. A small rugose process extends rostrally from the rostrodorsal corner of the laterosphenoid and forms part of the dorsal margin of the trochlear notch (Fig. 7D). An almost identical process, which underlies the frontal, is present in Massospondylus (BP/1/5231; Gow et al., 1990:fig. 5) Dentary—The dentary is complete with a maximum dorsoventral depth that is 23.5% of the total length (Fig. 2). It bears 18 alveoli along its dorsal margin, the first of which is inset from the rostral tip of the dentary by the distance equivalent to the diameter of one alveolus. In lateral and medial views, the dorsal margin of the dentary extends in a horizontal line to the rostral tip, unlike most other basal sauropodomorphs where the dorsal margin is ventrally curved at its rostral end. The symphyseal axis is strongly procumbent so that the rostral tip of the dentary is acutely pointed in lateral and medial views (Fig. 3). In ventral view, the rostroventral corner of the symphysis is inflected medially, indicating that the mandibular rami met to form a broader triangular snout tip than in other basal sauropodomorphs. The labial surface of the rostral tip is sculpted with numerous tiny foramina and one especially large neurovascular foramen, here called the rostral dentary foramen (Fig. 2). In other basal sauropodomorphs, there is a cluster of enlarged neurovascular foramina at the rostral end of the dentary (e.g., Massospondylus, BP/1/4934). An open canal appears to extend caudally from the rostral dentary foramen but because it connects to an area where the lateral surface of the bone has flaked away, it is possible that this canal is just an enlarged crack. A series of six neurovascular foramina in a horizontal row extend caudally from the rostral dentary foramen, with the caudal-most foramen opening between the level of the 13th and 14th dentary teeth. The lateral surface of the dentary is gently convex dorsoventrally and lacks a caudal-lateral ridge defining a dorsally facing shelf at the caudal end of the bone. Consequently the caudal end of the dentary tooth row is not medially inset relative to the middle and rostral parts of the tooth row. The caudal margin of the dentary is remarkably straight and lacks the strong embayment that defines the rostral margin of the external mandibular fenestra in other basal sauropodomorphs (e.g., Plateosaurus, SMNS 13200; Massospondylus, BP/1/4376). A short, spike-like process, located at about two-thirds the height of the dentary from the ventral margin, projects horizontally from the caudal margin. This process marks the rostrodorsal margin of the external mandibular fenestra. Medially the dentary bears a deep Meckelian sulcus that opens caudally into a broad triangular fossa that would have been covered by the splenial in life. A covering of matrix makes it difficult to determine if the sulcus extended rostrally to the symphyseal surface. Ventral to the Meckelian groove the dentary forms a thin, sharp-edged medially projecting ridge, whereas the area of the dentary dorsal to the groove is much Downloaded By: [Bonnan, Matthew F.] At: 17:42 11 May 2011 FIGURE 6. Nasals of Arcusaurus pereirabdalorum, gen. et sp. nov. A, isolated right nasal (BP/1/6844) in dorsal view. B, reconstruction of nasal pair in dorsal view. Scale bar equals 20 mm. the bone greatly exceeds its dorsoventeral height (e.g., Galton, 1985:pl. 6, fig. 12; Gow, 1990:fig 5). Like other nonsauropod sauropodomorphs (e.g., Massospondylus, BP/1/5276, and Melanorosaurus, NM QR3314), there is a large laterally projecting, capitate process that articulates dorsally with the rostral end of the ventral surface of the parietal (Fig. 7A, D). The process is compressed with a narrowly elliptical cross-section (Fig. 7B). The lateral process is canted rostrodorsally and its long base divides the remainder of the bone into two regions: a laterally facing caudal region, which forms part of the wall of the adductor chamber, and a rostroventrally facing rostral region, which forms part of the rear wall of the orbital cavity. In theropods and sauropods a tall antotic crest forms this division (Sampson and Witmer, 2007) but in Arcusaurus, as in other basal sauropodomorphs (e.g., Plateosaurus, Galton, 1985:pl. 6, figs. 11, 12; Massospondylus, Gow, 1990:fig. 3), the antotic crest is subsumed by the long base of the capitate process. The exits for several cranial nerves leave distinct notches in the margins of the bone. A well-developed tongue-shaped process extends from the caudoventral corner of the bone, between these two regions (Fig. 7B, C). A sharp-edged sulcus, marking the path of the ophthalmic branch of the trigeminal nerve (V1 ), extends across the lateral surface of the tongue-shaped process in a rostrodorsal direction (Fig. 7B, C). This indicates that caudal margin of the tongue- YATES ET AL.—NEW BASAL SAUROPODOMORPH 617 FIGURE 7. Right laterosphenoid (BP/1/6925) of Arcusaurus pereirabdalorum, gen. et sp. nov., in A, caudal, B, lateral, C, rostrolateral, and D, dorsal views. Scale bar equals 10 mm. broader (Fig. 3). Most of the interdental plates are missing but one is preserved between the seventh and eighth alveoli (Fig. 3). Like the interdental plates of the premaxilla, this plate is tall and narrow. Coronoid—The coronoid (supradentary) is a narrow, thin, strap-like bone that is less than half the length of the dentary, which is proportionally shorter than in other basal sauropodomorphs such as Plateosaurus (Brown and Schlaikjer, 1940:fig. 4) and Massospondylus (SAM PK K1314; Barrett and Yates, 2006:fig. 4). It also differs from these taxa in bearing a fine central longitudinal ridge for the caudal two-thirds of its length (Fig. 3). Dentition—Most of the dentary teeth of the holotype are preserved within their alveoli. Scattered around these is a group of larger teeth that have slipped out of their sockets. These appear to be teeth from the missing right maxilla. A further isolated tooth (BP/1/6842; Fig. 8A) also appears to be a maxillary tooth. Each maxillary tooth is lanceolate, with a straight, vertical apicobasal axis. The crown is mesiodistally expanded above its base, reaching a maximum width at about a third of the height of the crown. The crowns have a slenderness index (SI = length of a crown divided by its maximum mesiodistal width; Upchurch, 1998) of 1.69 to 1.89. The cross-sectional shape of the crown is elliptical with sharp carinae developed along the mesial and distal edges. There are shallow, narrow sulci on the lingual side that extend parallel to each carina (Fig. 8B). Both mesial and distal carinae bear numerous, coarse sharply pointed and apically angled denticulations at a density between 9 and 11.5 denticulations per 5 mm. The denticulations extend from the apex to the widest point of the crown, and so occupy more than half the height of the crown. The enamel surface is polished and smooth. The dentary teeth (Fig. 8C) are very similar to the maxillary teeth, differing only in the absence of the paramarginal sulci on the lingual side and slightly more slender crowns (SI is 2.13 in middle dentary teeth). Axial Skeleton Sacral Vertebra—The sacrum is represented by a single sacral centrum (BP/1/6851; Fig. 9) that had separated from its neural arch and both sacral ribs along their sutural surfaces, prior to final burial. The element probably derives from the caudal end of the sacrum based on the greater depth of the caudal face compared to the cranial face but it is not possible to determine if it represents a Downloaded By: [Bonnan, Matthew F.] At: 17:42 11 May 2011 second primordial sacral or a caudosacral. The centrum is a stout element and is slightly shorter than it is deep. The central body, below the sutural scars for the sacral ribs, is unusually strongly FIGURE 8. Teeth of Arcusaurus pereirabdalorum, gen. et sp. nov. A, isolated maxillary tooth (BP/1/6842) in labial view. B, loose rostral maxillary tooth of the holotype (BP/1/6235) in lingual view. C, 10th, 11th and 12th dentary teeth of the holotype (BP/1/6235) in labial view. Scale bar equals 10 mm. 618 JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 31, NO. 3, 2011 TABLE 2. Dimensions (mm) of the vertebrae and humerus of Arcusaurus pereirabdalorum, gen. et sp. nov. Distal Sacral caudal Humerus (BP/1/6851) (BP/1/6235) (BP/1/6850) Length Proximal centrum face: width Proximal centrum face: depth Distal centrum face: width Distal centrum face: depth Distal transverse width Depth of ulnar condyle Depth of Radial condyle Minimum transverse shaft width (as preserved) 31 35 36 ∗ 35 39 — — — — 18 10 7 9 7 — — — — — — — — — 45.2 16.4 17.1 19.6 Asterisk indicates an estimate based doubling the distance between the lateral margin and the midline. Downloaded By: [Bonnan, Matthew F.] At: 17:42 11 May 2011 The neural arch is separated from the centrum by a longitudinal sulcus that is bordered dorsally by a ridge. The prezygapophyses are short prongs that protrude beyond the centrum for a distance of 2.2 mm, which is equivalent to 12% of the total length of the centrum. Distally the postzygapophyses are borne on a single median process, from which a very small, caudodorsally oriented neural spine arises. Forelimb Humerus—The distal humeral fragment has a maximum transverse width of 45 mm, whereas the minimum preserved transverse shaft width is 19 mm (Table 2).<Table2> The strong difference between these two measurements suggests that the distal end of the humerus was transversely expanded relative to total bone length, a derived condition seen in most other basal sauropodomorphs. A cubital fossa is centrally placed on the distal end of the flexor surface (Fig. 10A) but is not as deep or well defined as in the humeri of other basal sauropodomorphs, such as Massospondylus (BP/1/4732), except Anchisaurus (Yates, 2004:fig. 10). However, this may well be due to the immaturity of both Anchisaurus and Arcusaurus. For most archosaurs, it is now known that juveniles possess thicker epiphyseal cartilage than adults, and that the fossilized calcified cartilage remnant of fossil long bones typically does not match the shape of the living joint surface (Bonnan et al., 2010; Holliday et al., 2010). These studies support our inference that much of the poorly developed distal articular anatomy described for the humerus in Arcusaurus are probably tied to the immature stage of development of this specimen. The radial and ulnar condyles are merged into a contiguous articular surface. In distal view, this articular surface consists of a lateral and distal expansion separated by a central constriction (Fig. 10C). The radial articular surface occupies the constricted zone and extends onto the lateral expansion. Unlike most other basal saurischians, such as Herrerasaurus (Sereno, 1994:fig. 4a), Saturnalia (Langer et al., 2007:fig. 6c) and Massospondylus (BP/1/4732), the radial articular surface does not curve up onto the flexor surface (Fig. 10A), another similarity shared with Anchisaurus (YPM 1883) that may be due to skeletal immaturity. The lateral expansion of the distal articular surface curves slightly up onto the medio-extensor surface (Fig. 10B). Laterally the articular surface merges smoothly with the ectepicondylar prominence without an intervening sulcus. The ulnar condyle makes up the medial expansion of the distal articular surface and is larger than the lateral expansion. Unlike mature specimens of basal saurischians, such as Herrerasaurus (Sereno, 1994:fig. 4b) and Massospondylus (BP/1/4732), the articular surface of this condyle does not curve up onto the FIGURE 9. Centrum of second primordial sacral vertebra of Arcusaurus pereirabdalorum, gen. et sp. nov. (BP/1/6851), in A, cranial, B, left lateral, C, caudal, D, right lateral, E, ventral, and F, dorsal views. waisted for a sauropodomorph and the ventral surface is narrow and pinched in transverse section. The central faces are concave and were apparently subcircular in outline when complete. The attachment scars for the sacral ribs are located against the cranial margin and extend for 90% of the length of the centrum. The scars form raised pedicels but are deeply hollowed. Each scar bears a caudodorsally located pit that is offset from the rest of the scar by a tall, obliquely oriented ridge. This morphology is present on both sides and is therefore likely to be a real feature rather than a chance configuration of an irregular surface. Although a similar morphology has not been reported in any other dinosaur, it must be noted that the rugose surfaces of sacral rib attachment scars are rarely described. Consequently it is not certain if this feature is diagnostic of Arcusaurus. Caudal Vertebra—The distal caudal vertebra has an elongate centrum that is 2.4 times longer than it is high. The articular faces of the centrum are dorsoventrally compressed ovoids that are 1.3 times wider than high. This morphology is not a result of deformation because the vertebra was buried on its side with the coronal plane at a steep angle to bedding, which should have resulted in asymmetrical transverse compression if any deformation had occurred. However, the vertebra remains bilaterally symmetrical. YATES ET AL.—NEW BASAL SAUROPODOMORPH 619 FIGURE 10. Distal end of the right humerus of Arcusaurus pereirabdalorum, gen. et sp. nov. (BP/1/6850), in A, cranial, B, caudal, C, distal, D, medial, and E, lateral views. Scale bar equals 20 mm. Downloaded By: [Bonnan, Matthew F.] At: 17:42 11 May 2011 extensor surface. In many other basal sauropodomorphs, the entepicondyle bears a rather flattened mediodistally facing surface of unfinished bone that is separated from the ulnar condyle by a marked ridge. This can be seen in Anchisaurus (YPM 1883), Massospondylus (BP/1/4732), and Plateosauravus (SAM PK 3342). This ridge is absent in Arcusaurus and the surface of the ulnar condyle curves gradually into the entepicondyle without a perceptible division between the two. Similarly, the sharp crest at the summit of the entepicondyle that divides the mediodistal unfinished facet from the periosteal surface above it is not present in Arcusaurus, thus rendering the entepicondyle as an indistinct rounded prominence on the medial side of the distal end of the humerus. This may also be a consequence of the immaturity of the specimen and the absence of an epiphyseal cap of calcified cartilage. A broad, shallow, weakly defined olecranon fossa is present on the extensor surface of the distal end. It is similar in size and depth to those of other basal sauropodomorphs such as Massospondylus (BP/1/4732). Pelvis and Hind Limb Ilium—The fragment of the ilium (BP/1/6926) reveals little about its anatomy. The ischiadic peduncle is a robust, ventrally projecting process, as is to be expected in a non-eusauropod sauropodomorph (Fig. 11). The distal caudal projection (‘heel’) of the ischiadic peduncle that is frequently seen in basal plateosaurians is absent. Medially, the striated scar for the articulation of the second primordial sacral rib is well developed and impressed into the surface (Fig. 11B). Non-Terminal Phalanges—The single complete non-terminal phalanx (BP/1/6846) is unusually stout for a non-plateosaurian sauropodomorph (Fig. 12A, B). Its total length is equivalent to the maximum width of the bone at its proximal end. The distal articular end is distinctly ginglymoidal and bears a ligament pit on each side. The proximal articular surface is roughly semicircular in shape and is evenly concave, indicating that it did not form a ginglymoidal joint. By comparison with Massospondylus (Cooper, 1981:fig. 81) and identification of phalanx IV-1 would seem likely. Ungual Phalanges—Both preserved unguals appear to be pedal unguals based on the lack of flexor tubercles and their close correspondence with the unguals of pedal digits I and III of Massospondylus (Cooper, 1981:figs 78–81) and Gryponyx (SAM PK 3358). One noticeable difference between these unguals and the latter taxa is that the ungual of digit I is not quite as strongly enlarged relative to the ungual of digit III. Although the tips are missing from both unguals, their total lengths can be estimated with reasonable confidence and it would appear that the third ungual was between 75% and 80% of the length of the first. By contrast, this proportion ranges from 71% to 72% in Massospondylus and Gryponyx (BP/1/4780; SAM PK 3358). The ungual of pedal digit I is deep and triangular in lateral/medial view, with a moderately curved ventral margin (Fig. 12D, E). In transverse section, the ventral margin forms a broad flattened surface that is weakly canted dorsolaterally. The medial edge of the ventral surface is marked by a weak, rounded carina, whereas the lateral side is rounded and smoothly grades up into the lateral surface. There are simple neurovascular grooves on each side that lack the proximal bifurcation seen in some morederived sauropodomorphs (Pol and Powell, 2007b). The groove on the medial side is much shallower than that on the lateral side and is set lower. The proximal articular surface is a tall ovoid with a weak central ridge (Fig. 12F). It is strongly concave dorsoventrally. The ungual of pedal digit III is wider, relative to height, than that of pedal digit I (Fig. 12I). The broad ventral surface is not as strongly curved in lateral view and is not canted dorsolaterally in FIGURE 11. Fragment of left ilium of Arcusaurus pereirabdalorum, gen. et sp. nov. (BP/1/6926), in A, lateral and B, medial views. Scale bar equals 20 mm. 620 JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 31, NO. 3, 2011 synapomorphies, along with those of the intervening nodes, are given as online supplementary material (Supplementary Data 1; www.vertpaleo.org/jvp/JVPcontents.html). To further test the position of Arcusaurus, in the light of the unusually derived characteristics it displays, a constraint tree was built that includes Arcusaurus amongst the Plateosauria. No other topology was imposed. The shortest tree that conforms to the constraint was then searched for. The shortest trees found were 1141 steps long (an increase of three steps) and place Arcusaurus in a relatively derived position, as the sister group to Aardonyx + the ‘quadrupedal clade,’ within Anchisauria. A randomly selected example from this set of trees was compared to a randomly selected member of the set of most parsimonious trees in a Templeton Test. The difference was found to be insignificant at the 0.05 level (P = 0.6434), indicating that the basal position of Arcusaurus is not a significantly better explanation of the data than a derived position within Anchisauria. It must be noted that the humeral characters 210 and 212 (development of the cubital fossa and the shape of the entepicondyle) are scored as preserved for Arcusaurus. However, it is noted in the description that these states might be due to ontogeny and it could be argued that these characters are more accurately scored as unknown in Arcusaurus. A modified version of the matrix that changes the scores for characters 210 and 212 to ‘?’ was analyzed in the same manner as the original matrix. These changes were found to have no affect at all on tree topology or length in both the unconstrained analysis and the constrained analysis. These results indicate that the humeral states displayed by Arcusaurus are fully compatible with either of the two positions found above. Finally Arcusaurus was included in the matrix from Upchurch et al. (2007) in order to determine if a different set of characters and assumptions would also support a basal position for Arcusaurus (see Appendix 1 for the character scores). The same analytical procedure used in Upchurch et al. (2007) was employed, as were the scoring revisions suggested by Yates et al. (2010). The strict consensus of the 600 MPTs (each with a length of 751 steps) is rather poorly resolved (Fig. 13B). Nevertheless Arcusaurus forms part of a basal polytomy with Saturnalia, Thecodontosaurus, and the clade of all other sauropodomorphs, adding further support that Arcusaurus is indeed a non-plateosaurian sauropodomorph. DISCUSSION Is Arcusaurus a Juvenile? The open sutures of the braincase and sacrum strongly suggest that the known specimens of Arcusaurus are juveniles, as do the poorly defined features and shallow cubital fossa of the humerus. Therefore, it is important to determine whether or not the differences observed between Arcusaurus and contemporary taxa are not simply due to ontogeny. It is particularly important to investigate the possibility that Arcusaurus does not represent a very early stage in the ontogeny of the newly described Aardonyx, also from Spion Kop. Because an ontogenetic sequence of Aardonyx is unknown, the best that can be done is to compare the features that do vary between the two taxa to known ontogentic sequences from other sauropodomorphs. The features that distinguish Arcusaurus from Aardonyx as listed in the differential diagnosis are examined here in turn for evidence that they may be the result of ontogenetic differences. 1. Absence of lateral plates. An embryonic premaxilla of a neosauropod, referred to Camarasaurus, bears a welldeveloped lateral plate (Britt and Naylor, 1994:fig. 16.1a), indicating that lateral plates, if present, appear well developed at extremely early stages of ontogeny. Thus the absence of this Downloaded By: [Bonnan, Matthew F.] At: 17:42 11 May 2011 FIGURE 12. Pedal phalanges of Arcusaurus pereirabdalorum, gen. et sp. nov. A–C, right non-terminal pedal phalanx, probably IV-1, (BP/1/6846) in A, lateral, B, dorsal, and C, proximal views. D–F, Right pedal ungual of digit I (BP/1/6848) in D, medial, E, lateral, and F, proximal views. G–I, Left pedal ungual of digit III (BP/1/6849) in G, lateral, H, medial, and I, proximal views. Scale bar equals 10 mm. transverse section. Both sides of the ventral surface have margins formed by rounded carinae, with a sharper medial carina. The medial neurovascular groove is more strongly impressed than in pedal ungual I but is still shallower than the groove on the lateral side. Both grooves terminate proximally without bifurcation but unlike pedal ungual I, the proximal ends of the grooves curve ventrally to meet the margins of the ventral surface. Proximally the articular surface is subcircular and nearly flat, with a weak central ridge. PHYLOGENETIC ANALYSIS Arcusaurus was included in a modified version of the character taxon matrix used by Yates (2007). The modifications were made by Yates et al. (2010) in order to test the phylogenetic position of Aardonyx and are discussed in the online supplementary information of that paper, where the matrix can be found. Here we simply provide the character state scores for Arcusaurus (Appendix 1). The new matrix was analyzed using PAUP 4.0b (Swofford, 2002). The multistate characters that were treated as ordered in Yates (2007) were again treated as ordered. The heuristic search option was used with the following parameters: 300 addition sequence replicates with TBR branch swapping. Euparkeria was set as the user-defined outgroup. The search resulted in 104 most parsimonious trees with a length of 1138 steps. The strict consensus of these 104 trees includes Arcusaurus in a basal position nested between Pantydraco and Thecodontosaurus. The strict consensus only includes polytomies within the relatively derived plateosaurian sector of the tree while the relationships outside Plateosauria are fully resolved (Fig. 13). Decay analysis indicates that most of the relevant nodes are weak and collapse with the addition of just a single step. The strongest clades (each with a decay index of three) are Pantydraco + all more-derived sauropodomorphs; Efraasia + all more-derived sauropodomorphs; and Plateosauravus + all more-derived sauropodomorphs. These nodes happen to bracket the position of Arcusaurus and their supporting YATES ET AL.—NEW BASAL SAUROPODOMORPH 621 Downloaded By: [Bonnan, Matthew F.] At: 17:42 11 May 2011 FIGURE 13. Results of the phylogenetic analyses. A, a portion of the strict consensus of 104 MPTs found after analysis of the modified version of the Yates (2007) matrix. Outgroups basal to Theropoda are omitted and several polyspecific higher taxa (Plateosauridae, Riojasauridae, Massospondylidae, and the ‘quadrupedal clade’) are collapsed into single terminals in order to save space. Two positions for Arcusaurus are shown: the position found in the MPTs (solid black line) and the suboptimal position found when Arcusaurus is constrained to fall within Plateosauria (dashed grey line). Decay index values appear above each branch. B, a portion of the strict consensus of 600 MPTs found after analysis of the modified version of the Upchurch et al. (2007) matrix. Outgroups basal to Theropoda are omitted and Anchisauria is collapsed into a single terminal. Stem-based taxa are marked with an open bracket and nodebased taxa are marked with a solid spot. structure in Arcusaurus is unlikely to be a consequence of its immaturity. 2. Four premaxillary teeth. The number of premaxillary teeth appears to be fixed through ontogeny of sauropodomorphs based on the presence of four premaxillary teeth in juvenile and adult Massospondylus (Sues et al., 2004), and embryonic, juvenile, and adult Camarasaurus (Britt and Naylor, 1994; Madsen et al., 1995), so this difference between Arcusaurus and Aardonyx is not due to ontogeny. 3. Straight ventral margin of the dentary with an acutely pointed rostral tip. The ontogenetic series of Massospondylus and Mussaurus demonstrate that the ventral curvature of these taxa does increase through ontogeny but is still weakly present in early ontogenetic stages. Furthermore, the shape of the rostral termination appears not to vary strongly in both of these taxa. Early-stage juvenile Mussaurus have a deep, eusauropod-like rostral expansion of the dentary, as do laterstage juvenile specimens over twice the size of the early stage juveniles (Pol and Powell, 2007a). Similarly, all ontogenetic stages of Massospondylus show a blunt rostral termination of the dentary, with a rounded ventral corner (Sues et al., 2004:figs. 1a, 5a). Therefore, it seems very unlikely that the straight and acutely pointed rostral tip of the dentary of Arcusaurus would grow into the squared-off and ventrally curved termination seen in Aardonyx (Yates et al., 2010:fig. 1f). 4. Single large neurovascular foramen at the rostral tip of the dentary. The bone surface at the rostral end of the dentary of BP/1/4376, the juvenile Massospondylus figured in Sues 622 JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 31, NO. 3, 2011 mains uncertain whether or not the smooth juvenile teeth could have been replaced by the fluted and finely wrinkled teeth of Aardonyx. Thus, of the eight characters (aside from the autapomorphies of Arcusaurus) to distinguish Arcusaurus from Aardonyx listed above, five appear to be invariant in known sauropodomorph growth series, one is unknown in any ontogenetic series, the ontogenetic significance of one is ambiguous, whereas the other may be the result of ontogeny but would represent a greater degree of change than is represented in known growth series from other taxa. Thus the hypothesis that Arcusaurus represents a very young juvenile stage of Aardonyx (itself represented by juvenile material) is rejected. It is easier to distinguish Arcusaurus from Massospondylus because the latter is known from juvenile specimens (BP/1/4376 and BP/1/6161) that are of equivalent size to the known Arcusaurus specimens. These juvenile Massospondylus display all of the traits used to distinguish the two taxa in the differential comparisons given above. Turning to its phylogenetic position, Arcusaurus displays an interesting mix of characteristics that suggest two rather different positions within sauropodomorph phylogeny. Many features support a position outside Plateosauria. Plateosauria is defined as the least inclusive clade containing Plateosaurus and Massospondylus (Sereno, 1998), which in our analysis also includes Sauropoda. Synapomorphies of Plateosauria, or more inclusive clades, that are absent in Arcusaurus are listed here by clade. Thecodontosaurus + more-derived sauropodomorphs (including Plateosauria): a caudal ridge on the lateral surface of the dentary; 20 dentary teeth or more. Plateosauria: ventral curvature of the alveolar margin of the dentary at its rostral end; the base of the rostroventral process of the nasal 50% broader than the craniodorsal process; ungual of pedal digit III reduced relative to the unguals of digits I and II. In addition to these characteristics, the maximum depth of the dentary is more than 20% of its length. This is a transient synapomorphy of Sauropodomorpha that is reversed in the clade of Efraasia + more-derived sauropodomorphs (including Plateosauria) and also supports a basal, non-plateosaurian position for Arcusaurus. In contrast to these characters are a number of derived characteristics that are otherwise only known from less inclusive clades within Plateosauria. These include two synapomorphies of Massospondylidae + more-derived plateosaurians: the absence of a distinct rim of the narial fossa impressed upon the premaxillary body; complete separation of the caudolateral process of the premaxilla and the rostroventral process of the nasal; and a synapomorphy of Aardonyx + the ‘quadrupedal clade’: the presence of at least some stout non-terminal pedal phalanges that are as wide as they are long. When Arcusaurus is analyzed in a modified version of the Yates (2007) matrix, it is the latter set of characters that are interpreted as convergences and a relatively basal position within the ‘thecodontosaurid grade’ taxa (namely Thecodontosaurus and Pantydraco) is upheld, albeit not significantly. A similar result is obtained with a different matrix (Upchurch et al., 2007, modified by Yates et al., 2010). It might be argued that the characters supporting the more basal position are the result of immaturity. However, an examination of the characters supporting a basal position shows that only the number of dentary teeth is possibly related to ontogenetic change. Arcusaurus as a Relictual Taxon 5. 6. 7. 8. et al. (2004:fig. 1), is damaged, making it difficult to discern the openings of the neurovascular foramina. However, a new specimen of M. carinatus (BP/1/6161), of approximately the same size as BP/1/4376, shows a clear cluster of neurovascular foramina at the rostral end of the dentary, as does the adult specimen (BP/1/4934), indicating that the single foramen seen in Arcusaurus is unlikely to be due to its immaturity. Absence of a dense band of fine pits surrounding the alveolar margins of the premaxilla, maxilla, and dentary. The band of dense fine pits around the alveolar margins of the jaws of Aardonyx are an autapomorphic characteristic, consequently nothing is known of its ontogeny. However, bone surface texture is known to vary with ontogeny (e.g., Brown et al., 2009), so it is possible that this character does vary with ontogeny. Absence of a strong medial curvature of the rostrodorsal ramus of the postorbital. The rostral ramus of the postorbital is more strongly inflected towards the medial side in mature specimens of Massospondylus than in juveniles of the same species (Fig. 14). Thus it is possible that the relatively straight rostral ramus of Arcusaurus in comparison to the strongly curved ramus of Aardonyx could be due to ontogeny, although it must be noted that if these do indeed represent ontogenetic stages, the change in curvature exceeds that exhibited by Massospondylus. Denticulations of the maxillary and dentary teeth extending into the basal half of the crowns. Comparison of juvenile teeth of Massospondylus with those of adults indicates that there is little ontogenetic change in the density and distribution of serrations in this taxon (Sues et al., 2004:fig. 8). Similarly, the teeth of early- and late-stage juvenile Mussaurus are both weakly serrated (Pol and Powell, 2007b). Both embryonic and adult teeth of Camarasaurus have smooth marginal carinae devoid of serrations (Madsen et al., 1995; Britt and Naylor, 1994). Consequently there is no evidence to support the hypothesis that the strongly, and extensively denticulated teeth of Arcusaurus are an earlier ontogenetic stage of Aardonyx, where denticles are weakly developed and are restricted to just a few at the apex of the crown. Smooth enamel of the tooth crowns. The embryonic Camarasaurus (Britt and Naylor, 1994) tooth also shows that the derived state of wrinkled enamel is expressed at hatching. On the other hand, early post-hatching Mussaurus have smooth enamel (PVL 4210), whereas a larger juvenile has faintly wrinkled premaxillary teeth (Pol and Powell, 2007a). Thus it re- Downloaded By: [Bonnan, Matthew F.] At: 17:42 11 May 2011 FIGURE 14. Ontogenetic change in the postorbital of Massospondylus carinatus. A, Right postorbital of juvenile, BP/1/4376, in dorsal view. B, Left (reversed for comparison) postorbital of adult BP/1/4779, in dorsal view. Scale bars equal 10 mm. Grey lines represent long axes of the rostromedial and caudal rami of the dorsal end of the postorbital. Arcusaurus is significant because it may indicate the extension of a basal lineage of sauropodomorphs from the Late Triassic into the Early Jurassic. Its most parsimonious position is as a member of a polytomy at the base of Sauropodomorpha in an analysis YATES ET AL.—NEW BASAL SAUROPODOMORPH based on a modified version of the Upchurch et al. (2007) matrix and as the sister group to the clade Thecodontosaurus + all morederived sauropodomorphs (including Plateosauria) in an analysis based on a modification of the Yates (2007) matrix. Efraasia, from the middle Norian of Germany, represents the earliest known occurrence of the latter clade. If a basal position is accepted, then Arcusaurus would have had to have diverged no later than the middle Norian, indicating a ghost lineage of up to 35 million years (depending upon which age estimate for the upper Elliot Formation is chosen). It has been suggested that the Triassic-Jurassic boundary may actually lie above the boundary between the lower and upper members of the Elliot Formation based on the occurrence of a cheirotheroid trackway in the upper Elliot Formation (Smith et al., 2009). Cheirotheroid tracks are usually attributed to non-crocodylomorph crurotarsans or more basal archosauriforms (Haubold, 1971), which are extinct by the close of the Late Triassic (Brusatte et al., 2010:fig. 8). Thus part, or all, of the upper Elliot Formation might actually be Late Triassic in age, and the relictual status of Arcusaurus would be falsified. However, we regard the trackway evidence as inconclusive because it is possible that it was produced by a large basal crocodylomorph crurotarsan. Basal crocodylomorphs are well known from the Early Jurassic (Wu and Chatterjee, 1993; Clark and Sues, 2002) and are known to reach sizes similar to rauisuchians (Nesbitt et al., 2005). Furthermore, some basal crocodylomorphs may well have retained a hooked fifth metatarsal given that this morphology is retained in the more-derived crocodyliform Orthosuchus (Nash, 1975:fig. 17), and therefore such archosaurs would have been capable of producing a cheirotheroid track. No other non-plateosaurian sauropodomorphs are known to have survived into the Jurassic. The recently described Ignavusaurus, also from the upper Elliot Formation, was posited to be a non-plateosaurian sauropodomorph (Knoll, 2010). The characters cited as suggesting a non-plateosaurian position were as follows: the longest proximal chevron twice the length of the preceding chevron; a rounded entepicondyle lacking a flattened distomedial facet; a square distal end of the tibia; and a femur less than 400 mm in length. None of these are especially compelling. Elongate proximal chevrons that are more than twice the length of the preceding centrum are widespread within basal Sauropodomorpha, including the plateosaurians Massospondylus (BP/1/4934) and Yunnanosaurus (Young, 1942:fig. 7). The lack of a distomedial facet of the entepicondyle may well be a function of the immaturity of the specimen and the poor preservation of the spongy articular ends of the long bones. A similar situation exists within the type series of Plateosauravus cullingworthi. The larger of the two humeri in this series displays a well-developed distomedial entepicondylar facet, whereas the entepicondyle of the smaller humerus is simple and rounded. The lack of transverse expansion of the distal end of the tibia is a plesiomorphic condition seen in non-plateosaurian sauropodomorphs such as Thecodontosaurus (e.g., BRSUG 23621) and Saturnalia (Langer, 2003:fig. 5h). However, the transverse flaring of the distal end of the tibia of Massospondylus is quite poorly developed and in some specimens the distal end of the tibia is as long as it is wide (e.g., BP/1/4789, 4998, 5006). A small femur may simply reflect the immature nature of the specimen. In any case, similarly smallsized Massospondylus specimens are frequently encountered in the upper Elliot Formation. Furthermore, the Ignavusaurus holotype shares a number of character states with the contemporary Massospondylus such as weakly developed basal constrictions of the teeth, coarse apically restricted denticulations of the teeth, and a relatively narrow pubic apron with a concave lateral margin. Many of the purported differences are quite subtle and do not take the range of variation present in Massospondylus into consideration. Consequently we regard the type and only specimen of Ignavusaurus rachelis to be a damaged juvenile specimen 623 of either Massospondylus carinatus or its close relative M. kaalae. This leaves Arcusaurus perierabdalorum as the sole candidate for a Jurassic non-plateosaurian sauropodomorph. ACKNOWLEDGMENTS We thank L. Pereira for discovery of the holotype specimen and F. Abdala for most of the referred specimens. The following people participated in the field work at Spion Kop: Z. AliJinnah, N. Barbolini, N. Bremmer Jr., J. Cisneros, G. DeVilliers, ¨ C. Dube, S. Fowell, J. Hancox, K. Lalla, C. McCrae, R. Morsner, M. Nicolas, L. Norton, S. Potze, N. Sithole, and C. Vasconcelos. The holotype specimen was skillfully prepared by C. Dube. We thank P. Barrett and D. Pol for their thoughtful reviews of the manuscript. The field work was funded by the National Geographic Society (CRE no. 7713–04). M.F.B. was supported in part by a Faculty Mentor grant from the College of Arts and Sciences at WIU and a Center for Innovation in Teaching Research, Faculty Research Developmental Activities Award. LITERATURE CITED Barrett, P. M. 2004. Sauropodomorph dinosaur diversity in the upper Elliot Formation (Massospondylus range zone: Lower Jurassic) of South Africa. South African Journal of Science 100:501–503. Barrett, P. M. 2009. A new basal sauropodomorph dinosaur from the upper Elliot Formation (Lower Jurassic) of South Africa. Journal of Vertebrate Paleontology 29:1032–1045. Barrett, P. M., and A. M. Yates. 2006. New information on the palate and lower jaw of Massospondylus (Dinosauria: Sauropodomorpha). Palaeontologia africana 41:123–130. Bonnan, M.F., J. L. Sandrik, T. Nishiwaki, D. R. Wilhite, R. M. Elsey, and C. Vittore. 2010. Calcified cartilage shape in archosaur long bones reflects overlying joint shape in stress-bearing elements: implications for nonavian dinosaur locomotion. The Anatomical Record 293:2044–2055. Britt, B. B., and B. G. Naylor. 1994. An embryonic Camarasaurus (Dinosauria, Sauropoda) from the Upper Jurassic Morrison Formation (Dry Mesa Quarry, Colorado); pp. 256–264 in K. Carpenter, K. F. Hirsch, and J. R. Horner (eds.), Dinosaur Eggs and Babies. Cambridge University Press, Cambridge, U.K. Brown, B., and E. M. Schlaikjer. 1940. A new element in the ceratopsian jaw with additional notes on the mandible. American Museum Novitates 1092:1–13. Brown, C. M., A. P. Russell, and M. J. Ryan. 2009. Pattern and transition of surficial bone texture of the centrosaurine frill and their ontogenetic and taxonomic implications. Journal of Vertebrate Paleontology 29:132–141. Brusatte, S. L., M. J. Benton, J. B. Desojo, and M. C. Langer. 2010. The higher-level phylogeny of Archosauria (Tetrapoda: Diapsida). Journal of Systematic Palaeontology 8:3–47. Clark, J. M., and H.-D. Sues. 2002. Two new basal crocodylomorph archosaurs from the Lower Jurassic and the monophyly of the Sphenosuchia. Zoological Journal of the Linnean Society 136:77–95. Cooper, M. R. 1981. The prosauropod dinosaur Massospondylus carinatus Owen from Zimbabwe: its biology, mode of life and phylogenetic significance. Occasional Papers of the National Museums and Monuments, Rhodesia, series B 6:689–840. Downs, A. 2000. Coelophysis bauri and Syntarsus rhodesiensis compared, with comments on the preparation and preservation of the fossils from the Ghost Ranch Coelophysis Quarry. New Mexico Museum of Natural History and Science Bulletin 17:33–38. Galton, P. M. 1984. Cranial anatomy of the prosauropod dinosaur Plateosaurus from the Knollenmergel (Middle Keuper, Upper Trias¨ sic) of Germany. I.Two complete skulls from Trossingen/Wurt. With comments on the diet. Geologica et Palaeontologica 18:139–171. Galton, P. M. 1985. Cranial anatomy of the prosauropod dinosaur Plateosaurus from the Knollenmergel (Middle Keuper, Upper Triassic) of Germany. II. All the cranial material and details of soft-part anatomy. Geologica et Palaeontologica 19:119–159. Gow, C. E. 1990. Morphology and growth of the Massospondylus braincase (Dinosauria, Prosauropoda). Palaeontologia africana 27:58–75. Downloaded By: [Bonnan, Matthew F.] At: 17:42 11 May 2011 624 JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 31, NO. 3, 2011 Smith, R. M. H., C. A. Marsicano, and J. A. Wilson. 2009. Sedimentology and paleoecology of a diverse Early Jurassic tetrapod tracksite in Lesotho, Southern Africa. Palaios 24:672–684. Sues, H.-D., R. R. Reisz, S. Hinic, M. A. Raath. 2004. On the skull of Massospondylus carinatus Owen, 1854 (Dinosauria: Sauropodomorpha) from the Elliot and Clarens formations (Lower Jurassic) of South Africa. Annals of the Carnegie Museum 73:239– 257. Swofford, D. L. 2002. PAUP∗ : Phylogenetic Analysis Using Parsimony (∗ and Other Methods). Version 4.10b. Sinauer Association, Sunderland, Massachusetts. Upchurch, P. 1998. The phylogenetic relationships of sauropod dinosaurs. Zoological Journal of the Linnean Society 124:43–103. Upchurch, P., P. M. Barrett, and P. M. Galton. 2007. A phylogenetic analysis of basal sauropodomorph relationships: implications for the origin of sauropod dinosaurs; pp. 57–90 in P. M. Barrett and D. J. Batten (eds.), Evolution and Palaeobiology of Early Sauropodomorph Dinosaurs. Special Papers in Palaeontology 77. The Palaeontological Association, London, United Kingdom. Vasconcelos, C. C., and A. M. Yates. 2004. Sauropodomorph biodiversity of the upper Elliot Formaion (Lower Jurassic) of southern Africa. Geoscience Africa 2004, Abstract Volume. University of the Witwatersrand, Johannesburg. Wu, X.-C., and S. Chatterjee. 1993. Dibothrosuchus elaphros, a crocodylomorph from the Lower Jurassic of China and the phylogeny of the Sphenosuchia. Journal of Vertebrate Paleontology 13: 58–89. Yates, A. M. 2003. A new species of the primitive dinosaur Thecodontosaurus (Saurischia: Sauropodomorpha) and its implications for the systematics of early dinosaurs. Journal of Systematic Palaeontology 1:1–42. Yates, A. M. 2004. Anchisaurus polyzelus (Hitchcock): the smallest known sauropod dinosaur and the evolution of gigantism among sauropodomorph dinosaurs. Postilla 230:1–58. Yates, A. M. 2006. A new theropod dinosaur from the Early Jurassic of South Africa and its implications for the early evolution of theropods. Palaeontologia africana 41:105–122. Yates, A. M. 2007. The first complete skull of the Triassic dinosaur Melanorosaurus Haughton (Sauropodomorpha: Anchisauria); pp. 9–55 in P. M. Barrett and D. J. Batten (eds.), Evolution and Palaeobiology of Early Sauropodomorph Dinosaurs. Special Papers in Palaeontology 77. The Palaeontological Association, London, United Kingdom. Yates, A. M., and P. M. Barrett. 2011. Massospondylus carinatus Owen 1854 (Dinosauria, Sauropodomorpha): proposed conservation of usage by designation of a neotype. Palaeontologia africana. Yates, A. M., M. F. Bonnan, and J. Neveling. 2007. A new diverse dinosaur assemblage from the Early Jurassic of South Africa. Journal of Vertebrate Paleontology 27(3, Supplement):169A. Yates, A. M., P. J. Hancox, and B. S. Rubidge. 2004. The first record of a sauropod dinosaur from the Upper Elliot Formation (Early Jurassic) of South Africa. South African Journal of Science 100:504– 506. Yates, A. M., M. F. Bonnan, J. Neveling, A. Chinsamy, and M. G. Blackbeard. 2010. A new transitional sauropodomorph dinosaur from the Early Jurassic of South Africa and the evolution of sauropod feeding and quadrupedalism. Proceedings of the Royal Society B 277:787–794. Young, C.-C. 1942. Yunnanosaurus huangi Young (gen. et sp. nov.), a new Prosauropoda from the Red Beds at Lufeng, Yunnan. Bulletin of the Geological Society of China 22:63–104. Submitted August 23, 2010; accepted January 31, 2011. Handling editor: Hans-Dieter Sues. Haubold, H., 1971. Ichnia amphibiorum et reptiliorum fossilium. Hand¨ buch der Palaoherpetologie 18:1–124. Holliday, C. M., R. C. Ridgley, J. C. Sedlmayr, and L.M. Witmer. 2010. Cartilaginous epiphyses in extant archosaurs and their implications for reconstructing limb function in dinosaurs. PLoS One 5: e13120. ¨ Huene, F. von. 1926. Vollstandige Osteologie eines Plateosauriden aus ¨ ¨ dem schwabischen Keuper. Geologische und Palaontologische Abhandlungen, Neue Folge 15:129–179. Huene, F. von. 1932. Die fossile Reptil-Ordnung Saurischia, ihre En¨ twicklung und Geschichte. Monographien zur Geologie und Palaontologie, Series 1 4:1–361. Kitching J. W., and M. A. Raath. 1984. Fossils from the Elliot and Clarens Formations (Karoo Sequence) of the northeastern Cape, Orange Free State and Lesotho, and a suggested biozonation based on tetrapods. Palaeontologia africana 25:111–125. Knoll, F. 2010. A primitive sauropodomorph from the upper Elliot Formation of Lesotho. Geological Magazine 147:814–829. Langer, M. C. 2003. The pelvic and hind limb anatomy of the stemsauropodomorph Saturnalia tupiniquim (Late Triassic, Brasil). Paleobios 23(2):1–40. Langer, M. C., M. A. G. Franca, and S. Gabriel. 2007. The pectoral gir¸ dle and forelimb anatomy of the stem-sauropodomorph Saturnalia tupiniquim (Upper Triassic, Brazil); pp. 113–137 in P. M. Barrett and D. J. Batten (eds.), Evolution and Palaeobiology of Early Sauropodomorph Dinosaurs. Special Papers in Palaeontology 77. The Palaeontological Association, London, United Kingdom. Madsen, J. H., J. S. McIntosh, and D. S. Berman. 1995. Skull and atlas-axis complex of the upper Jurassic sauropod Camarasaurus Cope (Reptilia: Saurischia). Bulletin of Carnegie Museum of Natural History 31:1–115. Nash, D. S. 1975. The morphology and relationships of a crocodilian, Orthosuchus stormbergi, from the Upper Triassic of Lesotho. Annals of the South African Museum 67:227–329. Nesbitt, S. J., R. B. Irmis, S. G. Lucas, and A. P. Hunt. 2005. A giant ¨ crocodylomorph from the Upper Triassic of New Mexico. Palaontologische Zeitschrift 79:471–478. Pol, D., and J. E. Powell. 2007a. Skull anatomy of Mussaurus patagonicus (Dinosauria: Sauropodomorpha) from the Late Triassic of Patagonia. Historical Biology 19:125–144. Pol, D., and J. E. Powell. 2007b. New information on Lessemsaurus sauropoides (Dinosauria: Sauropodomorpha) from the Upper Triassic of Argentina; pp. 223–243 in P. M. Barrett and D. J. Batten (eds.), Evolution and Palaeobiology of Early Sauropodomorph Dinosaurs. Special Papers in Palaeontology 77. The Palaeontological Association, London, United Kingdom. Sampson, S. D., and L. M. Witmer. 2007. Craniofacial anatomy of Majungasaurus crenatissimus (Theropoda: Abelisauridae) from the Late Cretaceous of Madagascar; pp. 32–102 in S. D. Sampson and D. W. Krause (eds.), Majungasaurus crenatissimus (Theropoda: Abelisauridae) from the Late Cretaceous of Madagascar. Society of Vertebrate Paleontology, Memoir 8. Seeley, H. G. 1887. On the classification of the fossil animals commonly called Dinosauria. Proceedings of the Royal Society of London 43:165–171. Sereno, P. C. 1994. The pectoral girdle and forelimb of the basal theropod Herrerasaurus ischigualastensis. Journal of Vertebrate Paleontology 13:425–450. Sereno, P. C. 1998. A rationale for phylogenetic definitions, with application to the higher-level phylogeny of Dinosauria. Neues Jahrbuch ¨ ¨ fur Geologie und Palaontologie, Abhandlungen 210:41–83. Sereno, P. C., C. A. Forster, R. R. Rogers, and A. M. Monetta. 1993. Primitive dinosaur skeleton from Argentina and the early evolution of Dinosauria. Nature 361:64–66. Downloaded By: [Bonnan, Matthew F.] At: 17:42 11 May 2011 YATES ET AL.—NEW BASAL SAUROPODOMORPH APPENDIX 1. Character state scores for Arcusaurus. Yates (2007) matrix ?0?1??0[12]1?0??????0?010?1??????????????????? ??0??????000?????????????????????????????????? 0????0?01010?????011?0101010000??????????????? ?????????????????????????????????????????????? ??????????00????????????????0?0??????????????? ?????????????????????????0???????????????????? ?????????????????????????????????????????????? ??????????????????????[12]???0?0????[12] 625 Upchurch (2007) matrix ????11???000?????????????0011????????0???????? ?0?????????????????????0?00001??????0101100001 00?1000??????????????????????????????????????? ????????????????????????????????1????????????? ?????????????????????????????00??????????????? ?????????????????????????????????????????????? ????1?????01??? Downloaded By: [Bonnan, Matthew F.] At: 17:42 11 May 2011