Thunder-Lizards
LIFE OF THE PAST James O. Farlow, Editor
INDIANA UNIVERSITY
PRESS
Bloominston and Indianapolis
Thundertizards
The Sauropodomorph Dinosaurs Edited by Virginia Tidwell and Kenneth Carpenter
This book rs a publication of L.rdiana University Press 601 North Morton Street
Bloomington, IN 47404-3797 USA http://iupress.indiana.edu
Telephctneorders 800-842-6796
orders Ordersbye-mail
Fax
812-8.55-7931
[email protected] O 2005 bv Indiana Universitr.Press
AII rights reserr.ed No part of this book may be reproduced or utilized in anv form or bt'any means, electronic or mechanical, including photocopving and recording, or b,v any informatior.r storage and retrieval system, without permission in rvriting fron'r the publisher. The Association of American Universit,v Presses' Resolution on Permissions constifutes the onlv exception to this prohibition. The paper used in this publication meets rhe minimum requirements of American National Standard for Information Sciences-Permanence of Paper for Printed Librarv Materials, ANSI
239.48-1984. N,{anufactr:red in the United States of America
Librarv of Congress Cataloging-in-Publicatron Data Thunder-lizards : the Sauropodomorph dinosaurs / edited by Virginia Tidwell and Kenneth Carpenter. p. cm.-(Life of the past) L.rcludes bibliographical references and inder. ISBN 0-253-345.{2-1 (cloth : alk. paper)
1. Saurischia. 2. Saurischia-Anaromy.
3. Saurischia-Aging. 4. Saurischia-Infancv. 5. Saurischia-Evolution. 5. SaurischiaMorphologv. 7. Animal mechanics. i. Tidr.vell,
Virginia. II.
Carpenter, Kenr.reth. QE862.53T48 2005
III.
Series.
567.913-dc22 2004018474
r2345100908070605
CONTENTS
Contributors
PART 1
ONE:
Sauropods Old and New
. Postcranial Anatomy of Referred Specimens of the Sauropodomorph Dinosaur Melanoroslurus from the Upper Triassic of South Africa Peter
2. 3
.
ix
L
M. Galton, JacquesVan Heerden, and Adam M. Yates
The Genus Barosaurus Marsh (Sauropoda, Diplodocidae) 38 John S. Mclntosh Reassessment of the Early Cretaceous Sauropo d Astrodon
johnsoni Leidy 1865 (Titanosauriformes) 78 Kenneth Carpenter and Virginia Tidwell
4
.
Osteology of Ampelosawrws atacis (Titanosauria) from Southern France 115 Jean Le Loeuff PART
T\7O:
Sauropods Young to Old
5 . New Juvenile Sauropod Material from Sfestern Colorado, and the Record of Juvenile Sauropods from the Upper Jurassic
Morrison
Formation
1'41,
John R. Foster 6 . New Adult Specimens of CamarasAurus lentus Highlight Ontogenetic Variation within the Species 154 Takehito lkejiri, Virginia Tifuaell, and Dauid L. Trexler
.
\Se -Related Cl'rar:rcterrstics For,rnd
in a Partial Pelvis of Camaraslurus \-irgini; Tidu,ell, Kenneth Stadttnan, and Allen Shaw S
11
Ontogeneric Variation and Isometric Growth in rhe Forelimb of the Early Cretaceous Sauropod Venenosaurus Virginia Tidwell and D. Ray Wilhite
PART
THREE: Body Parts: Morphology and Biomechanics
.
Neuroanatomy and Dentition of Camarasaurus lentus Sankar Chanerjee and Zhong Zheng
9
.
10
.
.
Neck Posture, Dentition, and Feeding Strategies in Jurassic Sauropod Dinosaurs Kent A. Steuens and J. Michael Parrish Neck Posture of Sauropods Determined Using Radiological Imaging to Reveal Three-Dimensional Structure of Cervical Vertebrae Dauid S. Berman and Bruce M. Rothschild 12
13
'
.
180
t87
199
21,2
L.).1
Evolution of the Hyposphene-Hypanrrum Complex within Sauropoda 248 Sebastidn Apesteguia
variation in the Appendicular Skeleton of North American
SauropodDinosaurs:Taxonomiclmplications 268 D. Ray'Wilhite
14
.
First Articulated Manus of Diplodocus carnegii 302 Malcolm'W. Bedell Jr and Dauid L. Trexler 15
.
Evolution of the Titanosaur
Metacarpus 3ZI
Sebastidn Apesteguia
16
.
Pes
Anatomr- in Sauropod Dinosaurs: Implications for Functional Morphology, Evolution, and Phylogeny 346 Mattheu F. Bonnan
17
vi .
Contents
.
Sauropod Stress Fracures as Clues to Activity Bruce M. Rothschild and Ralph E. Molnar
381
PART 18
.
FOUR: Global Record of Sauropods
Between Gondwana and Laurasia: Cretaceous Sauropods in an Intraoceanic Carbonate Platform
395
Fabio M. DaIIaVecchia 19
.
Sauropods of Patagonia: Systematic Update and Notes on Global Sauropod Evolution Leonardo Salgado and Rodolfo A. Coria
20
21
.
.
Observations on Cretaceous Sauropods from Australia Ralph E. Molnar and Steuen W. Salisbury
430 454
Late Cretaceous (Maastrichtian) Nests, Eggt, and Dung Mass (Coprolites) of Sauropods (Titanosaurs) from India D. M. Mohabev
466
Index
491,
Contents
.
vii
CONTRIBUTORS
Sebasti6n Apesteguia, Museo Argentino de Ciencias Naturales "B. Rivadavia," Av. Angel Gallardo 470, (1405) Buenos Aires, Argenuna. Malcolm'\i7. Bedell Jr., Big Horn Basin Foundation, P.O. Box 868,
Thermopolis,'V7yoming 82443.
David S. Berman, Section of Vertebrate Paleontology, Carnegie Museum of Natural History, 4400 Forbes Avenue, Pittsburgh, Pennsylvania 1.521.3.
-$Testern
^\latthew F. Bonnan, Department of Biological Sciences, Illinois University, Macomb, Illinois 61455. Kenneth Carpenter, Department of Earth Sciences, Denver Museum of Natural History, 2001 Colorado Blvd., Denver, Colorado 80205. Sankar Chatterjee, Museum of Texas Tech University, Box 43191, Lubbock, Texas 7 9 409-319 1. Rodolfo A. Coria, Direcci6n Provincial de Cultura-Museo Carmen Funes, Av. C6rdoba 55 (8318) Plaza Huincul, Neuqu6n, Argentina. Fabio M. Dalla Vecchia, Museo Paleontologico Cittadino di Mon-
falcone (Gorizia), Via Valentinis 134, I-34074 Monfalcone (Gorizia), Italy.
Iohn R. Foster, Museum of \Testern Colorado, P.O. Box 20000, Grand Junction, Coiorado 81502. Peter M. Galton, College of Naturopathic Medicine, University of Bridgeport, Bridgeport, Connecticut 06601,-2449. T.rkehito Ikejiri, Department of Geosciences, Fort Hays State University, Hays, Kansas 67601..
':an Le Loeuff, Mus6e des Dinosaures, LL260 Esp6raza, France. -,rlrn S. Mclntosh, 278 Court St., Middletown, Connectrcut 06457. l. \L Mohabey, Geological Survey of India, Seminary Hills, Nagpur-440 006,India. 3.:lph E. Molnar, P.O. Box 158, Flagstaff, Arizona 86002.
J. Michael Parrish, Department of Biological Sciences Northern
Illi-
nois University, DeKalb, Illinois 60115. Bruce M. Rothschild, Arthritis Center of Northeast Ohio, 5500 Market Street, Youngstown, Ohio 44572. Leonardo Salgado, Museo Universidad Nacional del Comahue, Buenos Aires 1400 (8300) Neuqu6n, Argentina. Steven \X/. Saiisbury, Department ol Zoologv and Entomology, University of Queensland, Brisbane, Queensland , 4072, Australia. Allen Shaw, Section of Vertebrate Paleontologl', Carnegie Museum of Natural History, 4400 Forbes Avenue, Pittsburgh, Pennsylvania 15213. Kenneth Stadtman, Brigham Young University Earth Science Museum, P.O. Box 23300, Provo, Utah 84602.
Kent A. Stevens, Department of Computer and Information Science, University of Oregon, Eugene, Oregon 97403. Vrrginia Tidrvell, Department of Earth Sciences, Denver Museum
of Natural History, 2001 Colorado Blvd, Denver, Colorado 80205.
David L. Trexler, Timescale Adventures, P.O. Box 356, Choteau, Montana 59422. Jacques Van Heerden, Biological and Nursing Sciences, Universitl' of Fort Hare, Alice 5700, South Africa. D. Ray \7ilhite, Department of Biomedical Sciences, Louisiana State University School of Veterinary Medicine, Baton Rouge, Louisiana 70803.
Adam M. Yates, Bernard Price Institute of Paleaontological Research, Witwatersrand University, Johannesburg 2050, South Africa. ZhongZheng, Museum of Texas Tech Universitn Bor 43191,Lubbock, Texas 79409-3791.
x .
Contributors
Part One Sauropods Old and New
L. Postcranial Anatomy of Referred Specimens of the Sauropodomorph
Dinosasr Melanorosaurus from the Upper Triassic of South Africa Psrpn M. GerroN, Jecquns VeN HnenopN, AND Aoeu M. Yarps
Abstract Postcranial remains of two referred specimens of the sauropododomorph dinosaur Melanorosaurus readi Haughton 7924 ate described from the Lower Elliot Formation (Upper Triassic, Norian) of South Africa. The two specimens, found together, are about the same size, but one is slightly more massive than the other. The bones include vertebrae, two complete sacra, scapulae, a humerus, an ulna, and most of the peivis and hindlimb. One autapomorphy noted is the presence of four vertebrae in the sacrum with incorporation of a dorsosacral. Therefore the sacral count is DS1 + S1 + 52
+
CS. Depending on the cladistic analysis followed, Melano' rosaurus readi ts either a prosauropod with a few characters convergent to Sauropoda, or it is a sauropod with several prosauropod characters. We tentatively regard Melanorosaurus as Sauropodomorpha incertae sedis pending further analysis of the holotype and of all the referred speclmens.
Introduction Four valid taxa of sauropodomorph dinosaurs have been described from the Lower Elliot Formation (Upper Triassic, Norian) of South
Africa (Olsen and Galton 1984; Lucas and Hancox 2001), the fauna of which is discussed in several papers (Olsen and Galton 1984; Kitching and Raath 1984; Galton and Van Heerden 199g; Lucas and Hancox 2001). The most common remains have been re-
ferred to the prosauropod Ewskelosaurus browni Huxley, 1g66 (Van Heerden 1.979; Kitching and Raath 1984). Howeveq as pointed out by Yates (in press a), the holotype of E. browni rs a nomen dubium. The holotype of the next available species name for
rhis material , Plateosauraerzs Huene 1932 (for plateosaurus cullingworthi Haughton 7924), has a unique combination of prosauropod
and sauropod characters, as well as two autapomorphies (yates in prep.). It is therefore considered as sauropodomorpha incertae sedis (Yates in press a). Blikanasaurus cromptoni (Galton and Van Heerden 1985), which is represented by a very stocky partial hindlimb (Galton and Van Heerden 1998), was originally described as a prosauropod, but is now considered to be a sauropod (Galton and Upchurch in Upchurch et al. 2002; Galton and Upchurch in press; Upchurch et al. in press; Yates 2003a). The partial skeleton of Antetonitrus ingenipes (Yates and Kitching 2003) is also a sauropod. Based on a revised concepr of the Sauropoda (yates in press b), Yates (in press a) considers an associated partial vertebral series and
femur (BPI 114953) to be the only true prosauropod described to
date from the Lower Elliot Formation. However, the Melanorosauridae (and Pldteosaurauus as Euskelosaurus\ are considered to be prosauropods by other workers (Fig. 1.1A) (e.g., Galton and Upchurch in Upchurch et al. 2002; Upchurch et al. 20041Galton and Upchurch 2004; Sereno 1999; IJpchurch 1998; Wilson and Sereno 1998; Wilson 2002). The Melanorosauridae Huene 1929 is based on Melanorosaurus readi Hatghton 1,924 (for biography see Raath 7994). Van Heerden
(1979) regarded
M. readi as a junior synonym of the
sympatnc
"Euskelosaurus browni," which he transferred from the Melanorosauridae to the Plateosauridae Marsh 1895. However, Galton (1985) pointed out that the femur of Melanorosdurus resembres that of Riofasaurus from the Upper Tliassic of Argentina (Bonaparre 7972), a raxon accepted by Van Heerden (1979) as a non_ plateosaurid, because the femur is straight in anterior or posrerlor viervs r'vith the fourth rrochanrer close to or on the medial margin of the shaft. In"EuskelosAttrLts" (Huene 1906; Van Heerden 1979lGalton 1985), the femur is sigmoidal in these views, with the fourth trochanter well removed from the medial edge of the shaft as in other plateosaurids. Concerning the femur of Melanoroslurus, Van Heerden (1979) noted that the straightness of the femoral shaft may be due to distortion because the prorimal end, which lacks a proper
head,
is unlike that found in
Rioiasaurus and
is similar to
"Euskelosaurus." Galton (1935) noted that the femur does not ap-
2.
Peter
M. Galton,
Jacques Van Heerden, and Adam
M. yates
Ancslor
An@stor
B
Blikanasaurus Kotasaurus Vulcanodon
Satumalia Ihecodonfosaurus Efraasia
Banpasautus Shunosaurus
Rlolsaurus
Thecodontosaurus
P/al@saurus
Satunalia Ammosautus Anchisautus Riojasautus Melanorosaurus Camelotia VatIIetUrIc Lessemsaurus
l/t ssGpondy,us LtJfengosaws
Aroiisaurus
e
fulelaf,ffiawus
Jingshanosautus
Antelonitrus
Yunnanosaurus MassosPondylus Mussaurus Coloradisaurus "G." sinensis Lulengoseurus
/sanosaurus
Euskelosaurus Plateosaurus Serrosaurus
Kolasaum
Vulcil&M Shun6aurus Barapasautus
Omsaurus Ne6auropoda
Fig. 1.1. Two recent cladistic dndlyses of bdsal Saurctpodotnorpha. (A) One of the two most parsimonious trees found by a Heuristic anall'sis using PAUP 4.0 [Swofford (1c)98); see Galtott and Llpcburch (2001) for details]. The other mrtst parsimctnious tree is identical to tbat shott'n, except that Nlassospondvlus azd Yur.rnanosaurus haue sutapped positictrts. Tree statistics: Length = 279 steps; Cl = 0.5'11; RI = 0.63.5; RCI = 0.J55. From Galton and Llpchurcb (in press)' who prouide a full discussion of the 136 chdracters, nodes, and genera, plus the chardcter matrix. Tbe s!*napomorphies cited for the nodes are based ttn NM QRi.551, with mention if these are also present in SAM 3149 or 34i0. "G." sinensis = G)'posaurus sinensis of Young (1911 a, 1L)48), Sellosaurus = Efraasta, the genus to which most of the Pfdffenhofen specimens dre now referted, see Yates (2003b). (B) Simplified cladogram of fiue mr,tst-parsimonictus trees. Tree statistics:
Length = 149 steps; CI = 0.513,1; Rl = 0.7288 after ttnstable Blikanasaurus /:as been pruned. After Yates and Kitcbing (2003), a'ho prouide details on 212 characters, nodes, and genera plus character matrix. Abbreuiations: p = Prosauropoda; s = SauroPodomorphd: sa = Sduropoda.
pear distorted (see stereo photographs in Van Heerden 1977, pls. 6-8; Van Heerden 1979,p\s.64,65), and the degree of development of the head varies in different-sized femora of Riojasauras (Bonaparte 7972,
fig. 68). The tabular listing of the constituent genera of the Plateosauridae and Melanorosauridae given in Galton (1990, table 15.1) is in error. In the table, Plateosauridae includes only Ammosdurus, Musslurus, Plateosaurus, and Se/iosaurus. Three other genera, Co/oradosdurus, "Euskelosaurus," and Lufengosaurzs, which are discussed as plateosaurids in the text, were incorrectly tabulated within the Melanorosauridae and this has lead to some confusion. \Wellnhofer (1993) used the genera listed in Galton (1990, table 'Wellnhofer) for the 15.1; sometimes mistakenly cited as 1985 by Plateosauridae and Melanorosauridae. Benton (1'993, L994) rncotrectly cited "Euskelosaurus" as an earliest record and Lufengosdurus as the latest record for the Melanorosauridae. However, Postcranial Anatomy of Referred Specimens of Melanorosaurus
'
3
these three genera are included in the plateosauridae in Galton (7992, table 15.1), in which the Melanorosauridae is restricted to three genera from the Upper Triassic, namely, Camelotia borealis Galton 1985, Melanorosaurus readi Haughton 1.924; and Riojasaurus incertus Bonaparte 1969, plus an unnamed melanorosaurid from Argentina (Bonaparte 1986). Camelotid was found in the \Testbury Formation (Upper Triassic, Rhaetian) of England, and the described remains consisr of vertebrae, a femur, and phalanges with fragments of a pubis, an is-
chium,
a tibia, and a
metatarsal (Galton 1998). Details on
Melanorosaurus are given below. Riojasaurus comes from the Upper Los Colorados Formation (Upper Triassic, Norian; Bonaparte 1972) of Argentina. It is the best-known melanorosaurid, with remains of approximately rwenty individuals. There are some compiete skeletons and remains of different age groups that illustrate growth changes (Bonaparte 7972), and a complete skull and associated skeleton (Bonaparte and Pumares 1995). An unnamed melanorosaurid from the Late Triassic (Norian) of Argentina was briefly described by Bonaparte (1986) on the basis of two neural arches, those of a posterior cervical and an anterior dorsal that now comprise part of the holotype of Lessemsaurus sauropoides Bonaparte 7999. Unfortunately, the holotype only consists of the neural arches of three cervicals, fourteen dorsals, and fivo sacrals, together with the centra from three cervicais and numerous dorsals. These vertebrae are much more sauropod-like than those of other prosauropods, and, as noted by Yates and Kitching (2003), they bear a close resemblance to those of Antetonitrus. Van Heerden and Galton (1997)provided a preliminary description of a new specim en of MelanosAurus readi. The characers of this specimen were used in cladistic analyses of the Sauropodomorpha, which placed Melanorosauras in the Prosauropoda (Fig. 1.1A; Galton and Upchurch in Upchurch et al. 2002; Galton and Upchurch 2004) and in the Sauropoda (Fig. 1.1B; yates 2003a; yates and Kitching 2003). These referred bones of MeldnorosAurus are irlustrated in detail in this paper and their affinities are considered on the basis of the cladistic analyses cited above.
Institutional abbreuiations. BPl-Bernard price Institute for
Palaeontological Research, University of \fitwatersrand, Johannesburg, South Africa; NM-National Museum, Bloemfontein, Free
State, South Africa; and SAM-South African Museum. Caoe Town, South Africa. Systematic Paleontology Prosauropoda Melanorosaurus r eadi Hatghton 1924 Syntypes. SAM 3449 includes a right ilium (Haughton 7924, fig. 44), a left pubis, a left tibia (Haughton 1924, fig.46; Van Heerden 1979, fig.22, pls. 66, 67), aleft fibula, metatarsals, a right ulna (Haughton 1924: fig.43), and both radii. SAM 3450 consists of a
4.
Peter -\I. Galton, Jacques Van Heerden, and Adam
M.
yates
right femur (Haughton 1924, fig.45; Van Heerden 7979, fig.21. pls. 64, 65) and the proximal half of an eroded right humerus. Type locality. Base of the Elliot Formation (formerly known as the Red Beds), under the first sandstone ridge u'est of the dolerite dike on the north slope of the Thaba 'N1'ama (Black Mountain). Thaba is iocated betr'veen Josana's Hoek and Josana's Nek near Bensonvale, Herschel District, Transkei (formerly in eastern Cape Province), South Africa (SAM Archives). Comments. Van Heerden (1979) cited Haughton (1924) that the femur r,vas found some distance a\ ray from the rest of the skeletal material. Haughton (7924, 429, 433) noted that the bones "were lying isolated and embedded in a soft red mudstone below a sandstone band," "together with a femur partiy embedded in the overlying sandstone and the proximal half of a humerus found weathered down the slope" and that the femur "was in doubtful association with the other remains and may possibly belong in another form." This material was restudied by Van Heerden (1979) and most of it was referred to Euskelosaurtts browrtL r'vith the exception of one sacral, possibly the tibia, and the weathered femur. The femur was thought to be possibly distorted and therefore unsuitable as the holotype of the genus and species, thereby rendering Meldnorosatrrus readi a nomen dubium. However, M. readi Haughton 7924 must be a valid taxon because the femur is, in all likelihood, undistorted, and the lack of a proper head is probably the result of weathering. Many additional bones, mostly of Plateosdurduus, were catalogued with SAM 3449 and SAM 3450 since 1921, bm apart from these, the remaining bones probably represent one individual, the syntype. Except for unfigured vertebrae, it would appear from the text that the right ulna and both radii form part of SAM 3449 (Haughton 1924). The bones of SAM 3450 were found separately twenty yards (18.3 m) east of SAM 3449 (SAM archives), with the right humerus downslope of the femur.
Referred Specimens SAM 3532-Referred Specimen of Haughton (1921)
Mdterial. Unfigured vertebrae, an almost complete left scapula, a right humerus (Haughton 1924, fig. 42), a complete left ilium, and a metatarsal lll (Haughton 1924,Fi1.47). Locality. About one-third up in the Lower Elliot Formation, at a higher horizon than the holotype. From below the Rooi Nek, between Kromme Spruit and Ma;uba Nek, Hershel. Cornments. A somelvhat smaller specimen than the holotype, it was referred to MelanorosaurLts readi bv Haughton (1924). The left scapula and right humerus were incorrectly referred to the holotype b,v Van Heerden (7979). They were briefly described but not figured, as was metatarsal III. The taronomic status of this specimen, which includes other prosauropod material given the Postcranial Anatomy of Referred Specimens of Melanorttsaurus
'
5
same catalog number srnce 1924, is still to be determined. therefore not be referred to further.
It will
NM QR3314-Specimen of 'Welman (1998, 1999) Material. Most of an articulated skeleton (see photo as preserved in MacRae 1999,202), including a complete skull (see photos in \felman 7999, fi1. a). Locality. Lower part of the Elliot Formation, near Ladybrand, Free State Province.
Comments. Of the skull, only the ventrai aspecr of the basicranium has been described (in Welman 1.999, as a basal prosauropod;
in \Welman 1998, as Euskelosaurers). Although a juvenile individual, this is the most complete specimen of Melanorosaurus discovered to date. A full description of the material will be given elsewhere (Yates, Van Heerden, and Galton in prep.)
NM QR1-t51-Referred Specimen of Van Heerden and Galton (1997)
Mdterial. A cervical verrebra, several dorsal and caudal vertebrae, four associated sacrals, and various girdle and limb bones belonging to two individuals of approximately the same size (but one slightly more robust).
Locdlity. Base of the Elliot Formation, Miiner Farm, 'Wodehouse (Dordrecht) District, Free State Province. Comments. The original excavation rn 1967 included a weathered femur (since lost). The rest of the material was ercavared rn 7971 from the banks of a narrow furrow and over a rectangular area of approximately 6 m. NM QR1551 lacks cranial remains, but there is a fair representation of the different elements of the postcranial skeleton. Most of the material is well preserved. bur some elements exhibit signs of transportation over a short distance prior to fossiiization. This material formed part of the thesis of Van Heerden 11977). A preliminary description was given by Van Heerden and Galton (1997) and is elaborated upon below. Description of NM QR1551 with Brief Comparisons To supplement the description of the bones of NM eR1551, brief comparisons are made with several other sauropodomorphs and basal sauropod genera. The genera, and the references involved
(unless indicated to the contrary), are: Anchisaurus and Ammoslurus (Lower Jurassic, Connecticut Valley, United States; Gal, ton 1976), Antetonitrus (Upper Triassic, South Africa; yates and Kitching 2003), Barapasaurus (Sauropoda, Lower Jurassic, Indra; Jain et al. 1975, 7979), Blikanasaurzs (Sauropoda, Upper Triassrc, South Africa; Galton and Van Heerden 1998), Camelotia (Upper Triassic, England; Galton 1998), "Gyposaurus" capensis (Lower Jurassic, China; Young L941,a, 1948), Kotasaurus (Sauropoda, Lower Jurassic, India; Yadagiri 2001), Isanosaurus (Sauropoda, Upper Triassic, Thailand; Buffetaut et al. 2000, 2002\, Lessem-
6.
Peter
M. Galton,
.|acques Van Heerden. and
Adam M. Yates
saurus (Upper Triassic, Argentina; Bonaparte 1999), Lttfengosaurus (Lower Jurassic, China; Young 1941.b, 1947 , 19 57), Massospondylus (Lower Jurassic, South Africa and Zimbabwe; Cooper 1981), Plateosaurus (Upper Triassic, Germany; Huene 1926,1932;
Galton 1990, 7992, 2001a), Plateosaurauus ("Euskeloslurus," Upper Triassic, South Africa; Van Heerden 1979), Rioiasaurtts (Upper Triassic, Argentina; Bonaparte 1972), Shunosaurus (Sauropoda, Middle Jurassic, China; Zhang 1988), and Vulcanodon (Sauropoda, Lower Jurassic, Zimbabwe; Raath 1972: Cooper 7984). Two proposed relationships for most of these taxa are shown in the cladograms (Fig. 1.1), and the systematic position of Melanorosdurus is discussed below In the section on the anatomy of prosauropods in Galton (7990,7992) and Galton and Upchurch (2004), the condition for Melanoroscturus is based on NM QR1551 (Fig. 1.1A), as it is in Yates and Kitching (2003)' in which this genus is considered to be a sauropod (Fig. 1.18). Vertebrae
Ceruical uertebrae. The centrum of cervical 6 or 7 (Figs. 1.2B, 1.3A-D) is short, broad, and high in comparison to that of cervical 6 of Rioiasaurus and Plateosdunrs (Fig. 1.2A, C, D), with the length:height ratio at about 2.0 rather than more than 3.0 (measurements given in Table 1.1). The diapophysis is better deveioped, but it is close to the parapophysis and is relatively low. The articular ends are not well preserved, but the centrum was probably amphicoelous or opisthocoelous. The articular ends are inclined at 60' (anteriorly) and 80' (posteriorly) to the long axis of the centrum' indicating a distinct anteroventral curvature in the neck. The ventral margin of the centrum is concave in lateral view. The parapophyseal facet is not preserved but it was probably close to the ventral margin. Above this ridge, the lateral side of the centrum is very distinctly concave. The neural arch is fairly low. The neural spine has a narro% short base and projects slightly anteriorly. The anterior half of the dorsal tip of the neural spine is transversely expanded. The zvgapophyses have flat, long, and broad facets inclined at 30-35'to the horizontal. The terminology of \(iilson (1999) for the laminae of sauropod vertebrae is followed for the laminae (or the equivalent ridges) associated with the processes of the vertebrae of NM QR1551. Each parapophysis has a sharp, low ridge extending a short distance posteriorly, but this is the base of the parapophysis. The posterior centroparapophyseai lamina would be more extensive and, although thought to be restricted to dorsal vertebrae in sauropods, it is present in cervicals of Sauroposeiden (Sauropoda, Lower Cretaceous, 'Wedel et al. 2000). The lateral edge of the prczyOklahoma; gapophysis is prolonged posteriorly to the base of the neural spine as a prominent ridge (spinoprezygapophyseal lamina), whereas the posterolateral edge of the postzygapophysis forms a more promir-rent edge (spinopostzygapophyseal lamina) that continues posterodorsally to the broken top of the neural spine. The diapophysis Postcranial Anatomy of Referred Specimens oi Melanorosdurus
'
7
Fig. 1.2. (A) Skull and ceruical ,,/ Riojr,.ruru' ineerru. irz lelt Ltteral t rcu (reucrscdt lrom Bonaparte and Pmnares (191),5). (B) Left lateral uiea, of sixth or
-.6y'i,5
scuctrth ccn'ical t crtclrrs
uf
N{elanorosaurus readi, NM QR 1{51 C-D: Lcfr ldternl uictt of cerui cal s o/ Plateosaurus, /rorz Huerte (19.)2): (C) sixth and (D) eighth. E-F: Leli lateral uietu of t h e tl orsals o/ Melanorosaurus readi, NM QRl.i.i7: (F,) neural arch ottl,- of the eighth 1?); (F) incomplete ninth (l). ()-H: Left Ittteral uietu of the dorsals of Camelotia borealis, reuersed from Seelet (1898): (G) anterior dorsdl
F-l
(?
fourth) and (H) the fourteenth.
I-J: Left lateral uieu of the dorsals o/ Plateosaurus , frotn Huene (1932): (l) eighth and (J) tenth. K-M: Left lateral uieu' of dorsals crf Riojasaurus,
from Bctnaparte
(1972): (K) fifth, (1.) ninth, and
(M) ta,elfth. Scale is tlpploximateb' .i0 mm.
is small compared to those of the dorsal vertebrae. Its base lies just
above the floor of the neural canal and it projects anteroventrafiy towards the parapophysis. There are rwo lamina supporting the diapophysis ventrally. The posterior centrodiapophyseal lamina, which is inclined slightly ventrallv. is much larger that the anterior centrodiapophyseal lamina. Dorsal uertebrae. Remains of four dorsal vertebrae are preserved (Figs. 1.2E, F; 1.3E-I; 1.4A) (measurements given in Table 1.1). A neurai arch possibly represents dorsal 8 (Figs. 1.2E; l.-iE,
8.
Peter N{. Galton, Jacques Van Heerder.r. and Adam
M.
yates
,ffi B
Flg. -1.3. Melanorosaurus readi, NM QR1.t51. A-D: Ceruical uertebra 6 or 7 in (A) left lateral, @) dorsal, (C) anterior, dnd (D) posteriol uiews. E-F: Neural arch ctf sixth or seuenth dorsal uertebra in (E) right lateral and (F) dorsdl uieu's. G-H: Neural arc.b of a posterior dorsal uertebra in (G) dntetior and (H) dorsal uietus. I-J: Centrum of dorsal uertebra in (I) lateral and (J) dorsal uiews. Scale is 50 mm.
H
w TABLE 1.1. Measurements (in mm) of Three Presacral Vertebrae (NM QR1551)
Descriotion Centrrrm
lenoth
6th?
cervical 9th? dorsal 11th? dorsal
+ 118
124
105
Centrum height anteriorly
49
96
r02
Centrum width anteriorly
55
72
+75
Centrum height posteriorly Centrum width posteriorly
61
68
+92 +71
-\{aximum length of vertebra
164
?145*
Maximun-r height of vertebra
||
l
?245"
102 L L
1/ /a
These figures based on a combination of the supposed 8th and 9th dorsals.
Postcranial Anatomy of Referred Specimens ol Melanorosaurus
'
9
'K K qw_
+ -=t €-
.r Le MA N df gl
v'-'r
ur o& *&t af -,
.-=- I g *
rc
jryl
,3f-
-,:sr
-
-'
Fig. 1.4. Melanorosaurus readi, NM QR75J1. (A) Ninth(!) dorsal uertebra in right lateral uietu.
B-D: First sacral uertebra in (B) left lateral, (C) dorsal, and (D) dnterior uieus. E-H: Second and third sacral uertebrae in (E) dnterior, (F) Ieft lateral, (G) right lateral, and (H) dorsal uiews. l-J: Fourth sacral uertebra in (I) dorsal and (J) anterior uiews. K-L: Anterior caudal uertebru in (K) right Iateral and 1Lt anterio, uiews. M-N: Middle caudal uertebra in (M) anterior and (N) Ieft lateral uieus. O-P: Middle caudal uerlebra in tO) posterior and (P) rigbt lateral uiews. Q-R: Middle caudal uertebra in (Q) anterior and (R) Ieft lateral wews. Scale is 50 mm.
10 .
Peter
M. Galton,
F). The neural spine is fairly high, but the ratio of the height to length of the transversely narrow base is less than 1.5. It thickens a little transversely at the dorsal edge and becomes thinner passing posteriorly. The rounded dorsal tip may be the result of weathering. The lower half of the anterior margin of the spine has a deep sulcus, whereas the upper half bears an indistinct keel. The prezygapophysis has a rather long, narrow facet facing dorsomedially at about 30" to the horizonral. Posterodorsally it forms a spinoprezygapophyseal lamina that extends to the base and onto the lower third of the neural spine. There is a distinct hypantrum between the
two prezygapophyses, which is as long axially as the prezygapophyseal facets (Fig. 1.3F). The postzygapophysis also has a long, narrow facet, which is inclined ventrolaterally at about 30" to the horizontal. The dorsoventrai height of the hyposphene equals that of the neural canal, as in sauropods, rather than being much lower, as in prosauropods (Yates and Kitching 2003). Dorsally, a spinopostzygapophyseal lamina ertends to the lorver part of the neural spine. The diapophysis is short, robust, and situated above the level of the neural canal; it projects slightly posterodorsally and is subcircular in axial section (Figs. 1.2E; 1.3E). It is connected ro the zygapophysis by the prezygodiapophyseal, the most prominent lam-
Jacques Van Heer.den, and Adam
M.
Yates
ina, and the postzygapophyseal lamina, rn'hereas ventrally there are posterior and anterior centrodiapophyseal laminae; the latter is the weakest lamina and terminates at the parapophysis. The most complete vertebra probably represents dorsal 9 (Figs. 1.2F; 1.4A). The centrum is distinctly amphicoelous, with no appreciable difference between the anterior and posterior surfaces. As in other prosauropods, the length:height ratio is less than 1'5' The sides are concave, especially in the upper half of the centrum. The ventral margin is concave in lateral view and the edge is rounded in cross-section. The neural arch is high. The base of the neural spine extends beyond the posterior articular surface of the centrum. What remains of the incomplete prezygapophyses are similar to those of dorsal 8. The postzygapophysis has a long, narrow facet inclined at about 40' to the horizontal. It has the same laminae as the preceding vertebra, and a hyposphene situated ventrally' the dorsoventral height of which equals that of the neural canal. The diapophysis has an elliptical base, the three lamina associated with it are slightly better developed, and it is closer to the parapophysis than in cervical 8. A centrum (Fig. 1.3I, J), possibly from dorsal 11, is similar to cervical 9 ercept that it is slightly longer. An incomplete neural arch (Fig. 1.3G, H) is probably from one of the last dorsals and, as in other prosauropods' there is no prezygapophyseal lamina. The base of the neural spine has the same axial length as dorsal 8, but it is broader posteriorly than anteriorll'. The postzygapophyseal lamina is somewhat stronger on this arch than it is in the more anterior dorsals. Sacrum. All the bones in the fossil pocket were disarticulated except for the four sacral vertebrae (Figs. 1.4B-J; 1.5A; 1'6A-D). whose articulated state indicates that they were originally united together as a functional unit. The four sacrals are a little shorter than the ilium, and the sacral ribs fit the rugose medial surface of the ilium (Fig. 1.10B) very well. Comparisons with other prosauropods show that the sacrum has the reptilian sacral vertebrae 1 and 2 and a caudosacral, the plesiomorphic form for the sacrum in prosauropods, and that a dorsosacral is incorporated anteriorly (Fig. 1.6A; see discussion in Galton 7999,2007b). The dorsosacral (Figs. 1.4B-D; 1.6A, B) lacks the neural spine and postzygapophyses, but the rest of this element is well preserved. The centrum is amphicoelous and has concave sides. The ventral margin is concave in lateral view and is less rounded in cross section than is the case in the middle dorsals. The facet of the prezygapophysis faces dorsomedially and is inclined at approximately 40' to the horizontal. It is almost as broad as it is long, it is slightly concave transversely, it has a robust base, and no spinal lamina can be distinguished. The prezygapophyses project a little beyond the centrum anteriorly. The postzygapophyses are broken off, but there is a robust, wedge-shaped hyposphene. The diapophr-sis is fused to its sacral rib. The suture is still visible on the lefthand side in anterior view (Fig. 1.4D). The dorsal surface of the diPostcranial Anatomy of Referred Specimens of Melanorosaurus
'
11
=F*
eE Fig. 1.5. Ivlelanorosaurus readi,
NM QRIiil. (A) Fourth sacral
uertebra in left lateral uieu. B-D: First caudal uertebra in (B) left lateral, (C) right lateral, and (D) posterior uieu,s. E-F: Third(?) caudal uertebra in (E) right lateral dnd (F) posterior uiews. G-I: Middle caudal uertebra in (G) antelior, (H) left lateral, and (I) dorsal uietus. U) Centrum of middle caudal uertebra in rigbt ldterdl t,iew. K-O: Progressiuely more posterior caudal uertebrae in leit lateral fK, I-, N, O) and right lsteral (trl) uiews. Scale is 50 mnt.
ww Ww -M G*
M
ffi"M apophysis-sacral rib complex is flat and horizontal, being extended transversely at its proximal end by the pre- and postzygodiapophy_
seal laminae. The anterior centrodiapophyseal lamina
fo.-,
" strong ventral buttress, extending onto the anterodorsal quarter of the side of the centrum. A low, rounded posteroventral ridge below the posteriorly projecting plate represents a remnanr of the posterior centrodiapophyseal lamina, which is much more prominenr in the preceding dorsals. Laterally the sacral rib has a short anrerior projection that lies in the horizontal plane; and a longer exrension that lies in the vertical plane. The latter is fused with the anterior extension of the second sacral rib. A deep, wide sulcus is presenr between the first sacral rib and the horizontal olate. The centra and sacral ribs of the second thi.d vertebrae of "trd
12 o
Peter
M. Galton,
Jacques Van Heerden, and Adam
M. yates
,i" ' ...-:,1 ;,,:!.,-,rlr, 'i !.'-:-, . i:-
| . \ rr',--!...,...-
;.' _._-l
' .il'_....-'-
_1.,:1,-
" -\l
/
Fig. 1.6. A-D: Left lateral uiew the sacral uertebrae of N{elanorosaurus readi, NM
of
QRl5-51: (A) outline drawing shoLt,ing the fottr sacral uertebrae in articulation; (B) /irst sacral; (C)
ankylosed second and third sacrals uith the sacral rib of the
tI'B
"/
{1L!t4"'-.j? --F
first fused to tbat of the second; dnd (D) fottrth sacral. (E) Left laterdl uieu of the sdcral uertebrae o/ Riojasaurus, from Bondpdrte (11)72); third sacral should be at
I
front
it
as a dorsosacral (Nctuas
1996). F-H: Left lateral uiew of
caudal uertebrae of Melanorosautus readi
IE
F -n\ t)
\nf
lNM
QR155/): (F) the first, (G) the
ltresumed fourth, and (H) a distal caudal uertebra. I and J: Left lateral uiew of caudal uertebrde of Camelotia borealis, from Huene
(1907-08): (I) second and 0) sixth. (K) Left lateral uiew of first
and second caudal uertebrae of Plateosaurus, from Huene (1932). Scale is approximately .50 mm.
the sacrum are fused together (Figs. 1.4E-H; 1.6A) and represent the reptilian sacrals 7 and 2. The centra have the same general shape as the dorsosacral, but that of sacral 2 is slightly longer. The neural spines are high, with transversely thickened tips and short, broad bases. The anterior surface of the neural spines was probably keeled, whereas posteriorlv the ventral two-thirds shows a narrow groove. The spines are directed slightly posteriorly. The prezygapophysis has a short, broad facet facing dorsoventrally and inclined at about 55" to the horizontal. The sacral rib is fused with the diapophysis. In sacral vertebra 1, the diapophyseal part of the complex has a flat dorsal surface in the horizontal plane. Almost all of this is directly supported by the very strong anterior centrodiapophyseal lamina, which extends slightiy farther ventrally than in rhe dorsosacral. The diapophysis proiects beyond the posterior cenPostcranial Anatomy of Referred Specimens of Melanorosaurus
.
1'3
trodiapophyseal lamina
to form a slight
ridge
in the horizontal
plane. There is a strong constriction between the diapophysis and the sacral rib. Just lateral to the diapophysis, the sacral rib divides into a dorsal and a ventral part. The dorsal part curves around to the anterior side and forms a thin plate, which is again yoined to the anterior extension of the ventral part of the rib. Below the dorsal extension, and behind the anterior plate, there is a deep, concave excavation. The ventral part of the sacral rib is very robust. It is expanded dorsally and ventrally and oriented anterodorsally and posteroventrally. The anterodorsal part is confluent rvith the anterior extension of the dorsal part of the sacral rib, whereas the posterior part ioins the anterior extension of the rib of sacral vertebra 2.The diapophysis-sacral rib complex of sacral vertebra 2 lacks the anterior extension of the dorsal part, and there is no lateral connection between the dorsal and ventral parts of the sacral rib. The anterior centrodiapophyseal lamina extends downward to just below the middle of the centrum height. The caudosacral (Figs. 1.4I, J;1.5A; 1.6.4, D) is, in general, similar to sacral 2. The centrum is a little shorter and the neural spine is situated more anteriorly. On each side the diapophysrs forms a shelf anteriorly and slopes posteroventrally. The postzygapophyseal facet is inclined at approximately 65" to the horizontal. Anteriorly it is confluent with the base of the neural spine (as in sacrals 7 and2), and between the paired facets there is a wide sulcus. There appears to have been no hyposphene. The pre- and postzygodiapophyseal laminae are small. The sacral rib has an elliptical end surface, with the long axis inclined obliquely ventralll., the anterior end being lower. There is only a small dorsolateral extension of the sacral rib. Berween this and the large lower part, there is a distinct sulcus on the anterior side. The complex has a single medial and ventral anterior centrodiapophyseal lamina that extends to just below the middle of the centrum height. The anterior surface of this lamina is concave. Caudal uertebrae. There are twenty-two caudal vertebrae (Figs. 1.4K-R; 1.5B-O; 1.6F-H). The proximai caudals (Figs. 1.5B-F; 1.6F) have a high, amphicoelous cenrrum that anteroposteriorly is extremely short with concave sides. The ventral surface is transversely broadened with demifacets for chevrons. The neural spine is tall, being almost as high as those of the sacrals, and it is inclined slightly posteriorly. The postzygapophysis is situated on the base of the neural spine. The facet is slightly longer than broad, and it is inclined between 55o and 60' to the horizontal. Each transverse process is incomplete, but they were probably fairly short; the base is not deep in posterior view (Fig. 1.5D, F) as in sauropods. Each lacks distinct ventral burtresses, but the base is nearly as long as the centrum. The immediately succeeding caudals (Figs. 1.4K-N; 1.6G)have comparatively longer, lower centra. The zygapophyses are smail and their facets are inclined at 50-60'to the horizontal. The neural spine is tall, directed slightly posteriorly, with the base situated on 1.4
.
Peter
M. Galton,
Jacques Van Heerden, and Adam
M.
Yates
the posterior half of the centrum. The spine is eiliptical in cross sec-
tion and the posterior surface is keeled. The transverse processes have a fairly long base, extending almost the whole length of the centrum. The transverse process is directed obliquely posteriorly in the horizontal plane. The mid-caudals (Figs. 1.4O-R; 1.5G-J) are approximately the same length as the anterior ones, but the centrum and neural arch become progressively lorver posteriorly along the series. The neural spine and transverse processes are placed posteriorly. The spine is directed more posteriorly, lvhereas the transverse processes are in-
clined slightly posteriorly and dorsally. The distai caudals (Figs. 1.5L-O; 1.6H) lack a neural spine, and the transverse process is represented only by a slight ridge on the lateral surface of the centrum. The transverse width of the ventral surface, which is slightly concave transversely, almost equals the centrum height. The terminal caudals are very small. Pectoral Girdle and Forelimb
The combined length of the humerus and ulna is 63"h of the length of the femur plus tibia, compared to 70"/' in Rioiasdurus. The relatively shorter forelimb, somewhat shortened trunk (through the inclusion of the last dorsal in the sacrum), and the concurrent strengthening of the sacrum may indicate that Melanorosaurts was more facultatively bipedal than Rioiasdurus. Scapula. The pectoral girdle is represented only by the scapulae: the right is almost complete (Figs. 1.7A, B; 1.8A-C) and the left is incomplete (Fig. 1.11D) (measurements given in Table 1.2). These two are very similar and are probably from the same individual.
The scapula is a long, broad element that would have been held more or less horizontallv. The posterior end measures about 90 mm across (as preserved) whereas the anterior (proximal) end measures at least 190 mm in width (185 mm preserved, restored width 210
mm, Fig. 1.8A). The major portion of the bone is flat, blade-like, and elliptical in cross section. The long edges of the blade are subparallel, being slightiy concave superiorly and straight inferiorly. The glenoid region is transversely thickened. The scapular fossa, adjacent to the glenoid fossa, is large and distinctly separated from the blade, especially in the acromial region. Humerus. There is one weathered right humerus (Figs. 1.7E; 1.8D-F) and the distal ends of a left and a right (not illustrated) (measurements given in Table 1.2). The humerus is 71."h of the length of the femur and its proximal and distal expansions lie in the same plane. The humerus is slightly sigmoidal in lateral view. The head lies just lateral to the midline of the bone. The proximal half is triangular with a concave anterior surface and a convex posterior surface (Fig. 1.8D, F). The deltopectoral crest is broken off, but its base indicates that it extended to the distal half of the bone as in Riojasaurus (Fig. 1.7F); in Anchisdurus, Plateosaurduus, and Plateosaurus (Fig. 1.7G) the deltopectoral crest is limited to the proxiPostcranial Anatomy of Referred Specimens ol Melanorosaurus
'
15
ffi W{
Fig. 1.7. A-B: Left scdpuld euersed) o/ Melanorosaurus
(r
readi, NM QR1J51, in (A) uentral (pctsterior) and (B) Iateral uielus. C and H: Left scapula and coracoid o/ Riojasaurus, from B ondp drte (1972), in (C) Iateral and (H) uentral (posterior) uiews. (D) Left scapula and coracoid of Plateosaurus, reuersed from Gabon (1990). E-G: Antenor uieus of right humerus of (E) Melanorosaurus readi lNM QR1551); /F/ Riojasaurus, reuersed from Bonaparte (1972);
\t /
G I
and (G ) Plateosaurus, /roz Galton (1990). I-J: Left ulna r:f
pl
o
x
i7 n
ffi $/ vt
N{elanorosaurus readi, NM QR1.t51 in (I) lateral and (J) medial uiews. K-L: Left ulna of Rio jasaurus, from B ondp drte (1972), in (K) lateral and (L) medial uieuts. (M) Lateral uietu of the left ulna ofPlateosaurus, from Galton (1990). Scale d[)
ItH h/
E [ lr 11 &\ It\
L%/
dt elt- 5 0 m m.
mal half. In most sauropods the deltopectoral crest is reduced to a low crest or ridge, but it is large in Antetonitrus. The shaft is short and subcircular in cross section proximally. More distally it is elliptical in section, with the long axis placed transversely. The distal end is slightly erpanded, with a flar posterior surface; the anrerior surface has a vague fossa that is not as rounded or as nicely sharpedged as inPlateosaurus (Fig. 1.7G) nor is it as flat as in sauropods. The articular surfaces at both ends are weathered. The humerus rs more massive than that of Plateosaurus (Fig. 1.7G) and, although eroded and incomplete, it is much less massive than that of Rio-
jasaurus (Fig. 1.7F).
16
.
Peter NI. Galton, Jacques Van Heerden, and Adam
M.
Yates
ffi
GW :""1)l'9
% Iig.
1.8. Nlelanorosaurus readi, N,\,I QR l ii L A-C: Right scaPula in (A) lateral, (B) superior tJnlcriot'). and tCt ntediol uitws. D-F: Right humerus in (D) anterior, (E) lateral, and (l) posterior uieuts. G-l: Lelt ulna it (G) dnterior, (H) lateral, and (l) p o sterctlat eral uiet,s. J -1, : Right metatarsal I in ([) laterdl, (K) rnedial, and (L) anterior uiews. Scole is 5() nrm.
Postcranial Anatomy of Referred Specimens of Melanorositurus
'
1'7
TABLE 1.2. Measurements ( in mm)
Element
?*Tii;trtllT: rl*.'
Total length
of Metanoros aurus
Proximal width
Distal width
Scapula
489
?
?
Humerus
450
155
r47
Ulna
269
Pubis
499
178
134
485
176
126
476
?
\t4
Ilium
476
Ischiurn
140
Femur
638
140 135
?
Tibia
640 / oo
217
143
Fibula
185
509
t76
120
485
190
130
481
100
)
Forearm. No recognizable part of a radius or manus rs preserved, but there is a fairly complete left ulna (Figs. 1.7I, J; 1.8G-I). In general it is similar to those of Plateosaurauus and Plateosaurus
(Fig. 1.7M), but it is less robust than in the former. The proximal end appears to be subtrianguiar in outline rather than triradiate, with a deep radial fossa as in Antetonitrus and sauropods; the lateral process is incomplete, so it cannot be determined if it was Ionger than the anterior process. The olecranon process is prominent, as in prosauropods and Antetonitrus, rather than reduced or absent as in sauropods. The incomplete distal end has a longitudinal ridge on the medial side of the posterior surface. Peluic Girdle and Hindlimb
Ilium.There is a well-preserved right ilium (Figs. 1.9A; 1.10A, B) that is not as robust as rhar of Riojasawras (Fig. 1.9B, C) (measurements given in Table 1.2). The dorsal margin is stepped, but not as distinctly as in Rioiasaurus, and the anterior process of the iliac blade is short and pointed, whereas the posterior process ls both deeper and much longer as in most prosauropods. In Anchisaurus and Ammosaurus the anterior process is long, and it makes an acute angle with the pubic process. The ischial process has about the same vertical extent as the pubic process, so it is
not reduced as in the
sauropods Bardpdsaurus, Kotasaurus,
Shunosaurus, and Vulcdnodon. The acetabulum is not backed by bone and the overail form is more similar to that of Plateosaurws than to that of Riojasaurus (Fig. 1.9B-D). The anterior part of the
18
.
Peter
M. Galton,
Jacques Van Heerden, and Adam
M.
yates
lig. 1.9. A-B: Lateral yiew of right ilium o/ /A) Melanorosaurus readi, N-&{ QR1551, and (B) Riojasaurus, reuersed front
OGR^,
,F$'. U'\
Bonaparte (1972). C-D: Reconstrttctiun oI right peluis in lateral uiew o/ /C/ Riojasaurus, from Bonaparte (1972), and (D) Scdle
fr;'a
''',{'t'''
',j
,iff;r*1,,,i,,i,
'
'4
,-\6&
l1!"';l ,,;'::': ''
:1.
" .'::a
*,,$'
'*"":
I
'*J
q
r"'
G dlton ( 1 9 9 0 ). dpploximately 50 mm.
Plateosaurus, frctrn
",@
H
*'Jl&
1' ,l'i r"l'l'llii
#',;
.,$
ffi
W: "*,,. -4 1ll'
'".t4
$
D,'$ ,E
i I *_
Ir *.
1
supra-acetabular ridge is better developed than the posterior part, Lr-rdicating that in life the dorsal margin of the ilium was orientated anterodorsally. Pubis. There are three pubes, one left and two right. One of the
right pubes is almost complete and virtually undistorted (Fig. 1.10C-E) (measurements given in Table 1.2). Laterally the pubic plate has a strong ridge in the proximal half of its length. Between ihis ridge and the obturator foramen, there is a distinct depression ,rnd the bone is very thin. The obturator foramen is large, as in ,rther prosauropods such as Plateosaurauus and Plateosaurus (Ftg. 1.9D), rather than small, as in Riojasaurzs (Fig. 1.9C). The proxi-
Fig. 1,10. tr{elanorosaurus readi,
NM QRl.t.tl. A-B: Right ilium in \At ldtcrcl ,tnd
1Bt
tnedial uiews.
C-E: Right pubis in (C) posterouentral and slightly lateral), (D) medial, and (E) dorsdl (and slightly medial) uieuts. F-G: t
More massiue left tibia in (F) lateral and (G) posterior uiea's. H-I: Incomplete left tibia in (H) medidl and (I) dnterior uietus. Scale is .50 mtn.
Postcranial Anatomy of Referred Specimens of Melanorosaurus
'
19
mal part is turned through 50' with regard to the distal part. The distal half of the pubic plate is triangular in cross section, the lateral edge forming the narrorv base of the triangle. The distal end is expanded anteroposteriorh, rl'ith its greatest thickness near the lateral margin. In this wa)'a posterioriy directed "foot" is formed, but this is much less prominent than in theropod dinosaurs. In anterior view the Iateral margin is slightly sigmoidal, not stepped as in Riojasattrus, whereas the medial margin (a1ong the pubrc symph,vsis) is straight.
Ischia (Measurements given in Table 1.2). The united distal ends of two ischia are preserved (Fig. 1.11A-C). The shaft of each is trianguiar in cross section. There is slight transverse and strong
anteroposterior expansion on the distal ends, but not as much
as
rn Pldteosdurauus, and the distal ends are not blade-like as in sauropods.
Femur. One of the elements found originall,v in 1967 was a weathered femur that was iost. In addition, NM QRl551 includes three other femora, one of r.vhich is virtually complete and undistorted. As far as can be ascertained, rhe femora are about the same length, but one pair is more massive (Fig. 1.11E-H; cf. Fig.
H
We B
Fig. 1.1 1. \lelanorosaurus readi, QR l -.-il. .{-C; Dist:rl part of iscl:l,t tri .\ .ilttertor ,L'entrJl), (B) leit lrter.tl. r11tl C l)()sterior ldctrs,tl: t tttt s. rDt Lelt scapula in
NII
ltter;l
t ietL.
I I
E-H: \Iore
rndssu,e
l,;t.e_H,lt.tnora in rE. G:nte.lr,rl ttt,l lF, H) lateral ti"-1..t
.t,t.l
L'ieu s. St,rle ts 5A
20 .
nnt.
Peter N{. Galton, Jacques Van Heerden, and Adam
M.
Yates
' ,lt
... ..,t:aalt ,,
.i*ei ':'t: ?-
:iE *i
a; ,1
* :s.
#"
ffi F
?
{ ,,i
:fc
i]]:lrr,:rl
l'"i):'
ffi ',1
:'
;'$$9
,i.
$.
iriiilrli
H&
ffi
'T
$ x
],,.tllt
1t.
tl,
;:.
ffi
w,
@
ir* .,t i: u: ,i
s ,$'
i
w
,d-
.
'*
*
l*i
w
$
s
&
€
$
1.12A-D) (measurements given in Table 1.2). The best-preserved femur, 638 mm long, is straight in anterior view and slightly sigmoidal in lateral view (Figs. 1.12A-D; 1.138, H).In these respects it resembles those of other members of the Anchisauria (Fig. 1.1A; -\ttchisaurus, Ammosaurus, Camelotia, and Riojasaurus), as well as Lttfengosdurus, "Gyposaurtts" capensls, and the sauropod Ante-
tonitrus; the femur is straight in lateral view in other sauropods, such as BarapdsdurLts, Kotasaurus, Isanosaurus, Shunosaurus, and \-tilcanodon. The head projects perpendicular to the shaft and posteriorly a shallow sulcus separates it from the rest of the proximal end. The medially directed head does not project from the center of lhe proximal part of the shaft but is shifted anteriorly. In these three respects it resembles the femora of SAM 3450, Riojasaurus and Cunelotia (Figs. 1.13A., G; 1.13D, J, E, K). Lateral to the head, on ihe posterior surface, there is a rounded ridge. The upper part of the :rochanter major is damaged. The anterior (lesser) trochanter is well developed, sheet-like, nore or less parallel to the lateral margin, and partly visible in pos:e rior view, as in Riojasaurus, Camelotia, and Antetonitrus,' it is not -i low ridge as in other prosauropods and sauropods. It is situated some 110 mm below the prorimal end with a quite abrupt proximal :e rmination, but distally it becomes gradually lower and disappears.
Fig. 1.12. Melanorosaurus readi,
NM QR1551. A-D: Left femur tn (A) lateral, (B) anterior, (C) mLdial. and (Dt posterior uteu's. E-H: Right libula in (E) lateral, (F) medial, (G) anterior, and (H) posterior uiews. I-L: Right tibia in (l) Lateral, (J) anterior, (K) medial,
and (L) posterior uieu's. Scale
.50
tntn.
Postcranial Anatomy of Referred Specimens of Melanorosaurus
.
21'
c
D
/-. \ '/ tl{ tl
tl ll
\ .,,
I
\,,,.1I
-
Iqli
t\J\.::
rc
ff\ f*P 5P
{/ \
\il ( ir
Fig. 1.13. Left femora
\
bii,
of
melanorosaurids (A-E, G-K) and Plateosaurus (F and L) in Poslerior tA-lt and medial tC-Lt uiews; (A, G) holotype of tr{elanorosaurus readi, SAM 3450; (8, H) lt{. readi, NM QR/;i/: iC. /l M. thabarren'is. from Gauffre (1993); (D, J) Riojasaurus, from Bonaparte (1972); (E, K) Camelotia, from Gabon (1985); (F, L) Plateosaurus. (F) from Gabon (1990); (L) from Weishampel and Chapman (1990). Scale dpproximately 50 mm.
>--1
ffi
/l
\vi
K
lr9D s
l.t, I
l'\
]{\
1
fI
I.d
l'1 1'
LIJ
\.\
\.l':
(tl"
ffi qh
!l
I
,ll
I
I
\t"/
}$ L
F} 111
axt
i'1
|
i\ { ('\ \I+,\.df
J {t inl ..,,,t !r. ll'1'
I
i:l
\l
J \i'1
Ltr
lt
F
E
\&rdF v
!\ t1 i
.t
#
The lower end of the fourth trochanter is 334 mm below the proximal end of the femur, so it extends to iust below femoral midlength. The fourth trochanrer:length ratio of 52% is similar to rhat of Riojasaurus.The upper end of the trochanter lies on the medial margin of the femur and its lower end is slightly removed from it; such a medial position occurs in Anchisauria (Fig. 1.1A.) and in sauropods. The trochanter is large as in other prosauropods and in Antetonitrus; it is not reduced to a crest or ridge as rn BdrdpasaLrrus, Isanosaurus, Kotasaurus, Shunosaurus, and Vulcanodon.
The medial surface of the trochanter and the adjacent medial surface of the femur are concave. There is also a distinct concavity proximolateral to the trochanter on the posterior surface. Above the fourth trochanter the femur is triangular in cross section, with the lateral margin forming the short base and the medial margin the apex of the triangle. Below the fourth trochanter the femur is elliptical in cross secrion, with the long axis oriented rransversely, as in SAM 3450, RioiaslLtrLts, Camelotia (Fig. 1.13A, D, E, G, J, K), Antetonitrus, and sauropods. However, the ratio of transverse to anteroposterior widths at midlength is 1.2 as against more than 1.5 in sauropods (\X/ilson 2002). The distal end of the femur is expanded medially, laterally, and posteriorly; the distal width exceeds the proximal width of the femur. The posteromedial condyle (pos-
22
.
Peter
M. Galton.
Tacques Van Heerden. and
Adam M. Yates
terior tibial condyle) lies on the medial margin. There is a deep, narrow sulcus between it and the fibular condyle. The sulcus is limited to the posterior surface. It is continued dorsally as a shallor,v depression. Halfway to the fourth trochanter, it passes over obliquely to the lateral side. The anteromedial condyle (anterior tibial condyle) is very incomplete. The fibular condyle is smaller than the posteromedial one. From the former extends a prominent, rounded ridge proximally on the posterior surface. Lateral to this ridge there is a distinct longitudinal concavity that forms a prominent "shelf." Although the surface is abraded, the holotype femur (SAM 3450) of Melanorosattrus readi is more or less complete (Fig. 1.13A, G; Haughton 1924, fig. 45; Van Heerden 1979, fig. 21, pls. 64,65). The femora of SAM 3450 and NM QR1551 are similar in that they are straight in anteroposterior views, with the head directed medially, the distal condyles projecting posteriorln the fourth trochanter on (or very close to) the medial margin in posterior view, and immediately distal to the fourth trochanter. The shaft is roughly suboval in cross section (the anteroposterior width is less than the transverse width). The femur of Riojasaurzs (Fig. 1.13D, J) is similar, except that the lesser trochanter is smaller. Flowever, in Plateosaurauus and Plateosaurus (Fig. 1.13R L), the distal part o{
the shaft is curved in anteroposterior views, so that the distal condyles face posterolaterally, the fourth trochanter is well removed from the medial margin, and the anteroposterior and transverse widths of the subcircular shaft are subequal. Based on a single large femur, Gauffre (1993) described a second species, Melanorosaurus thabanensis, from the Upper Elliot
Formation of Lesotho, southern Africa (Hettangian to Pliensbachian, Lower Jurassic; Olsen and Galton 1984; Olsen and Sues 1986). This femur (Fig. 1.13C, I) differs from NM QR1551 in being slightly more robust, but particularly in that the fourth trochanter is situated away from the medial margin (in posterior view) and extends slightly more distally than in either M. readi or Riojasaurus. Tibia. There are three more or less complete tibiae and the \\reathered proximal end of a fourth (Figs. 1.1OF-I;7.721-L 1.148 1.16J) (measurements given in Table 1.2). These represent two pairs of about the same length, but one pair is slightly more robust Fig. 1.10F, G; 1..1.6J) than the other (Figs. 1.10H, I; 1.12I-L; 1.148). The tibia is 79% of the length of the femur. It is compararir.ely slender and its two ends are mildly expanded compared to Rioiasawrus (Fig. 1.14C, E) and especially to the sauropod Blikanasauras, but still massive compared to Plateosaurzzs (Fig. 1.14D). The proximal articular surface is triangular, the anteromeJral margin forming the long base of the triangle, and it is flat, exrept for the condyles, sloping appreciably posterolaterally. Between :he two condyles there is a shallow sulcus (just behind the lateral :ondyle), which extends over the posterolateral edge onto the pos:crolateral surface. There is a prominent cnemial crest on rhe prox:rnal third of the bone. Lower down it is continued as a rounded
Postcranial Anatomv of Referred Specimens of Melanorosaurus
.
23
ffitr I Fig. 1.11. A-D: Right tibiae in lateral uieu. 14) Melanorosaurus readi, holotype SAM 3,150; (B/ M. readi, NM QR1.5.f 1; /C) Riojasaurus, from Bondpdrte
c)72 ) ; (D ) Plateosaurus, /rorz Gabr,tn (1990). (E) Ri1bt tibia and
(1
fibula oIRit:'jasJurus r/7 anterior uieru, from Bonaparte (1972). (F) Right fibula o/ Melanorosaurus readi, NM QR155l, in lateral uieru. (G) Right fibula of Plateu\aurur irt laterol uiew, lrom Gdlton (1990). H-J: Right dstrdgdlus in proximal uietu of (H) N,Ielanorosaurus readi, NM QRl551; (f Riojasaurus (r.rlrD calcaneum), from Nouas (1989); and (J) Plateosavrus, from Gabon (11)90). K-M: Left metdtdrsdls of Melanorosaurus readi, NM 8R/J51. (K. Lt llin tKt posterior and (L) lateral uieus; (M) lll in medidl uietu. N-O: Anteriur uiew of left pes o/ /N/ Riojasaurus, from BondpLlrte (1972), and (O) Plateosaurus, from Galton (1990). Scalc dpproximatelv .50 mm.
"11
o I ii
L$ &J
m
/ I' ii f l
ci{rl
{* ,'[i
rll ltl
Ir
1I
ilT{
u
t..'h
'J KL
r_.#
fl Tf,l
*s
ridge that extends to the medial margin of the bone. The distal third of the tibia is triangular in outline; the anterior surface is flat, and the medial and posterolateral surfaces are rounded. The distal end has the usual sauropodomorph configuration, with a prominent descending process that fits a corresponding depression in the astragalus. The lateral margin of the descending anteroventral process of the distal tibia protrudes lateraily as far as the anterolateral corner as in prosauropods (Fig. 1.12L) rather than being set well back from it as in Antetonitrus and sauroDods (Yates and Kitching 2003).
The holotype tibia (SAM 3449; Fig. 1.14A; Haughton 1.924, fig. 45; Van Heerden L979, fig.22, pls. 66,67) is somewhar more robust than that of NM QR1551 (Fig. 1.1a8). Like the femur, the
holotype tibia is weathered and the distal end is incomplete. Fibwla. There is a fairly complete right fibula (Figs. 1.12E-H; 1.14F) and the proximal half of a left one (Fig. 1.16K) (measurements given in Table 1.2). The prorimal end is flattened and has
24 .
Peter
M. Galton,
Jacques Van Heerden, and Adam
M.
Yates
Wff
ww
w cg
wffi, ffi &
w
Iq
&
H
w
KffiW
n
Y
W
ww
broad anteromedial and posterolateral surfaces. The latter is rounded, whereas the former has a knoblike elevation in the middle, with a shaliow groove on either side. The proximal half of the shelf is elliptical and the distal half subcircular in cross section. The distal end is less expanded than the proximal one. The posterolateral surface is slightly rounded, whereas the anteromedial surface has a distinct sulcus that becomes narrower proximally. Near the distal end the sulcus width is more than half that of the 6bula and
is confluent with a transverse groove just proximal to the distal condyle. The condyle itself is placed medially, as in PlateosdLtrus
(Fig. 1.14G). Tarsals. The only proximal tarsal represented are two right astragali, one well preserved (Figs. 1.14H; 1.15A-D) and the other weathered. It is a rounded, disk-like bone that articulates with the distal end of the tibia. In form it is similar to those of Rioiasaurus and Plateosaurws (Fig. 1.14I, J, N). The medial half of the dorsal surface is slightly concave, with the anterior margin turned dorsally. There is a very robust ascending process, with a rvelldeveloped excavation posteriorly that articulates with the descending process of the tibia. The astragalus thus closely interlocks with the tibia as in all sauropodomorphs. Anteriorly there is a fossa at the base of the anterior surface of the ascending process. A vascular ioramen is present here in prosauropods and Blikanasaurus (SAM K403), but it is lost in Vwlcanodon and other sauropods (Wilson :rnd Sereno 7998). The condition rn Melanorosaurus is unknown because matrix still fills the area. The lateral surface of the astragalus is slightly higher than broad, and there is a shallow concavity
W
Fig. 1 .15. Melanorosaurus readi,
NM QR1.t.t1. A-D: Right astragalus in (A) anterior, (B) proximal, (C) posterior, and (D) distal uiews. E-F: Distal tarsals in (E) proximal and (F) distal utea,s. G-H: Distal pdrt of left metdtarsdl IV in tGt lateral and tHt postcrior uiews. l-K: Left metatarsal lll in (I) medidl, (J) anterior, and (K) lateral uiews. phalanges
in
L-Q:
Pedal ungual
(1., P) dorsal and
(M-O and Q) side uiews. Scale is 50 mm.
Postcranial Anatomy of Referred Specimens of Melanorosaurus
'
25
TABLE 1.3. Measurements (in mm) of Metatarsals of Melanorosawrus readi
(NM QR1ss1) MT I (1) MT I (2) MT II t29 t20 180
Description Length
\fidth of proximal
MT III 225
70
a/
65
29
31
.)L
z5
69
63
59
63
surface
76
\)(/idth of proximal surface
(parasagitally) (para-rransversely
)
\fidth of distal end (
para-transversely)
where the calcaneum abutted against it. The ventral surface of the astragalus is strongly convex anteroposteriorly. A single, distal tarsal is known for Melanoroslurus, but not
from which side of the bodn nor if it was the only one present. The distal tarsals of Riojasdttrus are in articulation (Fig. 1.14N), thus limiting what can be seen, and those of the saurop od Blikanasdurus are poorly preserved (these elements are absent or unossified in other sauropods; \(ilson 2002). The distal tarsal of Meldnoroslurus has a gently convex proximal surface (Fig. 1.15E) and is mostly concave distally, with a more strongly concave part with a large foramen (Fig. 1.15F). The convex edge is sharp, whereas the opposite concave edge is beveled and pierced by another foramen (Fig. 1.1sE). Metatarsdls. There are remains of two metatarsals
I and one metatarsals II and III, and the possible remains of metatarsal IV (Figs. 1.SJ-L; 1.14K-M; 1.15G-K; 1.16A-D) (measurements given in Table 1.3). The metatarsals are similar to those of Rioiasaurus, Plateosaurus (Fig. 1.14N. O), and Pldteosaurauus, but are much less massive than those of the early sauropods Antetonitrus and B likanasaurus. One of the right metatarsals I (Figs. 1.8J-L) is slightly larger and more weathered than the other. The proximal end surface is elIiptical. The posterolateral surface, which abuts against the proximal end of metatarsal II, is divided into two subequal halves: the larger, aimost flat surface faces posterolaterally, whereas the smaller, flat surface faces laterally. The metatarsal is slightly constricted about two-thirds from the proximal end and broadens distally. The each
of
lateral distal condyle is much larger than the medial one, so that the first digit was directed anteromedially as in prosauropods (Galton and Cluver 1976). The proximal articular surface of left metatarsal II (Fig. 1.16A-D) is hourglass-shaped, with concave long sides an' teromedially and posterolaterally, whereas the other two sides are short and straight" All four surfaces of the proximal third are longitudinally concave. The largest and deepest of these concavities is the
26.
Peter
M. Galton,
Jacques Van Heerden, and Adam
M.
Yates
€
:'r
,
,h,l ut, !{lr,
;,,,
W
c
i|[
ffi
,''.St"'
llirl&r
Fig. 1.15. N{elanorosaurus readi,
NM QR15i1. (A-D:) Left
";$, k: ,.
I
,w
fr-*re ru ru
metatarsal II in (A) mediaL, (B) anterior, (C) lateral, and (D1 posterior uietus. E-C: Problenl bone in three uiews. H-l: Phalanx tprohahly digit lVt in tHt anteriur and (l) side uiews. (J) More massiue left tibia in lateral uietu. (K) Proximal part of left fibula in medial t icu'. Scalc is \o mrn.
anteromedial one. The shaft is short without being stocky. Distally, the lateral condyle is slightiy larger than the medial one. Metatarsal III is complete, but weathered (Figs. 1.14M; 1.15I-K). It is 36% of the length of the femur. The proximal articular surface is triangular,
rvith the apex of the triangie directed posterolaterally. The medial long side is slightly concave and the posterolateral long side is straight. The medial surface of the proximal end has a triangular concavity, whereas the anterior surface has a much smaller, circular concavity. The posterolateral surface is subdivided into a smaller, concave posterior part and a rounded lateral part. The shaft is comparatively long, slender, and subcircular in cross section. The two distal condyles are of equal size, but the lateral one proiects farther distalll'. The left metatarsal IV is incomplete proximally and poorly preserved (Fig. 1.15G, H). Phalanges. There are thirteen disarticulated phalanges, five of rvhich are unguals (Figs. 1.15L-Q; 1.16H, I), but it has not been possible to reconstruct any digit with certainty. Most of the phalanges are the same shape as in other prosauropods, such as Rioidsaurus and Plateosdurus. (Fig. 1.14N, O). However, some ot them are broader transversely than they are long anteroposteriorln Postcranial Anatomy of Referred Specimens of Melanorosaurus
.
27
as in Camelotia and sauropods including Blikanasaunzs. The unguals vary in size from rhe very large rrenchanr ones, which proba-
biy belonged to the first digit of the manus, to the less trenchant, smaller ones found on the outer digits of the pes (Fig. 1.1L-Q). Discussion Sacral types. Comparisons with
NM QR1551 (Fig. 1.6A) indi-
cate that the three cenrra illustrated by Van Heerden (1979, pls. 58,
59, 60) are those of sacral vertebrae 1 and 2 (with most of left rib) and a possible proximal caudal of Plateosaurauus. Comparisons with Plateosaurus and other prosauropods (Gaiton 1999, 200Ia, 2001b) indicate that the sacrum of Plateosaurauus probably had the derived prosauropod sacrum with three vertebrae, namely, sacral vertebrae 1 and 2 (Van Heerden 1979: figs.7,9, pIs. 13, L6, 17) with a dorsosacral (Van Heerden 1979: fig. 8, pls. 74, 15; misidentified as the third sacral, a caudosacral, see Galton 2001b); Novas \1996) came to the same conclusion for Riojasaurus (Fig. 1.6E). In prosauropods, the plesiomorphic condition of rwo sacral vertebrae are supplemented by a third, which can be incorporated from either the tail (S1 + 52 + CS as in Plateosaurus) or from the dorsal series (DS + 51 + 52 as in Massospondylzs) (Galton 1999, 2001b). The basal sauropodomorph Saturnalia (Santa Maria Formation, Upper Triassic, Brazil) has incorporated a caudosacral into the sacrum (Langer etal. 1999), as was probably also the case for the basal sauropodomorph Thecodontosaurus (Norian, Upper Triassic, England; Galton 1999;Benton et al. 2000). A cladistic analysis of the Prosauropoda (Galton and Upchurch
2004) indicates that a sacrum with a caudosacral (S.l + 52 + CS) represents the plesiomorphic state for the Prosauropoda, whereas a sacrum with a dorsosacral (DS + 51 + 52) represents rhe synapo, morphic state. However, a detailed cladistic analysis shorvs that the history of the sacrum in prosauropods is complicated (Galton and Upchurch 2000, Galton and Upchurch 2004; for figures of sacra see Galton 1999). Apart from Melanorosaurus, the only other prosauropod in which there were previously thought to be four sacral vertebrae is Massospondylus (as DS1 + 51 + 52 + CS, Cooper
1981). However, Galton (1999) reinterpreted the sacrum, suggesting that DS1 is probably an unmodified dorsal, 51 is a modified dorsal, and 52 and CS represent the two reptilian sacral vertebrae. This results in the sacrum being DS1 + 51 + 52, a concluslon confirmed by undescribed sacra of Massospondylus (BPI, Vasconcelos, pers. comm.). Basal sauropods have four sacral vertebrae but this represents a convergence, the extra vertebra being another caudosacral ('!7ilson and Sereno 1998;'Sfilson2002). so rhe sacrum is 51+52+CS1+CS2. Relationships o/ Melanorosaurus readi
The absence of a skull and manus and the incompleteness of the pes, the sources of many character states used in cladistic analy-
28 .
Peter
M. Gahon,
Jacques Van Heerden, and Adam
M.
Yates
of the Sauropodomorpha, greatly reduce the number of characters available to classify Melanorosdurus. The synapomorphies used to place Melanorosaurus within the cladogram (Fig. 1.1A) are based on the cladistic analysis of Galton and Upchurch (2004). NM QR1551 is referred to the Sauropodomorpha (Node 1) because the large ascending process of the astragalus keys into the distal articular surface of the tibia (Fig. 1.15A-C; tibia SAM 3449, Fig. 1.14A).It is a prosauropod (Node 2) because the centra of the posterior dorsal vertebrae are elongate (length:height ratio is greater than 1.0, Fig. 1.4A), the posterior dorsal vertebrae lack a prezygodiapophyseal lamina (Fig. 1.4A), and the pubis has a large obturator foramen (Fig. 1.10C-E). It is referred to Node 3 (unnamed) because the deltopectoral crest terminates at or below the midlength of the humerus and the distal end of the ischium is expanded dorsoventrally (Fig. 1.11B). Referral to Node 4 (unnamed) is based on the acetabulum not being backed medially by a sheet of bone (Fig. 1.10A; SAM 3449, Haughton L924,frg.44),the subtriangular outline to the distal end of the ischium (Fig. 1.11A-C), and the increased robustness of metatarsals II and III (Figs. 1.151-K; 1.16A-D). NM QR1551 is referred to the Anchisauria (Galton and Upchurch 2004) (Node 5, Fig. 1.1A) because the forelimb:hindlimb length ratio is greater than 0.60, the obturator foramen is completely visible in the anterior view of the pubis (Fig' 1.10E), the femoral shaft is distally straight in anteroposterior views (Fig, 7.1,28, D; 1.13B; SAM 3450, Fig. 1.13A), the fourth trochanter of the femur is displaced to the posteromedial margin of the shaft (Fig. 1.12D; 1.138; SAM 3450, Fig. 1.13A), and the proximal end of metatarsal II is hourglass-shaped. NM QR1551 and SAM 3449 are referred to the Melanorosauridae Huene 1929 (Node 7,Frg. 1.1A) because of the following apomorphies: the ilia have a step-like sigmoid profile to the dorsal margin in lateral view (Figs. 1.9A;1.10A) and, for the femora, the proximal and lateral margins meet at an abrupt right angle in anterior view (Fig. 1.128, C; SAM 3450). The lesser trochanter is a prominent, sheet-like structure (Fig' 1.12A' B; also SAM 3450 but eroded), which projects beyond the lateral margin of the femur so that it is visible in posterior view (Figs' 1'.12D; 1.13B; SAM 3450), the distal end of the fourth trochanter lies at or below femoral midlength (Figs. 1.11F; 1.12A, C, D; 1.138; SAM 3450, Fig. 1.13A, G), and below that the shaft is transversely rvidened and anteroposteriorly compressed (Figs. 1.12A, C; 1.13H; cf. Figs. 7.128, D; 1.13B; SAM 3450, Fig. 1.13A, G). NM QR1551 is referred to Node 8 (unnamed lCdmelotia + Melanorosaurusl) because the proximal caudal centra are high relative to their axial length (Fig. 1.5B, C) and at least some pedal phalanges, ercluding unguals, are broader transversely than their proximodistal length (Fig. 1.16H,I). An autapomorphy for Melanorosattrzs (NM QR1551) is a dorsosacral as the fourth sacral vertebra, so the sacrum is DS1 + S1 ses
Postcranial Anatomy of Referred Specimens of Melanorosaurus
'
29
+ 32 + CS. The sacrum of Massospondylus was rhoughr to consist of four vertebrae but, as noted above, several undescribed sacra show that it is DS + 51 + 2. The sacrum of NM QR1S51 differs from those of sauropods in which the extra sacral is a caudosacral (Wilson 2002), so the sacrum is 51 + 52 + CS1 + CS2 (Galron 1999\.
NM QR1551, however, does not possess the following synapomorphies of the Anchisaurudae Marsh, 1885 (Node 6, Anchisaurus and AmmosdLtrus, Fig. 1.1A), the sister group to the Melanorosauridae (Galton and Upchurch 2004): the iength:height ratio of
the longest post-axial cervical centrum is at least 3.0 (2.0, Fig. 1.3A), the deltopectoral crest terminates above 50% of humerus length (below midlength, Fig. 1.8D, E), the anterior process of the ilium terminares in front of the distal tip of the pubic process in lateral view (or rather behind it, because the process is short; Figs. 1.9A; 1.10A; SAM 3449,Haughton 1,924, fig. 44), and the area between the anterior and pubic processes of the ilium is acute in lateral view (not acute, Figs. 1.9A; 1.10A; SAM 3449, Haughton 1924, fig.44). NM QR1551 also lacks the following synapomorphies of the Plateosauria Tornier 1913 (Node 10, Fig. 1.1A), the sister group of the Anchisauria (Galton and Upchurch 2004). The length:height ratio of the longest post-axial cervical centrum is at least 3.0 (2.0, Fig. 1.3A) and the sacrum consists of a dorsosacral plus sacral vertebrae 1 and 2 with loss of the caudosacral (piesiomorphic caudosacral retained in the sacrum, Frg. 1.6A). The characters of the Sauropoda have been discussed by Wilson and Sereno (7998), Upchurch (1998), Sereno 11.199), and mosr
recently by rWilson (2002). Of the twenry-one synapomorphres listed by Wilson (2002), seven refer to elements nor preserved in NM QR1551. As in sauropods, the humerus-to-femur ratio is 0.70 or more (0.71) and the femoral midshafr is elliptical in crosssection but the rransverse diameter is only about I20"h of the anteroposterior diameter, not at least 150% as in sauropods. The
basal saurop ods Barapasdurus, Kotasdwrus, and Shunosaurus have
four sacral vertebrae, a synapomorphy for Sauropoda (Upchurch 1995; \iliilson and Sereno 7998; \flilson 2002\. but the additional sacral vertebra of NM QR1551 (Fig. 1.6A, B) is a dorsosacral, not a caudosacral as in Sauropoda (\X/ilson and Sereno 1998; Wilson 2002). Unlike the situation in sauropods, the posture of NM
QR1551 was probably nor that of a columnar, obligate quadruped; the base of the anterior caudal transverse processes is shallow, not
deep (Fig. 1.5D, F); the incomplete deltopectoral crest of the humerus is large, not reduced to a low crest or ridge (Fig. 1.8D-F); the proximai end of the ulna (Fig. 1.SG-I) lacks three sauropod characters (triradiate with deep radial fossa [appears to be subtriangular], ulnar proximal condylar processes unequal in length with anterior arm longer [unknown, forearm incomplete], olecranon process reduced or absent [large]); the ischial peduncle of the ilium
30.
Peter
M. Galton,
Jacques Van Heerden, and Adam
M.
Yates
is deep (Figs. 1.9A; 1.10A), not low; the distal ischium is triangular,
not bladelike (Fig. 1.11A-C); the fourth trochanter of the femur is large (Figs. 1.11E, F; 1.12C, D), not reduced to a crest or ridge; the vascular fossa at the anterior base of the ascending process of the astragalus is not lost (Figs. 1.14H;1.15A, B; foramina hidden by matrix in fossa); and there is an ossified distal tarsal (Fig. 1'15E' F); but present in BlikanasaurLts, a sauropod sensu (Upchurch et al. 2004\.
In Yates (2003a), Ancbisawrus and Meldnorosaurus are Prosauropoda in their traditional position outside of the Sauropoda (Fig. 1.1,{; e.g. Gaiton 1990; Sereno 1999; Benton and Storrs in Benton et al. 2000; Upchurch 1995,1998; Upchurch et aL.2002; 'Wilson and Sereno 1998, Wilson2002). However, Yates and Kitching (2003) include these genera in a modified concept of the Sauropoda (Fig. 1.1B; also Yates 2002), but the reasons for this change are yet to be published (Yates in press b). Yates and Kitching (2003: electronic appendix A) list seventeen unambiguous synapomorphies for Sauropoda, but onlv three of these characters involve bones preserved in NM QR1551, namely, the loss of a strong constriction between the transverse process and the sacral rib of sacral vertebra 1 (vs. constriction stiil present' Fig. 1.4H), reversal to a humerus with a transverse width of the distal end that is less than 33% of the length (vs. present, 24'/.), and a distal end of the tibia with a posteroventral process that does not extend as far laterally as the anterolateral corner (vs. no, prosauropod condition in which it projects at least as far, Fig. 1.L2J,L; coded as a sauropod because Yates previously only saw the incomplete tibia of NM QR1551)' Two of the three sauropod synapomorphies with DELTRAN of Yates and Kitching (2003) are for bones preserved in NM QR1551, namely, loss of semicircular fossa on the distai flexor surface of the humerus (vs. vague fossa present, not as flat as in sauropods and not sharp-edged as in prosauropods, but the distal surface is a little eroded; Figs. 1.7E; 1.8D) and the fourth trochanter is on the medial margin of the femur in posterior view (present, Figs. 1.12D; 1.138; SAM 3450, Fig' 1.13A). Only five of the twenty-five character states for the sauropod synapomorphies with ACCTRAN of Yates and Kitching (2003) involve bones present in NM QR1551, namely, reversal to a scapula blade rvith a midsection that has parallel margins (vs. present, Figs' 1.7B; 1.8A), reversal to a deltopectoral crest that extends to less than midlength (no, beyond midlength, Fig. 1.8D, E). For the proximal condyles of the ulna. the anterior process is much greater than the lateral process (not known as lateral process incomplete), and the transverse width of the ischial shaft is greater than its depth (vs' no, the re-
of the fourth trochanter of the femur is not rounded and symmetrical (asymmetrical' Figs. 1.11E, F; 7.IzC) as in sauropods. Yates and Kitching (2003) list four unambigous synapomorphies for the next node (Melanorosdurus + the rest' Fig. 1.1B), all
verse, Fig. 1.11A-C), and the profile
of which involve
bones present
in NM QR1551,
namelS the
Postcranial Anatomy of Referred Specimens of Melanorosaurus
'
31
dorsoventral height of hyposphenes equal to thar of the neural canal (present), and reversal to tall dorsal neural spines that are greater than 1.5 times the length of their bases (not the case but only one dorsal with a neural spine, Figs. 7.2E; I.3E). The sacrum includes a caudosacral vertebra (present, Fig. 1.6A, D), and there is a deep radial fossa on rhe proximal ulna (it appears to be subtriangular in NM QR1551, present in SAM 3449).
Yates and Kitching (2003) list thirty-three synapomorphies with ACCTRAN for this node, but only four of these characteis relate to bones preserved in NM QR1551, namely, loss of the free tip from epipophyses of all postaxial cervicals so epipophyses are joined to postzygapophyses along their length (present, Figs. 1.28; 1.3A), with a minimum width of the scapular blade greater than 20"/. of its length (present, 23ok,Figs.1.78; 1.8A), anreroposrerior length of distal pubic expansion is greater than 15% of the length of the pubis (vs. no, 1.2o/o, Fig. 1.10D), and a crestlike lesser trochanter on the femur with height exceeding basal width (present, Fig. 1.12A, B). The referred specimen NM QR1551 of Melanorosdurus is a prosauropod according to the analysis of Galton and Upchurch 1in Upchurch et aL.2002; Upchurch et al. 2004). Based on the synapomorphies given by V/ilson (2002), it is not a sauropod (Fig. 1.1A). However, this specimen is placed within a redefined Sauropoda as the sister group of the rest of the Sauropoda (Fig. 1.1B) by yates and I(itching (2003; also Yates in press b), but, based on the dererminable character states, NM QR1551 lacks several synapomorphies for both of these nodes. A more detailed cladistic analysis of the basal Sauropodomorpha is being prepared by Upchurch, Bar, rett, and Galton (in prep.). In addition, a more detailed analysis of
NM QR1551, NM QR3314, and the original material
of
Melanorosaurus (SAM 3449,3450,3532) of Haughton (1924) is being undertaken by Yates, Van Fleerden, and Galton (in prep.). It is hoped that these studies will clarify the systematic position of
Melanorosaurus readi that,
for the moment, is
referred
to
as
Sauropodomorpha incertae sedis. Acknowledgments. Jacques Van Heerden thanks Teresa Bos-
well (NM) for the drawings of Melanoroslurus. Adam M. yates thanks Elize Butler and John Nyaphuli (NM) and Sheena Kaal (SAM), for assistance with the collections; earlier, Anusuya Chinsamy-Turan (formerly at SAM, now ar University of Cape
Town, South Africa) kindly provided information and photographs of Haughton's material. Peter M. Galton thanks Michael euinn and Joseph Souza (University of Bridgeport) for printing the photographs.'We thank Cecilio Vasconcelos (BpI) for providing information on the sacrum of Massospondylus and appreciate the useful comments made by the editors and especially those by paul M. Barrett (The Natural History Museum, London, England), who provided excellent detailed reviews of the different versions of this paper.
32.
Peter
M. Galton,
Jacques Van Heerden, and Adam
M. yates
References Cited
Benton,
M.J. 1993. Reptilia. In M.J. Benton, ed.,The Fossil Record
2:
681-71.5. London: Chapman and Hall. 1.994. Late Triassic to Middle Jurassic extinctions among continental tetrapods: Testing the pattern. In N. C. Fraser and H.-D. Sues, eds., In the Shadow of the Dinosattrs: Earll' Mesoz.oic Tetrapods,
366-397. Cambridge: Cambridge University Press. Benton, M.J.. L.Juul. C. W. Storrs. and P. .V. Galton. 2000. Anarornr and systenratics of the prosauropod dinosaur Th ecodontosduru s dntiquLts from the Upper Triassic of southwest England. Journal of Vertebrate P aleontology 2Q: 7 1-102. Bonaparte, I.F. 1969. Dos nuevas "faunas" de reptiles Tri6sicos de Argentina. I Gondtuana Symposium, Mar del Plata, Ciencas Tierra 2: 283-306. 1,972.Los tetrdpodos del sector superior de la Formaci6n Los Colorados, La Rioja, Argentina (Triisico Superior). I Parte. Opera LilIiond 22:1-183. Bonaparte, J. F. 1986. The early radiation and phylogenetic relationships of the Jurassic sauropod dinosaurs, based on vertebral anatomy. In K' Padian, ed., The Beginning of the Age of Dinosaurs, 247-258. Cambridge: Cambridge University Press. 1999. Evoluci6n de las v6rtebras presacras en Sauropodomorpha' Amegh iniana 36: 1'1' 5-187 . Bonaparte,J. F., andJ. A. Pumares.I995. Notas sobre el primer craneo de Rictiasaurus incertus (Dinosauria, Prosauropoda, Melanorosauridae) del Triasico Superior de La Rioja, Argentina. Ameghiniana 32
34r-349.
Buffetaut, E., V. Suteethorn, G. Cunl', H. Tong, J. Le Loeuff, S. Khansubha, and S. Jongautcharlyakul. 2000. The earliest known sauropod dinosaur. Nature 407 : 72-74' Buffetaut, E., V. Suteethorn, J. Le Loeuff, G. Cuny, H. Tong, and S. Khansubha. 2002. The first giant dinosaurs: A large sauropod from the Late Triassic of Thailand. Comptes Rendus: Palectuolunte 1: 103-109. Cooper, M. R. 1981. The prosauropod dinosaur Massospondylus carina/r.rs Owen from Zimbabwe: Its biologn mode of life and phylogenetic significance. Occasional Papers, National Museums and Monuments of Rhodesia (Series B) 6: 689-840. 1984. A reassessment of Vulcanodon kdribaensis Raath (Dinosauria: Saurischia) and the origin of the Sauropoda' Palaeontologid Africana 25 203-231.
Galton, P. M. 1,976. Prosauropod dinosaurs (Reptilia: Saurischia) of North America. Postilla 1'69: 1'-98. 1985. Notes on the Melanorosauridae, a family of large prosauropod dinosaurs (Saurischia: Sauropodomorpha). Gdobios 19: 671'676.
1990. Basal Sauropodomorpha-prosauropods.
In D. B. Weis-
hanrpel, P. Dodson, and H. Osm61ska, eds., The Dinosauria' 320344. Berkeley: Universitl' of California Press' 7992. Basal Sauropodomorpha-prosauropods. In D. B. Weis-
hampel, P. Dodson, and
H. Osm6lska,
eds., Tbe Dinosauria, 320-
344. Paperback ed. Berkeley: University of California Press' 1998. Saurischian dinosaurs from the Upper Triassic of England: Camelotia (Prosauropoda, Melanorosauridae) and Aualonianus (Theropoda, Carnosauria). Paldeontographica A 250: 1 5 5-172' Postcranial Anatomy of Referred Specimens oi Melanorosaurus
'
33
1999. Sacra, sex, Sellosaurzs (Saurischia: Sauropodomorpha; Upper Triassic, Germany)-or \\,hy the character ..two sacral verte_ brae" is a plesiomorphic character for Dinosauria. Neues lahrbuch filr Geologie und Paliiontolctgie, Abhandlungert 213: 19-55. 2001a. The prosauropod dinosaur plateosaurus Meyer, 1g37 -. (Saurischia: Sauropodomorpha). II. Nores on referred species. Rerze de Pal1obiologie 20 435-502. 2001b. Prosauropod dinosaur sellosaurus graciris (Upper Triassic,
Germany): Third sacral vertebra as either a dorsosacral or a caudosacral. Neues Jahrbuch fiir Geologie und pakictntologie, Monat-
schefte 200I:6gg_704.
Galton, P. M., and M. A. Cluver. I976. Anchisaurus capensis (Broom) and a revision of the Anchisauridae (Reptilia, Saurischia). Annars of the South African Museutn 69: 121-159. P. M., and J. Van Heerden. 1985. partial hindlimb of Btik_ anasaurus cromptoni n. gen. and n. sp.. represenring a new famill.of prosauropod dinosaurs from the upper Triassic of South Africa. Gdo-
Galton,
bios 18:509-516. P. M., and J. Van Heerden. 1.998. Anatomy of the prosauropod di_ nosaur Blikanasdurus cromptoni (Upper Triassic, South Africa), with
Galton,
notes on the other tetrapods from the Lower Elliot Formation. Paliiontologische Zeitschrift 72: 163-177. Galton, P. M., ar.rd P. Upchurch. 2000. Prosauropod dinosaurs: Homeotic transformarions ("frame shifts") with third sacral as a caudosacral or a dorsosacral. Journal of Vertebrate palectntology 20 (supp. to 3): 43A.
2004. Basal sauropodomorpha-prosauropods. In
D. B.
lTeishampel, P. Dodson, and H. Osm6lska, eds.. The Dinosauria. 2d ed. Berkeley; University of California press.
Gauffre, F.-x. 1993. The most recent Melanorosauridae (sauriscl.ria, Prosauropoda), Lower Jurassic of Lesotho, with remarks on the prosauropod phylogeny. Neues lahrbuch filr Geologie und paltion_
tologie, Monatschefte 1993: 64g_6 54. s.H. 1924. The fauna and stratigraphy of the Stormberg series. Annals of the South African Museum 12: 323-497. Huene, F. von. 1906. Ueber die Dinosaurier der aussereuropdischen Trias. G eolo gisch e und P alaeontolo gis ch e. A b h a n d I u n ge n g, 97 _ t 5 6. 1907-1908. Die Dinosaurier der europdischen Triasformation mit Berticksichtigung der aussereuropdischen Vorkom nisse. G eologisch e und Palaeontologische, Abhandlungen supp. 1,: I-419.
Haughton,
1926. Vollstandige Osteologie eines plateosauriden aus der schwdbischen Trias. Geologische und palaeontorogische, Abhandlungen, Neue Folge 15: 129-179. 7929. Kurze Ubersicht iiber die Saurischia und ihre natiirlichen Zusammenhd nge.
p ahi o rttol o gi s c b e Z e it s ch r ift I 1 : 269 _27 3. 1.932. Dre fossile Reptil-Ordnung Saurischia, ihre Entwicklune und Geschichte. Monographie Geologie und paliiontologie (serie {: |
r-361.
Huxley, T. H. 1866. on the remains of large dinosaurian reptiles from the Stormberg Mountains, Sourh Africa. Geological Magazine 3:563. Jain, S. L., T. S. Kutty, T. Roy-Chowdhury, and S. Chatteriee. j.975. The sauropod dinosaur from the Lower Jurassic Koto Formation of India. Proceedings of the Royal Society of LondonB 1.88:221-22g.
34 .
Peter
M. Galton,
Jacques Van Heerden, and Adam
M.
yates
1979. Some characters of BarapasaurLrs tdsorei, a sauropod dinosaur from the Lolr'er Jurassic of Deccan, lndia. Proceedings of the IV International G ondwana Symp o s ium, Cal cutta 1' : 204-21 6. Kitching, J.'$7., 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. P alaeontologia Africana 25 : 1'1'I-125.
Langer, M. C., F. Abdala, M. Richter, and M.J. Benton. 1999. A sauropodomorph dinosaur from the Upper Triassic (Carnian) of southern Brazrl. Comptes rendus de I'Acaddmie des Sciences, Paris, Sciences de ld Terre et des planites 329: 511-517. Lucas, S. G., and P.J. Hancox. 2001. Tetrapod-based correlation of the nonmarine Upper Triassic of southern Africa. Albertina 25 5-9. MacRae, C. 1999. Life Etched in Stone. Johannesburg: Geological Society of South Africa. Marsh, O. C. 1885. Names of extinct reptiles. American Journal of Science (series 3)
29:169.
1895. On the affinities and classification of the dinosaurian repti\es. American Journal of Science (series 3) 50: 483-498. Novas, F. E. 1989. The tibia and tarsus in Herrerasauridae (Dinosauria, incertae sedis) and the origin and evolution of the dinosaurian tarsus. Journal of Paleontology 63: 677-690. 1996. Dinosaur monophyly. Journal of Vertebrate Paleontology 1.6:723-741.. Olsen, P. E., and P. NI. Galton. 1984. A review of the reptile and amphibian assemblages from the Stormberg of southern Africa, with special emphasis on the footprints and the age of the Stormbetg' Palaeontologia Africana 25: 87-l 10. Olsen, P. E., and H.-D. Sues. 1986. Correlation of continental Late Triassic-Jurassic tetrapod transition. In K. Padian, ed.,The Beginning of the Age of Dinosaurs,321-351. New York: Cambridge University Press.
Raath, M. A. 1972. Fossil vertebrate studies in Rhodesia: A new dinosaur (Reptilia; Saurischia) from near the Trias-Jurassic boundary-' Arnoldia (Rhodesia) 30: 1-37.
7984.
In
memoriam: Sidney Henry Haughton (1888-1982).
Palaeontogia Africana 25: l-3. H. G. 1898. On large terrestrial saurians from the Rhaetic beds of 'Wedmore Hili, described as Aualonia sdnfordi and Picrodon herueyi. Geological Magazine (series 4) 5: 1'-6. Sereno, P. C. 1,999. The evolution of dinosaurs. Science 284:2137-2747.
Seeley,
D.L. 1998. PAUP": Phylogenetic analysis using parsimon,v ("and other methods), Version 4.0. Sunderland, Mass.: Sinauer. Tornier, G. 1913. Reptilia (Paldontologie). Handtuorterbuch Naturwissensch aften 8: 337 -37 6. Upchurch, P. 1995. Evolutionary history of sauropod dinosar"rrs. P/rllosophical Transactions of the Royal Society of London B 3492 365-390. 1998. The phylogenetic relationships of sauropod dinosaurs. Zoological Journal of the Linnean Society of London 124: 43-103. Upchurch, P., P. M. Barrett, and P. Dodson. 2004. Sauropoda. In D. B. 'Weishampel, P. Dodson, and H. Osm6lska' eds., The Dinosauria' 2d ed. Berkeley: University of California Press. Swofford,
Postcranial Anatomy of Referred Specimens oi Melanorosaurus
'
35
Upchurch, P., C. A. Hunn, and D. B. Norman. 2002. An analysis of di_ nosaurian biogeography: evidence for the existence of vicariance and dispersal patrerns caused by geological events. proceedings of tbe Royal Society of London, Biological Sciences 269: 61.3-621. van Heerden , J. 1977 . The comparative anaromy of the postcranial skele, ton and the relationships of the South African xrlelanorosauridae (Saurischia: Prosauropoda ). Ph.D. dissertation. Bloemfontein: University of the Orange Free State. I979.The morphologv and raxonomy of Euskelosaurus (Repril\a: Saurischia; Late Triassic) from South Africa. Nauors Nasionai Mu_ seum, Bloemfontein 4: 2I-84. van Heerden, J., and P. M. Galton. 1997.The affinities of MeranorosaurLts-
a Late Triassic
J ah
'Wedel,
rbuch
fiir
G
prosauropod dinosaur from South Africa. Nezes
eologie und
p aliiontologie,
Monats
ch
efte
19 97
: 39_55.
J., R. L. Cifelli, and R. K. Sanders. 2000. Sauroposeidon oro_ teles, a new sauropod fror.n the Early Cretaceous of Oklahoma. I'cturnal of Vertebrate Palaeontology 20: I09-I 14. \7eishampel, D. B., and R. E. Chapman. 1990. Morphometric stud1, of P lateosaurtts from Trossingen (Baden-\fiirtemberg, Federal Republic of Germany). In K. C. Carpenter and p. J. Currie, eds., Dinosaur Systematics: Approaches and Perspectit,es, 43-5j.. Cambridge: Cam_ lVI.
bridge University Press.
rfellnhofer, P. 1993. Prosauropod dinosaurs from the Feuerletten (Middle Norian) of Ellingen near \)Teissenburg in Bavaria. Reuue de pat6obiotoPte i /-6J-/-/ l. Ifelman, J. 1998. Euskelosaurus and the origin of dinosaurs. Journal ctf African Earth Science 27 (IA)t 209.
1999. The basicranium
of a
basal prosauropod from the
Euskelosaurus range zone and thoughts on the origin of dinosaurs. Journal of African Earth Science 29(I):227-232. 'Wilson, J. A. 1999. A nomenclature for vertebral laminae in sauropods and other sau'ischian dinosaurs. .lournal of vertebrate paleontirogy 19: 639-653. 2002. Sauropod dinosaur phylogeny: critique and cladistic analy_
sis. Zoological .lournal 11-
1-/
of the Linnean
Society (London) 1.36:
rff/ilson,
J. A., and P. C. Sereno. 1998. Early evolurion and higher phy_ logeny of sauropod dinosaurs. Society of Vertebrate paliontoiogy, Memoir 5: 1-68. Yadagiri, P. 2001. The osteology of Kc.ttasaurus yamanpalliersrs, a sauropod dinosaur from rhe Early Jurassic Kota Formati on of India. Iour_ nal of Vertebrate Paleontology 21.: 242-252. Yates, A. M. 2002. A re-examination of the phylogenetic position of the
unusual sauropodomorph, Anchisaurus. p alaeontological Association 1\ewsletter )U: J.). 2003a. A new species of the primitive dinosaur, Thecodctntosaurus (Saurischia: sauropodomorpha), and its implicati.ns for the systematics of early dinosaurs. Journal of Systematic palaeontology 1: 1_42. 2003b. The species raxonomy of the sauropodomorph dinosaurs from the Lowenstein Formation (Norian, Late Triassic) of Germany. P a lae ontolo gy 4 6: 3 1.7 -3 37 . In press a. The first definitive prosauropod dinosaur from the
Lower Elliot Formation (Norian: Upper Triassic) of South Africa. Palaeontologia Africana 39 (Dec. 2003).
36 .
Peter
M. Galton,
Jacques Van Heerden. and
Adam M. Yates
In press b. Anchisdurus pctlyzelus Hitchcock: the smallest known sauropod dinosaur and the evolution of gigantism amongst sauropodomorph dinosaurs. P ostilla. Yates, A. M., and J. !7. Kitching. 2003. The earliest knorvn sauropod dinosaur and the first steps tovnards sauropod locomotion. Proceedings of the Royal Society of London, Series B, 270: 7753-1758. Young C.-C. I94Ia. Gyposailrus sinensis (sp. nor..), a ner'v Prosauropoda from the Upper Triassic Beds at Lufeng, Yunnan. Bulletin rtf the Geological Society of China 2I:205-253. 1941b. A compiete osteology of Lufengosaurus huenei Yotng (gen. et sp. nov.). Palaeontologica Sinica (series C) 7:1,-53. 1947. On Lufengosdnrus nrlgnus (sp. nov.) and additional finds of Lufengosaurtrs huenei Young. Palaeontologica Sinica 12: 1-53. 1948. Further notes on Gyposattrus sinensis Young. Bulletin of the Geoktgical Society of Cbina 28: 91-103. 1951. The Lufeng saurischian fauna in China. Palaeontolologica Sinica (series C) 13:1.-96. Zhang Y. 1988. The Middle lttrassic dinosaur fauna from Dashanpu, Zigong, Sichuan. Sauropod dinosatrrs. Pt. 1. Shunosaurus, 89. Chinese translation rvith English summary. Chengdu: Sichuan Science and Technology Publications House.
Postcranial Anatomy of Referred Specimens oi Melanorosaurus
.
37
2. The Genus Barosaurus Marsh ( Sauropoda, Diplodocidae) JonN
S.
McINrosH
Abstract The sauropod genus Barosaurus, which was long thought to be rare (in North America at least), is shown to be relatively common. Four partial skeletons (in addition to the holotype) are described, one of which also possesses the greater part of the appendicular skeleton. Barosaurus is closely related to Diplodocws and to the African species Gigantosaurus africanus Fraas, referred to Barosaurus by Janensch. The latter relationship is discussed briefly at the end of the paper, but the diagnosis of Barosaurzs presented here derives solely from the American species B. lentus Marsh. The three diplodocids from North America-Diplodocus, Barosaurus, and Apatosaurus-share the following postcranial characters (in contrast to CamarasaLlrLts, Bracbiosaurus, Haplocanthosdurus, etc.): (1) an increase in the number of cervicals at the base of the neck by cervicalization of the most anterior dorsals, (2) a deep, "V"-shaped cleft in the presacral spines in the shoulder region, (3) vertebral spines very high and slender in the sacral region, (4) anterior caudals with winglike transverse processes resembling sacral ribs, (5) median and posterior caudals elongated, and a 38
whiplash at the end of a very long tail (80 + vertebrae contrasted
with the usual 50+), (6) median and posterior chevrons developing an anterior extension in a typical Diplodocus pattern, (7) a marked shortening of the forelimb, with the humero-femoral length ratio = 2/s (compared to the typical 3/ra/s), (8) relatively short metacarpals, (9) an absence of a calcaneum, (10) the presence of lobe on the rear-lower corner of the lateral face of metatarsal I, (11) the longest metatarsals III and IV (rather than II and III in Camarasaurus, Brachiosarus, etc.), and (12) an expanded distal end of ischium. Diplodocus and Barosaurus differ from Apatosaurus in (1) overall lightness of the skeleton, (2) much longer cervical vertebrae,
with much lighter cervical ribs, (3) winglike, caudal
transverse
tail, (4) possession of pleurocoels on the anterior caudal centra, (5) ventral scuipturing of the anterior and anteromedian caudal centra, (6) much greater development of the typical diplodocoid chevrons, (7) pronounced lightness of the limb bones, particularly in contrast to the very robust forelimbs in Apatosaurus, and (8) possible retention of two carpals in contrast with one in Apatosaurus. The characters that separate Barosaurus from Diplodocus are derived almost totally from the vertebrae. The cervicalization of the presacrals, which results in fifteen cervicals and ten dorsals in Approcesses extending much further back in the
atosdurus and Diplodocws, reached its extreme in Barosaurus, where the number of dorsals was nine (or possibly eight). The cervicals are enormously elongated, fully 50% longer than those of Diplodocus in the postmedian part of the neck. The tops of the caudal spines, of which the first eight or nine bear notches in Diplodocus, are {l,at in Barosawrus. The winglike transverse processes and pleurocoels do
not persist as far back in the tail of Barosaurus. The tail itself is shorter. Although the number of caudals is unknown, the ratio of the sum of the lengths of the first twenty caudals to that of the ilium is 4.8 in Barosaurus and 5.6 in Diplodocus. Few chevrons are known to exist in Barosaurus, and their development may have been somewhat less extreme than in Diolodocus.
Introduction The genus Barosaurus (Sauropoda, Diplodocidae) was established by Marsh in 1890, but now, over one hundred years later, it is still not well understood. If reference to the genus of material from East Africa is disregarded, only one substantial, comprehensive treatment of the genus has appeared since the original descriptionnamely, Lull's 1919 memoir on the type specimen. There are six well-established Morrison sauropods, and it is widely assumed
that three were common-Camarasaurus, Diplodocus, and Apcttosdurus-and that three were rare-Haplocanthosaurus, Brachiosaurus, and Barosaurus. As will be shown, Barosaurus was far from rare, and several new skeletons will be described. lnstitutional abbreuiations. AMC-Amherst College Museuml AMNH-American Museum of Natural Historv: BMNH-The The Genus Barosaurus Marsh (Sauropoda, Diplodocidae)
. l9
Natural History Museum, London; CM-Carnegie Museum of
Natural History; DINO-Dinosaur National
Monumentl
DMNH-Denver Museum of Natural History; HMB-Museum fiir Naturkunde der Humboldt Universitdt Berlin; HMNS-Houston Museum of Natural Science; ROM-Royal Ontario Museum; SDSM-South Dakota School of Mines; USNM-National Museum of Natural History; UT-University of Texas; YPM-Yale Peabody Museum; and YPM-PV-Yale Peabody Museum, Princeton University Specimens.
History of Discoveries During the summer of 1889, Professor O. C. Marsh made one of his periodic trips to the 'Western United States. He visited J. B. Hatcher, who was collecting spectacular Triceratops skulls from the Upper Cretaceous in the Lance Creek area in Wyoming. Hatcher had learned of the discovery of dinosaur remains in the Black Hills of South Dakota, a hundred miles to the northeast. Marsh and Hatcher proceeded to the site, located one-half mile east of Piedmont, where they spent several days collecting part of the tail of a skeieton that was discovered by Mrs. E. R. Ellerman on land owned by Mrs. Rachel Hatch. Before leaving, Marsh obtained the promise of these two women to protect the remainder of the skeleton until he could send someone to collect the rest of it. Two boxes (Yale Accession number ), containing six caudal ver-
tebrae and a chevron, arrived in New Haven on November 4, 1889. The bones were immediately prepared for study so that Marsh could include a brief description in a paper submitted on Decenrber 21, 1889 (Marsh 1890). The specimen was assigned to a new genus and species of sauropod, Bctroslurus lentus. Marsh noted that it resembled Diplodocus, but differed from that animal by the possession of deep pleurocoeis in the caudal centra, relatively shorter caudal centra, and a lack of anterior extension of the chevrons characteristic of Diplodoczs. As will be shown below, most of these supposed differences arose from the comparison of more anterior caudals of Barosaurlzs with posterior ones of Diplodocus.
No further attempt was made to collect the rest of the skeleton until nine years later, when Marsh sent George \X/ieland to South Dakota (Wieland 1920). Wieland arrived in Piedmont on August 10, 1898.'When he learned that Mrs. Ellerman had died in 1895, he secured the help of her daughter in retrieving fragments of the skeleton that had been carried off by various people during the intervening years. Proceeding south and west from where Marsh and 'S7ieland
Hatcher had worked, collected during the next two months seven more caudals, fragments of the sacrum, seven dorsals, four posterior cervicals, a number of ribs, several chevrons, a sternal plate, and most of the right pubis. These were sent in sixteen boxes to New Haven (Yale Accession numbers , , and ). Sent with them were a number of fragments
49 .
John S. Mclntosh
from a different site in the area referred to as possibly belonging to Barosaurus (?) II by Wieland in his letters to Marsh. Among these were fragments of a femur, tibia, coracoid, and scapula. Although later attributed bv Lull (Lull 1919) to the Barosaurzrs skeleton, these fragments almost certainly do not pertain to it, and there is no reason to believe that they even belong to Bdrosaurus. Before obtaining this new material, Marsh referred briefly to Bctrostturtts in his memoir Dinosaurs of North America (Marsh 1896), placing it in the Atlantosauridae rather than the Diplodocidae. Then in 1898, l-re visited the American Museum of Natural History and observed the partial skeleton of Diplodoczs that H. F. 'Wyoming (Osborn 1899). Osborn had collected at Como Bluff, From this visit, he r,vas able to correct some misconceptions he had of Diplodoozs based on material from Colorado. Shortly thereafter, the \X/ieland material arrived in Nelv Haven, allor,ving Marsh to write one of his last papers (Marsh 1898), in which he corrected his diagnosis of the Diplodocidae and moved Bdrosaurus into it. In his very last paper submitted several months later, which rvas devoted to dinosaur footprints collected in the Black Hills, Marsh included a single sentence stating, "\Vith these remains [the rest of the Barosaurzs skeleton collected by Wieland] were found remains of a much smaller species, which may be called Barosaurus affinis" (Marsh 1899,228). His death less than a month later precluded any further statement on this species. It was not until twenty 1-ears after Marsh's death that the Barosdurus skeleton was fully prepared for study and description by R. S. Lull (Lull 1919). Lull determined that two small "metacarpals" found among the miscellaneous fragmentary material collected by Wieland must be the "smalier species" that Marsh named Barosaurus affinis.In the meantime E. Fraas (Fraas 1908) had described two sauropod specimens from East Africa as two new species of a new genus Gigantosaurus. One of these, G. africantts,
resembled Diplodocus
and was
subsequently referred to
Barosdwrus by Janensch (1922). In later papers, Janensch gave detailed descriptions of the skull parts (Janensch 1.935-1936) and bones of the appendicular skeleton (Janensch 1.961.) of several specimens he referred to this species. He also established a nerv subspecies, Barosaurus africanus gracilis for some very slender limb elements (Janensch 1961,). In another part of the rvorld, Earl Douglass was engaged by the Carnegie Museum of Natural History to excavate the magnificent quarry north of Jensen, Utah (now Dinosaur National Monument). A fine, articulated skeleton of Diplodocas, field #150 (now DMNH 7492), had been located at the end of the 1912 field season, and some disarticulated cervical vertebrae found near the dorsais were assumed to go with them. With so man,v specimens in the quarry', it was not until two years later that Douglass got around to fully uncovering these vertebrae, and he found to his amazement that they were over three feet long. This, coupled with the fact that the artic-
ulated Diplodocus cervicals extended from the front of the dorsal The Genus Barosaurus Marsh (Sauropoda, Diplodocidae)
.
41
series of #150, made it clear that the huge cen'icals belonged to a different individual, and they were assigned field #1508. In a letter to \X/. J. Holland (November 78,1974), Douglass speculated that the elongate cervicals might belong to Barosaurzs. This was a remarkable observation, because up to that time there was nothing in print to indicate that Barosaurushad a very long neck. In his annual report (March 31, 1915), Holland mentioned the "huge sauropod" but did nor try to identify it. These elongate cervicals (CM 1198) probably belong to a partial skeleton, field #155 (now ROM 3570), which was originaliy identified as Diplodocus. Several years later, Douglass made a diary entry on October 17, 1918, mentioning another long-necked sauropod (field #310), which he again suggested might be Barosaurus, but later questioned whether it might not be "Brachysaurus" [slc]. By this time, Lull had presented an oral paper attesting to the long neck of Barosaurus (Lull 1917). However, on receiving Lull's memoir (Lull 1919) on Barosaurus, Douglass wrote to Holland on February 14, 7920, confr,rming that #310 (now CM 77984) was indeed Barosaurus, and this information was published by Holland in his annual report on March 31, 1920. Surprisingly, when Gilmore published the map of the Dinosaur National Monument quarry in 1936, he misidentified the specimen as "?Brachyosaurus" (but in
all fairness it must be stated that he never saw the specimen because had not been prepared at that time). This fine specimen was finally worked out in relief by Allen McCrady in the 1970s and
it 1
980s.
In 1979, the death of Andrew Carnegie, the source of the financing of this very expensive operation, signified that quarry work would terminate soon. The easternmost part of the quarry, which was just then being worked, was producing the finest series of articulated skeletons. It was decided to take out only those skeletons needed for mounting in Pittsburgh and to leave the rest for other institutions. Among the latter group were two fine skeletons, #340 and #355, thought at the rime to be Diplodoazs. The National Museum of Natural History expressed interest in obtaining one of these, and Douglass intended to collect the other for the University of Utah, whose staff he now joined. Gilmore arrived from Washington, D.C., in the summer of 1923, and decided to collect #355. This was a splendid skeleton, but when it was discovered that the skull, cervicals, and orher parts had weathered awa5 the new director of the Carnegie Museum, Douglas Stewart, and Gilmore arranged to supplemenr the skeleton with parts from #340, including the cervicals. This placed Douglass and the University of Utah in a difficult position, but he cooperated fully with Gilmore and turned over to him the grearer part of the cervicals from #340, which the Carnegie Museum had already taken out. After the National Museum field crew had completed their work, the University of Utah party under Douglass proceeded to collect the rest of #340, which consisted of most of the dorsal vertebrae and ribs, the pelvis, the sacrum, the tail, and one hindlimb, but
42
.
John
S.
Mclnrosh
lacked the cervicals, anterior dorsals, pectoral girdle, and humerus, which had gone to $Tashington, D.C. The skeleton was prepared, but it was never mounted at the University of Utah. Ironically, when the cervicals were finally prepared in \Tashington, D.C., what had been thought to be two cervicals in the field was in reality one long cervicai. Thus the cervicals reaily belongedto Barosaurus and were thus of no use for the mounted Diplodoctts skeleton (Gilmore 1932). The latter skeleton, USNM 10865, was completed with casts of the cervicals of Diplodocus carnegii (CM 84) and went on
display in 1932.
Barnum Brown nolv entered the picture. After a trip to Salt Lake City in 1929 to evaiuate the University of Utah collection, Brown managed to arrange trades with the National Museum of Natural History, the University of Utah, and the Carnegie Museum of Naturai History (where a section of the tail had been sent previously), and so he was able to unite the Barosaurus skeleton in New York City. The American Museum of Natural History, which had done no work in the quarrl', ended up with one of its most important specimens (AMNH 6341). A cast of the skeleton was recently mounted in the American Museum of Natural History by Research Casting International. In a spectacular but controversial pose, the Barosaurus is shown rearing up on its hind legs to a height of over fifty feet, protecting its young from a marauding Allosdurus. Further Barosaurws material from South Dakota has been reported by John Foster (1996). It is now apparent that, far from being rare, Bdrosattrus is really a common Morrison sauropod. The limb bones are in many cases indistinguishable from those of Diplodocus and many of those previously identified as belonging to that genus undoubtedly belong to Barosauras. This paper will concentrate on a description of AMNH 6341, supplemented by CM 1,1.984 and other specimens. The characterization of Barosaurus is strictly limited to North American specimens. The status of the East African material will be discussed at the end. Systematic Paleontology Order Saurischia Suborder Sauropoda
Family Diplodocidae Marsh Genus Barosaurus Marsh 1890 Barosaurus lentus Marsh 1890 Synonym. Barosaurus affinis Marsh 1899 Type specimen. YPM 429: 31/z cervicals, 6+ dorsals, fragments of sacrum, 13+ caudals,T left and 2 right ribs,3 chevrons, sternal plate, right pubis. Referred specimens. AMNH 6341: cervicals 10-16 (?), dorsals 1-9, sacrals 1-5, caudals 1-29, six ribs and fragments, 1 chevron, complete pelvis, left scapula-coracoid and part of right scapula, Ieft humerus, right hindlimb and part of pes. CM 77984: cervicals The Genus Barosaurus Marsh (Sauropoda, Diplodocidae)
. {3
7-16 (?) (cervicals 4-6 have apparently been destroyed), dorsals 1-7, several ribs, left pes questionably associated. CM 1198:4 cervical vertebrae (others destroyed), likely associated with partial skeleton ROM 3670. YPM 419: left metatarsal I and a fragment of II (type specimen of so-called Barasaurus "affinis"). SDSM 25210: 2 dorsals, 15 caudals, vertebral fragments, 4 chevrons, right scapula-coracoid (incomplete), pelvis. SDSM 25331: 5 caudals. Age. Upper Jurassic, Morrison Formation. Locale. North America.
Diagnosis. Skull unknown. The vertebrae ciosely resemble those of Diplodocus but differ in (1) enormously elongated, very delicate cervicals up to 50% longer in the postmedian part of the neck; (2) the development of V-shaped, divided neural spines that commence in the middle, rather than in the anterior part of the neck, and do not continue as far back in the dorsal column; (3)the cervicalization of vertebrae in the shoulder region, reducing dorsals
to nine (or possibly eight); (4) the summits of the caudal
spines
lower in anterior part of tail and flat across top (first eight or nine are notched in Diplodoczs); (5) a proportionally shorter tail; (6) winglike transverse processes and pleurocoels that do not persist as far back in the tail; and (7) a ventral excavation of the caudal centra less pronounced than in Diplodocus. Chevrons are not well known, but the anterior process of the midcaudal chevrons appear to be less developed. The slender limb bones are virtually indistinguishable ftom Diplodocus, but the humero-femoral ratio is greater. The upper end of the scapula is very little expanded; the distal end of the ischium is exoanded.
Description Skull and Mandible
The skull and mandible have not been recovered with any Barosdurus specimens in North America. It is possible that some of the skuli materials referred to Diplodoczs actually belong to this taxon, but this possibility will be explored elsewhere. Vertebral Colttmn
The presacral vertebral formula is not yet known in Barosaurus. Because the other North American diplodocids, Drplodocus and Apatosaurus, are each known with 95% certainty to have 15 cervicals, 10 dorsals, and 1 dorso-sacral, as well as 3 primary sacrals, a caudo-sacral, and 80+ caudals, it is reasonable to assume that Barosaurzs, which is very closely allied to Diplodocus, probably had the same number of presacrals. I should point out, however, that the number of cervicals is based on one individual,
CM 84 for Diplodocus and CM 3018 for Apatosaurus.ln both trunk sections when they were found, relatively small in CM 84 but large in CM 3018. It is conceivable that a vertebra could be missing in either, although cases there was a break between the neck and
44 .
John S. Mclntosh
this appears unlikely because the vertebrae fit together so r,vell. Furthermore, in the more distantly related genus Mamenchisaurus, there are 19 cervicals and 12 dorsals and dorso-sacrals. The evidence for the vertebral count in Barosaurus is from two specimens, AMNH 6347 and CM 11984. Excluding the dorso-sacral, there are 18 presacrals preserved in the former and 19Vz in the latter, of which 17 remain. The quarry diagram for CM 11984 shows that all the vertebrae were articulated, a fact borne out by the blocks that fit together. The diagram for AMNH 6341 indicates that nine presacrais were preserved in articulation from the sacrum forward. At that point, the cervicals are separated and directed upward and
somewhat backward. The possibility that the multi-institutional collecting of this specimen might have resulted in the loss of a vertebra is dispelled b,v a letter written by Gilmore to Barnum Brorvn In this letter, Gilmore states that the Smithsonian party had disarticulated the third of the three anterior dorsals, which they collected, from the fourth and that the only break came between the
last cervical and the first dorsal. Regrettabll', the ninth presacral forward fron-r the sacrum is the one in the whole series that has suffered the most from distortion-the whole vertebra has been compressed downward and the arch rotated to the right. The eighth presacral is unquestionably a dorsal and the tenth a cervical, but lvhat is the ninth? At first glance it certainly appears to be a cervical-the parapophysis projects from the very bottom of the centrum well below the pleurocoel. However, in sauropods the cervical-to-dorsal transition is gradual, u,ith the parapophysis ascending frorn its position low on the centrum to high on the neural arch over a span of three or four vertebrae. The most abrupt change occurs in the ribs, and if these are present and articulated to their respective vertebrae it is not difficult to assign each vertebra to its place in the column. In the case of the ninth presacral, neither
rib is co-ossified to the parapophyses and diapophyses as in the cervical vertebrae anterior to it. Largely for this reason, I have concluded that it is the first dorsal. It is possible that future finds will indicate that the ninth presacral was attended by cervicai ribs and should be moved to that part of the column. \X/ith its assignment as a dorsal and the assumption of 25 presacrals (not including the dorso-sacral). there are 16 cervicals in Bdrosaurus. Ceruical uertebrae. Measurements are given in Tables 2.1 and 2.2. The two cervical series cover almost the same range. In AMNH 6341 there are cervicals 8-16 (Fig. 2.1), whereas CM 71984 has cervicals 7-16,bt the first is so poorly preserved that it has little value. CM 11984 is about 3VzTo larger than AMNH 6341. The massive distortion and/or incompleteness of the 3r/z cervicals belonging to the t.vpe YPM 429 makes their positioning difficult. The centra of Bdrosauras gradually increase in length, reaching a maximum at cervicals 13 and 14 (fourth and third predorsals). Cervicals 13 and L4 arc also the longest inDiplodocus carnegii (CM 84), but these represent the third and second predorsals in that species. In AMNH 6341, these cervicals are 3 5o/o longer The Genus Barosaurus N{arsh (Sauropoda, Diplodocidae)
.
{5
TABLE 2.1. Measurements (mm) of Barosaurus Cervicals (*' Length Cervical F
Length without
(max)
ball
21.0"
930
220
216
890
300
618
590
685
630
10
737
660
I1
775
7r5
T2
813
715
13
8s0
760
I4
865
745
15
840
731
I6
750
620
l-)
14
slightly distorted; e
:
estimated)
\7idth across Height prezyga- postzyga- diapobreadth height breadth height overall pophysis pophy'sis physis AMNH 6341 Centrum
anterior end
72 157 118 145 e 138 133 135 130" 143"
8
9
:
58
65 80 98 1
18
725 165
165*
Centrum posterror eno
130 I23 168 145 155 180 155', 160* 250
115
300
173
218
24s
135
375
205
215
305
147
390
220
245
330
165
413
230
275
330
190
450
265
283
363
230
,10(]
300
302
415
225*
s82
328
/-v 5
470
260"
61.5
303
280
500
240
657
297
300
540
273
560
YPl'4 429"
15 16
720
220 345
570
300 365 AMNH
220 200
760
6341
" Measurements from Lull 1919. These vertebrae have all suffered from crushing, often severe. Their assigned positions are uncertain. and the measurements themselves should be treated rvith extrene cauuon.
than in CM 84. On the other hand, the pelvis of AMNH 6341 rs 11% smaller than that of CM 84. Thus, based on equally sized pelves, the posterior cervicals of Bdrosdurus are nearly 50% longer than those of Diplodocus. The cervical centra in Barosaurus are all strongly opisthocoelous and all possess complicated systems of pleurocoels that dif, fer noticeably from those of Diplodocus. The general pattern is a
three-part division. There is the sharply margined pleurocoel proper, which is a deep, elongated depression sharply pointed at its anterior and posterior ends, situated just behind the overhanging diapophyses. It is subdivided in two by a vertical ridge in cervical 11 and perhaps in ali the cervicals, but the pleurocoels in these have not been prepared in sufficient detail to be certain. A somewhat similar, but even more complicated arrangement is evident in the posterior cervicals of Diplodoczs (CM 84). In Barosaurus, the lower margin of the pleurocoel proper is equaily divided by two sharp ridges of bone. Beneath the posterior ridge there is a broad depression covering much of the lower part of the posterior half of the centrum. A less well-defined deoression lies under the front
46 .
John S. Mclntosh
TABLE 2.2. Length without Anterior Ball of Barosautus Presacral Centra, CM 11984 (Figure 10.1)
cervical 7
585
cervical 16
654"
cervical
6200
dorsal
1
500 "
cervical 9
618
dorsal 2
401 *
cervical 10
dorsal
261
cervical 12
659" 689" 760"
cervical 13
857*
cervicai 14
800*
cervical 15
720
cervical
8
11
3
dorsal 4
228
dorsal 5
225
dorsal dorsal
240
6 ./
" These vertebrae have been partially rvorked out in relief in their blocks, and the blocks have not been put together, so accurate measufements are not possible. " - sum of measurements for vertebrae in separate blocks; should be rierved uirh oarrieuhr c.rution.
---'i'
-
,l'
*---'-.:,
-'': l\
{
w t B
Fig. 2.1. Barosaurus lentus,
AMNH 6311. (A-C) Ceruical t crtebra,'8 to l o. left sidc uieu s tiiilr,:
(or right side reuersed). Scale bar: 70 ctn.
ridge, and
it
extends beneath the cervical rib. The laminae extend-
ing to the zygapophyses and diapophyses are generally similar to those of Diplodocus. The neural spine of cervical 8 is flat across the top, and that of cervical 9 shows the first trace of a divided spine (FiS. 2.2A. This division increases gradually in sequential vertebrae, being moderately developed in cervicals 12 and 13, and as a deep V-shape in cervicals 15 and 16. This development is in sharp contrast to Diplodoclls, where cervical 3 already shows the first The Genus Barosaurus Marsh (Sauropoda, Diplodocidae)
' {7
'?'
ry?
A &.,. Fig. 2.2. (AJ Barosaurus lentus AMNH 5311 ceruical uertebra 8 in antcrior. le[t side. and pos!erior t iews. rB) Dipludur u' cerrregii CM 81 ceruical 8 in anterior, Ieft sitle, and posterior uieu,s (front
Httcher
1c)01).
t-
't
B trace of division, and where the division is already quite deep in cervical T (Fig. 2.28).By cervical 11 it is as well developed as cervical 16 of Barosaurus (AMNH 6341; Fig. 2.3A). In the spines, a further difference is that those of the last two cervicals (14 and 15) of Diplodocus project anterodorsally, whereas those of Barosaurus are ail directed dorsally (Fig. 2.3B). The cervical ribs are firmly coalesced to all the cervicals, and as
far as can be determined are indistinguishable from those of Diplodocus (the shafts have been severely damaged in AMNH 6347). An anterior extension of the rib can be seen in cervical 15 (AMNH 6341) and also in CM11984 (Fig. 2.4A), where some of the ribs of the more anterior cervicais are well preserved. The straight, slender rib does not extend beyond the posterior margin of the centrum as in Diplodocus, Apatosaurus, Dicraeosaurus, and Haplocanthosaurus, which is in sharp contrast to Bracbiosaurus, Eub elop us, and Cam ar as Aurus.
Dorsdl uertebrae. Measurements are given in Table 2.3 All the dorsals of AMNH 634L were articulated to one another and to the sacrum. In CM 11984 dorsals 1-7 were found articulated with one another and with the iast cervical. Assuming ten dorsals as in Diplodocus, Lull (1919) determined that those preserved in YPM 429 were dorsals 1, 4, 5; spine of 6,7, 9; and 10. Accepting nine as the correct number of dorsals, I would now identify these as dor-
2,4,5; spine of 7,8,9; and the dorsosacral. All of these have suffered from various distortions and mutilations (Fig. 2.5). The dorsals are very lighdy constructed and bear a strong resemblance to those of Diplodoczs. Dorsal 1 (presacral 9) is by far the longest. As stated above, it conforms in every way to a cervical, except that the ribs are not coalesced to it; they were probably thoracic. In AMNH 634I the vertebra is comolete but crushed downward, the arch having undergone torsion. Taken together with the sals
43 .
John S. Mclntosh
:f
t
L\ \
&
i
,,,.
,8s'
lxS & lJt
!E
Lig. 2.3. iA) Barosaurus lentus
AMNH 6317 ceruical uertebra
13
in rigbt side (reuersed) and dnterior uieu's. (B) CM 84 Diplodocus carnegii ceruical 13 in leit side dnd anterior uietL,s. (Some data from Hatcher 1901 .)
Iig. 2.4, Barosaurus lentus CM 11981. (A) Ceruical uertebra number 9 with complete rib, and (E;t dorsal tn rtSnl sldc dila pa)stertor uteL's.
The Genus Barosaurus Marsh (Sauropoda, Diplodocidae)
' {9
TABLE 2.3. Measurements (mm) of Barosauru.s Dorsals
Length
Cervical Length witliout
(max) ball
#
s65 422 332 265 280
1
2 3
4 5
)49, 265 z/) 270
6
7 8
9
5
470 260 310
d
270"
9
212
2
4
Dorso-
Centrum end
rWidth across Centrum end Height pr.r@ breadth height breadth height overall pophysis pophyiis
anterior
posterior
AMNH 500
290
310
278
250
2s5
zrs
LJ \)
240
235
))
s
248
243
220
/-.t)
220
215
2r6
328
165
290
220
320
200
240+
260
260
265
123 r82 160+ 215 220 209 215 212 248
physis
6341
34s 322 307 270 285 264 254 249 240
235 223
245
230 235 240 255 255 272
480* 640 723 674+ 778 828 865 906 901
377 413
378 386 300
438
285
485
200
467
169
290
465
720 720 730
177
464
150
472
147
370
YPM 429"
245
350 330 320 240" 30s 290
250 23s 2ss 280 280 280
590* 760
530 26s
460
*t
2;
300
840 800 850
240
sacral
" Measuremenrs from Lull 1919
moderate latero-medial crushing of cervical 16, this produces marked changes in the measurements from one vertebra to the next. Most striking is the total height, which is enhanced in cervical 16 and greatly reduced in dorsal 1. The dorsal centra, all of which bear prominent pleurocoels, decrease in length rapidly from cervical 15 to dorsal 4, at which point the length remains continuous. The ratio of the centrum length of presacral 10 to presac ral 6 is 2.8 in AMNH 6341, but only 2.0 in Diplodocrzs CM 84. The parapophysis is at the very bottom of the centrum on dorsal 1. It is slightly higher on dorsal 2, but it is still below the pleurocoel. On dorsal 3 it is directly in front of the pleurocoel, and on dorsal 4 it is just above the border of the centrum and arch. On dorsal 5 it is well up on the arch, where it is on dorsals 6-9. The first three dorsals are strongly opisthocoelous. The anterior ball becomes progressively less prominent on 4 and 5, and the last four dorsal centra are virtually plano-concave. As with the cervicals, the complicated laminae on the arch extending to the zygapophyses and diapophyses do not differ from those of Diplodocus (Hatcher 1901). A
59
.
John S. Mclntosh
TABLE 2.4. Measurements of the Barasaurtts Sacrum. AMNH 6341
Number
1
I 3 4 5
Heieht of Centrum Proximal Proximal Distal breadth height breadth prezygapoph,vses length 185
23-
455
229
'!flidth
across
diapophyses I q)
r80 t35 r80
205
lez
hyposphene-hypantrum articulation is present between dorsals 4 and 5, as well as the remaining dorsals. The V-shaped cleavage of the neural spine is quite deep in dorsal 1, gradually decreasingin2 and 3 and then rapidly in 4 and 5 (presacrals 6 and 5). A trace remains in dorsals 6 and7, but the tops of the spines of 8 and 9 are virtually flat. In CM 11984, the notch is still evident on dorsal 6 but has disappeared on 7 and also on the detached spine tentatively identified as that of 7 in YPM 429. A small, secondary median spine at the base of the cleft, shown by Hatcher (1901) to be present in the first three dorsals of Diplodocus carnegii CM 84, is also present in the first three dorsals of AMNH 6341 and can be traced forward in this specimen to the last two cervicals as well. These spines may serve as a rear anchor for the attachment of the long ligament that extends forward to the
skull (Colbert 1.961). From dorsal 5 posteriorly, the spines increase in length, with those of the last two dorsals being about the same height. The spine of the last dorsal is inclined forward as in Diplodocas (Gilmore 7932), but not to the same degree. Sacrwm. Measurements are given in Table 2.4. As in the other Morrison sauropods, Barosaurus possessed five functioning sacrals: a dorso-sacral; three primary sacrals and a caudo-sacral, or probably more correctln as Hatcher (1901) noted, a dorso-sacral; two primary sacrals; and two caudo-sacrals. However in all adult sauropods, the anterior most of these caudo-sacrals has become so modified as to appear more like a true sacral. In AMNH 6341,,the five centra and the heavy costal yoke on both sides are preserved, but the arches and diapophyses are largely restored and parts of the spines are missing. Lull (1919) noted that in the type specimen YP,l4 429 all that remained of the sacrum was part of one centrum and the co-ossified summits of the spines of the three primary sacrals. To this may be added the vertebra, which he described as dorsal 10, but which is most likely the dorsosacral. As far as can be seen from its state of preservation and preparation, the sacrum of AMNH 6341, drffers from that of Diplodocus The Genus Barosaurus Marsh (Sauropoda, Diplodocidae)
.
J1
,.{ I
ttlt,
":,
\
'.rT*
\.T "-
v
w
t,9.,.,,
% fu"-
"\,,,. :"jl,i
ii,-l
"f' ,r;"1,,i;
D
C!,,
* -#tr{'
E
Fig.2.5. Barosaurus lentus AMNH 6341 dorsdl uertebrae 1 to 9 (A-l) in anterior, lc[t sid,', aud p, 'sterior uiews.
52 .
John S. Mclntosh
'3
,$,
1
:t!
'w. ,.b
etr
,q\
I lrii, lil 1. :
,.]
The Genus Barosaurus Marsh (Sauropoda, Diplodocidae)
.
J3
only in minor respects. The anterior centrum face of sacral 1 is flat. The parapophysis or capitular facet for the modified thoracic rib is high on the anterior part of the lateral face of the centrum, in front of the large pleurocoel. The cenrra of sacrals 2-5 are firmly coossified, but the dorso-sacral is not in AMNH 6347: rt also is not in YPM 429 if the identification of that element is correct. The rib of sacral 1 is restored on the left side, as is the middle part of the right rib. It does not appear to send a lower spur to the pubic peduncle of the ilium as in Diplodocus (CM 94), but this may be due to incompleteness. The sacral spines of AMNH 6341 are damaged and have been restored. The overall heights of the dorso-sacral and first primary sacral (sacral 2) are about the same height as that of the last dorsal. Cdudal uertebrae. Measurements are given in Table 2.5. The caudals of AMNH 6341 were almost completely articulated. By comparison with the tail of Diplodocus, Lull (1979) determinec that the caudals of YPM 429 represented caudals 2-6, 1.3,15-1.7, 19-20; fragments of 23, 25 , 28, 32 and distal caudals 42 and 70 . \X/ith the articulated caudals of Barosaurzs available for AMNH 6341,I now believe that the five anterior caudals of YPM 429 are more likely caudals 1-5, 10, 1,2, 14,15,17,18; I shall not attempt to position the incomplete and the posterior ones. LuJl (1919) has described the anterior caudals in great detail. The anterior caudals resemble those of Diplodocws in that the centra have deep pleurocoels and are excavared ventrally. They are very slightly procoelous, less so than in Diplodocus. The anterior surfaces are mildly concave, the rear ones nearly flat or slightl,v convex. The transverse processes (more correctly caudal ribs) are of the winglike type, and are very similar ro the sacrals ribs characteristic of Diplodocus and Dicraeosaursus. These winglike ribs are only present in the first three or four caudals of Apatctsaurws (Ftg. 2.6A). The transition to the more normal type of transverse process occurs more rapidly in Barosaurus than rn Diplodoczs. By caudal 5, they have proceeded down onto rhe centrum and by caudal 7 have assumed almost the normal sauropod form, projecting from just anterior of the pleurocoel (Fig. 2.68). In contrasr, the winglike development is still evident on caudal 12 in Diplodoczzs, where the transverse process arises on the arch and extends down only as far as the border of the arch and centrum. The transverse process disappears on caudal 18 (AMNH 223) or 19 (USNM 10865), whereas it disappears by caudal 15 of Barosaurus AMNIH 6341. The pleurocoel also persists further back in Diplodocus, to caudal
18 (AMNH 223), caudal 19 (USNM 10865), or caudal 16 (DMNH 1.494); the last trace of a rrue pleurocoel in Barosaurus (AMNH 6341) is caudai 14 (Fig. 2.6C). A striking difference between the caudal vertebrae of Barosaurus and Diplodocus is evident in the degree of ventral sculpturing. In Barosaurus, this begins as a prominent, squarish pit in the anterior caudals (Fig. 2.6G) but soon becomes a broad,
rounded concavity covering the entire ventral surface of the cen-
54
.
John S. Mclntosh
TABLE 2.5. Measurements (mm) of Barosaurus Caudals
AMNH
6341
Centrum
Number length
Centrum
anterior face
breadth height
1
153
300
2
175
280
3
160
320
4
195
295
5
183
290
6
190
275
7
190
288
8
200
283
9
206
270
10
227
279
11
227
283
12
250
265
13
266
245
I4
263
235 +
15
263
230
t6
276
228
1a
278
209
18
195
I9
283 279
20
268
170
21
268
163
22
260
149
L.)
248 240
r13
24 25
220
125
LO
21.3
120
27
198
110
28
182
94
29
17l
9I
186
139
283 27s 270 252 247 256 236 223 22r 214 212 215 205 274 202 180 165 1.64 152 133 130 1.27 123 122 93 98 82 80 70
I
185
345
265
2
220
340*
230
3
210
370
345
.+
210
355
235
5
205
325
200
10
245
250
208
12
245
240
215
1.4
270 270
2t5
203
220
180
15
(" -
slightly distorted)
Centrum
face Overall Postzygapophysis '!ilidth across breadth height height to top of spine postzygapophyses AMNH 6341 posterior
26s 290 278 275 259 268 273 253 278 280 263 253 250 238 224 228 195 188 179 775 748 151 140 127 109+ 118 110 98 85
320 340 330 315 230 23s 230 270
29s 235
126
2s2
133
230
660
490
180
240
463
21.5
228
590 498
41s
178
225
520
446
t65
21.9
505
452
143
21.9
439
458
t)A
21s
433
390
135
189+
396
240
415 +
209
448
11)
113
r95
408
390
98
189
375
393
115
172
185 169
298
381
75
268
380
71
153 149
87 78
-/ /1
70
743 113
233
108
210
322
197
258+
76
132
185
48
57
113
186
42
195
670
180
185
200
640
220
215
585
245
155
109 109
65 63
60
99 85
49
71.+
YPM429230
185
200
21s
443
205
422
200 180
360
The Genus Barosaurus Marsh (Sauropoda, Diplodocidae)
.
JJ
Centrum
Number t7
Centrum length
285
anterior
200
a
b o'7
d
Centrum
face overall
posterior
breadth height breadth height
18
c
face
il3 77 47
e
6+
f
hejght
155 210 170 183 142 135 |]3 e- il92 70 90 70 38 38 43 21 12
Postzygapophysis \(idth across to top of spine postzygapophyses
310*
53
20 17
" These vertebrae are not available for re-measuring at this time. I have therefore used those made by C. C. I{ook in 191.7 for caudals 1 through 17. These differ slightlv from those published bv Lull in 1919. Lull's measurements were used for 18, d, and f, and all others were made by me. The anterior and posterior portions of caudals "b" and "c" do not connect and may belong to seDarate vertebrae.
g
ff* n';:' &"'
r'
.!
iri'rr':f
i
.v{,,,',.
4&,,,,'
\k's
f-
e;"-
Afli-i t
B
,.
* ""
-rc#t*q--
{9
.;
,;tot
i
''"
J,'"*"".S",,
J F
F ,rr,.;
"
t1;;.4. ril
Fig. 2.6. Barosaurus lentus uertebrae itt Ieft side uieus (A-F), dnd caudal
AMNH 5311 cattdal 5, uentral uiew (G).
56
.
John S. Mclntosh
tl
",|
H. .
trum. It persists throughout the entire series of 29 vertebrae in AMNH 6341, becoming gradually shallower until, on number 29,the ventrai surface has become almost flat. In Diplodocus, the ventral sculpturing is much deepeq particularly in caudals 1-20 (AMNH 223).In addition, Diplodocus caudals are greatly elongated, whereas those of Barosaurus are somewhat less so (Fig. 2.6D-F). Those of Apatosaurus are elongated still less so but are considerably longer than those in most sauropods, for example, Camarasaurus, and to an even greater degree in Haplocanthosaurus and Brachiosaurus. Another noticeable difference between Barosaurus caudals and those of Diplodoczs involves the height of the spines in the anterior part of the tail. In BarosAurus, the overall height of the spine in the few anterior caudals is almost twice the diameter of the centrum, whereas in Diplodocus it is almost three times as great. A less striking difference is seen in the broader transverse diameter of the spine in the most anterior caudals of Barosauras. Also, the caudal spines in Barosaurus show no trace of a cleft, whereas rn Diplodoczs a small, but definite cleavage is evident as far back as caudal 8 (CM 84) or caudal 7 (AMNH 223). A final difference is in the greater anteroposterior breadth of the spine in Barosaurzs beginning at caudal 16 or caudal 17 posteriorly. No caudals were found with AMNH 6341 posterior to caudal 29. but Lull (1.9L9) believed he found one, which he identified as caudai 32. If Holland's (1906) estimate of the positions of the caudals in Diplodocus (CM 307) is correct, the position of the YPM caudal is probably in the 40s. Nothing need be added to Lull's (1919) description of it except to state that it is relatively shorter than its counterpart rn Diplodours. Another fragment found with YPM 429 suggested the presence of a distal whiplash as in Diplodocus and Apatosaurus. To summarize, there are more similarities between the caudais of Barosaurus and Diplodocus than there are between those of any other sauropod genera. However, the specialization is more advanced in Diplodoozs in that it had a longer tail; higher neura spines in the sacral region; greater development of the anterior, winglike transverse processes; longer persistence of the pleurocoels and transverse processes; and more pronounced procoelous ante-
rior centra. Thoracic Ribs
A number of ribs, none complete, were found with both YPM 429 (Fig.2.7A) and AMNH 6341. These ribs resemble those figured by Gilmore (1936) for Apatosaurus ICM 3018). The proximal end likely belonging to rib 1 (?) shows the wide separation of the head and tuberculum. If it is truly rib 1 and not rib 2, there is a difference from that of Apatosaurus in that the head continues in a direct line from the shaft and is not directed outward. A similar fearure is noted in what is probably the same rib from the right side of
The Genus Barosaurus Marsh (Sauropoda, Diplodocidae)
.
57
AMNH 6341. Further description of the ribs in their
present frag-
mentary state would not be useful. Cheurons The two most anterior chevrons of the type determined by Luli (1919) to be chevron 5 and chevron 9 are typical anterior sauropod chevrons (Fig.2.78, C). They have been described by Lull (1919), and no further comment need be made except to note that the
hemal canal is strongly bridged over as is usual in the diplodocids, and in contrast with, for example, Camarasaurzs. The third chevron is from the transition region in which the anterior projection typical of Diplodocrzs has begun to form (Fig. 2.7D). Lull
it as about chevron 16, but rf Diplodocus (AMNH 223\ is taken as the guide, it is more probably chevron 13. However, it is unclear tf Diplodoclzs should be used as a guide based on one side of a more posterior chevron preserved with AMNH 6347 (FiS. 2.7E).Its position is not known, but it was collected from near caudals 22-28.It resembles most closely chevron 18 of Diplodoczs (AMNH 223) and has a fully developed anterior process. However, it is only about2/z as long (about 200 mm long when complete). This shortening is in keeping with the general trend noted in the caudals of the two genera. This chevron is important because it clearly demonstrates the chevrons in the mid-tail region of Barosaurlrs were of the typical Diplodocus-type, al(1.919) identified
though shorter, and that they were much more advanced than those
.\ril,i
ffia
W%
*.,w \
,,'1
'\
'.,
\ ,,'t
r$\ ,w , l'!i.
e$j ts'
p
Fig. 2.7. Barosaurus lentts (A) left ribs of YPM 129; (B-D) cheurons
of YPM; (E) cheuron of AMNH 6311.
53 .
John S. Mclnrosh
,ff f
.t
ffi, D
E
':,
"-4t
*'.."-*tt '*-:":
TABLE 2.6. Measurements (mm) of Barosaurus Pectoral Arch,
AMNH
63,11
Length of scapula, on the curve Length of scapula, in a straight line
1300
1240
Breadth of distal end Least breadth of shaft
375
Length of coracoid
297
\(/idth of
420
same
195
of Apatosaurzs (AMNH 339), where Diplodocus-like chevrons are present but smaller.
Pectoral Girdle
Scapula and cordcoid. The right scapula and coracoid of AMNH 634L are complete except for a portion of the anterior edge of the broad plate that forms the proximal end of the scapula. They are firmly co-ossified and resemble those of Diplodocus (Fig. 2.8A, B). The scapular blade, directed upward and backward, expands gently and continuousl,v toward the distal end, the rear border flaring slightly just below the end, but the erpansion is less than in Diltlodoozs (CM 94, AMNH 223). Except for the upper expansion and a small bulge at midlength, the posterior border is straight until it flares out ventrall,v. As in Diplodocus, the ridge on the lateral surface of the proximal plate between the great muscle fossa and the superior fossa makes an acute angle -40', contrasted with -60" in Cdmarasaurus and Apdtosaurws (measurements in this secrion refer to figures in Table 2.6). Otherwise, the scapula resembles that of the latter quite closely. The coracoid is not distinctive. It is a squarish bone, rounded at the corners) thickest at the posterior border where it forms part of the glenoid fossa for the humerus. The fully enclosed circular foramen is located below the scapular
articulation. Sterndl Plates
Lull's (1919) thorough description of the sternal plate need not be repeated, except to note that the anterior and posterior ends \\.ere reversed by Lull (19L9). Forelimb. Measurements are given in Tables 2.7-2.9. A right manus, field #310/L, was found near the junction of the cervicals and dorsals of CM 11984 and was given the same field number by Douglass. It is possible that it does belong to Bdrosaurezs, but I believe that unlikely. The manus is indistinguishable from that of ApJtosaurus and is the exact size of the left manus of Apatosaurus Iouisae (CM 3018) found about 21 m west of CM 11984. CM The Genus Barosaurus Marsh (Sauropoda, Diplodocidae)
.
59
'qr\.
,\
t\
s,* 'p ,,
'e
,,
".{tr :l{
c
A
ffi,'"
D ir..-
B
q
'",
\{ i'$
t$ .i,.y
E
frg. 1.8. Barosaurus lentus: /A-B)
t
sc.rPttla/corcoid in posterior L1l1Ll side uietus AMNH 5311; r tel:
tC-Dl right humerus in antertor and obliqtte uiews AMNH 5311; (E) carpal, top uiew CM 21711; (F) right ulna, radius, and iltet)(arpols I !o lV in anlerior uieu'.
69 .
John S. Mclntosh
TABLE 2.7. Ratio of Humerus: Femur Length in Diplodocids
Barosaurus lentus AMNH 6341,
.72
Diplodocus sp. AMNH 5855 Diplodocus "longus" USNM 10865 Diplodocus hayiHMS 175 (formerly CM 662) Gigantosaurus africanus HMB, quarry A Gigantosaurus africanus HMB, quarry k
.67 .67 .64 .72 .72
TABLE 2.8. Comparative Measurements (mm) of Diplodocid Forelimbs. Ratios given in ( ) are relative to bone length (100). Circumference is minimal shaft. Barosaurus
lentus
AMNH
CM
Diplodocus
"longus"
USNM 10865
Diplodocus hayi
HMNS 175
Gigantosaurus robustus
HMB A HMB K right left left right Lt / /a (100) 970 (100) (100) (100) 990 (100) 936 (100) 910 (-) 965" 1032 (100) Humerus, 1034 Iength (-) 429 147) 400 (43) 440 i44) 445 147) (-) 382 (37) 340 (33) Breadth, proximal (-) 162 1161 150 (16) 173 (19) 188 (20) r70 (17) 1s0 (1-s) 151 (15) Breadth, shaft - (-) (32) 328 (34) 265 (,29) 317 B4) 330 (33) 320 (33) 333 278 250 Breadth, distal 124) (45) 438 (,+5) 449 119) 470 (50) 502 (51) 430 144]) (-) 460 440 circumference 143) (-) 740 (100) (100) 720 (100) 728 (100) (100) (-) 735 940 ulna, length - (-) - (-) 210 Q6) 240 (33) 22S (33) 797 \27J - (-) (-) 255 (34) Breadth, 6341
proximal
shaft Breadth distal Breadth,
Circunrference Radius, length Breadth, prorimal Breadth, shaft Breadth, distal Circumference -
(-) (-) (-) (-) (-) (-) (-) (-)
(-) - (14) 302 (32) 930 (100) 117 (13)
\14) (18) 29-5 (40) (-) - (-) (-) 77 \8) - (-) 147 \16) - (-) 219 (27) -
139
102
129
(16) 83 (11) - (-) (22) 130 (18) - (-) (38) - (-) - (-) 275 632 (100) 686 (100) 660 (100) 148 (22J 151' 122) 7s7 124) 113
159
(13) 1,48 (22) 254 137) 88
83 (12) 106 157 123) 160 233 (34) 265
-
(-) (-) (-) (-) (-)
116) - (-) 124) - (-) \10) - (-)
1s5 (21)
21s
129)
249 (34)
(-) - (-) (-) - (-) - (-) -
" Extreme proximal end restored, measurement, probably too [ow.
3018 is the largest Apatosaurus, indeed the largest sauropod in the entire quarry at Dinosaur National Monument. I think that this specimen is the missing right manus of that otherwise almost complete skeleton, which was washed downstream from the Apatosaurus. Another specimen, CM 27744 (field #312)' comprising
part of the right forelirnb and foot, found 4 m west of the Barosaurus skeleton (CM 11984), may belong to that specimen (Mclntosh 1981). The iimb bones are very long and slender but cannot belong to Brachiosaurus or Camarasaurus because the ratio of the length of metacarpal II to that of rhe ulna is only 0.31 compared to 0.49 in Brachiosaurus and 0.44 in Camarasaurzs,' 0.31 is a reasonable ratio for a diplodocid, but the limb is immediately excluded from Apatosaurus whose forelimb is much more robust. The metacarpals are also much more slender than those of the rela-
tively stout-limbed Diplodocus Daf i (HMNS 175' formerly cM 662: see also Bedell and Trexler, this volume). The metacarpals are also more slender than those of AMNH 380, which are referred by Osborn and Granger (1901) to Diplodocus? and probably belong The Genus Barosaurus Marsh (Sauropoda, Diplodocidae)
'
(1
TABLE 2.9. Comparative Measurements (mm) of the Sauropod Manus (e
Element
Barosaurus lentus
hayi
cM
HMNS 175
Ulna Radius
Metacarpal I Metacarpal II Metacarpal III Metacarpal IV Metacarpal V Metacarpal II: ulna ratio Phalanr I-1
21774
:
estimate)
Diplodocus
940
728
sp. AN,INH 707
380
Camarasaurus grandis
sp.
AMC
695
840
AMNH
823
AMNH
573
716
930
686
645
810
549
672
230 distal
170
187
/_+ 5
213
288
293
212
226
260
247
322
225
./-l1')
242
-)
215
/-+
\
))4
300
965
L-)
257
191 182
188
220
202
)1n
0.31
0.29
0.32
0.31
0.42
0.45
54
Phalanx I-2 Phalanx III-1
58
Phaianx IV-1
48
Phalanx V-1 Phalanx V-2
aa .)L
78
to that animal. Interestingly, the limb of CM 21774 ls 28% longer than that of a large Diplodocws (USNM 10865) from the Carnegie Quarry. Since I think it plausible that CM 21774 does belong to Barosaurus, if not to CM 11984 itself, I shall describe it as such be1ow.
Humerus. The humerus of AMNH 6341 is a long, straight, slender bone whose shaft is not twisted (Fig. 2.8C, D). It is expanded at both ends, more so at the upper end, but nor ro the degree seen rn Camarasaurus er, particularly, rn Apdtosaurzs. As in most sauropod humeri, the radial and ulnar condyles are weakly developed. A prominent deltoid crest extends down the lateral margin of the bone from near the proximal end to just above half the length, where it ends abruptly. The bone overall is virtually indistinguishable from that of Diplodocus, and there can be little doubt that isolated humeri belonging to Barosaurushave been attributed to Diplodoczs. The ratio of its length to rhar of the femur is greater than in Diplodocws (see Table 2.7). The presence of a relatively longer forelimb than in Diplodocus is not unexpected in light of the longer neck and shorter tail of Barosaurus. ulna. The ulna of CM 21774 is very long and slender (Fig. 2.8F). The lower third is gently bent backward. The rwo processes on the proximal end, which cradle the proximal end of the radius, are, like Diplodocus,less prominenr than in most other sauropods.
62
.
John S. Mclntosh
Radius. The radius is very long and slender (Fig. 2'8F). Its proximal end is squarish with blunted corners. The lower half of the bone is slightly flattened latero-medially and expanded in this direction. It gently bends toward the ulna. Manus. The preserved parts of the right manus of CN{ 2I744 (Fig. 2.8F) include one large carpal; metacarpals I, II, and IV and the upper half of III; phaianges I-1 and parts of I-2,1V-1', V-1, and V-2. The carpal, which is often called the radiale in sauropods but is more likelv the first carpal of the second row, is crushed against the upper end of metacarpal I (Fig, 2.8F). It is massive and generally circular, covering the proximal ends of metacarpals I and II. It measures 159 mm by 135 mm by 48 mm. The metacarpals almost approach those of Camdrasaurzzs in their slenderness, but their length ratios with the ulna and radius are typical of the Diplodocidae. They are more slender than those of Apatosaurus or Diplodocus hayi and to a lesser degree to the two mani with accompanying radius and ulna, from Bone Cabin Quarrn \fyoming, which probably belong to Diplodocrzs (AMNH 380 and 695, the latter now AMC 658). Recently, an associated forelimb and manus from the Black Hills of South Dakota, SDSM 25277, was described by Foster \1996), and referred with a query to Barosdurus or perhaps Diplodocus. The proportions of the elements of this well-preserved limb to those of CM21774 indicate a somewhat more robust limb. Peluis
In AMNH 6341', the right ilium is complete except for a small part of the upper border; of the ieft ilium, only the acetabular portion was preserved. Both pubes and ischia were complete except for portions of the proximal ends of the pubes and the right ischium' The greater portion of the right pubis of the type YPM 429 lacked the disral end and part of the head (measurements are given in Table 2.10).
Iliwm. The most notable feature of the ilium, which is virtually indistinguishable from that of Diplodocus, is its long, slender public peduncle (Fig. 2.9A-C). As in the latter genus, the outwardly directed anterior lobe is a iittle larger than that of Apatosaurus. Pubis. The pubis, with its relatively slender shaft, likewise re-
that of Diplodocus. The circular pubic foramen opens down"vard from the lateral face. It is open in both pubes of AMNH 6341 but closed in those of YPM 429, and it is probably ontogenetic. The most prominent feature of the Diplodocus pubis is the strong development of the process for the attachment of the ambiens muscle on the anterior margin of the upper end of rhe bone. A hook-shaped process develops even in juveniles, as shown by a specimen on the face at Dinosaur National Monument (DINO 3783). A similar deveiopment is seen in the African genus Dicraeosaurus and also in Gigantosaurus africanus, but not in Apatosdurus. On the right pubis of AMNH 6341 most of the ambiens process has eroded away (Fig.2.10A). The upper end of the sembles
The Genus Barosaurus Marsh (Sauropoda, Diplodocidae)
'
(3
TABLE 2.10. Measurement (mm) of the Barosaurus Pelvic Bones, AMNH 6347 G: estimate) Length of ilium Length of pubic peduncie Diameter of acetabulum
940 285 310
Breadth across anrerior tips of both ilia Length of pubis
1010e 890
Extent of ischiac articularrons
315
Least breadth of shaft Breadth of distai end
220
155
Length of ischium Least breadth of shaft Breadth of distal end
219
Thickness of distal end
80
873 707
K.'ir.:
,
r,,x
qt:
{
Fig. 2.9. Barosaurus lentus
AMNH 6311 ilia and parrial in (A) right side, (B) uentral, and (C) anterior uiews.
sacTum,
64
.
John S. Mclntosh
B
A
t
i!i:
$
,,*-'ta""r,l. Fig. 2.10. Barosaurus lentus 6341 all lateral uieus. Left (A) and right (B) Pubis; YPM
AMNH
D
429 rigbt pubis (c): AMNH 6341 left iscbium (D).
left one is also incomplete but shows the beginnings of the process (Fig. 2.10B). The type pubis YPM 429 has a prominent process) but even here it is not complete, so it is not certain that a true hook was present (Fig. 2.10C). Lull's (1'91'9) estimated length of the incomplete pubis of YPM 429 is grossly inflated because he estimated the length from Hatcher's figure of the pubis of Diplodocus (CM 94), which was drawn at the oblique angle as the bone appears in an articulated pelvis. This reduced considerablv the apparent breadth of the shaft as determined by Lull. It seems more likely that the pubes of YPM 429 were about the same as those of
AMNH 6341. Ischium. The ischia also closeiy resemble those of Diplodocus, with their expanded distal extremities, which in AMNH 634L were co-ossified, a common occurrence in adult members of the Diplodocidae (Fig. 2.10D). The distal expansion is similar to that in Diplodocrzs, but less exaggerated than that in Apatosaurus (Ftg. 2.11). The manner in which the distal ends abut one another is in sharp contrast to the edge-to-edge articulation of the unexpanded distal ends in Camarasaurus and Brdchiosaurus' The Genus Barosaurus Marsh (Sauropoda, Diplodocidae)
' (5
Fig. 2.11. Distal ends of ischid.
/A/ Barosaurus lentus AMNH 6311 ; (B ) "Gigantosaurus" africanus k 14 and 4.t; (C) BMNH
M
specimen; /D/ Camarasaurus supremus
AMNH 5761; (E)
Dipodocus cargegii CM 81 (F) CM 91; (G) CM 91; (H) D.Iongus USNM 10855; {/ Apatosaurus ajax (Atlantosaurus immanis/r YPM 1810; /// A. excelsus YPM 1980; (K) A. louisae CM 3018.
Hindlintbs The hindlimb of AMNH 6341 is beautifully preserved and uncrushed (Fi9.2.12) and is exrremely similar ro those of Diplodocus. This similarity is most evidenr with AMNH 5855. an isolated specimen comprising all rhe limb bones from both sides and several qir-
dle bones, but no verrebrae. Ir was described by Mook 11917) as Diplodocws sp. but the similarities are so striking that I would have referred it unhesitatingly to Barosaurus were there not a striking difference in the humero-femoral ratio: 0.72 tn Barosaurus and 0.66 in AMNH 5855. Associated fore-and hind limbs provide a significant difference in distinguishing the two taxa (measurements are given in Tables 2.11 and 2.12\. Femur. The femur is readily distinguished from those of other non-Diplodoczs sauropods from the Morrison Formation by its overall slenderness. The femur of Barosaurus falls inro one of two variants of the "standard" Diplodocus-type femur. The standard type has a generally straight shaft, which becomes somewhar expanded in its upper third, and it also has an oval cross-section. Examples are D. carnegii (CM 94, CM 86 now at the Field Museum
in Vernal, Utah), D. hayi (HMNH 175), and a number of
speci-
mens from Dinosaur National Monument and Bone Cabin Quarry. In the first variant, which grades into the "standard" type, the expansion of the upper third is enhanced, giving the shaft a slight sig-
66 .
John S. Mclntosh
TABLE 2.11.
Tibia: Femur Lengths in Diplodocids Barosaurus lentzs AMNH 6341
-A ./a
Diplodocus carnegii CM 94 Diplodocus DINO 378, DINO 4236 Diplodoar DNM field #601I5-16 in Durham and Toronto Diplodocus AMNH 5855 Diplodocus "longus" DMNH 1494 Diplodocus hayi HMS 175 (formerly CM 662)
.72
Diplodctcus "longus" USNM 10865 Gigantosaurus africanus HMB Quarry k
.65 (or .75")
.72 .71.
.69 .68 .65 .64
" Two articulated sets of left tibia-fibula-astragalus r.vere found rvith DNM field #355. The shorter set r'vas used by Gilmore in the mounted skeleton, br'rt there is some reason to believe that the longer ones mav have been the correct sei.
moid curve. Examples are the Barosaurus femur (AMNF{ 6347 U.T. 993.1, YPM-PU 18710 [formerly AMNH 651], AMNH 5855). The type femur of Gigantosaurus africanus from Quarry A at Tendaguru also falls into this category, whereas the type of D. carnegii (CM 84) is intermediate with the standard type. The second, more striking of the two variations is the so-called Amphicoelids or stovepipe type, where the anteroposterior diameter is increased, resulting in an almost circular cross-section of the shaft; the head of the femur is also more weakly developed. Examples include the type femur of Amphicoelias ahus AMNH 5764, a specimen in the University of Utah collection from the Cleveland-Lloyd Quarry, and two Diplodocus femora (AMNH 223, USNM 10865 on the mounted skeleton). Finally, Lull's (1919) estimate of a femur over 2.5m long in YPM 429 is incorrect. The fragment he used in making the estimate does not belong to the type skeleton. Tibia. The shaft is straight and slender. It is little expanded at the distal end, more so at the proximal end, with a prominent cnemial crest typical of the diplodocids (Fig. 2.1.28), in contrast to CamdrdsAurus, where it is smaller. There is a wide variation in the ratio of the length of the tibia to that of the femur, which is not easiiy explained and will be treated more fully in a forthcoming publication. Fibula. This long, slender, flat bone is slightly expanded at each end. The usual triangular muscle scar appears on the medial face of the proximal end (Fig. 2.1,2C). The only other feature of the bone is a muscie scar about one-third of the wav down the shaft from the The Genus Barosautus Marsh (Sauropoda, Diplodocidae)
'
(7
TABLE 2.I2. Measurements (mm) of Hindlimbs
DiPlodorus
Barosaurus sp.
Ientus
AMNH
cL1l
634I
Femur, lc. oth
ne8tc
CM
94
A.\4NH
5855
left
rieht
- (_) 3es 127) 393 (27) 313 (28) - (-) 202 (14) 183 (13) 162 (r4) 150 (-)
1440
breadth,
proximal breadth, shaft breadth, distal
(100)
385
1410
(27)
(100) 113s (100)
369
(2.6)
274
(24)
263
*t^."^..-* ru.ta45 USNM 10865
t,-)
circunrference 540 (38) 510 (36) 410 (36) 408 (-) Tibia, Iength 1064 (100) 1010 (100) 783 (100) 770 (100) breadth,
27s
\26)
274
136
(13)
130 24s
proximal breadth,
shaft
(27)
195
(25)
215
(28)
(13)
100
(13) l23J
108
(14)
"
hayi
HMNS 175
(cM
562)
393 \27)
430
(31) - (-)
208 (13)
227 (16)
195
114)
208 (16)
301 (19)
382 \26)
420
(30)
387 \29)
563 (37)
590 (41)
548
1020 (100)
93.r (100)
362 (32J
301 (32)
-
))\ t)4\
circumference 368 (35) 358 (35) 273 (35) 279 136) 419 141) Fibula, length 1120 (100) 1055 (100) 800 (100) 798 (100) 107s (100)
444 147) 983 (100)
(24) (21t
178
(18)
223
179
\9) (16)
88 (8) 63 (S) 165 11.6) 1.28 (16)
circumference 267
(24J
240
breadth,
205
185
proximal
shaft breadth, breadth,
99
distal
\23)
1,75
768
(22)
(23) - (-)
\22)
63 (8)
214 t20)
)\) t)4\
(-)
17s \76)
-(-)
107 (11) 172 lr7)
180 (23)
)q\ t)7\
279 (28)
-
Quarry k
34e (23)
174 119)
(24)
A
(100)
1448 (100)
1Sg (15)
zss
Quarrv
1s75 (100)
247 (24)
breadth, distal
Gigantosaurus" africanus
1380
-
1340 (100)
(40) ss9 (42) (-) 860 (100) i-) 325 (38) (-) 1,40 \16) (-) 2s.5 (30)
- (-) -(-)-(-) -(-)-(-) -(-l-(-)
362 t43)
- (-) -(-)-(-)
(-)
Ratios given in ( ) are relative to bone length (100). circumference is minimu.-r shafr
top on the lateral face. The distal end is expanded mediaily opposite a concavity in the astragalus. Tarsals. The astragalus of AMNH 6341 (Fig. 2.12D) is virtually a carbon copy of AMNH 5855, but it differs from that of Diplodocus (CM 94) and those of the African species in that the medially directed process, which lies beneath the tibia, is somewhat longer in Barosaurzs. It resembles that of Apatosaurzzs closely. There was no calcaneum found with AMNH 6341 or with any other diplodocid. Indeed there was no room for one beneath the Iower end of the fibula, whose distal end extends well beneath that of the tibia. Bonnan (2000) has reported the presence of a calcaneum associated rvith a Diplodocws pes CM 30767. The specimen, Field Number 1,75 from Dinosaur National Monument. was as-
63
.
John S. Mclntosh
*
I '
l,
:t. '1,'
'M :'11''
,ttli
.
'.
:r tir6
.l 1
t,,,,$ti',',
li .i Y:,'
,,,,"
& |
.i
eq'
.-,il
signed to scattered bones of a number of individuals. Having studied the records carefully, I do not believe the association of this bone with the foot can be substantiated. Pes. Metatarsal I is short and massive, with the prominent process on the lower part of the posterior margin of the lateral face (that facing metatarsal II) so characteristic of the diplodocids Diplodocus, Apatosaurus, Cetiosawriscus, and Dicraeosaunzs; it is so far not known in others (measurements are given in Table 2.13).
'a
,
Fig. 2.12. Barosaurus lentus
AMNH 5311 right hindlimb bones: (A) femur posterior and distal uiews; (Bt tibia posterior; (C) fibula in lateral and distal uieuts; and (D) astragalus in
posterior uiew,
The Genus Barosaurus Marsh (Sauropoda, Diplodocidae)
' (9
TABLE 2.13. Measurements (mm) of the Pes of Barosaurus
AMNH
6341
Height of astragalus Maximum transverse breadth Anteroposterior breadth
139 246 167
lentus 6341 IIIV
Metatarsal length Metatarsai proximal transverse breadth Meratarsal proximal anteroposrerior
B.
B. "affinis"
AMNH
YPM 419 TTT
140 181
175
704 94 136 131 134 82
138 86
97
105
98
?cM 11984 III IV
il
208 2r7 242 239 231 142 132. 101 706 199 183 192 140 137 1r2
breadth
Metatarsal distal transverse breadth Metatarsal distal anteroposterior breadth
r21
Shaft least circumference Length of first phalanx
158
86 58
Length of second phalanx Length of claw
108 92 187
93
120
58
69
167 146 113 110
t75 77
;10; 265
A 'il:i),,,'' ' '|LY'&'fit'
B
Fig. 2.13. Barosaurus lentus bones of the right
AMNH 6341
(A-C) Metdtdrsdls I,II, and V in anterior, posterior, lateral, and dorsal uiews.
Pes.
70
.
John S. Mclntosh
c
119
*$d
; 188
91
79
i+ B
,i"u.
)t
'6'n t*'"'*-e; The bone is virtually identical to that of Diplodouzs, but a bit less massive than that of Apatosaurus. Metatarsal II possesses a similar, though smaller, distal process and the comparisons are the same as for metatarsal I. Metatarsal V cannot be compared with that of CM 94 (Diplodocus) because it was broken in life and is so distorted as to make comparisons worthless. Compared to that ol Apatosaurus (CM 89), metatarsal Y in Barosaurus is surprisingly more robust, particularly at the distal end. It bears a close resemblance to that of Camarasaurus grandls (YPM 1905). The phalanges do not differ noticeably in AMNH 6347 and CM 94. They may be tentatively identified as I-1, lI-I,ll-2,II-3, III-2, and III-3 (Fig. 2.13). As expected they closely resemble those of Diplodocus
Fig. 2.11. /A) Barosaurus lentus YPM 419 E. "affinis") left metatarsal I in anlcrior, fosterior, medial, and laterdl uiews. (B) CM 11981 (?) Ieft metdtdrsdls I to V in InteTK)r utew.
carnegii (Hatcher 1901) and "Barosaurus" africanus (Janensch 1961\.
Lull (1919) identified the type of Barosaurus affinis Marsh as two metapodials that he inexplicably called metacarpals I and II. One is clearly the left metatarsal I and the other appears to be the upper end of left metatarsal II (Fie.2.1a). They do not differ from those of AMNH 6341 except that they are a little smaller. A large diplodocid left pes was found just below the junction of the cervicals and dorsals of CM 11984 and assigned the same field number #310/N. Although the first metatarsal differs from that of AMNH 6341, in minor respects, the pes has been provisionally catalogued as part of CM 719B4.
The Genus Barosaurus Marsh (Sauropoda, Diplodocidae)
.
/1
Discussion Status o/ Gigantosaurus africanus Fraas
In 1908, E. Fraas described two species of a nerv genus of sauropod dinosaurs from Tendaguru, German East Africa (now Tanzania). He named them Gigantosaurus africanus and G. robustus, aware that Seeley had used the generic name for an English sauropod in 1869. Fraas argued that Seeley had never properly characterized this genus and had never figured it, and that therefore the name was open for use. Citing nomenclatural rules then in effect, Sternfeld in 1911 argued that the name Gigantosdurus was not available and renamed the African genus Tornieria. Then in 1.922, lanensch concluded that the two species belonged to separate genera and in a brief footnote referred to G. africanus as Barosaurus africanus. Until recently, following Nopcsa (1930), most writers incorrectly referred to the other species as Tornieria robusta. This reference is incorrect because Tornierid replaced Gi-
gantosdurus africanus and thus may still be available for that species if it proves not to belong to Barosaurus. GigdntosAurus robustws is properly called ldnenschia robusta, as proposed by \X/ild in 1991. However, Janensch never really made his case for assigning G. africanws to Barosaurzzs. Fraas (1908) noted the great similarity of the the material assigned to this species wtth Diplodocus, and my observations of the large collections at the Museum fiir Naturkunde in Berlin and the Natural History Museum in London reveal that the preponderance of the specimens belong to Bra-
chiosaurus and animals closely allied to Diplodocus. Does this diplodocid material belong to a single genus, does it belong to both Diplodocus and Barosaurus as rn North America, or does some or all of it belong to a third, as yet unnamed genus? A series of anterior caudal centra exhibit ventral sculpturing quite different from that seen in either Barosaurus or Diplodocus from North America. There is a central pit divided in two by an anteroposteriorly directed midridge. Some of the caudals do closely resemble those of Barosaurus, particularly as to rheir slight ventral sculpturing, but there are others that have a greater resemblance to those of Diplodocus. The only Barosaurus-like cervical figured by Janensch (1929; dd 1,79) is not overly elongated and resembles Diplodocus more than Barosaurus. The humerus:femur ratio is 0.72 for both the skeleton from Quarry k and that from Quarry A (Table 2.7).Both agree with that of AMNH 634l implying a long forelimb. These ratios for the Tendaguru material musr be viewed with caution, because the humerus from Quarry A was collected by the German expedition at the same site from which Fraas had obtained the type femur of Gigantosaurus africanus. Fewer than a dozen scattered bones were taken up there by the two expeditions, and although there is no duplication, at least one anomalous bone, a Brachiosaurus fibula, was taken from the site. As to skeleton k, the humerus:femur ratio is suggestive of Barosawrzs, but the
72
.
John S. Mclntosh
tibia:femur ratio of 0.64 as compared to 0.74 for AMNH 6341 (Table 2.10) suggests a very different animal. The conclusion I come to is that further study must be made of Gigantosaurws africanus material, particulariy of its vertebrae, before it can be assigned with any assurance to Barosdurezs. This may require preparation of more material in Berlin if indeed this material survived \forld War IL Barosaurus Compared with Other Diplodocids Barosaurus differs from ApatosAurus (1) in having one less dorsal vertebrae; (2) in having more slender cervicals with slender,
rather than robust, ribs; (3) in having winglike,
transverse
processes on the caudals extending much further posteriorly; (4) in
having both pleurocoels and ventral excavations in the anterior caudals; (5) in having more slender limb bones; and (6) in having a humerus:femur ratio of 0 .72, instead of 0.67 . Barosaurus differs from Dicraeosaurus (1) in having many more and much longer cervicals and fewer dorsais, 9 instead of 12; (2) in having pleurocoels in its dorsals; (3) in having much less cleavage in the spines of the posterior cervicals and anterior dorsals; and (4) as cited above for Apatosaurus, in having both pleurocoels and ventral excavations in the anterior caudals. Barosaurus differs from Cetiosauriscus (1) in having more complex sculpturing laterally and ventrally in the caudals; (2) in having a larger humerus:femur ratio; and (3) in having differently developed chevrons.
Barosaurws has both similarities and differences with Superslurus. Various specimens formerly assigned to other taxa may in fact belong to Supersaurzs, including the large anterior caudal (BYU 9045, formerly BYU 5002; Curtice et aI. 1996) and the gigantic cervical (BYU 9024, formerly BYU 5003), although it is so badly crushed that many important details cannot be seen. The type specimen, the right scapula-coracoid, is of the diplodocid type as noted by Jensen (1985). Supersaurus differs (1) in that the scapula has a greater expansion of the distal end from that of Diplodocus and Barosaurus, (2) the ischium is virtually indistinguishable except for size from those of both genera, and (3) the caudal vertebrae are more reminiscent of Barosawrzs. If BYU 9045 is caudal 1 or caudal 2 because of its very short centrum, then the spine is relatively low and transversely broad as in Barosaurus, in contrasted with Diplodocus. The midcaudals referred by Jensen (1985) to Supersawrus also exhibit the rounder, less exaggerated ventral sculpturing of Barosaurzs. Finalln the truly enormous size difference may be gauged from the length of the scapula of Supersaurus, which is 2.2 m in its somewhat flattened condition. This is 1.7 times as long as 1.3 m (measured on the curve) for the scapula of Barosaurzs (AMNH 6341). At one time, I thought that Supersaurus might be a gigantic species of Barosaurus, btt norv believe that there is evidence to indicate that it is a valid senus. A final de-
The Genus Barosaurus Marsh (Sauropoda, Diplodocidae)
.
7J
{/r tl\
Fig. 2.15. Comparison of lateral uieus of last dorsal: (A)
Arnphicoelias altus AMNH .5764; (B) Diplodocus "longus" USNM 10865 (after Gilmctre 1932); (C) Barosaurus lentus AMNH 5341 (data from Osborn and Mook 1e21 ).
f\
(*\
t
\*4 I
"-*\l
termination must await full preparation and study of the large Brigham Young University collection from the Dry Mesa Quarry. Comparing Barosaurus and Amphicoelias altus Cope is difficult due to the paucity of material of the latter. Its straighr, stovepipe-like femur, with an almost circular cross-section, is more reminiscent of Diplodocus (USNM 10865) than Barosaurus (AMNH 6341). On the other hand, the well-preserved posterior dorsal of A. altus (AMNH 5764) dilfers from both Barosaurws and Diplodocus (Fig. 2.15). Osborn and Mook (192I) concluded, probably correctly, that it is the last dorsal. It cannot belong to Diplodocus because its spine is straight and vertical, in contrast with Diplodoczs (USNM 10865), where it is angled anteriorly (Gilmore 1932). The spine of the last dorsal vertebra of Barosaurus is anteriorly angled but less so than in Diplodoczs. In the less likely event that the Amphicoelias dorsal is the next-to-the-last one, it cannot be Diplodoczs because its spine shows no sign of a cleft, which is present in that genus. The anteroposterior breadth of the spine of Ampbicoelids (AMNH 5764) greaily exceeds that of either Diplodocus or Barosdurus, and its pleurocoel is smaller than in either of these. How significant these differences are must await better material from the Cope locality, but it is doubtful whether Amphicoelias can ever be raised from the status of a nomen dubium. Finally, the large, incomplete scapula (AMNH 5764a), referred to Amphicoelias by Osborn and Mook (1.921), suggests that Amphocoelias is perhaps closer to Supersaurus than it is to either B arosaurus or D iplodo cus.
74 . lohn S. Mclntosh
Conclusions Bdrosaurus is a valid genus of the family Diplodocidae which is closely related to Diplodours, but it is more advanced in some respects and more primitive in others. Advance conditions are in the anterior part of the skeleton where Baros aurus has incorporated an
additional dorsal into the cervical series, leaving nine dorsals. Furthermore, the cervicals and anterior dorsals are up to 507o more elongated rn Barosaurus, and the forelimb is relatively longer. On the other hand, the tail of Barosaurtts is more primitive in that the centra are relatively shorter, the spines of the anterior caudals are shorter and unclefted, the caudal pleurcoels and transverse processes disappear several vertebrae anterior to those of Diplodoaus, and the ven-
tral sculpturing is less extreme. Finally, the typical Diplodocus-llke chevrons in the midcaudal region have a less developed anterior extension. The referral of Gigantosaurus africdnus to Barosaurus, though possible, should arvait further study. Acknowledgrnents. I am deeply indebted to Drs. Eugene Gaffney, Mary Dawson, the late Hermann Jaeger, the late Nicholas Hotton II, the iate Alan Charig, and John Ostrom for permission to examine materials in their care. I also want to thank Michael Brett-Surman and Robert Long for photographs of the Barosaurus cervicals. I have tried to measure personally as many of the bones mentioned in this paper as possible, but in some instances where this was not possible I have made use of published measurements of Lull, Janensch, and Foster as well as an unpublished measurement of the late C. C. Mook. I express my extreme gratitude to Allen McCrady for his painstaking, very skillful work on the neck of CM 11984. I am most grateful to Kenneth Carpenter and Virginia Tidwell for their help in preparation of the vertebral figures. References Cited
Bonnan, M. F. 2000. The presence of a calcaneum in a diplodocid sauropod. Journal of Vertebrate Paleontoktgy 20(l: 3f7423. Colbert, E.H. f96L Dinosaurs: Their Discouery and Their'Vlorld. New
York: Dutton. Curtice, B. D., K. L. Stadtman, and L.J. Curtice. 1996. A reassessment of Utrdsaurus macintosbi (Jensen, 19851. Museum of Northern Arizona
Bulletin 60:87-95. Foster, J. R. 1996. Sauropod dinosaurs of the Morrison Formation (Upper
Jurassic), Black Hills, South Dakota and \Wyoming. Contributions to Geology, Uniuersity of Wlyotning 31(1): 1-25.
E. 1908. Ostafrikanische Dinosauier. Palaeontographica 55: 105-144. Gilmore, C.W. L932. On a nen4y mounted skeleton of Diplodoats in the United States National Museum. Proceedings of the United States National Museum 81(18): 1-21. 1936. Osteology of Apatosaurus wrth special reference to specimens in the Carnegie Museum. Memoirs of the Carnegie Museum 1,1,: 1-63.
Fraas,
The Genus Barosaurus Mrrsh (Sauropoda. Diplodocidae)
.
75
Hatcher, .J. B. 1901. Diplodocus (NIarsh): Its osteologn taronomy and probable habits, with a restoration of the skeleton. Memoirs of the Carnegie Museum 1,: 1,-63.
Holland, W. J. 1906. The osteology of Diplodoars Marsh. Memoirs of the Carnegie Museum 17 : 225-27 8.
1915. Section, "Paleontology." Annual Report of the Carnegie Museum for 1914.30-33. 1920. Section, "Paleontology." Annual Report of the Carnegie Museum for L9L9. 34-37. 1,922. Das Handskelett von Gigantosaurus robustus u. Brachiosaurus brancai aus den Tendaguru Schichten Deutsch-Ostafrikas. Centralblatt Mineralogie, Geologie und Palaeontologie (1922):
464480. Janensch, W. 1,929. Magenstein bei Sauropoden der Tendaguru-Schichten.
Palaeontographica. Supplement 7 (1): 137-143.
L935-I936. Die Schadel der Sauropoden Brachiosaurus, Barosaurus, und Dicraeosdurus aus den Tendaguru Schichten Deutsch Ostafrikas. (Schluss). Palaeontographica Supplement 7(1'): 14s-298. 1,961,. Die Gliedmassen und Gliedmassengurtel der Sauropoden der Tendaguru-Schichten. Palaeontographica Supplement 7(Iir: 177-235. Jensen, J. A. 1985. Three new sauropod dinosaurs from the Upper Jurassic of Colorado. Great Basin Naturalist 45(4): 697-709. Lull, R. S. 1.917 . Barosaurus: A gigantic sauropod dinosaur. Btilletin of the Geological Society of America 28: 214 (abs.) 1919. The sauropod dinosaur Bdrosanrus Marsh. Memoirs of the Connecticut Academy of Arts and Science 6: 1-42. Marsh, O. C. 1890. Description of new dinosaurian reptiles. American Journal of Science 39(3): 81-86. 1896. The Dinosaurs of North America. U.S. Geoktgical Suruey Annual Report 16: 1.33-244. 1898. On the families of the sauropodous dinosaurs. American Journal of Science 6(4):487-488. 1899. Footprints of Jurassic dinosaurs. American Journal of Scien
ce / t+ )i
/_./_
/ -^!_ J/_,
Mclntosh, J. S. 1981. Annotated catalogue of the dinosaurs (Reptilia, Archosauria) in the collections of the Carnegie Museum of Natural History. Bulletin of Carnegie Museum of Natural History, r1o t$. Mook, C. C. 1,91,7. The fore and hind limbs of Diplodocus. Bulletin of the American Museum of Natural History 37: 815-819. Nopcsa, F. 1930. Zur Systematik und Biologie der sauropoden. Palaeobiologica 3:40-52. Osborn, H. F. 1899. A skeleton of Diplodocus. Memoirs of the American Museum of Natural History 1': 191'-21'4. Osborn, H. F., and C. C. Mook. 1.921. Cdmardsdurus, Amphicoelias and other sauropods of Cope. Memoirs of the American Museum of Natural History, new series 3(3): 249-387. Osborn, H. F., and'W. Granger. 1901. Fore and hind limbs of sauropoda from the Bone Cabin Quarry. Bulletin of the American Mttseum of
Natural History 14(13): 199-208. H. G. 1869. Index to the Fossil Remains of Aues, Ornithosauria, and Reptilia. Cambridge, UK: Deighton' Bell, and Co.
Seeley,
76 o John S. Mclntosh
Sternfeld, H. L911. Zur Nomenklatur der Gattung Gigantosaurus Fraas. Sitzungsberichte der G esellschaft naturforschender Freunde du B erlin
1911(8):398.
'Wieland,
G.L. 7920. The long neck sauropod Barosaurus.
series (1326):
'Wild,
Science, new
528-530.
R. 1.991-. lanenschia n.g. robusta (E. Fraas 1908) pro Tornieria robusta (E. Fraas 1908) (Reptilia, Saurischia, Sauropodomorpha). Stuttgarter Beitraege Zur Naturkunde (B\ 173: 14.
The Genus Barosaurus Marsh (Sauropoda, Diplodocrdael
'
/7
3. Reassessment of the Early Cretaceous Sauropod Astrodon iohnsoni Leidy 1865 (Titanosauriformes ) KeNNrru CenppNTER AND VrncrNre Trownrr
Abstract Sauropod material from the Arundel Formation (Aptian-Albian boundary) of Maryland has been variously referred to Astrodon johnsoni Leidy 1865 or to Pleurocoelus nanus Marsh 1888. Most of the specimens are juvenile as demonstrated by the small size of the bones, the lack of neurocentral fusion, absence of an olecranon, and underdevelopment of muscle scars. Contrary to some recent statements, the Arundel sauropod is diagnostic. Only a single sauropod taxon is present in the Arundel Formation, to which the name Astrodon jobnsoni must be used under the Principle of the
First Reviser of the Inrernational Code of Zooloeical Nomenclature.
Introduction Prior to 7996,Early Cretaceous sauropods were assumed to be rare in North America, and most specimens were referred to Pleurocoelus, a taxon from Maryland first described by Marsh (1888). Since 1996, the diversity and number of Lower Cretaceous sauro78
pod specimens from North America have increased dramatically, although not all of the specimens have been formally described. The brachiosaurid Cedarosaurus weiskopfae (Tidwell et aI. 1999) and Venenosaurus dicrocei (Tidwell et al. 2001) are known from sediments low in the Cedar Mountain Formation (Barremian-basal Cenomanian) of Utah. The brachiosaurid Sattroposeiden proteles ('Wedel et al. 2000) has been described from a single specimen recovered from the Twin Mountain Formation of Oklahoma. Isolated teeth from Utah and Texas have been referred to Astrodon (Cifelli et a|. 1997), although this is probably not correct because similar teeth occur in other taxa. The brachiosaurid Sonorosatrus thompsoni (Ratkevitch 7998), from the Turney Ranch Formation of Arizona. may be Cenomanian in age. In recent years, Pleurocoelus has become a wastebasket for fragmentary Jurassic and Cretaceous sauropod specimens. The referral by Langston (1974) of various Texas specimens to Pleuro' coelus has resulted in a great deal of confusion about Pleurocoelus (e.g., Gallup 1975; Mclntosh 1990). The recent discoveries of numerous taronomically diverse sauropods from the Lower Cretaceous of North America highlight the need for a reexamination of the original material of Pleurocoelzs, which we present beiow.
Taxonomic History The first sauropod named from North America was Astrodon johnstoni, found in the Arundel Formation near Bladensburg, Maryland (Fig. 3.1). Although Joseph Leidy (1865) is usually credited with the name, in fact the genus was proposed in 1859 by Christopher Johnson. However, because no species was named by Johnson, Leidy is considered the author (permissible under ICZN 1999 67.2.2:l "if a nominal genus . . . was established before 1931 .. . without included nominal species . . . , the nominal species that were first subsequently and expressly included in it are deemed to be originally included nominal species"). The syntype is a spatulate tooth and a thin section of a second tooth (YPM 798) (Figs. 3.2G; 3.4A, B). Additional specimens of sauropods from the Arundel Formation were recovered by John Hatcher in 1887 and 1888, including a number of teeth with a similar morphology. As was common for the time, Marsh did not refer any of his specimens to a previously named taxon) especially one created by someone else. Instead, he proposed new names, Pleurocoelus nAnus ancPleurocoelus altus (Marsh 1888). The type series (syntypes) consists of several specimens, some of which were illustrated by Marsh in lBBB and more rn 1896. The specimens were transferred to the U.S. National Museum,'Washington, D.C. (now the National Museum of Natural History) in 1898-1899. John Hatcher (1903) offered the first review of the Maryland sauropods while describing some iuvenile sauropod bones from the Morrison Formation (see also Carpenter and Mclntosh 1994). He notes that it was he who had collected the specimens described by Reassessment of the Early Cretaceous Sauropod Astrodon iohnsoni Leidy
1865
'
79
Washington D.C.
Hanover Jessup Fig. 3.1. Geographic dis*ibution of Astrodon localities in eastern
Gontee
Maryland. Specific sites around Muir ki rk include : Cb ero kee Sdnford Brick Qwarry ( = Maryland Clalt Products Brick
Muirkirk
Quarry,), Duuall's lron Mine, Dut,all's Bank, Engine Bank, Henson's Bank, Island Bank, Latcbford Dump, Shea's Bank, Srt,ttnp Poodle r = Co[fin's LnBine Bank, Coffin's OId Engine Bank).
Bladensburg Washlngton D.C.
r..//
0
50km
-E
Marsh and that most of the specimens came from the same general location and horizon (for a history of the Maryland iron industry and the discovery of the Arundel sauropods, see Kranz 1996). Hatcher also notes that there was no evidence to indicate the presence of more than one species of sauropod, a point on which we concur. Hatcher concluded that Astrodon johnsoni had prioriry and that this name should be applied to the Arundel sauropod. Nomenclature stability could have been insured for the Arundel sauropod had it not been for Richard Lull. Lull (1917\ described and figured a representative sample of the Arundel sauropods at the U.S. National Museum, as well as specimens in Gloucher ColIege (the specimens have since been transferred to the National Museum of Natural History). Based on the relative abundance of the two size classes, he concluded that "Pleurocoelus abus. . . could have been the possessor of teeth like those of Astrodon lohnstoni. .. . It is therefore quite possible that Pleurocoelus altus should be considered as synonymous with Astrodon johnsoni, in
which case the latter name would take precedence" (Lull 1911, 203). By inference, P. nanus would be a separate taxon. Despite his
80
.
Kenneth Carpenter and Virginia Tidwell
prootic antotica
frontal laterosphenoid
adductor fossa
@ffi@ A
c
B
premaxillary process
fenestra ovalis
vidian
E
ffiffiffiffiftffiffi&ffiffim H
conclusion, he left the three taxa separate, which has led to the present nomenclatural confusion regarding Astrodon versus Pleurocoelus. Charles Gilmore (1921,) presented a brief summary of the Arundel fauna and attempted to head off the confusion when he concluded that "I think it preferable to assign all to the genus Astrodon, which clearly has priority" (Gilmore 1921,588). He listed three species ol Astrodon: A. nanus, A. ahus, and A. iohnstoni. After Gilmore's study, little mention was made of the Arundel sauropods until Kingham (1,962) did a brief review. He recognized only a single species of Arundel sauropod, Astrodon iohnsoni, but oddly he also synonymized Brachiosaurus with Astrodon as well, a synonymy not followed by subsequent authors. Ostrom (1970) discussed the Arundel sauropod in the context of sauropod material from the Lower Cretaceous Cloverly Formation. He agreed with Hatcher, Lull, and Gilmore that only a single sauropod taxon was present in the Arundel. However, he chose to follow Lull in retaining temporarily all three taxa; he was apparently not aware of Kingham's study. Langston (1,974) followed Ostrom and also retained all three taxa. However, he referred the postcrania of a Lower Cretaceous Texas sauropod to Pleurocoelus sp', thereby Reassessmenr
J Fig.3.2. Cranial bones o/Astrodon johnstoni as illustrated b1'Lull (1911). Supraoccipital in (A) posterior, (B) dorsal, and (C) anterior uiews. Left or b ito sp h e n o i d- lat er o sp h eno i d (alisphenoid of Lull 1911) in lateral uieu (D); parts labeled as c.urrently identified (not labeled by Lull). Left maxilla in lateral uiew (E), showing the premaxillary prucess as originall1, preserued (now missing; figure not originalll'labeled by 1,vllt. Lcft dentary in lateral uiew (F). Teeth were restored b1' Lull, original is edentulous. Teeth (to scale) include the holotype in buccal, marginal, and lingual uiews (G); anterior(l) semi-spdtuldte tooth in lingual and marginal uieuts (H); small poste-
riortll. semi-spatulate lootb in buccal, marginal, lingual, and marginal uiews (I); posterictr(?) semi-sptfiulate tooth in buccal,
rnarginal, lingual, and marginal t,iews (J).
of the Early Cretaceous Sauropod Astrodon iohnsoni Leidy 1865
.
81
changing the diagnosis of the genus, as can be seen by that given by Mclntosh \1990). Salgado et al. (1995) presented a preliminary review of Pleurocoelws, accepting P, nanus as a valid taxon and "Plewrocoelus" altus as a separate taxon distinguished by the distal end of the tibia. Nothing was said regarding Astrodon, although they did question the identity of the Texas sauropod as Pleurocoelus. Later, Salgado and Calvo (7997) reversed themselves, accepting only a single taxon, P. nanus, and they referred to the Texas specimens as Pleurocoelus sp. More recently Kranz (1998) has referred to the Arundel sauropod as Astrodon johnsoni. This brief historical view shows that earlier authors tended tt-rward Astrodon, and some, but not all, recent authors tended toward Pleurocoelus. Most of the authors, Salgado et al. (1995) excepted, consider the Arundel sauropod to represent a single genus and possibly a singie species. Our examination of the material also leads us to conclude that a single taxon is present, as first stated by Hatcher (see additional discussion below). Because Hatcher (1903) was the first reviser of the taxon, we have accepted his determination that Astrodon iohnsoni is the correct nomen under the Determination of the First Reviser: "'When the precedence between names li.e., Astrodon vs. Pleurocoelusl . . . cannot be objectively determined, the precedence [i.e., the name to be used] is fixed by
the action of the first author citing in a published work" (ICZN 1999, 24.2). Specifically, Hatcher (1903, 7I-12) conciuded:
1. comparison of the syntype teeth of Astrodon jobnstoni with those studied by Marsh showed "a verv striking similarity . . ." 2. that the material was collected from the same deposits (i.e., Arundel Formation) as the syntype teeth "likewise was found in a bed of iron ore near Bladensburg
Maryland"
3, the remains are from the same area, "since these remains were found essentiallg and perhaps identically, the same locality and horizon . . ." 4. the preservation is the same, "the great similarity which they exhibit . . ." 5. only a single species is present, "there appears no good reason for considering them as pertaining to either different genera or species."
6. therefore, " Astrodon johnstoni Leidy having priority should therefore be retained, whlle Pleurocoelus nanus would become a synonym." 'We are cognizant that this decision
will not be universally accepted, but not to accept Astrodon iohnsoni as the valid name wiil upset nomenclature stability because of the implication it would have that more than one species of sauropod is present in the Arundel Formation. Because such an implication was not substantiated by us, the oldest name available (ICZN 1999,23.1) is Astrodon
82.
Kenneth Carpenter and Virginia Tidwell
iohnsoni, as acknowledged by Hatcher (1903), and subsequently
by Lull (1911), Gilmore
(7921.), Kingham (1962), and Kranz
(1998). Although as shown above, both Astrodon and Pleurocoelus have been used for the Arundel material, suppression of the name Astrodon is not possible under the non-use guidelines of the ICZN (1,999,23.9.1,, "the senior synonym or homonym has not been use as a valid name after 1899") because Astrodon has been used as recently as 1998 (Kranz 1998\.In addition, the use of the name Astrodon does not promote taronomic instability or cause confusion, thus it cannot be abandoned in favor of Pleurocoelus (ICZN 7999, 23.9.3 with 23.2). Finally, the question of whether or not the syntype teeth are diagnostic needs to be addressed. Although the teeth are said to have come from Bladensburg, Maryland, there is no indication that they were found adjacent to one another. Given that most specimens from the Arundel Formation occur as isolated specimens, ir seems more probable that the teeth were not found together. Nevertheless, both Johnson and Leidy treated the teeth as belonging to the same taxon, and today they comprise the syntype for Astrodon iobnsoni (ICZN
1999 72.1..I: "all specimens on which the author established a nominal species-group taxon . . . in the absence of holotype designation . . . all are syntypes and collectively they constitute the namebearing type"; Art.73.2: "Syntypes are specimens of a type series that collectively constitute the name bearing type . . . for a nominal species-group taxon established before 2000 . . . all the specimens of the type series are automatically syntypes if neither a holotype . . ' or lectotype . . . has been fixed . . . all have equal status nomenclature as components of the name-bearing type"). That the syntypes probably came from more than one locality is acceptable by the ICZN (1999,73.2.3: "if the syntypes originated from two or more locali-
ties (including different strata), the type locality encompasses all of the places of origin").
'!7hen Leidy described the syntype of Astrodon johnsoni in 1856, the characters were unique at that time. Since then, however, other teeth have been referred to either Astrodon or Pleurocoelus (which was based on the assumption that this was the correct nomen for the Arundel specimens). Some of these specimens do not resemble any of the Arundel specimens (e.g., specimen of Langston 1974), whereas others do have some resemblance (e.g., Cifelli et al. 1997). Thus, under current rules of the ICZN (1999,13.1.1), the syntypes cannot "be accompanied by a description or definition that states in words characters that are purported to differentiate the taxon." However, this ruling applies to names published after 1930 (Art. 13). Because Astrodon johnsoni was named before that time, all that is required is that the name satisfy Article 11 (which it does, e.g., properly published, created from Latin alphabet, etc.), and "be accompanied by a description or definition of the taxon." Differentiation is not a requirement for the taxon named prior to 1931. Hatcher (1903), however, has made it possible for us to dif
Reassessment of the Earlv Cretaceous Sauropod Astrodon iohnsoni Leidv
1865
'
83
ferentiate Astrodon iohnsoni by his nomenclatural act under Determination of the First Reviser (ICZN 7999,24.2) whereby ali of the Arundel specimens, including the syntypes of Pleurocoelus nanus, were referre d to Astrodon johnsoni. In our studS comparisons were made with various taxa from the literature (citations in parenthesis), casts, and actual speci-
mens (denoted by catalog numbers). In addition, one of us (V. Tidwell) examined specimens at various museums in Argentina and England (see Acknowledgments, end of this chapter). Specrmens considered include: Aegyptosaurus (Stromer 1932 Lapparent 1960); Aeolosaurus MACN RN 147 (Powell 1986; Salgado and Coria 1.993a); Alamosaurus UT WL 476 (Gtlmore 1946; Lucas and Sullivan 2000); Andesaurus MUCPv132 (Calvo and Bonaparte 1991); Antdrctosaurers MACN 6904 (Huene 1929); Argentinosaurus PYPH-I (Bonaparte and Coria 1993); Argyrosaurus MLP 77-V-29-1, MACN VH 217 (Huene 1929;Powel| 1986); Brachiosaurzs USNM 5730, FMNHP25107 (Riggs 1903; Janensch 1.935-1936, 1950); Camarasaurzs DMNH 2850 (Mc-
Intosh et al. 1996a, 1996b; Ostrom and Mclntosh 1966): Cedarosaurus DMNH 39045 (Tidwell et al. 1999): ChubutisAurus MACN CH1.8222 (Salgado 1993); Epacbthosaurus MACN 1.3689 paraplastotype (Powell 1986, 7990; Gimenez 1992); Eucamerotus /oxl BMNH R 90 (Blows 7995); Gondwa-
natitan (Kellner and Azevdo 1999); Jones Ranch unnamed SMU 61732 (Langston 1974; Gomani et al. 1.999); Laplatasaurzs MLP CS 1128 (Huene 1.929;Powell1.979); Lirainosaurzs (Sanz et al. 1999); Magyarosaurzs (Huene 1.932); Malauisaurzzs MAL 181, 182, 200 (Jacobs et al. 1993; Gomani er al. 1999); I'Iewquensaurus MLP Ly 1-6 (Huene 1929; Powell 1986; Salgado and Coria 1993b); Paralititan (Smith et al. 2001); Phuuiangosaurus (Martin et a|. 7999); Rapetosaurus UA 8698, FMNH PR 2209 (Curry, Rogers, and Forster 2001); Saltasaurus (Powell 1992); Sauroposeidez OMNH 53062 (-ifedel et al. 2000); Isisaurus colberti (Jain and Bandyopadhyay 1997; Wilson and Upchurch 2003); Venenosaurus DMNH 40932 (Tidwell et al. 2001). Institutional abbreuiations. BMNH-British Museum Narural History, !-ondon, United Kingdom; DMNH-Denver Museum of Natural History, Denver, Colorado; FMNH-Field Museum of Natural Historn Chicago, Illinois; MACN-Museo Argentino de Ciencias Naturales, Buenos Aires, Argentina; MIOW-Museum of the Isle of Wight, Sandown, Isle of \X/ight; MLP-Museo de La Plata, La Plata, Argentina; MUCP-Museo de Ciencias Naturales de la Universidad Nacional del Comahue, Neuquen, Argentina; NSMT-National Science Museum Tokyo, Tokyo, Japan; PVPH-Paleontologia de Vertebrados, Museo "Carmen Fumes," Plaza Huincul, Argentina; SMU-Southern Methodist UniversitS Dallas, Texas; USNM-National Museum of Natural History, '$Tashington, D.C.; UT-University of Texas, Austin, Texas; YPM-Yale Peabody Museum of Natural History, New Haven, Connecticut.
84.
Kenneth Carpenter and Virginia Tidwell
Systematic Paleontology
Order Saurischia Seeley 1888 Suborder Sauropodomorpha Huene 1932 Infraorder Sauropoda Marsh 1878 Titanosauriformes Salgado et aI. 1997 Family incerta sedis Astrodon johnsoni Leidy 1865 Astrodon johnstoni Leidy 1865 Pleurocoelus nanus Marsh 7888 Pleurocoelus a/ras N{arsh
18
88
Astrodon nanus Grlmore 1921 Astrctdon altus Gtlmore 1.921
Syntypes-YPM 798 tooth and a thin section of another. Locality dnd horizon Arundel Formation, Potomac Group, near Bladensburg, Maryland. Age. Aptian (Doyle 1992). Referred material. See Table 3.1. All specimens are from the Arundel Formation, Potomac Group, Muirkirk, Maryland. Diagnosis: Supaoccipital with low, wide crest and tall, narrow foramen magnum. Cervicals short, very wide, camerate centra with very large pleurocoels that leave the centrum a mere sheil. Dorsals with deep, well-defined pleurocoels. Sacrals with deep pleurofossa, posterior sacrals rvith prominent groove on ventral surface. Anterior caudal centra ver-v short, circular, amphiplatyan; neural spines low. Coracoids very thick along edges and with a prominent lip. Radius with distinct ridge extending obliquely down shaft, and two small postero-distal condyles.
Description Some of the bones illustrated by Marsh (1888) and Lull (1911) have been damaged and processes have been lost. Other bones, notably a small femur (USNM 2263), cannot be found. To show what
the bones used to look like, Lull's illustrations are reproduced below to ease comparison rvith the bones as they are today.
Skull The cranial material is from different localities, hence from different individuals. Horveve! as noted by Lull, the individuals were all very young, as indicated by the lack of fusion of the braincase elements.
Supraoccipital. The supraoccipital (Fig. 3.3A-D) is low and wide, rather than tall as in most other sauropods (e.g., Camardsdurus, Brachiosaurus, Jainosaurus). It forms part of the lateral wall and the roof of the foramen magnum. The foramen magnum is tail and narrow as it is in the titanosatr Jdinosaurus and an unnamed titanosauriform from Texas (Tidwell and Carpenter 2003). Reassessrnent
of the Earlv Cretaceous Sauropod Astrodon iohnsoni Leidv 1865
.
85
Table 3.1 Referred Specimens of Astrodon and References of Figured Specimens
Material
CatalogNumber
Reference
maxilla maxilla with tooth maxilla maxillary tooth
USNM 5667 USNM 7288
Lull 1911, pl. 14,
altisphenoid pterygoid, frag. laterosphenoid supraoccipital dentary
357091.
USNM USNM USNM USNM USNM USNM
9152 5668
Lull 1911, pl. 16, fig.
5670 5688 5692 5669
Kingham 1,962, fig. 3b
1
Lull 1911, pl. 15, fig.3
I;
USNM 6104
Marsh 1896, pl. 40, frg. Lull 1911, pl. 14, fig. 5 Kingham 1962, fig.2b
tooth
USNM 357052 USNM 5691
Marsh 1896, pl. 40, frg.2;
roorh roorh
USNM 6105 USNM 6106
toorh tooth, frag. tooth tooth, frag. toorh tooth tooth toorh tooth tooth, frag. tooth tooth tooth
USNM 8443 USNM 8457 USNM 8458 USNM 8459 USNM 8460 USNM 8461 USNM 8462 USNM 8463 USNM 8464 USNM 8465 USNM 8466 USNM 8467 USNM 8468 USNM 8469 USNM 8470 USNM 8471 USNM 8472 USNM 8516 USNM 357080 USNM 357085 USNM 435228
dentary, frag. angular, frag.
roorh
tooth tooth, frag. toorh
tooth tooth tooth tooth tooth tooth tooth tooth tooth
pl. 79, frg. 4
USNM 481078 USNM 481088
USNM 6094 USNM 6:I22
Kenneth Carpenter and Virginia Tidwell
191'1,,
USNM 438042
cervical centrunr cervical centrum
?
Lull
USNM 442452
cervical centrum cervical centrum cervical centrum
axis
Lull 1911, pl. 14, fig. 5 Lull 1911, pl. 14, fig. 8
USNM 437986
USNM USNM USNM USNM USNM
AXIS
86 .
usNM
frg. 6
5452 57OO
5640 5641
5678
Marsh 1896, pl. 40, fig. 3; Lull 1911, pl. 15, fig. 2
Material
CataloeNumber
Reference
dorsal centrum
USNM 4968
Marsh 1888, figs. 1, 2; Marsh 1896. pl. 40. figs. 4, 5; Lull 1911, pl. 15,
dorsal centrum dorsal centrum dorsal centrum dorsal centrum dorsal centrum dorsal centrum dorsal centrum dorsal neural spine rib, frag.
USNM 5675 USNM 5705 USNM 6092 USNM 6097 USNM 6103 USNM 8499 USNM 85OO USNM 6110 USNM 5679 USNM 6102 USNM 8509 USNM 8515
fie.4
rib rib, frag.
rib
head
sacral centrum
USNM 4969
Marsh 1888, figs. 3, 4; Marsh 1896, pl. 40, figs
6,7;LulI fie. sacral centrum sacral centrum sacral rib sacral rib caudal centrum caudal centrum
USNM USNM USNM USNM
caudal
USNM 5372
1911, pl. 15,
5
5666 8475
5672 6117 USNM 2266 USNM 4970
Marsh 1888, figs. 5, 6; Marsh 1896, pl. 40, figs. 8, 9; Lull 1,9'1,1, pl. '1,6, fre.2 Manh 1896. pl. 40. figs. 10.
ltr Lull l9lt. pl. 16.
fre.1 caudal centrum
USNM 5639
Marsh 18q6. figs.38-4 l:
Lull 1911, pl. 1,6, fi1. caudal caudal caudal caudal caudal caudal caudal caudal caudal caudal caudal caudal caudal caudal caudal caudal
centrum centrum centrum centrum centrum centrum centrum centrum centrum centrum centrum centrum centrum centrum centrum
USNM USNM USNM USNM USNM
2
5643 5644 5650 5651
5662
USNM 5663 USNM 5664
USNM USNM USNM USNM
5665 5680 5682 5683 USNM 5694 USNM 5702 USNM 7290
USNM 7291 USNM 7293 Reassessment of the Early Cretaceous Sauropod Astrodon iohnsoni Leidv
1865
.
87
Table 3.L (continued)
Refered Specimens of Astrodon and References of Figured Material
CataloeNumber
caudai caudal caudcI caudal caudal caudal caudal caudal caudal caudal caudal caudal caudal caudal caudal caudal caudal caudal caudal caudal caudai caudai caudal caudal caudal caudal caudal caudal caudal caudal caudal caudal caudal caudal caudal caudal neural neural
USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM
centrum
centrun
centrur centrum centrurr centrurr centrurr-
centrurr centrurn centrum centrum centrum centrum centrum centrum centrum centrum centrum centrurrt centrum centrum centrun-r
centruni centrum
centrur,
centrum centrum centrum centrum centrum centrum centrum centrum centrum centrum neural arch arch arch, frag. neural arch, frag. neural arch zygapophysis. frag. zvgapophl sis. frag. centrum centrum centrum, frag. centrum, frag, centrum, frag. centrum, frag. vertebra, frag. scapula, frag.
88 .
Kenneth Carpenter and Virginia Tidwell
73OO
7301
7302 7303
7319 8442 8455 8476
8477 8,178
8479
8480 8481
8482
8483 USNN,I 8484 USNM 8485 USNM 8486
USNM 8487 USNM 8488 USNM 8489 USNM 8490 USNM 8491 USNM 8492 USNM 8493 USNM 8494 USNM 8495 USNM 8496 USNM 8497 USNM 8498 USNM 8501 USNM 8506 USNM 3570,+5 USNM 357069 USNM 357084
USNM USNM USNM USNM USNM USNM USNM
357096
USNM USNM USNM USNM USNM USNM
6101
6111
8510 357039 357060 7287 357103 USNM 2264 357051 357083 357100 357101 357044 USNM 5677
Reference
Specimens
Material
CataloeNumber
scapula, frag. scapula, distal end coracoid, frag. coracoid, frag. sternal plate
USNM 8474 USNM 357093 USNM 6096
Reference
USNM 8473 USNM 357075 USNM 2263
humerus
Kingham 1,962, frg. 4 Lull 1911, fig. 16, fig. 5;
pl. USNM USNM USNM USNM
humerus, frag. hr,merrrc hrrmerrtc)
f
rr o frao
ulna, ulna, ulna, ulna, ulna,
6098
357066 357105 USNM 357106 USNM 5673 USNM 5674 USNM 9150 USNM 9153 USNM 357092 USNM 2263
humerus, prox. end humerus, prox. end
prox. end prox. end distal end distal end distal end
radius
1,7, frg. 2
5697
Lull 1911, fig. 16, fig.
5;
pI. 77, frg.2 radius, prox. end radius, prox. end metacafpal
USNM 9147 USNM 9149 USNM 2265
phalanx
USNM 2265
metacarpal metacarpal metacarpal, prox. end metacarpal metacarpal
USNM 5646 USNM 5648
--+^,-^-^..1
metacarpal
*-,^^^--^t lrrL rdLdrydr,
f-^-
f-^rra5.
metacarpal ntetacarpal, prox. end f*^'-^+^..^-^^lPdr! rra5. * -.^..^ -^^ f-^ y4'r| rraE.
ischium?
ischiurn, frag. ischium, frag. ischium femur
Marsh 1896, pl. 41, fig. Lull 1911, p1. 17, fig. 1 Marsh 1896, pl. 41, fig. Lull 1911, pl. 17, fig. 1
1; L;
USNM 5649
USNM 5658 USNM 5659 USNM 5686 USNM 5689 USNM 5695 USNM 5698 USNM 5699 USNM 8513 USNM 357056 USNM 6095 USNM 6118 USNM 357094
usNN{ 3s7107 USNM 2263
Lull 1911, fig. 16, fig.
5;
pL. 1.7, fi1. 2
femur, frag. t-emur,
distal end
femur femur, distal end
tibia tibia
tibia, distal end
USNM USNM USNM USNM
5696 6112 16710 452029 USNM 4971 USNM 5656 USNM 6114
Lull 1911, pl. 18, fig.3 Luil 1911, pL 1,7, frg. 3
Reassessment of the Earlv Cretaceous Sauropod Astrodon iobnsoni Leidv
1865
.
89
Table 3.1 (continwed) Refened Specimens of Astrodon and References of Figured Specimens Material
CataloeNumber
tibia, prox. end
USNM USNM USNM USNM USNM USNM USNM USNM
tibia fibula fibu1a
fibula 6bula, prox. end fibula, distal end metatarsal
Reference
6121 357T07 4971
Lull 1911, p1. 18, fig.3
5657 5676 6113 9155 2267
Lull 1911, pl. 17, fig. 3
Marsh 1896.
pl.4l.
3.4. s.6r Lull lell.
figs.
pl.
15, figs. 1, 2 metatarsal metatarsal metatarsal metatarsal metatarsai metatarsal
phalanx, pes
USNM USNM USNM USNM USNM USNM USNM
5660 5681 5687 6093 6119 6120 2267
Lull 1911, pl. 19, fig.1
Mar:h l8qo. pl. 41. fig'. 3, 4, 5, 6; Lull 1911, 15, figs. 1, 2
phalanx, indet. phalanx, indet. phalanx, indet. .holo.*
i.,-lot
phalanx, indet. nhalenw
inlct
phalanx, phalanx, phalanx, phalanx, phalanx,
indet. indet. indet.
frag. indet. frag. indet.
USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM
p1.
5642 5653 5654 5661 5690 6119 7294 7296 8514
357042 357049
Ventrally (Figs. 3.2C; 3.3C, D), the supraoccipital is open its entrre length, rather than partially constricted as in Cdmarasdurus (see Madsen et al. 1.995, fig. 27). The supraoccipital crest, seen on the posterior side above the foramen magnum (Fig. 3.3A), is low and very wide; it is narrow and sharp in Camarasdurus, and it is narrow and flat in Brachiosaurus and Jainosaurus. The sutures for the opisthotic and basiocciptal are triangular in outline, whereas they are more rectangular tn Camdrdsaurus. Two small facets above the foramen magnum are for the proatlas. C)rbitosphenoid and laterosphenoid. L:ull (1911) originally figured and illustrated these two bones as the alisphenoid (see Fig. 3.2D). These are separate today (Fig.3.3E-H), and we have been unable to articulate them as shown by Lull. In addition, the "hook-
90.
Kenneth Carpenter and Virginia Tidwell
Flg. 3.3. Cranial bones
of
Astrodon lohnstoni, Swpraoccipital /USNM 5592) in (A) pctsterior, (B) anterior, (C) dorsal, and (D) uentral uiews. Ahhreuialion: ps = proatlas shelf. Left laterosphenoid (USNM 5688 mislabeled as USNM 357091) in lateral (E) and medial (F) uieus. Left orbitosphenoid (USNM 5688 mislabeled as USNM 357091) in lateral 1C) and medial tH) uiews. I-eft maxilla /USNM 5667) m lateral (J) and medial (I) uiews; crowns of replacetnent teeth uisible at first and third alueoli.
9't :, '1*!r '€
G
Scale
in cm.
like" process illustrated by Lull, which is actually part of the laterosphenoid, can no longer be found. Neither element is complete, making comparisons with counterparts on other skulls difficult. If we accept the illustration given by Lull as accurate, then based on their relative sizes and positions, the fenestra include that for cranial n V and the vidian canal into the pituitary fossa. The problem with such an interpretation is the apparent absence of a fenestra for Reassessment of tl.re Earlv Cretaceous Sauroood Astrodon iohnsoni Leidv
1865
.
91
n II-IV. To resolve the problem will require that all of the pieces be relocated and rearticulated in the manner available to Lull. Maxilla. None of the maxillae (e.g., Fig. 3.3I, J) are as com-
plete as illustrated by Lull (see Fig. 3.2E). The maxillary body is wedge-shaped in profile, being deep anteriorly and tapered some-
what posteriorly. The sutural surface with the premaxillary is slightly convex and vertical in profile. In contrast, it is posteriorly sloped in Camardsaurus and Brachiosaurzs. In addition, the hooklike premaxillary process projects anteriorly and leaves little or no space for the subnarial fenestra. In Camdrasaurus, the process arises from the dorsal surface of the maxilla and does not protrude
anteriorly very much beyond the sutural surface. ln Brachiosaurus,
the process is similar to that in CamarasaurusJ but it originates more posteriorly on the dorsal surface of the maxilla. The alveoli of the Astrodon maxillae are damaged on all the specimens noq although Lull noted that a maximum of ten were present. Pterygoid. A pterygoid fragment was illustrated by Kingham (1962), but it was not found by us. Assuming that the illustration and identification are correct, the pterygoid is more arched dorsally than in Camarasaurus,but less arched than in Brachiosauras. It is difficult to compare the various processes (e.g., the quadrate process) because they appear damaged. It is therefore not known if there was originally a dorsomedially oriented, hook-like, basipterygoid process as in Camarls*urus. Dentary. The complete dentary illustrated by Lull (1911; see Fig. 3.2F) cannot be found, and those that remain are damaged (Fig. 3.3K). The dentary is most like that of Brachiosaurus, although it is proportionally deeper as compared to its length. Medially, the meckelian groove extends anteriorly towards the symphysis, which is missing. The coronoid process is long and low, as in Brachiosauras. Lull 11911) has noted that there are thirteen alveoli present. Lull also claims that none of the teeth present in the dentaries are spoon-shaped, a point we have been unable to substantiate. If true, then perhaps the slender, isolated teeth (e.g., Figs. 3.2H, I; 3 .4C, E) are from the dentary. All of the dentaries are edentulous. Teeth.Isolated brachiosaurid-like teeth from the Lower Cretaceous have frequently been referred to both Astrodon and Pleurocoelus (e.g., Langston 1.974; Cifelli et al. 1999; Kirkland et al. 1999) . Many of these identifications are based on superficial resemblance to teeth illustrated by Lull (1911; see Fig. 3.2G-J). Unfortunateln few of the identifications have been based upon direct comparisons with the holotype of Astrodon iobnstoni, nor have they been based on the large number of referred teeth from the Arundel Formation in the USNM. In fact, most of the non-Arundel teeth differ from the Arundel teeth, as noted by Maxwell and Cifelli (2000). Marsh's description of the teeth is, "their crowns are mainly compressed cones," thus not spoon-shaped. The teeth may actually be defined as semi-spatulate or semi-peglike (Figs. 3.2G-J; 3.4). The differences are probably due in part to the position in the mouth, with the broader teeth probably more anterior to the semi-
9r .
Kenneth Carpenter and Vrginia Tidwell
n*o
IL"
F
F E
Fig. 3.4. Teeth oi Astrodon johnsoni (A-F) and Langston Texas sauropod (G-H): wear facets denoted uith drrows.
F
h,l
Synttpe (YPM 798) in buccal (A) and lingual (B) ttieu,s (comldre u,ith Fig.3.3, G); stnall pctsterior(?) tootb (USNM 5105) in buccal (C) and lingual uiews (D) (compare tuith Fig, 3.3,l); larger, anterior(?) tooth (USNM 8516) in buccal (E) and lingual uiews (F). Tooth (USNM 187535) incorrectly referred by Langston (1974) as Astrodon from the Paluxy F ormation, Montague County, Texas (G and H). Note the prismatic angles on the crown (at lines) and tbe double uear facets. Scale in ctn.
peglike teeth based on comparison with other sauropods. All of the teeth share the feature of having the wear facet along the margin of the crown (Fig. 3.4), rather than centrally on the top or lingual side (Fig. 3.4G, H). Furthermore, most teeth show only a single wear facet, which begins near the top and expands toward the root as wear increases. The crown is typically smooth in the upper half of the buccal side, and faintly rugose or crinkled in the lower half, especially near the root. The crown may be very slightly constricted where it joins the root (e.g., Fig. 3.4E), but not always (e.g., Fig. 3.4C). On the lingual side, the crown is often finely crinkled under magnification, Reassessment
of the Early Cretaceous Sauropod Astrodon iohnsc,tni Leidy 1865
'
93
although this is highly variable. A weak vertical ridge may be present on the lingual side of rhe crown dividing the crown in half (e.g., Fig. 3.2I; 3.4F); its presence is probably dependent on its position
in the mourh. The tooth referred by Langston (I974) as pleurocoelzzs cannot be referred to that taxon. The crown has a distinct prismatic appearance on both buccal and lingual sides, with a sharp angle be_ tween "prisms" (Fig. 3.4G, H). This prismatic condition is unusual among sauropods. Furthermore, there are two wear facets, one at the lingual side of the tip and the other along one side of the tooth. Postcrania Few of the postcranial elements were found in articulation, and most were widely scattered. As a result, vertebral counts for the various segments are unknown. Nevertheless, most of the individuals were immature, as indicated by the near lack of fusion between the neural arches and their centra.
The cervicals and dorsals are characterized by very large cavi-
ties in the centra, especially in the dorsals, as noted by Marsh (1888): "The latter [dorsals] . . . have a very long, deep cavity on each side of the centrum, to which the proposed generic name refers." Such cavities have been variously referred to as ..pleurocoels" (Marsh 1888), "lareral depressions" (Lull 1911), and "pneumatic fossa" (Makovicky 1997). More recentll', pleurocoels in dinosaurs have been defined as "fossae and foramina,' (Britr 1997), that is, any depression and cavity in the lateral surface of the centrum (e.g., Mclntosh 1990; Wilson and Sereno 199g) or a deep excavation in the lateral sides of the centrum (Upchurch 1998). Bonaparte (1999) is one of the few modern authors who separates a depression in the lateral side of rhe centrum, which he calls a "lateral depression," from a cavity into the centrum, which he calls a "pleurocoel."'We have independently come to the same conclusion and note that it is not always clear whether an aurhor of a reference means that the "pleurocoel" is a shallow depression or a cavity. Considering that Marsh (1888), in establishing the name Pleurocoelws, created the term "pleurocoel" to mean a large cavity within the centrum, we propose that the term be restricted to such a cavity, and that the term "pleurofossa', (,,lateral depres_ sion") be used for the lateral depression on rhe sides of the centrum. In some sauropods (e.g., Argentinosaurus and Epacb_ thosaurws), a pleurocoel is developed within a pleurofossn in a manner analogous to the antoribital fenestra within the antorbital fossa in theropods. Ceruicals. Lull (1911) identified a damaged cervical centrum (USNM 5700) as possibly the axis, based on what appeared to be a coossified odontoid. Although he may be correct, he also notes that the posterior cotyle for articulation with the next cervical is shallow, suggesting that the centrum is a damaged dorsal. The cervicals are opisthocoelous, and the mid- and posterior cervicals have large pleurocoels separated medially by a thin vertical wall (Fig. 3.5A-D;
94.
Kenneth Carpenter and Virginia Tidwell
B
fr* 'r''lf :,
o
J
N
r--r--r--r--r-l cm
P
e$& S T
U
--.a-
r#p$
o
3.6,4, B). Internally, the pleurocoels are enormous, so that the cen-
trum is almost a shell of bone. Although known in some juvenile sauropods (Carpenter and Mclntosh 1994), they generally do not leave the centrum as a shell of bone.
large pleurocoels are
The centrum of the mid- and posterior cervicals are very wide compared to the height. The centrum is also wider than it is high in the titanosaurids lleuquensdurus and Sahosaurus. However, in these sauropods, the pleurocoels are not proportionally as large as in Astrodon. The neural canal in Astrodon is hourglass-shaped and has a thin ridge extending down its length (Fig. 3.5C).
Dorsals. Only anterior and mid-dorsal centra (Figs' 3.5E-I; 3.6C, D) are known, and these are opisthocoelous, with centra as tall as they are wide. This contrasts with the centra of Ewcamero/ns, whose centra are dorso-ventrally compressed. The pleurocoels ol Astrodon are very large and somewhat teardrop-shaped, being tallest anterioriy and tapering posteriorly as is characteristic of many titanosauriforms (Salgado et al. L997)' The pleurocoels of Cedarosaurus are lenticular and divided by lamina, unlike the simple pleurocoels of Astro don. The neural canal is also hourglassshaped in dorsal u;.v' (Fig. 3.5I), but it lacks the midline ridge seen in the cervicals. Although the posterior dorsals are not known,
R Fig. 3. 5. Representdtiue uertebrae
o/Astrodon johnsoni. P ostelrcr ceruical centrum (USNM 5678) in left lateral (A), anterior (B), dorsal (C), and uentral (D) uiews. Anterior dorsal centrum /USNM 4958) in left lateral (E), rilht lateral (F), anterior (C), posterior 1H1. and dursal rlt utews. Anlerior caudal neural spnte (USNM 5650) in right lateral (J), posterior (K), and le[t lateral 1Lt uiews. Anterior caudal TUSNM 8188t in posterior (M) and right lateral (N) (anterior tou,ard rigbt) uieuts. Mid-caudal centrum /U.SNM 4970) in left lateral (O), posterior (P), dorsal (Q), and uentrdl (R) uiews. Distal caudal in left lateral (S), antenor (T), and right lateral (U) uiews. Scale
in cm.
Reassessment of the Early Cretaceous Sauropod Astrodon iobnsoni Leidy
1865
'
95
.'_a.
I
.
n. \ trtt,brte o/. Astrodon :s illustrated bt' NIdrsh
(.:r:rs(rni
1596t
tnd Lull (1911). Posterior
cert,iccrl centrutn in left laterdl t'ietu (A) and cross sectiotl (B) shotuing the deuelopment of the
pleurococls. Dorsol centrttm in Ieft lateral (C) and posterior (D) uiews. Arrlerior sdcral cenlrum in Ieft lateral (E) and posterior (F) uiews. Anlerior caudal cenlrum in posterior (G), left lateral (H), and
dnterior (I) uieuts. Mid-caudal centrutll in left lateral 1J) and dorsdl (K) uiews. Distal caudal in anterior (L), left lateral (M), and posterior (N) uieuts.
based on the sacrals and mid-dorsals, the centra shourd be progressively more amphiplatyan and have smaller pleurocoels. Sacrals. The sacral centra (Figs. 3.6E, F;3.7) have an elongate
pleurofossa on each side, a character known in few titanosauriforms, although present in Diplodocus and variably found in camarasdurus. On the second(?) sacral (Figs. 3.6E, F; 3.7A_D), the pleurofossa is located posterior to the parapophysis and is partially excavated into it (Fig. 3.7A). The suture for the neural arch on the second sacral is located over the anrerior two-thirds (Fig. 3.7c). The anterior articular face is triangular (Fig. 3.7B), whereas the posrerior face is almost circular (Fig. 3.68). The last(?) two posrerior sacral cenrra are coossified (Fig. 3.7E-H). The centra are low and wide (Fig' 3'7E, G). The pleurofossa are low in height, elongated, and moderately deep. on the dorsal surface of the centra, the neural canal has a pair of circular fossa and another pair that are elongate (Fig. 3.7G). Ventrally; there is a shallow srlou. .rtending the length of the centra (Fig. 3.7H). This groove apparently did not extend the full length of the sacrum becauie it is not presenr on the second(?) sacral. caudals. Most of the vertebrae of Astrodon are caudars (Figs.
3.5M-U; 3.6G-N). The articular faces are circular and am_ phiplatyan. The anterior caudal centra are short anteroposteriorly and become progressively longer (compare Fig. 3.6Ft with Fig. 3'5N), reaching their maximum length in the mid-caudals (Figs. 3.5o; 3.6J). None of the caudal centra have either pleurocoels or pleurofossa. The neural arch of all caudals is locared on the anterior half of the centra, a titanosauriform character (Salgado et al. 7997). However, the anterior edge of the neural arch is set back
slightly as in Andesaurus, cedarosauru.s, and venenosauru.s. The floor of the neural canal is constricted between the pedicels (Figs. 3.5Q; 3.6K). The facets for the chevrons are small, short, and not
96
.
Kenneth Carpenter and Virginia Tidwell
k
ro".
A
' -#rl
c
m
cm Fig.3.7. Sacral uertebrde of Astrodon johnsoni in right lateral
F
uiew (A) shou'ing Pleurofossa (arrou'), anterior uiew (B), dorsal uiew (C) (anteriol rigbt), and uentral uiew (D). Coossified posterior sacrals in right lateral
uieu (E) shouing Plettrofossa (arrouts), cross section (F), dr,trsal uiew (G) showing paired shalloa' fossa in the floor oi the neural canal prrows). and uentral uicw (H) showittg longitudinal grooue. Scale in cm.
very prominent. The neural arch of the most posterior caudal (Figs. 3.5S-U; 3.6L-N) is coossified to the centrum' suggesting that fusion of these two structures in the caudals occurs progressively
from back to front as in crocodiles (Brochu 1,996)' The anterior caudal neural spine is somewhat triangular, being anteroposteriorly longer along its top than along its base (Fig. 3'5J-L). The spine is simple. The postzygapophysis is developed at the base of the spine, rather than projecting posteriorly as in all other sauropods. However, neural spines are poorly represented elements in the very small juvenile specimens recovered to date. Further discoveries may allow us to determine if this feature is restricted to very young individuals. Ribs
A fragment of a large, hence adult, dorsal rib, was briefly deLull (1911). The tuberculum is very prominent and separated from the capitulum by a deep notch suggesting that it is an scribed by
Reassessmenr
of the Earlv Cretaceous Sauropod Astrodon iohnsoni Leidy 1865
'
97
Fig. 3.8. Large rib /USNM 8515/ o/ Astrodon ;ohn'oni shou'ing pleurocoel (arrotu) in web between capitulum dnd tuberculum. Scale in cm.
anterior rib (Fig. 3.8). A V-shaped depression in the web between the heads of the rib extends into a shallow pleurocoel on the posterior side, a condition also known in Venenosaurus, A Dneumarrc rib is characteristic of the titanosauriforms as defined tv '$Tilson and Sereno (1998), but this characer is also found in nontitanosauriforms, including Swpersaurus (Lovelace et al. 2003), and in a rib referred to Apatosaurus by Marsh \1896) (see Tidwell et ai. 2001; this rib may actually belong to Brachiosaurus). Forelimb Scapula. Only a fragment of the distal end of the scapular blade is known, and this is not very informative. Coracoid. The coracoids are incomplete (Fig. 3.9), but they do show unique features. First is the unusual thickness of the coracoid as compared to its size, even at the distal edges. The coracoid foramen must be set high, as in many titanosaurs, because it is nor visible on the fragments. Lull (1911) alleges that a small portion of the coracoid foramen is present, but we were unable to substantiate his claim. Furthermore, the anterior edge of the glenoid forms a promi-
98
.
Kenneth Carpenter and Virginia Tidwell
scapula facet
Fig. 3.9. Coracoid o/Astrodon jolrnsoni (USNM 6096) in uentral (A) and glenoid (B) uiews. Scale in cm.
nent lip more pronounced than that in Camarasaurus, and it is bordered by a deep notch. Hwmerus. The humerus is long and slender, although less so
than in Brachiosaurws and Cedarosaurus (Figs. 3.10; 3.11A-D). The deltopectoral crest is small in titanosaur fashion and is located a little over a quarter down the shaft; it proiects perpendicularly almost at a right angle, as rn Brachiosaurus, rather than medially, as in Cedarosaurus and titanosaurids. The proximal end of the humerus is slightly more expanded relative to the width of the shaft than it is in Cedarosdurus, but considerably less than it is in Chubutisaurus, Camarasdurus, or Brachiosaurus. Both the medial and lateral sides of the shaft are concave so as to give the humerus an hourglass shape. This profile makes the medial edge of the humeral shaft appear straighter than in Brachiosaurws. The shaft is not as slender in proportion to its length as in Brachiosaurus. The proximal end lacks the well-developed dorsal tuberosity for the M. subscapularis seen in many titanosaurids. This absence makes the proximal end appear slightly narrower than the distal end' Distally, the radial and ulnar condyles are not developed, possibly because of the immature nature of the bone. In ventral view, the articular surface is more rectangular than in Brachiosaurus. The supracondylar ridges on the posterior surface are not well developed. U/za. Neither ulna is complete (Fig. 3.12). The olecranon is Reassessment of the Early Cretaceous Sauropod Astrodon iobnsoni Leidy
1865
.
99
s'
;**q
%W
'%,sd
@ ',]1|,.
':i*,.
ry|)
w
ctraryG F tg. 3.
1
[t. Hurnerus
o/Astrodon
johnsoni / USNM 2263) in left l.rt, rtl .f . pruxintal \B). anterior 1Ct. distal tD), and medial (E) L'ieu s. Scale in cm,
absent, in marked contrast to the low olecranon in Cedarosaurus, Bracbiosaurus, and Camarasaurzzs. However, the ulna of a juvenile Venenosaurus has only a slightly developed olecranon (see Tidwell and ril/ilhite, this volume), whereas in the adult, it is rvell developed (see Tidwell et al. 2001). In dorsal view, the radial notch is well developed. The medial side of the ulna is almost straight and the lateral process almost perpendicular (Fig. 3.12A) as in Cedarosaurus and Brachiosaurus.
100
.
Kenneth Carpenter and Virginia Tidwell
s
b B
ff1 F
fB
{h
v'{ \'{
fl\. I tlt lnq
lk
n\
ftL$ c
K
l;{ I[\1 ,i# I/ \1fl-/tl \esid;ry \r"LJ
l,l /\
A trqq \4t
J
}
\lt
lil
'\d @
l* .ln
E
H
ffi G
I
/-\ffi
G@^S\2/
L
Iig. 3.11. Appendicttlar material
Lull (19111. Hurnerus in posterior (A), proximal (B), distal (C), and lateral (D) uiews. Femur in posteric;r (E), proximal (F), uentral (G), and lateral (H) uteuts Tibia and fibula in anterior (l), proxinnl as illustrated (J) and corrected $) (fibula should not arttcttlata with the cnential tresl'. distdl (L), and posterior (M) as illustrated by
ulews.
Radius. The radius is very slender and straight (Fig.3.13). As Rapetosaurus, a prominent ridge extends obliquely down the posterior side. This feature contrasts with Cedarosdurus and Brachiosdurus, which have two ridges extending the length of the radius. The ridge begins just below the proximal end and terminates just above the distal ulnar expansion. The distal end has two small condyles on its posterior surface and is wide transversely. These features have not been found in any other sauropod and give the Astrodon radius a unique profile.
in
Metacarpals. The metacarpals are long and slender (Figs. 3.14A-H; 3.19A), a camarasauromorph character according to Salgado et al. (1997). However, the distal ends are undivided and lack ventrally distinct condyles, suggesting the absence of phalanges, a titanosaurid feature for Salgado et aL. (7997) and a titanosauriformes character for Wilson and Sereno (1998). However. some Astrodon metacarpals show poorly developed condyles on the posterior sides, similar to those in Venenosaurus and Antarctosaurus. Peluis
All that is known of the pelvis is an ischium. Ischium. The ischium is damaged (Fig. 3.15); nevertheless, it is complete enough to show that it is sharply bowed (Fig. 3.15B, D), and it is twisted along its long axis as in Camarasaurus and Brachiosaururs. The articular surface for the ilium is set on a short peduncle as in Andesdurtrs, Gondwandtitan, and Aeolosaurus; the peduncle is poorly developed in Camarasaurus and Brachiosaurus. Reassessment
ofthe Earlv
Fig. I.12. U/ra u/'Arrrudor. johnsoni. USNM J67J nr proximal (A) and left lateral (B) uiews. Scale in cm.
Cretaceous Sauroood Astrodon iohnsoni Leidv
1865
.
101
\
/\
I
.l
&,
I
ffiw
tI T
%
&e
#
w
A
E
F
Fig. 3.11. Metacarpals o/Astrodon johnsoni (USNM 2265) in anterior (A, C, E, G) and posterior (8, D, F, H) uiews. Scdle in cm.
l,l
I -fu
n
fl It f,'1
il '\r,,,
i-(". ti;,
B
, r.
:.
R.rilfu,s /LTTNM 22631 :,,r L'ietu showirtg the
,:.,:t oblique ridge (A) and
.:i:L
c
Flg. 3.15. lschium o/Astrodon johnsoni in posterior (A), Iateral (B), anterior (C),
:nd tncdinl
1D1 uictt s. Scalc
in cm.
tB). Scdle in cm.
Hindlimb Femur. The right femur has been damaged since illustrated by
Lull (1911; compare Figs.3.11E and 3.16). The shaft is slightly constricted just below midlength (Fig. 3.16), thus making it more slender than in Brachiosdunzs. The greater trochanter is slightly lower than the femoral head, as in Brachiosaurus and I'Jeuquensdurus, and it is not offset by a step as in Cedarosaurus. Just below the greater trochanter, the femur has a prominent, titanosauro1
I .
Kennetir Carpenter and Virginia Tidwell
morph-like lateral bulge. The fourth trochanter is a slight swelling located at midlength of the shaft. The distal condyles are more prominent than in Cedarosaurus and are separated by deep
+- greater trochanter
grooves.
\7e would like to take this opportunit,v to correct a common error in discussions of the greater trochanter in sauropods. This trochanter has, in recent years, been identified as the region dorsal and lateral to the femoral head (i.e., the top of the femur; see Borsuk-Bialynicka 1977, fig. 17A). However, as pointed our by Gregory (1918,535), "the greater trochanter in most repriles (including the Sauropoda) remains on rhe outer side of the shaft more or less near the proximal end. The outer portion of the head itself has sometimes been wrongly called 'great trochanrer,' especially in the femur of the Sauropoda." In a footnote he continues, "the proximal surface of the so-called great trochanter was co\rered by bursa . . . and that the gluteal muscles [i.e., M. iliofemoralis externus] were attached on the outer side of the great trochanter and not upon its top" (see Fig.3.16). Indeed, this area on the sauropod femur is often rugose, showing the insertion for various muscles (see further discussions by Carpenter and Kirkland 1998); it is also the region that corresponds to the proximal half of the "lateral bulge." In fact, development of the bulge is probably analogous to the development of the raised fourth trochanter, which, in sauropods, can appear as a broad ridge. The lateral bulge, is in fact, an extension of the greater trochanter; it is so variable in sauropods that its utility in phylogeny is questionable without quantification. Tibia. The tibia and fibula of the holotype Pleurocoelus altus (USNM 4971) cannot be found. Based on the illustrations of Lull
,q ,$
u'g'''
'Wr'
n
t I t t
Fig. 3.16. lentur of Astrodon johnsoni /USNI4 5696) irt posterinr uiew. Note large lateral
bulge (arrout). The trLte greater trochanter is noted. Scdle in crn.
*;.$, &*
*mu
ry
Fig. .3.17. Tibia of Astodor johnsoni /USNM .56.57) irtmedial (A), anterior (B), and laterttl (C) uiews. Distttl entl of tihh fUSNM 61.11) in onterior (D) and uetttral
E
B Reassessment
tr T
rLt t tcu s. lthrtlo irr tncdiLl t'ieu
.
Scdle in cm.
of the Early Cretaceous Sauropod Astrodon johnsoni Leidl' 1865
.
103
Fig. 3.11I-M), we find no reason to maintain this species as distinct contrary to Salgado et al. (1995; but apparently no longer distinct in Salgado and Calvo 1997). The distal end of the complete tibia of Astrodon (Fig.3.17A-C) is abraded, as indicated b1'the absence of the astragalar notch (this may be the tibia that (7911;
see
Salgado et al. (1995) used to separate P. dltus from "P. nanws"). This feature is well preserved on another, partial tibia (Fig. 3.17D, E). The tibia is more slender in proportion to its length than in Bra-
chiosaurus, but this may be a size- or weight-related difference and might change ontogenetically. The cnemial crest is ver1, prominent, much more so than rn Brachiosaurus, and it extends almost to midshaft (Fig. 3.1,7 A, C). The astragalar facet is not as open as in Brachiosaurus. The ventral surface of the distal end of the tibia resembles that of Brachiosaurus, contrary to Salgado and Calvo (1.997), and it is transversely wider than anteroposteriorly long. Fibula. The fibula is rather robust for its length (Figs. 3.11I, M; 3.17F). The fibula found r,vith a tibia (USNM 4971) cannot be located, but based on the illustration by Lull (1911), it does not differ markedly from another we refer to Astrodon (Fig. 3.17F). The fibula is slightly bowed in lateral view. The scar for M. iliofibularis is a well-defined ridge rather than the low, broad oval found in Camarasaurus. The proximal end is slightly C-shaped, rather than Dshaped as in Brachioslurus.
Metatarsals. The metatarsals are mostly of iuveniles (Figs. 3.18A-E;3.19B), but there is also one large aduit (Figs.3.18F G; 3.19D). They are the typical short, wide form of sauropods. Some of these show an articular surface on the distal end, indicating that 'We phalanges were present. do not accept the referral by Gallup (1989) of a pes from the Lower Cretaceous of Texas as Pleurocoelus for reasons discussed below. Thus, the number of claws on the pes remains unknown for Astrodon (= Pleurocoe/zs), contrary to Salgado et al. (7997).
Comparisons with Other Sauropods Sauropods from the Lower Cretaceous of Texas have been referred
to Pleurocoelws by Langston (1974), as well as more recently by Salgado and Calvo (1997). However, as Gomani et al. (1999) have
noted, the most proximal caudals of the Teras specimens are slightlv procoelous and grade posteriorly to plani-concave (see Table 3.2), rvhereas all the caudals of Astrodon are amphiplatyan. The hypantrum and hyposphene are also well developed in the Texas sauropod, but are absent in Astrodon (contrary to Tidwell et al. 1999, fig. 14; see Fig. 3.5K). Gomani et al. (1999) conclude that the Texas sauropod is a basal titanosauriform more derived than Brachiosauras, but less derived than Somphospondyli. Vertebrae belonging to other titanosauriformes show a variety of similarities and differences with Astrodon. The cervical vertebrae of Astrodon are proportionally much shorter and wider than those of Brachiosaurus and Sauroposeiden, suggesting a relatively 104
.
Kenneth Carpenter and Virginia Tidwell
l:',1.,"',,1,:$ n,,'::
proximal
la'
anterior ril
l!,
\..W
!
@ ",* 1 I...,j..,**lS
I
_&.
s
qtr
anterior
,&*,ul,,l'r . ,.,]ir,,-
posterior
-
=w
posterior
! {;,
ffi
W F
,q6ffi
ffi
jW..
';t,,S@"''
A
BCD
ry
ffi "/: distal E
short neck compared with other North American titanosauriformes. All presacral centra of Astrodon exhibit a degree of internal excavation due to the extremely large pleurocoels beyond anything found in the titanosauriformes Brachiosaurus, Cedarosaurus, Andesaurus, or the 'Wealden brachiosaurs of England. This difference is even more noticeable in comparison with the basal titanosaur Malawisaurus and with more derived titanosaurs like Argentinosaurus, Aeolosaurus, and Saltasaurws, all of whom show very shallow pleurocoels in the presacral vertebrae. The unusually deep pleurocoels of Astrodon are sometimes ascribed to its immature development, as large pleurocoels have been considered a juvenile character in the past (Carpenter and Mclntosh 1,994).Indeed, presacral vertebrae in baby Cdmarasdunzs specimens from Oklahoma show deep pleurocoels (Carpenter and Mclntosh 1994), although not as deep as those in Astrodon. However, pleurocoels in the somewhat larger juveniles Cdmarasauras CM 11338 and YPM 1.91.0 are smaller in proportion to the centrum size, a trait that is also found in Bellusaurus and Phuwiangosaurus (Martin 1994). Astrodon shares with CedarosaLtrus, Epachtbosaurus, ChubutiReassessment
G
Fig. 3.18. Metcttalsals of Astrodon johnsoni. (A) Metatdrsdl i /USNM -5660) in proximal, dntelior, posterior, and distal uietus. (B-E) Melalarsnls r( SNM 22t,-1 in proximdl, dnterior, posterio/, and distal uiews. Scale in cm. Metatarsal (USNM 5687) irt dnterior dnd posterktr (l) ueus. Ungual oI Astrodon jol.tnsuni /USNM 2267) in laterdl and nredial uiews (G). Scale in cm.
of the Early Cretaceous Sauropod Astrodon iohnsctni Leidv 1865
.
105
lrl !\
{-}
ffi tifl t' '[
ru
J,ffi
ffi
4,Uer\ lf,lr*;{Cr?
WA
l'{ (1
Pwn
\_$
fh a$
ffi
Fig. 3.1 9. Metapodials of Astrodon johnsoni as illustrated
B
qcr*-/
bt Lull. Metdcdrpdl (A) in proxinlal, anterior, distal, and ntetlial t.,iews. Metatarsdl (B) in
proxinnl, anterior, distal, and nteLllrtl Ltie&'s. Distal ungual (C) in
ltroxirntl, lateral, dorsal, and tttteriol' ttieu,s. Metatdrsal (D) in
protittnl, anterior, and distal t,iett
ffiWW
s.
sAurus, and Andesau,,as pleurocoels that are teardrop-shaped and
posteriorly acumate. The dorsai centra of Astrodon are camerate in structure, in contrast to the camellate vertebrae of Cedarosaurtts, ChubutisAurus, and Malawislurus. The amphiplatyan anterior caudal centra of Astrodon are similar to those in Brdchiosaurus and Venenosaurus, though differing from the plani-concave anterior caudals of Cedarosaurus, "PelorosdLtrus," Chubwtisaurus, and Andesaurus. Although Calvo and Bonaparte (1991) describe the ante106
.
Kenneth Carpenter and Virginia Tidwell
TABLE 3.2. Terminology for Vertebra Types Based on Articular Surfaces of Centrum Etymology Term
G:
Greek;L
amphicoelous (G) double
- Latin
cavit_v
Definition deep cotvle on both articular surfaces.
amphiplatyan (G) double flat amphicyrtian (G) double convex
both articular surfaces flar.
biconvex
ball on both articular surfaces.
(L) two curves outr,vard opisthocoelous (G) behind cavity
plani-concave (G) flat concave
plani-convex
(G) flat convex
platycoelous (G) "flat"
cavity
-
biconvex.
ball'and-socket, with ball on
anterior surface and cotyle on posterior surface. one articular surface flat, the other concave. one articular surface f1at, the other convex. shallow cot\,le on both articu-
lar surfaces.
procoelus
(Ct hefore caviri
ball-and-socket, with cotl'le on anterior surface and ball on posterior surface.
Adapted iron-r'$(illiston 1925; Hoffstetter and Gasc 1969. Cot,vle articulatine surface of the cenrrum.
: cavit)'on
the
rior caudals of Andesaurzs as amphiplatvan, Salgado et al. (1997) call them slightly procoelous. A reexamination of the specimen shows that they are plani-concave, as in Cedarosaurus.In addition, three of the most anterior caudals are laterally compressed, producing the slight distal convexity that led Salgado et aI. (1997) to call them procoelous. The caudal vertebrae of Astrodon lack the lateral pleurofossa found in Venenosaurus, Cedarosdurus, and Malawisaurzs; however, it is possible that these features may develop later in life. In those Early Cretaceous titanosauriformes where the humerus is preserved, most display the elongate shaft found in Astrodon, in-
cluding Cedarosaurus, " P elorosaurus," Phuwiangosaurus, Chubutisanffus, and Malawisaurus. This elongate morphology persists into the Late Cretaceous, and it is found in the titanosaurs Laplatasaurus, Rapetosaurus, Lirainosaurus, Isisdurus colberti, and the juvenile titanosaur referred to Alamosaurus by Lehman and Coulson (2002), although it is not found in the adult specimen described by Gilmore (1946).The Astrodon humerus lacks a dorsal prominence lateral to the humeral head, as do most Early Cretaceous titanosauriformes; this process is only reported in Malawisaurus and Paralititan.In contrast, most Late Cretaceous titanosauriformes developed this process Reassessment
of the Early Cretaceous Sauropod Astrodon iohnsoni Leidv 1865
.
107
to varying degrees, except for Laplatasaurus, Isisaurtts colberti, and the juvenile material of Alamosaunts. All of the very slender radii in Astrodon more closely resemble
in Ceddrosaurus, Maldtuisaurus, and Rapetosaurus than those in the much heaver radii found in Venenosattrus, Chubutiscturus, and I'Jeuquenscturus. Neither the strongly developed, oblique ridge extending along the shaft nor the prominent distal condyles are found rnVenenosaurus or Chubwtisdurus. A complete set of metacarpals, unfortunately, is not known for Astrodon; just individual elements, all of which are very slender, as those
in Ve n en o s aur u s, Mal aw i s aur u s, Lap I a t a s dunzs, and Rap et o s auru s. Many of the Astrodon metacarpals possess a posterior ridge or process arising near the proximal end, and extending distally just past the mid-shaft region. Similar processes are found in Venenosdurus, Laplatasaurus, and R(tpetosaurus. However, in Antdrc' tosaurus and Chubutisdurws, the posterior process is located on the distal portion of each metacarpal, rather than proximally. Two characters often assigned to titanosauriform femora, a weildeveloped lateral bulge on the proximal end of the femur and a medial deflection of the femur head, are developed to varying degrees among the broad range of titanosauriform taxa' The lateral bulge shows an uneven distribution in titanosauriformes. Although clearly present in Chubutisdurws, I:Ieuquensdurus, Sabasaurus, and Rapetosdurus, in other taxa it is developed only moderately or poorly, such as in Cedarosaurus, Phuwiangosdurus, Aegyptosaurtts, and Argyrosaurus. Although this character is often difficult to assess, Salgado et al. (1997) attempt to quantif it among several sauropod taxa. It should be noted that moderate development of this character is also present in Apatosaurus femora (YPM 1980; NSMT 20375; casts on display at the Wyoming Dinosaur Center, Thermopolis, 'Wyoming). The medial deflection of the femur head is even more difficult to assess between taxa due to a number of factors: the relative length of the medial and distal condyles, which influences the inherent tilt of the femur; difficulty in pinpointing the greater trocanter on weathered or eroded proximal ends; and compression or taphonomic distortion on the proximal end. In only a few titanosauriform taxa is this character unmistakable: Astrodon, Cbttbutisaurus, Phuwiangosaurus, and Neuquensaurus. ln Antdrctosdurus it is variably developed, whereas in most others it is less apparent. Marsh (1898) iisted several characters in establishing the family Pleurocoelidae; however, all of the characters occur in other sauropods. Although Astrodon can be separated from other sauropods, there is no suite of unique characters that would war'We aiso note that there has been an errorant a separate family. neous assumption beginning with Marsh ( 1 8 8 8 ) that Astrodon is a small sauropod. However, as Kranz (1996) has noted, considerably larger specimens are known, including a femur that, when whole, would have been 1.25-1.50 m long. Such a large size piaces Astrodon well within the size range of another Lower Cretaceous brachiosaur, Cedarosaurus (see Fig. 3.20).
108
.
Kenneth Carpenter and Virginia Tidwell
Historically, there has been some concern regarding comparison of the juvenile Astrodon elements with those belonging to other adult sauropod taxa. Only a ferv studies have looked at ontogenetic variation in sauropods (Carpenter and Mclntosh 1994; Vilhite and Curtice L998; Martin 1994 Wilhite 1999; Tidwell and Wilhite this volume; Ikejiri et al. this volume). However, these studies suggest some general trends in morphologic change that may be applied to understand possible ontogeny in Astrodon An adult specimen of Astrodon is predicted to have more elongate cervical vertebrae, although never reaching the proportional length found in Brachiosaurus or Sauroposeiden; increased complexity of the pleurocoel cavity; or possible development of pleurofossa in the anterior caudal vertebrae. Less variability between juveniles and adults is found in the limbs, suggesting that adult Astrodon lin.'bs would closely resemble those of juveniles. Because a phyletic definition of the titanosaurids is presently underway by Curry-Rogers, no attempt has been made to replicate or impinge upon her work. Acknowledgments.'We thank Dr. Mike Brett-Surman and Mr. Robert Purdy (National Museum of Natural HistorS'Washington, D.C.) for access to the sauropod specimens from the Arundel Formation, and for the Brdchiosattrus skull from Marsh-Felch Quarry I. We also thank Mr, Cliff Miles (\Testern Paleontological Labs, Lehi, Utah) for a cast of the disarticulated skull of Camdrdsaurus grandis. V. Tidwell would also iike to thank the many individuals -"vho gave her access to the sauropods in their collections: Dr. Jose Bonaparte and Dr. Fernando Novas (Museo Argentino de Ciencias
F
ig. 3.20.
Skel etal r econstr uctioTt
o/Astrodon johnsoni. To
sc.ale
skeleton as a juuenile, use scale A; as an adult, use B. Courtesy of Greg Paul.
Reassessment of the Earlv Cretaceous Sauropod Astrodctn iohnsoni Leidy
1865
.
109
Naturales, Buenos Aires, Argentina); Dr. Oliver Rauhut (Museo Paleontologico Egidio Feruglio, Trelew, Argentina); Dr. Marcelo Requero (Museo Argentino de Ciencias Naturales, La Plata, Argentina); Dr. Leonardo Salgado and Dr. Jorge Calvo (Museo de ia Universidad Nacional del Comahue, Neuquen, Argentina); Dr. Rodolpho Coria (Museo Municipal Carmen Funes, Plaza Huincul, Argentina), Ms. Sandra Chapman (Natural History Museum, London, England), Dr. Martin Munt (Dinosaur Isle, Isle of V/ight, England). References Cited
Blows, \7. T. 1995. The Early Creraceous brachiosaurid dinosaurs Ornithopsis and Eucamerotus ftom the Isle of Wight, England. Palaeontology 38: 187-197. Bonaparte, J.F. 1999. Evoluci6n de las v6rtebras presacras en Sauropodomorpha. Ameghiniana 36: 115-187. Bonaparte, J. F., and R. A. Coria. 1993.Un nue\ro )'gigantesco sauropod titanosaurio de la formacion Rio Limay (Albiano-Cenomaniano) de ia provincia del Neuquen, Argentina. Armeghiniana 30: 27 1-282. Borsuk-Bialynicka, M. 1977. A new caffrarasaurid sauropod Opisthocoelicaudia skarzynskii, gen.n.sp.n. from the Upper Cretaceous of Mon, golia. Palaectntologicd polonica 37: 5-64. Britt, B. B. 1997. Postcranial pneumaticity. In P.J. Currie and I(. Padian, eds., Encyclopedia of Dinosaurs, 590-593. San Diego: Academic Press.
Brochu, C. A. I996. Closure of neurocenrral surures during crocodilian ontogeny: implications for maturity assessment in fossil archosaurs. Journal of Vertebrate Paleontology 16: 49-62. Calvo, J. O., and J. F. Bonaparte. 1991. Andesaurus delgadoi gen et sp. nov. (Saurischia-Sauropoda), dinosaurio Titanosauridae de la formacion Rio Limay (Albiano-Cenomaniano), Neuquen, Argentina. Ameghiniana 28: 303-3 1 0. Carpenter, K., and J. L Kirkland . 1998. Review of Lower and Middle Cretaceous ankylosaurs from North An-rerica. In S. G. Lucas, J. I. Kirkland, and ril/. Estep, eds., Lower and Middle Cretdceous Terrestrial J.
Ecosystems, 249*270. New Mexico Museum of Natural History and Science Bulletin, no. L4. Albuquerque: New Mexico Museum of Nat-
ural Hisrory and Science. Carpenter, K., and J. Mclntosh. 1994. tJpper Jurassic sauropod babies from the Morrison Formation. In K. Carpenter, K. Hirscl-r, and J. Horner, eds., Dinosattr Eggs and Babies, 265-278. New York: Cambridge University Press. Cifelli, R. L., J. D. Gardner, R. L. Nydam, and D. L. Brinkman. 1997. Additions to the verrebrate fauna of the Antlers Formation (Lower Cretaceous), southeastern Oklal.roma. Oklahoma Geology No/es 57(4):
124-131.
Cifelli, R. L., R. L. Nydam, J. D. Gardneq A. \(/eil, J. G. Eaton, J. I. Kirkland, and S. K. Madsen. 1999. Medial Cretaceous vertebrares from the Cedar Mountain Formation, Emery Count1,., Utah: Tl-re Mussentuchit local fauna. In D. Gillette, eds., Vertebrate Paleontologl, in utah, 219-242. Utah Geological Survey Miscellaneous Publications, no. 991. Salt Lake Cit,v: Utah Geological Survey.
110 . Kenneth Carpenter and Virginia Tidwell
Curry Rogers, K., and C. A. Forster.2001. The last of the dinosaur titans: A new sauropod from Madagascar. Nattte 41'2: 530-534. Doyle, J. A,. 1992. Revised palynological correlations of the lower Potomac Group (USA) and the Cocobeach sequence of Gabon (Barremian-Aptian). Cretaceous Research 73: 337 -349. Gallup, M. R. 1974. Lower Cretaceous dinosaurs and associated vertebrates from north-central Texas in the Field Museum of Natural History. Master's thesis, University of Teras at Ausrin. 1989. Functional morphology of the hindfoot of the Texas sauropod Pleurocoelus sp. indet. In J. O. Farlow, eds., Paleobiology of the Dinosaurs, TL-74. Geological Societ.v of America Special Paper, no. 238. Boulder, Colorado.
Gilmore, C. \7. 1921. The fauna of the Arundel Formation of N{aryland. Il.S. National Museum Proceedings 59: 58 1-594. 1946. Reptilian Fauna of the North Horn Formatktn of Central Utah. tJ.S. Geological Survey Professional Paper, no. 21 9-C.'Sfashing-
ton, D.C.: U.S. GPO. Gimenez, O. 1992. Estudio preliminar del miembro anterior de
1os
sauropodos titanosauridos. Ameghiniana 30 754. Gomani, E. M., L. Jacobs, and D. A. \7inkler. 1999. Comparison of the African titanosaurian MaldwisaLtrus, with a North American Early Cretaceous sauropod. In Y. Tomida, T. H. Rich, and P. Vickers-Rich, eds., Proceedings of the Second Gonduanan Dinosaur Symposium, 223-233. National Science Museum Monograph, no. 15. Tokyo: National Science Museum.
Gregory, \7. K. 1918. Note on the morphology and evolution of the femoral trochanters in reptiles and mammals. American Museum of Natural Histctry Bulletin 38: 528-538. Hatcher, J. 1903. Discovery of remains of Astrodon lPleurocoelus) in the Atlantosaurus beds of \ff/yoming. Annual Report Carnegie Museum 2: 9-74. Hoffstetter, R., and J.-P. Gasc. 1969. Vertebrae and ribs of modern reptiles. In C. Gans, ed., Biology of the Reptilia, 20I-310. London: Academic Press.
Huene, F. i. t92q. Los saurisquios v ornitisquios del Creticeo Argentino. Anales del Museo de La Plata 3: 1,-1'94. 7932. Die fossile Reptil-Ordnung Saurischia, ihre entwick; und geschichte. Gectlogie und Palaeontologie Monographien 1(4): 1-361. 'Sflinkler,'!7. R. Downs, and E. M. Gomani. 1993' Ner.t' Jacobs, L. L., D. material of an Early Cretaceous titanosaurid sauropod dinosaur from Malawi. Palaeontology 36: 523-534. Iain, S. L., and S. Bandyopadhvay. 1997. New titanosaurid (Dinosauria: Sauropoda) from the Late Cretaceous of central lndia. Journal of Vert
ebr ate
P aleontolo
gy
1,7
:
1,
t 4-1' 3 6.
Janensch, W. 1935-1936. Die schddel der sauropoden Brachisaurus, Barosaurus und Dicraeosdurus aLts den Tendaguruschichten DeutschO stafrikas. P al a e ctnt o gr ap h i ca, Supplement 7 ( 3 ) : 1 47 -29 8. 1950. Der t'irbelsaule von Brachiosatrus brdncai. Palaeontogrctp h icd, Supplement 7 (3\ : 27 -9 3. Kellner, A. \(. A., and S. A. K. de Azevedo. 1999. A new sauropod dinosaur (Titanosauria) from the Late Cretaceous of Brazii. In Y. Tomida. T. H. Rich and P. Vickers-Rich, eds., Proceedings ctf the Second
Gonduanan Dinosaur Symposium, 111442' National Science Museum Monograph, no. 15. Tokyo: National Science Museum.
Reassessment
of the Early Cretaceous Sauropod Astrodon iohnsoni Leidy 1865
.
111
Kingham, R. F. 1962. Studies of the sauropod dinosaur Astrodon Leidy. Proceedings of the V/ashington Jtutior Academy of Sciences 1: 38-43. Kirkland, J. I., R.L. Cifelli, B. B. Britt, D. L. Burge, F. L. DeCourten, J. G. Eaton, and J. trI. Parrish. 1999. Distribution of vertebrate faunas in the Cedar Mountain Formation, easr-cenrral Utah. In D. Gillette, ed.. Vertebrate Paleontobgy in Utah, 20I-218. Utah Geological Survey Miscellaneous Publications, no. 99-1. Salt Lake Cit,v: Utah Geological Survey.
Kranz, P. M. 1996. Notes on the sedimentary iron ores of Maryland and their dinosaurian fauna. Maryland Geological Sttruey Special Pttblica-
tion 3:87-775. 1998. Mostly dinosaurs: A revierv of the vertebrares of the Potomac Group (Aptian Arundel Formation), USA. In S. G. Lucas, J. I.
Kirkland, and J. \7. Estep, eds., Lower and Middle Cretaceous Terrestrial Ecosystems,235-238. New Mexico Museum of Natural Histr,try and Science Bulletin, no. 1.4. Albuquerque: New Nlexico Museum of Natural Historv and Science. Langston, W. 1974. Nonmmmalian Comanchean tetrapods. Geoscience and Man 3:77-102. Lapparent, A. P. 1960. Les Dinosariens du "Continenral Intercalaire" du Sahara Central. Mdmoires de la Soci,lte G|ologiqtte de France 88 1-56.
M., and A. B. Coulson. 2002. A juvenile specimen of the sauropod dinosaur Alamosaurus sanjuanensis from the Upper Cretaceous of Big Bend National Park, Texas. Journal of Paleontology
Lehman, T.
76(7\:156-172.
Leidn J. 1865. Cretaceous reptiles of the United States. Smllbsonian Contribution to Knowledge I92: L-1,35. Lovelace, D., ril/. ril/ahl, and S. Hartman. 2003. Evidence for costal pneumaticit)' in a diplodocid dinosaur (Strpersaurus uiuanael. Journal of Vertebrate P aleontoktgy 23 (31 7 3 A. Lucas, S. G., and R. M. Sullivan. 2000. The sauropod dinosaur Alamosaurus from the Upper Cretaceous of the San Juan Basin, New Mexico. In S. G. Lucas and A. B. Heckert, eds., Dinosaurs of New Mexico, 147-156.
New Mexico Museum of Natural History and Science Bulletin, no. 17. Albuquerque: New Mexico Museum of Natural History and Science. Lull, R. S. 1911. The Reptilia of the Arundel Formation. Maryland Geological Suruey: 17 1-211. Madsen, J.H., J.S. Mclntosh, and D. S. Berman. 1995. Skull and atlasaxis complex of the Upper Jurassic sauropod Camardsaurus Cope (Reptilia: Saurischia). Bnlletin of tbe Carnegie Museum of Natural History J l: l- I 15. Makovicky, P. 1997. Postcranial axial skeleton, comparative anatomy. In P. J. Currie and K. Padian, eds., Encyclopedia of Dinosaurs, 579-590. San Diego: Academic Press.
Marsh, O. C. 1888. Notice of a new genus of Sauropoda and other new dinosaurs from the Potomac Formation. American Journal of Science
135:89-94.
1896. The Dinosaurs of North America. U.S. Geological Survey Annual Report, no. 16. lWashington, D.C.: GPO. 1898. On the families of sauropodous Dinosauria . American .lournal of Science 136: 487-488. Martin, V. 1994. Baby sauropods frorn the Sau Khua Formation (Lower Cretaceous) of northeastern Thailand. Gaia L0: I47-153. Martin, V., V. Suteethorn, and E. Buffetaut. 1999. Description of the type 11.2 . Kenneth Carpenter and Virginia Tidwell
and referred material of Phttwidngosattrus sirinhornae Martin, Buffetaut and Suteethorn, 1994, a sauropod from the Lower Cretaceous of Thailand. Oryctos 2: 39-97. Maxwell, \7. D., and R. L. Cifelli. 2000. Last evidence of sauropod dinosaurs (Saurischia: Sauropodomorpha) in the North American midCretaceous. Brigham Young University. Geology Studies 45: 79-24Mclntosh, J. S. 1990. Sauropoda. In D. B. \feishampel, P. Dodson, and H. Osm6lska, eds., The Dinosauria, 345-401. Berkeley: Universit.v of
California, Press. Mclntosh, J. S., \7. E. Miller, K. L. Stedtman, and D. D' Gillette. 1996a. The osteology of Camarasaurus lewisi (Jensen, 1'9881. Brtghant Young [Jniuersity Geology Studies 41 73-115.
Mclntosh, J. S., C.A. Miles, K. C. Cloward, and J. R. Parker. 1996b. A new nearly complete skeleton of Camarasaurus. Gunma Museum of Natural History, Bulletin 1: 7-32. Ostrom, J. H. 1970. Stratigraphy and paleontology of the Cloverly Formation (Lower Cretaceous) of the Bighorn Basin area, \Tyoming and Montana. Peabody Museum of Natural History Bulletin 35 7-234. Ostrom, J. H., and J. S. Mclntosh. 1966. Marsh's Dinosaurs. New Haven, Conn.: Yale University Press. Powell, J.E.7979. Sobre una asociacion de dinosaurios y otras evidencias de vertebrados del Cretacico superior de la region de la Candelaria pror.. de Salta, Argentina. Ameghiniana 76: 191-204. 1986. Revision de los titanosauridos de America del Sur. Doctoral thesis, Universidad Nacional de Tucuman Facultad de Ciencias Naturales, Tucuman. Argenrinir' 1.990. Epachthosaurus sciuttoi (gen. et. sp. nov') un dinosaurio
sauropodo del Cretacico de Patagonia (Provincia de Chubut, Argentina). V Cortgreso Argentino de Paleont euru> Quaruy in Santloual Countt,' 1.0 = Stouall Pits in Kenton; and 11 = Grand Valley in Mesa County.
-l;-;;Js.* 40"/o aduh size (YPM 1910 and
Fig. 6.1 1. Ocarrences of Camarasaurus species in the Morrison Formation (Modifi ed from Turner and Peterson 7999; Ikeiiri 2002). Symbols = identified species. X = Camarasaurus sP. Abbreuiations for quarries: 1 = Hou,e Quarry in Big Horn
CM 11338). Our conclusions rein Table 6.4 and elaborated
garding ontogeny are summarized below.
The cranial skeleton shows few ontogenetic changes besides (CM 11338, USNM 17863) are already tightly fused (Madsen et al. 1995,11). The braincase is also fused and there is no significant difference from the suture pattern of the WDC specimen. The juvenile CM 11338 does show a taller anterior margin of the premaxillae than the subadult USNM 77863. However, the'WDC specimen also shows a size. The frontals and parietals of the juveniles
New Adult Specimens of Camarasattrus lentus Highlight Ontogenetic Variation
.
171
TABLE 6.4. List of Ontogenetic Characteristics and Four Representative Growth Stages in Camarasaurus lentus and Other Species in the Genus except Embrvo Stage Stage General
age.
Representative
I
Srrgc
Subadult
.jur enile
speci-
YP\,I
CM11338,
mens in Cdmarasau- 1910 rus lenttts
USNN.I 17863, 1
1901, OMNH specimens"" Pleurocoel complex- Sirnple, no iossae (in ity in cervicals all specirnens)
of
Other species tndrdsdLt/us
(rrqe
2
Ca- YPM baby
CNI
1393
Stro. 4
l
Adult \(/DC-B, Clt{
8492,
Verv old individurl \(/DC-A
71,069*
YPN1 190,{,
ONINH
190.5,
1173
Simple, no fossae all)
AMNH 5760, 5767*"*, GNINH
(in
101
and/or
Shallor.v rooms fossae (in all)
ANINH 5761*"*. BYU 9047 Nlanv fossae and
tinl' laminae
(ir.r
all)
11338)
Atlas knob and hook- Absent (CM like process ("see
Absent
(YP\I 1905,
Srnall IUUVP
usN\,r 17863, ON,INH GNINH-PV
text)
2983,
101)
Developed iBYU 90,+7,
\ilDC-A)
r173)
Over-grorvth bone on cervical neural
Absent (in
all)
Absent (in
all)
Absent (GMNH-P\/ 101), present (NfDC-B)
Absent (in
all)
Absent (in
all)
Absent (in
all)
Present (in all)
Absent (in
all)
Absent (in
all)
Absent (in
all)
Prcscnt (in all)
Present (in all)
spine
Ossified ligament on lateral neural spine
of posterior dorsals Over-gror.vth bone on cleft of bifurcated dorsal neural spines Tin.v laminae on
sal neural
arch
dor- No (YPl,l
Ossified ligamer.rts
in
sacrum
1901, CNI 11338, ON{NH spec-
No (USNN{ 17863, CM 11393, YPIvI
lmens)
1904,1905) No (USNM 17863, YPM 1904, 190s) Identical sutures (usN\,I 17863, CM
No (YPNI 1901, 1
Clvi
1338)
Neurocentral closure Opened (YPM in cen icals and dor- CM11338)
1901,
sa]s
1
closure in
rib
sacrum
Fused sacral
ribs;
closure Opened (CN'I 11338, caudals YPI,I 1910)
Neurocentral
closure
in mid-caudals ("rvith tall neural splnes
Cllosed (in
101)
all)
9047 }'es (\(rDt--A, BYU 9047) Closed (in all)
17863)
Fused
101,
(GNINH-PV 690)
AMNH
F'uscd {WDC-A,
BYU 90,+7)
s
Neurocentr:al
in anterior
No {GNINH'PV
\VDC-A, BYU
)
Fused (USNN.{
identical neurocentral closure
CNI 101)
8492, GX{NH-PV
1393, YPIVI 1904,
190.5
Neural arch and
No (\(DC-B,
Nearl,v closed (USNM Closed (in 17863, YPM 1904, 1905)
all)
Closed (in all)
(WDC-B, 101)
Visible (YPNI
1901, Closcd (USNNI 17863, OI,{NH YPM 1904, 1905)
Closed
CN'I 11338,
GNINH-PV
all)
Closed and no line (in all)
Ckrsed (\fDC-A,
BYU 9047)
specimens)
)
Rib closure in
ante-
Opened (in
rior caudals
Opened but visible tures (USNtrI
su-
17863,
suture
Closed (in all)
YPlvl 190.+, 1905) Limb articulated
sur-
faces
Pubic
foramen
Coracoid
772
.
foramen
Shallow rugosit,v only Shallorv pitted rugosity Relatrvely deeplv pitted Decply pirted on around external mar- on the nearlv entire ar- rugosity on thc entire the entire articular gin of articular sur- rrcular surface (in all) articular surface (in al1) surface (in all) face (in all) Opened
Opened
1905, 17863)
Closed (YPM 1910), Closed (YPM opcned (CNI 11338) USNI,{
Takehito Ikejiri, Virginia Tidwell, and David L. Trexler
Closed Closed (AMNH 5767)
Closed
5760,
Closed (S7DC-A)
Steoe
Coracoid-scapula closlrre Size of chevron
facets
I
(t:ue
Separate (YPM 1910)
Visual suture (USNNI
Small (in all)
Small (in all)
1
7863
Srr caudal facet. 'White (1958) Tooth Replacement Pattern. briefly discussed the tooth repiacement pattern in Camarasdzrurs. DINO 28 includes the left premaxilla and marilla rvith intact tooth rows, providing an opportunity to study the replacement pattern in detail. Because of crown-to-crown occlusion, various wear facets are present throughout the tooth rows. The morphology and wear facets of the tooth series from front to back are described here (Fig.9.a). There are four teeth in the premaxilla. Since the interdental plates are intact in CAmtrasaurus, the size and location of replacement teeth are not clearly visible as rn Shunosattrus. The first premaxillary tooth is a replacement tooth; it has erupted but not reached the occlusal level. Its crown lies below the labial margin. The crown in mediai aspect is roughly triangular with a pointed apex. From this shape, we can determine the degree and nature of wear in the following tooth series. The second tooth is fully erupted, with extensive wear. The occlusal facet is small and relatively flat. The rostral facet is moderate near the top of the crown. The caudal facet is extremely excavated with a nearly horizontai surface. The third tooth is shorter than the preceding one; it has just reached the occlusal level and the crown tip has suffered wear. The fourth functional tooth is missing, but was present in life. The replacement tooth can be observed in the caudal section that articulates with the rnaxilla. This replacement tooth is far beneath the interdental plate of the premaxilla and is impossible to see medially. The crown, compietely formed, is the same length as in the functional tooth during its initial development. The replacing tooth generally migrates to the position of the old tooth during its development. Caudally of the four premaxiilary teeth, there are ten teeth in the maxilla that show a gradual reduction of height caudally. The fifth tooth (first maxillary tooth) crown is young, with a little occlusal wearing. The sixth tooth is fully erupted and has a long root. Crown wear is so extensive that it has produced a horizontal platform. The seventh tooth is also fully functional, but the wear is not Neuroanatonrv and Dentition of Camarasdurus lentus
.
207
TABLE 9.1. Tooth morphology and measurements oI Camarasaurus
Tooth
Occlusal
facet
Rostral
facet
Caudal
facet
'Wear
stages
T
no
no
no
R3
2
yes
yes
,ves
F3
3
yes
no
no
F2
+
no
no
no
R1, F4
5
yes
no
no
F2
6
yes
r'la
nla
F5
7
yes
yes
)'es
F3
8
yes
nla
nla
F5
9
yes
no
no
F4
10
no
no
no
F1
11
yes yes
nla nla
F4
12
nla nla
13
no
11C)
F1
74
no
no
no no
F3
R2
extensive. The occlusal facet is small and the rostral moderate. The caudal facet is deeply truncated to produce a pointed tip on the
crown. The eighth tooth is probably the oldest tooth in the series and shows an extensive wear facet that forms a horizontal surface. Rostral and caudal facets have lost their identity and all are merged into a single platform. The wear surface has reached the maxrmum area. The ninth tooth is also fully functional, with a long root. The wear facets in the crown have just begun to merge. The caudal facet is sharply truncated. The tenth tooth is young with an intact crown and lacks wear. The eleventh tooth is functional, with a prominent facet on the occlusal surface. This facet is obliaue forward and more outward than those on other teeth. Rostral and caudal facets can not be identified. The twelfth tooth is smaller than the preceding with a narroq elongate occlusal facet. As in the eleventh, there are no rostral or caudal facets. The thirteenth tooth is a new replacement tooth and lacks a wear facet. The fourteenth tooth is also a replacement tooth and occupied its position after the loss of the old tooth. The tooth replacement pattern (Table 9.1) is inferred from the degree of wear and the relative height of the root. Five stages of
functional teeth (F1-F5) and three stages of replacement teeth (R1-R3) are recognized on the left maxilla, based on the tooth exposure, tooth size, length and relative position in the alveoli, and wear facet. In order of increasing age, the stages are: R1, small incipient tooth showing the tip of the crown. R2, fully erupted crown.
208
.
Sankar Chatterjee andZhongZheng
1234567A91011121314 F5
\
F4
\
F3
\
F2
\
F1
\
R3
R2
\
\
\
\
\
\ \
\
RI
Fi1.9.5, Z-spdcing diagram for Camarasaurus lentus (DINO 2B). (See Table 9.1 for rau' data.)
R3, the crown has reached the labial margin of the maxiila and is completely exposed. Similarlv. for functional teeth:
tooth is young but fully developed, with a long crown and long root, but it lacks a wear facet. F2, the crown tip is beginning to show a wear facet. F3, a wear facet is present not only on the occlusal, but also on the rostral or caudal side; however) the crown apex is still triangular. F4, the wear facet is expanded so that the apex of the crown becomes rounded. F5, three facets are merged to form a horizontal wearing F1, the
surface. Based on these stages, the Z-spacing was measured (Table 9.1).
Camarasaurzs shows a fairly well organized replacement pattern. The Z-spacing varies between 2.0 and 3.0 (Fig. 9.5). These values indicate that the replacement waves moved from back to front (Osborn 1972). The individual Zahmethen have a consistent slope. This slope becomes steep in the lower part, but the angle abruptly decreases in the last two stages. The Zahnreihen slope varies from 1 to 4 in Cantarasaurzls. In most reptiles, the Z-spacing lies in the range of 1.56 to 2.80 (DeMar 1972). The tooth replacement pattern of Camarasaarzs provides the fol lowing information:
1. The replacement series (odd and even) are from the back to front; the replacement wave direction is independent of the tooth nrorphology and tooth number, but relies on the Z-spacrng. 2. The average Z-spacing is 2.5 in Camarasdurus; the individual or u'hole Zahnreihen are highly organized. The slope between them is of slight variatron. 3. The number of tooth-wear facets strongly suggests the upper and lower teeth must occlude in alternate fashion and that the occlusion is orthal. 4. In his Camarasaunrs stud-v. \7hite (1958) observed anomalies and suggested that the replacement pattern in the odd series of Neuroanatomv and Dentition of Camarasaurus lentus
.
209
maxillary teeth is from back to front, whereas in the even series, it is from front to back. However, Z-spacing analysis gives a better picture. Using the same measurement data (see White L958, 491, table) but on the Z-spacing mapping, the replacement pattern in both series becomes back to front, as in most reptiles. Following Osborn's equation (1,972\, the average Z-spacing in Camarasaurus is 2.5 (Z = 2.5), and counting alternatively (R = 2), the replacement
wavelength should equal five teeth. This analysis contradicts White's observation. 'lfhite (1958) estimated a replacemenr wave of three, with each wave length corresponding to rwo or three teeth.
Acknowledgments. \X/e thank Kenneth Carpenter and Virginia Tidwell for inviting us to conrribute rhis paper and also for edito-
rial
assistance. This paper resulted
in part from Zhong
Zheng's
Ph.D. thesis research, conducted under the supervision of S. Chatterjee at the Museum of Texas Tech University. V/e thank Dan
Chure
of Dinosaur National Monument for the opportunity
ro
study the beautiful skull of Camarasaurus lentus. We thank Mike Nickell and Jeff Martz for the drawings. The research was supported by Texas Tech University. Appendix: Abbreviations for Figures
Braincase angular aof: antorbital fenestra ar: articular bo: basioccipital bpt: basipterygoid process bs: basisphenoid bt: basisphenoid rubera ch: choana d: dentary ec: ectopterygoid en: external naris eo: exoccipital eov: foramen for external occipital vein f: frontal fm: magnum fenestra ic: internal carotid artery idp: interdental plate j: jugal l: lacrimal ls: laterosphenoid t*r. t-.^-^l .--^oral fenestra m: maxilla n: nasal o: orbit op: opisthotic os: orbitosphenoid p: parietal Skull and an:
210 . Sankar Chatterjee andZhongZheng
popr: paroccipital pl: palatine pm: premaxilla po: postorbital ppf: post palatine fenestra pra: prearticular
prf: prefrontal pro: prootic ps: parasphenoid psf: postfrontal pt: pterygoid ptf: post temporal fenestra ptr: pterygoid ramus q: quadrate qj: quadratojugal sa: surangular snf: subnasal fenestra soc: supraoccipital sp: splenial sq: squamosal sym: symphysis
utf: upper temporal fenestra v: vo[ler Endocast cel: cerebellum cer: cerebral die: diencephalon fo: fenestra ovalis
lnt
e:
internal ear region
u, Lvr ] rr ar ^l'.r. vrr4r ^lf^-'^-".-^^+
L
op l: optic lobe pf: parietal fenestra pit: pituitary body vcm: vena cerabralis medius Cranial Nerues
IV trochlear V trigeminal \/1. ^-hth"l-i.
I: olfactory II: optic
VI: abducens VII: facial IX: glossopharyngeal X: vagus XI: spinal accessory
III: occulomotor
XII: hypoglossal
References Cited
Chatterjee, S., and Z. Zheng. 2002. Cranial anatomy of Shunosaurus, a basal sauropod dinosaur from the Middlle Jurassic of China. Zoological Journal of the Linnean Society 1,36: 1,45-1,69. DeMar, R. 1.972. Evolutionary implications of Zahnreihen. Euolution 26:
435+50. P. M. 1984. Cranial anatomy of the prosauropod dinosaur Pldteosdurus from the Knollenmergel (Middle Keuper, Upper Triassic) of
Galton,
Germany. I. Two complete skulls from Trossingen'Wurtt, with comments on the diet. Geologica et Pdlaeontologica 1,8:139-171. 1985. Cranial anatomy of the prosauropod dinosaur Plateosaurus from the Knollenmergel (Middle Keuper, Upper Triassic) of Germany. II. A1l the cranial material and details of soft-part anatomy. Geologica et P alaeontologica 19 : 119-1 59. Hopson, I. A,. 1979. Paleoneurolog.v. In C. Gans, ed., Biology of the Reptilia. Yol. 9A,39-146. London: Academic Press. Madsen, J.H., J. S. Mclntosh, and D. S. Berman. 1995. Skull and atlasaxis complex of the Upper Jurassic sauropod Camarasaurus Cope (Reptilia: Saurischia). Bulletin of the Carnegie Museum of Natural
History 31: 1-115. Mclntosh, J. S. 1990. Sauropoda. In D. B. 'Weishampel, P. Dodson, and H. Osm61ska, eds., The Dinosauria, 345-401. Berkeley: University of California Press. Osborn, J.W. 1972. New approach to Zahnreihen. Nature 225:343-346. 'S7hite, T. E. 1958. The braincase of Camarasaurus lentus (Marsh). Journdl of Paleontolology 32: 477494. \filson, J. A., and P. C. Sereno. 1998. Early evolution and higher phylogeny of sauropod dinosaurs. Society of Vertebrate Paleontology, Memoir 5: 1-68. Zheng, Z. I996. Cranial anatomy of Shunosaurus and Camarasaurus (Dinosauria: Sauropoda) and the phylogeny of the Sauropoda. Ph.D. dissertation, Texas Tech University.
Neuroanatomv ar-rd Dentition of Camarasaurus lentus
.
271,
L0. Neck Posture, Dentition, and Feeding Strategies in Jurassic Sauropod Dinosaurs KpNr A. SrpvpNS AND J.
MrcHerr
PenRrsH
Abstract The extreme elongation of sauropod necks, some of which were 9 m long, has been used to suggest that they engaged in highly specialized feeding. This common hypothesis of brorvsing on high vegetation is examined by reconstructing the curvature of the necks of representative sauropods in the undeflected, or neutral, pose. The cervical columns of the sauropods Apdtosaurus, Brachiosaurus, Camarasaurus, Cetiosaurus, Dicrdeoscturus, Diplodocus, and Euhelopus were found to be straighr exrensions of their dorsal vertebral columns. None presented the osteological specializations for a more erect neutral pose, such as seen in the base of the giraffe and avian neck. Hence, the elevation of the head, cantilevered far anterior to the shoulders, was determined primarily bv the length of the neck and its slope at the base. All sauropods examined held their heads at or below the height of the shoulder in neutral posrure. Furthermore, most of the sauropod necks studied curved downward cranially. The ventral inclination of the skulls relative to the cervical column in diplodocids and nemegtosaurids may have been a further adaptation for downward feeding at ground level or for
grasping aquatic vegetation ar or near the surface of the r,vater. Previous studies of sauropod dentition have made a distinction between taxa rvith a full batter,v of spatulate teeth, which presumably employed some oral processing and specialized on more resistant forage, and those with a reduced battery of peg-like teeth, which may have fed predominately on softer plants or possibly were used to strip branches. The current knou.ledge ofJurassic floras suggests that the most abundant potential forage for sauropods would be found within the lor,v- to medium-browsing range thar is predicted by our cervical reconstructions.
Introduction Sauropod dinosaurs represent the extremes of both gigantism and neck elongation in the history of terrestrial vertebrares. Neck lengths were extreme in both absolute and relative measures. The necks of the Jurassic sauropods Barosdurus and Brachioslurus were about 9 m long, well over twice the length of their dorsal columns. Even the relatively short-necked Camarasaunzs, r,vith a cervical column about 3.5 m long, had a neck substantially longer than its trunk. Because of their elongate necks, massive bulk, abundance, and diversity, the feeding habits of the sauropod dinosaurs of the Jurassic period have been a subject of inquiry and dispute since relatively complete fossils of this clade were first discovered over a hundred years ago. Analysis of the functional morphology and paleoecology of sauropods has led to a dir.ersity of interpretations of their feeding habits. Sauropods have been interpreted as high browsers, low browsers, aquatic, terrestrial, bipedal, and tripodal, and they have even been restored with elephant-style proboscises. Furthermore, differences in neck length, body shape, cranial anatomlr, and dental morphology indicate significant morphological (and inferred behavioral) variation among the Sauropoda. Here rve will review various lines of evidence relating to feeding in sauropods and reconsider them in light of our own studies on neck pose, mobilit,v, and inferred feeding envelopes in Jurassic sauropods. Sauropods varied considerablv in body plan and size, resulting in a range of head heights across the different taxa, when feeding in a neutral, quadrupedai stance. Reconstructions have often differed in the depiction of a sauropod taxon in neutral pose, both in the placement of the eiements of the appendicular skeleton and the curvature of the axial skeleton. More precise determination of neurral feeding height would permit sorring sauropod raxa into a vertical range of feeding niches, follor,ved with an analysis of extrernes of neck mobi|ty and postural repertoire for more specialized forms of feeding. Dentition morphology and microwear provide additional indirect evidence of feeding preferences (e.g., Barrett 2000; Barrett and Upchurch 1994; Calvo 1994; Fiorillo 7991,1998). Evidence regarding the vegetation consumed by different sauropod taxa can then be correlated with their specific range of feeding heights.
Institutional abbreuiations. CMNH-Carneeie Museum of Neck Posture, Dentition, and Feeding Strategies in Jurassic Sauropod Dinosaurs
.
213
Natural History, Pittsburgh, Pennsylvania; DINO-Dinosaur National Monument, Jensen, Utah; and LCM-Leicester Civic Museum, Leicester, United Kingdom.
Neutral Pose and Intrinsic Curvature along the Axial Skeleton Careful analysis of the undeflected state of sauropod necks is of central importance to understanding their feeding habits. Of all aspects of sauropod biology, perhaps the greatest divergence of opinion has concerned the curvature of the neck. The early reconstructions of most sauropods depicted the necks as cantilevered ahead of the animal, generally descending at the base due to the arch of the back (e.g., Holland 1906; Gilmore 1925,1936;Hatcher 1901; Osborn and Mook 1,927). Only a few sauropod taxa were initially reconstructed as having giraffe-like necks, with sharp upward curvature at the base (Wiman 1929;Janensch 1950b, pls. 6-8). Over the decades since their initial descriptions, however, there has been a general trend toward depicting sauropods as having ascending necks, some with necks much more steeply curved than originally depicted. For instance, Opisthocoelicdudia, a taxon for which the neck is unknown, has been rendered with a swanlike neck by default (Paul 2000,406) contrary to the original description that concluded the neck would have been horizontal or downward-curving (Borsuk-Bialynicka 1977, fig. 19). Paul (2000, 92) suggests that some sauropod necks had thick intervertebral discs, effectively wedged between successive centra, which induced an upward curve at their base. Sauropod necks, however, were strongly opisthocoelous, with central articulations that closely resemble the mammalian opisthocoelous biomechanical design, consisting of condyles that insert deeply in cotyles of matching curvature, leaving little room for cartilage. In modern quadrupeds with opisthocoelous cervicals, such as the horse, gtaffe, and rhino, the central
condyle and cotyle are separated by only a few millimeters. In avians, heterocoely is similarly associated with very precisely matching articular facets and tight intervertebral separations. Across a large range of extant vertebrates, while substantial intervertebral separations are associated with platycoelous vertebrae, vertebrae with nonplanar central articular geometry generally have little intervening cartilage (pers. obs.), and thus little room for conjecture regarding their undeflected state. Neutral deflection. The neutral state of deflection between successive vertebrae is defined geometrically by the alignment of the zygapophyses and by nulling the deflection at the central articula-
tion (Stevens and Parrish 1999). The pre- and postzvgapophyses, if present, are generally centered within their range of dorsoventral travel when the two vertebrae are in the undeflected state. Simultaneously, the central facets will be in a neutral or undeflected state. For platycoelous vertebrae, the two planar articular surfaces are parallel when undeflected, a state particularly easy to verify in lat214
.
Kent A. Stevens and
T.
Michaei Parrish
eral view. Determining the neutral position for opisthocoelous vertebrae requires closer scrutiny of the margins of the central articulation. The synovial capsule surrounding the condyle-cotyle pair at
the centrum generally exhibits circumferential attachment
scars
surrounding the condyle and cotyle. These ridges are parallel when the joint is undeflected, and especially apparent when viewing osreological mounts of extant vertebrates in lateral aspect. Note that
the gap across these margins at the centrum is necessarily wider than the actual intervertebral separation deep within the ball and socket in order to accommodate the displacement of the cotyle during mediolateral and dorsoventral deflection.
The intrinsic curvature of the vertebral column for a given taxon can be determined by placing successive elements in neutral deflection. This procedure can be performed using photographs or engravings in lateral view (see below), provided the illustrations are of verified dimensional accuracy. Whereas two-dimensional analysis is sufficient to establish the neutral pose along a vertebral column, a three-dimensional reconstruction is required to estimate the range of motion and curvature achievable (Stevens and Parrish 1996, 1999; Stevens 2002; Stevens and Parrish in press; Stevens in prep.). There is consistency between the geometrically defined neutral posture and the pose habitually held by the behaving animal, determined by direct manipulation of the cervical vertebral columns of a variety of extant vertebrates. For example, the neutral pose reveals the sigmoid curvature characteristic of avian and equine necks, the catenary shape of the camel's neck, and the sharp upturn at the base of the otherwise straight giraffe neck (see Fig. 10.1). The giraffe neck is particularly relevant ro the reconstruction of some sauropod necks, owing to the historical and persisting interpretation of some sauropods as giraffe analogues, especially as regards the presumed upturn at the base of the neck. The adult giraffe neck is sharply angled at its base while held in the undeflected, neutral position (Stevens and Parrish in press, fig. 1).This elevation arises not from deflection at the intervertebral joints, but from keystone-shaped cervicothoracic vertebrae, the most wedge-shaped being C7 in giraffe.'With no known exception, the curvature characteristic of the axial skeleton of a given vertebrate arises, not from chronic flexion out of the neutral position, but from the morphology of the vertebrae in the undeflected state.
Neutral Posture of the Presacral Axial Skeleton For a number of well-preserved sauropods, the original descriptions provided accurate illustrations of the vertebrae, often as steel engravings based on photographs taken of the prepared vertebrae prior to mounting. The individual lateral views provide a valuable resource for reconstructing the neutral posture of their axial skeletons (Stevens in prep.). For example, Figure 10.2 shows a digital composite of the individual steel engravings of the presacrals of DlNeck Posture, Dentition, and Feeding Strategies in Jurassic Sauropod Dinosaurs
.
215
Fig. 1 0.1 . Extant uertebrates tend to l:abitLtally hold their necks in d neiltrdl stdte of deflection. The characteristic cdtendr,y shape of Il e c.tntcl neck tA1 and tbe ot ian sigtnoid neck (domestic turkey, B) is tntluced by ker*stone-shaped t' e rte br ae. Keystone- sh ap ed centra are also found in non-auian theropod dinosaurs, but haue not
bte,t,th sert'ed i n
s,t u
ropod
s.
216 . Kent A. Stevens and J. Michael Parrish
:l ,l , ,l
crdeosaurus hansemanni (from Janensch 1,929a, pl. 1). The individual images were adjusted for scale, then rotated and translated into neutral deflection using the layering option in PhotoshoprN{. The opacity of each layer was decreased to make the vertebrae partially transparent. The composite image thus reveals the insertion of the
central condyle within cotyle, and the centering and superposition of the zygapophyses. The dorsal column of Dicraeosaurus is distinctly arched, but the cervicodorsal transition is straight, as we have found in other sauropods. The neck is gently curved ventrally, a droop also observed in digital composites of the diplodocids Apatosaurus louisae (Fig. 10.5) and Diplodocus carnegii (Stevens in prep.) and in some other sauropods including Brachiosaurus brancai (see below) and Cetiosaurus (Fig. 10.3A). This technique, using illustrations of the original material, can be used to revise the interpretation of the seemingly giraffe-like sauropods. For instance, Euhelopus zdanskyi (lX/iman L929; Fig. 10.3B) was depicted with a neck ascending at about 38' from the horizontal; in subsequent reconstructions this slope increased to about 65' (Mclntosh et aL.7997,Fig.20.9; Paul 2000, appendix A). The remarkable linearity in the neck from C1 to C16 (\fiman 7929, pl.3) in fact extends through the cervicodorsal region when
Fig. 1 0.2. Dicraeosaurus hansemanni presacral uertebral column reconstruction in neutral pctsition, a composite of the p r e s a cr a I u erte b r d e indiu idual ly figured in Janensch (1929, pl.1). lnset: superposition uf compusite onto reconstruction in Janensch (192e, p1. 1 6).
Neck Posture, Dentition, and Feeding Strategies in Jurassic Sauropod Dinosaurs
.
217
":""dplrlr *"}9"'
/-\tr-4sj - /t)a1i-_
Fig.
1
0.3. Digital reconst/uctions
of tbe neutral pose of the necks of tbree sauropods ftrom Steuens tin
prep.). (A) Cetiosaurus /lCM C468.1968) is shoun in neurdl
pose by compositing the ceruical u er t e b r ae indiu idually illustr ate d b1, John Martin. (B) Euhelopus zdanskyi (from'V{iman 1929, pl. 3), with C1.7 and D1 rotated
digitally to remoue the postmortem " death pose" dor siflex ion. /C/ Brachiosaurus brancai specimen SII neutral pose composite of ceruicals C3-D2 (from Janensch 1950a, figs. 32-19). The original ,ndterial shows none of the giraffe-like keystone shape depicted in Mclntosb et al. (1997, fr7. 20.16) eleuated neck in neutral positiun
shuwn) (sec Steuens 2002. fig. 70.2; Steuens and Parrish in press).
218
.
Kent A. Stevens and
T.
care is taken to remove the postmortem dorsiflexion "death pose" posterior to C17, as shown in Figure 10.3B (Stevens and Parrish in press; Stevens in prep.). Eubelopus was apparently a low browser, not a "Juraffe." Death-pose dorsiflexion was also responsible for the swanneck pose of the juvenrle Camarasdurus lentus (Gilmore 1,925, pl.
14). The postzygapophyses are in fact displaced posteriorly far from the neutral position, with many completely out of articulation (Parrish and Stevens 1998). The posterior cervicals and anterior dorsals show no evidence of the wedge shape needed to induce curvature (Osborn and Mook 1921, pI. 68, DINO 28; CM 11338; CM 11069). Despite the popular depiction of Camarasaurus with a sharply upturned neck, the original reconstruction (Fig. 10.4) showing the cervicodorsal transition as horizontal and linear is consistent with recent mounts of actual fossil material (e.g., at the 'Wyoming Dinosaur Center and "Annabelle" at the Natural History Museum, University of Kansas). Similarly, Brachiosaurus brancai was originally illustrated (and mounted) with a giraffe-like ascending neck (Janensch 1950b, pls. 6-8), by providing the last few cervical vertebrae and the first dorsal vertebra with distinctly wedge-shaped centra (Stevens 2002, figs. 2-3). Some subsequent reconstructions in fact depict the neck as nearly vertical (Mclntosh et al. 1.997, fig.20.16; Paul 1988, 2000, appendix A), and arguments have been presented in support of this dramatic interpretation (Christian and Heinrich 1998; but cf. Czerkas and Czerkas 1991. \32: Martin et al. 1998). The neuMichael Parrish
tral pose of the neck can be reconstructed in Figure 10.3C (from Stevens in prep.) by compositing the original steel engravings from Janensch (7950a, figs. 32-49) in neutral deflection between each successive pair of vertebrae. The result is a very gentle, downward-
curving neck extending from a straight cervicodorsal transition. The centra at the base of the neck show no evidence to suggest this sauropod had a giraffe-like elevated neck. In particular, the centra of the vertebrae from C10 to D2, which were found articulated within a single block, are spool-shaped, not wedge-shaped, and their resulting neutral pose is straight, not ascending (Stevens in prep.). Again, as in some other sauropods, the anterior cervicals of Brachiosaurus cLtve downward in neutral position. This droop is likely important in orienting the head ventrally in support of downward feeding (see below).
Fig. 10.4. Original Camarasaurus reconstruction from Osborn and Mook (1921 , fig. 28) showing an
entially str aigh t c eru ico dor s al transition, t,ith no trace of the
ess
gir affe-
li
ke,
ne ck- ele u atio n
adaptation figured in more recent p opular re constructions,
Skeletal Reconstructions The acetabulum can be regarded as the fulcrum about which the
axial skeleton would pivot, according to differing estimations of glenoid and acetabular height and the placement of the pectoral girdles upon the ribcage. Amongst sauropod taxa limle variation is apparent in the basic design of the hindlimbs, and reconstrucrions for a standing posture differ only marginally. Sauropod forelimbs show more variability across reconstructions in the articulation of the digitigrade manus, the degrees of elbow flexion (Christian et al. 1999), and the orientation of insertion of the humerus into the glenoid fossa. For a given neck reconstruction, the two factors having the greatest effect on head height are the degree of arch to the back and the placement of the pectoral girdles. A digital model of ApAtosdurus, created using DinoMorphr\{, will illustrate. The neck is held constant in neutral deflection based on the composite in Figure 10.5, whereas the arch to the back and the placement of the pectoral girdles is varied. The low arch condition is consistent with that in Gilmore (7936, pl. 3a) and Vilson and Sereno (1998, foldout 1). The high arch condition and pectoral girdle orientation approximates that in Mclntosh et al. (1997, fig.20.12) and Paul (2000, appendix A). Neck Posture, Dentition, and Feeding Strategies in Jurassic Sauropod Dinosaurs
.
219
lt
Fig. 10.5. The presacral uertebral column of Apatosaurus louisae
rirotn Cl to D9) in neutral ion, re constructed by ,l i g it a I ly comp o siting indiu idual
p os it
i
I I
tr str ations
pl. 21-25).
from
G
ilmor e
(
19 3 6,
The similarly posed DuroMorphrM model will be used
:,) tllustrdte abernatiue trunk :,tierpretdtions (also see Fig.
.'.61.
Feeding enuelopes and inferred browsing heights. A "feeding envelope" is the range of head positions that can be reached by a tetrapod standing in one place and simply moving its neck relative to its body. Such an envelope might be estimated from the range of motion along the neck (Martin 1987, fig. 3; Stevens and Parrish 1,999). The assumption of this approach is that the undeflecred or neutral pose approximates the center of the feeding envelope for eacn taxon. Martin (1,987) estimated a curved feeding envelope for Cetiosaurws approximately 4.5 m wide by 3.5 m above ground level, based on a reconstruction in which the base of the neck slopes slightly downward at the shoulder. In an earlier study of ours (Stevens and Parrish 1999), the longer necks of Apatosaurws and Diplodocus were estimated to sweep through a lateral arc about 8 m wide, and surprisingly, to permit reaching dor,vnward below ground level, an adaptation perhaps related to aquatic feeding. Fig-
ure 10.7 shows the range of dorsoventral deflection for Ap-
atosdurus and Diplodocws. The dorsal flexibility of Apatosaurus, 6 m, was somewhat greater than that of Diplodocus, 4 m) attributable primarily to the larger zygapophyses at the base of the neck of Apatosaurus. Both were capable of low to moderately high browstnq
Figure 10.8 shows Brachiosdwrws brdncai (SII specimen) mod-
eled from quantitative data in Janensch (1929b, 1935-1936, 1950a, 1950b), with the neck in the neutral pose shown in Figure 10.3C. While the range of dorsoventral movements cannot be estimated due to the lack of preservation of the neutral arches, the head would reach over 9 m above ground level with a modest dorsiflexion of approximately 3" per joint, and could reach ground
.
Kent A. Stevens and J. Michael Parrish
Art
t $-lJ**{#
i,t
rl
,J$i
level (without requiring a giraffe-like splay of the forelimbs to drink) by ventriflexion of slightly less than 8' per joint proximally' It is not necessary to postulate osteological adaptations, such as wedge-shaped centra, for Brachiosaurus to have reached remarkable heights and to achieve a huge feeding envelope, even if it had negligible ability to elevate the neck above its neutral pose (for muscular or cardiovascular reasons). Sauropod Dentition The study of feeding modes in sauropods extends beyond the esti'We will now briefly review inferences mation of cervical positions. skull structure' plant distribudentition, drawn from studies of
Fig. 10.6. DinoMorPhrv re
constru ct i ons illu str ating th e
uariation in head height resulling front ttuo interpretations of tbe body plan o/ Apltosaurus louisae. The neck is held constant in the neutrdl position shown in Figure 10.5. The low dorsal arch condition corresponds to the neutral pose (Fig. 10.5). The higb dorsal arch dnd steePel Placement of tbe pectoral girdles lsee insetst correspctnds to that in Mclntosh et al. (1997, fig. 12; also see Paul
2000).
Neck Posture, Dentition, and Feeding Strategies in Jurassic Sauropod Dinosaurs
.
221
t
{
4" .,,1,r
ltliiUil
Fig. 10.7. DinoMorpbr\ models
o/Apatosaurus loursae (A) and Diplodocus carnegii (B) presaoal axial skeletons in neutral pose plus lou-contrast auerlay of the extremes of dorsal dnd uentrdl flexion (Steuens and Parrish 1999).
tions, and previous analyses of neck posirions and discuss how our models impact previously proposed models of sauropod feeding. The basal members of all dinosaur lineages for which an her-
bivorous diet is generally inferred (Sauropodomorpha and Ornithischia) share the same basic tooth-form, consisting of a leaflike shape featuring an expanded crown and a serrated row of denticles occurring along the ridge dividing the labial and lingual surfaces of
the tooth. In Sauropodomorpha, the leaf-shaped form predominates in prosauropods, particularly if one acceprs Barrerr's (1999, 2000) reassignment to the Sauropoda of the blunt teeth originally designated by Simmons (1965) as belonging to the prosauropod Yunnanosaurus.Yates and Kitching's recent (2003) analysis of Triassic sauropodomorphs has placed several taxa formerly consid222 . Kent A.
Stevens and T. Michaei Parrish
#
s
of a monophyletic Prosauropoda inside of the sauropod lineage, including Melanorosaurus (but see Galton et al', this volered part
ume) and Anchisaurus.
The taxonomic reassessment of Anchisaurus, which is the most basal sauropod in Yates and Kitching s (2003) phylogeny, is significant because, unlike the other Triassic and Lower Jurassic sauropods, Anchisaurus is well represented by dental and cranial material. The teeth of Anchisdurus are leaf-shaped, serrated, and recurved medially (Galton 7976). Dentition among other basal sauropods is represented by isolated teeth of the Early Jurassic genera Barapaslsaurus and Kotasaurus (Yadagiri 1988). Both possess
Fig. 10.8. DinoMorPhr\t model Brachiosaurus brancar with ceruical uertebral column in neutral pose (see Fig. 10.3, C).
of
coarse denticles and exhibit the expanded spoon-shaped and lingually concave crown pattern that is characteristic of most nondiplodocid sauropods. Similar teeth are present in members of the Chinese sauropod clade Euhelopodidae, although some variation occurs in the size and distributions of denticles, which are absent altogether in Euhelopu.s. No sauropod teeth are serrated, other than those of Barapasawrus, Kotlsaurus, euhelopodids, and some unworn examples of the Tendaguru Brachiosaurzs. The teeth of the Middle Jurassic sauropod Patagosaurus are similar in shape to those of Euhelopodidae (Bonaparte 1986). Camarasaurus teeth are well known, and are similar in basic shape to those in Patagosaurus and euhelopidids, although the expansion of the crowns relative to the tooth base is less pronounced, as is the concavity of the lingual surface of the teeth (Madsen et al. 1995). The teeth of Brachiosaurus share the general spatulate configuration with Vulcanodon, Euhelopodidae, and Camarasarurus (Janensch 1935-1,936). Denticles are reported in some unworn Brachiosaurus teeth (Janensch 1935-1936) but are not visible in worn teeth. In contrast to those of Camarasaurus, the crowns of Brachiosaurus teeth exhibit minimal expansion relative to the base. The teeth of Diplodocidae and Titanosauridae are both nearly circular in diameter, without any expansion of the crowns' In both families, the crowns taper gently to a point in the unworn condiNeck Posture, Dentition, and Feeding Strategies in Jurassic Sauropod Dinosaurs
.
223
tion, but form planar occlusal surfaces when worn (Holland 1924; Barrett and Upchurch 1994; Curry Rogers and Forster 2001). Dental macroweAr.'Wear facets are absent in the teeth of An_ chisaurus, Bardpasdurus and Kolasaurus, but a varietv of wear facets are apparent in other sauropods for which dentition rs known. In the euhelopodids Omeisaurus and Shunosaurus, srepshaped wear facets occur on the cranial and caudal margins of the teeth, which appear to be the result of significant tooth-tooth wear (Upchurch and Barrer, 2000), with the greatest amount of wear on the cranial facet. No such wear is apparent in the skull of Mamenchisaurus sinocanadorzru (Russell and Zheng 1993). Step-like tooth wear is also evident in Patagosawrzs (Bonaparte 19g6i. Sig_ nificant concave wear facets are also visible o.t .ith.. side of the apices of the teeth rn Camarasaurus (Madsen et al. 1995), although here the amount of wear is more symmetrical than in patagosdurus and the euhelopodids.
In Brachiosaurus, small amounts of wear are observed, but generally on the lingual and labial sides of the crown rather than on the cranial and caudal margins. As in diplodocines and titanosaurs, this has been interpreted as evidence of tooth-to-tooth occlusron, rather than interdigitation of upper and lower teeth. In titanosaurs and diplodocines, this type of occlusion produces high-angle wear facets that resemble the point of a chisel (Fiorillo 1998; Curry Rogers and Forster 2001; Upchurch and Barrett 2000). Dental microwear. In studies of the teeth of mammals, and par_ ticularly those of fossil primates, dental microwear has been studied via scanning elecrron microscopy as a method of inferring diet in extinct vertebrates. The logic behind this approach is that different food types will create differenr, and distinctive, striations on the enamel of herbivore teeth that may be indicative of diet. This approach has been applied to sauropods by Fiorillo (1991,799g), Calvo (1994) and Barrett and Upchurch (1994, IggS), although studies to date have focused only on the teeth of camarasaurus (Fiorillo 1991, 1998), Diplodoczs (Fiorillo 199I, 1999; Calvo 1994; Upchurch and Barret 2000), and the titanosaurid Rabeto_
sdurus (Upchurch and Barrett 2000). All studies of diplodocid teeth showed labiolingual scratches across the wear facets, whereas the studies of camarasaurus indicated both pits and scratches in adult teeth (calvo 1994; Fiorlllo 1998; Upchurch and Barrett 2000). Fiorillo (199I) interpreted the absence of pits in juvenile camarasawrzs teeth as evidenie for ontongenetic switching of diets. Upchurch and Barrett's (2000) study of one Rapetosauru.s tooth indicated both coarse scrarches and prtting on the wear surface.
Cranial Characters Plesiomorphicalll',
the skulls of sauropods resemble those of
prosauropods such as Plateosauru.s (Huene 1926;Galton 1984). In the most basal sauropods for which relatively complete skulls are
.
Kent A. Stevens and J. Michael Parrish
known-for example, Ancbisdurus, Shunosaurus, and Omeisdurus (the latter two of which are either basal Euhelopodidae [in Upchurch 19981 or basal Sauropoda ['Wilson 2002] in the two most recent, comprehensive phvlogenetic analyses of sauropods)-the skull is essentially convex in anterodorsal profile, with a modest snout.
lnclination of tbe skull. Because of its presence in Euhelopodidae, Camarasaurus, Brachiosauridae, and sauropod outgroups, the plesiomorphic pattern for sauropod cranial inclination appears to be one with the tooth row positioned horizontally relative to the long aris of the brain cavity extending from the foramen magnum into the braincase. In Diplodocidae, Nemegtosauridae, and the Titanosauridae for which cranial material is known, the tooth row is inclined cranioventrally relative to this axis, such that the head would naturally tilt dorvnward in a neutral position (here defined as the situation where the long axis of the brainstem cavity and the neural canal of the atlas/aris are horizontal). Fiorillo (1998) noted these differences when comparing Camarasaurzs with Diplodocus, interpreting the nearly 90' angulation of the head relative to the neck in Diplodocus as an indication that it might have had a more erect vertical neck than that of Camarasaurzs. Howet'er, the alternate interpretation, that the head was directed more ventrally in Diplodocus to facilitate low browsing' seerns equally plausible. The diplodocids Apatosaurus and Diplodocus had sufficient neck flexibility ventrally to reach far below ground level (Stevens and Parrish 1,999), a capability consistent with lacustrine feeding. Taken in conjunction with the presence of prognathous, peglike cropping or sieving dentition, dorsally placed nostrils, and a ventrally trending neutral position for tl-re cervical column, the dor,vnward curvature of the head could have sen'ed as a means of grazing on or under the surface of the water while maintaining visual and olfactory vigilance. The absence of these features in Cdmarasaurus and Bracbiosaurus would be consistent with a more generalized feeding envelope for these genera.
Configuration of dentition and role of iaws in food processing' The basal arrangement of dentition in Sauropoda is not dissimilar from the condition that is plesiomorphic for Sauropodomorpha and Dinosauria, consisting of an essentially isodont array of teeth that projected from most of the ventral surfaces of the maxillae and premaxillae without any diastema. The teeth are either mildly prognathous or essentially perpendicular to the long axis of the tooth row, and the plane defined by the bases of the teeth is parallel to that of the horizontal long axis of the braincase. The cranial end of the dentarf is convex upward in Anchisaurzs (Galton 1976), Plateosdurus (Gaiton 1984), and Diplodocidae (Hatcher 1901), so the possibility exists that the everted lower jarv is a piesiomorphic feature for Sauropoda and reversed in lineages such as the Euhelopodidae and Titanosauroidea.
In
Omeisaurus, Mdntenchisaurus, and within the Diplodo-
coidea and Titanosauroidea, the dentition is restricted to the front Neck Posture, Dentition, and Feeding Strategies in Jurassic Sauropod Dinosaurs
.
225
of the jaws and, particularly within the Diplodocinae, the teeth are strongly prognathous, such that the teeth could have served only to
collect a mouthful of vegetation by nipping, raking, or sieving, rather than to facilitate extensive oral processing of the fodder. By comparison, the condition in the remaining sauropods, with expanded crowns, interdigitation of the teeth, and a more extensive dental battery, would potentially have facilitated more masticarion.
History of Inferences of the Use of Sauropod Necks in Feeding The notion that sauropods used their necks to achieve a significant lateral sweep can be traced back to Hay (1908). The concept of sauropods as low browsers has persisted for some taxa over the years. The idea of sauropods-as-molluscivores was first proposed by Holland (1,924), who believed that the blunt, prognathous den-
tition of Diplodocus could have been utilized to crack
open
unionid bivalves. Haas (1963), in his study of diplodocid jaw musculature, inferred Diplodocus was an aquatic filter feeder, specializing on floating crustaceans and/or mollusks. Both visualized the long neck as an adaptation for sweeping the head through a broad lateral arc without moving the body'. Alexander (1985) restored the
neck of Diplodocus in an essentially horizontal orientation, contending that the cervical musculature would have been insufficient to allow the neck to be raised significantly, and suggesting that a Iarge nuchal ligament would have been instrumental in maintaining the neck in a horizontal orientation. Martin (1987) manually articulated the neck of the Leicester specimen of Cetiosaurlrs, and concluded that the neurral position for its neck was near horizontal, with a slight downward curvature (see also figure 10.3A). Martin envisioned the neck of Cetiosaurzs as primarily an adaptation fa-
cilitating alateral sweep of the head. Dodson (1990) cited the broad, vertical feeding ranges made possible by the elongate necks of sauropods, and suggested that neck length and mobility might facilitate niche partitioning of different genera. Barrett and Upchurch (1994) proposed that Diplodocus might have served as both a high browser and a low browser, stripping vegetation in the high browse by pulling stems through its tooth comb. They cite the different types of wear observed on upper and lower teeth as evidence for these two types of feeding, and suggested that propalinal scratches on the teeth might be an artifact of branch stripping during high browsing. They held that the greatest amount of mobility in the cervical vertebrar column of Diplodocus was in the most cranial vertebrae, and that this flexibility close to the head facilitated their branch stripping mechanism. In a subsequenr review of sauropod feeding mechanisms, Upchurch and Barrett (2000) suggested that vulcanodontids, most euhelopodids, brachiosaurs, cetiosaurs, Camdraslurus, and Brachiosaurus were high browsers, whereas Shunosauru.s and the Dicraeosauridae were low browsers. 226
.
Kent A. Stevens and
T.
Michael Parrish
Martin et al. (1998) proposed that sauropod necks were held essentially horizontally, and suggested that the cervical ribs served as a ventral compressive member that, along with the dorsal nuchal ligament, would have held the neck as a segmented, flexible horizontal beam. They identified Dicraeosaurus and Apatosaurws as taxa that were predominately dorsally braced, and Euhelopodidae (sensu Upchurch 1998), Brachiosaurus, and Camarasaurus as taxa that were braced more ventrally. Jurassic Plant Communities and Implications for Sauropod Feeding Globally, Jurassic floras are dominated by herbaceous piants and small trees, most significantly bennettitalean cycadeoids (tree ferns), ferns, horsetails, cycads, and ginkgoes (Behrensmeyer et al. 1992; Coe et al. 1987). The Morrison Formation of the western United States and the Tendaguru Formation of Tanzania represent the two major accumulations of Late Jurassic sauropod fossils. The Jurassic climates of both regions have been interpreted as strongly seasonal (Russell et al. 1980; Dodson et al. 1980; J. T. Parrish et al. in press). Paleoclimate modeling based on biome distributions (Rees et al. 7999) interpreted both regions as "winterwet." The more thoroughly studied of the two formations, the Morrison has most recently been interpreted as savanna-like, dominated by herbaceous vegetation and traversed by large, everflowing rivers along which the greatest concentrations of trees would have occurred (J. T. Parrish et al. in press), although some other recent studies interpret the Morrison as a whole as more humid (e.g., Tid-
well et al. 1998). Inferred sawropod diets. Because of the massive bulk of sauropods, most studies have assumed that their primary food source would be both highly nutritious and abundant (Weaver 1983; Farlow 1987). 'Weaver (1983) measured the caloric densities of extant members of plant groups that were abundant in the Late Jurassic, and reported a range of wet weights of 0.97-2.89 kcaUg for the various herbaceous and arborescent groups, with the highest values yielded for cycads and conifers, somewhat lower values for ferns and ginkgoes, and the lowest values for horsetails. On the basis of her analysis, 'Weaver concluded that endothermy in Brachiosaurus was unlikely because the relatively low caloric content of Jurassic plants and the sauropod's small mouth relative to body size would preclude sufficiently rapid intake to maintain an elevated endothermic metabolism.
Krassilov (1981) suggested a diet of ferns and horsetails for Diplodocids and cycads and shrubs for camarasaurids, hypothesizing that the retraction of the nares in diplodocids was an adaptation for breathing while obtaining forage underwater. Dodson (1990), utilizing arguments of their abundance in Jurassic landscapes, cited ferns as the most likely candidate for a predominant sauropod food source, but also noted that these giant herbivores Neck Posture, Dentition, and Feeding Strategies in Jurassic Sauropod Dinosaurs
.
227
were not likely to have been specialists in particular plant types. Fiorilio (1998) suggested that micror,vear patterns favored an inrerpretation of Diplodocas specializing on cycads, whlIe Camdrasaurus might have specialized on ginkgoes. Chin and Kirkland (1998) described what appear to be herbivorous dinosaur coprolites from the Mygatt-Moore Quarry of the Morrison Formation. Although determination of the taxonomic identity of the dinosaurs producing the coprolites is problematic, they do include significant components (ranging from 8To to 52"/") of organic matte! the identifiable components of which include woody tissue (5-147"), cuticle (0-8%), and seeds (0-6%). Taxa represented include cycadophytes, ferns, and conifers. J. T. Parrish et ai. (in press) cite the presence of significant detrital matter in these coprolites as evidence for low browsing, although both the taxonomic uncertainty of the dinosaurs involved and the possibility of taphonomic disturbance of the coprolites makes direct inference of sauropod diets from these structures highly speculative. Combining the current state of knowledge of the paleoecology
of the Morrison Formation and Tendaguru with our reconstructions of the feeding envelopes of Late Jurassic sauropods leads to
the foliowing conclusions: 1. At least in the Morrison, the greatesr abundance of trees were found along the riparian corridors, and herbaceous floras dominated elsewhere. 2. Feeding envelopes of the principal Late Jurassic sauropods overlapped broadly', with dipiodocids, euhelopodids, and dicraeosaurids clearly earmarked as low browsers with the potential for a broad lateral sweep of their necks. Camdrasaurus and Brdchiosaurus both had straight necks that appear to have pointed slightly downward in the neutral position, but the fleribility of the neck in Cdmarasaunts and the height of the base of the neck in Brachiosaunzs indicate that these taxa would have been capable of high as well as low browsing. 3. Studies of craniai morphology, gross toorh shape, and den-
tal microwear indicate that the
narrow-toothed sauropods
(diplodocids, nemegtosaurs, and at least some titanosaurs and euhelopodids) predominately fed by cropping relatively soft vegetation and/or by straining planktonic plants and animals. The broadtoothed forms (Cantarasaurus, Brachiosaurus, and potentially vulcanodontids and cetiosaurs) apparently fed on more durable plant material, including cycads and perhaps conifers. 4. The vertical feeding envelopes of Jurassic sauropods overlapped broadly, suggesting that feeding height alone rvas not a predominant mode of niche parritioning among the abundant and speciose sauropods of the Morrison and Tendaguru.
Conclusions Varying patterns of dentai morphology, cranial anatomy, cervical design, and appendicular specialization indicate that sauropods
228
.
Kent A. Stevens and I. Michael Parrish
similarly differed in their modes of feeding. Reconstructions of the neutral position of the vertebral column for six well-known Jurassic and Cretaceous sauropods (Apatosaurws, Brachiosdurus, Camarasaurus, Dicraeosaurus, Diplodocus, and Euhelopus) indicate that all of these taxa had necks that were inclined slightlv downward in
the undeflected position. Morphological evidence for the nearvertical inclination of sauropod necks favored by some contemporary restorations of sauropods (such as rrapezoidal-shaped cranial dorsals or caudal cervicals, indicators of thick intervertebral discs or other adaptations to create neck elevation) were not observed in any taxa analyzed for this study. Acknowledgments. For access
to material, the authors are
grateful to Dave Berman, Mike Brett-Surman, Don Burge, Kenneth Carpente! Mary Dawson, Jim Madsen, Mark Norell, Ken Stadtman, and David Unwin. The authors are grateful to many of their colleagues, including Paul Barrett, David Berman, Matt Bonnan, Kenneth Carpenter, Peter Dodson, John Hutchinson, Jack McIntosh, Kevin Padian, Greg Paul, Paul Sereno, Paul Upchurch, Matt Wedel, Jeff \X/ilson, and Fred Ziegler for fruitful discussions on sauropod anatomy, phylogeny, and paleoecology. Judy Parrish, A1Iister Rees, Bob Spicer, and Fred Ziegler provided helpful perspectives on Jurassic and Cretaceous floras and paleoclimates. Matt 'Wedel provided a meticulous and helpful review. Philip Platt very kindly provided access to the dimensional drawings he has painstakingly compiled of Apatosaurus louisae and assisted us on issues of the reconstruction of the appendicuiar skeleton. Thanks to John Martin for the illustrations of the cervicals of Cetiosaurus. Eric Wills assisted in the software development for the DinoMorphrM project. Matt Bonnan, Phil Senter, and David Allen assisted in dissections, research, and specimen analysis. This research was supported by NSF Grant G1A-62082, as well as by Northern Illinois University and the University of Oregon.
References Cited
Alexandeq R. M. 1985. Mechanics of posture and gait of some large dinosaurs. Zoological Journal of the Linnean Societl, 83: 1-25. Barrett, P. M. 1999. A sauropod dinosaur from the Lower Lufeng Formation (Lower Jurassic) of Lufeng Province. Journal of Vertebrate Pale-
ontology 19 785-787.
2000. Speculations on prosauropod diets. In H. D. Sues, ed., Euo-
lution of Herbiuory in Terrestrial Vertebrates, 42-78. Cambridge:
Cambridge University Press. Barrett, P. M., and P. Upchurch. 1994. Feeding mechanisms ol Diplodocus. GA/A l0: 195-204. 1995. Sauropod feeding mechanisms: Their bearing on palaeoecology. In Sun A. and !(/ang Y., eds., Sixth Syrnposium on Mesozoic Terrestrial Ecosystems, Short Papers,107-1 10. Beijing: China Ocean Press. Behrensmeyer, A. K., J. D. Damuth, !7. A. Dimichele, R. Potts, H.-D. Sues, and S. L. Wing. 1992. Terrestrial Ecosystems through Time. Chicago: University of Chicago Press. Neck Posture, Dentition, and Feeding Strategies in Jurassic Sauropod Dinosaurs
.
229
Bonaparte, J. F. 1986. Les Dinosaures (Carnosaures, Allosauroides, Sauropodes, Cetiosaurides) du Jurassic Moyen de Cerro Condor (Chubut, Argentine). Annales de Paldontologie 72: 325-386. Borsuk-Biaiynicka, M. 1,977. A new camarasaurid sauropod, Opisthocoelicaudia skarzynskii, n. sp. n., from the Upper Cretaceous of MongoIia. Paleontologica Polonica 37: 5-64. Calvo, J. O. 1994. Jaw mechanics in sauropod dinosaurs. GAIA 10 183-194. Chin, K., and J. I. Kirkland. 1998. Probable herbivore coprolites from the Upper Jurassic Mygatt-Moore Quarry, western Colorado. Modern Geology 23 249-27 5.
Christian, A., and \7. Heinrich. 1998. The neck posture of Brachiosaurus
fur Naturkunde, Berlin, Geowissenchaften l: 73-80. Christian, A., W.-D. Heinrich, and !7. Grlder. 1999. Posture and mechanics of the forelimbs Brachiosaurus brancai (Dinosauria: Sauropoda). Mlttelungen Museum filr Naturkunde Berlin, Geowissenschaftliche Reihe brancai. Mittelungen Museum
2: 63-73.
Coe, M. J., D. L. Dilcher, J. O. Farlow, D. M. Jarzen, and D. A. Russell. 1987. Dinosaurs and land plants. In E. M. Friss,'S7. G. Chaloner, and P. R. Crane, eds., The Origin of Angiosperms and Their Biological Consequences, 225-258. Cambridge: Cambridge University Press. Curry Rogers, K., and C. A. Forster.2001. The last of the dinosaur titans: A new sauropod from Madagascar. Nature 4L2:. 530-534. Czerkas, S.A., and S.J. Czerkas. 1991'. Dinosaurs: A Global Vietl,. New
York: Mallard Dodson,
P.
Press.
M. 1990. Sauropod paleoecology. In D. B. \Teishampel, P. Dod-
son, and H. Osm6lska , eds., The Dinosauria, 402-407. Berkeley: Uni-
versity of California Press. P. M., A. K. Behernsmeyer, R. T. Bakker, and J. T. Mclntosh. 1980. Taphonomy and paleoecology of the dinosaur beds of the Jurassic Morrison Formation. Paleobiology 6: 208-232. Farlow, J. O. 1987 . Speculations about the diet and digestive physiology of herbivorous dinosaurs. Paleobiology 13: 60-72. Fiorillo, A. R. 1991. Dental microwear on the teeth of Camarasaurus and Diplodocus: Implications for sauropod paleoecology. Contributions of the Paleontological Museum, Uniuersity of Oslo 364:23-24. 1998. Dental microwear patterns of sauropod dinosaurs Camardsdurus and Diplodoarsr Evidence for resource partitioning in the Late Jurassic of North America. Historical Biology 13: I-16. Galton, P.M. 1,976. Prosauropod dinosaurs (Reptilia: Saurischia) of North America. Postilla 169 1,-98. 1,984. Cranial anatomy of the prosauropod dinosaur Plateosaurus from the Knollenmergel (Middle Keuper, Upper Triassic) of Germany. 1. Two complete skulis from Trossingen/!7urtt with comments on the diet. G e o I o gi cd et P a I e onto I o gi c a 18 : 1 39 -I7 l. Gilmore, C.W. 1925. A nearly complete articulated skeleton o{ Camarasdurus, a saurischian dinosaur from the Dinosaur National Monument. Memoirs of the Carnegie Museum I0: 347-384. 1,936. The osteology of Apatosaurzs with special reference to specimens in the Carnegie Museum. Memoirs of the Carnegie Museum 1.1.:175-300.
Dodson,
Haas, G. 1963.
230 . Kent A. Stevens and
T.
Michael Parrish
A proposed reconstruction of the jaw musculature of
Diplodocus. Annals of the Carnegie Museum of Natural History 36: 139-157. Hatcher, O. P. 1901. Diplodocus (Marsh): Its osteology, taxonomy, and probable habits, with a reconsrruction of the skeleton. Memoirs of the Carnegie Museum 1: 1,-63.
Ha5 O. P. 1908. On the habits and pose of the sauropodous dinosaurs, pecially ol Diplodocus. American Naturalist 42: 6672-681. Holland,
X7.
es-
J. 1906. The osteology of Diplodocus Marsh. Memoirs of the
Carnegie Mwseum 2: 225-27 8.
1,924.The skull of Diplodocus. Memoirs of the Carnegie Museum 15:119-138. Huene, F. von. L926. Vollstandige Osteologie eines Plateosauriden aus der schwdbischen Trras. Geologische und P alaeontologisch e, Abh andlungen, Neue Folge 15: 139-179. Janensch, V. 1929a. Die \Tirbelsaule der Gattung Dicraeosaurus hausemanni. Palaeontographica (supp.7) 3: 39-133. 1929b. Magensteine bei Sauropoden der Tendaguru-Schichten. Palaeontographica (supp. 7) 2: 13 5-144.
1935-1936. Die Schadel der Sauropoden Brachiosaurus, Barosaurus und Dicraeosaurus aus den Tendaguruschichten DeutschOstafrikas. P al a e ont o gr ap h i c a (stpp. 7 ) 2: 1 47 -29 8. 1950a. Die Virbelsaule von Brachiosaurus brancai. Palaeontographica (supp. 7) 3:27-93. 1950b. Die Skelettrekonstrucktion von Brachiosaurus brancai. Palaeontographica (supp. 7) 3: 97-103. Krassilov, V. A. 1981. Changes of mesozoic vegetation and the extinction of the dinosaurs. Paleogeogrdphy, Paleoclimatology, Palaeoecology 34:207-224. Madsen, J.H., Jr., J.S. Mclntosh, and D.S. Berman. 1995. Skull and atlas/axis complex of the Upper Jurassic sauropod Camarasaurus Cope (Reptilia: Saurischia). Carnegie Museum of Natural History Bulletin 31: 1-1 15. Martin, J. 1987. Mobility and feeding of Cetictsdurus: !7hy the long neck? Occasional Paper of the Tyrell Museum of Paleontology 3: 1
50-1 55.
Martin, J., V. Martin-Rolland, and E. Frey. 1998. Not cranes or masts, but beams: The biomechanics of sauropod necks. Oryctos 1,: 1,1,3-720. Mclntosh, J.S., M.K. Brett-Surman, and James O. Farlow. 1997. Sauropods. In J. O. Farlow and M. K. Brett-Surman, eds., Tbe Complete Dinosaur, 264-290. Bloomington and Indianapolis: Indiana University Press.
Osborn, H. F., and C. C. Mook. 192L. Camardsdurus, Amphicoelias, and other sauropods of Cope. Memoirs of the American Museum of Natural History 3: 244-287. Parrish, J. M., and K. A. Stevens. 1998. Undoing the death pose: Using computer imaging to restore the death pose of articulated dinosaur skeletons. Journal of Vertebrate Paleontology 18: 69A.. Parrish, J.T., F. Peterson, and C. Turner. 2004. Jurassic "savannah": Plant taphonomy and climate of the Morrison Formation (Jurassic,'Western USA). Sedimentary G eology 67 : 163-17 6. Paul, G. S. 1988. The brachiosaur giants of the Morrison and Tendaguru with a description of a new subgen:us, Giraffatitan, and a comparison of the world's largest dinosa:urs. Hunteria 2(31: 1-14.
Neck Posture, Dentition, and Feeding Strategies in Jurassic Sauropod Dinosaurs
.
231
Paul, G. S., ed. 2000. The Scientific American Book of Dlnosaars. New York: Br1,on Press and Scientific America. Rees, P. M., A. N{. Ziegler, and P. J. Valdes. 1999. Jurassic phytogeography and climates: New data and model comparisons. In B. T. Huber, N. Macleod, and S. L. \fing, eds., \X/arm Climates in Edrth History, 297-318. Cambridge: Cambridge University Press. Russell, D. A, P. Beland, and J. S. Mclntosh. 1980. Paleoecology of the dinosaurs of Tendaguru (Tanzania). Memoirs of the Geological Society of France 139 169-17 5. Russell, D. A, and Z. Zheng. 1993. A large mamenchisaurid from the Junggar Basin, Xinjiang. Canadian .lournal of Earth Sciences 30: 2082-209 s. Simmons, D. J. 1965. The non-therapsid reptiles of the Lufeng Basin, Yunnan, China. Fieldiana, Geologl, 15: 1-93. Stevens, K. A. 2002. DinoMorph: Parametric modeling tures. Senckenbergiana lethaed 82(1): 23-34.
of skeletal struc-
K.A. and J.M. Parrish. 1996. Articulating three-dimensional computer models of sauropod cervical vertebrae. Journal of Verte-
Stevens,
brate Paleontology 16(3): 67A. 1999. Neck posture and feeding habits of trvo Jurassic sauropod dinosaurs. Science 284 798-80Q. In press. Digital reconstructions of sauropod dinosaurs, and implications for feeding. In D. Chure, K. Curry Rogers, and J. Wilson, eds., Sauroprtd Euolution and Palectbiology. Berkeley: University of Cali-
fornia Press. Tidweli, \(/. D., B. B. Britt, and S. R. Ash. 1998. Preliminarv floral analysis of the Mygatt-Ntoore Quarry in the Upper Jurassic Morrison Formation, west-central Colorado. Modern Geology 22: 341-378. Upchurch, P. 1998. The phylogenetic relationships of sauropod dinosaurs. Zoological lournal of the Linnean Society 124: 43-1,03. Upchurch, P., and P. M. Barrett. 2000. The evolution of sauropod feeding mechanisms. In H. D. Sues, ed., Euolution of Herbiuory in Terrestrial 7 9-f22. Cambridge: Cambridge University Presse. C. 1983. The improbable endotherm: The energetics of
Verte brat es,
'Weaver,
J.
the
sauropod dinosaur Brachiosaurus. Paleobiology 9: 173-182. Wilson, J. A., and P. C. Sereno. 1998. Early evolution and higher-level phylogeny of sauropod dinosaurs. Society of Vertebrate Paleontology Memoir 5: 1-68. \ffilson, A. 2002. Sauropod dinosaur phylogeny: Critique and cladistic J. analysis. Zoological Jcturnal of the Linnean Society 136:21.7-276. Wiman, C. 1929. Die Kriede-dinosaurier aus Shanrung. Palaeontr:logica Slrica (Series C) 6: 1-67. Yadagiri, P. 1988. A new sauropod, Kotasaurus yamanpalliensis from the Lower Jurassic Kota Formation of India. Records of tbe Geologicdl Suruey oflndia 11:3-8. Yates, A. M., and J. !7. Kitching. 2003. The earliest known sauropod dinosaur and the first sreps towards sauropod locomotion. Proceedings
of the Royal
Sr,tcietl,
1525: 1753-17-58.
232 . Kent A.
Stevens
and
T.
Michael Parrish
of London,
Series
B, Biological Sciences 270, no.
1"
1. Neck Posture of Sauropods
Determined Using Radiological Imaging to Reveal ThreeDimensional Structure of Cervical Vertebrae Devro S. BennAN AND Bnucp M. RotrscHrlD
Abstract Two types of sauropod cervical cenrra are distinguished based not only on external features, but also on internal features revealed through a modified method of computerized tomographic X-rays, with three-dimensional reconstructions. Using external morphology the two types of cervical cenrra are: (1) a robust type that is short relative to its width, does not become strongly waisted or narrower at its midlength, and lacks prominent ridge-like buttresses; and (2) a gracile type that is long relative to its widrh, becomes strongly waisted or narrower lateraliy and ventrally toward its midlength, and possesses prominent, ridge-like buttresses. Of the seven different sauropods whose cervical centra were studied, those of Camarasaurus and an unidentified titanosaurid exhibit the robust-type morphology, whereas those of Diplodocus, Ap at
o s auru
s, H ap I o c anth
o
s
auru s,
B dr o
s
auru
s, and
B r a cb io
s
duru s
exhibit the gracile-type morphology. Three-dimensional, radiological images of the cervical centra of Diplodocus and Camarasaurus reveal that the robust and gracile types of centra each possess a distinct structural design based on the distribution parterns and relaL-) -)
tive abundances of the compact and cancellous bone. Functional stress analyses of the structural designs of the gracile and robust centra types are proposed based on an analogy of the neck as a cantilevered beam. On this basis we conclude that the robust-type cen-
trum supported a neck held in a vertical, or near-vertical, pose, whereas the gracile-type centrum supported a neck held in a horizontal, or near-horizontal, pose.
Introduction Essential to a full understanding of the functional design of skeletal
elements is a detailed knowledge of the three-dimensional structure, which includes the distributions, patterns, and relative abundances of the different types of bone. \7ith the advent of newer and more refined radiological instruments and the accessibility of those instruments to a wider range of researchers, such data is now more easily realized. This sort of approach was important in providing a possible functional and behavioral explanation for the fusion of caudal vertebrae in certain Jurassic sauropod dinosaurs, which was the result of ossification of ligaments laterally spanning consecutive
centra (Rothschild and Berman 1991). The goal of the present paper is to explore the potential of three-dimensional, radiological imaging in identifying major stress patterns of skeletal bone as a function of the distribution patterns and relative abundances of the two basic types of bone, compact (or laminar) and cancellous (or trabecular).
For any major, supportive skeletal element in which strong gravitational or mechanical forces must be overcome) structural design can usually be explained in terms of maximum strength attained with a minimum cost in weight (Alexander et aL. 1979). 'With this simple principle as a guide to the structural design of a bone, the much denser, heavier, and stronger compact bone should be present in relatively greater amounts only where resistance to
concentrated mechanical forces is most critical. The effect of mechanical stress on the distribution patterns and relative abundances of compact and cancellous bone in a limb bone, such as a femur, has become the standard textbook example, because of the ease by which the phenomenon can be demonstrated using histologic sec-
tions and routine X-rays (Rockoff et al. 1969; Mow and Hayes 1991; Einhornl9961' Carter and Hayes 1997). Undoubtedly, other skeletal elements have been excluded (e.g., cervical vertebrae) from this sort of investigation because of their complex structure. A major challenge has been the mathematical complerities of trying to describe vertebral structure (Hannson et al. 1.987;. Panjabi et al. 1991). A new technique of determining the major distribution patterns and relative abundances of compact and cancellous bone in an element is introduced here, which employs an advanced technique in radiological imaging that produces three-dimensional images. This was accomplished using computerized axial tomography
(CT), which permits the application of post-scanning density win-
234
.
David S. Berman and Bruce M. Rothschild
dows (ranges of sensitivity). An important advantage of this technique is that it eliminates confusion with pleurocoels, which have been well described (Longman 1933; Janensch 1.947,1950; Wedel 2003). Radiological imaging, therefore, provides an indirect means of studying the mechanical stress forces in elements having com-
plex morphologies without physically altering (e.g., sectioning) tnem.
In order to explore the potential of radiological imaging in identifying the major patterns of structural stresses in skeletal elements, the centra of cervical vertebrae of sauropod dinosaurs were selected for two important reasons. First, it is presumed that the long, heavy necks of the sauropods undoubtedly sublected the cervical vertebrae to considerabie gravitational force. If it is also accepted that the cervical centra were subjected primarily to a single, one-dimensional force, the downward bending of the neck due to gravitation, this should manifest itself in an obvious emphasis of compact bone distributed in a precise pattern. Admittedly, the nuchal ligament undoubtedly played an important role in supporting the neck, but only to lessen the magnitude without altering the primary direction of the gravitational force (Francots 7975; Wedel and Sanders 1999). and therefore need not be considered further as affecting the stress analysis presented here. In addition, muscles attaching to the cervical ribs are discounted as contributing to neck support. Ratheq as suggested by }Tedel and Sanders (1999), the principal muscle attaching to the ribs of sauropods was probably homologous to the M. longus colli ventralis in birds and therefore provided for lateral movements of the neck. The second reason for using sauropod cervical vertebrae in this study was the hope of offering an additional line of evidence in resolving the long, historical, and continuing speculation or debate regarding sauropod neck posture (Coombs 19751'Martin1987; Paul 1988, 1998; Chatterjee and Zheng 7997; Frey and Martin 1997; Chrrstian and Heinrich 1998; Martin et al. 1998; Wedel2003). Materials and Methods The characterization of mid-cervical sauropod vertebrae using gross, external features (Table 11.1) was based on examination of the following list of specimens, with the total number of vertebrae examined included parenthetically Apatosaarrzs, AMNH 460, CM 3390, FMNH7163, (78); Barosaurus, CM 1198 and 11984, (5); Brachiosauras, FMNH p25107, (3). Although the cervical verrebrae of FMNH p25707, the holotype of B. altithorax, are entirely plaster, they were considered suitable for this study with regard to external morphology, because they were accurately modeled after
the holotype of
Bracbiosaurus brancai at the Museum fiir Naturkunde der Humboldt Universitdt, Germany; Camarasaurus, AMNH 5761,BYU 9047 and 5604, CM 17069 and 11338, KU 729776, USNM 15492,I7DC BS1, YPM 1905 and 1910, (25); Diplodocus, AMNH 223 and 608, CM 94, DMNH 1494, USMN Neck Posture of Sauropods Determined Using Radiological Imaging
. l,lJ
Table 11.1 Correlation between Gracile and Robust Proportions with Presence or Absence of Buttressing (Transverse Central Buttress) in Sauropod Cervical Vertebrae
vertebrae Centrurrr morphotr-pe eramined #
Taxon Apatosaurus
Buttressing
18
gracile
presenr
5
gracile
present
Brachiosaurus
3
gracile
presenr
Camarasaurus
25
robust
absent
l.)
gracile
present
Barosaurus
Diplodocus Haplocanth osaurus
A
gracile
present
titanosaurid
3
robust
absent
'$7DC
BB1, (13); Haplocdnthosaurus, CM 572, CMNH 10380, USNM 405612, (4); unidentified titanosaurid, SB 655 and 10865,
(uncatalogued), (3). Cervical vertebrae 7 andS of CamarasaurusBYLJ 6504,8 and9 of Diplodocas CM 94, aod 7 of Haplocdnthosdurus CMNH 879 and of titanosaurid SB were subjected to computerized axial tomographic (CT) X-rays, using S-mm-thick slices and employing human thorax protocol to produce not only cross-sectionai scans, but also three-dimensional reconstructions (using General Electric, Sytec-i 3000 at Southwoods X-ray, Youngstown, Ohio). Contrast and density (windows) of the resulting images were adjusted so as to record only the major deposits of the denser compact bone, thus excluding the cancellous bone. However, in the absence of significant amounts of compact bone, sensitivity was increased to reveal the distribution pattern and relative abundance of the cancellous bone.
Institwtional abbreuiations. AMNH-American Museum of Natural History, New York, New York; BYU-Brigham Young University, Provo, Utah; CM-Carnegie Museum of Natural History, Pittsburgh, Pennsylvania; CMNH-Cleveland Museum of Natural History, Cleveland, Ohio; DMNH-Denver Museum of Science & Nature, Denver, Colorado; FMNH-Field Museum of Natural History, Chicago, Illinois; KU-University of Kansas Museum of Natural History, Lawrence, Kansas; SB-State University of New York at Stony Brook, New York; USNM-National Museum of Natural Historn'Washington, D.C.; WDC-Wyoming Dinosaur Center, Thermopolis, Wyoming; and YPM-Peabody Museum, Yaie University, New Haven, Connecticut.
Description of Sauropod Cervicals
A twofold approach was taken to reveal the basic, structural designs of the cervical centra in sauropod dinosaurs that may be in236
.
David S. Berman and Bruce M. Rothschild
dicative of neck posture: (1) mid-cervical centra were categorized on the basis of their gross external morphology, particularly overall shape and proportions, and the presence or absence of prominent buttresses or laminae. It should be pointed out that the term "buttress" is used here in a biomechanical sense to describe what'Wilson (1999) referred to as laminae, which has precedence in usage (Phillips 7871,, 255); and (2) examples of centra exhibiting contrasting external morphologies were subjected to CT X-rays with three-dimensional reconstruction capability to determine whether they also exhibit contrasting internal morphologies in the distribution patterns and relative abundances of compact and cancellous bone. On the basis of the three-dimensional structure of a cervical centrum, a functional stress analysis is proposed to suggest whether the neck was heid in a nearly horizontal or vertical position. Thus, it is hoped that the external morphology of a sauropod cervicai centrum could be used as indirect evidence of neck posture' Using gross external morphology, two morphotypes of sauropod cervical centra, robust and gracile, are recognized as exhibiting three contrasting sets of features (Figs. 11.1, 11.2). The robust-type centra are characterized as (1) being relatively short, with a posterior transverse width that ranges from about 60'/" to 90'/, of the
centrum length, (2) lacking prominent midlength waisting or narrowing, and (3) lacking well-developed, ridgelike buttresses. In contrast, the gracile-type centra are characterized as (1) being relatively long, with a posterior transverse width that ranges from about 25"/" to 42o/" of the centrum iength, (2) exhibiting prominent
midlength waisting or narrowing iaterally and ventrally, and (3) possessing well-developed, ridgelike buttresses. A survey of numerous specimens of seven different sauropods (Table 11.1) reveals 100o% association of the features of robust-type cervical centra in Camarasaurus and an unidentified titanosaurid from Madagascar, on the one hand, and those of gracile-type centra in Diplodocus, Ap at o s auru s, H ap I o canth o s aur u s, B ar o s duru s, and B r a ch i o s lur us, on the other. In Camarasaurzzs the eighth cervical centrum is easily distinguished from the other cervical centra in being not only noticeably ionger in absolute size, but also relative to its transverse width (Fig. 11.28) (Osborn and Mook 1921). However, the width-
to-length percentage of the eighth centrum falls well short of the range of values recorded for the gracile-type centrum. In order to determine whether robust and gracile-type cervical centra can be characterized by significant differences in internal structure, CT images of vertebral centra were subjected to threedimensional reconstructions rvith the density and contrast adlusted so as to record only the distribution pattern and relative abundances of the more opaque compact bone. The CT image of the robust centra \e.g., Camarasaurus) revealed no significant structural features of compact bone. Readjustment of the image, however, to a wider density window revealed a diffuse distribution of the less opaque cancellous bone throughout the centrum interior (Fig. 1,1,.2A, B). This is in marked contrast with images of a Diplodocus Neck Posture of Sauropods Determined Using Radiological Imaging
'
237
Fig. 11.1. Lateral uietus of (A) mid-ceruical ueftebra
of
Diplodocus carnegri CM 91 (anterior to right and with rib missing) and (B) ceruicals 7 and 8 of Camarasaurus lentus CM 11059 (anterior to left dnd rib of ant eri ormo st u er t e br a mi s sing). Note absence of transuerse central bwttress (TCB) in Camarasaurus centrd.
centrum (Fig. 11.2C), which exhibited a pattern of four distinct areas of concentrated compact bone. Two of these areas consist of bands extending the length of the centrum along its dorsolateral and midventral surfaces. The dorsolateral bands are much broader anterior to the base of the neural arch, whereas the midventral band erhibits a pronounced broadening posteriorly to include the entire ventral surface of the centrum. The second area of compact bone concentration forms a prominent, ridgelike buttress that extends posteroventrally across the anterior half of the lateral surface of the centrum and is referred to here as the transverse central but'lTilson tress (see also fig. 39, and Sereno 1998). Lastly, circumferential bands of compact bone are present at the ends of the centrum. Collectively, the bands and buttresses of compact bone of the
238
.
David S. Berman and Bruce M. Rothschild
Fig. 11.2. Computerized axial omo gr ap h ic cr o s s - s e ctio nal s cans of mid-ceruical uertebrae of sauropods: 1At transuerse section through centrum (dorsal toward top of p,tgat: rBt threedimensional r e construction of centrum in lateral uietu (anteilor to left) of Camarasaurus BYU 6504 tuith sensitiuity adjusted to reueal cancellous bone (white areas) ; (C) three-dimensional reconstructions of mainly centrum in oblique lateral uiew (anterior to Ieft) of ceruical uertebrae of Diplodocus carnegii CM 94 ruith sensitit/ity adiusted to reueal comPdct bone. Abbreuiation: TCB = transuerse central buttress, t
Neck Posture of Sauropods Determined Using Radiological Imaging . 2J9
centrum are arranged so as to define partially a hollow, hourglassshaped structure.
Analysis and Discussion The hypothesis being tested here contends that rhe magnitude of the mechanical stresses and their orientation to the cervical centrum should be observable in the distribution parterns and relative abundances of the compact and cancellous bone. The structure of the centrum is assumed to minimize the mass of bone utilized without sacrificing its ability to withstand the srresses encountered under normal movements of the neck. Paramount to this analysis is an understanding of the exrremely important differences in the structural properties between compact and cancellous bone, which are related primarily to porosity. Compact bone is essentially solid, with very low porosity limited to Haversian canals, canaliculi, capillaries, and erosion sites, whereas cancellous bone can be characterized as exhibiting large spaces or cavities that are interconnected by fenestra. Differences in porosity are related apparently to strikingly different physical properties between compacr and cancellous bone, with the former being denser, stiffer, heavier, and, most importantly, stronger. Both types of bone, however, are typicalry found directly associated, but difference in porosity allows them to be very easily discernible to the naked eye, and the transition from one to the other occurs usuaily over a very short distance. Governed by the widely accepted axiom that dictates that the mass of any element is maintained at the minimum required for it to perform properly within the range of its normal structural demands, extreme differences in structural demands should be observable in noticeable differences in the distribution patterns and relative abundances of the compact and cancellous bone (Mow and Hayes
1997; Einhorn 1996). That is, the question being posed here rs whether the cervical centra of sauropods exhibiting distinctively different structural designs that use greatly differing proportions of compact and cancellous bone can be correlated with distinctly different neck postures of horizontal versus vertical. Assuming that the two centra morphotypes, gracile and robust, reflect contrasting neck postures, then an analysis of the probable, primary, mechanical srress forces rhat acted on them should provide an explanation for their structural differences. That is, if rhe ielarive magnitudes of the stress forces a sauropod centrum was subjected to are directly related to neck orientation, then the centrum structure would be expected to provide a basis for determining whether neck
posture was predominately horizontal or vertical. Cervical vertebrae in forms such as birds, deer, or giraffes fail as modern anaIogues in such an analysis, because their necks are not as massive as those of sauropods and would produce a scaling artifact that would likely negate their value as a model (McMahon 1975; Alexander et aL.1979; Alexander et al. 1985; Alexander 1989; Hokkanen 1986; Biewener 1989; Bertram and Biewener f990).
240 . David
S. Berman and Bruce
M. Rothschild
tiirection of-fS
direction of TS ctuvertecl t0 CS alnus TCB
dilection of CS i I
neck anchored at shoulder girdle
I
Iig. 1 1.3. Diagrantmatic draning of the neck dnd head of a longneck sauropctd held irt a ltorizontal posc to illuslrate its analogy to d cantileuered beam supported at the shoulder girdle dnd *bjected tr-t a grauitational loadtng lorct cnncen!ratcd at its free or distal end. The centrd dre considered the principal elements of the beam (neural arches and ribs omitted), tuhich are of the gr d c il e m orp
glar itati onal loacling tblce
A rather simple analysis of the forces acting on the neck, or its cervical centra, is performed here that assumes a few generalizations: (1) the neck is compared to a cantilevered beam that is anchored or supported at the shoulder girdle; (2) the principal loading force on the neck is gravitational and acts along the entire length of the column, but the loading effect is regarded as if concentrated at the distal (cephalic) end of the neck; (3) the centra are considered the principal elements forming the cantilevered beam, and the influence of the neural arches in counteracting the stress forces is not considered significant enough to otherwise alter the basic structural requirements of the centra (Francois 1975; Stevens and Parrish 1999): and (4) each vertebral unit, or centrum, can be
h
otyp e
(r e latiu
e
I.t
Iong compttred to didmeteL narrow tou'ard midlength, and possesslng a lateral tr.tnsuerse central buttress ITCBI). Ottlines of the cenlrn corrcspond l, n1o1u, deposits of comPdct bone and where tensile or tension and compressictn stresses due tct grauitatktnal lodding are concentrdted. Tensile stress (TS) is directed Llnteriorly along dorsal surfaces rtf centra and comltression st/ess (CS) is directed posleriorlt' along uenlral surfacr's. Portion,'I the tcnsile s!rt'ss is redirected b1- the transuerse central buttress (TCB) as contpression stress.
analyzed as if subjected to the same stress forces as the whole of the
neck, That is, each centrum can be viewed as a separate' cantilevered beam anchored at its posterior end to the preceding centrum. Following these generalizations, two types of structural design of the centra in sauropods can be hypothesized that would
meet the stress demands anticipated if the necks were held either at or near a horizontal or verticai posture. In the standard analysis of a cantilever beam that is held horizontally and loaded at its free end, two equal and opposite principal stress forces are described as passing through the length of the
beam (Fig. 11.3). Tensile or tension forces are transmitted anteriorly along the dorsal surface of the beam and compression forces are transmitted posteriorly along the ventral surface of the beam. The stress forces are typically visualized as lines or trajectories that gradually become more closely spaced toward the dorsal and ventral margins of the beam to indicate areas of greatest stress. Therefore, the greater the magnitude of the loading, the greater is the
number and peripheral concentration of the stress lines. On the other hand, increased spacing of the tensile and compression stress Iines toward the center of the beam indicates a drop in the magnitude of the stresses to a theoretical zero in a neutral zone' If this theoretical pattern of stress lines is used to predict the distribution patterns and relative abundances of compact and cancellous bone Neck Posture of Sauropods Determined Using Radiological Imaging
'
241
in a cervical centrum loaded as a cantilevered beam, then the more tightly placed lines of stress at the dorsal and ventral margins of the bone, where the greatest tension and compression stresses are concentrated, is where compact bone would be expected to be present
in relatively substantial
amounrs. Internalln where the lines of
stress become fewer and more wideiy spaced, cancellous bone, bone marrow, or simply voids would be expected. However, cancellous bone is always intimately associated with the compact bone and therefore with loaded surfaces where stresses are reasonably constant. In this situation the cancellous bone generally lies internal to the compact bone and has the appearance of very porous, compact bone with interconnecting holes between an array of little struts or beams referred to as trabeculae. The trabeculae of cancellous bone functions structurally by their orientation to direct stresses outward to the compact bone (Hansson et al. 1987; Mow and Hayes 1991; Einhorn 1996). The above theoretical distribution pattern of compact bone
of the Diplodocus midof Figure 77.2C. Most importantly, it is what would be expected if the neck were held at or near a horizontal matches that exhibited by the CT X-rays
cervical centrum
pose and had to resisr strong gravitational loading (Martin et al. 1998). The pronounced deposits of compact bone along the dorsolateral and midventral surfaces of the centrum would have ct-runtered strong tensile and compression stresses, respectively. The compact bone of the prominent, transverse centfal buttress can also be explained as having played an important role in helping to support the neck. Because the buttress is oriented posteroventrally and unites the dorsolateral and midventral bands of compact bone, a significant portion of the tensile stress directed anteriorly along the dorsal surface of the neck would have been redirected posteroventrally as compression stress to the ventral surface of the column. The advantage of this strategy becomes obvious when comparing the ability of bone ro resisr compression and tensile forces. Bone, as
is the case in most materials, can withstand a much greater compression force than tensile or rension force. This is especially true of bone tissue, in which the compression strength (the maximum compression force it can sustain before breaking) exceeds tensile strength by approximately twice (Pugh et aL. 1.975; Currey 1984, 2002). This functional explanation of the rransverse cenrral buttress also accounts for the more extensive development of compacr bone on the dorsolateral surfaces of the centrum anterior to the base of the neural arch and on the ventral surface of the oosterior half of the centrum. Both areas lie on the trajectory pnih of th. redirected compression stresses passing through the transverse central buttress.
Significant deposits of compact bone enclosing the margins of the anterior and posterior rims of the centra are also considered as areas indicative of major structural stresses. In contrast to gravitational loading, however, these stresses are probably due mainly to compression or tensile forces exerted by the pull of stretched liga-
212
.
Davtd S. Berman and Bruce M. Rothschild
ments spanning vertebrae and contracted muscles during lateral flexion of the neck. It could be argued that the significance of the robust or gracile character of cervical vertebrae relates simply to airlrespiratory sac cavities (pleurocoels) (Wedel 2003), but an alternative explanation can be offered. The development of gracile vertebrae could repre-
sent an anatomical modification
to
lessen the osseous,
or
bone
weight, of the neck. However, given the relatively small mass of the neck, as compared to that of the entire bodS its reduction would have minimal effect if the animal's neck were held at or near a vertical pose. Neck mass would have a greater impact, however, if the neck were held horizontally. A longer neck places the moment arm farther from the body (Fig. 11.3), which magnifies the effect of the neck weight. Increasing the graciie structure of cervical vertebrae significantly reduces the torque of the moment arm and, therefore, the amount of cervical muscle activity necessary for maintaining a
horizontal neck posture.
If the neck of a sauropod were held in a vertical or, more realisticalln in a near-vertical pose, gravitational loading of the neck would be predominately axial, with the overwhelmingly dominant stress being compression. Under these conditions compression forces would be most effectively countered if distributed uniformly over the entire cross-sectional area of the centrum and if acted on the lighter cancellous bone as the structural material of the centrum (Hansson 7987). Tensile stress due to bending would be essentially nonexistent or negligible, and the necessity for compact bone would be minimal and mainly to resist the pull of ligaments and muscles. This is the pattern revealed by CT X-rays of the Cama-
rasaurus mid-cervical centrum (Fig. 11.2A, B) when adjusted to a sensitivity suf6ciently high enough to record cancellous bone. The differences in the distribution patterns and relative abundances of the compact and cancellous bone that distinguish the robust- and gracile-type centra also provide an explanation for the
prominent and diminished midlength waisting or narrowing of their centra. According to Frost (1'964), when a centrum is loaded in axial compression there is a potential for outward bulging deformation of the outer walls due to pressure exerted by the fluid-like marrow (including fat, blood-forming tissue, and blood) contained in internal spaces, a phenomenon he referred to as the internal hydraulic effect. This can be observed indirectly in vertebral compression fractures prior to the reparative remodeling that usually restores the compact bone margins (Recker 1993; Keller et al. 2003). degree of influence of the internal hyexpected between robust- and gracile-type would be draulic effect is the robust type found in Camarasaurus of centra. If the centrum core of cancellous bone larger cross-sectional and has a relatively hydraulic effect internal the spaces, with small, marrow-filled
A strong difference in the
would be minimal. In this instance, according to Frost (1964), the cancellous bone provides the necessary resistance to the compression loading of the centrum, and outward bulging would not be a Neck Posture of Sauropods Determined Using Radiological Imaging
'
243
problem. On the other hand, if the centrum is of the gracile type found in Diplodocus, rvith a relatively narrow width and a muchreduced core of cancelious bone, the internal hydraulic effect due to compressional loading, despite the probable presence of relatively small, marrow-filied spaces, may have been potentially great enough to cause an outward bulging of its outer walls. If, howevcr, the outer walls of the centrum are waisted, any internal, outwardly directed pressure could be more effectively countered, as it must first act to reverse the curvature of the outer wails. This would result in the internal pressure of the centrum being redirected by the compact bone of the outer walls of the centrum anteriorly and pos-
teriorly as compression force (Frost 1964). Vertebral pneumaticiry (pleurocoels) would not affect the response to the internal h1'draulic pressurer as they are self-contained systems (\X/edel 2003). The above discussion, therefore, suggests that when cervical centra are loaded primarily in compression, as in the case of vertical neck posture in sauropods, the greater their internal content of cancellous or spongy bone and the lesser need for waisting, whereas the reverse would be expected in centra of necks held in a horizontal pose.
Based on the above observations, the contrasting features of the distribution patterns and relative abundances of the compact and cancellous bone in the cervical centra of Diplodocus and Camtrasdurus, as revealed by CT X-rays, suggest correlations between gracile and robust-type centra with horizontal and vertical neck postures, respectively. This also provides corroborarive evidence for the widely accepted interpretations of horizontal neck
posture tn Apatosaurus, Barosaurus, and Haplocanthosaurus, and vertical neck posture in Camardsdurus and titanosaurids (Frey and
Martin 1997; Marrin et al. 1998; Stevens and Parrish 1999). An unexpected implication of this study, however, is that the gracile nature of Brdchiosaurus cervical vertebrae suggesrs a horizontal, rather than the widely portrayed, near-vertical neck posture (Christian and Heinrich 1998). The recently described (Zimmer 7997) initial results of investigations of neck posture in sauropod dinosaurs by K. A. Stevens and M. Parrish (1999; this volume) make the same conclusions as those given above, but based on different morphological data and analyticai techniques. Measurements on the range of movements at the articulation contacts between successive cervical vertebrae, particularly between pre- and postzygapophyseal facets, were processed in a three-dimensional, graph-
ics computer program to determine not only the range of
movements of the necks in several sauropods, but also their neutral, undeflected postures. The results of their study indicate that Apatosaurus and Diplodocus, on the one hand, and Brachiosaurus, on the other, held their necks at about the horizontal and 20o above (suggesting ground feeding or low browsing), respectiveiy, whereas Camarasdurzs held its neck nearly vertically (suggesting high browsing). These results were reaffirmed subsequently by Steven and Parrish (19991 for Apatosaurus and Diplodocus in a more de-
.
Drr-id
S.
Berman and Bruce M. Rothschild
tailed account of the same research. The phylogenetic significance of this variation is, however, beyond the scope of the present discussion, but has been addressed in studies by Salgado et aI. (7997), \Tilson and Sereno (1998), Upchurch (1.998), and Wilson (2002). Inasmuch as the resistance of a cantilevered beam to bending varies inversely as the square of its length, there is a particularly interesting relationship between relative neck length and probable
neck posture among the sauropods considered here. Those sauropods with proportionally the longest necks relative to trunk srze, Apatosaurus, Barosaurus, Brachiosaurus, and Diplodocus, with neck lengths of approximately 6 m, 9 m, 8.5 m, and 8 m, respectively (Gilmore 1936; Zimmer 1.997), apparently held their necks in a near-horizontal pose, whereas the one prominent, relatively short-necked example, Camarasaurzs, is assumed to have held its 3-4-m-long neck in a near-vertical pose (Zimmer 1977). This comparison further explains the strong differences noted in the structural designs betlveen the cervical centra in forms believed to have held their necks at or near horizontal and vertical poses. Acknowledgments. \7e are grateful to Kenneth Carpenter, Larry D. Martin, Mark Norrell, Burkhard Pohl, Robert Purdy, Bill Simpson, Mary Ann Turner, Ken Stadtman, Scott Sampson, and the 'Sfilliams late Mike who qraciouslv allowed us access to their collections.
References Cited
Alexander, R. M. 1989. Mechanics of fossil vertebrates. Journal of the Geological Society 146: 47-52. Alexander, R. M., A. S. Jayes, G. M. Maloiy, and E. M. Vathuta. 1979. Allometry of the limb bones from shrews (Sorex) to elephants (Loxodonta). Jcturnal of Zoology 189: 305-314. Alexander, R.M., J.F. A. Hall-Martin, and D. A. Russell. 1985. Long bone circumference and weight in mammals, birds and dinosaurs.
Journal of Zoology A207: 53-61.
Bertram, J. E., and A. A. Biewener. L990. Differential scaling of the long bones in the terrestrial Carnivora and other mammals. Journal of Morph ology 204: 1 57 -1,69. Bieweneq A. A. 1989. Scaling body support in mammals: Limb posture and muscle mechanics. Science 250: 45-48. Carter, D. R., and \7. C. Hayes. 1997.The compressive behaviour of bone as a two-pl-rase porous structure. Journal of Bone and Joint Surgery
954:954-962.
Chatterjee, S., and Z. Zheng. 1997. The feeding strategies in sauropods. Journal of Vertebrate Paleontology 77: 37 A. Christian, A., and W.-D. Heinrich. 1998. The neck posture of Brachiosaurus brancai. Mitteilung aus dem Museum fiir Naturkunde zu Berlin, Geowissenschaftenliche Reibe 1: 73-80. Coombs, !7. P., Jr. 1975. Sauropod habirs and habitats. Palaeogeography, Pdlaeoclimatology, Paleoecology 17 : 1,-33. 1984. Mechanicdl Adaptation of Bones. Princeton, N.J.: Princeton
University
Press.
Currey, J. 2002. Bones. Princeton, N.J.: Princeton University Press. Neck Posture of Sauropods Determined Using Radiological Imaging
.
24J
Einhorn, T. A. 1996. Biomechanics of bone. In L. Bilezikian, L. Raisz, and G. Rodman, eds., Principles of Bone Biology,25-37. San Diego: Academic Press. Francois, F.J.1975. Ligament insertions into the human lumbar vertebral body. Acta Anatomica 91 467480. Frey, E., and J. G. Martin. 1997. Long necks of sauropods. In Dinosdurs, 406409. San Diego: Academic Press. Frost, H. M. 1964. The Laws of Bone Structure. Springfield, Ill.: Charles C. Thomas. Gilmore, C. Vi. 1936. Osteology of Apatosaurus, with special references to specimens in the Carnegie Museum. Memoirs of the Carnegie Museum 1.1:175-300.
Hannson, T. H., T. S. Keller, and M. M. Panjabi. 1987. A study of the compressive properties of lumbar vertebral trabeculae: Effects of tissue characteristics. Spine 12: 56-62. Hokkanen, J. E. 1986. The size of the largest land animal. Journal of Theoretical Biology 118: 491499. Janensch, W. 1947. Pneumatizitat bei Wirbein von Sauropoden und anderen Saurischier. Palaeontogrdphica, supp. 7: 1-725. 1950. Die Wirbelsiiule van Brachiosaurus brancai. Palaeontograpbica,supp. 7: 27 -93. Keller, T. S., D. E. Harrison, C. J. Colloca, D. D. Harrison, and T. J. Janik. 2003. Prediction of osteoporotic spinal deformity. Spine 28: 455462. Longman, H. A. 1933. A new dinosaur from the Queensland Cretaceous. Memoirs of the Q.ueensland Museum 13: 133-144.
Martin, J. 1987. Mobility and feeding of Cetiosaurrs (Saurischia; Sauropoda): \fhy the long neck? In P. J. Currie and E. H. Koster, eds.,
Fourth Symposium of Terrestrial Mesozoic Ecosystems, Short Paper, 154-I59. Drumheller, Alberta, Canada: Royal Tyrrell Museum of Paleontology.
Martin, J., J.V. Martin-Rolland, and E. Frey. 1998. Not cranes or masts, but beams: The biomechanics of sauropod necks. Oryctos 1: 113-120. McMahon, T. A.
197 5. Allometry and Biomechanics; Limb bones in adult ungulates. American Naturalist I09: 547-563.
Mow, V. C., and Sf. C. Hayes. L99L. Basic Orthopaedic Biomechanics. New York: Raven Press. Osborn, H. F., and C. C. Mook. L921. Camardsdurus, Amphicoelias, and other sauropods of Cope. Memoirs of the American Museum of Natural History 20: 181-190. Panjabi, M. M., K. Takata, V. Goel, D. Federico, T. Oxland, J. Duranceau, and M. Krag. 1991. Thoracic human vertebrae: Quantitative threedimensional anatomy. Spine 1,6: 888-901. Paul, G. S. 1988. The brachiosaur giants of the Morrison and Tendaguru with a description of a new Subgenus, Giraffititan, and a comparison of the world's largest dinosa:urs. Hunteria 2: 1-14. 1998. Limb design, function and running performance in ostrichmimics and tyrannosaurs. Gaia 15:257-270. Phillips, J. 1877. Geology of Oxford and tbe Valley of the Thames. Oxford: Clarendon Press. Pugh, J.'W., R. M. Rose, and E. L. Radin. 1975. Buckling studies of single human trabeculae. Journal of Biomechanics 8: 1'99-201. Recker, R. R. 1993. Architecture and vertebral fracture. Calcific Tissue International 53: 5139-142.
246 . David
S. Berman and Bruce
M. Rothschild
Rockoff, S. D., E. Sweet, and J. Bleustein. 1969.The relative contribution of trabecular and cortical bone to the strength of human lumbar vertebrae. Calcific Tissue Research 3: 163-17 5. Rothschild, B. M., and D. S Berman. 1991. Fusion of caudal vertebrae in Late Jurassic sauropods. Journal of Vertebrate Paleontolctgy 11: 29-36. Salgado, L., R. A. Coria, and J. O. Calvo. 1997. Evolution of titanosaurid sauropods. I: Phylogenetic analysis based on the postcranial evidence.
Ameghininiana 34t 3-32. Stevens, K. A., and J. M. Parrish. 1999. Neck posrure and feeding habits of two Jurassic sauropod dinosaurs. Science 284:798-300. Upchurch, P. 1998. The phylogenetic relationships of sauropod dinosaurs. Zoological Journal of the Linnean Society 724 43-103. 'Wedel, M. J. 2003. The evolution of vertebral pneumaticity in sauropod dinosaurs. Journal of Vertebrate Paleontology 23: 344-357. 'Wedel, M. J., and R. K. Sanders. \999. Comparative anatomy and functional morphology of the cervical series in Aves and Sauropoda. /ozrnal of Vertebrate Paleontology 19: 83A. J. A. 1999. A nomenclature for vertebral laminae in sauropods and other saurischian dinosaurs. Journal of Vertebrate Paleontolr,tgy
'$?ilson,
19:639-653.
2002. Sauropod dinosaur phylogeny: Critique and cladistic analyZoological Journal of the Linnean Society 136:21,7-276. 'Wilson, J. A., and P. C. Sereno. 1998. Early evolution and higher-level phylogeny of sauropod dinosaurs. Memoir 5. Journal of Vertebrate Paleontology (supp. to no. 2): 1-68. ZimmeqC. f997. Dinosaurs in motion. Discouer 18(11): 96-109. sis.
Neck Posture of Sauropods Determined Using Radiologicallmaging
. 2!/
12. Evolution of the Hyposphene-Hypantrum Complex within Sauropoda SpeesrrAN Appsrecuie
Abstract The hyposphene-hypantrum accessory articulation complex is present in several archosauromorphs (e.g., rauisuchids). However, because it is absent in early dinosauriomorphs (e.g., Marasuchus) and ornithischians, it can be regarded as a saurischian synapomorphy, and thus plesiomorphic for sauropods. The shape of this complex within Sauropoda is extremely variable and its homologies far from clear. The rhomboidal primitive configuration remains present in
Diplodocoidea. In Macronaria, conversely, the hyposphenehypantrum complex is sharply different, as it is in some basal titanosaurs (e.g., Andesaurus delgadoi and Phuwiangosaurus sirindbornae) that bear hollow rhomboidal hyposphenes. In basal Macronaria, the hyposphene is ventrally widened, and rn Brachiosaurus brancai and other basal titanosauriforms there are both ventrally widened and bifid hyposphenes. Several Late Cretaceous sauropod lineages have independently
lost the hyposphene-hypantrum complex in the dorsal vertebrae (e.g., rebbachisaurids in Diplodocoidea and titanosaurs in Macronaria). In the titanosaur lineage, the lost hyposphene-hypantrum is replaced 248
with a new accessory structure deveioped from a bifid hyposphene plus the medial centropostzygapophyseal laminae. Advanced titanosaurs are characterized by the absence of a hyposphenehypantrum complex; a light, camellate vertebraeq a skeleton much replaced with cartilage; and a wide anterior blade on the ilium for muscies. These adaptations allowed them to develop a relatively free and rapid locomotion, even by heavily armored forms. The titanosaur specializations are predictable in the context of growing angiosperm importance and the relatively low ornithischian diversity in South America, and perhaps throughout the Gondwanan terrestrial ecosystems during the Late Cretaceous period.
Introduction Accessory articulations are fairly common in reptile dorsal verte-
brae (e.g., hyposphene-hypantrum and
zygosphene-zygantrum
complexes). The hyposphene-hypantrum articulation involves a "positive" structure, the hyposphene, on the posterior side of the vertebra below the postzygapophyses, and a "negative" structure, the hypantrum, betr,veen the prezygapophyses of the next vertebra where the hyposphene fits. This complex is rather common in unrelated archosauromorph lineages (e.g., rauisuchids), especially large forms, and is thought to give rigidity to the vertebrae: dorsals and proximai caudals in sauropods and caudals in theropods (Powell 2003). Because the hyposphene-hypantrum complex is absent in early Dinosauromorpha (e.g., Marasucbus) and ornithischians, Gauthier (1986) postulated that it was a saurischian synapomorphy, and thus plesiomorphic for sauropods. The hyposphene-hypantrum complex is present in prosauropods, such as Lessemsaurus sauropoides (Bonaparte 1999), and also in eusauropods, such as Omeisaurws tianfuensls (He et al.
1988), Patagosaurus fariasi (Bonaparte 1986), Bdrapasaurus
tagorei (Jain et al. 1997), Shunosaurus lii (Zhang 1988), Diplodoci-
dae (Hatcher 1901; Janensch 7929), Mamenchisaurus (Young
19 541, Lapparentosurus (Bonaparte 1,999), Camarasauridae (Osborn and Mook 7927), Brdchiosaurus brancai (Janensch 1950a), Phuwiangosaurtts sirindhornae (Martin et al. 1994), and Andesaurus delgadoi (Calvo and Bonaparte 1991). Its presence has also been reported for the titanosaurs Epachthosaurus sciuttoi (Martinez et al. 1991) and Argentinosdurus hwinculensis (Bonaparte and Coria 1.993). The shape of the complex within Sauropoda is extremely variable and its homologies are far from being resolved. However, there is an independent trend in Late Cretaceous lineages of sauropods toward the loss of hyposphenehypantrum articulations as seen in rebbachisaurid diplodocoids
and titanosaur macronarlans.
Institutional abbreuiations. BYU-Brigham Young University; MACN-Museo Argentino de Ciencias Naturales "B. Rivadavia," Argentina; MLP-Museo de La Plata, La Plata, Argentina; and MPEF-Museo Paleontol6gico "E. Feruglio," Trelew, Argentina. Evolution of the Hyposphene-Hypantrum Complex
.
249
Materials and Methods Specimens examined. Material examined included numerous specimens of titanosaurs and the cast of Diplodocus cdrnegii housed
in the Museo de La Plata, Patagosdurus fariasi (MACN 250-326), Chubutisaurus insignis (MACN 18222), cf , Epachthosdurus (UNPPV-920), Andesaurus delgadoi (MUCPV-132), and Neuquensaurus australis, and various specimens at Brigham Young University. Specimens from the literature include diplodocoids (e.g., Apatosaurus excelsus, A. louisae, Barosaurus lentus), basal macronarians (e.g., CamardsAurus supremus) and titanosauriformes (e.g., Brachiosaurus, Venen o s dur u s, Atld s aur u s, and P h uw i an go s dur u s s ir in dh o rnae) . The phylogenetic framework used is based on Salgado et al. (1997), \X/ilson and Sereno (1998), \7ilson (2002), Wilson and Upchurch (2003), and Salgado (2003). Lamindr homologies.In order to understand the homologies of laminae and structures related to the hyposphene-hypantrum compler, it is necessary to refer to'Wilson (1999). He analyzed, correlated, and discussed the nomenclature of sauropod vertebral laminae as used by different authors. This nomenclature is very useful for describing and comparing these complex structures, which not only vary among closely related taxa, but aiso for their position on the vertebrae in a series and ontogenetic stage. Furthermore, lamina development can also vary on each side of the vertebra (e.g., I'Jeuquensaurtts australis, Salgado et al. submitted). The hyposphene involves several laminae in its structure. The posterior centrodiapophyseal lamina laterally frames the region, extending from the dorsal edge of the centrum, to the postzygapophyses dorsally. The laminae are here described, characterized and typified in the complex in order to avoid future confusion. I propose to use "type" laminae when referring to taxa that typify a structure. 'Sfithin a phylogenetic context, the use of "lamellotypes" will be useful in order to avoid wrong comparisons, especially because laminae are used as characters of high phylogenetic value. They are: C e ntr op o st zy gap op h y s e al I amin a e (\X/ilson 79 9 9) . This consists of paired laminae developed between the postzygapophyses and the dorsal edge of the centrum. \flilson (1999) stated that these laminae originate far from the midline in anterior dorsal vertebrae, whereas in posterior dorsals, with a well-developed hyposphene, they originate in the middle. However, anterior dorsals of A. louisae show where the centropostzygapophyseal lamina meers the posterior centrodiapophyseal laminae, a poorly preserved lamina in the position of the centropostzygapophyseal lamina. Furthermore, almost all dorsals of A. excelsus shor,v the centropostzygapophyseal lamina in their normal placement and another pair of laminae in the position of \il/ilson's centropostzygapophyseal laminae. The same occurs in Brdchiosaurus brancai, B. altithorax, the titanosaur Ampelosdurus atacis (see Le Loeuff, this volume), and in Argentinosaurus huincwlensis. Clearly, there exists an additional distinctive lamina. The centropostzvgapophyseal laminae are therefore redefined
250 .
Sebastidn Apesteguia
to lateral extent of the postzygapophyses articular surface to the dorsal edge of the centrum, but not close to the neural canal. The "lamellotlrpe" of this lamina is present in the ninth dorsal of A. louisae (Fig. 12.1). The centropostzygapophyseal iamina is differentiated from another lamina with a relatively close location, the medial centropostzygapophyseal lamina (described below). The centropostzygapophyseal lamina forms the main pillars in posterior dorsals of A. excelszs and A. louisae, but not in the most anterior vertebrae, where the medial centropostzygapophyseal lamina has such a function. The centropostzygapophyseal iamina are the main infrapostzygapophyseal pillars in Haplocanthosaurus (Hatcher 1906), Brachiosaurus brancai, and Brachioslurus dltithorax; however, they are absent in titanosaurs, where they are replaced by the medial centropostzygapophyseal laminae. The basal titanosaurs have both laminae, as seen in AntpelosaLtrus, Maldwisaurus, and Argentinosaurus (Fig. 12.2K-M). Medial centropostzygapophyseal laminae (new). Paired laminae that are developed between the medial-most part of the postzygapophyses and the dorsal edge of the centrum, closely bounding the neural canal. The "lamellotype" of this lamina is present in the first dorsal of A. excelsus (scheme in Fig. 12.2).The medial centropostzygapophyseal lamina forms the main pillars in the anterior dorsals of A. excelsus, A. louisae, Diplodocus carnegii, Barosaurus, and Tendaguria (Bonaparte et al. 2000). In the titanosaur lineage, mid-dorsals of Phuwidngosdurus sirindhornae bear well-developed medial centropostzygapophyseal laminae, which are also the main as paired laminae, developed from the mid-
infrapostzygapophyseal pillars
in
derived titanosaurs, such
as
Isisdurus colberti, Saltdsaurus loricatus, and Opisthocoelicaudia. \fhere the hyposphene is not well developed, as in the third dorsal of A. excelsus (scheme in Fig. 12.2), these laminae exist but have a separate origin at the medial-most ends of postzygapophyses. In specimens where the hyposphene is developed, these laminae could be called the centrohyposphenal laminae, but this probably corresponds to the medial centropostzygapophyseal laminae (Fig. 12.3). The oniy reason to suggest that the medial centropostzygapophyseal lamina and the centrohyposphenal lamina could be different in origin is the presence of a poorly developed centrohyposphenal in the eighth dorsal of A. excelszs, in addition to the presence of the medial centropostzygapophyseal lamina. In the fourth dorsal of A. excelsus is also seen an incipient development of medial centropostzygapophyseal lamina under the hyposphene, The centrohyposphenal lamina (i.e., medial centropostzygapophyseal lamina, where the hyposphene is present) is clear in the ninth 'Where dorsal of A. excelsus. the hyposphene is present, in posterior dorsals, the medial centropostzygapophyseal laminae arise from their base as parallel, tail pillars that bound the neural canal, acting as paired laminae developed between the ventrolateral margins of the hyposphene and the dorsal edge of the centrum, directly boundine the neural canal. Evolution of the Hyposphene-Hvpantrum Complex
.
25
|
,.,.tils @li
w
Fig. 12.1. Primitiue sauropod and Diplodocoid dctrsal uertebrae in posterior uieu. (A) Close-up o/r/:e Apatosaurus lotisae seuenth dorsal uertebra shouing the hyposphene zone and rclated laminae (frotn Gilmore 19.J6). (B-C) Jut,enile and adub o/ Patagosaurus fariasi showing the ontogenetic euolution of th e byposphene in non-neosauropod saurop ods. (D) Barapasaurus tagorei (from Jain et al. 1977). (E/ Apatosaurus louisae seuenth dorsal uertebra. lF) Apatosaurus louisae ninth dorsal uertebra. (G) Haplocanthosaurus pnscus thirteenth dorsal uertebra hnodified front Dalla Vecchia 1998). (H) Dicraeosaurus sattleri posterior dorsal (from Janensch 1929). I-K: Rebbachisaur posterior dorsal uertebrae: 0/Histriasaurusboscarollii (modifiedfromDallaVecchialggS);(J)Rebbachisaurusgarasbae QnodifiedfrontBondpdrte 1999); (K) "Rebbachisaurus" tessonei (from Dalla Vecchia 1998); (.1) dnd (K) Iack hyposphene. Abbreuiatktns: cpr.tl = centropostzigdpophyseal lamina; mcpol = medial centropostzigdpophyseal lamina; tpol = intrdpostzigdpophyseal lamina.
252 .
Sebastidn Apesteguia
$; A
1
}ll|j, , ",
45
ll.---
:fllr 'llli htr
l1l.
'lYl lrl
\YT
|:l
'X; irli
V l
m
ilt
l'l
r
,l
,f{''. H
iTl,
5
789
u-
)i lrt
'l: r-i
r )t ltrt"
i
lll ' lil
rr
,r ''' r\ I
:
]fi
G1 It
23
_rr.-
H il, 1
lif-, r I I r'
lll\ il
r.,{1;
:.] i
I
?34
O1
\$[ ;,;i'
.:i':
. i: ,-; -'t /'
34
Il/. )T/ ;$l l$!
I
fII
il
,.t
2
iilfrw w ,-l .- I 1
r1i', lll i ;l}:i
'
2
t-'
,
2
,,ry
iXt,i
A7
3
2
'1
\lU:int ,lllt
,,1uJi
,,'tI
,
1
,'
; ;'l;
,\T
,
I
'
Fig. 12.2. Sketches shcnuing the distribution of laminae that surrotntd the bypospbene in sauropod dorsal uertebrae. Numbers cctrrespond utith position. (A) Apatosaurus excelsus (modified from Gilmore 193 5) . (B) Apatosaurus louisae (modified front Cilmore 1936). (C) Diplodocus carnegii (modilied from Hatcher 1901). (D)Barosaurus lentus (modified from Lull 7911)), dorsals 1,1, 5, and 9 respectiuelt. (E) The rebbachisaurs Histriasaurus boscarollliand "Rebbachisaurus"tessonei (modifiedfromDallaYecchial99S;CaluoandSalgadol995),mid-posteriordorsals.(F) Camarasaurus grandis (modified from Osborn dnd Mook 1927), anterior, mid, and posterior dorsals respectiuely. (G) Haplocanthosaurus priscus (tnodified from Hatcher 1906), tuo posterior dorsals. (H) Brachiosaurus brancai (modified from Janensch 1950), dorsals 4, 6, 7, and B respectiuely. /1) Eucamerotus foxi (modilied from Hulke 1880), posterior dorsal uertebra. l/ Brachiosaurus altithorar (modified from Riggs 190'1), dorsals 6 and 12 respectiuely. (K) The titdnosaurs Phuwiangosaurus sirindhornae, Andesaurus delgadoi, and Malawisaurus dixeyi respectiuely (modified from Martin et d. 1999; Cdluo and Bondpdrte 1991; and Jacobs et al. 1993), posterior dorsals. lL/ Argentinosaurus huinculensis (modified from Bondparte and Coria 1993), anterir.r and mid dorsals respectiuely. (M) The deriued titanosaurs Ampelosaurus atacis, Isisaurus colberti, Opisthocoelicaudia skarzynskn, and Saltasaurus loricatus (modified frotn Le Loeuff 1995; ldin and Bandl'opddhyal' 1997; and Powell 2003), posterior dorsals. Tbe shape and size and relatiue height of postq'gdpophyses utere not considered.
The medial centropostzygapophyseal lamina most commonly reaches the centrum at the outer borders of the neural canal, but in
'Wilson's
some cases has an oblique development and reaches (1999) centropostzygapophyseal lamina base. This feature unites Brachiosaurus altithorax and Eucamerotus foxi, but it is also present in the last dorsal of A, excelsus.ln Camarasaurus grandis they are the main connection between the postzygapophyses and the centrum via the hyposphene. The medial centropostzygapophyseal Evolution of the Hyposphene-Hypantrum Complex
.
253
A
,6rye\=V.; :i
'':"..
')r-\--^
../1".:1.
h..\ .---
-\
$ .''.,'n
.:
V ,,\
i '',;'-..---)) -rV-'-!
ig. 1 2. 3. Camarasaurus supremus (from C)sborn and Mook 1921), dorsal uertebrae m posterior uien. (A-B) Third and fourth dorsal uertebrae. (C) Closeup of tbe byposphene region of the sixth, seuenth, and eightb uertebrae showing ( e nt ro p os ! 4, go p o p h y sea I I am i n a and medial F
entr op o stzy gdp op b y s e al lamm a. (D-F) Sixth, seuenth, and eighth
c
dor sal uertchro e. A bbreui ations:
ol = centrop o stzigapop lamina; mcpol = medial cp
c
entr op o stzy gdp op
2.5.1
.
h yse
hy seal
al lamtna.
Sebasti6n Apesteguia
lamina is especially well developed under the hyposphene in Diplodocus carnegii, Supersaurus, Cdmarasaurus supremus ante-
rior and mid-dorsal vertebrae, the complete
series of B. brancai, and at least the mid-dorsals of B. altitholarc. Considering that the basal rebbachisaur Histriasdurus has well-developed medial centropostzygapophyseal laminae and no other marked laminae (Dalla Vecchia 1998), the probably homologous laminae present in "Rebbachisaurus" tessonei, which lacks a hyposphene, are also medial centropostzygapophyseal laminae.
Intrapostzygapophyseal lamina (Osborn and Mook L921,). This is developed as an unpaired vertical ridge between the junction of the postzygapophyses and the dorsal edge of the neural canal. As shown by its development in the third dorsal of A. excelszs, it may be paired in origin, but it is commonly fused into a single structure. The "lamellotype" of the intrapostzygapophyseal lamina is present in the second dorsal of A. excelszs. Upchurch (I998) considered this single median lamina a Diplodocoidea character (#107) that supported and buttressed the hyposphene from below on dorsal neural arches of Barosaurzs and DiOlodocus. Thrs
lamina can be present with or without a hyposphene.
'!(here
a hyposphene is absent, the intrapostzygapophyseal lamina is developed between both the medial centropostzygapophyseal laminae, as in the first dorsals of A. louisae, A. ajax, Barosaurus, and Diplodocus carnegii. \fhen a hyposphene is present, the intrapostzygapophyseal lamina is absent or remains as an addition to the base of the hyposphene giving it a rhomboidal shape. \Tithout the intrapostzygapophyseal iamina, the hvposphene is a broadbased triangie. In some cases, this lamina is only developed above the hyposphene and creates the impression that the hyposphene actually hangs from it. Further studies are necessary to determine whether or not the epihyposphenal part of the intrapostzygapophyseal lamina is a different and independently derived lamina. This review of vertebral lamina demonstrates that hyposphenes are the result of the hypertrophy of the intrapostzygapophyseal lamina pius the addition of the medial centropostzygapophyseal lamina.
The Hyposphene-Hypantrum Complex in Phylogeny
Prior to the development of extensive cladistic analyses of titanosaurs, Bonaparte and Coria (1993) erected the family Andesauridae based on several plesiomorphic features, including the presence of hyposphene-hypantrum. This clade, now considered paraphyletic, included Andesaurus and Argentinosaurus (Bonaparte and Coria 1993). However, Andesdurus has a typical hyposphene, which differs from the supposed hyposphenal structure present in Argentinosaurus and the paraplastotype of Epachthosdurus. The cladistic analysis by Salgado et aI. (1.997) resulted in an unresolved polytomy including Epachthosaurus sciuttoi, Malawisaurus dixeyi and derived saltasaurids. Although Epachthosaurus bears procoelous anterior caudal vertebrae, it also has welldeveloped hvposphenes. Conversely, Malawisaurzzs has some procoelous and some amphyplathian anterior caudal vertebrae and no hyposphenes on the few recovered dorsal vertebrae. Salgado et al. (7997) and Saigado and Martinez (1993) recognized that the Argentinosaurlls structures are not real hyposphenes (contra Bonaparte and Coria 1993), and they coded them as different from
in Epachthosdurus. There is no hyposphenehypantrum complex in derived titanosaurs, such as Opisthocoelicaudia skarzynskii (Borsuk-Bialynicka 7977, 71) and saltasaurines (Powell 2003). The loss of the hyposphene-hypantrum complex in posterior dorsal vertebrae was noted by Salgado et al. (1.997) for Titanosauridae and by Sanz et al. (1999) for Eutitanosauria. Upchurch \1999) considered the loss in both mid- and posterior dorsal vertebrae as synapomorphic for his node "Q." A rhomboidal hyposphene is characteristic of prosauropods and most sauropods. It remains relatively similar in primitive eusauropods (e.g., Patagosaurus fariasi, Bonaparte 1986) and in derived Diplodocoidea (e.g., Diplodocus cdrnegil, Hatcher 7901; Apthose present
Evolution of the Hyposphene-Hypantrum Complex
.
255
cttoslurus louisae, Gilmore 1936). Ontogenetically, dorsal vertebrae of Patagosaurus fariasi show a progressive change \Fig. 12.28, C). The juvenile form shows a small rhomboidal hyposphene that ends ventrally in a short intrapostzygapophyseal lamina hanging over the neural canal. The hyposphene bears two concave ventrolateral faces, which surround a par of fossa and are externally limited by curved ridges. The adult form exhibits an elongate rhomboidal hyposphene with a concave ventrolateral face that is reduced or absent. Although both hyposphenes exhibit a slightly different shape, both can be described as a rhomboidal configuration. Diplodocids commonly exhibit a rhomboidal hyposphene (Fig. 12.2),which is clearly seen in Diplodocus. However, there are sev-
eral variations, especially in the anterior dorsals of
some
diplodocids (e.g., Apatosaurus excelszs,. fourth and fifth dorsal vertebrae, YPM 1980), rvhere the hyposphenes can be hollowed, massive, basally widened, or slightly bifurcated ("whale tail-shaped"). In A. excelsus, the hyposphene margin matches the opposite postzygapophyseal articular surface. This species has hyposphenes developed between dorsals four and nine. Although in the fourth and seventh dorsals, the hyposphene is rhomboidal, the fifth and sixth dorsals bear relatively wide hyposphenes due to the lack of intrapostzygapophyseal laminae. The eighth dorsal has a wide, ercavated hyposphene similar to those of macronarians. The ninth dorsal is small and hangs from an intrapostzygapophyseal lamina. In A. louisae, the hyposphene is alrvays rhomboidal, suggesting a major difference with A. excelsus.In the ninth dorsal, however, rhe hyposphene is small and hanging from a lamina. Diplodocus, Barosaurus, and Histriasaurus also have rhomboidal hyposphenes, hanging from the intrapostzl'gapophyseal lamina. Diplodocid hyposphenes can be opened ventrally (e.g., A. excelsus, eighth dorsal) or closed ventrally with a horizontal ridge (e.g., A. excelsus, sixth dorsal).
Histridsaurus boscarollir, a European diplodocoid, is represented by an isolated vertebra (\7N-V6) characterized by the pres-
ence
of a
hyposphene-hypantrum complex (Fig 12.2I), lvell-
developed outer spinopostzygapophyseal laminae laterally limiting the spine, and highly inclined and long diapophyses. Dalla Vecchia (1,998) considered this taxon a rebbachisaurid, but the presence of
a hyposphene-hypantrum in the dorsal vertebra suggesrs that this taxon is a basal form of a more inclusive group (Gallina and Apesteguia
in
press). Neither Rebbachisaurus garasbae, from the
Albian of Morocco, Africa (Lavocat 1954), nor "Rebbachisaurus" tessonei, from the early Cenomanian of Argentina, bear a hyposphene-hypantrum. Furthermore, Bonaparte (1999) characterized the "Rebbachisaurid" type of dorsal vertebra by the ab, sence of a hyposphene. Other non-diplodocid diplodocoids, such as Dicraeosaurzs (Janensch 1929), bear rhomboidal hyposphenes. \Tilson (2002) concludes that Haplocantbosaurus (Fig. 12.2G) is a basal Diplodocoidea, but the results of Majority Rule Consen-
256
.
Sebastidn Apesreguia
sus are not clear, and the phylogenetic position of this taxon is still uncertain between the basal branches of Diplodocoidea and Macronaria. Haplocanthosrurus, described by Hatcher (1906), has
very long, medial, centropostzygapophyseal laminae. These laminae reach a very reduced rhomboidal hyposphene that hangs from the contact of the subhorizontal postzygapophyses. This feature resembles the condition seen in Cdmardsdurus and Diplodocus, and is different from Apatosaurus. Haplocanthosaurus has both rhomboidal and basally wide hyposphenes.
In Macronaria, the hyposphene-hypantrum complex is different from that of most diplodocoids. As remarked by Bonaparte (1999), Camarasaurus grandis has characteristic, double-spined, anterior dorsal vertebrae that exhibit excavated, ventrally widened hyposphenes (YPM 1901 and 1902) supported by two strong ridges, the medial centropostzygapophyseal laminae that are columnar in shape (Fig. 12.3). They are especially well developed in HaplocanthosdLtrLts, Brachiosaurus, and titanosaurs (Fig. 12.1K-M), and are also present in Diplodoczs (Bonaparre 1999). The lateral margins of the hyposphene in Camarasaurzzs differ from Apatosaurus in that they are narrower and do not mirror the postzygapophyses of the following vertebra. In basal Macronaria (r.e., Camarasaurus), the hyposphene is ventrally wide. However, in Cdrnarasaurus grandis, the single-spined, posterior dorsal vertebrae exhibit wide rhomboidal hyposphenes (e.g., YPM 1901 and 1902).
ln
Brachiosaurus brancai and other basal titanosauriforms there are both ventrally wide and bifid hyposphenes, the latter especially evident in the fourth and seventh dorsals. This is also seen in the Sozorasaurus anterior dorsal vertebrae (Ratkevich 1,998), where the hyposphene bears a wide convex base. The hyposphenes are much better developed in B. altithorax (Riggs 1904) than in B. brancai (Janensch 1950). At the sixth dorsal of both
(Fig.0.a)
species of Brachiosaurus, the medial centropostzygapophyseal lam-
inae are well developed, but the centropostzygapophyseal lamina is only developed in the B. brancai and the hyposphene is proportionally larger in B. dltithorax. Comparing the last dorsals, the laminae
development is rather similar, but B. altithorax shows a large hyposphene and B. brancai does not (see Fig. 12.1). They are similar to the posterior dorsals of the titanosauriform Eucamerotus foxi (= Ornithopsis, Seeley, in Hulke 1880, pl. IV fig. 7). The hyposphene complex in Eucamerotus (Fig. 1,2.6C) looks different from that of diplodocoids. The hyposphene is not rhomboidal, but it is basally wide as in basal macronarians. Furthermore, on each side, the hyposphene base is continuous with a ventrolateral medial centropostzygapophyseal lamina, which meets the centropostzygapophyseal lamina before reaching the centrum (Fig. 12.6C, modified from
Hulke 18 80, pl. lY, frg. 7). The highly pneumatic titanosauriform posterior dorsal neural arch from Istria, Croatia (MPCM-V3), which is described by Dalla Vecchia (7998), bears a strong non-bifid hyposphene, hanging from Evolution of the Hyposphene-Hypantrum Complex
.
257
poz hcc
mcpol
Fig. 12.1. Brachiosaurus brancar (from Janensch 1950), dorsal uerlebrae in posterior uiew: tA-B) anterior dorsals; (C) close-up of the hypospbene region; (D) poslerior dorsal. Abbreuiatictrts: c = centrum; hy = hyposphene; hcc = hyposphenal concauity; mcpol = me dial centrop o stzl gdp op hy s eal lamina; nc = neurdl canal; poz = p ostzygap op hysls; spol = s p i n o P o s t zy ga p o p h1'se al lam i na.
the contact of the postzygapophyses (Dalla Vecchia 1998, fig. 118). The postzygapophyses bear a very wide articular surface, as is ryp-
ical of most titanosaurs. The hyposphene is rhomboidal and dorsoventrally elongate, although the distal portion is strongly
weathered (Dalla Vecchia 1998). Dalla Vecchia (1998) suggested that MPCMV3 and Eucamerotus foxi could actually be closer to titanosaurs than to Brachiosaurus brancai.
The Hyposphene-Hypantrum Complex within Titanosauria The hyposphene and hypantrum in some basal titanosaurs are apparently derived from the bifid structure initially seen in Bracbiosaurus brancdi. A bifid hyposphene-hypantrum articulation is also present in the posterior dorsals of a new taxon from the Aptian of Neuqu6n (Bonaparte et al. submitted). This feature could be diagnostic for Titanosauriformes, but Camarasdurus already has a ventraliy widened hyposphene, and advanced Titanosauriformes,
258 .
Sebastidn Apesteguia
t DtD db& basal
eusauropod
basal
macronaraan
Patagosaurus Camarasaurus
W ww@w basal
iilanosauriform Brachiosaurus
basal
litanosaur Andesaurus
-
basal
saltasaurid
eutitanosaurs
Argentinosaurus Epachthosaurus Neuquensaurus
o'oou"
such as Andesaurws and Phuwiangosaurus, bear apparently plesiomorphic rhomboidal hyposphenes.
The family Andesauridae, including Andesaurus, Argentinoslurus, and Epachthosaurus, supposedly share the hyposphenehypantrum accessory articulation, a character not seen in other titanosaurs, plus other features in the pleurocoels of the centra. Although Andesawrus (MUCPv 132) bears hyposphenes, this is not true for Argentinosaurws, where the presacral vertebrae (PVPH 1' Fig.1,2.6D, E), only show a system of hyposphenal bars formed by hypertrophy of the medial centropostzygapophyseal laminae (Bonaparte 1999). These structures originated from the strong lateroventral expansions of typical sauropod hyposphenes (e.g., Patagosaurws Bonaparte 1986; Bonaparte and Coria 1993). This bifid structure, or "hyposphenal bars," of basal eutitanosaurs are formed by both medial centropostzygapophyseal laminae, variably reinforced but always doubled, descending from the center point where postzygapophyses split, leaving a gap between them. As expected, the external walls of the medial centropostzygapophyseal laminae would fit in a hypantrum, the central cavity between both prezvgapophyses (Fig. 12.7C).If these hyposphenal bars formed by
Fig. 12.5. Cladogram shouing the different euolutionary paths of tbe hypospbenes and related lammae
(modified from Osborn and Mook 1921; Bonaparte and Coria 19 9 3 ). P hylogenetic analy sis consistent with Salgado et al. (1997) and.V/ilson (2002), among others.
medial centropostzygapophyseal laminae actually represent the Iateral borders of the complete but hollowed, primitive hyposphenes seen in Phuwiangosaurus or Andesaurus, this would mean a functional continuity among basal titanosaurs and derived saltasaurids. The hypertrophy of the Argentinos*urus "hyposphenal bars" may correlate to the increase of body size and provide a more solid, but not rigid, intervertebral connection (Bonaparte and Coria 1993). The paraplastotype of Epachthoslurus sciuttoi shows hyposphenal bars that resemble those seen in Argentinosaurus. However, they are not present in the holotype specimen, and the paraplastotype should be considered the type of another species. In those titanosaurs that have lost the hyposphenes, the medial centropostzygapophyseal laminae remain present but less developed than in ArEvolution of the Hyposphene-Hypantrum Complex
.
259
;.
') 1
._+_-
mcpol
//I Fig. 12.6. Titanosaur dorsal uertebrae in posterior uiew: (A) Phuwiangosaurus sirindhornae (from Martin et al. 1999); (B) Andesaurus delgadoi (fron Salgado 2001); (C) Eucamerotus foxi (modified from Hulke 1880); (D-E)Argentinosaurus huinculensis (modified from Bonapdrte and Coria 1993); (F) Saltasaurus loricatus (modified from Powell 2003); (G) neural arch of a basal titanosaur from the Aptian of Centdl Patagonid (from Apesteguia dnd Gimdnez in prep.); (H) Mendozasaurus neguyelap (from Gonzdlez Riga 2003); [l Opisthocoelicaudia skarzynskii (from Salgado 2001). Abhreuiations: mcpol = medial c e
260
.
Sebasti6n Apesteguia
n f ro
po st 4' 8d p
up
b\'sea I lam
in
a.
,1 r:
\ .i
.,i{ t,{
t:
l'
Fig. 12.7. Sauropod dorsal uertebrae in anterior uiew showing
h1'pantra: 1A/ Dicraeosaurus sattlerr (from .lanensch 1929; pl. V,
fig.
4t1 1Bl
Aplro:aurus loui>ae
(frctm Gilmore 1936; pl. XXV, fi1. 6/; /C/ Apatosaurus ercelsus (from Cilmore 1936; pl. XXXII, fis. 8); (D)Amphicoelias altus (from Osbom and Mook 1922; pl. 119, fiS.a); (E) Barosaurus lentus (from Lull 1c)19; pl. IV fi4. a); (F) Camarasaurus supremus (from Osborn and Mook 1922; pl. LXXI, fis. e); G) Camarasaurus grandis (from Ostrom dnd Mclntosh 1966; fig. 23). Abbreuiations: hylta = bt-pdtttrum.
gentinosauras. From the ventrolateral ends of the hyposphene originate thin, medial centropostzygapophyseal laminae, or ,,infrahyposphenal laminae," extending ventrally. Similar laminae can be observed in a basal titanosaur (Fig. 12.6G) from Chubut (Apesteguia and Gim6nez 2001). Salgado et aL (1997) defined the Titanosauridae as the clade including the most recent common ancestor of Epachthosawrus sciutEvolution of the Hyposphene-Hypantrum Complex
.
261
toi, Maldwisaurus dixeyi, Argentinosaurus huinatlensis, the titanosaur DGM
*B' from Periopolis,
Opisthocoelicaudia skarzyn-
skii, Aeolosaurus rionegrinus, Alamosaurus saniuanensis, and Saltasaurinae and all of its descendants. Within this clade, the basal
forms are known to have some kind of accessory structure on the dorsal vertebrae that were once mistaken as hyposphenes. As said before, true hyposphenes are not present within this group of derived titanosaurs, the Eutitanosauria (Sanz et aL 1999). The disappearance of hyposphenes from the titanosaur lineage could be related to other processes, such as the dorsoventral shortening of the neural arch (Tidwell, pers. comm.). Discussion Bonaparte and Coria (1993) questioned if the condition in Argentinoscturus has phylogenetic implications, if it is associated with amphiplatyan caudals, if it is associated with the early development of procoelous caudals, or if it was just a functional solution to the extreme body size. The presence of a true hyposphene-hypantrum complex does correlate with those taxa having amphiplatyan caudals (e.g., Andesaurus), because they are both plesiomorphic features for eutitanosaurs, as are broad teeth and phalanges in the manus. On the other hand, those titanosaurs with dorsal vertebrae lacking hyposphenes have procoelous caudals. In derived and chronostratigraphically late eutitanosaurs, such as Argyrosaurws, traces of posteroventral bony edges probably represent remnants of the structure that originated from the medial centropostzygapophyseal laminae. True hyposphenes are not present, however, within derived titanosaurs (except perhaps Ampelosaurus and Epacbthosauras), since loss of the hvposphene-hypantrum compler is characteristic of the group. Several Late Cretaceous sauropod lineages have independently lost the hyposphene-hypantrum articulation in the dorsal vertebrae (e.g., rebbachisaurids in the Diplodocoidea and titanosaurs in the Macronaria). This loss is accompanied by other convergent features that are plesiomorphic to several diverse clades (e.g., narrowcrowned teeth restricted to the anterior region of the snout, narial
retraction, square symphysis). The loss of the hyposphenehypantrum complex would have provided both unrelated taxa a higher mobility of the back. An explanation for these shared features is outside this work; however, the abundance of these convergences in unrelated taxa is remarkable.
Basal titanosaurs, such as Andesaurus and Phuwiangoslurus,
still bear the plesion-rorphic rhomboidal hyposphenes, whereas
a
new taxon from the Aptian sediments of Neuqu6n Province (Bona-
parte et al. submitted) shows
a bifid hyposphene, plus
well-
developed medial centropostzygapophyseal laminae or "infrahyposphenal ridges." The relative development of the hyposphenal bars seems to be related to the vertebral shape. Riggs (1904) noted that there was a relationship between the 'n'
e (ehacti4n
'Ane.rer 'r ''' 'JUIX
height of the neural spines, the breadth of the zygapophyses, and the hyposphene structur e. In Camardsaurus, the postzygapoph.vses are moderately broad, but are nowhere placed far apart, whereas
the hyposphene is relatively large. In Brdchiosaurus, the
zygapophyses are reduced and crowded together near rhe midline. The hyposphene-hypantrum complex is well developed, prevenring
'Within
any lateral displacement of the body. the Titanosauria, rhe vertebrae are rather wide and low, and accordingly they have more widely separated prezygapophyses when compared to the tall vertebrae in other clades. This becomes evident when comparing diplodocoid sauropod dorsal vertebrae with Opisthocoelicaudia and Argentinosdurus. In Saltasaurrs, posrzygapophyses are wider and closer to the midline (as in Apatosaurus), but they lack hvposphenes. Although Brachiosauras has relatively wide posterior dorsal vertebrae, these have very reduced, cup-like postzygapophyses. The neural arch MPEF-PV 1133 (Fig. 12.6G), belonging to a basal titanosaur (Apesteguia and Gim6nez 2001), shows a complicated arrangement formed by a pendant structure, with a constrained base and a rather flattened, basally erpanded hyposphene, which bears a smaller additional structure at the end. From the lateral margins of the hyposphene, two hyposphenal ridges (medial centropostzygapophyseal laminae) are developed lateroventrally to meet each other around the neural channel. Straight, tall, and parallel columns reach the postzygapophyses at the middle (centropostzygapophyseal laminae). The meeting of both extrahyposphenal columns close a small, peri-hyposphenal fossa.
Life without Restrictions: Paleobiological Significance of Hyposphene-Hypantrum Loss Although the hyposphene-hypantrum complex was presenr in primitive sauropods, this probably allowed them to acquire a large body size by stabilizing the vertebral column. However loss of the hyposphene-hypantrum complex occurred independently in the Late Cretaceous rebbachisaurid diplodocoids and titanosaur macronarians, and reached its maximum loss in saltasaurines. The Ioss allowed greater flexibility of the vertebral column (V/ilson and Carrano 1999). Basal titanosaurs such as Argentinosdurus and Epachthosaurzs developed a new kind of accessory structure, probably from the remains of the bifid hyposphene plus a part of the medial centropostzygapophyseal laminae, forming strong "hyposphenal ridges." In advanced eutitanosaurs, no hyposphenehypantrum complex is present. In addition, they developed light camellate vertebrae, a skeleton rvith articulations vastly replaced by cartilage and other calcified tissues, and r,vide iliac blades for locomotor muscles. These adaptations allowed them to have a relatively free and rapid locomotion, despite bearing dorsal scutes. Even in the tail, there is a trend toward the reduction of caudal accessory articulations, especially in titanosaurs. This loss results in
a
greater mobility
of the tail and
supporrs the use
of
the
Evolution of the Hyposphene-Hvpantrum Compler
.
263
"whiplash" tail as a weapon in derived titanosaurs. This suggestion was presented by \X/ilson et al. (7999), based on the development of a procoelic anterior to the mid-caudal vertebrae; the rod-shaped, biconvex distal vertebrae; and the biconvex last sacral (Salgado et al. submitted), rvhich allowed maximum mobility of the tail. Furthermore, increasing the lateral projection of the iliac blades would have also allowed improved control of the tail. All of these specializations in titanosaurs are better understood by considering the growing importance of angiosperms in the Late Cretaceous, and the relatively low ornithischian diversity in the South American and perhaps Gondwanan terrestrial ecosystems during this time. Acknowledgments. My thanks to Leo Salgado for advice and friendly "sauropod talks"; Jos6 F. Bonaparte for encouragement to follow the "wide gauge" of sauropod evolution; Pablo A. Gallina for hearing and restraining an easily flying mind; Fernando E. Novas for support and trust at my beginnings on sauropod studies; Virginia Tidwell for encouraging me to publish; Kristy Curry and Ray Rogers for very useful talks and friendship; Jaime Powell for access to Lillo Institute specimens; and Pablo Puerta for giving me the chance to study an Aptian sauropod from Chubut. To Virginia Tidwell and Kenneth Carpenter, thanks for their help, especially with the writing style of this chapter. Thanks also to Eva L6pez for help in illustration processing, and to the Jurassic Foundation and PaleoGenesis for field support. References Cited
Apesteguia, S., and O. Gim6nez. 2001. A titanosaur (Sauropoda) from the Gorro Frigio Formation (Aptian, Lower Cretaceous), Chubut Province, Argentina. Ameghiniana 37(4\ Suplemento: 4R. Bonaparte, J. F. 1986. Les Dinosaures (Carnosaures, Ailosaurid6s, Sauropodes, C6tiosaurid6s) du Jurassique Moyen de Cerro C6ndor (Chubut. Argentina). Annales de Pal1ontologie 72: 325-386. 1999. Evoluci6n de las v6rtebras presacras en Sauropodomorpha. Amegh iniana 36(2): fi 5-187. Bonaparte, J. F., and R. A. Coria. l993.Un nuevo y gigantesco saur6podo titanosaurio de la Formaci6n Rio Limay (Albiense-Cenomaniense) de la Provincia del Neuqu6n, Argentina. Ameghiniana 30:271,-282. Bonaparte, J. F., \(/.-D. Heinrich, and R. !7i1d. 2000. Revierv ol lanenschia \7i1d, with the description of a new sauropod from the Tendaguru beds of Tanzania and a discussion on the systematic value of procoelous caudal vertebrae in the sauropoda. Paleontograpbica A.256:
25-76. Borsuk-Bialynicka, M. 1977. A new camarasaurid sauropod Opisthocoelicaudia skarzynskii, gen. n., sp. n. from the Upper Cretaceous of Mongoha. Pdleontologia Polonica 37: L-64. Calvo, J. O., and J. F. Bonaparte. 1,99I. Andesaurus delgadoi gen. et sp. nor,-. (Saurischia-Sauropoda), dinosaurio Titanosauridae de la Formaci6n Rio Limay (Albiano-Cenomaniano), Neuqu6n, Argentina. Ame ghiniana 28: 303-3 1 0. Calvo, J. O., and L. Salgado. 1995. Rebbachisaurus tessoneisp.nov. a new
.
Sebastidn Apesteguia
Sauropoda from the Albian-Cenomanian of Argentina: New evidence on the origin of Diplodocidae. GAIA 11: 13-33. Dalla Vecchia, F. 1998. Remains of Sauropoda (Reptilia, Saurischia) in the Lo-"ver Cretaceous (Upper Hauterivian-Lou,er Barremian ) Limestones of SW Istria (Croatia). Geologia Crc,tdtica 51(2): 105-134. Gauthier, J. A. 1986. Saurischiar.r monoph,vly ar.rd the origin of birds. In K. Padian, eds., The Origin of Birds and the Euolution of Flight, 1-55. N{ernoirs of the California Academy of Sciences, no. 8. San Francisco:
California Academ,v of Sciences. Cilmore, Charles. 1936. Osteologv of Apatctsaunts with special reference to specimens in the Carnegie N1useum. Memoirs of the Cdrnegie Mu-
seumll:175-300.
Gonz6lez Riga, B. l. 2003. A new titanosaur (Dinosauria, Sauropoda) from the Upper Cretaceous of Mendoza Province, Argentina. Azeghiniana 40(2): I 5 5-772. Hatcher, J. B. 1901. Diplodocus Marsh: Its osteologv. taxonomy and prob-
a restoration of the skeleton. Memoirs of the Carnegie Museurn 1: 1,-63. Hatcher, l. B. 1906. Osteolog,v of Hapktcanthosaurus, rvith a description of a new species, and remarks on the probable habits of the Sauropoda and the age and origin of the Atlantosaurus beds. Memoirs of the Carnegie Museum 2(7): 1-7 5. able habits, with
He X., Li K., Cai K., and Gao Y. 1988. The Middle Jurassic Dinosaur Fauna from Dashampu, Zigong, Sicbudn.Yol. 4: Sauropod Dinosaurs (2), Omeisaurus tianfuensis. Chengdu: Sichuan Scientific and Technological Publishing House. Hulke, J. 1880. Supplementar,v note on the vertebra of Ornithopsls (See-
ley) = Eucdmerolas (Hulkel. Quarterly Journdl of the Geolctgical Society 36: 31-34. rWinkler, $7. R. Downs, and E. Gomani. 1993. New Jacobs, L. L., D. A. material of an Early Cretaceous titanosaur from Malawi. Palaeontology J6t3):521-ii4. Jain, S. L., and S. Bandyopadhya,v. 1997. New titanosaurid (Dinosauria: Sauropoda) from the Late Cretaceous of Central India. lournal of Vertebrate Paleontolctgy 17 : 114-136. Jain, S. L., T. S. Kuttl', T. Roychowdhury', and S. Chatterjee. 1977. Some characteristics of Barapasattrus tagorei, a sauropod from the Lower Jurassic of Deccan, lndia. In IV International Gonduana Symposium, Calcuttd, India, 204-216. Delhi: Hindusran Publishing Co. 'W. 1929. Die Wirbelsauleder Gattung Dicraeosaurtts. PalaeonJanensch, tographica, supp. 7(2): 37-133. 1950. Die Wirbelsaule von Bracbiosaurus brancai. Palaeontogrdphicct,supp. 7(3): 27-93 (Stuttgart). Lavocat, R. 1954. Sur les dinosauriens du Continental intercalaire des Kem-Kern de la Daoura. Comptes Rendus de la Dix-Neuuiime Sessiort, Congris G1ologique International, Alger. ASGA 21:65-68. Le Loeuff, J. 1995. Ampelosaurtts atacis (nov. gen, nov. sp.), un nouveau Titanosauridae (Dinosauria, Sauropoda) du Cr6tac6 Sup6rieur de la Haute Vallde de I'Aude (France). Comptes Rendus de I'Academie des Sciences de Paris, serie I,321: 693-699. Lull, R. S. 1919. The sauropod dinosaur Barctsaurtts Marsh. Memoirs of the Connecticut Academy of Arts and Sciences 6: I-42. Martin, V., E. Buffetaut, and V. Suteethorn.1994. A new genus of sauropod dinosaur from the Sao Khua Formation (Late Jurassic or Early
Evolution of the Hyposphene-Hypantrum Complex
.
265
Cretaceous) of northeastern Thailand. Comptes Rendus de I'Academie des Sciences de Paris 319, s6rie II: 1085-1092. Martin, V., V. Suteethorn, and E. Buffetaut. 1999. Description of the type and referred material of Phuwiangosaurus sirindhornae, a sauropod from the Lower Cretaceous of Thailand. Oryctos 2: 39-91,. Martinez, R., O. Gim6nez, J. Rodriguez, and M. Luna. 7991. Un titanosaurio articulado de1 g6nero Epachthosauras de la Formaci6n Bajo Barreal, Cret6cico del Chubut. Ameghiniana 26(3-41:246. Osborn, H. F., and C. C. Mook. 1,921,. Camarasaurus, Amphicoelias, and other sauropods of Cope. Memoirs of the American Museurn of Natural Histort, 3: 24--387. Ostrom, J. O., and J. S. Mclntosh. 1966. Marsh's Dinosaurs: The Collections from Como Bluff. New Haven, Conn.: Yale University Press. Powell, J. E. 2003. Revision of South American titanosaurid dinosaurs:
Palaeobiological, palaeobiogeographical and phylogenetic aspects. Records of the Queen Victoria Museum 111 (Launceston). Ratkevich, R. 1998. A new Cretaceous brachiosaurid dinosaur frorn southern Arizona. Journal of the Arizona-Neudda Academy, of Science
31(1):71-81.
Riggs, E. S. 1904. Structure and relationships of the opithocoelian dinosaurs. Part 2; The Brachiosauridae. Field Columbian Museum of' Geology 2:229-248.
Salgado,
L. 2003. Paleobiologia y
Evoluci6n de los saur6podos Ti-
tanosauridae. Thesis, Universidad Nacional de La Plata. Salgado, L., and R. Martinez. 1993. Relaciones filogen6ticas de los titanos6uridos basales Andesaurus delgadoiy Epachthosaurussp. Amegbiniana 30(3 ): 339. Salgado, L., R. A. Coria, andJ. O. Calvo. 1997. Evolution of titanosaurid sauropods. I: Phylogenetic analysis based on the postcranial evidence. Ameghiniana 34 3-32. Sanz, J. L., J. E. Powell, J. Le Loeuff, R. Martinez, and X. Pereda Suberbiola. 1,999. Sauropod remains from the Upper Cretaceous of Lafro (Northcentral Spain): Titanosaur Phvlogenetic Relationships. Estudios del Museo de Ciencias Naturales de Alaua 14 (Ndm. Esp. 1): 23 5-25
5.
Upchurch, P. 1998. The phylogenetic relationships of sauropod dinosaurs. Zoological Journal of tbe Linnean Society 124 43-703. 1999. The phylogenetic relationships of the Nemegtosauridae. Journal of Vertebrate Paleontology 19 706-1,25. 'Wiison, J. A. 1999. A nomenclature for vertebral laminae in sauropods and other saurischian dinosaurs. Journal of Yertebrate Paleontolctgl' 19(4\: 639-653. 2002. Sauropod dinosaur phylogeny: Critique and cladistic analysis. Zoological Journal of the Linnedn Society 136 277-276.
\filson,
J. A., and M. T. Carrano. 1999. Titanosaurs and the origin of "wide-gauge" trackways: A biomechanical and systematic perspecrive on sauropod locomotion. Paleobir.tlogy 25(2): 252-267. Vilson, J. A., R. N. Martinez, and O. Alcober. 1999. Distal tail segment of a titanosaur (Dinosauria) from the Upper Cretaceous of Mendoza, Argentina. Jottrnal of Vertebrate Paleontology 19(31: 59I-594. Wilson, J. A., and P. C. Sereno. I998. Early evolution and higher-level phylogeny of sauropod dinosaurs: Memoir 5. Journal of Vertebrate Paleontology 18 (supp., no.2): 1-68. rWilson, A., and P. Upchurch. 2003. A revision of Titanosaurzs Lydekker J.
266 . Sebastiln
Apesteguia
(Dinosauria-Sauropoda), the first dinosaur genus with a "Gondwanan" distribution. lournal of Systematic Palaeontology 1(3):
125-160. Young, C. C. 1954. On a new sauropod from Yiping, Szechuan, China. Scientia sinicd 3: 491-503. Zhang Y. 1988. The Middle Jurassic dinosaur fauna from Dashampu, Zigong, Sichuan. Journal of the Chengdu College of Geology, supp. 2:
r-12.
Evolution of the Hyposphene-Hypantrum Complex
.
267
"1.3.
Yariation in the Appendicular Skeleton of North American Sauropod Dinosaurs:
Taxonomic Implications D. Rev \Trrnrrn
Abstract Morphological variation in the appendicular skeleton (exclusive of the manus and pes) of North American Upper Jurassic Morrison sauropods was examined in detail in order to determine the range of variation, and to identify taxonomically significant characters within the appendicular skeleton. A clear understanding of morphology and morphological variation within raxa is essential for both phylogenetic and morphometric analvses. The scapulocoracoid, humerus, ischium, and femur were found to be the best elements for taxonomic identification at the generic level, but given good preservation any appendicular element may be identifiable at some significant taxonomic level. Aithough a clear understanding of morphological variation is essenrial to assessing the value of morphological characters, very little is known about variation in all but three tara examined (Apatosaurus, Diplodocus, and Camarasaurus) and care should be taken when using morphological characters of the appendicular skeleton to identify taxa.
268
Introduction Historically, sauropods have been classified and described based primarily on axial elements (Mclntosh 1.990a, 1990b), but recent cladistic analyses have focused much more on the appendicular skeleton (Upchurch 1995;'Wilson and Sereno 1998). However, no comprehensive comparison of morphological diversity in the sauropod appendicular skeleton exists. This study is a first step toward a comprehensive guide to sauropod appendicular material. This paper describes in detail the morphology of the sauropod appendicular skeleton, as well as the morphological differences among the most common North American sauropod genera. Specifically, I will focus on the fauna of the Upper Jurassic Morrison Formation of western North America because the Morrison sauropod assemblage is the most diverse and plentiful currently known. One major difficulty in identifying disarticulated sauropod specimens is the presence of large, multigeneric bonebeds. One such bonebed is represented by the Dry Mesa Quarry in 'uvestern Colorado, in which as many as eight genera of saurpods are present (Curtice and'Wilhite 7996). 'Within the deposit almost every bone is disarticulated and unassociated. Initial studies of the Dry Mesa Quarry by the author revealed the necessity of a "guide" for the identification of isolated iimb bones, since isolated elements are virtually useless if they cannot be assigned to at least the generic level. Understanding morphological variation in individual sauropod appendicular elements is also important for studies of ontogenetic variation, functional morpholog"v, and phylogeny because few associated or complete sauropod skeletons exist. Studies of ontogenetic variation in sauropod limbs would lack significant sample sizes if only articulated or associated skeletons were required (Wilhite 1999). Wilhite (2003b) relied heavily on well-preserved, undistorted limb bones that rvere digitized to examine functional anatomy of the appendicular skeleton. The project would have been impractical without the use of isolated, weil-preserved elements to complement missing elements in associated or articulated specimens. Phylogenetic analyses would be greatly improved if characters based on individual specimens could be documented in isolated elements as well. Therefore, an understanding of variation within individual elements aliows the use of isolated elements to increase sample sizes and improve the overall quality of a given proj-
Institutional abbreuiations. AMNH-American Museum of Natural History, New York, New York; BYU-Brigham Young University Earth Science Museum, Provo, Utah; CM-Carnegie Museum of Natural History, Pittsburgh, Pennsylvania; DNM-Dinosaur National Monumenr, Jensen, Utah; FMNH-Field Museum of Natural History, Chicago, Illinois; KUVP-University of Kansas, Lawrence, Kansas; M\7C-Museum of 'Western Colorado,
Variation in the Appendicular Skeleton of North American Sauropod Dinosaurs
.
269
Fruita, Colorado; SDSM-South Dakota School of Mines, Rapid CitS South Dakota; UUVP-University of Utah, Salt Lake CitS 'Wyoming, Utah; UW-University of Laramie, Wyoming; WDCWyoming Dinosaur Center, Thermopolis,'Sfyoming; YPM-Yale Peabody Museum, New Haven, Connecticut.
Materials and Methods Genera examined include Camarasaurus (Cope 7877), Apatosdurus (Marsh 7877), Diplodocus (Marsh 1878), Barosaurus (Marsh 1890), Brachiosaurus (Riggs 1903a), and Haplocanthosaurws (Hatcher 1903a). Unfortunately, reasonable sample sizes of individual elements (N > 10) are only available for three of these genera: Cdmarasaurus, Apatosaurus, and Diplodocus. Sufficient data for Brachiosaurzs is available from African specirnens; however, the amount of morphological variation between the African species, B. branchi (Janensch 1,974), and its North American counterpart, B. altithorax (Riggs 1903a), is not clear, and I have chosen not to attribute the characters of B. branchi to B. altithorax. Currently, manuscripts on both Bdrosaurus (Mclntosh, this volume) and Haplocanthosaurus (Bilbey and Hall, in prep.) are in preparation that will add much to our understanding of the morphology of these two genera. A recently discovered species of Apatosattrus, A.
ydhnahpin (Filla and Redman 1994), from the Lower Morrison Formation of Wyoming, will be cited frequently as an example of interspecific variation. The following sections focus on the six major limb elements (humerus, radius, ulna, femur, tibia, and fibula) and the six girdle elements (sternal, scapula, coracoid, ilium, ischium, and pubis). Each section begins with a general description of an element. Muscle scars were named based on dissections of numerous specimens of Alligator mississippiensls. The specific explanations and rationale for the naming of various muscle scars, as lvell as detailed descriptions of the corresponding structures in Alligator, can be found in Wilhite (2003b). Next, morphological differences among genera are described in detail. Many elements are defined based on robustness relative to other taxa. The measurements used to define relative robustness in each taxon can be found in \X/ilhite (2003b). Except where noted, it is impossible to distinguish appendicular material at the species level. The importance of each element as a taxonomic indicator based on the morphological differences nored in this study is also assessed. Approximately one hundred elements from twelve different institutions were digitized using an Immersion Microscribe threedimensional digitizer. Criteria used to select individual bones for digitizing and digitizing methods are given in Y/ilhite (2003a). Dig, itized elements rvere assembled by Arthur Andersen of Virtual Surfaces and articulated using Discreet's 3D Studio Max modeling software. Observations and measurements of over five hundred additional limb elements were also collected (see Wilhite 2003b).
270.
D.Ray\(ilhite
Measurements and observations were made of any appendicular element for which length could be determined. Additional measurements of the forelimb and hindlimb bones included: greatest proximal breadth, least breadth, greatest distal breadth, and least circumference. These measurements were made according to the guidelines presented in'Vfilhite (1,999).
Carpenter and Mclntosh (1994) noted that juvenile Apdtosdurus humeri were proportionally similar to adult specimens. Wilhite (1999) demonstrated that limb elements in Camarasdurus exhibit isometric growth patterns. Further measurement of numerous limb elements of Apatosaurus and Diplodocus weli supports an interpretation of isometric growth in these taxa ('Wilhite 2003b), and morphological differences within and across taxa are unrelated to size. Therefore, figures were generated and scaled to the same length using three-dimensional models of digitized elements (see \X/ilhite 2003a for digitizing technique). The advantage of using digitized elements is that bones can be placed in a precise orientation that is easily duplicated. Digitally generated figures also can be scaled and transformed so that all views are of the same size and side.
The Pectoral Girdle Sternal. The sternal in sauropods is a medial paired element situated between the coracoids (Fig. 13.1).It is generally broadly flattened with a dorsoventrally expanded anterior end (Fig. 13.1B).
Anterior
Anterior
Fig. 1 3.1 . Digitized Camarasaurus (VrDC BS-104) left sternal (length = 527 mm): (A) dorsal uietu; (B) lnteral uieu : tC) uenlral uiet,.
Anterior
Lateral
Medial
I I I
__J B Posterior
Posterior
Posterior
Variation in the Appendicular Skeleton of North American Sauropod Dinosaurs
.
271
Anterior
j],,t::.:rt::.)1:::,.::t:::,,:.:.:.:,:tt"l
atlt:taa,a,
:taa:1:t:::?|ti
| : :al
Fig. 13.2. Photograph of a pair of diplodocid sternal plates in uentral uiew on the quarrl'face at D in o saur N at ional Mon Ltment, Utah. Photo courtesy of .lohn S. Mclntosh.
Posterior
The anterior, posterior, and medial borders of the sternal are rugose and imbedded in cartilage in life. The lateral border of the sternal, however; is smooth and may have had a thin, but dense, cartilage covering, forming a tight joint with the coracoid. There is some debate as to the actual orientation of the sternal plates and their position relative to the coracoid. Filla and Redman (1994, fig. 10) illustrate a cartilaginous presternum, which is shown as three quarters of the length of the ossified sternals in a \Tyoming apatosaur. This restoration is in keeping with the hypothesis of Bakker (1987, fig. 13) that the scapulacoracoid was mobile in quadrupedal dinosaurs, increasing the arc of rotation of the forelimb. However, the true ex-
tent of the cartilaginous presternum is unknown. Figure
13.2
shows a pair of articulated sternals from Dinosaur National Monument with the orientation assumed here in the descriptions of the various genera that follow. The sternal plates of diplodocids and camarasaurids are distinct from one another. In general, the diplodocid sternal is relatively short and massive, whereas the camarasaurid sternal is elongate and gracile (Fig. 13.3).'S7ithin the diplodocids, the sternum of
Apatosaurus is more robust than that of Diplodoctts (Fig. 13.3A-D); however, the shape of the sternals is variable within genera. This may be due to the cartilaginous nature of the majority of the sternum, with the observed variation represenring different levels of ossification throughout ontogeny. The morphology of a juvenlle Diplodoczzs sternal in the collections of the Brigham Young University Earth Science Museum, BYU 681*12534 (Fig. 13.48) differs greatly from that of an adult Diplodocus in the collections
272.
D.Ray\filhite
Anterior
Anterior
Anterior
tr
ff
&
p
re
Posterior
Posterior Anterior
D.
Posterior
Anterior Ytaaaa&
tr ?gg t:4,&
1::&
& I.1
E.
Posterior
F.
Posterior
Fig. 13.3. Digitized sduropod right sternals (not to scale). Apatosaurus (BYU 681-4600) in (A) uentral uietu and (B) Iateral uiew, Drplodocts (AMNH61 5 ) in (C) uentral uietu and (D) lateral uie*', and Camarasaurus /WDC BS-104) in (E) uentral uieu and (F) lateral uiew.
of the American Museum of Natural History, AMNH 615 (Fig. 13.4A). Similar ontogenetic differences are observed between the juvenile Camarasauras, CM11338, and the adult specimen, \7DCBS 104.
Coracoid. In sauropods, the coracoid is a square to oval element whose posterior margin is joined to the scapula (Fig. 13.5). The dorsal and anterior borders of the coracoid are thin, but the element thickens posteroventrally. This posteroventral surface forms the anterior portion of the glenoid fossa, and a coracoid foramen is usually located near the element's posterior edge. In all Morrison genera, the coracoid foramen has a closed posterior border in adult specimens, but juvenile Apatosaurus specimens from the Cactus Park Quarry in northeastern Utah show that the posterior border of the coracoid foramen was not closed early in ontogeny in at least one taxon. Fusion of the coracoid to the scapula is related to ontogeny, and none of the juvenile Apatosaurzs scapulae from Cactus Park have fused scapulocoracoids. The coracoids probably articulated with the sternals based on the curvature of the lateral margins of the sternals (Wilhite 2003b). The anterior edge of the coracoid is curved medially, and the anterior edges of the coracoids probably lay close to one another in life, with the pectoral girdle wrapped around the front of the thorax.
The coracoid of Apatosaurus lFig. 13.6) is the most distinct morphologically of the common Morrison genera, featuring a squared anterodorsal margin, compared to the coracoid of Camdrasaurus, which has a rounded anterodorsal margin. The most notable feature of the Diplodocus coracoid (Fie. 13.7A) is the low Variation in the Appendicular Skeleton of North American Sauropod Dinosaurs
.
273
Medial
Anterior
A.
Posterior
Lateral 20 cm
Medial
Anterior
Posterior
B. Fig. 1 3.1. Photograph oi right sternals o/ Diplodocus in uentral uiew: (A) 615: iB) BYU 681-12531.
271
.
D. Ray Wilhite
AMNH
,,
ct'
ol A.
B.
:
Not preserved ,Flg. /3.-i. Digitized rigbt coracoid
o/Camarasaurus /KUVP 129714) in (A) lateral, (B) posterior, dnd (C) tnedial uiews (length = 255
mnl. Abbreuiations: cf = coracoid foratnen, gl = glenoid fossd.
A 20 cm
Anterior
B. 20 cm
Fig. 13.6. Photograph of Apatosaurus right coracoid (BYU 681-4599) in (A) ldteral and (B) medial uietus. Abbreuiations: cf = coracoid foramen, gl = glenoid fossa.
Variation in the Appendicular Skeleton of North American Sauropod Dinosaurs . 275
\
glenoid
Fig. 13.7. Digitized sauropod rigbt coracoids in lateral uiew shotuing uariation in the angle of the scapular suture. (A)
Diplodocus (DNM 1028); (B) Apatosaurus (BYU 681 -4.t99) ; /C/ Camarasaurus l.KUVP 12e71 4).
Not Preserved
angle of the scapular articulation (-50') compared to the higher angle observed in Apatosaurus (Fig. 13.78) (-70') and the verrical
articulation observed in Camarasaurus \Fig. 13.7C).
Scapula. The scapula in sauropods is a massive element with a broad proximal end and a long scapular blade (Fig. 13.8). The distal end of the scapula is variably expanded. All sauropod scapulae have a prominent acromion ridge lying approximately perpendicular to the long axis of the scapula. The acromion ridge divides a large anterior fossa from a much smaller posterior fossa. The anteroventral surface of the scapula is laterally expanded to form the posterior surface of the glenoid fossa. The scapular portion of the glenoid fossa accounts for about two-thirds of the glenoid's total surface area. The cranial edge of the scapula articulates with the caudal edge of the coracoid. The scapula and coracoid fuse during ontogeny (\Tilhite 2003b). The life position of the scapula has been a source of much debate. Some authors have reconstructed the scapula as lying nearly vertical (> 60') against the ribs in sauropods (Riggs 1903b; Osborn and Mook 1919,7921 Hallett 1987). This arrangement leads to a posteriorly facing glenoid, which seems poorly suited for articulation with the almost vertically oriented humerus found in articulated specimens (Gilmore 1925; Bonnan 2001) (Fig. 13.9A). In
276
.
D. RayVilhite
^€ dl
al
sb
\pf
+-+-* q
\'
lcs
\
f-)r A.
",t .
ot-
tl .),
?t
*,*
qr 4[: f\ '* l*.* o.. ,'
.-
z
.rit.ii:: -rl:
.ll:llllllll
..::::::,,,.,
.r)
'\
'loT/.r B.
al I
^E dl
......i...l.i..tll>.
llr'lllllll'l',]:'i
:li:l:r 'l
J\
r\ J
,/ ,l',:'..:, ]\
c.
tr"/ *.;,
.,,1 |
.t@/< r _
-*gf
Fig. 13.8. Digitized sauropod ight scapulae in lateral uiett. (A) Diplodocus (DNM 1028); (B) Apatosaurus (BYU 68 1 -1 061 8) ; f C) Camarasaurus (KU\zP 12971,4) (not to scale). Abbreuiations: af = anterin, fnsso. al = Ltcrotnion ridge, cs = corttcoid sttture, gf = glenoid fossa, pf posterior fossa, sb = scapular blade.
many restorations, the distal end of the scapula sits near the dorsal margin of the neural spines. This orientation leaves little room for the cartilagenous suprascapula indicated by the rugose distal ends of sauropod scapulae. The scapula was almost certainlv oriented more horizontally relative to the ribs (Fig. 13.98). Perhaps the best evidence of this more horizontal orientation is the articulated juvenrle Camarasaurus skeleton. CM 11338 (Gilmore 1925) from Dinosaur National Monument, Utah. The right scapula of CM 11338 appears to be only slightly displaced from its life position and lies -30' from horizontal. More recentln Parrish and Stevens (2002) have shown that apparent modifications in the ribs indicate a subhorizontal orientation for the scapula as well. While this is not definitive evidence, it is consistent with a near-horizontal orientation Variation in the Aooendicular Skeleton of North American Sauroood Dinosaurs
.
277
/ scapula _ q
@"Ph
,
glenoid fossa
- Fig. 13.9. Articulated digitized Camarasaurus (AMNH 664) (bwmerus length = 773 mm) Ieft forelimbs in lateral uieu. (A)
Articulated forelimb with scapula at -50" from horizontal. (B) Articulated forelimb with scapula at -28' from horizontal.
- -
humerusradius
A.
of the scapula in sauropods. Figure 13.9B represents the orientation assumed here for the scapula, with reference to the humerus, in the descriptions of the various genera that follow Scapula morphology is distinctive in sauropods. Apatosaurus scapulae can be distinguished from those of all known Morrison sauropods by the lack of a distally expanded scapular blade (Fig. 13.88) (Mclntosh 1990b). The distal end of the scapular blade in all other diplodocids (and all other known Morrison sauropods) is considerably expanded (Mclntosh 1,990b). The recently described Apatosawrus yahnahpin (Filla and Redman 7994), however, demonstrated that basal apatosaurs also had an expanded scapular blade. Diplodocids can be distinguished from Camarasaurus and Brachiosaunzs by the angle created by the scapular blade and the
acromion ridge (Mclntosh 1990b). In diplodocids the angle is acute, whereas in Camarasdurus and Brachiosaurus this angle is nearly 90' (Mclntosh 1990b). The anterior fossa in Diplodocus is the Iongest (anteroposteriorly), relative ro rhe length of the scapula, of any of the known Morrison genera (with the possible exception of Barosaunzs) (Fig. 13.8A). Camdrasaurzzs has a relatively massive scapula with a short blade and a much-expanded distal end (Fig. 13.8C) . Brachiosdurus has a scapula morphology similar to that of Camarasaurus; however, the scapular shaft is longer, with a thinner "waist" in the scapular blade just caudal to the acromion ridge. The described scapula of Haplocanthosaurus (Hatcher 1903b) resembles that of Camarasauras with the exception that the distal end of the scapular blade is equally expanded 278
.
D. Ray \Tilhite
dorsally and ventrally whereas tn Cdmarasaurus the blade is prima-
rily expanded dorsally. Forelimb Humerus. The humerus in sauropods is expanded both proximally and distally, with weak development of the distal condyles (Fig. 13.10). The head of the humerus is composed of an anteroposteriorly thickened central portion, which thins laterally and mediallS thus forming a raised triangular area on the caudal face of the humeral head (Fig. 13.108, D, F). This raised area articulates within the anteroposteriorly expanded glenoid fossa. The most prominent feature of the humerus is the large, laterally placed deltopectoral crest (Fig. 13.10A, C, E). The humerus is oriented so that the deltopectoral crest faces craniomedially. The distal end of the sauropod humerus is narrower than its proximal end, but the extent of narrowing varies among tara (Fig. 13. 10). The poorly defined distal condyles of the humerus are divided anteriorly by two small longitudinal intercondylar ridges (Fig. 13.10A, C, E) and posteriorly by a broad shallow anconeal (olecranon) fossa (Fig. 13.10B, D, F). The distal condyles of the humerus articulate with the lateral and medial processes of the ulna, with the radius lying directly beiow the intercondylar ridges on the humerus (Bonnan 2003:'S7ilhite 2003b).
The four key characters that help separate Apatosaurus, Diplodocus, and Camarasaurus humeri are overall robustness, proportion of the distal condyles, orientation of the anconeal fossa, and symmetry of the humeral shaft. ApatosAurus humeri are the
mc
C
"4*j4;
. .*te--6 €
-hh
1'
AF G
4drii E;
qP \#'-""" Ff,WX BAHi
*lc t ,.f .i" f
='€!,,'
:4&,8,,', lr
dnn
Fig. 13.10. Digitized sauropod left bumeri (not to scale). Apatosaurus (AMNH 6111) in (A) anterior uiew and (B) posterior uiew; Dtplodocus lBYU
68 l-4-42) in 1C) anterior uiew and (D) posterior uieu; and Camarasaurus (M\YC 2812) in (E) anterior uiew and (F) posterior
uiew. Abbreuiations: af = anconeal fossa, dpc = deltapectoral crest, hh = humeral head, ir =
intracondl,lar ridges, lc = Iateral condyle, mc = medial condyle.
Variation in the Appendicular Skeleton of North American Sauropod Dinosaurs
.
279
Anterior Posterior
Flg. 13.11. Digitized left radius of Drplodocus (BYU 681 -1726) It,tgth = {t,l mntt tn 1At posterior, lBt prctximal, (C) anterbr, and tDt rnedial uiews. Abbreuiation: uls = ulnd ligdment scdr.
most robust (Fig. 13.10A, B), whereas those of Diplodocus are the
most gracile of the three taxa (Fig. 13.10C, D). Cdmardsdurus humeri are intermediate between Apdtosaurus and Diplodocus (Fig. 13.10E). The anconeal fossa divides the distal condyles into subequal portions in diplodocids (Fig. 13.10A-D), but in Camarlsaurus the medial condyle is considerably larger than the lateral condyle (Fig. 13 . 108, F). In diplodocids, the shaft of the humerus is virtually straight in anterior vieq forming an hourglass shape (Fig. 13.1OA-D). Alternatively, Camarasaurzs exhibits a distinct bend in the medial shaft (Fig. 13.10E, F) due to the medially offset humeral head. Apatosaurus is both more robust than Diplodocus andhas a more prominent deltopectoral crest (Fig. 13.10A, C). The humerus of Brachiosaurus can be distinguished from all other North American Jurassic sauropods by its overall gracility alone (least breadth:length = 0.12) (Wilhite 2003). Radius. The sauropod radius is a relatively straight element (Fig. 13.11). The proximal end is triangular in shape, and the posteroprorimal border is contoured to fit into the anterior fossa of the ulna (Fig. 13.118). The radial shaft is gently borved in medial or lateral view so that the posterior aspect is slightly concave (Fig. 13.11D). The distal end of the radius has a distinct ulnar ligament scar on the posteromedial surface where the radius articulates with the ulna (Hatcher 1902; Gilmore 7936; Bonnan 2001) (Fig. 13.11A). Typicallg the distal end of the radius is rectangular in shape.
The radius is a very difficult element to diagnose at the generic level; however, some general trends can be noted. The radius of Apatosdurus (Fig. 13.12A, B) is the most robust of the three primary
280
.
D. Ray \Tilhite
-l
I I
Late
Medial
l .l
Latera
| ,,tJ
__.t
I
__l
l I
I
t
Med ra I
I
I I
_1,.
F
taxa, whereas the radius of Diplodocas (Fig. 73.12C, D) is the most gracile. Unfortunately', the robustness of the radius tn Cdmardsaurus (Fig. 13.12E, F) falls in between that of Apatosaurus and Diplodocus and, indeed, falls within the morphospace of both tara. Morphologically, all three taxa are very similar. The feature that has been considered to be the most useful character for distinguishing Camdrasawrus from the other taxa is its bowed radial shaft (\Wilson and Sereno 1998). However, this character is based on the diagenetically altered radius of YPM 1901 (Fig. 13.13A). Examination of numerous Camarasaurus radt, including P 25182 (Fig. 13.138), has demonstrated that they are no more bowed than the radii of any other North American Jurassic taxon (Wilhite 2003b; Bonnan 2001) (Fig. 13.12). Ulna. The ulna in sauropods is a robust element (Fig. 13.14). The proximal end is triradiate (Fig. 13.148) and composed of three
Fig. 13.12. Digitized sauropod left
radii (not to scale). Apatosaurus (BYU 681-1711) in (A) antertor uicw ond 1St posterinr uieu: Diplodocus (BYU 681-1726) in (C) anterior uiett,and (D) posterior uie*,; and Camarasaurus /AMNH ttt,l) itr ,E) anteriur Iietu and (F) posterior uiews. Abbreuidtion: trls = ulnar Iigament scar.
Variation in the Appendicular Skeleton of North American Sauropod Dinosaurs
.
281
Fig.13.13. Photograph of (A) YPM 1901, Camarasaurus /e/t radius in posterior uiew and (B) F
MNH
25 1
82, Camarasaurus
right radius in posterior uiew. Scale bar: 20 cm. Abbreuiation:
uls = ulnar ligament scar.
^a dl
rl0 ' I
,,r!riu:ll'lll:liii'lllllll!11tr,,,..
/' pp
/
I
B.
mp
Fig. 13.14. Digitized left ulna of Apatosaurus (BYU 681-471 9) (length = 510 mm) in (A) anterior, (B) proximal, (C) distal, and @) p
ostelior uiews. Ab br euiations
:
mp = medial process. a[= anlerior fossa, lp = lateral process, pp = posterior process, rls = radial ligament scar,
282
.
D. Ray\Tilhite
lp
/
mP
t:
processes (lateral, mediai, and posterior). The medial process is the longest of the three and articulates with the humerus so that it is
oriented anteromedially (Fig. 13.14B). The lateral process is usu-
ally thicker than the medial process and shorter overall (Fig. 13.148). The proximal end of the radius articulates in the anterior fossa (Fig. 13.14A, B) created by the lateral and medial processes of the ulna, so that proximally the radius is oriented cranial to the ulna (Bonnan 2001, 2003). The term "posterior process" is used here instead of the more common term "olecranon process" (i.e., Wilson and Sereno 1998), because the ulnae in ali North American Jurassic sauropods lack an olecranon process projecting dorsal to the articular surface of the ulna. The posterior process is the portion of the ulna posterior to the humerus. This process is triangular in shape and is angled caudoventrally in all the taxa discussed here (Fig. 13.14B, D). A prominent dorsal ridge, dividing the articular portion of the ulna (sloping cranioventrally) from the posterior process (sloping caudoventrally), is visible in lateral view (Fig. 13.15C, F, I). This ridge is analogous to the anconeal process of mammals and marks the caudal extent of the antebrachial joint so that the posterior process is not included in the antebrachial joint.
pp
Fig. 13.15. Digitized sauropod rigbt ulnae (not to scale). Apatosaurus (BYU 681-4719) in (A) anterior uiew, (B) proximal uieu, and (C) lateral uiew; Camarasaurus (AMNH 661) in
(D) d?Tteriol uiea (E) proximal uiew, and (F) lateral uieut; and Diplodocus (BYU 681-1726) in (G) anterior uiew (H) proximal uiew, and (I) Iateral uiew. Abbreuiations: IP = Iateral pr()cess, mP = medial ptocess, pp = posterior process, rls = radial ligament scar.
The ulnar shaft decreases in size below the ulnar Drocesses to the Variation in the Appendicular Skeleton of North American Sauropod Dinosaurs
.
283
pp?
qpp
Lateral
j
Fig. 13.16. Digitized right iliunt of Carnarasaurus lDNM I 253/ (length = .\50 mm) in (A) lateral, (B) dorsal, (C) dnterior, and (D) u entral uiew s. Ab br euiat ion s : acet = acetdbulttm, ip = ischiatic pedtrncle, ppd = pubic peduncle, ppp = prcacctobular process.
t
Lateral
c.
-\
'1
'ppd
\,'
acet'
D.
distalend of the ulna (Fig. 13.14C). A radial ligament scar located on the lateral surface of the distal end of the ulna indicates the distal articulation surface with the radius and corresponds with the ulnar ligament scar on the radius rvhen the two bones are articuIated (Fig. 13.14A). The shaft of the ulna has a triangular cross section and the distal end appears triangular when viewed from below (Fig. 13.14C). The main characters of the ulna that aid in taxonomic identification include overall robustness, orientation of the posterior process, size of the medial and lateral processes relative to one another, and development of the dorsal ridge. As with all other foreIimb elen-rents, the ulna of Apatosaurus is the most robust of all known Jurassic taxa (Fig. 13.16A-C). The ulnae of Cantarasaurus and Diplodocus are indistinguishable from one anorher based on robustness. Camdrasdwrzzs differs from all known diplodocids
in having a
posterolaterally directed posrerior process (Fig. 13.15D-F). In Diplodocus and Apatosdurus, the posterior process
without a lateral component. In Camarasaurus, the medial process is noticeably longer than the lateral process (Fig. 13.15H). ln Diplodoczs, both processes are nearly equal in length is directed caudally
(Fig. 13.15H). The dorsal ridge in Camarasaurzs is somewhat more prominent than in diplodocids (Fig. 13.15I). There are no described ulnae from North American Jurassic brachiosaurs, Haplo284
.
D. Ray \Tilhite
cdnthosaurus. For a description Mclntosh (this volume).
of the ulna in
Barr.tscturus, see
Peluic Girdle
Ilium. The ilium in sauropods is expanded both dorsallv and anteriorly, and has a prominent, preacetabular blade (Fig. 13.16). The pubic peduncle is elongate and gracile relative to the size of the ilium (Fig. 13.16A), and it is mediolaterally expanded (Fig. 13.16C, D). The ischiatic peduncle is poorly developed as a smail, rounded process at the posterior margin of the acetabulum (Fig. 13.16A). The ventral surface of the ilium forms the dorsal half of the acetabulum. The ventral surface of the ilium (dorsal surface of the acetabulum) is expanded medially to form a widened {lange (Fig.
13.16D). The dorsal border of the ilium is rounded and forms
a
gentle curve rvhich varies in height among taxa (Fig. 13.16A). In dorsal view, the preacetabular blade of the ilium curves iaterall;', creating a sacral profile that is wide anteriorly and narrow posteriorly (Fig. 13.168). A variable number of longitudinal ridges mark the sacral rib attachment points on the medial side of the ilium. The outer edges of the ilium are relatively thick, but the bone thins toward the center. Areas between the sacral rib attachments are only a few millimeters thick, and many ilia lack much of this surface due to poor preservation. Sauropod ilia differ generically primarily in the relative length of the pubic peduncle, relative height of the ilium, and the shape of the preacetabular process. The ilium of Camarasaurzs is the most distinctive of all the w'ell-known genera from the Morrison Formation. Based on personal observations, the pubic peduncle tn Camarasaurus is the longest of the known North American Jurassic sauropod taxa, with the exception of Brachiosaurus (Fig. 13.17C) (Mclntosh 1,990b\. The ilium ol Brachioscturus can be distinguished from that of Camarasaurus, however, by its erpanded preacetabular process and the high angle of its iliac blade (Mclntosh 1990b). Camarasaurus lha can also be distinguished from diplodocid and Haplocdnthosdurus ilia by the height of the iliac blade when measured from the dorsal border of the ilium to the dorsal border of the acetabulum (Hatcher 1903a). ln Haplocantbosaurus, the ilium is very low and the dorsal surface is almost straight (Hatcher 1903a). In diplodocids, the iiium is intermediate in height between that of Camarasdurus and Haplocanthosdurus (Fig. 13.17A, B). The preacetabular process in Camarasaurus rs
"hooked" ventrolaterally, while most diplodocid ilia have
a
straight, ventral, preacetabular border (Fig. 13.17). Ischium. The ischium in sauropods is posterior to the pubis and is directed caudally to caudoventrally. The proximal end of the ischium is composed of a dorsoventrally elongate iliac process thar articulates with the ischiatic peduncle of the ilium (Fig. 13'18). Just anterior to the iliac process, the anterior surface of the ischium flares laterally to form the posteroventral quarter of the acetabulum (Fig. 13.18A, C, E). The anterior surface of the ischium conVariation in the Appendicular Skeleton of North American Sauropod Dinosaurs
.
285
ppp
ppp
l t
,*l\
A.?
\
/
'-'"-.-\
9pp Fig. 13.17. Digitized sauropod right ilia in lateral uieu (not tct scale). (A) Apatosaurus /CM 21716); /B) Diplodocus /DNM
1018 ; (C) Camarasaurus /DNM 2; I . Ahbreuiatiuns: dce! = acetabulum, ip = ischiatic petluncle, ppd = pubic peduncle; PpP = predcetabular process.
tlooo
I
., vip
\/
!cet'
sists of an expanded flange that articulates with the posterior surface of the pubis, Posteriorly the ischium narrows to articulate with
its mates at the distal end, along the ventromedial distal articular surface (Fig. 13.18A, C, E).The distal end of the ischium is variably expanded dorsoventrally (Fig. 13.18A, C, E). Key features of the ischium include fusion of the distal ends of the ischia, expansion of the distal ends of the ischia, orientation of the ischial shaft, morphology of the cranial border of the ischium, and the angle of articulation of the distal ends of the ischia. In all diplodocids examined, the distal ends of the ischia are fused in adult individuals; however, the ischia are unfused in adult specimens of Camarasaurzs. \While the sample size is very small (-5), it would appear that fused ischia are common in Haplocanthosaurus (Mclntosh, pers. comm.). Diplodocid ischia have expanded distal ends (Fig. 13.18A, C), whereas in Camarasaurus, Haplocanthosaurus, and Brachiosdltrus, the distal end of the ischium is not expanded (Fig. 13.18F). The shaft of the ischium in diplodocids is directed caudoventrally (Fig. 13.18A, C), whereas jn CamarasAurus and Haplocanthosaurus the ischial shaft is rotated 90. relative to the prorimal end and directed posteriorly nearly parallel to the body axis (Fig. 13.18E). Diplodocid ischia are morphologically similar, and both Apatosaurus and Diplodocus have a hook-shaped process on the anteroventral margin of the pubic symphysis (Fig. 13.18A, C). The distal ischia of Apatosaurus and Diplodocws fuse at different angles, with that of Diplodocrzs fusing at an obtuse angle (-100'; Fig. 13.18D) and that of Apatosaurzs fusing at an
.
D. Ray \Tilhite
acet-
,,
ilF..r
right
A
B
sym
drt
acet \
,ilP't-
-z*
t .r(
oi)
right
':t,. :':tlr):
ilp
I
das
acet
,.ilp..-
1"/
t
a
right
pal
E
ohs
Fig. 13.18. Digitized sauropod ischia (not to scale). Apatosaurus (BYU 681-10687) (A) m right medial uiew and (B) articulated in posterior erler-a, Diplodocus /DNM 1227) (C) in right medial uiew and (D) articulated in posterior uiew; and Camarasaurus (UUVP 13.t0 [cast]) (E) in right medial uiew and (F) articulated in posterior uiew. Abbreuiations: acet = acetabulum, das = distal articular surface, ilp = i1io" peduncle, pa = pubic articular surface, sym = ischial symphasts.
Variation in the Appendicular Skeleton of North American Sauropod Dinosaurs
'
287
acute angle (.'80'; Fig. 13.188). The distal ischia of both Camar.tsaurus and Haplocantbosaurus meet in a symphysis along the medial edge of the distal shaft (Fig. 13.18F). In one diplodocid taxon, Seismosaurus, the distal end of the ischium is hook-shaped (Gillette 7991). Although this feature is cited as a generic charaiter (Giliette 1997), it seems more likely to be an ontogenetic effect due to the ossification of the prepubic cartilage in that individual because the outline of the original distal end of the ischium is evident in the figures of the bone (see Gillette 1991,\. Pwbis. The pubis in sauropods is located anterior to the ischium and is directed anteroventrally in all known sauropod genera. The prorimal end of the pubis is expanded mediolaterally and articuiates with the pubic peduncle of the ilium (Fig. 13.19).In diplodocids, the anterior margin of the pubis has a noticeable ambiens process (Fig. 13.19A, C). The posterodorsal border of the pubis forms the anteroventral quarter of the acetabulum. The obturator foramen is incorporated into the caudodorsal margin of the pubis (Fig. 13.194, C, E). Based on personal observations, the posterior margin of the obturator foramen is usually open in young individuals, but it is always closed in adults. The posterior margin of the proximal end of the pubis forms the arricular surface with the ischium (Fig. 13.19A, C, E). The shaft of the pubis is mediolaterally compressed and is much more massive than the ischial shaft. A flange of bone beginning below the articular surface with the ischium wraps around the shaft ventrolaterally to form an enclosed pubic "apron" (Fig. 13.198, D, F). The distal end of the pubis is greatly expanded relative to the pubic shaft with a roughly triangular cross section. A large, trianguIar, articular surface is present on the medial side of the distal end for articulation of the pubes (Fig. 13.19A, C, E). The main characteristics of the pubis useful for taxonomic differentiation include the presence or absence of an ambiens process, the shape of the ambiens process if present, and the relative size of the obturator foramen. Diplodocids are the only North American Jurassic sauropods that have an ambiens process (Fig. 13.19A., C) (Mclntosh 1990b). Diplodocus pubes have a hook-iike ambiens process and Apatosdurus has a rounded ambiens process. Haplocanthosdurus and Camarasaurus show no indication of an ambiens process on the pubis (Figs. 13.19E, 13.20). Haplocdnthoslurus, however, has a very large obturator foramen that distinguishes it
from Camarasaurus and diplodocids (Fig. 13.20).
Hindlimb Femur. The femur in sauropods is a massive, anteroposteriorly compressed bone (see beiolv for the exception; Fig. 13.21). The proximal end of the femur has a poorly defined head but iacks an anterior trochanter. A poorly developed fourth trochanter is located about one-third of the way down the femoral shaft on the posteromedial margin (Fig. 13.21B, C). The femoral shaft narrows beiow the fourth trochanter, reaching its narrowest point about two-thirds of the way down the femur. The distal end of the femur
288 . D. Ray \filhite
lPr I
of
I
amb/
ras
\/
-das p
lP'.
/o
{
amb'
ras
rrr.if
v
,.
.-\/
19
of
.-tipi.-.
Fig. 13.19. Digitized sauroPod pubes (not to scale). Apatosaurus (CM 563) (A) in righnnedial uiew artd tBt articulated in puslerior
rlear; Diplodoe u' iBYU o8ll)ql5' rCt in right medial uiew and rD) articttlated in posterior uieu; and Camarasaurus /UUVP 4939 [castl) (E) in right medial uiew and (F) articulated in ster i or uiew. Abbr euiations: anrb = ambiens process, das = distal articular sur[acc, ias = ischial art icrrlar surftcc. tp = ischiatic peduncle, of = obturator
p o
foramen.
is expanded mediolaterally and the distal condyles are expanded posteriorly (Fig. 13.21B). The lateral (fibular) condyle is the largest and has an intracondylar groove, which separates the lateral condyle into posterior and lateral subcondyles (Fig. 13.21C). The medial (tibial) condyle is smaller than the lateral condyle and lacks an intracondylar groove (Fig. 13.21C). The medial and lateral
Variation in the Appendicular Skeleton of North American Sauropod Dinosaurs
'
289
ras
Anterior
Posterior
Fig. 13.20. Photograph of left ub is of Haplocanthosaurus priscus (CM 10380) in lateral uieu. Abbreuiatioft: ids = ilial
p
articular surface.
20 cm
condyles are separated anteriorly by a weak intercondylar groove and posteriorly by a deep intercondylar groove (Fig. 13.21C). The diagnostic aspects of the femur include the shape of the femoral head, the robustness of the shaft, and the shape and relationship of the distal condyles. The two most difficult femora to tell apart among the common Jurassic taxa are those of Apatosaurus and Camarasdurus. Camarasaurzs has a more distinct femoral head than Apatosaurus (Bonnan 2001), and the distal condyles are perpendicular to the femoral shaft (Fig. 13.22G-I).
190
.
D. Ray Wilhire
fh
fh
1\
I
Posterior
Anterior
intra
ntc Apatosaurus lacks a distinct femoral head, and the medial (tib-
ial) condyle is longer than the lateral (fibular) condyle (Fig. 13.22A-C). An elongate tibial condyle is, in fact, characteristic of all known North American diplodocids (Foster, pers. comm.). The femur of Diplodocus is gracile relative to that of Apatosaurus and Camarasaunzs, but is very difficult to distinguish from that of Barosaurus (Fig. 13.22D-F). Wilson and Sereno (1998) contend that one diplodocid taxon, Amphicoelias, is characterized by having a circular femoral cross-section. However, my analyses of femora from the Dry Mesa Quarry of western Colorado shows that, in at least one quarry, approximately half of the Diplodocus femora have a circular or sub-circular cross-section. The most likely explanation for this observed ratio is sexual dimorphism, because there is no other evidence that more than one species of Diplodocus is present based on vertebral morphology, and no known Amphicoelia.s elements have been identified from the site
inter
"q
Fig. 13.21. Digitized right femur o/Diplodocus (BYU 681-17A14) tlength = 9SS mm) in tAt an7rrlor. (B) lateral, and (C) posterior uiews. Abbreuiations: fc = ltbular condyle, fh = femoral head, ft = fourth trochanter, inter = intercundylar groouc, inlra = intracondylar grooue, tc = tibial condyle.
(Curtice and \Tilhite 1,996). Variation in the Appendicular Skeleton of North American Sauropod Dinosaurs
.
291
:$ Fig. 13.22. Digitized sauropod right femora (not to scale). Apatosaurus (BYU 601-17103) in (A) anterior uiew, (B) ldteral uiea', and (C) posterior uieu; Diplodocus (BYU 681-17011) in (D) anterior uietu, (E) lateral uietu, dnd (F) posterior uiew; and
C.
D.
f
fb
lm:
/\
tc
tc
J-
Catnarasaurus ( cast of YP M 5723) in (C) anterior uie4 (H) lateral uieu, and (1) posterior uieu. Abbreuiatbns: fc = fibular ,;nnd1le. fh = fentural heod. Tt = fourth trochanter, tc = tibial condyle.
Tibia. The tibia in sauropods is quite massive relative to the fibula and articulates with the medial condyle and medial portion of the lateral condyle of the femur (Bonnan 2001) (Fig. 73.23). The proximai end of the tibia is broad and flat, tapering distally to a laterally compressed shaft. A prominent cnemial crest exrends from the proximomedial border of the tibia, beginning just distal to the proximal articular surface and ending about one-quarter of the way down the shaft (Fig. 1,3.23A, C, D, R G, I). The shaft of the tibia maintains a relatively constant width from the base of the cnemial crest to a point about three-quarters of the way dovvn the tibia, where the distal end starts to flare anteriorly as well as laterally. The distal end of the tibia consists of an anterior and a posterior process divided by a relatively broad lateral groove (Fig. 13.23B, E, H). The posterior process is the longer of the two, and the anterior process has a significant cranial expansion. The shape of the distal end of the tibia closely mirrors that of the dorsal surface of the astragalus with which it articulates. The best taxonomic indicators of the tibia include the size of the cnemial crest relative to the width of the proximal end of the tibia, the curvature of the shaft, overall robustness, and the shape of the distal end. The cnemial crest in diplodocids is proportionally larger than in Cdmarasaurus and Haplocantbosdurus (Fig. 1 3.23A, 'Within C, D, F). the diplodocids, Apatosawrws has the largest cnemial crest of all the North American taxa. In lateral view. the ante-
.
D. Ray Wilhite
ut' ;Medial--1
pp-red
lrd l
B
ial
ap
LiLateral--l
icc 1-Medial-i
ppiw: E I'i f"o
ap
L- Lateral-l
D.
I ap
Lateral
F. pp
pp'
t)..
Fig. 13.23. Digitized sduropod right tibiae (not to scale).
pp'
Apatosaurus (CM 556) in (A) lateral uiew, (B) distal t,iew, dntl (C) anterior llez,; Diplodocus (BYU 581-1718) in (D) Iaterdl t,iett', (E) distal uien, and (F) antetior uiet,; and Camarasaurus tYl\4 i6tr// in rC/ lttcral uicu. rHt distol t'iert', and tl) .trrteri,'r t'icw. Abl,reuiottoils: 0P =.til teriut process, cc = cnemiol crest, lg = lateral grooue, pp = posterior process.
Variation in the Appendicular Skeleton of North American Sauropod Dinosaurs
.
293
rior margin of the tibia in Apatosaurars forms a long shallow bow due to the large cnemial crest (Fig. 13.23A). ln Cdmarasaurus, the anterior margin of the tibial shaft is straighter below the small cnemial crest (Fig. 13.23G). Diplodocus and Baroscturus tlbiae can be distinguished from Camardsaurus and Apatosaurus by the gracile nature of the tibial shaft (Fig. 13.23D). Whereas Camarasdurus and Apatosaurus tibiae have very similar robustness, their distal ends are expanded mediolaterally and craniocaudallS respectively (Fig. 13.23B, H). Fibula. The fibula in sauropods is a long, slender element and is always longer than the corresponding tibia (Fig. 13.2a). The proximal end of the fibula is expanded craniocaudally, with the most prominent feature being a large, triangular, tibial ligament scar on the medial side extending from the proximoposterior border diagonally to the anterior border of the fibula, between one-fifth and one-third of the way dou.n the shaft (Fig. 13.24C, R I). In lateral or medial view, the posterior edge of the fibular head is relatively straight, but the anterior edge expands noticeably. Below the wellwidened head, the shaft of the fibula narro\vs until the slightly expanded distal end.
The most noticeable feature of the fibular shaft is the prominent muscle scar for the insertion of M. iliofibularis located just short of halfway down the lateral side (Wilhite 2003b; for alternate interpretation see Bonnan 2001) (Fig. 13.24A, B, D, E, G, H). Below this scar, the anterior edge of the fibula is very narrow compared to the posterior edge. The distal end of the fibula is expanded siightly anteroposteriorly with a large medial expansion forming an astragalar process (FiS. 13.248, E, H) that fits into the fibular fossa on the astragaius. The distinguishing characters of the fibula include the shape of the proximal end, the nature of the transverse ligament scar, and overall robustness. As with the tibia, the robustness of the fibula in Cama-
rlsdurus and Apatosaurus is very similar. ln Camarasaurus the proximal end of the fibula is anteriorly divergent from the main shaft of the fibula (Fig. 13.24G-I). In diplodocids there is no noticeable diver-
of the proximal end of the fibula from the shaft (Fig. !3.244-F). One exception to this character ts Camarasaurus grandis (YPM 1901 and YPM 1905 from Como Bluff,'Wyoming), in which the prorimal end of the fibula does not diverge from the shaft. Another feature that may help distinguish Camarasaurus and Apatosaurus is the nature of the transverse ligament scar on the medial side of the fibula. In Camarasaurus, the anterodistal portion of the ligament scar forms a medial prominence (Fig. 13.24H, I), whereas in Apdtosaurus, there is no prominence (Fig. 13.248, C). In addition, the transverse ligament scar of Camarasaurus extends farther down gence
the medial shaft of the fibula (about one-third of the way; Fig. 73.241) than it does in Apatosaurus (about one-fifth of the way; Fig. 13.24C). Diplodocus fibulae are most easily identified by the extremely gracile
fibula shaft (Fig. 1.3.24D-F). Otherwise the features of Diolodocus fibula are verv similar to those seen in A,atosaurus.
.
D. Ray Wilhite
the
tlt) - '( ,.-'"{
Late ral I
__l
c tls
-
Later al I
I
I
Fig. 13.21. Digitized sauropod right fibulae (not to scale). Apatosaurus (BYU 681-12801) in (A) Iateral uieu, (B) posterior
mp_
:I t:l
uieu\ and (C) medial uieu,; Diplodocus IMV/C No #) in (D) lateral uiew, (E) posterior uiew, and 1F) medial uiew; and Camardsaurus (KUVP 573) in (G) lateral uiew, tHt postcriur uicu'. and (I) medial uiew. Abbreuiations: ap -- astragalar process,
if
=
M. iliofibularis
insertion scdr, mp = medial prominence.
Discussion The utility of any element as a taxonomic indicator is at least partially related to its preservation potential. Flat bones such as the scapula, coracoid, ilium, and sternals are frequently distorted during the diagenesis of the rocks in which they are found. Also, they are very thin in places and these areas are frequently preserved poorlv. Appendicular elements with excellent preservational potential include the humerus, radius, femur, and tibia. These bones are all relatively short and robust compared to the other appendicular Variation in the Appendicular Skeleton of North American Sauropod Dinosaurs
.
295
elements, rvith few thin processes. The remaining appendicular elements (ulna, pubis, ischium, and fibula) are frequently well preserved, but each has peculiarities that limit its preservation potential. The sauropod ulna has medial and lateral processes that are relativeiy thin and at approximate right angles to one another. These processes are easily deformed by sediment compaction. The pubis is a massive element, but the posterior border is thin and many times is not preserved. The ischium is a long, thin (relative to its width) bone rn'ith a thin anterior edge susceptible to diagenetic effects, such as warping, and poor preservation. The fibula is very long and thin relative to the other appendicular elements, and it is very common to see fibulae with the shaft bent in several places due to diagenetic alteration of the bone after burial. Knowing these po-
tential preservational problems and the taxonomically significant morphological features of each bone, it is possible to consider the usefulness of each element as a generic identifier. Sternals are of litle taxonomic significance below the level of family. The probable cartilaginous nature of the majority of the sauropod sternum (based on the rugose borders) suggests that ossification of the sternal plates varied with age. Juvenile sauropod sternal plates differ in shape from those of adults and appear to have been almost totaily surrounded by cartilage, since they are rugose even on the lateral margin (Fig. 13.a). Also, individual variation within genera is not well documented. The utility of the coracoid for taxonomic identification is limited because considerable variation can be observed within genera. Also, many characteristics of the coracoid are based on relative differences between taxa and are difficult to assess in single specimens; however, the quadrilateral coracoid ol Apatosauras is unique, allowing most apatosaur coracoids to be identified with confidence even in the absence of other elements. Filla and Redman (1994), however, have shown that at least one Apatosaurus species, A. yahnabpin, from the lower third 'Sfyoming, had an oval coracoid with of the Morrison Formation of rounded margins, a discovery that calls into question the generic identification of isolated coracoids that do not exhibit the typical Ap ato s aur u s morpholo gy.
Scapulae are very useful for taxonomic identification. Even partial scapulae can, in many cases, be identified to genus; however, uncommon taxa such as Swpersaurus and Seismosaurus are verv likely to be identified as more common Morrison taxa, such as Diplodocus and Barosaurzs, which they closely resemble. Filla and Redman (1994) documented such a case by pointing out that the scapula of A. yahnahpin was originally identified as a cetiosaur based on its morphology, aithough subsequent material clearly demonstrated the scapula belonged to an apatosaur. Humeri are excellent elements for generic identifications; however, the identification of incomplete diplodocid specimens can be difficult because taxonomic affinities are based on robustness and no single portion of the humerus is diagnostic. In addition, A. yahnahpin has a gracile humerus like that of Camarasaurus (Ieast
296 . D. Ray\(ilhite
breadth:length = 0.16) so robustness alone will not help separate diplodocid and Camarasaurus humeri. Nonetheless, as more specimens are described and prepared, it may prove possible to separate some species based on robustness. For instance, Apatosaurus louisae from Dinosaur National Monument is much more robust than any other Apatosaurus species thus far known (Mclntosh 1990a). Based on morphological features and measurements, the radius is not useful for differentiating sauropods at even rhe family level, and identifications should be based on other, more diagnostic elements. However, within a given quarry it is sometimes possible to distinguish between genera based on robustness alone, as can be '!flilhite seen in radii from Dinosaur National Monument (see 2003b). The ulna is of some use as a taxonomic indicator. Overali robustness is the key to distinguishing between Apatosaurus and Diplodocus; however, there are definite morphological differences between Camarasaurus and diplodocids. Given well-preserved material, it is possible to assign at least family-level identifications to isolated elements. Given the presence of all the key features noted above, many can be identified at the generic level as well. The ilium is a moderate to poor taxonomic indicator in most cases due to the poor preservation of most specimens, which hinders identification of many specimens beyond the level of sauropod. \Well-preserved ilia, however, may be easily identified to family based on the suite of characters given above. The ischium is very useful as a taxonomic indicator because of the taxonomic significance of the ischium's shape. Even partially preserved ischia can almost always be identified to family. If the distal ends are present, Diplodocus and Apatosaurws ischia can be distinguished by the angle at which the ischia join; however, there are not enough specimens of most of the rarer sauropods (i.e., Sezpersaurus) to truly assess morphological variation of the ischia in those taxa. The pubis is a relatively poor element for raxonomic identification. Unless the proximal end is well preserved, it is virtually impossible to assign even a family identification to many pubes.'Wellpreserved elements, however, can often be identified to genus. The femur is a very reliable taxonomic indicator; however, Camardsaurus and Apatosaurus femurs are easily confused and, when possible, these taxa should be identified based on a suite of femoral characters. Because identification of many femora relies on the morphology of the prorimal and distal ends, it is important to make sure that these surfaces are relativel)' complete and uneroded if identification is to be attempted. The tibia is a fairly diagnostic element for taxonomic differentiation; however, based on the tibiae examined in this study, it is still unclear how to separate Diplodocus and Barosaurus tlbiae. Also, if the cnemial crest is broken or poorly preserved, the differences between Apatosaurus and Camarasaurus trbiae can be very
Variation in the Appendicular Skeleton of North An-rerican Sauropod Dinosaurs
.
297
difficult to discern because tibial robustness overlaps ('Wilhite 2003b). As with most other limb elements, identifications should be based on more than one feature of that element. Fibulae can be used as taxonomic indicators but should not be relied on for a positive generic identification. As mentioned above for other elements, it will always be difficult to tell Diplodocus from Barosaurus. I have observed that fibulae are easily distorted during diagenesis and it is often difficult to distinguish the subtle features that differentiate genera; however, given well-preserved material, fibulae can frequently be identified to genus with confidence based on overall robustness (Apatosaurus vs. Diplodocus) or morphological features (Camarasaurus vs. diplodocids). Conclusions The two key factors that influence the usefulness of a given appendicular element as a taxonomic indicator are preservation potential
and the presence of taxonomically significant morphological features. Based on observations of over five hundred individual appen-
dicular elements, the best appendicular elements for taxonomic identification of North American Jurassic sauropods at the generic level appear to be the scapulocoracoid, humerus, femur, and ischium. Given good preservation, however, any appendicular element may be identifiable at the genus level. The most important aspect of sauropod research involving appendicular material is to observe and measure enough material to determine the range of 'While variation. it may be said that variation within the appendicular skeleton of Diplodocus, ApatosaLtrus, and Camarasaurws is well understood, numerous taxa known only from a single specimen or small numbers of specimens remain enigmatic. The detailed morphological descriptions given above are intended to be a first step in understanding the range of qualitative morphological variations in the appendicular skeletons of sauropods. Acknowledgments. I would like to thank the Jurassic Foundation, the Louisiana State University chapter of Sigma Xi, and the Louisiana State University Museum of Natural Science for their support of this project. I am also grateful to Art Andersen of Virtual Surfaces for the use of the Microscribe digitizer as well as for editing of data for the project. This research represents a portion of my dissertation work and I am most grateful to my advisor, Judith Schiebout, for keeping me focused on the task at hand. I would also like to thank the rest of my dissertation committee: John 'Wrenn, Laurie Anderson, Barbara Dutroq Paul Farnsworth, and Dan Hillmann for their support. I am also grateful to Ken Carpenter, Virginia Tidwell, and Matt Bonnan for their insightful and constructive reviews of this manuscript, which have greatly improved its content. Thanks also go to the collection managers and curators of the many institutions r,vhere I have worked over the years, including:
Dan Chure, Dinosaur National Monument; Ken 298
.
D. Ray Wilhite
Stadtman,
Brigham Young University Earth Science Museum; Don Burge, College of Eastern Utah; Burkhard Pohl,'lfyoming Dinosaur Center; David Brown, Tate Geological Museum; John Foster and Rod Sheetz, Museum of 'Western Colorado; Larry Martin and Dave Burnam, University of Kansas; Ken Carpenter and Virginia Tidwell, Denver Museum of Natural History; Pete Reeser, New Mexico Museum of Natural History; Jack Horner and David Vericchio, Montana State University; Peter Dodson, Matt Lamana, and Josh Smith, Philadelphia Academy of Natural Science; David Berman and Betty Hill, Carnegie Museum; Mike Brett-Surman, National Museum of Natural History; Mark Norell, American Museum of Natural History; Vicki Yarborough, Yale Peabody Museum; Bill Simpson, Field Museum of Natural Historn Chicago; and Matt Wedel, Sam Noble Oklahoma Museum of Natural History.
Finally,
I would like to thank my mentor and friend, Jack
Mclntosh, for nurturing my initial interest in sauropods and for al-
ways being willing
to
discuss perplexing aspects
of
sauropod
anatomy. References Cited
Bakker, R. T. 1987. The return of the dancing dinosaurs. In S. J. Czerkas and E. C. Olson, eds., Dinosaurs Past and Present. Vol. 1, 39-69. Los Angeles: Natural History Museum of Los Angeles County in association with University of 'Sfashington Press. Bonnan, M. 2001. The evolution and functional morphology of sauropod dinosaur locomotion. Ph.D. dissertation, Northern Illinois University. 2003. The evolution of manus shape in sauropod dinosaurs: Implications for functional morphology, forelimb orientation, and phylogeny. Journal of Vertebrate Paleontology 23(3): 595-613. Carpenter, K., and J. S. Mclntosh. 1994. Upper Jurassic sauropod babies
from the Morrison Formation. In K. Carpenter, K. F. Hirsch, and J. R. Horneq eds., Dinosaur Eggs and Babies, 265-278. New York: Cam-
bridge University Press. Cope, E.D. 1877. On a gigantic saurian from the Dakota epoch of Colorado. Paleontology Bulletin 25 5-10. Curtice, B. D., and R. Wilhite. 1996. A re-evaluarion of the Dry Mesa Quarry sauropod fauna with a description of new juvenile sauropod elements. In Geology and Resources of the Paradox Basin: IJtah GeoIogical Association, 1996 Field Symposium,325-338. Utah Geological Association Guidebook, no. 25. Salt Lake City: Utah Geological Association. Filla, J., and P. D. Redman. I994. Apatosaurus yahnahpfu: Preliminary description of a new species of diplodocid sauropod from the Late Jurassic Morrison Formation of southern Wyoming, the first sauropod dinosaur found with a complete set of "belly ribs." In The Dinosaurs of 'W1,oming, 159-178. Wyoming Geological Association 44th Annual Field Conference GuidebooA. Casper: Wyoming Geological Associa-
tlon. Gillette, D. 1991. Seismosaurus halli (n. gen., n. sp.) a new sauropod dinosaur from the Morrison Formation (Upper Jurassic-Lower Cretaceous) of New Mexico, U.S.A. Jottrnal of Vertebrate Paleontology 1L: 417-433. Variation in the Appendicular Skeleton of North American Sauropod Dinosaurs
.
299
A nearly complete articulated skeleton of CamarasdLtrus, a saurischian dinosaur from the Dinosaur National Monument. Memoirs of the Carnegie Museum 10:347-384. 1936. Osteology of Apatosdurus with special reference to specimens in the Carnegie Museum. Memoirs of the Carnegie Museum I1:
Gilmore, C. \7. 1925.
75-300. M. 1987. Bringing dinosaurs to life. In S.J. Czerkas and E. C. Olson, eds., Dinosattrs Past and Present. Yol. 1,97-71.3. Los Angeles: Natural History Museum of Los Angeles County in association with 'Washington Press. University of Hatcher, J. B. 1902. Structure of the forelin.rb and manus o{ Brontosaurus. Annals of the Carnegie Museum 1.: 356-376. 1903a. A new sauropod dinosaur from the Jurassic of Colorado. Proceedings of the Biological Society (\Tashington) 1'6: 1'-2. 1903b. Osteology ol Haplc,tcanthosanrus, with description of a new species, and remarks on the probable habits of the Sauropoda and the age and origin of the Atlantosaurus beds. Memoirs of the Carnegie Museum 2: 7-7. 'Wirbeltierfauna der TendaguruJanensch, W. 1914. Ubersicht irber die 1
Hallett,
Schichten, nebst einel Kurzen Charakterisierung der neu aufgefirhrten
Arten von Sauropoden. Archiu fiir Biontologle 3: 81-1 10. Marsh, O. C. 1,877. Notice of some new dinosaurian reptiles from the Jurassic Formation. American lournal of Science, ser. 3, 14: 514-576. 1878. Principal characters of American Jurassic dinosaurs. Part I American Journal of Science, ser. 3, 16: 41'1'41'6. 1881. The sternum in dinosaurian reptiles. American lournal of Science, ser. 3, 19: 395-396. 1890. Description of new dinosaurian reptiles. American Journal of Science, ser. 3, 39: 81-86. Mclntosh, J. S. 1990a. Species determination in sauropod dinosaurs with tentative suggestions for their classification. In K. Carpenter, and P. J. Currie, eds., Dinosaur Systematics: Approaches and Perspectiues, 53-69. New York: Cambridge University Press. 'Weishampel, P. Dodson, and H. Os1990b. Sauropoda. In D. B. m6lska, eds., The Dinosauria, 345-401. Berkeley: University of California Press. Osborn. H. F.. and C. C. Mook. 1919. Reconstruction of the skeleton of tlre sauropod dinosaur Camarasaurus Cope (Morosaurus Marsh). Proceedings of the National Academy of Sciences 6: 15. 1,921,. Camarasaurus, Amphicoelias, and other sauropods of Cope. Memoirs of the American Museum of Natural History 3: 246-387. Parrish, M., and K. Stevens. 2002. Rib angulation, scapular position, and body profiles in sauropod dinosaurs. Journal of Vertebrate Pdleontology 22(3A): 95A. Riggs, E. S. 1903a. Brachiosaurus abithordx, the largest known dinosaur. American lournal of Science, ser. 4, 15: 299-306. 1903b. Structure and Relationshil2s of Opisthocoelian Dinosaurs. Part I: Apatos avus Marsb. Field Columbian Museum Geology Series, no.2. Chicago: n.p. Upchurch, P. I995. The evolutionary history of sauropod dinosaurs. Pi:i/o-
sophical Transactions 36s-390.
300
.
D. Ray $Tilhite
of the Royal
Society
of London B,
349:
Wilhite, R. 1999. Ontogenetic variation in the appendicular skeleton of the genus Camarasaurus. Master's thesis, Brigham Young University.
2003a. Digitizing iarge fossil skeletal elements for threedimensional applications. Palaeontologia Electronica 5(2\. http:ll
palaeo-electro nica.or gl 20 02-2lscan/issue2-02.htm.
in
2003b. Biomechanical reconstruction of the appendicular skeleton
three North American Jurassic sauropods. Ph.D. dissertation,
Louisiana State Universitl.. http://etd02.1nx390.lsu.edu/docs/available/ etd-0408 103-003549/. 'Wilson, J. A., and P. C. Sereno. 1998. Early evolution and higher-level phylogeny of sauropod dinosaurs. Journal of Vertebrate Paleontology, memoir 5, 18(2, supp.): 1-68.
Variation in the Appendicular Skeleton of North American Sauropod Dinosaurs
.
301
1-4.
First Articulated Manus of Diplodocus carnegii Mercornr W. Bpoprr Jn. eNo Devro L. Tnpxrpn
Abstract Though Diplodocws carnegii is one of the more familiar Jurassic sauropods, the manus of this animal is known only from isolated elements. In existing mounts, the manus has been created by substi-
tution from the remains of other taxa. In 1997, the articulated manus of a single, subadult Diplodocus cf. carnegii (V'DC-FSO01A) was recovered from a Morrison Formation location (FS Quarry) in northern '$7yoming. This includes five
metacarpals, one phalanx, and an ungual; all are short and of gracile proportions. Also present were two articulated pes, and many bones of the appendicular and axial skeleton. After careful comparison with corollaries from Sauropoda known to be extant within that time
frame and area, alI \7DC-FS001A elements are interpreted as Diplodocidae. Diplodocus hayi, the only previously recognized member of this group with a reasonably intact and articulated distal forelimb, is too robust to account for the N7DC-FSOO1A manus. Since nothing in the observed anatomical features of 'ff/DC-FS001A is in unquestioned agreement with a corresponding D. longus hoiotype (YPM 1920) model, but does agree with the known D. carnegii mate302
rial, the manus is referred to D. carnegif providing the first objective information on the structure and arrangement of this assemblage for
D. carnegii.
Introduction From the first discovery (D. longus), by Benjamin Mudge and 'Wendell Williston in 1877, the various species of Samuel 'With the Diplodocidae have enjoyed a large amount of publicity. rush to describe new taxa of dinosaurs, the "bone wars" of the late nineteenth century (particularly the competition between Cope and Marsh) resulted in some errors of description. Some of the initial information regarding the type specimen Diplodocws longws has proved to be incorrect (Mclntosh and Carpenter 7998). For example, the manus found in loose association with other bones of the type Diplodoczs (Marsh 1896) was dismissed as not definitively part of the D. longus specimen (Mclntosh and Carpenter 1998). Since that time, other species of Diplodocus have been recovered. Diplodocus carnegii is known from several reasonably complete 'Western Interior of the United States but, until skeletons from the recently, none were found with an articulated manus (Mclntosh 1990a\. Diplodocus lacustris, diagnosed from only a few skeletal elements, also was found to be missing the manus (Marsh 1884; Mclntosh 1990a). Holland (1,924) described Diplodocus hayi, the only specimen shown conclusively to be in the family Diplodocidae and also to have a fairly complete, articulated manus. The manus of Apatosaurus has been recovered, but that animal is more distantly related (Marsh 1883; Mclntosh 1990a). With sauropods in general, metapodial elements are often lacking upon recovery.
Good manus preservation exists for other North American sauropods of the Western Interior. Several camarasaurids, including
BYIJ 9047, GMNH-PV 101, and holotype YPM 1910, have produced complete, or almost complete, maniis. Several brachiosaurid skeletons (Janensch 1961; Tidwell 2001) have also been recovered with relatively intact, articulated manus. No manus is currently known for Haplocantbasaurus (Hatcher 1903a; Mclntosh and Carpenter 1998) or Barosaurus (Lull1919 although see Mclntosh this volume). In the course of this study, various taxa of Sauropoda, including those from other continents, were examined and compared. This was done by means of actual bones, where practicable, and
with published descriptions supplemented, in some instances, with good color photographs taken from many angles. These include: Camarasaurzs AMNH 711; YPM 1901,4633, L9L0;BYU 9047; GMNH-PV 101; a Camarasaurus lentws from WDC described in this volume; Barosaurus YPM 429, 479; Haplocanthasawrus CM 572,879; Apatosaurus YPM 1980; Tate 001; CM 563,3018; brachiosaurid DMNH 39045 Brachiosaurus brancai (from Janensch 1,961); Tornieria robusta (from Janensch 1961): titanosauriformes DMNH 40932; Diplodocws HMNS 175; CM 84, 94, 307, 662; NMMNH 3690; USNM 10865; YPM 1920; and DMNH 1494. First Articulated Manus of Diplodocus carnegii
.
303
Fig. 14.1. Reconstructed natural position of manus (\X{DCFS001A) in proximal uiew, with carpal in place, and anteriur uieu. tuithout carpal.
The metacarpals of specimen \fDC-FS001A (Fig. 14.1), recovered from the FS Quarry, are not as elongate as those of camasaurids or brachiosaurids studied. Diplodocws hayi metacarpals 'WDC-FS001A, appear much more robust than those of having relatively expanded prorimal ends, and width:height ratios more like those of apatosaurids. Apatosaurs, in turn, have width: height:breadth ratios of their corresponding bones indicating a robustness greater than that in all other sauropods considered here for which good manus exist. Measurements of the \X/DC-FS001A manus elements may be found in Tables 1,4.1, 74.2, and 14.3. Along with the manus, several articulated mid-caudal vertebrae (FS 20,746,1.47,748,150A, 150B), and other associated cau-
dal vertebrae have been recovered from the FS Quarry. \7hen viewed laterally, the articulated caudals show a general elongate appearance, with deep latero-medial excavation, and a relatively flat area along the ventral surface punctuated by a significant cranial-caudal groove (Fig. 14.2). There is also strong lateral constriction. One caudal in this series, FS 181, shows a complete neural spine, with pre- and postzygopopheses intact (the other caudals also have fairly complete neural spines, but with some damage). 304
.
Malcolm \7. Bedell
Tr.
and David L. Trexler
TABLE 14.1. Dimensions (in mm) of the Right Manus of Diplodocus carnegii
(wDC-FS001A)
il
III
146
208
198
191
153
63
71
72
731.r
83
Metacarpal Greatest
Proximal
length breadth
852'
Prorimai depth
33
47
25
2r
21
Distal breadth
65
76
65
551r
71
602)
30 Least circumference 130
Distal
depth
37
26
24
.)L
139
104
112
125
I rX/ithout fragment.
2lWith fragrnenr in approrimare po'itiorr.
TABLE 14.2. Dimensions (in mm) of the Carpal oI Diplodocus carnegii (WDC-FS001A)
Transverse breadth
96
Anteroposterior diarneter
64
Thickness
.t1
TABLE 14.3. Dimensions (in mm) of the Phalanges ol Diplodocus carnegii (wDC-FS001A)
Greatest length Proximal breadth
Distal breadth
Phalanr I-2 tungualt 105 (distal end missing) 66
Phalanx [-
| {proximalt 69 34
I7
This caudal neural spine clearly angles toward the posterior of the tail, and slightly beyond the posterior rim of the centrum. No caudal vertebra has erect neural spines. Taken together, these characters are diagnostic for Diplodocus carnegii (Giimore 1932 Holland 1906; Hatcher 1.901,7903b). Also, none of the vertebrae in this series display the extreme, anteroposterior reduction of the centrum seen in haplocanthasaurs (Hatcher 1903a, 1903b). First Articulated Manus of Diqlodocus carnegii
.
305
Fig. 14.2. Articulated mid-caudal uertebrae t\v/ DC-1500 I A1 in sittt. offering a left lateral uiew, caudals to right being more distal. Note ektngation of centra and neural spines (where exposed on the three most distal caudals) leaning distally.
Three double-beamed chevrons were locared near the caudals.
One of these (FS 168), completely prepared, shows a gracility of appearance not seen in similar chevrons of the apatosaurs. A partially prepared cervical vertebra (FS 56) has a measured ratio of three to one (510 rnm to 170 mm) of the length of the centrum to its diameter. This would indicate a ratio not compatible with those of the centra of barosaurs (Lull 1,91,9). FS 56 also appears to resemble very closely cervical #9 or #10 in Hatcher's 1901 Diplodocus morphology (both attributed to Diplodocus carnegii). A complete scapulocoracoid is articulated, but unfused (FS 182, 186). The scapula has an expanded proximal plate, with a ridge separating the inferior and superior fossae, forming an acute angle with the shaft. This was seen as a Diplodocus character by Osborn and Mook (1921), and Mclntosh (1990a). Lack of fusion is viewed as an indication of subaduir status (Tidwell, pers. comm.). It is of a correct size to belong ro this animal (Vz to 2/z that of observed adults), and is very closely associated with a humerus (FS 151), as as the articulated caudals. As with FS 56, the outline and interior features of this bone are indistinguishable (with the exception of its being unfused to rhe coracoid) from the Hatcher description (1901) attributed to Diplodocus carnegii. A left humerus, FS 151, still in siru, directly covers a section of the articulated caudals. \fell exposed, its dimensions match perfectly the fully prepared right humerus (FS 221), and both are r/z to 2/s the size of adult diplodocids. There is a strong angulation of the bisecting planes (nearly 45') passing through the greatest diameter of each end. This torque, in turn, throws the deltoid crest far under the main axis of the bone. Gilmore (1932) thought this angulation to be a diplodocid character; easily distinguished from the much
well
306
.
Malcolm'i7. Bedell Ir. and David L. Trexler
i
,;"i
:
::.t
r*,'f.ii
*"
't*
Fig. 14.3. Complete right pes WD C-F 5001 A), witb arti culated astrdgalus in upper right, partially exposed in situ, frorn only known photogrdph (35mm), circa 19c)7. (
straighter appearance of apatosaurids (holotype YPM 1980), or camarasaurids (Mclntosh et aI. 1996b). A radius (FS 185) and an ulna (FS 194) were found in perfect articulation with each other, and with FS 151. They are very slender, and short, as with other diplodocids. Two compiete pes were also recovered from the FS site. One pes was found articulated with an astragulus, and was less than a meter from the manus (Fig. 1a.3). There was no calcaneum. The
astragulus
is much more gracile than those of examined
ap-
atosaurids and camarasaurids, with a very short medial projection. 'Whereas this is diagnostic to the genus Diplodocus, the diagnostic utility of these features to species level is not assumed (Wilhite, pers. comm.). Overall, the metatarsals are very gracile and thin, with a small latero-distal process on the plantar edge of metatarsal I partially damaged by pathology (Rothchild, pers. comm.) on one pes, but more evident on the second. Some workers consider this to First Articulated Manus ol Diplodocus carnegii
.
307
be a diagnostic dipiodocid feature (Averionov et al. 2002; McIntosh et al.7992:164).The longest metatarsals are III and IV, with
a
phalangeal formula of 2-3-3-2-1, all listed characters of Dipiodocidae (Mclntosh 1990a). Institutional abbreuidtions. AMNH-American Museum of
Natural History, New York, New York; BYU-Brigham Young University, Provo, Utah; CM-Carnegie Museum of Natural History, Pittsburgh, Pennsylvania; DMNH-Denver Museum of Nature and Science, Denver, Colorado; GMNH-Gunma Museum of Natural HistorS Tomioka, Gunma, Japan; HMNS-Housron Mrrseum of Natural Science, Houston, Texas; NMMNH-New Mexico Museum of Natural History, Albuquerque, New Mexicc.,; Tate-Tate Geological Museum, Casper, Wyoming; USNM-National Museum of Natural History, Smithsonian, 'Washington, D.C.; WDC-Big Horn Basin Foundation/Wyoming Dinosaur Center, Thermopolis, Wyoming; and YPM-Yale Peabody Mrrseum. New Haven. Connecticut.
Geology and History The specimen was discovered at the FS Quarry, located approximately 3 miles southeast of the town of Thermopolis, Wyoming, on the \farm Springs Ranch. The quarry is on the Thermopolis Anticline, and within the Morrison Formation (Fig. 1a.a). In this localit1', the Morrison Formation is bounded above by the unconformity of the cliff-forming Pryor Conglomerate Member (Ostrom 1970) of the Cloverly Formation (Lower Cretaceous), and overlies the marine Sundance Formation (Upper Jurassic). No formal subdivision has been applied to the Morrison of the Big Horn Basin in this area, r,vhich here is 63 meters thick (Carson 1999). This is informally divided into three generalized units: a lorver calcareous mudstone unit; a middle, fine-grained, quartzose sandstone unit; and an upper calcareous mudstone, or "carbonate mud" (Jennings 2002) unit, with sandstones interbedded (Bjoraker and Naus 1996). Subsequent studies by Cleaves and Carson (Carson 1999), Houck (pers. comm.), and Turner (pers. comm.) suggest that the paleoenvironment at the time of deposition was likely dry and monsoonal. Paleotopography would have been characterized by anastomosing streams and scattered small lakes, whose levels change when crevasse splays allow recharging of stream water during the monsoon (Jennings 2002). The bone-bearing stratum of the FS Quarry is a calcareous, nodular mudstone (Carson 1999),hmestone (Houck, pers. comm.), or "carbonate mud" (Jennings 2002). Houck (pers. comm.) interprets this environment as "probable la-
custrine." Other workers offer "para-lacustrine origin" (Jennings 2002), referring to a possible epherneral nature for the water body. Recent petrographic work utilizing X-ray diffraction, thin-probe chemicai analysis, and SEM on thin sections (Jennings in press) has confirmed a lacustrine interpretation. Ostracodes are clearly visible in the thin sections.
.108. Malcolm \7. Bedell
Tr.
and David L. Trexler
Geologic Section Warm Springs Ranch (neaf FS Quarry) Carson.
1
999
Geologic Sectlon Detail including FS Quarry (rt.) Houck, K., 2001 (unpublished)
tlu
fl Conglomerale [-- sandstone fl CalcareousSandstone El
F_!
section
Yot]*|.. scare (n)
f
50
I
L0
SilIStOne
Mudstone
e Bones r." Concretions
Gray siltstone, 1/4" beds: blocky fracture Gray fine sandstone, weathers tan, in 4" beds. Some layers splintery; some more nodular and €lcareous Gray siltstone,
FS Quarry
Section Detail
Dark gray mudstone, hard, massive, non-calcareous
Houck, K., 2001 (unpublished)
Bone-bearing beds at
FS
bl6ky in beds, 1/4" thick
Gray flne sandstone, weathers tani slightly cal€reous, massive
Dark gray, llne sandstone, nonralcareous
lI
L
Sandy limestone, nodular, fubbly, with mottles and small nodules, as below Gray, tine cal€reous sandstone to sandy limestone, quartzose with small nodules 1/4" to 1/8", & white mottles
Fig. 11.4. (A) V/arm Springs Ranch geologic section, from a location near FS Quarry. (B) Ce,tlogic Sertion Detatl. shnwing
position. (C) Quarrl, Detail, taith bonebearing beds from tuhich elements
FS site stratigraphic FS
of WDC-FS001A were recouered, including the right manus.
Although several researchers have studied the geology of this area, until recently none had noted the dinosaur bones here (Bjoraker and Naus 1996; Ostrom 1970). Darton (1906) mentions dinosaur bones in the region, but not in the Thermopolis area. Oil and gas geologists (Horn 7963) had also failed to notice the bones. In 1993, Dr. Burkhard Pohl became aware of the potential for dinosaur remains on the'Warm Springs Ranch. \7hile investigating this potential, several important discoveries were made. To date, over fifty fossil-bearing locaiities have been identified on the ranch. Research on these sites has been conducted with the supervision of the Big Horn Basin Foundation/$Tyoming Dinosaur Center since 1995. The FS Quarry was discovered by Ty Naus in 1993. 'Work was not attempted vntrl 1.997 due to the dangerous 30o slopes lit'WDC tered with large boulders. Since that year, staff and volunteers from the'Western Interior Paleontological Society (WIPS), and other organizations, have begun excavations and research at this site. Nearly a hundred bones were recovered during the 1997 seaFirst Articulated Manus of Diplodocus carnegii
.
309
Fig. 11.5. Quarry Map of the Loot Site showing distribution of bones witb manus assemhlage ntdgnified by a factor of 3 in blown-up
image (in box). Otber assemblages, as tuell ds indiuidual elements refeted to in text .lre labeled. Grid in .S tn squares.
son (Fig. 14.5), including articulated podal elements of two pes and
one manus. Logistics prevented further work at the quarry until 1999. Since then, work has continued, with the exception of the year 2000 field season (May-October) only.
Specimen Description The right manus (Fig. 1a.6) was discovered as an articulated group of five metacarpals (FS 45, 46,47,85,44), one phalanx (FS 42), and one carpal (FS 43). Closely associated, but not in direct articuIation, was an element tentatively identified as the digit one ungual phalanx (FS 220). Other phalanges are absent. The manus is be-
lieved mesaxonic, with metacarpal III being the longest, and III, and IV. Metacarpal II (Fig. 14.7), however, is in fact longer than metacarpal III (Table 14.1), but it is thought missing material at the distal end of metacarpal III would make it slightly longer than metacarpal II. The manus was originaily articulated but somewhat collapsed from a digitigrade living posture, though the metacarpals were still touching at both proximal and distal ends. Metacarpal V is slightly shifted from its natural position. A reconstruction of the manus with some of these distortions removed is shown in Figure 14.1. Its pose is presumed to be digitigrade because all previous analogues (Upchurch 1995, 1998) of articulated sauropod manus have been recovered in positions suggesting a semicircular, columnar arrangement of the metacarpals. As well, a recent functional morphological study of sauropod manus (Bonnan 2003) supports this. U-shaped manus trackways have also been found (Farlow et al. 1989; Lockley metacarpals V and I shorter than II,
1997\.
Metacarpal I (FS 45) is not as elongate as that of any camarasaurid metacarpals (Mclntosh, et al. 1996a; Mclntosh et al.
310 . Malcolm \7. Bedell Ir. and David L. Trexler
Fig. 14.6. WDC-FS001A manus, in situ (from phcttograpbs, uith cdsts of dctudl bones) demonstrating articulation of fiue metacarpals, anterior uiew (left to
right, V-l1, single phalanx (l-1, below and dttdched to metctcdrpal l), single carpal (aboue, and resting on metacarpals ll and I), and the closely associated, but dis ar ti cula t e d un gual p h a lanx placed near the lower right of phaldnx I-1.
.m
qffi
#w
Hffi E* X
W w
K
=*
.ffi:*i w
xsw
tr
W W
ffi
$
& 'r
Fig. 11.7. Metacarpals (\X/DCFS00 I At photographed
'ffi,'
digitally
in (A) anterior, (B) posterior, and Y lll ail r.. -ur...-*-.-l
(C) laterdl uieus; in order I-V, left to right, u'ith the only rrrorrrr, phaldnx left attached to metdcarpal l. Scale in cm.
First Articulated Manus of Diplodocus carnegii
. 3I1
Ant.
Post,
"":
&
Ycnt.
'-"w l\{etl.
'-"w Fig. 14.8. tA) Carpal (WDLFS001A) in (top to bottom) anterior, posterior, medial, and proximal uiews. (B) Manual ungual (WDC-FS})LA) in (top to bottom) dorsal, uentral, medial, and proximal uiews. (C) Metacarpals (WD C-F5001 A) I-V (top to bottom) in proximal uiews. Scale in cm. Dotted lines represent missing material.
'-'ffi l{} cm
r--I::I::r'--r1996b; Ostrom and Mclntosh 1999). Brachiosaurids (Janensch 1961; Tidwell et al. 2001) are distinctly more elongate, whereas apatosaurs were much more robust (Marsh 1883; Hatcher 1902; Ostrom and Mclntosh 7999;Fllla and Redman 1994). Diplodocus bayi (Holland, 1924) is also considerably more robust than this specimen.
The proximal end of FS 45 (Fig. 1a.8) is rugose, and slightly crescent-shaped, with a convex anterior edge, a concave interior edge, and anteroposterior elongation. Some bone is apparently miss-
ing from the posterior proximal end of metacarpal I, although not 312
.
Malcolm !7. Bedell Ir. and David L. Trexler
enough to alter this interpretation of its shape (see Fig. 14.7).Upon examination by Rothchild (pers. comm.), this was not thought to be the result of pathology. The bone may have been lost before permineralization was accomplished. There appears to be more of a "twist" between the proximal and distal ends of metacarpal I than is ordi-
narily noticed in camarasaurids (Ostrom and Mclntosh 1999). An angle of 38" is described by the intersection of the long axis of the proximal end of metacarpal I and a mediolaterai line parallel to the distal end of the bone, in proximal view. The lateral surface of the shaft of metacarpal I is fairly straight, with a slight bulge in the middle, whereas the medial surface is slightly concave. The medial surface of the shaft exhibits a central concavity, roughly one-third of the
distance from the proximal end. On the anterior surface of the distal end is a distinct articular surface for phalanx I-1, rising toward the medial side and culminating in an abrupt bulge. This articular surface angles downward and outward toward metacarpal II, forming a laterodistal process. The bulge at the apex of the anterodistal articular surface forms a ridge, rising through the middle of the anterior surface to the highest point on the crescent of the proximal end. Posteriorly, there is also a bulge at the middle of the laterodistal process, extending up the middle of this face of the bone, and terminating in a cavity formed partially by the bone missing near the proximal margin. This concavity most likely provided an attachment surface for a ligament located at the center of the metacarpal arcade. On the lower anterodistal articular surface of metacarpal I, both closely appressed to it and attached by matrix, is proximal phalanx I-1 (FS 42). This phalanx is badly distorted, and appears to have been somewhat crushed anteroposteriorly. There is an articular surface for the ungual on the anterior side of FS 42, but it is not well defined. Metacarpal II is the most robust of the group, and the longest. The shaft of this element is nearly as wide mediolaterally as it is deep anteroposteriorly throughout its length. It is believed to have been originally shorter than metacarpal III, but this cannor be veri-
fied yet. The rugose proximal surface is subtriangular in shape, with some missing one medial proximal margin. The surface of this element is damaged, and small pieces are missing from the medial side. If restored, the medial side would be an almost straight surface, with a slight middle bulge. There is a striated rugosity along the proximal upper-half of this surface reminiscent of the C. lentus holotype (YPM 1910). Laterally, metacarpal II has a smooth, nearly flat surface. One corner of the proximal surface lies lower than the others. This lower portion borders the anterodorsal border of the lateral metacarpal surface. There is also a subtle middle bulge on this lateral surface.
Most distal, anterior articular surfaces in sauropods examined are more pronounced than the posterior surfaces. The anterior articular surface here is correspondingly higher as well, but seems to form a very thin, laterodistal edge, which, in turn, extends downward and away from the center of the distal condyle. Vierved distally, metacarpal II exhibits a great deal of anteroposterior elongaFirst Articulated Manus of Diplodocus carnegii
.
3L3
tion. There are at least two marks, which appear on the lowerdistal anterior edge of metacarpal II. These marks extend across the distal surface to the medial edge, and rnay be toorh marks. In any case, they appear postdepositional. Middle lateromedial bulging of the anterior shaft surface of metacarpal II is exaggerared somewhat by bone missing at both the upper lateral and lower medial edges. The entire anterior surface, however, is concave between the proximal and distal ends. Posterior surfaces are smooth, rvith a small bulge about one-third of the distance from the proximal to the distal margins. Overall, this surface is straight, lvith a flaring at each end, viewed mediallv. toward the proximal and distal articulating surfaces. Metacarpal III, the second longest of the group, also suffers from missing bone fragments and eroded surfaces. A large quantity of bone is missing from along the upper anterior surface adjoining the anterior edge of this element. The proximal surface is rugose and subtriangular in contour. The shaft is elongare, though not nearly as much as in examined camarasaurid and brachiosaurid specimens (Mclntosh et al. 7996b; Janensch 1961; Ostrom and Mclntosh 1999; DMNH40932). Because of the missing bone, the anterior surface looks more rugose than it was origrnally. The anterodistal end exhibits a very high articular surface. This articular surface obviously provided a well-defined joint face for articulation with a phalanx. No such phalanx, however, was discovered in the quarry. The medial edge of the anterior surface is relatively straight, and the lateral edge markedly concave, whereas the lateral edge of the proximal end continues to a laterally protruding point. There is a proximal-distal groove extending across the middle of the anterior surface, from the apex of the anterodistal articular surface to the area of the missing bone. In lateral view, the edge of this groove thickens distoproximally to a point about
two-thirds of the distance from the distal end into a kind of
process. Due to missing portions of the distal and proximal ends of the shaft, an accurate determination of this feature is difficult. A
groove is also present on the posrerior shaft, bisecting a bulge rn the middle of the shaft (viewed medialiy), and with slight concavities between this bulge and both the prorimal and distal ends of the bone. Metacarpal IV has a smooth, slightly concave anrerior surface, and there is some evidence of crushing of the lower mediodistal
edge. The prorimal end is missing a significant portion of bone from its posterior side. Restoration of this area would likely show a lateromediaily elongated crescent, rvith the anterior edge describing
a concavity. As with
metacarpal
III, the proximal
surface
is
rounded on the medial side when viewed anteriorly, and becomes a laterally protruding point on its lateral edge. This area is slightly damaged. \fere the missing secrion present, this laterally protruding point would be even more pronounced. On the medial edge, a triangular piece of bone juts from the otherwise rounded surface. This fragment matches perfectly the area missing from the upper
314 . Malcolm
S7. Bedell Jr.
and David L. Trexler
lateral edge of metacarpal III. The fragment was left attached to metacarpal IV during preparation, since it provides a "key" to the original in situ articulation. The posterior surface of the shaft bulges out distinctly toward the middle, in either lateral or medial views. In these views, metacarpal IV is quite slender throughout its length. A jagged, 11-mm protuberance of bone on the anterodistal surface close to lateral edge may be a preparation relic. In general, the distal end has a narrow and subrectangular aspect (viewed distally), and the articular surface for the missing proximal phalanx is damaged.
Metacarpal
V
has an anteroposteriorly expanded proximal
surface, its greatest width exceeding that of the distal face by a ratio of 1.17 to 1.00. If a line is taken through the long axis of the proximal and distal ends, there is a "twist," forming an angle of 28". This feature becomes even more apparent in the upper medial edge of the proximal surface, where it bends toward the anterior f ace ol the bone. This anterior face is smooth and slightly concave when seen mediolaterally. There is some bone missing on the lower anteromedial face, and a very minor offset may be due to a preparation error. Posteriorly, the surface is smooth from the proximal to distal ends, with the exception of a transverse groove extending from the mediodistal to the lateroproximal corners of the bone. The proximal end is rugose, and it tapers smoothly on the posterior edge. It appears, in proximal view, as a rather thin wedge, somewhat wider laterally than medially. A distal view of metacarpal V reveals a much more robust, elliptical shape, with a smoother surface and a distinctly high articular surface on the anterior edge for the probable placement of the missing proximal phalanr. As with metacarpal II, there are grooves present that may be postdepositional in nature. The medial surface of metacarpal V is a thin edge
bending anteriorly at the proximal end. The lateral edge of metacarpal V is much thicker than the medial edge, smooth, and gradually curves toward the posterior surface.
A carpal found in articulation with the proximal ends of
metacarpals I and II (Figs. 14.2, 14.4) is rhomboidai and slightly rugose when proximally viewed, with pronounced medial tapering from a flat surface. Viewed distally, the same general shape is observed with an emarginate medial edge. This emargination is likely due to postmortem bone loss. This distal surface erhibits a more even surface than the proximal, with no obvious protuberances.
The anterior aspect is wedge-shaped, with the lateral edge wider (by approximately a 2:1 ratio) than the medial edge. There is some bone absent from this lateral edge. A thin medial surface gradually increases to the much thicker lateral surface. Proximal tapering is visible on the lateral edge. Also, the lateral edge shows grooves sim-
ilar to those described for metacarpals II and V. The only manual unguai recovered from the quarry thus far was discovered in close association rvith the other bones of the articulated manus. This ungual appears to be of the correct size and shape to belong to the manus. It is a right ungual phalanx (I-2). There is a First Articulated Manus o{ Diplodocus carnegii
.
315
sharp-edged anterior surface, lateromedial hook-shape, broader proximal surface, and, at the expanded proximal end, a rounded depression for ligament attachment to the proximal phalanx. There is bone absent from the anteroproximal and posteroproximal edges, and the entire medial surface is generally smooth. Tapering from proximal to distal edges is pronounced, and the tip of this claw core (roughly the bottom-fourth portion of the bone) is missing.
Discussion and Conclusions These bones represent the first report of articulated metacarpals from this species of dinosaur. Additionally, WDS-FS001A is identified as a subadult because the manus and all other associated skele-
tal elements recovered to date are of a size and morphology consistent with the interpretation of a single subadult animal present in the quarry. For example, the humeri and the articulated radius/ulna are only about one-half to two-thirds the size of the Diplodocus
DMNH 7494, identified
as an adult. The scapulocorato 2/s adult size, was found unfused, but articulated. This lack of fusion is an accepted character of immaturity in specimen
coid, also
Vz
sauropods.
Regarding the placement of the bones as to anrerior and poste-
rior surfaces, particularly in their natural articulation, several factors were considered. These include: (1) proximal end shapes and how they fit together because of ground distortion or missing bone and (2) shaft surfaces definitively shaped to suggest ligament attachment on a posterior surface. The obvious osteological keys were not always present. Some postmortem distortion may have occurred. Also, much bone is missing from several elements of the manus, enough to significantly distort some views (see Fig. 14.7). The placement of the articulated carpal, especially regarding the shape of the proximal end of metacarpal I, was considered important. How the ungual phalanx could properly articulate with the proximal phalanx was yet another consideration, though the ungual was only closely associated with the rest of the manus. In any instance, where the articulation was ambiguous, the in situ articulation was used
as the defining factor. Observations of unusual features include that the "twist" between the proximal and distal ends of metacarpal I is greater than is present in camarasaurids (Ostrom and Mclntosh 1999). The 28'
angle formed between the proximal and distal long axis of metacarpal V is also something unnoticed in other sauropod taxa. Marks present on metacarapals II and V, and the carpal, may be tooth marks and are likely not part of the original metacarpal structure. If so, this might explain the missing phalanges (Fiorillo 1997).It is thought that metacarpals II, III, IV and V each had a single phalanx, despite their lack of recovery in WDC-FS001, due to the presence of corresponding articular surfaces at the distal ends of each bone, as well as the prevalence of this arrangement in the Sauropoda.
316 . Malcolm W. Bedell
Tr.
and David L. Trexler
There are three accepted species of Diplodocus D. longus, D. carnegii, and D. hayi (Mclntosh 1990a). The holotype of Diplodocus longus is now thought to consist of eight caudal vertebrae with one chevron, with other bones assigned by Marsh described as questionable (Mclntosh and Carpenter 1998). Entire specimens tentatively labeled D. longus show enough differences with this holotype to possibly constitute a different species, or were tentatively attributed in the first place (Gilmore 1.932; Mclntosh and Carpenter 1998). At least one of these (DMNH 7494) is a composite skeleton. There are five skeletons and two skulls, as well as "hundreds of postcranial elements" recognized for D. carnegii (Mclntosh 1990a). As can be seen from the previous descriptions and introductory remarks, the FS specimen cannot be easily separated taxonomically from Diplodocws carnegii. Many of the \X/DCFS001 elements have no accepted D. longus counterparts, though they do closely resemble those available from D. carnegii specimens, and differ from the few available D. hayi bones. However, the taxonomic classification within the genus Diplodocus itself is problematic (Gilmore 1932; Mclntosh and Carpenter 1998; McIntosh 1990b). Unfortunately, the paucity of diagnostic material available for D. longus prevents an accurate assessment of interspecific versus intraspecific variation between the two taxa. Since the FS specimen cannot be taxonomically separated from D. carnegii, and a resolution of tbe D. carnegii-D. longus issue is beyond the scope of this paper, rve have identified the FS specimen as D. carnegii.
Acknowledgments. \7hen a project approaches its seventh yeat the difficulty in affording all deserving parties proper gratitude becomes overwhelming. From Mr. Ake Sawa of Japan, who, with his cheerful group of Japanese volunteers, courageously helped to reopen a dangerous quarry under the guidance of geologist Bill Stein, to the raft of Florida Atlantic University graduate students who labored in 105" heat for months without ever seeing a bone, dozens of people have made invaluable contributions to the
fieldwork. Here,
I would particularly like to thank the Greater
Denver Gem and Mineral Council and Joanne Passmore for financial support; the'Western Interior Paleontological Society Big Horn Basin Foundation, and the Schiele Museum of Gastonia, North Carolina, for volunteers and equipment; as well as Karen Houck, Christine Turner, and Debra Jennings for their geological acumen; and also Carla Smith, who first noticed the manus. Gratefully acknowledged for both their assistance and inspiration are Jack Mclntosh, Robert Bakker, Jim Kirkland, and Lou Taylor. Thanks to Ray \X/ilhite, Matt Bonnan, and Bruce Rothchild for their expert technical opinions; Judy Peterson for her accuracy and artistic talent; Rich Barclay for his computer-drawing skills; John Rising for excellent photography; and Ray Jones for his specially shielded scintillometer. Appreciation is due to the Denver Museum of Nature and Science for specimen access and, especially, Virginia Tidwell and KenFirst Articulated Manus of Diplodocus carnegii
.
31,7
neth Carpenter, who, along with an anonymous South American reviewer, provided extremely insightful critiques of this paper's early iterations. Houston's Museum of Natural Science was notably helpful, as was the Morrison Museum of Natural History. I would also like to thank Susan Passmore for continued advocacy of this effort over the years as well as technical expertise with the 69ures, and Burkhard Pohl, without r,vhose multitude of encouragements and support none of this could have happened. And finally, Shirley Bedell, who will never see the end result she helped nurrure. References Cited
Averianon A. O., A. V. Voronkevich, E. N. Maschenko, S. V. Keshchinskiy, and A. V. Fayngertz. 2002. A, saulopod foot fron'r the Early Cretaceous of Western Siberia, Russia. Acta Palaec,tntc,tktgica Polonica 47(11 117-124. Bjoraker, C. A., and M. T. Naus. 1996. A summary of Nlorrison Formation (Jurassic: Kimmeridgian-Tithonian) geology and paleontology with notice of a new dinosaur locality in the Bighorn Basin. In C. E. Bowen, S. C. Kirkwood, and T. S. Miller, Resources of the Bighorn
Basin, 297-426. \7,voming Geological Association Guidebook, no. 47. Casper: lfyoming Geological Association. Bonnan, M. F. 2003. The evolution of manus shape in sauropod dinosaurs: Implications for functional morphology, forelimb orientation, and phylogen.v. lourndl of Vertebrate Paleontology 23(3): 595-613. Carson, C. 1999. Stratigraphy at the Warm Springs Ranch. Master's thesis, Oklahoma Stare Universirv. Darton, N. H. 1906. Geology of the Bighorn Mountains. United States Geological Survey Professional Paper, no. 51. 'Washington, D.C.: GPO. Farlow, J.O., J.G. Pittman, and J. M. Hawthorne. 1989. Brontopodus birdi, Lower Cretaceous sauropod footprints from the U.S. Gulf Coastal Plain. In D. D. Gillette and M. G. Lockley, eds., Dinosaur Tracks and Traces, 372-394. New York: Cambridge University Press. P. D. Redman. L994. Apatosaurus yahnahpln: Preliminary description of a new species of diplodocid sauropod from the late Juras'$fyoming, sic Morrison Formation of southern the first sauropod dinosaur found with a complete set of "be1ly ribs." In Gerald E. Nelson, 'lfyoming Geological Ased., The Dinosaurs of Vlyoming, 159-178.
Fiila, J., and
sociation 44th Annual Field Conference Guidebook.
Casper:
'l7yoming Geological Association. Fiorillo, A. R. 1991. Prev bone utilization by predatory dinosaurs. Paleo-
geography, Paleoclimatolctgy, Paleoecology 88: 157-f66. C. 7. 1932. On a newl,v mounted skeleton ol Diplodoats in the
Gilmore,
United States National Museum. United States Natbnal Musettm Proceedings 81: 1-21. Hatcher, J. B. 1901. Diplodocus (Marsh): Its osteology, taxonomn and probable habits, with a restoration of the skeleton. Memoirs of the Carnegie Musettm 1(1): 7-63. 1902. Structure of the forelimb and manus of Brontosaurus. Annals of the Carnegie Musetmt 1,: 356-376.
1903a. Osteologv of Haplocanthosaurus, with description of a species, and remarks on the probable habits of the Sauropoda
318
.
Malcolm W. Bedell
Tr.
and David L. Trexler
and the age and origin of the Atlantosaurus beds. Memoirs of the Carnegie Museum 2(1): 1-72.
1903b. Additional remarks on Diplodocus. Memoirs of the Carnegie Museum 2(1): 7 3-77. Holland, \(/. J. 1906. The osteoiogy of Diplodoctts N{arsh. Memoirs of the Carnegie Museum 2: 225-264. 1924.The skull of Diplodocus. Memoirs of the Carnegie Museum
9:379403. Horn, G. H. 1963. Geolog,v of the east Thermopolis area, Hot Springs and 'lfashakie Counties, 'lfyoming. United States Geological Sun,e,v, Oil and Gas Investigations Map OM-213. Washington, D.C.: GPO. 'W. 1961. Die Gliedmassen und Gliedmassenguortel der SauJanensch, ropoden der Tendaguru-Schichton. P aleontograph ica, supp. 7, 3: 1 80235. Jennings, D. S. 2002. Detailed sedimentary and taphonomic analysis of a dinosaur quarr\., Hot Springs Ranch, !(yoming. Journal of Vertebrate Paleontology 22 (supp. to 3): 71A. Locklev. M. L. 1991. Tracking Dinctsaurs: A New Look at an Ancient World. Neu, York: Cambridge University Press. Lull, R. S. 1919. The sauropod dinosaur Barosaurus Marsh redescriptions of the t,vpe specimens in the Peabody Museum, Yale University. Memoirs of the Connecticut Academy of Arts and Sciences 6: 542 Marsh, O. C. 1883. Principal characters of American Jurassic dinosaurs. Part 6: Restoration of Brontosaurtrs. American .[ourndl of Science, ser.
3,26: 81-85. 1884. On the Diplodocidae, a new familv of Sauropoda: Principal characters of American Jurassic dinosaurs, Part 7. American Journal
of Science,
ser.
3,27: 161-167.
1896. The dinosaurs of North America. United States Geological Suruel, Annual Report 16:135-244. Mclntosh, J. S. 1990a. Sauropoda. In D. \Teishampel, P. Dodson, and H. Osm6lska, eds., The Dinosaurid, 345-401 . Berkelev and Los Angeles: Universitl' of California Press.
1990b. Species determir.ration in sauropod dinosaurs. In K. Carpenter and P. J. Currie, eds., Dinosaur Systematics: Approaches and Perspectiues, 53-69. Cambridge: Cambridge University Press. Mclntosh, J. S., \)tr P. Coombs, Jr., and D. A. Russell. f992. A new diplodocid sauropod (Dinosauria) from \ff/yoming, USA. lournal of Vertebrate Paleontology 12: 1 58-167. Mclntosh, J.S., \7.E. Miller, K. L. Stadtman, and D. D. Gillette. 1996a. The osteologl' of Camdrasaurus leuisi (Jensen, 1988). BYU Geologl'
JtttAtes+li /J-lIJ. Mclntosh, J.S., C.A. Miles, K. C. Cloward, and J. R. Parker. 1996b. A new nearl,v complete skeleton o{ Camarasdunts. Bulletin of Gunma Museum of Natural History 1: 1-87. Mclntosh, J. S., and K. Carpenter. 1998. The holotype of Diplodocus longus, with comments on other specimens of the genus. Modern Geology 23: 85-1 10. Osborn, H. F., and C. C. Mook. 7921,. Camardscturus, Amphicoelias, and other Sauropctds of Cctpe. Memoirs of the American Museum of Natural Histor1., new series, vol. 3, part 3. New York: American Museum of Natural History. Ostrom, J.H. f970. Stratigraphy and paleontology of the Cloverly Formation (Lower Cretaceous) of the Bighorn Basin area, Wyoming and
First Articulated Manus of Diplodocus carnegii
.
31.9
Montana. Peabodl, Museum of Natural History Bulletin
35
162-
164.
Ostrom, J. H., and J. S. Mclntosh. 1999. Marsh's Dinosaurs: The Collections from Como Bluff. New Haven, Conn.: Yale University Press. Tidwell, V., K. Carpenter, and S. Meyer. 2001. New Titanosauriform (Sauropoda) from the Poison Strip Member of the Cedar Mountain Formation (Lower Cretaceous), Utah. In D. H. Tanke and K. Carpenter, eds., Mesozoic Vertebrate Lfe, 139-166. Bloomington: Indiana University Press. Upchurch, P. I995. The evolutionary history of sauropod dinosaurs. p/:l/osophical Transactions of the Royal Society of London, series B, 349: 365-390. 1998. The phylogenetic relationships of sauropod dinosaurs. Zoological Journal of tbe Linnean Society 124:43-1,03.
320 . Malcolm \fl. Bedell Tr. and David L. Trexler
L
5. Evolution of the Titanosaur Metacarpus SeeesrrAN Appsrecuie
Abstract The manus in the neosauropods trended toward a tubular configuration very early in their evolution. Although basal titanosauriforms show a trend toward lengthening of the metacarpals and reduction of the phalanges, titanosaurs went further. Their metacarpais all assumed a similar shape, became long and robust, and they completely
lost their digits. The metacarpals were vertically arranged into a completely tubular structure, and were connected to each other by their proximal and distal epiphyses. This graviportal adaptation was enhanced by enlarging metacarpal V so that it was subequal to metacarpal I, and modifying the distal ends of all metacarpals so that
they were square, widened laterally, profusely pitted, and flat. In some titanosaurs, metacarpal I is bowed and retains an asymmetrical, short first phalanx and ungual, which is lost in derived species. This bowing occurs in Andesaurus and Argyrosaurus, and might be a synapomorphic character for Titanosauria, which was reversed in more derived forms. If true, the persistence of a large ungual in the digit I of some species and the bowed metacarpal could be related to each other both phylogenetically and functionally. .)-L 1
The eiongate metacarpal V faced metacarpal I on the posterior side of the manus, thereby forming a tubular arrangement. Intermetacarpal contacts developed flanges distally on metacarpals II and III. In proximal view, meracarpal I varies, from a D-shape to flat outlines; the second and third metacarpals are robust and wedge-shaped; the fourth metacarpal is biconcave and squareshaped; and the fifth metacarpal is a twisted, clepsydra-like bone
with flat epiphyses.
The titanosaur carpus resembles that of basal titanosauriforms and seems not to be homologous to that of diplodocoids. It comprises both proximal and distal elements. A number of features make titanosaurs unique among Sauropoda, including redevelopment of the olecranon, lateral expansion of the ilia. bowed opisthocoelous dorsals, loss of hyposphene-hypantrum articulations, reduction of the forelimbs, enlargement of the sacrum, and the presence of osteoderrns, and some of these features might be relatec to the nest-excavation by advanced titanosaur sauropods.
Introduction The appendicular structure of most vertebrates is different between the manus and the pes. This is particularly true of bipeds, where the manus developed specialized functions (e.g., prehensility, manipulation, digging). In quadrupeds, however, the fore- and hindlimbs are more similar. Descended from bipedal ancestors (Bakker 1977; Bonaparte 1982), quadruped dinosaurs have secondary modifications for quadrupedal locomotion, thus departing from the original archosaurian bauplan. In bipedal basal theropods and facultative bipedal prosauropods, the primitive saurischian manus has short metacarpals that are no more than 40o/" of the radius length (Sereno 1993). The manus, not significantly different in srructure from the pes, is composed of sprawling metacarpals that allowed it to resist multidirectional stress under moderate weight. In non-neosauropod sauropods, the pes and manus also are not well differentiated (Fig. 15.1C), although this has changed in neosauropods. To give context to these differences, the sauropod pes is described first. The pes apparently arose from a sprarvling prosauropod-iike autopodiai, as evidenced by tracks (Fig. 15.1A) and skeletons (Fig. 15.1B). However, the sauropod pes differs from that in prosauropods in that it is semi-plantigrade. The pes also resembled that of ertant elephants in possessing a large cushioning pad that made the pes a large, semicircular, stabilizing device (Bonnan 1999;'this volume). The ungual phalanges are twisted laterally in a manner similar to that of tortoises for traction (see Bonnan this volume). The weight-induced spread of the feet is prevented by well-developed posterior flanges on the metararsals (e.g., Patagosaurus). The astragalus is the only functional rarsal, allowing pes mobility in Apatosaurus, but it is slightly reduced in Camarasaurus (Bonnan 1999). Expanded articular facets and loose-fitting articulations show that the pes was able to walk on uneven terrain, with
322 .
Sebastidn Apesteguia
A
B
1':i'i\19-
#'
f
j; [i(r ,,ir'' " ?t' Ylt-
a\
Itlri
E
D
ii
\J
tdJ)
Jffiq
ffi,
,4-* - f-\
@
i,k f-S
1;*#fi ,S{#-
Fig. 15.1. A: Pes (top) and rnanus b ottr,tm ) of diu er s e q uadr ttp e da I tetrapods. (A) Prosaurctpod tracks (from Thulborn 1990). (B) lvlassospondylu s atrt op o dia (fr om Thulborn 1990). (C) Cetiosaurus leed:i rBMNH Rl0-81 manus in dorsal uietu. D-G: Mammalia (
H
autopodia (from Marsh 1876, 1893 ): (D) Brontotherium; /E/ Coryphodon; fF) Dinoceras; (C)
f#
K
r-*\ 1\
'd,V
Elephas. H-N: Sauropod dinosaur atiopodid. H, J, L, M:
Apatosaurus (from Gilmore 1936): (H) Pes in dorsal uieu,; 0) metdcarpals in proximal uiew; (L) tnonus ds found; (M) ftranus as reconstrLtcted with no tttbular arrdngenlent. l, K, N; Janenschia (from Bonaparte et dl.2000; Jdnensch 1961): (I) Pes in dorsdl uiew; (K) metdcdrpdls in prr:ximal uieu': (N) nldnus .ts reconstructed. Figttres were re-scaled dnd reuersed fur ease of comparison.
the claws as anti-slip devices (Bonnan 1999). The sauropod pes,
with its non-uniform metatarsal length and asymmetrical entaxonic structure, is not comparable to that of ungulate or subungulate mammals. Even in derivative sauropods, such as Titanosauriformes, the pes underwent no major changes during their phylogenetic history despite major changes in the hip, development of a wider gait (Wilson and Carrano 1999), and rotation of the tibiae (Salgado et al.7997). Evolution of the Titanosaur Metacarpus
.
323
In contrast to the pes, the manus followed a very different trend. The sauropod metacarpus looks more similar to that of the hippopotamus, rhinoceros, or Brontotherium (Fig. 15.1D) rhan it does to that of the more plantigrade elephant manus (Coombs 197 5). Only in the abbreviated phalangeal segment does the sauropod manus resemble that of the eiephant (Bakker 7971). Other differences can be found in the loss and/or fusion of carpals, which are flat and wide in sauropods, and in loss of the olecranon on the ulna so that the forelimb is more graviportal. Because the radius and ulna are interlocked proximally in sauropods, pronarion of the manus was achieved by the medial position of the radius with respect to the ulna (Huene 7929; Bonnan 1999) and may have led to the development of the tubular arrangement of the manus (Bonnan 2003). Early in sauropod evolution, the radius migrated medially to form part of a columnar forearm, differing from the crossed radius and ulna seen in other tetrapods (Hatcher 1902). Because digit I migrated with the radius, the sauropod manus acquired a unique U-shape in cross-section (Bonnan 7999), formed by long metacarpals rearranged vertically (Figs. 15.1J, K; 15.3). The analysis of Bonnan (2003) concluded that several features of the forelimb considered independently in phylogenetic analyses constitute an integrated functional suite, which probably evolved together; and the presence of one feature strongly suggests the presence of others.
These features include obligatory quadrupedal posture with columnar limbs (Wilson and Sereno 1998); proximal end of the ulna tri-radiate, with deep radial fossa (\Tilson and Sereno 1998); distal condyle of radius subrectangular, with a flat posterior margin for ulna (\X/iison and Sereno 1998); and the proximai end of the metacarpals are subtriangular, so that the articulated manus is Ushaped in proximal view (Mclntosh 1990; Upchurch 7995,1998;
\Tilson and Sereno 1998). Although Bonnan (2003) may be correct in a broad view, the small individual changes in the forelimb anatomy and their functional consequences resulted in important differences along the sauropod lineages. These differences will be analyzed here within the context of the metacarpus, with oniy occasional mention of the implication to the whole forelimb. I will show the distinct features present in the titanosaur manus, and consider the carpus, metacarpus, and phalanges in a phylogenetic context. The absence of manual phalanges in titanosaurs, mainly based on negative evidence, is discussed, as weli as the possible of claws. The relationships of these features and the acquirement of graviportalism and large body size are evaluated.
presence
Institutional abbreuiations. MACN-Museo Argentino
de
Ciencias Naturales "Bernardino Rivadavia." Buenos Aires. Argentina; MLP-Museo de La Plata, La Plata, Argentina; MPCAMuseo Provincial "Carlos Ameghino," Cipolletti, Argentina;
MPCF-Museo Provincial "Carmen Funes." Plaza Huincul. Neuqu6n, Argentina; MUCPV-Museo de la Universidad del Co-
mahue, Neuqu6n, Argentina; UNP-Universidad Nacionai de la
324
.
Sebasti6n Apesteguia
Patagonia "San Juan Bosco," Comodoro Rivadavia, Chubut, Argentrna.
Materials and Methods Known articulated or closely associated metacarpi include part of the right and left metacarpi of Chubutisaurus insignls (MACN 18222), the left manus of Argyrosaurus superbzzs (MLP 77-V-291), the right of Antarctosdurus tuichmannianus (MACN 6804a), the left of Laplatasaurus (MACN 6804b), the right and left of Epachthosaurzs (UNP-PV-920), the right of Aeolosaurus rionegrinzs (MJG-R1), and an isolated but articulated right titanosaur metacarpus (MPCA 110-51). Poorly preserved but associated metacarpals include the first and fifth metacarpals of Andesaurus delgadoi (MUCPV-132), the second and third metacarpals of a new basal titanosaur species from the Aptian of Neuqu6n (MCF-PVPH-
233), a poorly preserved metacarpus from the Turonian of Neuqu6n (MCF-PVPH-638), and the single metacarpal of Aeolosaurus sp. (MPCA-27774). Materials assigned by Huene to Neuqttensaurus australis also consisted of isolated elements. For outgroup comparison, examples from the literature r,vere used, including Cetiosaurus /eedsl (BMNH R3078), diplodocoids (e.g., Ap(ltosaLrrus excelsus, CM 563), basal macronarians (i.e., CantaraslLrrus supremus, AMNH 711) and titanosauriformes (e.g., Brachiosaurws, Venenosdurus, Atlasaurus and Alamosaurus). The phylogenetic framework used is based on Salgado et a|. (7997), Sfilson and Sereno (1998), 'Wilson (2002), Wilson and Upchurch (2003), and Salgado (2003).
Considering that every metacarpal has a different position within the metacarpus, and in order to follow a consistent description, the metacarpals are considered to be parallel. Thus, for a given metacarpal, the proximal and distal ends refer to the ends that face the carpal and phalanges respectively, even when the latter elements are absent. The anterior and posterior sides in the tubular structure of the sauropod manus are equivalent to the outer and inner (palmar) sides. The terms "medial" and "iateral" are considered in the same way and refer to the complete forelimb position. No comparative methods were used, such as the Extant Phylogenetic Bracket; that will be done somewhere else. Reconstructions of the metacarpal structure were made from my own observations. Titanosaur Manus Carpus. Because of its block shape (Wilson and Sereno 1,998, ch. 42) and reduced number (Mclntosh 1990; Upchurch 1998; \7ilson and Sereno 1998, ch. 79), the identification of the carpal bones in sauropods is problematic. The studies of Osborn (1904) and
Hatcher 11902) show the carpals are not always homologous. In Apatosaurus, the carpal has two proximal facets for the radius and ulna, and it is only loosely placed above metacarpals II-IV (Osborn 1904; Fig. 15.2A). In Camarasaurus, on the other hand, the larger Evolution of
tl-re
Titanosaur Metacarous
.
325
-U
H ffir$ LJ
"_u Fig. 15.2. Sauropod left wrist in dorsal uiew. /A/ Apatosaurus excelsus (CM 563). (B) Camarasaurus supremus (AMNH 711). (C) Atlasaurr-rs imelakei. /D) Argyrosaurus superbus (MLP 77V-29-1). Not to scdle.
{rl
frdnn\\
!
U
frdnnt1
of two carpals fit closely to metacarpals I and II. This placement identifies it as a distal carpal. However, the smaller carpal between the ulna and metacarpal V makes its identification problematic (Fig. 15.2B). Other macronarians with carpals are the basal titanosauriform Atlasaurus (Monbaron et al. 7999) and the advanced titanosaur Argyrosaurws (Lydekker 1893). In the former (Fig. 15.2C), as in Camarasauras, there are also two carpals, and the position of the larger carpai over metacarpals I and II is similar to that of Camardsaurus (actualln there is a "step" in the carpal because metapodials range in height). The smaller carpal is closely associated to metacarpal III, thus is different from that of Camdrdslurus. Two carpals were described in titanosaurs early in the twentieth century (Huene 7929), but successive authors have ignored this information and proposed that "no carpals are known in titanosaurs, and it is possibie that they were not ossified" (Borsuk-Bialynicka 1977); "titanosaurs appear to have eliminated any ossification of the carpus, as evidenced by the lack of carpals in
that preserve all of the other bones of the forelimb" (Wilson and Sereno 1998). Two carpals are present in the well-preserved left forelimb of
specimens
Argyrosdurus (Fig. 15.2D). Unfortunateln the humerus and metacarpals remained in situ, but the ulna, radius, and carpals twisted prior to burial. The smaller carpal (the ulnar of Huene 7929) rs still attached to the distal end of the ulna and was probably associared with metacarpal V. The larger carpal (composite radial of Huene 1929) is attached to the proximal ends of 326
.
Sebasti6n Apesteguia
III-V, but this was the result of the twisting of the forearm, and should be associated with metacarpals I and II or I-III, as well as the radius. Porvell (2003) mentioned the existence of a carpal in the distal ulnar side of the radius, but I do not know if he is referring to a previously unidentified carpal. The close association between the smallest carpal and the ulna suggest that this carpal is the ulnar, a proximal carpal, whereas the larger carpal, closely associated to the metacarpals, represents the fusion of smaller, distal carpals. Tubular mlnus and metacarpal enlargement. The metacarpals of most sauropods are arranged in a half-cylinder (Fig. 15.3B-L) or 270" arc ('Sfilson and Sereno 1998). This U-shape was acquired early in the sauropod evolution and is related to weight support metacarpals
(Christiansen 1997). This unique organization prevents the metacarpals from splaying apart distally and is a graviportal adaptation that is well documented in the ichnological record as an
open half-moon track (Fig. 15.3A; Lockley 1991). However, two additional and probably related changes took place: lengthening of the manus and loss of the digits. In the early sauropod Vtrlcanodon, the longest metacarpal is about 32'/. the length of the radius, and this low value is also seen in the Dipiodocoidea (Gilmore 1936). The manus, already differentiated from the pes, was too short to be considered as a separate segment in the forelimb, but served as a wide, supporting base for the forelimb. Within the Macronaria, the metacarpals are twice the length of the metatarsals and are almost half the length of the radius (47% in Camarasaurus). Macronaria metacarpals are long and robust (Fig. 15.3C), but they do not achieve the substantial enlargement seen in titanosaurs, nor is there
a reduction in the digits. Titanosauriformes exhibit the
largest
metacarpal:radius ratio (see Table 15.1). Furthermore, most of the basal titanosauriform metacarpals not only are proportionally long and slender but also are not yet as robust as those of advanced titanosaurs (compare Figs. 15.3D, E, F, and 15.3I-K). The small, complete metacarpus MACN 6804b of Laplatasaurus (Fig.15.4G) is formed by five remarkably slender metacarpals that are heterogenous. The metacarpals are as slender as those of Bracbiosaurus, AtIasaurus, and other basal titanosauriforms (Fig. 15.3D, E, F), suggesting that this taxon could be a relatively basal titanosaur despite its young stratigraphic position (Lower Campanian). In most of the crown-group Titanosauria, the metacarpals are very long and secondarily robust (Fig. 15.3J, K, L, M), except in the very derived saltasaurines and Opisthocoelicaudia, where the metacarpals apparently became secondarily short (Fig. 15.3J, K). Oueruiew of the manus. Some degree of tubular arrangement of the metacarpal was already present in early eusauropods (e.g., Omeisaurus) as an open arc (Bonnan 2003). Some neosauropod groups still retain this primitive open arc, as evidenced by an
unidentified Late Cretaceous ichnotaxon having a 1:3 to 1:4 heteropody (Fig. 15.3A). This track may have been made by a basal diplodocoid \e.g., "Rebbachisaurus"). Other neosauropods show F,volution of the Titanosaur Metacarpus
.
327
Ea&
F ('l' ' _)'
'/ "t tt",
c ,,,i.r-, ''
ulti
''.
\
\.,
m \
/
\/K
W
\..,
\
\,/' \
,/ cor*'l
./ \/ \/
,_/\ B ',o1 --rLV
.tH 'iil ,4f ..-i.-l.-lL/ \/
\
ced\
-)'
,i
J-llrt
\
"- !' '.'i-
fi".,'-7 'l =i ' ttl f|rll ,li' { t'l " ''
fli
,1 1*;8
i
H''..',..-,.-,:,
Y
"=
t) r---', p4+q
tv
mAl
/
utu
ventral
rfro Ur--r n \)
328 . Seha'ri;in
Apesteguia
25 cm) (Dalla Vecchia 1998b). Despite its small size, the bone sample shows a diversified association of smali to large sauropods and small to (possibly) large theropods. The large vertebra of Fig. 18.8 beiongs to an individual at least 17-18 m long, fol-
420 . Fabio M. Dalla
Vecchia
Iowing sauropodan length estimations based on single bones in Naish and Martill (2001). Aiso, the Haurerivian-early Barremian footprint sample from the Apulia platform, berween ihe Adriaric-
Dinaric Platform and Afro-Arabia (Fig. 18.3A), is represented mainly by large tridactyl footprints (prevailing footprint lengths are 30-35 cm), mostly theropodan (Gianolla et al. 2000).
In conrrast, all the late Albian and late Cenomanian sauropods, identified by footprints from different horizons and localities, are rather small for sauropod standards. The largest manual print in the Albian and cenomanian sample of Istria is smalrer than the manual print from the upper Hauterivian-lo,uver Barremian of northeast Italy. Although other ichno-associations in the world present small sauropodan footprints, they are always associated rvith prints of large individuals. The Albian-Cenomanian tridactyl footprint record of Istria shows the same pamern of size reduction, with most footprint lengths around 18-20 cm, and verv few specimens longer than 25 cm (Dalla Vecchia and Tarlao 2000). Also the single "iguanodontian" trackway, from the late Albian of
Istria, shows footprints only 28 cm long (Dalla Vecchia et al. 2002). The hadrosaurians from the upper Santonian of Italy are only half the length of hadrosaurids from North America and Asia and, since body mass is proportional to the cube of total length, they were much smaller. Teerh from rhe Maasrrichrian of th. Slovenian Kras are all of small size. Furthermore, rhe footprints of quadrupedal dinosaurs in the upper Coniacian-lower Santonian of
the Apulia Platform are also small.
Dalla Vecchia and Tarlao (2000) and Dalla Vecchia (20021 noted that this reduction of dinosaur size in the peri-Adriatic platforms occurred between the late Barremian and the beginning of the late Albian, coinciding with the Aptian opening of the Eastern Mediterranean ocean between the northern Afroarabian shelf anc the peri-Adriatic platforms (Fig. 18.38). This isolation and the reduction of the emergenr surface in the platform (Adriatic Island), because of eustasy or tectonics, might have resulted in dwarfism in the dinosaur population. There is a graded trend from giantism in the smaller species to dwarfism in the larger species in living insular mammals G.g., Lomolino 1985). Large herbivores characteristically are smaller than their relatives on the mainland (e.g., Roth 1990). This is observed for Pleistocene mammal faunas (e.g., Kotsakis 1985; Roth 1990). The insular faunas are also characterized by lower species diversity rhan the main[and, rerenrion of primitive features, fewer species of carnivores, and the absence or extreme rariry of iarge piedators (e.g., Kotsakis 1985; Roth 1990; Burness et aI.2001). The main cause of dwarfism in large herbivores seems to be the resource lim,
itations of an island (e.g., Lomolino 1985; Roth 1990). Also, the advantages of large size as defense against predation do not exist where predators are absent or small (Lomolino 1985). The concept of dwarf dinosaurs dares back to 1912 (Nopcsa 1974). The titanosaurian MagyarosctLtrus, the hadrosaurian Tel-
Between Gondwana and Laurasia
.
421
mAtosaurus, and a nodosaurid ankylosaurian (" StrwthiosAurus") from the Maastrichtian of the Hateg Basin, Transylvania, Romania, have been considered insular dwarfs because of their diminu'Weishampel
et al. tive size and primitivity (Nopcsa 1974, 1.91.5; 799t, 1993;Jianu and Weishampel 7999 Pereda Superbiola 1999; Mussel and \Teishampel 2000). Humeri attributed to Magyarosaurus are only 37-40 cm long (Huene 1932) and all the bones recently excavated near Sanpetru by the Museum of Deva belong to very small, adult individuals (pers. obs.).Telmatosaurus was around 5 meters long (more or less the same as the late Santonian hadrosaurians from Italy) and its estimated weight is 500 kg, about 10% the body mass of the Campanian-Maastrichtian 'Weishampel et al. hadrosaurids of North America according to (1991). Like the emergent parts of the Adriatic-Dinaric Platform and the Apulia Platform, Transylvania was part of an island of the Late Cretaceous European archipelago (Sfeishampel et al. 1991; Philip et al. 2000b; Dalla Vecchia 2002). Other hadrosaurians, ankylosaurians and sauropods from that archipelago, found in Austria, Germann Netherlands, Belgium, France, and Spain, are unusually small and have been considered as possible insular dwarfs (Wellnhofer 1994; Pereda Superbiola et aI. L995; Pereda Superbiola 7999 Dalla Vecchia 2003). Thus, the dinosaurs of the Adriatic Isiand show just the same trend as those of the other islands of the archipelago. Acknowledgrnents. I am grateful to Dario Boscarolli, Alceo Tarlao, Maurizio Tentor, Giorgio Tunis, and Sandro Venturini, with whom I have worked on the dinosaurs of Istria since 1993 and whose contribution has been fundamental to the discovery and study of the specimens mentioned in this paper. I thank Jos6 Bonaparte, Rub6n Martinez, Jorge Calvo, Leonardo Salgado, SebastiSn Apesteguia, and Rodolfo Coria for support and access to the collections under their care in Argentina; Alexander Kellner and Diogenes Campos for support in Brazil; Alexander Kellner for help at the American Museum of Natural History of New York; and Mary Dawson for her support at the Carnegie Museum, Pittsburgh. I thank Jim Fariow for useful advice and also the reviewers of the manuscript, Kenneth Carpenter and Virginia Tidwell. Part of this work is the result of my post-doctoral project at the Dipartimento di Geologia, Paleontologia e Geofisica, University of Padua during the years 1995-1997. References Cited
Astibia, H., E. Buffetaut, A. D. Buscalioni, H. Cappetta, C. Corrall, R. Estes, F. Garcia-Garmilla, J. J. Jaeger, E. Jimenez-Fuente, J. Le Loeuff,
M. Mazin, X. Orue-Etxerbarria, J. Pereda-Superbiola, J. E. Powell, J. C. Rage, J. Rodriguez-Lazaro, J. L. Sanz, and H. Tong. 1990. The fossil vertebrates from Lano (Basque Country', Spain): New evidence on the composition and affinities of the Late Cretaceous continental faunas of Europe. Terra Noua 2: 460-466. J.
422 . Fabio M. Dalla
Vecchia
Barrett, P. M., Y. Hasegarva, M. Manabe, S. Isaji, and H. Matsuoka. 2002. Sauropod dinosaurs from rhe Lower Cretaceous of eastern Asia: Tax-
onomic and biogeographical implications. palaeontologl,
45(61:
1197-1217. Benton, M.J., S.Bouazrz, E. Buffetaut, D. tr{artill, M. Ouaja, NL Soussi, and C. Trueman. 2000. Dinosaurs and other fossil verrebrares fronr fluvial deposits in the Lower Cretaceous of southern Tunisia. palaectge o gr dp h
y,
P
ala
e
o
cli matol
ct
gy,
P a Ia e o e c o Io
g,
15
7:
227 -24 6.
Blows, 'W. T. 1995. The Early Cretaceous brachiosaurid dinosaur Ornithopsis and Eucamerotus from the Isle of \7ight, England. paleontology 38(1):187-197. Bonaparte, I.F. 1999. Evoluci6n de las v6rtebras presacras en Sauropodomorpha. Ameghiniana 36: 1 15-1 87. Bonaparte, J. F., \X/.-D. Heinrich, and R. \7ild. 2000. Revieu. of Janenschia \fild, with the description of a nerv sauropod from the Tendaguru beds of Tanzania and a discussion on the systematic value of procoelus caudal vertebrae in the sauropoda. Paldeontographica Abt. A
256:25-76.
Boscarolli, D., and F. M. Dalla Vecchia. 1999. The Upper HauterivianLower Barremian dinosaur site of Bale/Valle (SW Istria, Croatia). Ndtura Nascosta 18: 1-5. Boscarolii, D., M. Laplocrna, M. Tentor, G. Tunis, and S. Venturini. 1993, Prima segnalazione di resti di dinosauro nei calcari hauteriviani di piattaforma dell' Istria meridionale (Croaria). Natura Nascos/a 7: 1-20. Britt, B. l.997.Postcranial pneumaticitl'. In P.J. Cunie and K. padian, eds., Encyclopedia of Dinosaurs, pp. 590-593. New York; Academic press. Britt, B., and K. L. Stadtman. I996.The Early Cretaceous Dalton Wells di. nosaur fauna and the earliest North American titanosaurid sauropod. lournal ctf Vertebrate Paleontology 16 (supp. to 3): 24A. Britt, B., K. L. Stadtman, R. D. Scheetz, and J. S. Mclntosh. 1997. Camarasaurids and titanosaurid sauropods from the Early Cretaceous Dalton \7ells Quarry (Cedar Mountain Fomration). Jc.turnal of Vertebrate Paleontology 17 (supp. to 3): 34A. Britt, B., R. D. Scheetz, J. S. Mclntosh, and K. L. Stadtman. 1998. Osteological characters of an Earli' Cretaceous ritanosaurid sauropod from the Cedar Mountain Formarion of Utah. Journal of Vertebrate paleontology 18 (supp. to 3): 29A. Bulot, L. G., E. O. Amon, F. Bergerat, E. Baraboshkin, R. Guiraud, F. Hirsh, J. P. Platel, A. Poisson, M. Sandulescu, M. Ziegler, D. Vaslet, S. Bouazrz, and G. Rimmele.2000. Map 12.-Early Hauterivian (123 to 121 Ma). InJ. Dercourt, M. Gaetani, B. Vrielynck, E. Barrier, B. BiiuDuval, M. F. Bruner, J. P. Cadet, S. Crasquin, and M. Sanduiescu, eds.,
Atlas Peri-Tetbys, Palaeogeogrdphicdl Maps. paris: CCGM/CGMI7. Burness, G. P., J. Diamond, and T. Flannery. 2001. Dinosaurs, dragons, and dwarfs: The evolution of maximal body size. proceedings of the National Academy of Sciences 98: 14518-14523. Cati, A., D. Sartorio, and S. Venturini. 1989. Carbonate platforms in the subsurface of the Northern Adriatic Area. Memorie della Societit Geologica Italidnd 40: 295-308. Cherchi, A., and R. Schroeder. 1973. Sur la biogeographie de l,association i, Valserina du Barr6mien et la rotation de la Sardaigne. Compte Rendues de I'Academie de Sciences, Pdris 277D:829-832. N. 1990. Sequence strarigraphy and sea-level changes in Cretaceous times. In R. N. Ginsburg and B. Beaudoin, eds., Creta-
Christie-Blick,
Between Gondwana and Laurasia
.
423
l-27. NATO ASI Series, no. C 304. Dordrecht: Kluwer. Dalla Vecchia, F. M. 1998a. Remains of Sauropoda (Reptilia, Saurischia) in tl-re Lower Cretaceous Upper Hauterivian/Lower Barremian) limestones of SSf Istria (Croatia). Geologia Croatica 51(2): 105-134. 1998b. Theropod footprints in the Cretaceous Adriatic-Dinaric carbonate platform (Italy and Croatia). In B. Perez-Moreno, T. Holtz, J. L. Sanz, and J. J. Moratalla, eds., Aspecls of Tberopctd Paleobiology, 355-367. Gdia, no. 15. Lisbon: Museu Nacional de Hist6ria Natural. Universidade de Lisboa. 1999a. A sauropod footprint in a limestone block from the Lower Cretaceous of northeastern Italy. Ichnos 6(4): 269-27 5. 1999b. Atlas of the sauropod bones from the upper Hauterivceous Resources, Euents and Rhythms,
ian-lower Barremian N.7scosf,7 18:
of
Valle/Bale (SW Istria, Croatia). Natura
6-41.
2001a. A vertebra of a large sauropod dinosaur from the Lower Cretaceous of Istria (Croatia). Natura Ndscosta 22: 1'4-33. 2001b. Terrestrial ecosystems on the Mesozoic peri-Adriatic carbonate platforms: The vertebrate evidence. In H6ctor A Leanza, ed', VII International Sl,mplsium on Mesozoic Terrestrial Ecosltstems,
Buenos Aires, 77-83. Asociaci6n Paleontol6gica Argentina, Publicaci6n especial, no. 7. 2002. Cretaceous dinosaurs in the Adriatic-Dinaric carbonate platform (Italy and Croatia): Paleoenvironmental implications and paleogeographical h,vpotheses. Memorie della Societd Geologica ltdl-
iana 57 89-100. 2003. I dinosauri nani dell'Arcipelago europeo. Le Scienze 423: 86-94. Dalla Vecchia, F.M., ar.rd S. Venturini. 1995. A theropod (Reptilia, Dinosauria) footprint on a block of Cretaceous limestone at the pier of Porto Corsini (Ravenna, kaly). Riuista ltdlidnd di Paleontologia e Stratigrafia 101 ( 1 ): 93-98. Dalla Vecchia. F. M.. and A. Tarlao. 2000. New dinosaur track sites in the Albian (Early Cretaceous) of the Istrian peninsula (Croatia). Part IIPaleontology. In F. M. Dalla Vecchia, A. Tarlao, G. Tunis, and S. Venturini, Nezr Dinosaur Track Sites in the Albian (Early Cretaceous) of the Istrian Peninsula (Croatia), 227-292. Estratto da memorie di Scienze Geologiche di Padova 52, no. 2. Padova: Universiti di Padova. Dalla Vecchia, F.M., A. Tarlao, G. Tunis, and S. Venturini. 2000a. Nea; Dinosaur Track Sites in the Albian (Early Cretaceous) of the Istrian Peninsula (Croatia). Estratto da memorie di Scienze Geologiche di Padova 52, no. 2. Pador.a: Universiti di Padova. Dalla Vecchia, F. M., A. Tarlao, G. Tunis, M. Tentor, and S. Venturini. 2000b. First record of Hauterivian dinosaur footprints in southern Istria (Croatia). In I. Vlahovic and R. Biondic, eds., Proceedings of the Second Croatian Geological Congress, Zagreb, Croatia, 143-149. Dalla Vecchia, F.M., I. Vlahovic, L. Posocco, A. Tarlao, and M. Tentor. 2002. Lare Barremian and Late Albian (Early Cretaceous) dinosaur track sites in the Main Brioni/Brijun Island (S\( Istria, Croatia). Natura Nascosttt 25: 1-36. Dalla Vecchia, F. M., G. Tunis, S. Venturini, and A. Tarlao. 2001. Dinosaur track sites in the upper Cenomanian (Late Cretaceous) of the Istrian peninsula (Croatia) . Bollettino della Societd Paleontologica Itdliand 40(1\:25-54.
424 . Fabio M. Dalla Vecchia
D'Argenio, B. I974. Le piattafor.me carbonatiche periadriatiche. Una rassegna di problemi nel quadro geodinamico mesozoico dell'area mediterranea. Memorie della Societa Geologica ltdlitlnd j,3(2)l 137-159.
de Lapparent, A. F. 1960. Les dinosauriens du "Continental Intercalaire" du Sahara central. Mdnrcires de la Socidtd Gictlogique de France (n.s.) 88A: 1-57.
Dercourt, J., L.E. Ricou, and B. Vrielynck, eds. '1993. Atlas Tethys P alae
oenuir ctnm ental Map s.
Dercourt, J.,
P
aris: Gauthier-Villars.
M. Gaetani, B. Vriel,vnck, E. Barrier, B. Biju-Duval, M.
F.
Brunet, J. P. Cadet, S. Crasquin, and M. Sandulescu, eds. 2000. Atlas P eriTethys: Palaeogeographical Maps. Paris: CCGN{/CGMW. Dini, M., G. Tunis, and S. Venturini. 1998. Continental. brackish and marine carbonates from the Lower Cretaceous of Kolone-Barbariga (Istria, Croatia: stratigraph)', sedimentology and geochemistry. palaeoge o gr dp h y, P alae o climatolo g1,, P a I ae o e colo gy 1 40 : 24 5 -269. Eberli, G. P., D. Bernoulli, D. Sanders, and A. Vecsei. 1993. From aggrada-
tion to progradation: The Maiella Platform-Abruzzi, Italy. In
T.
Simo, R. W. Scott, and J. P. Masse, eds., Cretaceous Carbonate Plat-
of Petroleum Geologists Memoir, no. 56. Tulsa, Okla.: American Association of Petroleum Gefctnns, 213-232. American Association ologists.
Farlow, J. O. 1992. Sauropod rracks and trackmakers: Integrating the ichnological and skeletal records. Zubia 1.0:89-138. Farlow, J.O., J G. Pittman, and J. Nt. Hawthorne. 1989. Brontctpodus birdii, Lower Cretaceous sauropod footprints from the U.S. Gulf Coastal Plain. In D. D. Gillette and M. G. Lockle.t-, eds., Dinosattr Tracks and Traces,135-153. Cambridge: Cambridge Universitl' Press" Fourcade, E., J. Azema, F. Cecca, J. Dercourt, B. Vrielvnck, Y. Bellion, M. Sandulescu, and L.-E. Ricou. 1993. Late Tithonian palaeoenvironments (138 to 135 \,{a). InJ. Dercourt, L. E. Ricou and B. Vrielvnck, eds., Arlas Tethl,s: Palaeoenuironmental Maps. Rueil-Malmaison: BEICIP-FRANLAB. Gianolla, P., M. D. Morsilli, F. M. Dalla Vecchia, A. Bosellini, and A. Russo. 2000. Impronte di dinosauri in facies di piattaforma interna nel Cretaceo inferiore del Gargano (Puglia, Italia meridionale). In Riassunti delle comunicazioni orali e dei posters-S0" Riunione Estiua
S.G.I.:265-266. Gimenez, O. 1992. Estudio preliminar del miembro anterior de los saur6podos titanosdurios. Ameghiniand 30: 154. Haq, B. U., H. Hardenbol, and P. R. Vail. 1987. Chronology of fluctuating sea-level since the Triassic. Science 235 1156-1,1,67. Huene, F. F. 1932. Die fossile Reptil-Ordnung Saurischia, ihre Entr,vicklung und Geschichte. Mr.tnographien zur Geologie und Palaeontologie 4:
1-361.
Hulke, J. Y/. 1879. Note (3rd) on (Eucamerorzs Hulke) Ornithopsis H. G. Seeley = Bothriospondl'/r.rs Owen = Chondrctsteosdurus magnus Owen. Quarterly .lournal of the Geological Societl, of London 35: 7 52-762. Janensch, \(/. 1950. Die Wirbelsdule von Brachiosdurus brancai. Palaeontographica, supp. 7, 3:27-93. Jaquin, T., and P.-C. De Gracianskv. 1998. Ntajor rransgressive-regressive cycles: The stratigraphic signature of European basin development. In P.-C. De Graciansk_v, J. Hardenbol, T. Iaquin, and P. R. Vail, eds.,
Between Gondwana and Laurasia
.
42.5
Mesozoic and Cenozoic sequence stratigraphy of European Basins. Society of Economic Paleontologl, and Mineralctgy, Special Publica-
tion 60:15-29. Jianu, C.-M., and D. B. \Teishampel. 1.999. The smallest of the largest: a new look at possible dwarfing in sauropod dinosaurs. Geologie en Miinbouw 78: 335-343. Kotsakis, A. 1985. Vertebrati insulari e paleogeografia: alcuni esempi. Bollettino della Societit Paleontologica Italiana 24(2-31: 225-244. Langston, W., lr. 1974. Nonmammalian Comancl-rean tetrapods. Geoscience and Man 8:77-102. Lockley, M. G., A. P. Hunt, and S. G. Lucas. 1994. The distribution of sauropod tracks and trackmakers. In M. G. Locklel', V. F. dos Santos, C. A. Meyer, and A. P. Hunt, eds., Aspects of Sauropod Paleobiology, 233-248. Gaia, no.10. Lisbon: Museu Nacional de Hist6ria Natural, Lomolino, M. V. 1985. Body size of mammals on islandsl the island rule reexamined. The American Naturalist 125: 310-316. Marsh, O. C. 1888. Norice of a new genus of Sauropoda and other new dinosaurs from the Potomac Formation. American Journal of Science, series 3, 35:89-92. Martill, D. M., and. D. Naish. 2001.The global significance of the Isle of \7ight dinosaurs. In D. M. Martill and D. Naish, eds., Dinosaurs of the Isle of \Yight, 44-48. The Palaeontological Association Field Guide to Fossl/s, no. 10. London: Palaeontological Association. Masse, J.-P. 2000. Early Aptian (II2-114 Ma). In S. Crasquin, Atlas PeriTethys: Pttlaeogeographical Maps-Exl1lanatoyy No/es, 119-127. Paris: CCGMiCGM\(/. Masse, J.-P., S. Bouaziz, E. O. Amon, E. Baraboshkin, R. Tarkowski, F. Bergerat, M. Sandulescu, J. P. Platel, J. Canerot, R. Guiraud, A. Poisson, M. Ziegler, and G. Rimmele. 2000. Early Aptian (112-1 14 Ma). In J. Dercourt, M. Gaetani, B. Vrielynck, E. Barrier, B. Biju-Duval,
M.
F. Brunet, J. P. Cadet, S. Crasquin and
M. Sandulescu, eds., Atlas
Peri-Tethys: Palaeogeographical Maps, M"p 13. Paris: CCGM/ CGM\(/. Mclntosh, J. S. 1990a. Sauropoda. In D. B. rWeishampel, P. Dodson, and H. Osm6lska , eds., The Dinosaurid, 345-401. Berkelev: University of California Press. 1990b. Species determination in sauropod dinosaurs with tentative suggestions for their classification. In K. Carpenter and P. J. Currie, eds., Dinosaur Systematics: Perspectiues and Approaches, 53-77. Cambridge: Cambridge University Press. Mezga, A., and Z. Bajraktarevic. 1999. Cenomanian dinosaur tracks on the islet of Fenoliga in southern Istria, Croatia. Cretdceous Research
20:735-746. Mussell, J., and D. B. I(/eishampel. 2000. Magyarosaurzs and the possible role of progenesis in sauropod development. Journal of Vertebrate Paleontology 20 (supp. to 3): 59A-60A. Naish, D., and D. M. Martill. 2001. Saurischian dinosaurs. 1: Sauropods. In D. M. Martill and D. Naish, eds., Dinosaurs of the Isle of \Yight,
185-241. The Palaeontological Association Field Guide to Fossils no. 10. London: Palaeontological Association. Nicosia, U., M. Marino, N. Mariotti, C. Muraro, S. Panigutti, F. M. Petti, and E. Sacchi. 1999. The Late Cretaceous dinosaur tracksite near Altamura (Bari, southern Italv). ll-Apulosauripus federicidnus new ichnogen. and new ichnosp. Geologica Romana 35: 237-247.
426 . Fabio M. Dalla
Vecchia
Nopcsa, F. 1914. Uber das Vorkommen der Dinosaurier in Siebenbiirgen. Vehrssammlung Zoologischen Botanischen Gesellschaft Vlien 54:
12-14. 1915. Die Dinosaurier der siebenbiirgischen Landesteile Ungarns. Mitteilungen aus dem Jahrbucb der IJngarischen Geologischen Reichsdnstah 23: 1,-24. Norman, D. B. 1985. The Illustrated Encyclopedia of Dinosaars. New York: Crescent Books. Owen, R. 1875. Monographs on the Fossil Reptilia of the Mesozoic Formations. Part 2, Bothriospondylus, Cetiosaurus, Omosaurus. Palaeontographical Societ,v Monograph, no. 29. London: Palaeontographical Society. 1876. A Monograph on tbe Fossil Reptilia of the Mesozoic Formations. Supplement 7: Crocodilia (Poikiloplew.on) and Dinosauria? (Chondrosteosattrus). Palaeontographical Societ,v Monograph, no. 30. London: Palaeontographical Society. Pereda Superbiola, X. 7999. Ankylosaurian dinosaur remains from the Upper Cretaceous of Laflo (Iberian Peninsula). Estudios del Museo de Ciencias Naturales de Aldua 14 273-288. Pereda Superbiola, X., H. Astibia, and E. Buffetaut. 1,995. Neu'remains of the armoured dinosaur Struthiosaurus from the late Cretaceous of the Iberian peninsula (Laio localitl., Basque-Cantabric basin). Bulletin de ld Socidtd Gdctlogique de France 1.66(2t:207-211. Philip, J., J.-F. Babinot, G. Tronchetti, E. Fourcade, J. Azema, R. Guiraud, Y. Bellion, L.-E. Ricou, B. Vrielvnck, J. Boulin, J.-J. Cornee, and J.-P. Herbin. 1993. Late Cenomanian palaeoenvironments (94 to 92 Mal. In J. Dercourt, L. E. Ricou and B. Vrielynck, eds., Atlas Tethys: Palaeoenuironmental Maps. Rueil-Malmaison: BEICIP-FRANLAB. Philip, J., M. Floquet, J. P. Platel, F. Bergerat, M. Sandulescu, E. Bara-
boshkin, E. O. Amon, R. Guiraud, D. Vaslet, Y. Le Nindre, M.
Ziegler, A. Poisson, and S. Bouaziz.2000a. Late Cenomanian (94.7 to 93.5 Ma). In J. Dercourt, M. Gaerani, B. Vrielynck, E. Barrier, B. BiyuDuval, M. F. Brunet, J. P. Cadet, S. Crasquin and M. Sandulescu, eds., Atlas Peri-Tethys: Palaeogeographical Maps, Map 14. Paris: CCGNI/
CGM!r. Philip, J., M. Floquet, J. P. Platel, F. Bergerat, M. Sandulescu, E. Baraboshkin, E. O. Amon, A. Poisson, R. Guiraud, D. Vaslet, Y. Le Nindre, M. Ziegler, S. Bouaziz, and J. C. Guezou. 2000b. Late Maastrichtian (69.5-65 Ma). In J. Dercourt, M. Gaetani, B. Vrielynck, E. Barrier, B. Biju-Duval, M. F. Brunet, J. P. Cadet, S. Crasquin, and M. Sandulescu, eds., A//as Peri-Tethys: Palaeogeographicdl Maps, Map 16. Paris: CCGM/CGMI7. Powell, J. E. 1986. Revisi6n de los Titanosduridos de Am6rica del Sur. Ph.D. dissertation, Universidad Nacional de Tucumdn. Ratkevich, R. 1998. New Cretaceous Brachiosaurid dinosaur, SonordsdLtrus thompsoni gen. et sp. nou. from Arizona . Journal of the Arizona-Neuada Academl' ctf Science 31(1): 71-81. Roth, L. V. 1990. Insular dwarf elephanrs. A case study in body mass estimation and ecological inf-erence. In J. Damuth and B. J. MacFadden, eds., Bctdy Size in Mammalian Paleobiology, 151-179. Cambridge: Cambridge University Press. Salgado, L., R. A. Coria, and J. O. Calvo. 1997. Evohtion of titanosaurid sauropods. I. Phylogenetic analysis based on the postcranial evidence. Ameghiniana 34( 1 ): 3-32.
Between Gondwana andLaurasra
.
427
Sanz, J.
L., J.E. Powell, J. Le Louff, R. Martinez, and X.
Pereda-
Suberbiola. 1999. Sauropod remains from the Upper Cretaceous of Laio (northcentral Spain): Titanosaur phylogenetic relationships. Estudbs del Museo de Ciencias Naturales de Alaud 14 (ndmero especial 1):235-255. Sereno, P. C., A. L. Beck, D. B. Dutheil, H. C. E. Larsson, G. H. Lyon, B. 'Wilson, and Nfoussa, R. W. Sadleir, C. A. Sidor, D. J. Varricchio, G. P. J. A. Wilson. L999. Cretaceous sauropods from the Sahara and the uneven rate of skeletal evolution among dinosaurs. Science 286:
1342-t347. Sereno, P. C., J.A. \(/ilson, H. C. E. Larsson, D. B. Dutheil, and H.-D. Sues. 1994. Early Cretaceous dinosaurs from the Sahara. Science 266:
267-271. Smith, J. B., M. C. Lamanna, K. J. Lacovara, P. Dodson, J. R. Smith, J. C. Poole, R. Giegengack, and Attia Y.2001. A giant Sauropod dinosaur from an Upper Cretaceous mangrove deposit in Egypt. Science 27:
1704-1706. Thierrn J. 2000. Earlv Tithonian (141-139 Ma). In P
S. Crasquin, ed., A//as eri-Tethys: Palaeogectgraphical Maps-Explanatory Notes, 99-1 10.
Paris: CCGMiCGM\f. Thulborn, R. A. 1990. Dinosaur Tracks. Andover, Mass.: Chapman and
Hall. Tidwell, V. A., and K. Carpenter. 2002. Bridging the Atlantic; New correlation of Early Cretaceous Titanosauriformes (Sauropoda) from England and North America. Jc,turnal of Vertebrate Paleontology 22 (supp. to no. 3): 114A. Tidwell, V., K. Carpenter, and S. Nleyer. 2001. New titanosauriform (Sauropoda) from the Poison Strip Member of the Cedar Mountain Formation (Lower Cretaceous), Utah. In D. Tanke and K. Carpenter, eds., Meso:oic Vertebrate Life, 139-165. Bloomington and Indianapolis: Indiana University Press.
Tidwell, V., K. Carpenter, and \X/. Brooks. 1999. New sauropod from the Lower Cretaceous of Utah, USA. Oryctos 2: 21,-37. Tisljar, J., I. Velic, J. Radovcic, and B. Crnkovic. 1983. Upper Jurassic and Cretaceous peritidal, lagoonal, shallow marine and perireefal carbonate sediments of Istria. In L. J. Babic and V. Jelaska, eds., Contributions to Sedimentology of Some Carbonate and Clastic Units of the Coastdl Dinarides: Excursion Guide-book, Fourth I.A.S. Regional Meeting, Split, Yugoslduia, 18-20 April, 1983,1,3-35. Zagreb: International Association of Sedimentologists. Tunis, G., M. Sparica, and S. Venturini. 1994. Lower Cretaceous from Bale (Istria, Croatia): Stratigraphical, sedimentological and palaeoenvironmental problems. Fonrteenth Intenational Sedimentological Congress, AOstl'a ct s. )) :,tz+-I ). Upchurch, P. 1995. The evolutionary history of sauropod dinosaurs. P/:l/osophical Transactions of the Royal Society of London B 349: 365390. rWedel, M.J., R. L. Cifelli, and R. K. Sanders. 2000a. Sduroposeidon proteles, a new sauropod from the Earl,v Cretaceous of Oklahoma. Jour' nal of Vertebrate Paleontology 20(1): 709-11'4. 2000b. Osteologn paleobiology and relationships of the sauropod dinosaur SctLrroposeidon proteles. Actd Palaer:ntologica Polonica 45l.
343-388.
'Weishampel, D. B.,
428
.
Fabio M. Dalla Vecchia
D. Grigorescu, and D. B. Norman. 1991. The di-
nosaurs of Transylvania: Island biogeographv in the Late Cretaceous. National Geographic Research and Explctratbn 7: 68-87.
\Teishampei, D.
B., D. B. Norman, and D. Grigorescu. 1993.
Tel-
mdtosaLffus transsT,luanicus from the Late Cretaceous of Romania: Tl.re rnost basal hadrosaurid dinosaur. P alaeontologl, 36 (2): 36 1-3 85. 'Wellnhofer, P. 1994. Ein Dinosaurier (Hadrosauridae) aus der Oberkreide (X,{aastricht, Heivetikum-Zone) des ba.verischen Alpenvorlandes. Mitteilungen der Bayeriscbe Stddtssdmmlung filr Paliiontologie und historis ch e G eolo gie 3 4 : 227-23 8. 'Wilson, A. 1999. A nomenclature for vertebral lamrnae in sauropods J. and other saurischian dinosaurs. Journal of Vertebrate Paleontology 19\41: 639-653.
2002. Sauropod dinosaur phylogeny: Critique and cladistic anal.vJournal of the Linnean Society 136:217-276. 'Wilson, J. A., and P. C. Sereno. 1998. Earlv evolution and higher level phvlogeny of sauropod dinosaurs. Society of Vertebrate Paleontology, Nlemoir 5: 1-68. 'Winkler, D., L. Jacobs, and P. A. Nlurry.. 1997. Jones Rancl.r: An Early Cretaceous sauropod bone-bed in Teras. Journal of Vertebrate Paleontctlogy 17 (supp. to 3): 85A. Yilmaz, P. O., I. O. Norton, D. A. Leary. and R..J. Chuchla. 1996. Tectonic evolution and paleogeography of Europe. In P. A. Ziegler and F. Horvath, eds., Strttcture and Prospects of Alpine Basins and Forelands, 47-60. M6moire Musee nationale de Histoire naturelle, Peri-Tethys Memoir, no. 2. Paris: Edirions du Museum. Zappatena, E. 1990. Carbonate paleogeographic sequences of the periAdriatic region. Bollettino della Societd Geologica Italiand 109: 5-20. sis. Zoological
Berrveen Gondrvana and Laurasia
.
429
L9. Sauropods of Patasonia: Systematic Update and Notes on Global Sauropod Evolution LpoNenoo SercADo AND Rooorpo A. Conre
Abstract Patagonian sauropods are known from the Middle Jurassic through the Late Cretaceous. The Jurassic record consists of Amygddlodon patagonicus and the basal eusauropods Volkheimeria chubutensis and Patagoslurus fariasi. Early to mid-Cretaceous sauropods include diplodocoids represenre d by Amargasaurus cazaui, "Rebbacbisaurus" tessonei, and Rayososaurus agrioensis, and the first Patagonian titanosauriforms, Chubutisaurus insignis, Andesaurus delgadoi, and other species still under study. The first titanosaurians with procoelus caudal vertebrae appeared in the eariy Cenomanian. The South American titanosaurid radiation coincided with the extinction of the diplodocoids. The Patagonian titanosaurid record is continuous from the Cenomanian up to the end of the Cretaceous, and includes several different forms, like the gigantic Argentinosdurus huinculensis and the relatively small saltasaurines, which are possibly exclusive to the CampanianMaastrichtian of South American.
430
/lt)
/
16
2
4 5
10 12 17
Fig. 19.1. Majr:r sauropod localities in Patagonia. Lago Pellegrini-Cinco Sabos (1), Plaza Huincul Q), Bajdda del Agrio (3), La Amarga (4), El Choc6n (5), Salitral Moreno (6), General Roca
r{
r_\
t-,. Los Aldmitos tSt,lngcniero Iatubacci (')1. Pampo de Agnia 1101. Cerro Condor tl I t, Lstancia Fernindez 1 I 2 t. Cerro Barcino (13), Estancia "Ocho Hermanos" (14), Rio Senguer (15), Chos Malal (16), and Las Horquetas (17 ).
Introduction "Patag6n" is the name given to the aonikenk aborigines by the first Europeans exploring southern South America. Today, "Patagonia" refers to the land once inhabited by the "patagones." Convenrionally, Patagonia is the region south of the Rio Colorado, a territory
that actuaily comprises the provinces of Neuqu6n, Rio Negro, Chubut, Santa Cruz, and Tierra del Fuego (Fig. 19.1). In popular
culture, "Patagonia" is not associated with indigenous peoples, but with dinosaurs and other prehistoric creatures. Indeed, bones of enormous sauropods are common in Patagonia due in part to favorable, erosion-producing "badlands." The first report of sauroSauropods of Patagonia
.
431
pod bones were those dug up by soldiers, settlers, and farmers in the late nineteenth century (Coria and Salgado 2000). Soon afterward, there were scientific commissions organized by the Museo de La Plata, with subsequent research carried out by specially invited European professors R. Lydekker and F. von Huene. These men laid the basis of our current knowledge on the evolution of the sauropods in South America. Beginning in the 1980s there was renewed interest in Patagonian sauropods, with research by J. Bonaparte and J. Poweil from the Miguel Lillo Institute (San Miguel de Tucum6n, Argentina), and later by J. Bonaparte from the Museum Bernardino Rivadavia (Buenos Aires, Argentina). These researchers made successive trips to numerous sites in northern, central, and southern Patagonia (Bonaparte, 1986; Powell, 1.987,1990). Since
then, numerous erpeditions have collected sauropod remains, which we summarize in context with other taxa not found in Patagonia. After all, this region of South America was not isolated during the Mesozoic, so it is impossible to document the evolution of Patagonian faunas outside of the wider context of the global sauropod record. -We organize the chapter geochronologically, describing the diverse Jurassic and Cretaceous taxa and discussing their geological record and phylogenetic relationships. Notes on anatomy, systematics, and evolution of each group are added. Finally, we outline the rnajor features of sauropod evolution during the Cretaceous, the period that has produced the largest amount of sauropod fossils.
Institutional abbreuiations. MACN-Museo Argentino
de
Ciencias Naturales, Buenos Aires, Argentina; MCF-PVPH-Museo
"Carmen Funes," Paleontologia de Vertebrados, Plaza Huincul, Neuqu6n, Argentina; MCS-Museo de Cinco Saltos, Rio Negro, Argentina; MLP-Museo de La Plata, Buenos Aires, Argentina; MPCA, Museo Provincial "Carlos Ameghino," Cipolletti, Rio Negro, Argentina; MPEF-Pv-Museo Paleontol6gico "Egidio Feruglio," Paleontologia de Vertebrados, Trelerv, Chubut, Argentina; and PVl-Instituto "Miguel Lillo," Paleontologia de Vertebrados, Tucuman. Argentina. Jurassic Sauropods
The oldest record of a Patagonian sauropod is the holotype of Amygdalodon pdtagonicus from the Bajocian of Chubut Province, Central Patagonia (Cabrera, 1947; Casamiquela, 7963) (Fie. 79.2). This species, known from a few bones (MLP-46-VIII-27-712,MLP36-XI-10-3/9, and MLP-36-XI-10-312), which represent more than one individual (Rauhut, pers. comm. ,2002), is a probable basal eusauropod (sauropods more related to Sdltasauras than to Vulcdnodon, Wllson and Sereno 1998). Amygdalodon rvas traditionally
included in the "Cetiosauridae" (Mclntosh 1990), a family of sauropods whose monophyly has been repeatedly questioned (Upchurch 1995;'$Tilson and Sereno 1998). The absence of pleurocoels in the cervical and dorsal vertebrae reveals that Amygdalctdon is, in ,132
.
Leonardo Saleado and Rodolfo A. Coria
d
fact, a primitive eusauropod (\X/ilson and Sereno 1998), although the seeming lack of marginal denticles on its teeth suggests it is closer to the neosauropods. The Callovian sauropods include two species from the Cerro C6ndor (Chubut Province), Volkheimeria cbubutensis (southern Cerro C6ndor) and Patagosawrus fariasi (northern Cerro C6ndor) (Fig. 19.3). Both are found in the Caflad6n Asfalto Formation. Rauhut (2002) suggests that the material assigned to Patagosaurus
Fig. 19.2. Amygdalodon patagonicus. Holotype, MLP 3 6-Xl-1 0-3/2. (a) Ceruical centru?n, (b) and (c) dorsal centra, (d) and (e) caudal centra, (f) and (g) rib fragments. Scale bar: 10 cm.
fariasi may belong to more than one species. These sauropods were also initially assigned to the Cetiosauridae. Volkheimeria chubutensis,
which is the more primitive of the two, was considered by Bonaparte (1986) to be related to Lapparentosaurus madagascariezsls from the Jurassic of Madagascar based on the neural laminae of the dorsal vertebrae. Patagosaurus fariasi was considered by him as related to other cetiosaurids, such as Barapasaurus and Cetiosawrus. Rauhut (2002) has correctly pointed out that both of these species from Cerro C6n'$Tilson (2002) dor are basal eusauropods. stated that Patagosaurus is
a basal eusauropod that is more derived than Barapasaurus. .N1though the teeth figured by Bonaparte (1986, figs. 33 and 34) lack marginal denticles on their crowns, they are clearly present in an in situ tooth in the dentary (MACN-CH933, Bonaparte 1986), and in a new dentary that is most probably referable to Patagosaurzs. On the other hand, denticles are absent on the marillary teeth of MACNCH-934 (Rauhut, pers. comm.,2002). The absence of denticles on some teeth may be due to wear because these teeth all show strong apical wear of the labial margins (Bonaparte, 1986, frg.34; Rauhut, pers. comm., 2002). The proximal end of the Patagosaurus trbia is transversely compressed as in other non-neosauropod sauropods (\Wilson and Sereno 1998,48), except Joban z (\Tilson 2002). Sauropods of Patagonia
.
433
iil(ry l
Fig.
19.3 . Patagosaurus
fariasi.
Mounted cdst dt the Museo Argentino de Ciencias Naturales " B ernar dino Riua ddu id, " Buenos Aires,
The holotype
of
Tehuelchesaurus benitezi (Fig. 19.4) was
found in the Caflad6n Calcdreo Formation at Estancia Ferndndez (Kimmeridgian-Tithonian, Rauhut, pers. comm. 2002), a locality northwest of Cerro C6ndor. This species might be related to Omeisaurus tianfuensis from the Middle Jurassic of China (Rich et al. 1999). However, Rauhut (20021 considers T. benitezi as the sister taxon of the Titanosauriformes as defined by Wilson and Sereno (1998). In fact, the opisthocoelous condition of the dorsal vertebrae suggests an affiliation with the Camarasauromorpha as defined by Salgado et al. (1997).
A fragmentary knee (distal end of a femur and proximal
tibia and fibula) of a juveniie sauropod has been collected from the Tordillo Formation (Kimmeridgian) in Chos Malal, Neuqu6n Province, northern Patagonia. Based on the transverse narrowness of the proximal end of the tibia, Garcia et al. (2003) concluded that this fossil belongs to a basal eusauropod. The presence of a basal eusauropod in the Late Jurassic of Patagonia, coupled with the record of the neosauropod sister group (lobaria) in the Neocomian of Niger (Sereno et al. 1999), strongly suggests that non-neosauropod dinosaurs survived in Gondwana, together with more derived groups, after the extincends of the
434 . Leonardo
Saleado and Rodolfo A. Coria
ig. 1 9.1. Tehuelchesaurus benitezi. Holot r-p e, MPEF-P u1 125. Dorsdl uertebrae in left lateral uiew. Scdle har: 10 on. F
tion of these sauropods in other regions (Canudo and Salgado in press ).
Cretaceous Sauropods The Cretaceous has the best record of Patagonian sauropods. Until recently, most sauropod bones were known from uppermost Cretaceous formations. However, in the last few years, a number of new specimens were collected in Lower and mid-Cretaceous strata in northern and central Patagonia. These specimens have substantially modified our view of the evolution of these enormous animals (Sciutto and Martinez 1994; Calvo and Salgado 1995). Patagonian Cretaceous sauropods belong to two major groups, the Diplodocoidea and the Titanosauriformes. The latter clade consists of a series of nested subgroups, the Titanosauria, Titanosauridae, and Saltasaurinae, which are discussed below. The phylogenetic relationships of the Titanosauriformes are debated in Salgado et al. (1,997), Wilson and Sereno (1998), Currn Rogers, and Forster (2001), and'il/ilson (2002). Diplodocoids. \Tilson and Sereno (1998) consider the Diplodocoidea as all neosauropods more closely related to Diplodocus than to Saltasaurus.In the Hauterivian-Barremian, the only sauropod recorded in Patagonia is the diplodocoid Amargasaurus cazaui, a bizarre member of the Dicraeosauridae (Salgado and Bonaparte 1991) (Fig. 19.5). This 1O-meter-long dinosaur, found in Neuqu6n Province, is closely related to Dicraeosaurws sattleri from the Upper Jurassic of East Africa (Salgado 1999).If Amdrgasaurus cazaui is phylogenetically closer to Dicraeosaurus sattleri than it is to Dicraeosaurus hansemtnni, also from East Africa, Dicraeosaurus is paraphyletic and D. sattleri should be replaced, potentially by Amargasaurus sattleri.In fact, none of the ten autapomor'Sfilson (2002, 273) can be phies of Dicraeosaur;as presented by verified in Amdrg.dsdurus cdzaui. Sauropods of Patagonia
.
435
Fig. 19.5. Amargasaurus cazaui. Skeletal restortttion. Scale bar: 1 meter.
The most noticeable feature of the postcranium of Amargasaurus cazaui is the exceptional length of the cervical and anterior dorsal neural spines that are, as in other derived diplodocoids, deeply bifurcated. These spines could have functioned as defensive weapons (Salgado 1999). The cervical and dorsal centra have shal-
low or no pleurocoels, as in other dicraeosaurids. The most
un-
usual feature of the skull is the persistence in the adult stage of the parietal and postparietal fenestrae. Both openings are possibly derived characters common to all dicraeosaurids (Janensch 7929 Salgado and Calvo 1992). As in other dicraeosaurids, the basipterygoid processes, which ordinarily connect the braincase with the roof of the mouth, are extremely long and somewhat divergent. In turn, the supratemporal fenestrae are reduced, as also occurs in rebbachisaurids and titanosaurians (Calvo and Salgado 1995: Salgado and Calvo 7997\,. Besides dicraeosaurids, Early and mid-Cretaceous sauropods include the Rebbachisauridae, a group of basal diplodocoids. The Patagonian rebbachisaurids include Rayososaurus agrioensis from the Aptian (Bonaparte 1996) and "Rebbachisaurws" tessonei front the Cenomanian (Calvo and Salgado 1995) of Neuqu6n province. The Rebbachisauridae are the sister group of the clade comprised by Dicraeosauridae and Diplodocidae (Sereno et al. 1999). Unlike diplodocids and dicraeosaurids, rebbachisaurids have plesiomorphically undivided presacral neural spines. One of rhe most conspicuous synapomophies of the group is the wide erpansion of the posterior blade of the scapula (Wilson 2002), as can be seen in the Patagonian species Rayososaurus agrioensis and " Rebbachisaurus" tessonei (\X/ilson 2002). In the skull, rebbachisaurids have slender, peglike teeth, and they have basipterygoid processes anteriorly oriented as in other diplodocoids (Calvo and Salgado 1995; Sereno et a|. 7999 ; \filson 2002).
A probable rebbachisaurid caudal was reporred by Sciutto and Martinez (1994) from the Upper Cretaceous Bajo Barreal Forma-
436 . Leonardo
Salgado and Rodolfo A. Coria
tion in Las Horquetas near central Patagonia. The
transverse
processes of this anterior caudal are winglike, and the neural spine is formed by four laminae as in other diplodocoids.
Bdsal titanosauriforms. Diagnostic features of this group have been given by Salgado et al. (1997) and N7ilson (2002). The group is widely distributed throughout the world, and the oldest record is
from the Upper Jurassic. Salgado and Calvo (1997) claimed that many species previously assigned to the Brachiosauridae are actually basal titanosauriforms, comprising a series of successive sister taxa of the Upper Cretaceous Titanosauridae. These past misidentifications are due to characters previously considered as diagnostic of Brachiosauridae, which are actually applicable to other, more inclusive groups. Chubutisaurus insignis, a titanosaur-related sauropod from the Aptian of Chubut Province (central Patagonia), is one of these species (Fig. 19.5). Wedel et al. (2000) defined the Brachiosauridae by the elongation of the cervical vertebrae and ribs (although cervical vertebrae are not preserved in most basal titanosauriforms). They suggest that certain genera from the Early Cretaceous of North America, such as Cedarosawrus and Sonorasaurus, might belong to this lineage. The putative sister group of the Titanosauriformes, Tehwelchesattrus benitezl, is from the Upper
Fig. 19.6. Chubutisaurus insrgnrs.
Holotype. MACN-I 8222, Anterior caudal centrum in (A) posterior and (B) lateral uietus; mid-caudal centrum in (C)
posterior and (D) lateral uiews; mid-caudal centrum in (E)
posterior and (F) lateral uieuts; and mid-distal caudal centrum m (G) posterior and (H) lateral uiews. Scale bar: 10 cnt.
Sauropods of Patagoma
.
437
Jurassic of Chubut Province (Rich et aI. 1999; Rauhut 2002). All later Patagonian titanosauriforms are more closely related to the Titanosauria or are Titanosauria. Regarding the evolution of Titanosauriformes, Salgado et al.
(1997) argued that the protrusion on the lateral margin of the femur of the Titanosauriformes and the dorsal expansion of the preacetabular blade of the ilium could be biologically connected. The iliofemoralis muscle originates proximally on the preacetabular lobe of the ilium and distally it inserts on the lateral edge of the femur, below the greater trochanter (Borsuk-Bialynicka 1977, fig. 17). This association of a dorsally expanded preacetabular blade and lateral bulge of the femur suggests an increase in mass of that muscle. The increase of the iliofemoralis muscle may be a consequence of body mass redistribution resulting from the displacement of the center of gravity posteriorly (Salgado et aL. 1,997, fig. 9). This explanation for the lateral bulge is compatible with another explanation-that the bulge is the result of the medial deflection of the femoral diaphysis (\X/ilson and Carrano 1999). The development of a pubic peduncle of the ilium, which is perpendicular to the long axis of the sacrum in titanosauriforms, could be correlated to the posterior shift of the center of gravity and the lateral bulge of the femur (Salgado et al. 1997).If the titanosaunform femur is vertical, that is, arranged in its "resting phase," then the pubic peduncle of the ilium still retains its plesiomorphic anteroventral orientation. Instead, it is the long aris of the ilium and of the sacrum that has changed its orientation (Salgado et aL.1997, fig.9). The diverse modifications reported in the hip of titanosauriforms (and, to some extent, in their hindlimbs) could be related to the lengthening of the forelimbs and the concomitant anterodorsal inclination of the vertebral column (Salgado et alr.1.997, fig. 9). This lenorhino of rhe forelimb is seen in the humerus:femur ratio of basal titanosauriforms, which in Chubtttisaurus insignls is 0.86 and in Andesaurus delgadoi is 0.87. However, the lengthening of the forelimbs also occurs in the Camarasauromorpha, in which the metacarpals are long with respect to the length of the radius (Salgado et a|.7997). Thus, on the one hand, lengthening of the forelimb involves the humerus in titanosauriforms and the metacarpals in the Camarasauromorpha.In Camarasaurus there is ventral rotation of the ilia and the consequent posterodorsal inclination of the sacrum axis (Mclntosh et al. 1,996, plate 10, fig. c). These features may be a synapomorphy of the Camarasauromorpha instead of the Titanosauriformes as has been thought previously. Basal titdnosaurians. Sereno (1998) has defined the titanosaurians. Among the characters are the procoelous anterior caudal vertebrae. Powell (1986,242) conjectures that this condition offers biomechanical advantages related to occasional bipedal posture and the use of the tail as a third supporting point. However, he recognizes that procoely is advantageous only when compared to opisthocoely (Powell 1986,243). However, we note that the in the common ancestor of the titanosaurians (i.e.. all non-titanosaurian 438
.
Leonardo Salgado and Rodolfo A. Coria
titanosauriforms) the caudal vertebrae are amphiplatyan. Therefore, we do not agree with the supposed increase in biomechanical efficiency as the cause of the caudal procoely. It can only account for the greater efficiency of the procelous tail with reference to the opisthocoelous one, as Powell recognizes. Furthermore, Opisthocoelicdudia skarzynskii, whose caudal vertebrae are opisthocoelous, has modifications in the hip that Borsuk-Bialynicka (1,977) believes are for occasional bipedal posture. This means that occasional bipedalism may not necessarily require the attainment of caudal procoely. Titanosaurians are present on all continents except Antarctica. \Tilson and Sereno (1.998, fig. a9) place the origin of the group in the Upper Jurassic. Janenscbia robusta, from the Upper Jurassic of Tendaguru, Tanzania, is possibly a basal titanosaur (see Bonaparte et a1.2000 for an alrernative inrerpretation). The anrerior caudal vertebrae referred to Janenschid are procoelous (Mclntosh 1990) as in other titanosaurians, but the appendicular skeleton of the holotype specimen shows some plesiomorphic characters. For example, the distal end of the tibia is anteroposteriorly expanded (Janensch L9 61 , fig. 6, 21,2) and the referred metacarpals are short and robust (Janensch L961, fig. 2, 194; Bonaparte et al. 2000). It is possible therefore, that the materiai assigned to that taxon may be a mixture of different species. As stated above, many basal titanosaurians or species related to the Titanosauria, most of them from Lower Cretaceous strata, have been included within the Brachiosauridae (Mclntosh 1990). For example, Pleurocoelezs (see Tidwell et al. 2001; Carpenter and Tidwell. this volume) has teeth with wear facets inclined with respect to the labio-lingual axis, as in other titanosauriforms (Calvo 1994). However, in crown width, they present a condition intermediate between Brachiosaurus and Upper Cretaceous titanosaurids. The hoiotype ol Chubutisaurus insignis (MACN-18222) from the Aptian of Chubut Province (Corro 1975) (Fig. i9.6) is the oldest record of a titanosaurian-related sauropod in Patagonia. Recent discoveries corroborate that titanosaurians were well represented in Patagonia by that time (Apesteguia and Gim6nez 2001). Perhaps the recently described Agustinia ligabuei (Bonaparte 1999), as well as new discoveries from the Lohan Cura Formation (Aptian), also belong to this clade (Apesteguia, pers. comm. 2002a). Bdsal titanosaurids. The phylogeneric relationships of the Titanosauridae have been discussed by Salgado et al. (1997), Sanz et al. (1999), Smith et ai. (2001), and Curry, Rogers, and Forster (2001). In this group, the procoelous condition of the anterior caudals of basal titanosaurians has extended farther backward through the vertebral series. Additionally, this group of titanosaurians acquired a series of modifications in the cranial, axial, and appendicular skeleton. One modification is the lateral expansion o{ the ilia, which Wilson and Carrano (1999) hypothesized would have displaced outwardly the origin of femoral protractor muscles and the abdominal oblique muscles. Sauropods of Patagonia
.
439
One of the best-known basal titanosaurids comes from the Cenomanian of central Patagonia (Bajo Barreal Formation, Chubut Province). It is a nearly complete skeleton of Epachthosaurus sciuttoi (Martinez et al.2004); another specimen from northern Patagonia (Neuqu6n Province, upper Cenomanian) is also referable to Epachthosaurus (Calvo
1999 Sim6n and Calvo
20021.
Epacbthosaurus retains plesiomorphic characters that separate this taxon from the derived titanosaurids (Salgado and Martinez 1993). For instance, the hyposphene-hypantrum complex, which works as an accessory articulation in all saurischians, is lost in all titanosaurids except Epachthosawrws. Powell (7986, 301) and Apesteguia (this volume) believe the hyposphene-hypantrum complex gives the vertebral column greater rigidity and that its loss would increase the fleribility. As yet, the skull of a basai ti-
tanosaurian is unknown, but material recently found in the Bajo
ig. 1 9.7. Argentinosaurus huinculensis. Holotyp e, MCF PVPH-1. Postelior dorsal uertebra in lateral uieu. Scale bar: 50 cm. F
Barreal Formation will provide much insight into the cranial anatomy of basal titanosaurids (Martinez 1,998). Deriued titanosaurids. Titanosaurids with dorsal vertebrae lacking the hyposphene-hypantrum complex, and a prespinal lamina extending down to the base of the neural spine, first appear in the upper Cenomanian. Argentinosaurus huinculensis (Fig. 1'9.7) from Plaza Huincul, Neuqu6n Province, and Argyrosdurus superbus from southern Chubut, are huge titanosautids of possible Cenomanian age (however, new evidence suggests that the holotype
forelimb of Argyrosaurws superbus may be Campanian and a questionably referred specimen, PVL 4628, may be Cenomanian; Lamanna, pers. comm. 2003). Argentinosaurus is perhaps the largest sauropod and terrestrial animal ever found. The presence of accessory intervertebral articulations in the dorsal vertebrae (analogous to the hyposphene) and its hollow ribs may be adaptations
for its large size (Bonaparte and Coria 1993). Mazzetta (1999) has estimated the weight of Argentinosdurus as 120 tons. The Senonian (Coniacian-Maastrichtian) of Patagonia is characterized by the presence of numerous derived titanosaurids (Salgado et al. 7997). From the Anacleto Formation comes Antarctosaurus wichmannianus (General Roca, Rio Negro Province; Huene 1,929), Laplatasaurws araukaniczs (Cinco Saltos, Rio Negro Province; Huene 1.929; Powell 1986), l{euquensaurus australis (Cinco Saltos, Rio Negro Province), and Pellegrinisaurus powelli (Lago Pellegrini, Cinco Saltos, Rio Negro Province; Salgado 7996); lrleuquensaurus australis is also known from the Bajo de la Carpa Formation, Neuqu6n City, Neuqu6n Province (Salgado et al., submitted). The Allen Formation has produced Rocasdurus muniozi (Salitral Moreno, Rio Negro Province; Salgado and Azpilicueta 2000). Finalln the Angostura Colorada Formation has produced Aeolosaurus rionegrinus (Ingeniero Jacobacci, Rio Negro Province;
Powell 1986). The evolutionary changes in the skulls of derived titanosaurids include slender teeth and their confinement to the anterior extremity of the mandibles (Coria and Salgado 1999). The attenuation of
440 . Leonardo
Salgado and Rodolfo A. Coria
the teeth commenced with the rise of the titanosaurians, if not earlier (Salgado and Calvo 7997). The slenderness of the teeth and their terminal position in the jaws, also present in diplodocids, are linked to a particular chewing style (Calvo 1994) and to the adoption of a specific diet. The Saltasdurinae.Whrle all cladistic analyses agree on the existence of the Saltasaurinae, they differ in the taxa included in this subgroup (Salgado et aI. 1997,32; Wilson 2002,269). Currently, this clade includes the Patagonian species I'Jeuquensaurus australis (Figs. 19.8, 79.9), RocasAurus muniozi (Salgado and Azpilicueta 2000) (Fig. 19.10), and Sdltasdurus loricatus (from El Brete, Salta Province, northwestern Argentina; Bonaparte and Powell 1980).
g. 1 9.8. Neuquensaurus australis. Leit, skeletal restoration mounted at Museo de La Plata, Argentina; (A) MCS-5/28, right femur, (B)MCS-S, bttmerus, (C) MCS-6, tibia. Scale bar: 70 cm. Fi
lxleuquensaurus australis, one of the species identified by R. Lydekker in 1893, is the best-represented sauropod in the Upper Cretaceous of North Patagonia. Saltasaurines are characterized by the presence of caudal verte-
brae of camellate inner structure ("spongy texrure" of \Tilson 2002). Powell (1986) has pointed out that the bony tissue observed in the dorsal vertebrae of non-saltasaurine titanosaurids has extended caudally in saltasaurines. In Rocasaurus munioTi, the intricate system of inner spaces of the caudal vertebrae is connected with the exterior through small holes (which could be interpreted as true pleurocoels) piercing the ventral and lateral faces of the ver-
tebra (Salgado and Azpilicueta 2000). Although the relative development of cavernous bone has obvious taxonomic value, as discussed by Salgado and Azpilicueta (2000) for R. munioer, there is notable variation within a single taxon, including ontogenerically. Powell (1986,300) understood that the vertebral centra composed by "macrocells" were adaptive, in relation to the lightening of the axial skeleton. An alternative to this explanation is that the camellate texture of the vertebrae is not itself adaptive, but the result of some physiological process that required the mobilization of bony Sauropods of Patagonia
.
441
-
Fig. 19.9. Neuquensaurus australis, MCS-5 /22. P osterior dorsal uertebra in (A) anterior and (B) left lateral uiews. Scale bar: 10 cm.
Fig. 19.10. Rocasaurus muniozi. (A) MPCA-PV-.S6, ischium and Pubis; (B) MPCA-PV-S8, Postelior caudal in left lateral uiew; (C) MPCA-PV-46, holotype, posterior caudttl it right lateral uiew; (D) mid-caudal in uentrdl uiew; and (E) anterior caudal in uentral uiew. Scdle bar: 'l0 cm.
442 . Leonardo
calcium, particularly under situations in which that element cannot be taken directly from food (Salgado 2000). Saltasaurines are among the smallest sauropods (Salgado 2000; Apesteguia 2002a), and they are comparable to dicraeosaurids and Magyarosaurus from the latest Cretaceous of Rumania (Jianu and l7eishampel 7999). There are two aspects about the evolution of size that should be differentiated. One concerns the evolutionary mechanisms responsible for a phyletic decrease in size, and the other is the adaptive benefit for such a reduction. Small size could be positively selected for, or it could simplv be a secondary effect or a by-product of some other process. The evolutionar)' mechanism for size change can be defined in terms of heterochrony, and has been discussed by McKinney and McNamara (1997). In saltasaurines. it is possible that the reduced size is the result of natural
Salgado and Rodolfo A. Coria
selection pressures imposed by the presence of large predators. An-
other explanation for the decrease in size is ecological and associated with life in coastal environments (Apesteguia 2002b). Regardless, paedomorphosis resulted in insular dwarfism in saltasaurines, as it did in other d-uvarf sauropods in Europe, like Magyaroslurus and Ampelosaurus (Jianu and \Teishampel 1999).
Patagonian Sauropod Faunas throughout the Cretaceous Neocomian dinosaur faunas are poorly known in Patagonia (Fig. 19.11). Amargasaurus cazdui reveals the persistence of the dicraeosaurids. whose first representatives are from the Upper Jurassic of Africa (Salgado and Bonaparte 19971.In Lower Cretaceous faunas elsewhere in western Gondwana, Jurassic relicts are a recur-
rent component (Rauhut 7999; Sereno et al. 1999 Canudo and Salgado in press). By the end of the Early Creraceous, riranosauriforms were widely distributed in North America, South America, Eurasia, Australia, and Africa. In Europe, Africa, and South America, they lived along side rebbachisaurids and other basal diplodocoids (Salgado et a1. submitted). Titanosaurian-related sauropods (titanosauriforms more derived than Brachiosdurus, Gomani et aI. 7999,231) r,vere widespread by the end of the Early Cretaceous and were probably the only sauropods in North America and probably in Asia. In Patagonia, titanosaurian-related sauropods appeared in the Aptian (Chubutisaurus insignis fuom Chubut Province; Salgado 1993), whereas the first undoubted titanosaurian (Andesaurus delgadoi) is known from the lower Cenomanian (Calvo and Bonaparte I99 L). The Titanosauridae, with a profuse record in Patagonia, evolved from a group of basal titanosaurians, probably by the lowermost Cretaceous. In Patagonia, the oldest record of a procoelous caudal vertebra (the earliest indisputable titanosaurid) is Cenomanian (Calvo and Salgado 1998). The geographic origin of the Titanosauridae is a matter of debate. Traditionally, they were believed to have originated in Gondawana based on their abundant record in India, Madagascar, and, especially, South America (Powell 1986; Bonaparte and KielanJaworowska 1987). Le Loeuff (1991) considered titanosaurids as a central component of the "Eurogondwana" paleoprovince. Salgado and Calvo (1997), in contrast, suggested that the first titanosaurids lived in Europe, based on the record of Iwticosdurus ualdensis in the Barremian of the Isle of Wight. Sanz et aL. 17999) suggested that titanosaurians originated in "Neopangea," with a strictly Euroamerican lineage (represented lry luticosaurus) and another lineage repre-
sented
by Malawisaurus (these authors do not refer to
the
Titanosauridae but to the Titanosauria, although these two genera are interpreted here as members of Titanosauridae). According to these authors, titanosaurids would have reached Asia from Euramerica during the Barremian-Aptian. Titanosaurids (and ail other sauropods) apparently became extinct in Europe after the Cenomanian, and reentered from Africa in the Late Campanian (Le Loeuff Sauropods of Patagonia
.
443
to
3 \
(t,
t4
\ q, (! a
(!
\
t4
o
t4 G
o
o (! o o cr,
Maastrichtian
o
= q Ilr \ .n 5 \ rg t! t4 t4 t4 o (! 5 o s t4 fr *ra o L 5 q (! \ o o tt, L G P o -4. P q tn
Campanian
t4
5 L
: (!
q
Santonian
a 3 o uJ o
Coniacian Turonian
IJJ
u
(J
to
()
(!
5 La ao $ qq o t4 o qa
G
an tha
t4
s\ I
G ttl G
Aalenian
lJJ
\ \ C
t *i
\ {c
G
t4
{ q i. G q o
0-' Bajocian
R
\
o) G tha (!
Ia
q
{
c
t4
h
Hauterivian Valanoinian Berriasian
I}
3
a
P G
3 a
s
!t
fl
o u
c c
o ()
o .o t\
F?
r .R
.Ff,
O*, h
o
-J
1991 Le Loeuff and Buffetaut
199 5; see Wilson and Sereno 1.998
a reply). The European extinction of the titanosaurids coincided
for
with
the extinction of North American sauropods (Lucas and Hunt 1989; Le Loeff 1991).In North America, the decline of the sauropods occurred before the Albian (Wedel et al. 2000). Inexplicably, the midCretaceous extinction of sauropods (by the Cenomanian-Turonian
boundary) in Patagonia seems to have affected basal diplodocoids and many, but not all, groups of titanosaurians. The many genera of derived titanosaurids in the Neuqu6n Basin are the only sauropods recorded above the level of the Huincul Formation (upper Cenomanian), and presumably they evolved from one or more species that survived the extinction. A similar situation occurred in the San Jorge Basin of central Patagonia, where an isolated diplodocoid specimen is recorded as late as the Turonian-Coniacian (Lamanna et al. 2001; Martinez et aI.2001).'We hypothesize that the regional extinction of all sauropods in Europe and North America by the middle of the Cretaceous, and the disappearance of diplodocoids and certain groups of titanosaurians in Patagonia at the same time, are different facets of a single or a series of events occurring on a global scale' The key to understanding this rather dramatic faunal turnover is probably linked to variations of the mid-Cretaceous floras and the decline of many plants that were integrated into the regular diets of these animals (Salgado 2000). The two major groups of sauropods that inhabited Patagonia in the mid-Cretaceous had different masticatory styles that most
likely correlated with different diets (see Calvo 1994). Diplodocoids have slender, cylindrical teeth that are restricted to the front of the snout, whereas basal titanosauriforms have compressed, cone-shaped, chisel-like teeth (Calvo 1999, fig. 10; Sim6n 2001; Sim6n and Calvo 2002). Thus, the mid-Cretaceous floral turnover would have affected the two sauropod groups in different rvays. In northern Patagonia, the youngest record of diplodocoids (upper Cenomanian) come from levels lower than those containing the oldest titanosaurids (Calvo and Salgado 1995). The latter obviously evolved from groups that were present in the mid-Cretaceous (Aptian-Cenomanian interval). Did these middle Cretaceous ancestors have cone-shaped, chisel-like teeth, as do the basal titanosaurids? In other words, did typical titanosaurids acquire a cylindrical dentition once the diplodocoids became extinct, or does the extinction of the diplodocoids predate the expansion and diversifi-
cation of titanosaurids with cylindrical dentition? The occurrence of slender-toothed titanosaurids in the Cenomanian of Africa agrees better with the second interpretation (Kellner and Mader
1997; Rauhst 7999). Upchurch (1995) has suggested that titanosaurids and diplodocoids survived into the Late Cretaceous by having slender teeth and a particular chewing style. However, as already discussed, the dominant group of sauropods with cylindrical teeth during the Lower Cretaceous was the Diplodocoidea (dicraeosaurids and rebbachisaurids). which did not survive into the Late Cretaceous. Sauropods of Patagonia
.
445
Wilson and Sereno (1998) claimed that the slender teeth in sauropods evolved long before the rise of the angiosperms. As they
note, diplodocids of the Late Jurassic already possessed slender teeth with subcircular wear facets perpendicular to the axis of the tooth (Salgado and Calvo 7997; \Wilson and Sereno 1998). \X/edel et al. (2000) claim that the spread of angiosperms is not a satisfactory causal explanation for the exrinction of North American sauropods in the middle Cretaceous. However, if titanosauriforms are considered alone, we see that replacement of taxa having compressed, cone-shaped teeth by forms with cylindrical teeth coincides roughly with the beginning of the rapid expansion of these
plants (Sues and Wing 1992). In Europe, the broad-toothed titanosaurid Ampelosaurus dtdcis (Le Loeuff et al. 7994; this volume) and the slender-toothed Lirainosaurus astibide (Sanz et al. 1999) arc present in the uppermost Cretaceous. Hence, in Europe, uniike in Patagonia, sauropods rvith different feeding styles are present at the end of the Cretaceous. It has yet to be established whether the broad teeth of Ampelosaurus represents the plesiomorphic condition as'Wilson and Sereno (1998) suppose, or; as seems more probable, that it represents a derived state that is autapomorphic of that genus. If rffilson and Sereno are correct, then Ampelosaurus could be a vicariant lineage rhat persisted in Europe from a broader Early or mid-Cretaceous fauna that included Europe, Africa, and Amenca. \il/hat caused the mid-Cretaceous turnover? Lucas and Hunr (1989) have explained the extinction of sauropods in North America as the result of climatic change produced by a regression of the sea that occurred in the late Albian. Extinction affecting both nonmarine and marine faunas at the Cenomanian-Turonian boundary has been recorded in southwestern Utah (Eaton et aL. 1997), and ir may be responsible for the extinction of the North American sauropods. In southern Patagonia, Archangelskv (2001) recorded that, near the beginning of the Aptian, a floral assemblage characterized by the disappearance of the Benettitales and most of the Cvcadales and Ginkgoales was replaced by gleicheniacean ferns. This change in flora cornposition was due to volcanic acrivity that produced environmental changes. \7e hypothesize that toward the end of the Lower Cretaceous and the beginning of the Upper Cretaceous, in Patagonia as well as in continents other than South America, there were successive changes in mean temperatures (with multiple causes) and a concomitant alteration in flora1 comoosition. This resulted in floral turnover that reverberated in rhe sauroood faunas, as weil as in other groups of dinosaurs (F. Novas, pers. comm. 2002). The extinction of the diplodocoids ar the end of the Cenomanian paved the way for the dominance of the titanosaurians, and the radiation of the angiosperms was ar least partiallv responsible for the expansion and diversificarion of titanosaurids with slender, peglike teeth. Derived titanosaurids are abundant in northern Patagonia, but they are scarce in central and southern Patagonia. We do not know
446 . Leonardo
Salgado and Rodolfo A. Coria
if this is due to unequal sampling, the
incompleteness of the fossil record, or ecological factors. Recent discoveries in Chubut (Sciutto and Martinez 7994, UNPSJB-Pv 581; Casal et al. 2002) and Santa Cruz provinces (Novas et al. 2002) seem to support the first alternative. In addition, the Saltasaurinae seem to be restricted, both spatially and chronologicall,v, to the uppermost (Campanian-Maastrichtian) Cretaceous of southern South America. They are found in Rio Negro, Neuqu6n, and the Salta provinces, north of 42" south latitude. Their absence south of 42" is interesting, and although it might be explained by the imperfection of the paleontological record, there is the possibility that the Norpatagonian High had established a barrier that impeded the saltasaurines (Salgado and Azpilicueta 2000). Ho'uvever, Curry Rogers (2002) reported on a probable saltasaurine from the Upper Cretaceous of Madagascar, which would suggest no barrier was present. In Patagonia, saltasaurines are never the exclusive, nor even the dominant, group of sauropods, since in most localities they are associated with other titanosaurids, such as Laplatasaurusu araukdnicus, AntarctosAurus wicbmannianus, Aeolosaurus sp. and Pellegrinisaurus powelli, ar'd other indeterminate species (Garcia and Salgado 2002).
Late Cretaceous Diplodocoids? Many authors (Jacobs et al. 7993; Upchurch 1995; Wilson and Sereno 1998; Sereno et al. 19991 have suggested that Antarctoscturus wichmannidnus, from the Anacleto Formation in Rio Negro Province, is not a titanosaurid, as historically regarded (Huene 1929; Bonaparte and Gasparini 1980; Powell 1986; Salgado and CaIvo 1.997), but is a diplodocoid. Sereno et al. (1999) more precisely suggested that some of the cranial materials of Antarctosdurus wichmannidnus, particularly the lower jaw, belong to a rebbachisaurid based on the lower jaw of Nigersaurus tdqueti. Antarctosaurus and diplodocoids do share characters of the skull including cylindrical, peglike teeth restricted to the anterior extremity of the snout, which \Tilson and Sereno (1998) accept as convergent, but which Jacobs et aL. (1993) do not. Furthermore, both taxa share a symphyseal margin set at a right angle, and a slender basipterygoid process. However, there are characters in the skull of Antarctosaurus that confirm its titanosaurian affinities, including the presence of semilunate paraoccipital processes (Wilson 2002) that are also seen in Saltasattrus loricdtus and Rapetosaurus krausei. It could be argued, then, that the lower jaw of Antarctosdurus wichmannianzs belongs to a rebbachisaurid and that the other parts of the skeleton, including the braincase, pertain to a true titanosaurid. However, the absence of diplodocoid postcranial remains in the Coniacian-Maastrichtian of Patagonia would argue against this alternative. Although no derived titanosaurid with a mandible can resolve this problem of Antarctctsdurus, a new titanosaurid specimen from Rinctin de los Sauces (northern Neuqu6n Province) has a skuli rvith slender, cylindrical teeth restricted to the Sauropods ofPatagonia
.
447
anterior ends of the jaws as in the dipiodocoids (Coria and Salgado 1999, in prep.). Furthermore, this skuil reveals that the degree of convergence in the skulls of titanosaurids and diplodocoids is greater than previously suspected (e.g., Salgado and Calvo 7997). For instance, the quadrate is clearly slanted forward and the infraremporal fenesrra is extended below the orbit as in diplodocoids. These features negate previous statements that the titanosaurid
skull was more similar to that of Brachiosaurws (Salgado and Calvo 1997). Further supporr for this hypothesis is seen with the
skulls of lJemegtosaurus and Quaesitosaurus from the Upper Cretaceous of Mongolia. Although Upchurch (1995, 1999) placed these two taxa within the Diplodocoidea, Salgado and Calvo
to be titanosaurians, and Wilson and Sereno (7998) placed rhem together with the titanosaurians u,ithin (1997) considered them
Macronaria. Elucidating the phylogenetic relationships of these three sauropods, Antarctosaurus, I'tremegtosaurus, and Quaesitoslurus, is crucial because, if they are true titanosaurians, this clade is the only group of posi-Cenomanian sauropods, with the probable exception of an isolated record in the TuronianConiacian of the Chubut Province. This would srrengthen the hypothesis that the mid-Cretaceous extinction would have affected all sauropods with the exceprion of derived titanosaurians. Acknowledgments. We thank S. Apesteguia, K. Carpenrer, O. Rauhut, M. Lamanna, and V. Tidrvell for sharing information and usefui comments. References Cited Apesteguia, S. 2002a. Successional structure in continental tetrapod faunas from Argenrina along the Creraceous. Boletim do 6,, simp6sict sobre o Cretdceo do Brasil/ 2. Simposict sctbre el Cretdcico de Amlrica del Sur, Sao Pedro, Brasil, 135-141. Rio Clar.o, SP, Brazil; DIVISA, Editora e artes grlficas Ltda.
2002b. Greater Gondwana and the Kawas sea coastal tetrapod fauna (Campanian-Maastrichtian). In Boletim do 6,, sintp6sio sobre o Cretdceo do Brasil/2" Simposio sobre el Cretdcicct de Am1rica del Sua 143-147. Rio Claro, SP, Brazil: DIVISA, Editora e artes grdficas Ltda. Apesteguia, S., and O. Gim6nez.2001. A tiranosaur (Sauropoda) from the
Gorro Frigio Formation (Aptian, Lower Cretaceous), Chubut
Province, Argentina. Ameghiniana 37, supp. 4R; 4. Archangelsky, S. 2001. Evidences of an Earlv Creraceous floristic change in Patagonia, Argentina. In H6ctor A. Leanza, ed., V1I Internatk:nal Symposiutn on Mesrtzoic Terrestrial Ecosystems, 15-19. pub[caci6n Especial, no. 7. Buenos Aires: Asociaci6n Paleontol6gica Argentina. Bonaparte, J. F. 1986. Les Dinosaures (Carnosaures, Allosaurid6s, Sauropodes, C6tiosaurid6s) du Jurassique Moyen de Cerro C6r.rdor (Chubut, Argentina). Annales de Paldontologie 72: 32j-J86. 1996. Cretaceous Terrapods of Argentina. Miinchner Geowissenscbaftlicbe Abhandlungen, Geologie und Palaeontologie 30:
73-130. 7999. An armoured sauropod from the Aptian of Northern Pata-
448
.
Leonardo Saleado and Rodolfo A. Coria
gonia, Argentina. In Y. Tomida, T. Rich, and P. Vickers-Rich, eds., Proceedings of the Second Gondwanan Dinosaur Symposiu.m, 1-72. National Science Museum Monographs, no. 15. Tokyo: Natronal Science Museum. Bonaparte, J. F., and R. A. Coria. 1993. Un nuevo y gigantesco saur6podo titanosaurio de la Formaci6n Rio Limay (Albiense-Cenomaniense) de la Provincia del Neuqu6n, Argentina. Ameghiniana 30:271-282. Bonaparte, J. F., and Z. Gasparini. B.1980. Los saur6podos de los grupos Neuqu6n y Chubut y sus relaciones cronol6gicas. Actas del VII Congreso Geol6gico Argentino, Neuqudn 2: 394-406. Bonaparte, J. F., S7.-D. Heinrich, and R. \7ild. 2000. Review of Janenschia \7ild, with the description of a new sauropod from the Tendaguru beds of Tanzania and a discussion on the systematic value of procoelous caudal vertebrae in the Sauropoda. Paleontographica Abt. A.
256:25-76. Bonaparte, J. F., and Z. Kielan-Jaworowska. 1987. Late Cretaceous dinosaur and mammal faunas of Laurasia and Gondwana. In P. Currie and E. H. Koster, eds., Fourtb Symposiutn on Mesozoic Terrestrial Ecosystems: Short Papers, 24-29. Drumheller, Alberta, Canada: Royal Tyrrell Museum of Palaeontology. Bonaparte, J. F., and J. E. Powell. 1980. A continental assemblage of tetrapods from the Upper Cretaceous beds of El Brete, northwestern Argentina (Sauropoda-Coelurosauria-Carnosauria-Aves). Mdmoires de Ia Societd Gdologique de France 139: 19-28. Borsuk-Bialynicka, M. 1977. A nelv camarasaurid sauropod Opisthocoelicaudia skarzynskii, gen. n., sp. n. from the Upper Cretaceous of Mongolia. P aleontologia P olonica 37 : 1,-64. Cabrera, A. 1947. Un saur6podo nllevo del Jur6sico de Patagonia. No/as del Museo de La Plata 12 Paleontolctgia 95: I-17. Calvo, J. O. 1,994. Jaw mechanics in sauropod dinosaurs. Gaia I0: 1
83-1 93.
1999. Dinosaurs and other vertebrates
of the Lake
Ezequiel
Ramos Mexia Area, Neuqu6n-Patagonia, Argentina. In Y. Tomida, T. H. Rich, and P. Vickers-Rich, eds., Proceedings of tbe Second Gondwanan Dinosaur Symposium, 13-45. National Science Museum Monographs, no. 15. Tokyo: National Science Museum. Calvo, J. O., and J. F. Bonaparte. 1,991,. Andesaurus delgadoi gen. et sp. nov.(Saurichia-Sauropoda), dinosaurio Titanosauridae de la Formaci6n Rio Limay (Albiano-Cenomaniano), Neuqu6n, Argentina. Ameghiniana 28: 303-3 10. Calvo, J. O., and L. Salgado. 199 5 . Rebbachisaurus tessonei sp. nov. a new Sauropoda from the Albian-Cenomanian of Argentina; new evidence on the origin of Diplodocidae. Gaia 11: 13-33. 1998. Nuevos restos de Titanosauridae (Sauropoda) en el Cret6cico Inferior de Neuqu6n, Argentina. VII Congreso Argentino de Paleontologia y Bioestratigrafia, Bahia Blanca, Argentina. Resimenes, 59.
Canudo, J. I., and L. Salgado. In press. Los Dinosaurios del Neocomiense (Cretdcico Inferior) de la Peninsula Ib6rica y Gondwana Occidental: implicaciones biogeogr6ficas. Zubia. Casal, G., M. Luna, L. Ibiricu, E. Ivany, R. D. Martinez, M. Lamanna, and A. Koprowskv. 2002. Hallazgo de una serie caudal articulada de Sauropoda de la Formaci6n Bajo Barreal, Cretdcico Superior del sur
Sauropods of Patagonia
.
449
de Chubut. XVIII Jornadas Argentinas de Paleontologia de Vertebrados, Ameghiniana 39, supp. 8R: 8. Casamiquela, R. M. 1963. Consideraciones acerca de Amygdalodon Cabrera (Sauropoda, Cetiosauridae) del Jurdsico Medio de la Patagonia.
Amephtntana J: -9-95. Coria, R.A., and L. Salgado,
L.
1999. Nuevos aporres a la anaromia
craneana de Ios saur6podos titanosduridos. Ameghiniana 36:98. 2000. Los dinosaurios de Ameghino. In S. Vizcaino, ed., Obra de los Hermanos Amegbino,43-49. Lujdn; Publicaci6n Especial Universidad Nacional de Lujdn. Corro, G. del. 1975. Un nuevo saur6podo del Cretdcico Superior,
Chubutisaurus insignis gen. et sp. nov. (Saurischia, Chubutisauridae, nov.) del Cret6cico Superior (Chubutiano), Chubut, Argentina. Aclas Primer Congreso Argentino de Paleontologia y Bioestrattgrafia, Tv cumdn, Argentina, 12-16 de agosto de 1974,229-240. Tucumdn: Asociaci6n Paleontol6gica Argentina. Curry Rogers, K. 2002. Titanosaur tails tell rwo rales: A second Malagasy titanosaur. Journal of Vertebrate Paleontok:gy,22 (supp. to 3): 47A. Curry Rogers, K., and C. A. Forster. 2001. The last of the dinosaur titans: A new sauropod from Madagascar. Nature 412: 530-534. Eaton, J. G., J. I. Kirkland, J. H. Hutchison, R. Denton, R. C. O'Neill, and J. M. Parrish. 1997. Nonmarine exrincion across the CenomanianTuronian boundary, southwestern Utah, with a comparison to the Cretaceous-Tertiary extinction event. Geological Society of America
Bulletin 709: 560-567. Garcia, R., and L. Salgado. 2002. Titanos6uridos de la localidad de Salitral Moreno (Cret6cico Superior, Provincia de Rio Negro). VII Congreso
Argentino de Paleontologia
y
Bioestratigrafia, Bahia Blanca, Ar-
gentina. Resimenes, 62. Garcia, R. A., L. Salgado, and R. A. Coria. 2003. Primeros restos de drnosaurios saur6podos en el Jurdsico de la Cuenca Neuquina, Patagonia, Argentina. Ameghiniana 40 L23-126. Gomani, E. M., L. L. Jacobs, and D. A. Vinkler. 1999. Comparison of the African titanosaurian Malawisaurus, with a North American Early Cretaceous sauropod. In Y. Tomida, T. H. Rich, and P. Vickers-Rich, eds., Proceedings of the Second Gondwanan Dinosaur Symposium, 223-233. National Science Museum Monographs, no. 15. Tokyo: National Science Museum. Huene, F. von. 1929. Los Saurisquios y Ornitisquios del Cret6ceo Argentino. Anales del Museo de La Plata, ser. 2, no. 3:1-196. Jacobs, L.L., D.A. \)finkler,'W.R. Downs, and E. Gomani. 1993. New material of an Early Cretaceous titanosaurid sauropod dinosaur from Malarvi. Palaeontology 36: 523-534. Janensch,'W. 1929. Material ur.rd Formengehalt der Sauropoden in der Ausbeute der Tendaguru-Expedition. P aleonto grapblca, supp. 7 (1,), 2: 1-34. 1.961. Die Gliedmaszen und Gliedmaszengiirtel der Sauropoden der Tendaguru-Schichten. Palaeontographica, supp. 7 (11, 3: 177-325. Jianu, C. M., and D. B. \Teishampel. 1.999. The smallest of the largest: a new look at possible dwarfing in sauropod dinosaurs. Geoktgie en Mi-
inbouw -8: 335-34J. Kellner, A.'W., and B.J. Mader. 1997. Archosaur reeth from the Cretaceous of Morocco. Journal of Paleontology 71: 525-527. Lamanna, M. C., R. D. Martinez, M. Luna, G. Casal, P. Dodson, and J. B. Smith. 2001. Sauropod faunal transition through the Cretaceous
450 . Leonardo
Saleado and Rodolfo A. Coria
Chubut Group of Central Patagonia. .lournal of Vertebrate Paleontology
2l:711'.
Le Loeuff, J. 1,99L. The Campano-Nlaastrichtian vertebrate faunas from southern Europe and their relationships with other faunas in the world: Paleobiogeographical implications. Cretaceous Research 72 93-114. Le Loeuff, J., E. Buffetaut, L. Cavin, NI. Martin, V. Martin, and H. Tong. 1"994. An armoured titanosaurid sauropod from the Late Cretaceous of Southern France and the occurrence of osteoderms in the Titanosauridae. Gaia 10 1 55-159. Le Loeuff, J., and E. Buffetaut. 1995. The evolution of Late Cretaceous non-marine vertebrate fauna in Europe. In Sun A. and\fang Y. (eds.), Sixth Symposium on Mesozoic Terrestrial Ecosystems and Biota: Short Papers, pp. 181-1 84. Beijing: China Ocean Press. Lucas, S. G., and A. P. Hunt. 1,989. Alamosaurus and the sauropod hiatus in the Cretaceous of the North American 'Western Interior. In J. O. Farlow, ed., Paleobiology of the Dinosaurs, T5-85. Special Paper/The Geological Society of America, no. 238. Boulder, Colo.: Geological Society of America. Lydekker, R. 1893. Contributions to the study of the fossil vertebrates of Argentina. I. The Dinosaurs of Patagonia. Anales del Museo de La Pldta 2: 1,-1,4. Martinez, R.D. 1998. An articulated skull and neck of Sauropoda (Dinosauria: Saurischia) frorn the Upper Cretaceous of central Patagonia, Argentina. Journal ofVertebrate Paleontology 18: 61A. Martinez, R. D., O. Gim6nez, J. Rodriguez, M. Luna, and M. C. Lamanna. 2004. An artictilated specimen of the basal titanosaurian (Dinosauria, Sauropoda) Epachthosaurus sciuttoi from the early Late Cretaceous Bajo Barreal Formation of Chubut Province, Argentina. Journal of Vertebrate
P
aleontology 24: 107 -1,20.
Martinez, R. D., M. C. Lamanna, G. Casa1, M. Luna, P. Dodson, C. Tiedemann, and A. Koprovski. 2001. Dinosaurios de la Formaci6n Bajo Barreal, Cret6cico Superior Temprano del sur del Chubut. Ameghiniana38, supp. 12R: 12. Mazzetta, G.V. 1999. Mecdnica locomotora de dinosaurios saurisquios del Cretdcico Sudamericano. Ph.D. dissertation, Universidad de la Repiiblica Oriental del Uruguay. Mclntosh, J. S. 1990. Sauropoda. In D. lfeishampel, P. Dodson, and H. Osm6lska, eds., The Dinosauria, 345-401. Berkeley: University of California Press. Mclntosh, J.S., \7.E. Miller, K.L. Stadtman, and D.D. Gillette.1996. The osteology of Camarasaurus lewisi (Jensen, 1988). BYU Geolctgy Studies 41:73-11,5. McKinnen M. L., and K. J. McNamara. 1.991. Heterochrony: The Euolution of Ontogezy. New York: Plenum Press. Novas, F. E., E. Bellosi, and A. Ambrosio. 2002. Los "Estratos con Dinosaurios" de1 Lago Viedma y Rio La Leona (Cretdcico, Santa Cruz): Sedimentologia y contenido fosilifero. Actas del XV Congreso Geol6gico Argentino 1,: 596-602. Powell, J.E. 1986. Revisi6n de los titanosduridos de Am6rica del Sur. Ph.D. dissertation, Universidad Nacional de Tucumdn. 1987. The Late Cretaceous fauna of Los Alamitos, Patagonia, Argentina. Part 6. The titanosaurids. Reuista del Museo Argentino de Ciencias Nalurales
3: l4--
153.
Sauropods of Patagonia
.
451
1990. Epacbthosaurus sciuttoi (gen. et sp. nov.) un dinosaurio Saurop6do del Cretdcico de Patagonia (Provincia del Chubut, Argentina). Actas V Congreso Argentino de Paleontologia y Bioestratigrafia l: 123-128. Rauhut, O. \)f. M. 1999. A dinosaur fauna from the Late Cretaceous (Cenomanian) of northern Sudan. Paleontologia Africana 35: 61-84. 2002. Los dinosaurios de la Formaci6n Cafrad6n Asfalto. Ameghiniand 39(4\:15R-16R. Rich, T. H., P. Vickers-Rich, O. Gim6nez, R. Crineo, P. Puerta, and R. Yacca. 1,999. A new sauropod dinosaur from Chubut Province, Argentina. In Y. Tomida, T. H. Rich, and P. Vickers-Rich, eds., Proceedings of the Second Gondwanan Dinosaur Symposium, 61-84. Na-
tional Science A,{useum Monograph, no. 15. Tokyo: National Science Museum.
Salgado,
L. 1993. Comments on
Chubutisaurus insignis
Del Corro
(Saurischia; Sauropoda). Ameghiniana 30:26 5-270.
1996. Pellegrinisaurus pouelli nov.gen. et sp. (Sauropoda, Titanosauridae) from the Upper Cretaceous of Lago Pellegrini, norrhwestern Patagonia, Argentina. Ameghiniana 33: 355-365. 1999. The macroevolution of Diplodocimorpha (Dinosauria; Sauropoda): A developmental model. Ameghiniana 36: 203-216. 2000. Evoluci6n y paleobiologia de los saur6podos Titanosaundae. Ph.D. dissertation, Universidad Nacional de La Plata. Salgado, L., and C. Azpilicueta. 2000. Un nuevo saltasaurino (Sauropoda, Titanosauridae) de la Provincia de Rio Negro (Formaci6n Allen, Cretdcico Superior), Patagonia, Argentina. Ameghiniana 37: 259264. Salgado, L., and J. F. Bonaparte. 1991.. Un nuevo saur6podo Dicraeosaundae, Amargasdurus cazctul gen. et sp. nov. de la Formaci6n La
Amarga, Neocomiano de 1a Provincia del Neuqu6n, Argentina. : 333-34 6. Salgado, L., and J. O. Calvo. L992. Cranial osteologv of Amargasaurus cazaui Salgado y Bonaparte (Sauropoda, Dicraeosauridae) from the Ame gh iniana 28
Neocomian of Patagonia. Ameghiniana 29: 337-346. 1,997. Evolution of titanosaurid sauropods. II: The cranial evidence. Ame gh iniana 34 : 33-48. Salgado, L., R. A. Coria, and.f. O. Calvo. 1997. Evolution of tiranosaurid sauropods. I: Phylogenetic analysis based on the postcranial evidence. Ameghiniana 34: 3-32. Salgado, L., and R. D. Martinez. 1.993. Relaciones filogen6ticas de los titanosduridos basales Andesaurus delgadoi v Epachthosaurus sp.
Ameghiniana 30 339.
Sanz, J. L., J. E. Powell, J. Le Loeuff, R.
D. Martinez, and X. Pereda Suber-
biola. 1999. Sauropod remains from the Upper Cretaceous of Laiio (northcentral Spain): Titanosaur phylogenetic relationships. Estudios
del Museo de Ciencias Naturales de Alaua 14 (Niim. Esp. 1): 235-25-5. Sereno, P.C. 1998. A rationale for phylogenetic definitions, with applica-
tions to the higher-level taxonomy of Dinosauria . Neues Jahrbuch fiir Geologie und Paliiontologie Abhandlungen 2I0: 4i-83. Sereno, P. C., A. L. Beck, D. B. Dutheil, H.C.E. Larsson, G. H. Lyon, B. Moussa, R. \J7. Sadleir, C. A. Sidor, D. J. Varricchio, G. P. \filson, and J. A. Xfilson. 1999. Cretaceous sauropods from the Sahara and the uneven rate of skeleral evolution amons dinosaurs. Science 286 1.342-1347.
452 . Leonardo
Saleado and Rodolfo A. Coria
Sciutto, J. C., and R. D. Martinez. 1994. Un nuevo yacimiento fosilifero de la Formaci6n Bajo Barreal (Cret6cico Tardio) y su fauna de saur6podos. Ndturalia Patag6nica, Ciencias de la Tierra 2: 27-47. Sim6n, E.2001. A giant sauropod from the Upper Cretaceous of El Choc6n, Neuqu6n, Argentina. Ameghiniana 38, supp. 19R: 19. Sim6n, E., and J. O. Calvo. 2002. Un primitivo titanosaurio (Sauropoda) del Choc6n, Formaci6n Candeleros (Cenomaniano temprano), Neuqu6n, Argentina. Ameghiniana 39, supp. 77R: L7. Smith, J. B., M. C. Lamanna, K. J. Lacovara, P. Dodson, J. R. Smith, J. C. Poole, R. Giegengack, and Y. Attia. 2001. A giant sauropod dinosaur from an Upper Cretaceous mangrove deposits in Egypt. Science 292:
1704-1706. H.-D., and
Sues,
trial
S.
L. 'Wing. 1,992. Mesozoic and Early Cenozoic terres-
ecosystems.
In
A.
K.
Behrensmeyer, J.
D. Damuth, \f. A.
DiMichele, R. Potts, H.-D. Sues, and S. L. 'Wing, eds., Terrestrial Ecosystems through Time, 32741,6. Chicago: University of Chicago Press.
Tidwell, V., K. Carpenteq and S. Meyer. 2001. New Titanosauriform (Sauropoda) from the Poison Strip Member of the Cedar Mountain Formation (Lower Cretaceous), Utah. In D. H. Tanke and K. Carpen-
te!
eds., Mesozoic Vertebrate Life, 139-165. Bloomington: Indiana University Press. Upchurch, P. 199 5 . The evolutionary history of sauropod dinosaurs. P/rilosophical Transactions of the Royal Society of London 349:365-390. Upchurch, P. 1999. The phylogenetic relationships of the Nemegtosauridae (Saurischia, Sauropoda). lournal of Vertebrate Paleontology 79:1.06-1.25. \fedel, M.J., R. L. Cifelli, and K. Sanders. 2000. Osteologn paleobiology, and relationships of the sauropod dinosaur Sauroposeidon. Acta Paleontologica Polonica 45: 343-388. 'Wilson, J. A. 2002. Sauropod dinosaur phylogeny: critique and cladistic analysis. Zoological Jotrrnal of tbe Linnean Society 136:2I7-276.
Wilson, J. A., and M. T. Carrano. 1,999. Titanosaurs and the origin of "wide-gauge" trackways: A biomechanical and systematic perspective on sauropod locomotion. Paleobiology 25: 252-267. Wilson, J. A., and P.C. Sereno. 1998. Early evolution and higher-level phylogeny of sauropod dinosaurs. Society of Vertebrare Paleontology Memoir, no. 5. Journal of Vertebrate Paleontology 18 (supp. to no. 2): 1-68.
Sauropods of Patagonia
.
453
20. Observations on Cretaceous Sauropods from Austraha RerpH E. MoTNAR AND SrpvpN \7. SerrsBURY
Abstract The Cretaceous record of Australian sauropods from the Albian and Cenomanian of Queensland and New South \Wales includes possibly five taxa, most of which are titanosauriforms. Although some pertain to the Titanosauria, one seems to be a brachiosaurid. Furthermore, some evidence suggests that a non-titanosauriform was also present. No relict sauropod taxa can be presently substantiated in the eastern Australian Cretaceous.
Introduction Sauropod material was first collected in 1913 in Queensland, Australia, from Blackall in the east-central part of the state. Addi-
tional material was sporadically collected in the early 1930s to the late 7970s (Molnar 2000). This material was described by Longman (1933) and Coombs and Molnar (i981). Further sauropod material (not discussed here) has been recently found in central Queensland (Salisbury 2002). So far, all eastern Australian sauropod material (Table 20.1) derives from Albian and Cenomanian 4s4
TABLE 20.1. Australian Cretaceous Sauropod Material Discovered prior to Mid-2000 Specimen
Locality
Stratigraphic unit
Material
QM F6142
Stewart Ck., "Dunraven," near Hughenden, north-central
Toolebuc Fm., Albian
incomplete cervical
QM F13712 QM F40347
near Stewart Ck., "Dunraven"
Toolebuc Fm., Albian ?, Albian?
worn caudal centrum
Allaru Mudstone, Al-
at ieast 7 incomplete
bian
dorsals, rib fragments
unrecorded
6 proximal caudals, middle and 1 distal
Queensland
QM F2316
"Silver Hills," near Richmond, north-central Queensland "Clutha," near Maxwelron, north-central Queensland
QM F2470
unrecorded
QM F3390
"Alni," near \finton, central
QM F6737
QMF7291
QM F7880 QMF7292
l7inton Fm., terminal Aibia n- Cenoma
Queensland
n ia
n
"Lovelle Downs," near ITinton, central Queensland
t{/inton Fm., terminal
"Lovelle Downs"
\Tinton Fm., terminal
Al bian- Cenomani an
Al bia
n - Cenoma
nian
"Elderslie," near 'Winton, cen-
I(/inton Fm., terminal
tral Queensland
Albian- Cenoma n ia n 'Winton Fm., terminal Albian- Cenomania n
"Elderslie"
distal humerus
1
caudal centrum humerus, metacarpals, femur dorsal pieces and proximal and medial caudals, scapula, metacarpals, ischium dorsals, ulna?, meta-
carpal, femur coracoid, incomplete femur dorsal? pieces and proximal caudals, rib fragments, scapula, hu-
meri, ulnae, radii, 4
QM F10916
metacarpals
Chorregon, central Queensland
'$Tinton
Blackall, central Queensland
\Tinton Fm.?, termi-
Fm.?, termi-
nal Albian-
3 proximal and 1 distal caudals
Cenomanian?
QM
F311
incomplete humerus
nal AlbianCenomanian?
QM L380
"Bymount," near Surat, southeast Queensland
AM F66769 AM F66770
Lightning Ridge, northern New South Sflales Lightning Ridge ? (see text)
Griman Creek Fm., Albian Griman Creek Fm., Albian Griman Creek Fm.?,
incomplete ischium isolated tooth isolated tooth
Albian?
(not collected
)
Dampier Peninsula, near Broome, Western Australia
Broome Sandstone, Neocomian
trackways
Observations on Cretaceous Sauropods from Australia
.
455
Fig.20.1. Lateral uiew of one of th e
t'
ert e b ra e
o/ Austrosaurus
mckillopi (QM F2316) showing the btses of seueral laminae. (A) restotittiotl of dorsal, based on Lottgntut's specimen A (modified lirttn \Iolnar,2000). (B) Diagram ol t ettebra, to shou/ laminae. (C) \ e,'r.,pr-,r. leuersed for companson rt t:i 1. f hbret'ialions: acdl = ,7
1".', :' ) 1
Li;::i:-i:
ce
n tr o
,tCPl =
diap op hy seal
, ::.:r,ietl b), matrix; pcdl
=
=
didpop b),se,:tl i-r,'::ti.t tanterior end onlr- seen m c
entr
o
;:,!ebra on right); podl? l'',-.-.ible anterior end
!,'s t )'
go diap op
h y
se
of
=
al lamina;
p,.l I = prezy godiap op hy seal i.i ttt r n a ; sp dl = sp ino diap op hy seal
l:nrirta; x = unndmed ridge c otute cting uith anterior
b y s eal I amina. ing indicates br o ken bone, uertical batcbing indicates nntrtx. Scale bar: 5 cm.
t
o
etttr
Cr
o
s
be re-examined.
Institutional dbbreuiations. AM-Australian Museum, Sydney; DGM-Museum of the Divisio Geologia e Mineralogia, Departamento Nacional de ProduEio Mineral, Rio de Janeiro; and QMQueensland Museum, Brisbane.
dnterior
tet::,' ,L.trtPoPhyseal lamina; d p,;5;1,tia;l position of missing .i;;r' pl:r -si-s' P = pleurocoel,
I .-(:.ror
units (Dettmann et a|.7992). Previously, these specimens were referred to the Cetiosauridae (Longman 1933 Coombs and Molnar 1981). Recent work, however, indicates that most or all of this material pertains to titanosauriforms (Salgado 1993; Salgado and Calvo 7997; Molnar 2000). This contribution presents the results of work on this material subsequent to Coombs and Molnar (1981) and Molnar (2000), of particular significance in regard to the taxonomic conclusions reached in those papers. In addition, although no evidence was available to indicate that more than a single taxon was represented among the Queensland sauropod material previously studied, other material, new observations, and recent studies (e.g. \Tilson 2002) indicate that this issue needs to
diap op
s-h atch
Description
Albian material. Austrosaurws mckillopi is the only named sauropod from the eastern Australian Cretaceous (Longman 1933). It is represented mainly by incomplete dorsal vertebrae (QM F231,6) from the Allaru Mudstone, near Maxwelton, north-central Queensland. Longman's (1933) description mentioned three individual blocks, but in June 1933, after his publication, five large and more than ten small further pieces were received, thus at least eight vertebrae are known. One of these more recently discovered specimens, a dorsal vertebra (Fig.20.1) has the most complete laminae of this specimen. The incompleteness of the material makes identification of the laminae difficult as the nomenclature is based on the structures connected by the laminae (\Tilson 1,999). Furthermore, it
-\;pcdr a:\ 1\,
l\
456
.
Ralph E. Molnar and Steven 'i7. Salisbury
is unclear which end of the specimen is anterior. Longman's specimen A retains the ventral portions of three laminae (Longman 1933, pl. 15), which appear to be the anterior centroparapophyseal lamina (acpl) anteriorly, the anrerior centrodiapophyseal lamina (acdl) behind it (weathered to a low ridge), and the posterior centrodiapophyseal lamina (pcdl) angled into the posterior break (Fig. 20.1). These identifications, although plausible, are renrarive because the dorsal ends of the laminae are not preserved. These laminae seem to have been shallow, although the amount of weathering makes this impossible to confirm. Comparing the "new" vertebra with the laminar pattern of specimen A, and assuming that the identifications rhere are correct, suggests that the right side is seen in Figure 20.1, and the laminae present are the anterior and posterior centrodiapophyseal laminae, the prezygodiapophyseal lamina (prdl), and base of the spinodiapophyseal lamina (spdl) and possibly the anterior end of the postzygadiapophyseal lamina (podl). The posterior centrodiapophyseal lamina is represented only by its anterior termination at the (nolv lost) diapophysis. These observations support the phylogeneric placement of A. mckillopi as a member of the Titanosauria (Molnar 2000). Unidentifiable titanosaurid material includes QM F2470, consisting of six proximal caudals preserved as three pairs, rvith one middle and one distal caudal. The specimen came from an unrecorded locality probably in north-central Queensland (Molnar 2000). That six of the vertebrae were found as pairs indicates that this specimen was at least partially articulated when buried (Figs. 20.2,20.3,).In addition the chevrons remain in articulation. The general form of the vertebrae are shown in Figures 20.2 and 20.3. The neural arches are positioned anteriorly on rhe anterior caudals, and hence presumably also on the middle and posterior caudals. Neither laminae nor pleurocoels are present. The centra are amphicoelous and. in anrerior view. taper venrrally. The rario of rheir Iength to their height is about 0.6. At breaks, cancellous internal structure is revealed, but not the spongy bone found in eM F23I6 and QM F6737. The proximal articular facets of the chevrons are not joined by horizontal bars of bone. The middle caudal cenrrum is clearly depressed with dorsoventrally convex sides, the distal centrum less so. The ventral surfaces of the caudals are shallorviy ex, cavated, with each fossa bounded by a venrrolareral ridge on each side. Assuming rhese all represent a single individual, which is consistent with their relative sizes, the lengths of centra increase posteriorly. Molnar (2000) conciuded on rhe basis of the caudal vertebral character states that this specimen represents a titanosaurid. The analysis of Wilson (2002) would place this as probably a mem-
Lig. 20.2. Articulated anterior caudals of the titanosaurid indet.
(QM F2470), probably from near Ric b mond, nortb - central
Queertsland (see text), with cheuron. Scale bar: 5 cm.
ber of the Titanosauria from the presence of ventral longitudinal sulci on the anterior and middle caudals and the absence of forked chevrons.
QM F40347 is the distal part of a left humerus (Fig. 20.4), from the Toolebuc Formation, near Richmond, north-central Queensland. The radial condyle is prominent, but the ulnar condyle Observations on Cretaceous Saulopods from Australia
.
457
H
H Fig. 20.3. Associated dtfierior caudals of the titanosaurid indet.
(QM F2170) that are
less
displaced than those of Figure 20.7. Left, posterior uiew; right,
right latrrol uirw..\calc hnr: I cnt.
is not apparent anteriorly. Both condyles are prominent posteriorly,
with a well-developed olecranal fossa extending from the distal end. The bone flares distalll', so that the maximum width across the epicondyles is more than 1.5 times the transverse diameter at the break. The distal end is flat, presumably perpendicular to the long axis of the shaft. Although previous workers (Upchurch 1995, 1998; Salgado and Calvo 1.997; Salgado et al. 1997; 'Wilson and Sereno 1998) do not use humeral characters or only those of the 'S7ilson (2002) recognizes two from the distal end proximal region, of the humerus. One of the condyles is clearly anteriorly exposed in this specimen, and is a derived character state according to \Tilson
(2002), and the distal margin, at least in anterior and posterior vielvs, is flat. The latter character diagnoses sauropods, but the former, diagnoses an unnamed clade of the Titanosauria (including nemegtosaurids and saltasaurids).
Trvo teeth, one (AM F66769) certainly and one (AM F66770)
458 . Ralph
E. Molnar and Steven
!/.
Salisbury
Fig.20.1. The distal left humerus uI an indelcrnrtnatc litdnosaurian (QM F10347) found near Rich mond, north -c entral Quccnslarrd. !A, Ct onterior t'iL'u ; (8, D) posterioT ltist!; (E) distal uiew; (F) medial uiew. Scdle bar: 20 cm.
Fig.20.5. Sauropod teeth from the Griman Creek Fonnation dt Lightning Ridge, nrtrthern New South Wales. (A) Titanosauriform indet. (AM F66769) in mesiodistdl and lingual uiews; (B)
@ g
E E
x g
J"
T itanosaur iiorrn ( ) in det. ( AM F66770) in mesiodistal and Lingual uieus. Scale bar: I cm. ?
probably from the Griman Creek Formation (Albian) at Lightning Ridge, New South '$7ales, are ritanosauriform. AM F66770 rncludes the crown and upper part of the root, about two-thirds as long as the crown (Fig.20.5B). AMF66769 comprises the crown and upper part of the root, about as long as the crown (Fig. 20.5A). The tips of both are incomplete, presumably from wear. Both teeth are preserved in miiky, pale blue and nonprecious green opal
("potch")
as pseudomorphs.
Observations on Cretaceous Sauropods from Australia
.
459
Tlre upper half of the crown of AM F66770, in mesiodistal view, is flexed lingually at about 25o, and the labial margin of the crown is bulbous and more strongly curved than the lingual (Fig. 20.5B). This labial face is strongly convex, and the broad lingual face is concave with a slightly curved, strong central ridge. The labial face erhibits shallow grooves along the mesial and distal margins. The crown has an almost semicircular form in section. A distinct neck, marked lingually by a rvell-defined narrow groove, joins the crown to the cylindrical root. There is no indication of marginal denticles. A mesiodistal view shows 80% of the crown of
AM
F66769 is flexed lingually at about 20" (Fig. 20.5A). The crown is rounded polygonal in section, with a flattened, but narrow, lingual face. The neck is larger than the crown, but abruptly constricts to the faceted root. As in AM F66770, mar:ginal denticles are absent. In AM F66770 the wear facet at the tip, viewed mesiodistally, is inciined at about 45o to the labio-lingual axis. Salgado and CaIvo (1997) suggest that such wear is synapomorphic for the titanosauriforms. The crown of AM F66769, although narrow, is not "pencil-1ike" in the sense used to describe the teeth of Diplodocus. Applying the characters of Upchurch (1998), the "parallel-sided" form of the crown (K2) indicates that it derives from the radiation including Brdcbiosaurus, Lapparentosdurus, P huwian gosaurus, and the titanosaurians, in other words, the Titanosauriformes of Wilson and Sereno (1998).It is possible that both teeth (AMF66769 andF66770) derive from the same taxon, but there is no evidence to support this and, if so, this sauropod would have had an unusually wide range of variation in tooth form. A single, incomplete cervical vertebra (QM F6142) from the Toolebuc Formation, near Hughenden, in north-central Queensland, may represent a brachiosaurid titanosauriform (Fig. 20.6). The vertebra is represented only by its posterior portion. This specimen was mentioned by Coombs and Molnar (1981) and figured by Molnar (1991). The deepiy concave, dorsoventrally compressed, posterior, central articular face is ventrally and laterally flared. At the anterior break, the centrum is deeply excavated ventrally and edged with obliquely descending ridges. The spinopostzygapophyseal lamina is inclined at about 45" in lateral view (assuming the posterior, central articular face to be vertical). The postzygapophyseal facets are inclined at about 25' to the horizontal. The postzygapophyseal processes are separated by a deep fossa. The deep pleurocoel is divided by an almost horizontal lamina. The posterior end of the pleurocoel is beneath the postzygapophyseal facet.
Two character states of Upchurch (1998) are present, deep pleurocoels in cervical centra and cervical vertebrae with concave ventral surfaces. The infradiapophyseal lamina system is probably present on cranial and middle cervical vertebrae (Upchurch's infradiapophyseal lamina corresponds to Wilson's [1999] posterior centrodiapophyseal lamina). The presence of accessory oblique lamina in cervical pleurocoeis is unclear, because the lamina in QM F6742 460 . Ralph E. Molnar and Steven
S7.
Salisbury
Fig. 20.6. The posterior pdrt of the ceruical of a ?brttchiosaurid indet. (QM F61-+2), from near Hu gh end erq north - c erir al Qtteensland: (A) right lateral uiett'; ,B,lcft Itteral t icu': rCt pr6!svir'v
uieu; (D) tieu' ctf the anterior ltrtak. lrr D. tlsc cradle srrfporring the specimen can be seen beneath the centrum. Scale bar: 10 cm.
l_d is oriented quite differently from that figured by Upchurch (1998, fig. 8). However, a similar horizontal lamina is present in the last cervical of Brachiosaurus brancal (Janensch 1950, fig. 49). Three '!Tilson states of and Sereno (1998) are clearly or plausibly present: cervical centra opisthocoelous; mid-cervical neural arches deep, greater than centrum diameter; and cervical pleurocoels divided.
'Wilson
(2002) diagnoses the Titanosauria as having undivided cer-
vical pleurocoels, which indicates that this specimen does not derive from a titanosaurian because the centrum has complex pleurocoels divided by bony septa (83). Neither has this specimen the bifid neural spine characteristic of diplodocids and dicraeosaurids. This, combined with the results of Upchurch (1998), suggests that this cervical represents a brachiosaurid.
Late Albian-Cenomdnian mdterial. Several specimens, almost
all described and figured b.v Coombs and Molnar (1981),
have
been recovered from the'Winton Formation near \il/inton. central
Observations on Cretaceous Saurooods fron-r Australia
.
461
Queensland. These have been generall,v attributed to Austrosaurus sp. QM F3390, consisting of a humerus and femur, both represented by proximal and distal ends not sharing a contacr, and the proximal ends of three metacarpals (Coombs and Molnar 1981, pls. 3 and 6). The articular surfaces, especially those of the proxrmal part of the humerus, are probably the best preserved of the Queensland sauropod material. The form of the prorimal end of the humerus is not matched by any of those figured by Mclntosh (1990, fig. 16.10). The well-developed internal tuberosity, or proximomediai corner, is characteristic of sornphospondyls (Wilson 2002). The proximal bulge of the lateral margin of the femur is considered by Salgado et a|. (1997; \filson 2002) as indicaring a member of the Titanosauriformes. Several specimens are ritanosaurian, but also represent a single taxon or closely related taxa. A series of incomplete dorsais, proxrmal and middle caudals, pieces of ribs, an incomplere scapula, and the proximal part of an ischium are from one individual (QM F6737). Oddln although anterior parts of the dorsal centra were collected, there is no indication of any posterior portions. Three roughly discoid pieces of bone are taken to represent anterior portions of dorsal centra. These pieces have a shallowly convex arricular surface forming one side of the "disk," the orher being a broken surface. This break shows evidence of pleurocoels (character 8 of Salgado et aL. 1997; character 68 of Wilson and Sereno 1998), between rvhich is spongv bone (Fig. 20.7) (character 102 of ril/ilson and Sereno 1998). The conver articular faces and the pleurocoels are taken to indicate that these pieces represent parts of opistho-
coelous dorsal centra (characters 1 and 9 of Salgado et al. 1997\. Opisthocoelous dorsals are also knorvn for QM F2316. QM F6737 and QM F10916 have centrum length divided by centrum height approximately 0.5-0.6 (character C27 of Upchurch 1998). More imporrantll', rhe ventral surfaces of cranial caudal centra of eM F6737 and QM F2470 are mildly excavated, with the excavarion bounded by a venrrolateral ridge on each side (character Q5 of Upchurch 1998). The centra of middle caudals display a dorsoventrally compressed transverse cross-section (character p1 of Upchurch 1998), and the neural arches are positioned anteriorl.v in mid- and posterior caudal centra (character 1S of Salgado et al. 1997\. Two specimens were recovered from "Elderslie" (Coombs and Moinar 1981). QMF7292 is the most complete of the eueensland Cretaceous sauropod specimens, including pieces of ribs, seventeen middle and distal caudals, an incomplete scapula, substantial parts of both humeri, both radii and both uinae, and four metacarpals (Coombs and Molnar 1981, pls. 7-4,6). The characters indicate a titanosaurian (\Tilson 2002). Expansion of the distal end of the radius to twice the midshaft diameter characterizes a clade within the Titanosauria. The second specimen from "Elderslie," QM F7880, is substantially less complete than QM F72792.It consists onlr. of an incom-
462
.
Ralph E. Molnar and Steven \X/. Salisbury
xrl
t6
\1
t*'**-s
#.* \B
-.,-t
r)r), n rr rJ t0
*t r,
t':..g
r b{J{ r
nryAq
Fig. 20.7. Anterior part of a dorsal centrutn of an indeterminate titanosaurid (QM F5737), from "Louelle Dotuns," nctr Winton, cortral Qtteensltnd, together u,ith the same pdrt of Austrosaurus mckitlopi (QM F2316, holot.^,,pe). Aboue, right lttcrtl uicws: beloru. I,re'ken [aces, seen from behind. In the lateral uiews, the arrous mark the anterior margins of the pleurocoels. In the posterior uiews, tbe curued drrows mark the right pleurocoels, and the broad arroa,s indicate the positictn of the nettral canol ,fillcd with tnatrix in QM I-2316). ln QM F6737 the surficial bone remains, btx in QM F2llo tt bds hccn Iost. leauing only the calcdreous fill of the interndl cduities. The spctng; internal structure of the centrum rntt\ be seen in hoth specimens, Scole bar: 5 cm.
plete femoral head and an incomplete coracoid. Although the femoral head is uninformative, the coracoid is elongate. An elongate coracoid, about twice the length of the scapular contact surface, is characteristic of a clade u'ithin the Titanosauria (Wilson 2002).
QM F10916 is from Chorregon, central Queensland, and is most probabl,v frorn the \Tinton Formation. Three incomplete proximal caudals, and one more distal caudal, make up this specimen (Fig. 20.8). A1l the centra are amphicoelous and show some degree of constriction. The ventral surfaces of the anterior elements bear longitudinal sulci, bounded by prominent ridges, and are probabll' Tiranosauria. Other sdtrropod materia/. A specimen from "Lovelle Downs," QMF7291, consists of a metacarpal, distal femur; and unidentified fragments (Coombs and Molnar 1981, pls.5 and 6), one of which was suggested by Coombs and Molnar (1981, pl.5C, D) to possibly be the proximal end of an ulna. The distal surface of metacarpal I has two condyles. The preserved part of the femur of this specimen appears comparable in form to the distal part of that of Q\,I F3390, but is about one-third larger. Thus this specimen is one of the larger of the Queensland sauropods. The existence of two condyles, on the phalangeal articular face of metacarpal I, suggests that this specimen is not a titanosauriform (Wilson2002). Observations on Cretaceous Sauropods from Australia
.
463
Fig. 20.8. A distal caudal of a titanosauridn indet. (QM F10916), from near Chorreg,ttt, central Q.ueensland, in presumed left lateral (left1 and dr,;rsal (right) uietus. Scale bar: 70 cm,
l-t Conclusions Is more than a single taxon represented bv the eastern Australian Cretaceous sauropod material? The transverse processes of the proximai caudal are substantially deeper in QM F7292 than in
QM F2470, QM F6737, or QM F70976. Furthermore, the centra of QM F7292 are noticeably broader than those of the other three specimens. This may be merely due to different verrebrae being preserved in these four specimens, bur it could also indicate taxonomic difference. In QM F2470 and QM F6737 the lengths of the caudals increase posteriorly, but in QM F7292 they decrease posteriorly. The evidence is incomplete, but the difference in trends in lengths of the caudal centra, matched by the difference in proximal caudal form in the same specimens, suggest that at least two sauropod taxa are represented in the Winton material. There appear to have been nvo sauropod taxa during the pre-XTinton time (Albian), Azstrosdurus and a possible Brachiosauridae, and probably three dur'Winton ing the tirne, two of them titanosaurians. The argument of Moinar (2000) and the evidence presented here shows that Australia can no longer be regarded as the only Gondwanan continent without titanosaurids, and that there is no evidence for relict, plesiomorphic sauropods in the Australian Cretaceous. But Australia does seem to lack any advanced titanosaurids with procoelous caudals, at least through the Cenomanian. Acknctwledgments. \We particularly appreciate assistance during various aspects of this project by Graham Anderson (Lightning Ridge), Laurie Beirne (Queensland Museum), Sandy Swift (Northern Arizona University), Tony Thulborn (Monash University), Mary \7ade (then at the Queensland Museum), and Zhou X.-T. (Academia Sinica). References Cited
Coombs,
'W.
P., Jr., and R. E. Molnar. 1981. Sauropoda (Reptilia,
from the Cretaceous of Queensland. Memoirs of the Queensland Museum 20: 351-373. Dettmann, M. E., R. E. Molnar, J. G. Douglas, D. Burger, C. Fielding, H. T. Clifford, J. Francis, P. Jell, T. Rich, \,{. Sfade, P. V. Rich, N. Pledge, A. I(emp, and A. Rozef elds. L992. Australian Cretaceous terSaurischia)
464 . Ralph E. Molnar and
Steven
\7. Salisburv
restrial faunas and floras: Biostratigraphic and biogeographic implications. Cretdceous Research 13: 207 -262. Janensch, !7. 1950. Die \Tirbelsdule von Brachiosdurus brancai. Palaeontographica, supp. 7(3) 2:27-93. Longman, H. A. 1933. A new dinosaur from rhe Queensland Cretaceous. Memoirs of the Queensland Museum 1,0: 131,-144. Mclntosh, J. S. 1990. Sauropoda. In D. B. Veishampel, P. Dodson, and H. Osm6iska, eds., The Dinosauria, 345-401. Berkeley: University of
California
Press.
Molnar, R. E. 1991. Fossil reptiles in Australia. In P. Vickers-Rich, J. M. Monaghan, R. F. Baird, T. H. Rich, E. M. Thompson, and C. 'Williams, eds., Vertebrate Pdlaeontology of Australasia, 605-702. Melbourne: Pioneer Design Studio and Monash University Publications Committee.
2000. A reassessment of the phylogenetic position of Cretaceous sauropod dinosaurs from Queensland, Australia. In H. A. Leanza, ed., VII International Symposium on Mesozoic Terrestrial Ecosystems, 139-144. Asociacion Paleontologica Argentina Publicacion Especial, no. 7. Buenos Aires: Asociacion Paleontologica Argentina. Salgado,
L. 1993. Comments of
Chubutisaurtts insignis
Del Corro
(Saurischia, Sauropoda). Ameghiniana 30: 265-270. Salgado, L., and J. O. Calvo, 1997. Evolution of titanosaurid sauropods. II: The cranial evidence. Ameghiniana 34:33-48. Salgado, L., R. A. Coria, and J. O. Calvo. L997. Evolution of titanosaurid sauropods. I: Phylogenetic analysis based on the postcranial evidence.
Ameghiniana 34:3-32. Salisbury, S. 2002. Clash of the titans. Nature Australia 27(7): 44-51,.
Upchurch, P. L99 5 . The evolutionary history of sauropod dinosaurs. Pbllosophical Transactions of the Royal Society, London 349 365-390. 1998. The phylogenetic relationships of sauropod dinosaurs. Zoological Journal of the Linnean Society 724 43-703. 'Wilson, J. A. 1999. A nomenclature for vertebral laminae in sauropods and other saurischian dinosaurs. Journal of Vertebrate Palectntoktgy
19 639-653.
2002. Sauropod dinosaur phylogeny: Critique and cladistic analyJournal ofthe Linnean Society 136 215-275. 'Wilson, A., and P. C. Sereno 1,998. Early Euolution and Higher-Leuel J. Phylogeny of Sauropod Dinosaurs. Society of Vertebrate Paleontology Memoir, no. 5. Chicago: Society of Vertebrate Paieontology. sis. Zoological
Observations on Cretaceous Sauropods from Austraiia
.
465
2t. Late Cretaceous (Maastrichtian) Nests, Eggr, and Dung Mass (Coprolites) of Sauropods (Titanosaurs) from India D. M. MoHeeEy
Abstract Late Cretaceous dinosaurs from India are represented by at least twenty species of sauropods, theropods and, ornithopods. The dinosaur fauna is dominated by a titanosaurid and abelisaurid as'Well-preserved
dinosaur nesting sites, clutches, and eggs are abundant in the dinosaur skeleton bearing Upper Cretaceous Lameta sediments in central and western India. No dinosaur eggs are so far known from the pre-Cretaceous sediments in India. A majority of the Indian Late Cretaceous eggs have been assigned to oofamily Megaloolithidae, believed to be of titanosaurs. At least eight Megaloolithus oospecies have been established, Nesting and social behavior of the titanosaurs is inferred from eggs, nests, and nesting sites. Evidence suggests community nesting along the riverbank with the eggs buried in the river sand. Recently, weil-preserved, large-sized (diameter up to 100 mm) coprolites containing undigested plant tissues, pollen grains, spores, and other ingested organic matter have been found. The coprolites occur in association with skeletal remains of Titanosaurus indicus, T. blandfordl, and pelomedusid turtles. Based on the large semblage.
466
size of the coprolites, their association with titanosaur skeletal remains, and the occurrence of prolific plant tissues in the fecal mass, the coprolites of Type-A are assigned to titanosaurs. The floral analysis of the coprolitic mass has provided insight into the dietary habit of the titanosaurs. The evidence suggests that titanosaurs preferred cropping the soft tissues of higher plants such as pteridophytes, gymnosperms, and angiosperms as their main solid diet. The study of the Indian sauropod eggs, nesrs, and coprolites has shown that the Indian Late Cretaceous ecosystem offered an ideal habitat for titanosaurs wherein they reached the acme of their breeding and nesting. The prolific occurrence of the skeletal remains, eggs, and dung mass of titanosaurs in the Maastrichtian ecosystem
of India and their total absence in the older sediments
suggests a sudden turnover during the Lameta time, and was a pre-
lude to their extinction just before the Cretaceous-Tertiary Boundary (KTB) with the advent of Deccan volcanic eruprion.
Introduction The earliest report of sauropod skeletal remains in India is by Hislop (1859). The material r,vas collected from the Upper Creraceous (Maastrichtian) Lameta Formation of Pisdura in central India (Fig. 21.1).ln contrast, the report of sauropod eggs is more recent, with the first discoveries in 1981 (Mohabey 1983) from the Lameta Formation of the Kheda area in Gujarat in western India (Fig. 21.1). The Late Cretaceous dinosaur fauna of India is represented by sauropods (titanosaurs), theropods (abelisaurids, allosaurids, coelurosaurs), and questionably identified ornithischians (sregosaur
Drauidosaurus blandfordi and Brachypodosaurus grauis\. Although Lydekker established the sauropod genus Titanosaurus in 1877 (Lydekker 7877,1879),little is known about the species because of the fragmentary narure of its skeletal remains. Most of the taxonomy of the Indian Titanosaurus species is based on vertebral systematics that form the diagnostic characteristics (Huene and Matley 1933). The named species are Titanosaurus indicus (Lydekker 1877), Titanosaurus blandfordi (Lydekker 1879), Antarctosdurus septentrionalis, Laplatasaurus madagascariensis (Huene and Matley 1933), and Isisaurus colberti (Jain and Bandyopadhya 1997; \X/ilson and Upchurch 2003). Not all of these species are considered valid (Jain and Bandyopadhyay 1997;Wllson and Upchurch 2003). Of all the sauropod specimens, the holotype of I. colberti represents the best species among the Indian titanosaurs described to date. Since the discovery of dinosaur eggs in India, a large number of dinosaur nesting sites with well-preserved eggs have been located in the Lameta Formation in western India (Mohabey 1984a, 1987; Srivastava et al. 1986; Mohabey and Mathur 1989) and central India (Sahni and Tripathi 1990; Mohabey 1990, 1996a; Khosla and Sahni 1995). Though the dinosaur skeletal remains are well documented from the pre-Cretaceous sediments in India, no eggs Late Cretaceous Nests, Eggs, and Dung Mass of Sauropods ftom India
.
467
/'
t,Joiaat..,
l
(.
\
-}., .,
t@".9 ,) i,^ -' ' .{.0.,^g3 _ ,1.,-, \llobd1.dq Bhopor @- S4; -' ;|l )' " Ft"j'i't1-:"r".*g g^r*:l'--oo'*',, a.o=,
AR o
Anror
A
trmoooiiT
,
AHA RASHTR4
I
oC6onorocur'''i'
.^.'€,9, (Molrri.l
;*"f,ff--';i-;"' r '\' I
\'-' i
--'i"-_-:t,{
IOO O l#
)t.
e
Nesting sates (Megoloolilhido. eggs of f itonosours)
uctrctriroirot ti
{9
(Elongotoolithidoe eggs of obclisourids)
r\
A single
H
Cretoceouo titonosours skclcton sitas
- -./ Kollom.du a AriYolu' .O
?a )rir
\v' jra MtL /
g
x Fig. 21.1. Geological map showing important dinosaur fossil Iocalities in lndia.
2OOKm
,: