Bulletin 45

New Mexico Museum of Natural History & Science

A Division of the
DEPARTMENT OF CULTURAL AFFAIRS

The Paleobiology of Coelophysis bauri (Cope)

from the Upper Triassic (Apachean) Whitaker
quarry, New Mexico, with detailed analysis of a
single quarry block

by
Larry F. Rinehart, Spencer G. Lucas, Andrew B. Heckert,
Justin A. Spielmann and Matthew D. Celeskey

New Mexico Museum of Natural History & Science

Printed with the support of the

Vortman Fund

Albuquerque, 2009
 STATE OF NEW MEXICO
Department of Cultural Affairs
Stuart Ashman, Secretary

NEW MEXICO MUSEUM OF NATURAL HISTORY AND SCIENCE

Hollis J. Gillespie, Executive Director

BOARD OF TRUSTEES
Bill Richardson, Governor, State of New Mexico, ex officio
Hollis J. Gillespie, Executive Director, ex officio
Gary Friedman, President
Peter F. Gerity, Ph.D.
Laurence Lattman, Ph.D.
Morton Lieberman, Ph. D.
Imogene Lindsay, Emerita
Viola Martinez
Nancy J. McMillan, Ph.D.
John Montgomery, Ph.D.
Joseph Powell, Ph.D.
Dennis P. Trujillo, Ph.D.
Steve West

Cover illustration: Coelophysis bauri growth curve and skeletal reconstructions. Reconstructions by
Matt Celeskey.

Associate editor for this bulletin: Robert M. Sullivan.

Original Printing

ISSN:1524-4156

Available from the New Mexico Museum of Natural History and Science, 1801 Mountain Road NW,
Albuquerque, NM 87104; Telephone (505) 841-2800; Fax (505) 841-2866; www.nmnaturalhistory.org

NMMNH Bulletins online at: http://econtent.unm.edu/cdm4/browse.php?CISOROOT=%2Fbulletins

 BULLETIN OF THE NEW MEXICO MUSEUM
OF NATURAL HISTORY AND SCIENCE

EDITORIAL STAFF

Spencer G. Lucas, Editor-in-Chief

Justin A. Spielmann, Managing Editor

EDITORIAL BOARD

Guillermo Alvarado Asociación Costarricense de Geotecnica, San José, Costa Rica

Andrea Arcucci Universidad Nacional de San Luis, San Luis, Argentina
Marco Avanzini Museo Tridentino di Scienze Naturali, Trento, Italy
David Berman Carnegie Museum of Natural History, Pittsburgh, PA, USA
John R. Foster Museum of Western Colorado, Grand Junction, CO, USA
Gerard Gierlinski Polish Geological Institute, Warsaw, Poland
Jean Guex University of Lausanne, Lausanne, Switzerland
Jerald D. Harris Dixie State College, St. George, UT, USA
Andrew B. Heckert Appalachian State University, Boone, NC, USA
Adrian P. Hunt Flying Heritage Collection, Everett, WA, USA
Hendrik Klein Neumarkt, Germany
Heinz Kozur Budapest, Hungary
Karl Krainer University of Innsbruck, Innsbruck, Austria
Martin G. Lockley University of Colorado at Denver, Denver, CO, USA
Spencer G. Lucas New Mexico Museum of Natural History and Science,
Albuquerque, NM, USA
Claudia Marsicano Universidad de Buenos Aires, Buenos Aires, Argentina
Jesper Milàn Institute for Geography and Geology, Copenhagen, Denmark
Gary S. Morgan New Mexico Museum of Natural History and Science,
Albuquerque, NM, USA
Donald R. Prothero Occidental College, Los Angeles, CA, USA
Silvio Renesto Università degli Studi dell’Insubria, Varese, Italy
Vincent L. Santucci National Park Service, McLean, VA, USA
Joerg W. Schneider Technical University BergAkademie of Freiberg, Freiberg, Germany
Jingeng Sha Nanjing Institute of Geology and Palaeontology, Nanjing, China
Edward Simpson Kutztown University, Kutztown, PA, USA
Justin A. Spielmann New Mexico Museum of Natural History and Science,
Albuquerque, NM, USA
Robert M. Sullivan State Museum of Pennsylvania, Harrisburg, PA, USA
Lawrence H. Tanner Le Moyne College, Syracuse, NY, USA
Dana Ulmer-Scholle New Mexico Tech, Socorro, NM, USA
Ralf Werneburg Naturhistorisches Museum Schloss Bertholdsburg, Schleusingen, Germany
Richard S. White, Jr. International Wildlife Museum, Tucson, AZ, USA
 NEW MEXICO MUSEUM OF NATURAL HISTORY AND SCIENCE BULLETINS
1. Meniscotherium (Mammalia, “Condylarthra”) from the Paleocene of western North America, 1992. Thomas E.
Williamson and Spencer G. Lucas, 75 pp.
2. Vertebrate paleontology in New Mexico, 1993. edited by Spencer G, Lucas and Jiri Zidek, 338 pp. (Out of Print)
3. The nonmarine Triassic, 1993. edited by Spencer G. Lucas and Michael Morales, 478p +59 pp. road log
4. Late Triassic (Carnian and Norian) tetrapods from the southwestern United States, 1995. Robert A. Long and Phillip A
Murry, 254 pp. (Out of Print)
5. Keuper (Triassic) flora from middle Asia (Madygen, southern Fergana), 1995. Inna A. Dobruskina, 49pp.+49pl.
6. Early Permian footprints and facies, 1995. edited by Spencer G. Lucas and Andrew B. Heckert, 301 pp. (Out of Print)
7. The postcranial morphology of Paleocene Chriacus and Mixodectes and the phylogenetic relationships of archontan
mammals, 1996. Frederick S. Szalay and Spencer G. Lucas, 47 pp.
8. The beginning of the age of mammals in the San Juan Basin, New Mexico: Biostratigraphy and evolution of
Paleocene mammals of the Nacimiento Formation, 1996. Thomas E. Williamson, 141 pp.
9. Paleontology of the Greenhorn Cyclothem (Cretaceous: Late Cenomanian to middle Turonian) at Black Mesa,
northeastern Arizona, 1996. James Ian Kirkland, 131 pp. + 50 pl.
10. Land snails of New Mexico, 1997. edited by Artie L. Metcalf and Richard A. Smartt, 145 pp.
11. New Mexico’s fossil record I, 1997. edited by Spencer G. Lucas, John W. Estep, Thomas E. Williamson, and Gary S. Morgan,
143 pp. (Out of Print)
12. Permian stratigraphy and paleontology of the Robledo Mountains, New Mexico, 1998. edited by Spencer G. Lucas, John
W. Estep, and Jerry M. Hoffer 130 pp.
13. Reanalysis of Acrocanthosaurus atokensis, its phylogenetic status, and paleobiogeographic implications, based on a
new specimen from Texas, 1998. Jerald D. Harris, 75 pp.
14. Lower and middle Cretaceous terrestrial ecosystems, 1998. Spencer G. Lucas, James I. Kirkland, and John W. Estep (eds.),
330 pp. (Out of Print)
15. A new skull of Parasaurolophus (Dinosauria: Hadrosauridae) from the Kirtland Formation of New Mexico and a
revision of the genus, 1999. Robert M. Sullivan and Thomas E. Williamson, 49 pp.
16. New Mexico’s fossil record II, 1999. edited by Spencer G. Lucas, 284 pp.
17. Dinosaurs of New Mexico, 2000. edited by Spencer G. Lucas and Andrew B. Heckert, 230 pp.
18. Volcanology in New Mexico, 2001. edited by L. S. Crumpler and Spencer G. Lucas, 150 pp.
19. Walter Granger, 1872-1941, paleontologist, 2002. by Vincent L. Morgan and Spencer G. Lucas, 58 pp.
20. Dinosores: An annotated bibliography of dinosaur paleopathology and related topics—1838-2001, 2002. by Darren H.
Tanke and Bruce M. Rothschild, 96 pp.
21. Upper Triassic stratigraphy and paleontology, 2002. edited by Andrew B. Heckert and Spencer G. Lucas, 301 pp.
22. Notes from Diary-Fayum Trip, 1907, 2002. by Vincent L. Morgan and Spencer G. Lucas, 148 pp.
23. Paleoecological analysis of the vertebrate fauna of the Morrison Formation (Upper Jurassic), Rocky Mountain
Region, U.S.A., 2003. by John R. Foster, 95 pp.
24. Paleontology and geology of the Upper Triassic (Revueltian) Snyder quarry, New Mexico, U.S.A., 2003. edited by Kate
E. Zeigler, Andrew B. Heckert, and Spencer G. Lucas, 132 pp.
25. Carboniferous-Permian transition at Carrizo Arroyo, central New Mexico, 2004. edited by Spencer G. Lucas and Kate E.
Zeigler, 301 pp.
26. Paleogene Mammals, 2004. edited by Spencer G. Lucas, Kate E. Zeigler, and Peter E. Kondrashov, 230 pp.
27. Late Triassic microvertebrates from the lower Chinle Group (Otischalkian-Adamanian: Carnian), 2004. by Andrew
B. Heckert, 170 pp.
28. New Mexico’s Ice Ages, 2005. edited by Spencer G. Lucas, Gary S. Morgan, and Kate E. Zeigler, 300+ pp.
29. Vertebrate Paleontology in Arizona, 2005. edited by Andrew B. Heckert & Spencer G. Lucas, 210 pp.
30. The Nonmarine Permian, 2005. edited by Spencer G. Lucas & Kate E. Zeigler, 362 pp.
31. Permian of Central New Mexico, 2005. edited by Spencer G. Lucas, Kate E. Zeigler & Justin A Spielmann, 176 pp.
32. Genética y Mamíferos Mexicanos: Presente y Futuro, 2006. edited by Ella Vázquez-Domínguez and David J. Hafner, 73 pp.
33. Skeletal Impact of Disease, 2006. by Bruce M. Rothschild and Larry D. Martin, 226 pp.
34. America’s Antiquities: 100 Years of Managing Fossils on Federal Lands, 2006. edited by Spencer G. Lucas, Justin A.
Spielmann, Patricia M. Hester, Jason P. Kenworthy and Vincent L. Santucci, 185 pp.
35. Late Cretaceous Vertebrates from the Western Interior, 2006. edited by Spencer G. Lucas and Robert M. Sullivan, 410 pp.
36. Paleontology and Geology of the Upper Jurassic Morrison Formation, 2006. edited by John R. Foster and Spencer G.
Lucas, 249 pp.
37. The Triassic-Jurassic Terrestrial Transition, 2006. edited by Jerry D. Harris, Spencer G. Lucas, Justin A. Spielmann, Martin G.
Lockley, Andrew R.C. Milner and James I. Kirkland, 607 pp.
38. Pennsylvanian-Permian Fusulinaceans of the Big Hatchet Mountains, New Mexico, 2006. by Garner L. Wilde, 331 pp.
39. Upper Aptian-Albian Bivalves of Texas and Sonora: Biostratigraphic, Paleoecologic and Biogeographic Implications,
2007. edited by Robert W. Scott, 39 pp.
40. Triassic of the American West, 2007. edited by Spencer G. Lucas and Justin A. Spielmann, 247 pp.
41. The Global Triassic, 2007. edited by Spencer G. Lucas and Justin A. Spielmann, 415 pp.
42. Cenozoic Vertebrate Tracks and Traces, 2007. edited by Spencer G. Lucas, Justin A. Spielmann and Martin G. Lockley, 330
pp.
43. The Late Triassic archosauromorph Trilophosaurus, 2008. by Justin A. Spielmann, Spencer G. Lucas, Larry F. Rinehart and
Andrew B. Heckert, 177 pp.
44. Neogene Mammals, 2008. edited by Spencer G. Lucas, Gary S. Morgan, Justin A. Spielmann and Donald R. Prothero, 442 pp.
 Rinehart et al., 2009, The Paleobiology of Coelophysis bauri. New Mexico Museum of Natural History and Science, Bulletin 45.

i
TABLE OF CONTENTS

Abstract 1

Introduction 2

History of the Whitaker quarry 2

Photographs of selected Whitaker quarry blocks 3

Turning over the NMMNH Whitaker quarry block 5

Preparation of the NMMNH Whitaker quarry block 8

Geology of the Whitaker quarry 21

Invertebrate fossils from the Whitaker quarry 27

Fishes from the Whitaker quarry 30

Tetrapods other than Coelophysis from the Whitaker quarry 32

Description and biometrics of the NMMNH Coelophysis material 41

Postcranial osteology of Coelophysis bauri 65

Vision characteristics of Coelophysis bauri 85

The furcula of Coelophysis bauri 96

Growth, allometry and population dynamics 100

Paleobiology of Coelophysis bauri 117

Acknowledgments 126

References 127

Appendix 1. 118 Coelophysis bauri datasheets 133

Appendix 2. Measurements and descriptive statistics of Shipingia n. sp. 255

Appendix 3. Test for statistical significance for two cohorts of Shipingia 256

Appendix 4. Sclerotic ring metrics 257

Appendix 5. Large element measurements of a Coelophysis bauri growth series 258

ii

Appendix 6. Tabulated survival data 259

Appendix 7. Compare two regressions: Coelophysis bauri skull length vs. dorsal series length 260

Appendix 8. Compare two regressions: Coelophysis bauri femur head width vs. femur length 260
 Rinehart et al., 2009, The Paleobiology of Coelophysis bauri. New Mexico Museum of Natural History and Science, Bulletin 45.

1
THE PALEOBIOLOGY OF COELOPHYSIS BAURI (COPE) FROM THE
UPPER TRIASSIC (APACHEAN) WHITAKER QUARRY, NEW MEXICO,
WITH DETAILED ANALYSIS OF A SINGLE QUARRY BLOCK

LARRY F. RINEHART1, SPENCER G. LUCAS1, ANDREW B. HECKERT2,

JUSTIN A. SPIELMANN1 AND MATTHEW D. CELESKEY1
1
New Mexico Museum of Natural History and Science, 1801 Mountain Rd. NW, Albuquerque, NM 87104-1375;
2
Department of Geology, Appalachian State University, ASU Box 32067, Boone, NC 28608-2067

Abstract— Here, we investigate the paleobiology of a single population of the Late Triassic dinosaur Coelophysis
bauri from the Whitaker quarry at Ghost Ranch, New Mexico. The quarry, discovered by George Whitaker of the
American Museum of Natural History in 1947, has produced about 30 large blocks of fossil material during two
extended excavations by the American Museum during the late 1940’s and early 1950’s, and by the Carnegie
Museum of Natural History during the 1980’s. At least a few hundred dinosaur skeletons are contained in these
blocks. We have studied numerous quarry blocks at several institutions, but we concentrate on the block at the
New Mexico Museum of Natural History (NMMNH).
The Whitaker quarry lies near the center of the Upper Triassic Rock Point Formation of the Chinle Group in
the Chama Basin of northern New Mexico, 30 m above the Upper Triassic Painted Desert Formation and 35 m
below the Middle Jurassic Entrada Sandstone. Based on lithostratigraphy, tetrapod biostratigraphy and conchostracan
biostratigraphy the quarry is of Apachean (late Norian-Rhaetian) age.
The NMMNH block was turned over and prepared from the bottom-up, thus revealing new details about the
paleoecology of the site. The quarry apparently preserves a single very large flock of Coelophysis bauri together
with other larger and smaller animals that were washed into a topographic low containing a small pond, where they
probably drowned and were buried by a sheet flood event from a nearby river. The quarry fauna comprises
ostracodes and conchostracans from the pond deposit, large coelacanth and small redfieldiid fish, apparently from
the river, and small through large reptiles represented by a sphenodontid and drepanosauromorphs (small),
dinosaurs and other archosauromorphs (medium) and phytosaurs (large). We list and describe the quarry fauna, but
the paleobiological analysis is confined to C. bauri.
Our research generally agrees (and occasionally disagrees) with some previous workers, but also breaks consid-
erable new ground. We find that Coelophysis bauri was a bipedal, carnivorous, theropod dinosaur that was a fast,
agile runner and a diurnal, visually-oriented hunter. Based on skull and orbit morphology, and on a complete
sclerotic ring present in one specimen, we assess the vision characteristics of C. bauri. Using bird and reptile
outgroups for comparison, particularly in the sclerotic ring morphology and the degree of binocular overlap of the
visual fields, we convincingly demonstrate that the eyes of C. bauri ally with those of hawks and eagles, and
probably had similar acuity, power of accommodation and depth perception. Thus, they were the eyes of a fast,
agile hunter of small, fast-moving prey.
Coelophysis grew quickly, in a more bird-like than reptile-like manner, and displayed numerous sexually
dimorphic characters, particularly in the skull and neck, and pelvis and sacrum. Although Coelophysis had many
bird-like features, including its overall appearance, hollow bones and the oldest documented furcula (wishbone), its
reproduction was probably more “reptilian” than bird-like. We demonstrate that Coelophysis probably invested
about as much effort in reproduction as modern reptiles of similar size, producing a large number of relatively small
eggs and providing at least some parental care. Yet at least on occasion it cannibalized its young. The Ghost Ranch
flock of Coelophysis contained a great many more yearlings and juveniles than adults; about 20 yearlings for each
of the largest adults. Sexual maturity probably onset between two and three years of age and the largest animals in
the flock were seven years old, or older. The babies had a suite of characters that is common to many tetrapods.
Compared to the adults, they had short snouts, large eyes, small hands, long legs and large pedes. The survivorship
curve of Coelophysis indicates that juvenile mortality was very high, but if they survived for a few years, the
chances for a long life were greatly increased.
2
INTRODUCTION
In 1947, George Whitaker discovered what became one of the
most famous of all dinosaur bonebeds. Working as part of an expedition
of the American Museum of Natural History, Whitaker found dinosaur
bones in Upper Triassic strata at Ghost Ranch in Rio Arriba County,
New Mexico (Fig. 1). Subsequent excavation of the site uncovered hun-
dreds (perhaps, ultimately thousands) of skeletons of a small theropod
dinosaur long recognized as Coelophysis bauri (Fig. 2). The Ghost Ranch
bonebed (here referred to as the Whitaker Quarry) thus provides a re-
markable window on the anatomy and paleobiology of a Late Triassic
dinosaur based on one of the most extensive population samples known
of any dinosaur.
Many dinosaur (and other) fossils have been excavated from the
Whitaker Quarry, first by the American Museum of Natural History
during the 1940s-1950s and subsequently, during the 1980s by the Carnegie
Museum of Natural History. One of the large blocks collected from the
quarry by the Carneige Museum was given to the New Mexico Museum
of Natural History and Science (NMMNH). Preparation of the block
began during the 1990s and was completed to install in the NMMNH’s
Triassic exhibit hall, which opened to the public in May 2008.
Although much has been written about the Whitaker Quarry
Coelophysis (most notably Colbert’s [1989] monograph), much research
remains to be done. This monograph attempts to answer several research
questions about Coelophysis based on study of the Whitaker Quarry
block at the NMMNH, as well as comparative studies of Whitaker
Quarry fossils at various other institutions. These questions mostly
concern behavior and population biology that the large sample of
Coelophysis skeletons from the Whitaker Quarry provide a unique op-
portunity to answer. Our effort to undertake such research questions is
embedded with documentation of the preparation of the NMMNH block,
and a necessary review of other studies of the geology and paleontology
of the quarry. This monograph thus attempts a comprehensive study of
the fossils in the NMMNH block, and in combination with other data,
including outgroups of extant animals, to understand the lifestyle, behav-
ior, growth and reproductive strategy of the Late Triassic dinosaur FIGURE 1. Top, National and state map showing location of New Mexico
Coelophysis. and the Chama basin (star). Bottom, Regional map of Upper Triassic
Institutional abbreviations: AMNH, American Museum of Chinle Group outcrops in the Chama basin, north-central NM. Key landmarks
Natural History, New York; CLMNH, Cleveland Museum of Natural are highlighted. Cerro Blanco is likely the “Gallina Canyon” fossil locality
History, Cleveland; CMNH, Carnegie Museum of Natural History, Pitts- of Baldwin; Ghost Ranch is in the Arroyo Seco drainage.
burgh; MCZ, Museum of Comparative Zoology, Harvard University,
In 1887, Cope described the dinosaur fossils Baldwin had col-
Cambridge; MNA, Museum of Northern Arizona, Flagstaff; NMMNH,
lected in 1881. He assigned them to two new species of the previously
New Mexico Museum of Natural History and Science, Albuquerque;
described genus Coelurus: C. longicollis and C. bauri (Cope, 1887a). In
RHMP, Ruth Hall Museum of Paleontology, Ghost Ranch, Abiquiu;
an article published later that year, Cope (1887b) then assigned the
RTM, Royal Tyrrell Museum, Drumheller; SMP, State Museum of Penn-
species to Tanystrophaeus and named a third species, T. willistoni. How-
sylvania, Harrisburg; USNM, National Museum of Natural History,
ever, Cope (1889) changed his mind again, coining the name Coelophysis
Smithsonian Institution, Washington, D.C.
(“hollow form,” in reference to the hollow bones of the fossils Baldwin
HISTORY OF THE WHITAKER QUARRY had collected) for the three species. Hay (1930) designated Coelophysis
bauri the type species of the genus (Cope had failed to do this), and
The history of the dinosaur Coelophysis is as old as the study of Huene (1911, 1915) first illustrated the fossils of Coelophysis collected
vertebrate paleontology in New Mexico. In 1881, David Baldwin, a hired by Baldwin.
fossil collector working for E. D. Cope, discovered various dinosaur In 1947, an American Museum of Natural History field party led
bones in Upper Triassic strata in the Chama basin of northern New by Edwin H. Colbert discovered the dinosaur bonebed at Ghost Ranch,
Mexico at two localities: (1) “Gallina Canyon,” which according to New Mexico. The locality, variously known as the Ghost Ranch,
Baldwin, lies “400 feet below gypsum horizon, 180 feet above grey Coelophysis or Whitaker quarry (we prefer the latter, named for George
sandstone;” and (2) “Arroyo Seco,” in a layer “about four hundred feet Whitaker, who actually discovered the site) yielded hundreds of skel-
below gypsum stratum” (Colbert, 1989, p. 5) (Fig. 1). The gypsum etons of a small theropod dinosaur that Colbert (1947) immediately
stratum Baldwin referred to is obviously the gypsum of the Middle assigned to Coelophysis and later (Colbert, 1964) to C. bauri (also see
Jurassic Todilto Formation. Sullivan et al. (1996) further explained the Colbert, 1989). Eventually, Colbert (1989) would designate a lectotype
geographic locations (as best as can be determined) of Baldwin’s sites, specimen of C. bauri from Cope’s original material, something that had
and their stratigraphic level, which appears to have been in the Painted not previously been done (Padian, 1986).
Desert Member of the Petrified Forest Formation of the Chinle Group Hunt and Lucas (1991), however, argued that the lectotype of
(Lucas et al., 2003). The problem is that the exact location of the dino- Cope’s species Coelophysis bauri, four sacral vertebrae and a pubic
saur bones Baldwin collected (whether they were from “Mesa Gallina” process of the ilium (AMNH 2708), is not diagnostic rendering C. bauri
or “Arroyo Seco”) was lost long ago (e.g., Huene, 1915). a nomen dubium. Furthermore, the Whitaker quarry is not one of the
 3

FIGURE 2. An example of one of the AMNH Whitaker quarry blocks collected in the 1940’s. Two Coelophysis individuals are readily apparent, though
much of the bone-bearing surface was covered during preparation to highlight the specimens for exhibition purposes. Photograph courtesy of the late
Stuart A. Northrop.

localities where Baldwin collected the holotypic material of C. bauri. bonebed on 22 June 1947 as part of an AMNH expedition to collect
Those localities are much lower stratigraphically (in the Petrified Forest Triassic fossils in the Chama basin. The members of the expedition, led
Formation; the Whitaker quarry is in the Rock Point Formation) and by Colbert, included Whitaker, Carl Sorensen, and Thomas Ierardi. The
kilometers (Arroyo Seco) or tens of kilometers (Mesa Gallina) distant AMNH subsequently excavated the bonebed in 1947, 1948, 1949, 1951
from the Whitaker quarry (Sullivan et al., 1996). Consequently, there and 1953 (Colbert, 1989). In addition to the large jackets illustrated in the
was little justification for Colbert to refer the theropod dinosaur fossils quarry map, numerous smaller jackets and isolated specimens were re-
from the Whitaker quarry to C. bauri. moved, and there are dozens of small, unopened jackets at the AMNH
Hunt and Lucas (1991) proposed a valid name for the Whitaker (ABH, pers. obs.).
quarry theropod—Rioarribasaurus colberti—with the type specimen a The Carnegie Museum of Natural History, under the leadership of
skeleton (AMNH 7224) from the quarry. This, however, was not ac- David S Berman, re-opened the Coelophysis bonebed and excavated dur-
cepted by Colbert, who together with several supporters, petitioned the ing 1981 and 1982 (Fig. 3). The NMMNH block, C-8-82, from the
International Commission on Zoological Nomenclature to transfer the quarry was collected by the Carnegie team in 1982 and donated to the
type specimen of Coelophysis bauri from its lectotype (AMNH 2708) NMMNH.
to the holotype of Rioarribasaurus colberti (AMNH 7224) (Colbert et
PHOTOGRAPHS OF SELECTED
al., 1992). This step would validate C. bauri, invalidate R. colberti, and
confirm Colbert’s identification of the Whitaker quarry theropod as C. WHITAKER QUARRY BLOCKS
bauri through legislation rather than scientific consensus. Although it In this section we present a photographic record of various blocks
violated the principle of priority and was proposed without the neces- from the Whitaker Quarry at Ghost Ranch (Figs. 4-13). All measurable
sary revisionary work called for by the International Code of Zoological specimens in these blocks are included in this study (Appendix 1) except
Nomenclature, the Commission approved the petition of Colbert et al. for the MNA block, which was under study by other workers and was
(1992). Thus, the Ghost Ranch dinosaur is now “legally” called unavailable to us.
Coelophysis bauri. The type specimen of Coelophysis bauri, AMNH 7224 from
George Whitaker’s discovery of the Coelophysis bonebed at Ghost AMNH block VIII, is a complete robust (short neck and skull) morph
Ranch, and its subsequent excavation was reviewed by Colbert (1989). (Fig. 5A). We examined casts of this fossil but were unable to examine
The actual method and process by which the original AMNH excava- this block because it is on display at the AMNH. AMNH blocks (Figs.
tions of the 1940s-1950s were undertaken is detailed in Whitaker and 5B, 6) were excavated in 1947-48. The CMNH blocks (Fig. 7) were
Myers (1965). In brief, George Whitaker discovered the Coelophysis excavated in 1981-82. The CLMNH block (Fig. 8) was excavated by the
4

FIGURE 3. Photographs of the Carnegie Museum excavation of the Whitaker quarry during 1981-1982. A, Bulldozer begins clearing overburden. B-D,
Excavating and plaster-jacketing. E-F, Hauling and loading the plaster jackets. Photographs courtesy of David S Berman.
 5

FIGURE 4. Quarry map of the Whitaker Quarry with jackets from both the AMNH and CMNH excavations outlined. The darkened jacket is the NMMNH
Whitaker Quarry block. Roman numerals label AMNH blocks. CMNH blocks have a C prefix.

AMNH field crew in 1948. The block at RHMP (Fig. 9A-B), Ghost lies the principal bonebed, and thus its removal would impede the prepa-
Ranch, NM was excavated by the CMNH field crew in 1982 and re- ration process. Once the decision was made to prepare the block from
mained at Ghost Ranch. The first AMNH block (I) is also on exhibit at the bottom up it was necessary to devise a method to rotate the block
RHMP (Fig. 9C). MCZ blocks (Fig. 10) are probably parts of the origi- without damaging the specimens or endangering personnel. Because of
nal AMNH block X excavated by the AMNH field crew in 1948 (Fig. 4). the large size and friable nature of the block, we document here the
The MNA block (Fig. 11) was excavated by the CMNH field crew in methodology involved in rotating the block. Measurements reported in
1981. The NMMNH block (Fig. 12) was excavated by the CMNH field this section are often in English units as they pertain to tools and mate-
crew in 1982. The RTM block (Fig. 13) is approximately half of a large rials commonly available in the United States.
block excavated by the CMNH field crew in 1982. Prior to preparation the maximum dimensions of the NMMNH
TURNING OVER THE NMMNH block were ~ 2 m long, ~ 1.4 m wide, and ~ 1.1 m deep, yielding a volume
of ~ 2 m3 (Fig. 14A-B). The calculated maximum mass, including jacket-
WHITAKER QUARRY BLOCK
ing materials was ~ 5,440 kg (~ 12,000 lb). The rock was extremely
The NMMNH block from the Whitaker quarry was stored at friable and had intersecting desiccation cracks and joint sets throughout.
NMMNH until 1997, when a decision was made to begin preparation of With these potential problems in mind, we decided to design and
the block for the NMMNH’s upcoming Triassic hall. It was then decided construct a large wooden rocker to gently roll the block 180o. Accurate
to prepare the NMMNH block from the bottom up in hopes that no estimation of the center of gravity and analysis of nonconcurrent forces
bone would be found below the principal bonebed, and that this would (forces that are not balanced about the center of gravity), rolling resis-
expedite preparation. Based on the preparation of other blocks, it was tance, traction, material strengths, and minimization of vibration and
known that poorer quality, disarticulated (probably reworked) bone over- flexing of the specimen were essential to the successful roll. The block
6

FIGURE 5. A, The neotype specimen of Coelophysis bauri. Photo of the RHMP cast of AMNH 7224 from AMNH block VIII. B, AMNH Block V,
excavated by Colbert et al. The block contains at least 14 specimens.
 7
was rolled on a time scale of 1 minute with no damage to its jacket or
contents.
Materials and Methods
The block’s original field jacket of plaster, burlap, lumber, and
steel packing straps had deteriorated somewhat during the ~15 years of
storage at NMMNH and required strengthening in preparation for the
roll. A complete new jacket was applied over the old one using ~ 500 kg
(~ 1,100 lbs) of plaster-saturated burlap (Fig. 14A-B). More than half of
the plaster was applied to the top and upper 30 cm of the sides to
provide a thick, stable base after the roll. The jacketing process took
three to four people an entire afternoon (~ 4 hours).
Maintaining control of the heavy block throughout the roll was
essential to avoid damage to the block and insure the safety of the work-
ers. We planned to use both a pulling motive force to roll the block and a
restraining force to keep it under complete control.
Critical to this process was an accurate initial estimate of the
center of gravity (CG). To minimize nonconcurrent forces the block’s
axis of rotation had to pass through the CG. Incorrect estimation of the
CG would have produced nonconcurrent forces that would have added
to the rolling resistance and opposed or assisted the roll. An error of a
few tens of cms could have produced a torque of several thousand foot-
pounds that would have had to be compensated for by the pulling and
restraining devices.
We carefully estimated the block’s center of gravity by first pho-
tographing the top, side, and end views of the block. From the top and
side views, we estimated a point on the surface that had equal area above
and below it, and on each side of it. Lines perpendicular to the block
surface from each of these points should therefore intersect inside the
block at the approximate CG. Using the end view, a straight line was
projected parallel to the block’s axis that passed through the CG. This
line should represent the balanced axis of rotation.
The block had one approximately square corner that was parallel
to the axis of rotation (the upper left corner in the end view, Fig 14B).
This corner was the most natural place to mount the rockers. The radius
of the rockers was thus defined to be the distance from the axis of
rotation to this corner of the block.
The rockers were constructed of 2 x 6 inch (actually ~ 1.5 in x 5.5
in or ~ 3.75 cm x 13 cm) Douglas Fir lumber laminated with construction
adhesive and joined with 3.5 inch-long, number 10 steel woodscrews.
Douglas Fir was chosen as the strongest (compressive strength and shear
parallel to grain) of the available construction woods based on Marks
(1951). Wooden box beams made from 2 x 6 Douglas Fir and 0.75 inch
plywood were mortised into the rockers to solidly connect them and
provide lateral strength (Fig. 14C-D). The finished rockers were attached
to 4 foot-wide x 6.5 foot-long plywood panels using high strength (4140)
steel all-thread rods. Each panel comprised two sheets of 0.75 in ply-
wood laminated with construction adhesive and joined with woodscrews.
The rockers and their plywood panels were joined at 90o using 10 cm (4
in) angle iron and bolts (Fig. 14D). The assembled rockers were then
fastened to the block using five cm (2 in)-wide, 1360 kg (3000 lb)-rated
ratchet cargo straps (Fig. 14E-F). A total of 13 straps were used. Wooden
shims were placed in vacant areas between the uneven block surfaces and
the rocker panels to transfer force directly to the rockers and prevent
flexing of the block. A steel hardpoint was constructed for attachment of
the pulling and restraining cables and chains (Fig. 15A-B). The plywood
panel and 2 x 6 reinforcement were sandwiched between two 0.25 in-
thick steel plates with a braced 0.75 in steel half-loop welded to the top
plate.
Rolling the Block
FIGURE 6. AMNH block IX, collected by Colbert, et al. A, Overview of the The block was towed outside onto a concrete driveway. Jacks
block, which contains approximately 20 specimens. B, A bone map of block were placed under the “roll side,” the wheels were removed (Fig. 15C),
IX provided by AMNH. C, Close-up view of the skulls and necks of AMNH and this side was lowered by jacks to the concrete surface. The other side
7227 and 7228 in block IX.
8

FIGURE 7. Carnegie Museum of Natural History blocks from the Whitaker Quarry, excavated by Berman et al. A, Part of block C-3-82, which went to the
USNM in 1986, where it was prepared; it now resides at the CMNH. B, CM 81765, a large, isolated skull in block C-3-82. Skull length = 245 mm. C, Part
of block C-4-81 containing the nearly complete postcrania of Postosuchus (upper left to upper center), now on exhibit at the CMNH. D, Another part
of block C-4-81, also on exhibit at the CMNH.
was jacked up until an angle of ~ 30o was reached and the rocker was Palm, and one of the authors (LFR). Herb, possibly the better advised of
nearly in contact with the concrete (Fig. 15D). A small Toyota forklift the two, soon moved on to other projects. Initially, the block was to be
was used to pull the block through the roll, while a 2720 kg (6000 lb)- prepared right side up, as was the typical procedure for these large
rated bumper-mounted winch on a military HMMWV (Hum Vee) pro- blocks.
vided restraint (Fig. 16A). The winch cable was slowly played out to When the top of the field jacket was removed it was apparent that
restrain the block and maintain control as the roll proceeded. The roll was the block was unstable due to several large joints and these needed to be
completed in less than one minute with no damage to the block (Fig. 16B- repaired before preparation could begin. Additionally, the rock had pulled
H). The block was then jacked up (Fig. 17A-B) using two pairs of away from its jacket in some areas, and debris (including fossil bones)
railroad jacks with steel box beams passing between them under the was shedding into the joints in the block and spaces between the jacket
block and its plywood panel (Fig. 17B). Finally, the steel cart was reas- and the block. The joints and spaces apparently resulted from desicca-
sembled under the block, and the jacks were removed (Fig. 17C). tion and shrinkage of the siltstone. The largest joint was 10 cm wide and
ran east-west across the block, bisecting it from top to bottom. Although
PREPARATION OF THE NMMNH scattered, disarticulated bone was present on the block surface, the main
WHITAKER QUARRY BLOCK bonebed was deeper in the block. We inspected the inside surfaces of the
Because of the large size of the NMMNH block, the delicate larger joints prior to filling them and established that the principal bone
deposit was approximately 30 cm below the upper surface of the block.
nature of the specimens, and the widespread distribution of Ghost Ranch
Whitaker Quarry jackets that remain unprepared or partially prepared, We later corroborated this observation by inspecting the RHMP block at
Ghost Ranch, in which a section on one side of the jacket was removed,
we describe the preparation of the NMMNH block (C-8-82) in some
detail here. At first, we stabilized the block and began work from the top, exposing a profile of the sediments.
The joints and spaces were repaired by incrementally filling them
but eventually we determined that a more productive approach would be
to prepare the block from the bottom up, so the block was rotated (as with crushed matrix mixed with Glyptol®, a translucent, amber-colored
insulating varnish and sealant manufactured by General Electric. Only
described above). Measurements in this section are often given in English
units because they pertain to tools and implements commonly available about 2 cm per week of fill could be introduced as thicker applications of
the Glyptol would not dry. Approximately six months were required to
in the United States.
completely fill the joints and stabilize the block. Meanwhile, some prepa-
Initial Stabilization and Preparation of the Block Surface ration was begun on the most stable area near the northeast corner using
standard mechanical preparation techniques and magnification using a
Preparation of the NMMNH Whitaker Quarry block, C-8-82, Carl Zeiss® ophthalmological microscope.
began in June of 1997 under the supervision of Pete Reser, then NMMNH The sediments above the principal bonebed were apparently re-
preparator. Initial preparation was begun by two volunteers, Herbert worked as they contained a substantial amount of bone, all of which was
 9

FIGURE 8. Block AMNH XII, collected by Colbert et al. in 1948, went to the Cleveland Museum of Natural History where it was prepared and is now
housed. A, This densely fossiliferous block contains more than 12 dinosaur specimens. The Cleveland exhibit specimen was removed from this block
(upper right area in the photo). B, A bone map of the Cleveland block provided by CLMNH, was drawn by the block’s preparator, Dale Zelinski.
10

FIGURE 9. The Ruth Hall Museum of Paleontology block, C-9-82. A, Overview of the block. B, Part of C-9-82, removed from the main block to facilitate
the preparation of a complete Vancleavea skeleton, GR-138. The Vancleavea skull has been removed for study and is replaced by a cast in the photo. C,
AMNH Block I on exhibit at the RHMP, contains an incomplete phytosaur tail, hip, and ribs, plus some Coelophysis material.
 11

FIGURE 10. Museum of Comparative Zoology blocks, probably parts of AMNH block X excavated by Colbert et al. A, MCZ 4331, cm and inch scale on
calipers. B, MCZ 4332.
12

FIGURE 11. Museum of Northern Arizona block C-5-81, excavated by Berman et al., containing at least seven specimens including a nearly complete
juvenile Coelophysis.

FIGURE 12. New Mexico Museum of Natural History block C-8-82, excavated by Berman et al. contains more than 24 incomplete to nearly complete
Coelophysis individuals.
 13

FIGURE 13. Royal Tyrrell Museum block, part of C-1-82, excavated by Berman et al. The large specimen just below center with its leg flexed in a reversed
Z-shape is not Coelophysis; it pertains to Eucoelophysis.
14

FIGURE 14. A, Side view of the NMMNH Whitaker Quarry block C-8-82 on its steel cart after rejacketing. B, End view. C, A pair of finished rockers,
connected by their box beam, ready to bolt to their plywood panel. D, The two rocker panels are joined at 90o using angle iron to reinforce the corner. E,
The finished rocker is fastened to the block with ratchet straps and shimmed with wooden blocks. End view down the axis of rotation. F, Side view.
 15

FIGURE 15. A, Oblique view of the steel hardpoint for attachment of the pulling and restraining cables and chains. B, Side view. C, The wheels were
removed from the roll-side of the steel cart. D, The cart and block are jacked up to ~ 30o in preparation for the roll. The cable on the right restrains the
block.

disarticulated and scattered. Several mud-clast ripups were evident on more silty layers, so a small (5/16 in, ~ 8 mm) cold chisel and a ball peen
the block surface. As preparation proceeded, numerous disarticulated hammer (2 ounce, 57 g) worked sufficiently well to prepare most of the
vertebrae, ribs, and phalanges were exposed in relief, together with a lower portion of the block. Matrix was removed with care to maintain a
scapula, tibia-fibula pair, and a disarticulated skull. We recorded the relatively level surface (Fig. 18). Depth below the original upright sur-
orientation of the bones and ripups on the block surface using the con- face of the block was marked on the plaster jacket in several locations to
ventional protocol of measuring only elements that are at least four times aid in judging depth as preparation proceeded. Lithologic samples were
longer than they are wide as recorded in the section on the block stratig- continuously collected and their depth levels recorded. These were later
raphy and taphonomy (Table 1). The number of elements is few, but a used to reconstruct the microstratigraphy of the block (see below). Depth
strong west-northwest orientation is obvious. was also noted for all the fossils removed.
In October of 1998, shifting priorities at the NMMNH dictated a Invertebrates and fish bones in beds 3 and 4 (defined below) were
temporary halt in the preparation. removed on small matrix slabs. The first Coelophysis bones were en-
countered at a depth of approximately 90 cm below the original surface
Preparation From the Bottom Up
of the block, in bed 5. Most of these were articulated, and many were
In July, 2003 it was decided to turn the block over and prepare it well preserved, but the bone density was much lower than at the depth
from the bottom up (see above). A full-time contract preparator and later of the main bonebed. This region of high quality, but low-density bone
museum preparator (LFR) was hired to roll the block over and prepare it. was 30 to 40 cm thick, spanning most of beds 5 and 6. We decided to
In September 2003, all exposed bones on the top surface of the block remove theses bones in smaller “field” jackets made directly on the
were removed and cataloged before the block was turned over and full- working surface and save them for later preparation. In all, 11 jackets
time preparation begun. were removed from the block (Fig. 19). Existing joints were used to
In the following notations concerning stratigraphy of the block, define the jackets rather than trenching undisturbed areas, greatly lessen-
the beds are numbered conventionally, from the bottom up with respect ing damage to the fossils. The fractured, delta-shaped area between jack-
to the original upright position of the layers. The sediments in the lower, ets J2 and J3 contained Vancleavea osteoderms and bone. The slabs of
sandier portion of the block were somewhat more fissile than the upper, fossil-bearing matrix in this area were removed without jacketing and
16

FIGURE 16. A, The block is ready to roll. A small forklift (foreground) operated by M. Pierce will provide the pulling force. An electric winch on the
HMMWV (HumVee) (background) will provide restraint. B, Ready. C, Roll begins. As the forklift begins to pull, the winch cable is played out. D, Roll
continues; cart falls away. E-G, Roll continues. H, Roll complete. Photographs by R. Ungnade.
 17

FIGURE 17. A, After the roll, the straps, side panel, and rockers are removed. B, Four railroad jacks lift the block on its plywood panel. C, The steel cart
is reassembled under the block and the jacks removed.
18
TABLE 1. Orientation of exposed bones and ripups on the block surface at the top of bed 7 (Fig. 26).

FIGURE 18. The block rests on its original top. The jacket has been opened, exposing the original block bottom, and preparation has begun. The “pond
bottom” sediments of beds 1 and 2, are under preparation here. “Jake stick” increments are 0.25 m in the lower portion and 0.1 m in the upper portion.
 19

FIGURE 19. Sketch map of jackets removed from the NMMNH Coelophysis
block C-8-82. The lower right corner in this map represents the original
northeast corner of the block as it was oriented in the quarry.

marked for later reassembly and preparation. We illustrate the removal

of jackets J3 and J4 as an example of the process (Figs. 20-21).
After removal of the jackets, approximately 10 cm of siltstone
remained over the principal bonebed. The densest portion of the bonebed
was in the lower half of bed 7 and extended in places down into bed 6.
The entire Coelophysis bonebed was prepared under magnifica-
tion, not only to minimize damage to bone surfaces, but to ensure that no
microfossils would overlooked. The sphenodontid, Whitakersaurus
(Heckert et al., 2008), for example, would almost certainly have been
unnoticed and destroyed. A jeweler’s binocular visor of 3x or 5x power FIGURE 20. A, Jackets C-8-82-J1, J1B, and J2 have been removed from the
was used during airscribing the matrix, and a Leica® 7x to 30x stereo block. The areas that will comprise J3 and J4 are defined and ready for
zoom microscope was used while preparing the bone surfaces. jacketing with plaster and burlap. Note depth marks on the jacket. The
The bulk of the matrix in the Coelophysis bonebed was removed principal bonebed is at 0.3 m depth. The northeast corner of the block is at
using a Paleotools® Microjack 4 airscribe. Only hand tools were used the far right. B, Jackets J3 and J4 are finished and ready to be broken loose
from the block and removed. “Jake stick” increments are 0.25 m.
directly on the bone. These hand tools consisted of tungsten carbide rod
segments of ~0.5-2 mm diameter, sharpened with a fine or superfine taken to not apply the consolidant to the matrix.
diamond grinding wheel and held in a pin vise handle. The rods were Where necessary, the matrix was stabilized with Vinac B-15 (poly-
sharpened by grinding the ends off at a very acute angle to leave a rounded vinyl acetate). Vinac beads were dissolved in equal parts by volume of
tip with a very keen edge. Generally, the rods needed resharpening only acetone to form a “mother” solution, which was further diluted with
in the event of flaking the fine edge as they seldom became noticeably dull eight parts acetone to one part “mother.” This low viscosity consolidant
in normal use. They were very effective in carving the matrix. was applied very liberally to the block edges, which were generally much
A small air blower was fabricated to keep the immediate work area less stable than the more internal portions of the matrix. The bulk of the
free of dust during the use of hand tools. The blower consisted of a 6.4 matrix was solid siltstone and required no consolidation.
mm (0.25 in) diameter brass tube with the end cut off diagonally and Once the densest part of the Coelophysis bonebed was reached,
connected to the air supply by a rubber tube and control valve. A flat no more fossil material was removed from the block. The bones were
steel “foot” with a foam pad cemented to its underside formed the base of prepared in relief and left as preserved on the block surface.
the blower and was connected to it by a heavy flexible wire which al- After the block was fully prepared it was readied for exhibit. We
lowed the blower to be aimed at the work area. decided to tip the block up at a 20º angle for best visibility in the exhibit
Only Paleobond® cyanoacrylate glue and PB-002 Penetrant and hall. Because the remaining portion of the original huge block was only ~
Stabilizer consolidant were used on the bone. The use of consolidant was 30 cm thick and now resided in a shallow bowl formed by the plaster
held to the minimum amount necessary to stabilize the specimens. How- jacket, there was a small but finite possibility of the block shifting or
ever, only a few square mm of most bone could be exposed without sliding inside its bowl-like container as it was tipped up for display. To
consolidation. Application of the consolidant with precision and in small, prevent any motion between the block and its jacket, numerous 6-inch-
controlled amounts was an important consideration. Precision applica- long masonry anchors were screwed through the jacket and deep into the
tors were fabricated from very fine (0.22 mm diameter) acupuncture rock matrix. Additionally, ~5 pints (~2.4 L) of Paleobond® PB-750
needles with a small (<1 mm to ~3 mm diameter) loop or hook bent at cyanoacrylate were fed into the joint between the matrix and jacket to
their ends to hold the appropriate size drop of consolidant. Care was solidly glue the jacket to the rock.
20
From the resumption of work on the block in July, 2003 until it
was ready for exhibit in 2008, approximately 5200 person-hours of
preparation time were invested.

FIGURE 21. Jackets J3 and J4 have been removed from the block. Primary
preparation began on the lower surface exposed at right.
 21
Assignment of the strata that contain the Ghost Ranch Whitaker
GEOLOGY OF THE WHITAKER QUARRY quarry to the Rock Point Formation is well supported because: (1) these
Stratigraphy and Age strata closely resemble the type Rock Point strata in the Four Corners
region (Harshbarger et al., 1957; Lucas et al., 1997); (2) Rock Point strata
The Whitaker quarry at Ghost Ranch is located stratigraphically in the Chama basin are at the top of the Chinle Group, as they are
high in the local Upper Triassic stratigraphic section (Fig. 22) in strata elsewhere (e.g., Stewart et al., 1972; Lucas, 1993); and (3) Rock Point
first assigned to the Rock Point Formation of the Chinle Group by Lucas strata at Ghost Ranch yield Apachean-age vertebrate fossils (Lucas and
and Hunt (1992) and recently mapped by Koning et al. (2006) as the Hunt, 1993; Lucas et al. 2003, 2005; Lucas and Tanner, 2007). Stewart et
Rock Point Formation (Fig. 23). A measured stratigraphic section at the al. (1972) recognized this correlation, equating the “siltstone member”
Whitaker quarry (Figs. 24-25) demonstrates that it is approximately 30 (our Rock Point Formation) to strata they assigned to the Rock Point
m above the contact of the Rock Point Formation with the underlying Formation (Member) elsewhere.
Painted Desert Member of the Petrified Forest Formation and 35 m Indeed, the Rock Point strata at Ghost Ranch are demonstrably
below the base of the Middle Jurassic Entrada Sandstone (Lucas and different from underlying strata of the Painted Desert Member. Rock
Hunt, 1992; Hunt and Lucas, 1993a,b,c; Sullivan et al., 1996; Lucas et al., Point strata are dominantly siltstones and fine-grained sandstones lack-
2003, 2005; Lucas and Tanner, 2007; Heckert et al., 2008). The Rock ing bentonite or smectite clay minerals. These strata are massive or
Point Formation vertebrate fossil assemblage at the Whitaker quarry is repetitively bedded in sets typically ranging from 1 to 5 m thick and
distinct from that of underlying units and is considered Apachean (late weather to shades of orange (Koning et al., 2006). In contrast, Painted
Norian-Rhaetian) on the nonmarine Triassic tetrapod timescale of Lucas Desert Member strata are dominantly bentonitic mudrocks deposited as
(1998; Lucas and Tanner, 2007; Lucas et al., 2007b). Because of the long thick (3-10 m), tabular beds with occasional channel deposits of ripple-
history of mistakenly attributing a lower stratigraphic position and thus laminated to planar-laminated sandstone with minor intraformational
an erroneously old age to the quarry (e.g., Colbert 1989, 1990; Dubiel conglomerate and occasionally well-developed paleosols that collectively
1989a; Rowe and Gauthier 1990; Schwartz and Gillette, 1994), we briefly weather to shades of red and brown (Lucas, 1993; Lucas et al., 2003,
review the stratigraphy of the Rock Point Formation in the Chama basin 2005; Koning et al., 2006).
and the evidence for the age of the Whitaker quarry. Dubiel (1989a) and Goldstein et al. (1996; see also coauthored
The Whitaker quarry has been variously placed in the upper Pet- abstracts with Goldstein by Hargrave et al., 1996 and Trinh et al., 1996)
rified Forest Member (= Painted Desert Member of the Petrified Forest assigned upper Chinle strata in the Ghost Ranch area to the Owl Rock
Formation in our usage) (e.g., Colbert, 1989, 1990; Schwartz and Gillette, “Member” of the Chinle Formation based on sedimentological models
1994), the “siltstone member” (Stewart et al., 1972; O’Sullivan, 1974, (e.g., Dubiel, 1989a,b), postulating that these strata are a lake margin
1977; Nesbitt, 2007), Chinle Formation undifferentiated (e.g., Nesbitt facies of that unit. Lucas and Hunt (1992, p. 158) presented a detailed
and Norell, 2005; Nesbitt et al., 2006), the Owl Rock Member (= Forma- refutation of Dubiel’s assignment, and subsequent work on Owl Rock
tion of our usage) (Dubiel 1989a; Goldstein et al., 1996), and the Rock sedimentology (Tanner, 2000) has rejected the sedimentological model of
Point Formation (= “Member” of previous usage) (e.g., Lucas and Hunt, Owl Rock deposition advocated by Dubiel (1989a,b). Furthermore, Lucas
1992; Hunt and Lucas, 1993a,b,c; Long and Murry, 1995; Sullivan et al., et al. (2003, 2005) have demonstrated that the correlations of Goldstein
1996; Harris and Downs, 2002; Tykoski and Rowe, 2004; Koning et al., and colleagues are contrary to available outcrop data.
2006; Heckert et al., 2008). We follow Lucas (1993, 1997) and correlate Recently, Zeigler et al. (2008; also see Zeigler and Geismann,
strata at Ghost Ranch identified and recognized by Stewart et al. (1972) 2009) referred to the strata at Ghost Ranch that include the Whitaker
as the “siltstone member” to the Rock Point Formation. quarry as the Rock Point Formation or the “inferred” Rock Point Forma-

FIGURE 22. Summary of evolution of Upper Triassic stratigraphic nomenclature in the Chama basin of northern NM.
22

FIGURE 23. Geologic map of the area around the Whitaker quarry (after Sullivan et al., 1996). The stratigraphic sections in Figure 24 are labeled: A,
Canjilon Quarry, B, Whitaker Quarry and C, Orphan Mesa. Topographic base map (1.6 km.2 grid) from U.S. Geological Survey, Ghost Ranch, NM (1:
24,000). Abbreviations: TrP, Poleo Formation, TrPF, Petrified Forest Formation, TrRP, Rock Point Formation, JE, Entrada Sandstone, JT, Todilto
Formation, QC, Quaternary colluvium. Numbers are fossil localities near Orphan Mesa. CA = Canjilon quarry; T = Typothorax locality; W = Whitaker
quarry.
 23

FIGURE 24. Measured stratigraphic section at the Whitaker quarry correlated to nearby measured stratigraphic sections (modified from Lucas et al., 2005).
Section lines illustrated in Figure 23: A, Canjilon Quarry, B, Whitaker Quarry and C, Orphan Mesa.

tion, but correlated these strata to the Moenave-Wingate interval to the Redondasaurus we recognize are restricted to uppermost Triassic rocks,
west based on paleomagnetic pole similarities. Zeigler et al. (2008, p. including the Redonda Formation in east-central New Mexico, the Rock
236) thus stated that “strata previously interpreted as Rock Point For- Point Formation, and the basal Wingate Sandstone in Utah (Hunt and
mation in northern New Mexico are most likely time equivalent to the Lucas 1993c; Lucas et al., 1997; Heckert et al., 2001; Hungerbühler,
Moenave-Wingate formations and are not part of the Rock Point Forma- 2002; Lucas and Tanner, 2007). Whereas some authors have questioned
tion.” They further concluded that “based on the paleomagnetic data, the the validity of this taxon (e.g., Long and Murry 1995), others (e.g.,
‘Rock Point’ Formation of the Chama basin should be termed Moenave Hungerbühler 2002; Spielmann et al., 2006) recognize it as valid, which,
Formation.” Clearly, this confusion of chronostratigraphy with combined with its stratigraphic distribution, validates its utility as an
lithostratigraphy runs counter to basic stratigraphic principles by as- index taxon of the Apachean land-vertebrate faunachron (Lucas, 1998;
signing a lithosome to a rock formation based on its inferred age and not Lucas et al., 2007b). Unfortunately, the other vertebrates from the
on its lithologic characteristics and stratigraphic position. Coelophysis quarry do not provide a precise age within the Late Triassic,
Coelophysis is by far the most abundant vertebrate known from as all of the fish, Hesperosuchus, Postosuchus, and Vancleavea have long
the Whitaker quarry. Other vertebrates include an indeterminate temporal ranges (e.g., Huber et al., 1993; Long and Murry 1995; Lucas
chondostrean, redfieldiid and coelacanthid osteichthyans, at least two 1997; Milner et al., 2006; Hunt et al., 2002, 2005). Coelophysoid
drepanosauromorph taxa, the sphendont Whitakersaurus, the phytosaur theropods are extremely difficult to identify at the generic level and have
Redondasaurus, the crocodylomorph Hesperosuchus, the enigmatic a stratigraphic range from the Carnian to the Hettangian (early Late
archosauromorph Vancleavea, the rauisuchian Postosuchus, and the Triassic to Early Jurassic) (Heckert and Lucas, 2000; Tykoski and Rowe,
shuvosaurid Shuvosaurus (=Effigia) (Schaeffer, 1967; Colbert, 1989; 2004; Spielmann et al., 2007). Shuvosaurus okeeffeae and Whitakersaurus
Hunt and Lucas 1993a,b,c; Long and Murry 1995; Clark et al., 2000; bermani have only been found at the Coelophysis quarry, and thus are
Harris and Downs, 2002; Lucas et al., 2003; Nesbitt and Norell 2005; not biostratigraphically useful, and other records of Shuvosaurus (S.
Lucas et al., 2007; Nesbitt 2007; Heckert et al., 2008). Of these verte- inexpectatus) are from strata of older, Revueltian, age in West Texas
brate taxa, Redondasaurus is the most age diagnostic – records of (Lucas et al., 2007a).
24

FIGURE 25. Photograph of the local stratigraphy around the Whitaker quarry, with stratigraphic units labeled.

Litwin (1986) and Litwin et al. (1991) assigned an early Norian microstratigraphy that allows us to infer that the bone-bearing interval at
age to the Rock Point Formation based on palynostratigraphy. In con- the Whitaker quarry represents two cycles of deposition (Fig. 26). In
trast, the conchostracan assemblage from the Whitaker quarry yields a the following description of the block microstratigraphy grain size was
new species of Shipingia that is of late Norian (Sevatian) age, according determined microscopically using the standard American/Canadian Strati-
to the conchostracan biostratigraphic zonation of Kozur and Weems graphic grain size card. Colors are according to the GSA Rock Color
(2005, 2007). This is consistent with the vertebrate biochronology, which Chart (eighth printing, 1995).
suggests that Apachean time correlates to the late Norian-Rhaetian (Lucas Seven distinct microstratigraphic units are represented in the block
and Tanner, 2007; Lucas et al., 2007b). In contrast, Zeigler and Geissman (Fig. 26). All are relatively massive and poorly laminated, and thus do
(2009) recently assigned a late Rhaetian-Hettangian age to the Rock not split easily along bedding planes. We interpret that unit 4 was depos-
Point Formation in the Chama basin based on the similarity of its ited in a pond bottom based on the presence of freshwater invertebrates
paleopoles to those of the Moenave-Wingate interval to the west. Clearly, (ostracods and conchostracans) and its lithology of massive siltstone.
the Whitaker quarry is Late Triassic in age (not Hettangian), and either Units 4 and 5, followed by units 6 and 7, represent two upward fining
late Norian or Rhaetian. We favor the late Norian age for the Whitaker overbank flood sequences (Fig. 26).
quarry supported by vertebrate and conchostracan biostratigraphy. Unit 1: The lowermost unit of the block comprises 11 cm of fine-
grained silty litharenitic sandstone that is moderate red (5R 4/6) to mod-
Microstratigraphy erate yellowish brown (10YR 5/4) in color, with small (1-3 mm diameter)
The NMMNH Whitaker quarry block has a distinct pale green (5G 7/4) reduction spots.
 25
Unit 2: This unit pinches out laterally and has a maximum thick-
ness of 6-7 cm. This bed of moderate red (5R 4/6), silty, fine-grained
litharenitic sandstone contains numerous small (1-5 mm diameter), sub-
spherical, pale green (5G 7/2) reduction spots that make up nearly one-
third of the volume of the unit.
Unit 3: Ten cm of pale green (5G 7/2), reduced, poorly cemented,
fine sandy siltstone contains a few small pebbles of unit 2 lithology that
are up to 5 mm in diameter.
Unit 4: Eighteen cm unit comprising pale reddish brown (10R 5/4)
muddy siltstone with dark reddish brown (10R 3/4) stringers. There is
some moderate olive brown (5Y 4/4) mottling. The uppermost portion
of unit 4 intertongues somewhat with unit 5. This is the first vertebrate
fossil-bearing unit and contains sparse, isolated fish scales throughout.
The lower 5 cm of the bed contain conchostracans and ostracodes. The
upper 13 cm contain scattered tetrapod teeth, disarticulated fish bones,
including bones and scales of the coelacanth fish Chinlea, whereas the
uppermost, intertongued portion contains complete articulated skeletons
of the redfieldiid fish, Synorichthys. Above the basal invertebrate-bearing
portion of this bed, elongate, dark reddish brown (10R 3/4) mudstone
ripup clasts are present in a matrix-supported sedimentary breccia. The
ripup clasts are tightly aligned ESE-WNW (Fig. 27A). Throughout the
block the ripups and carbonaceous material are often surrounded by a
pale green (5G 7/2), halo-like reduction zone (Fig. 27B-C).
Unit 5: Fourteen cm of pale reddish brown (10R 5/4) mudstone
and siltstone containing pale green (5G 7/2) reduction zones. Dark red-
dish brown (10R 3/4) mudstone ripups with a maximum diameter up to
1.5 cm in diameter are common. The unit contains tetrapod bones through-
out, including disarticulated Vancleavea osteoderms and bones, and both
disarticulated and articulated Coelophysis bones. Most of the dinosaur
bones represent very small juveniles.
Unit 6: This 36-cm-thick unit comprises pale greenish yellow
(10Y 8/2) to moderate olive brown (5Y 4/4) conglomeratic silty litharenite
sandstone with sparse calcrete pebbles, some small quartzite pebbles,
and numerous moderate brown (5YR 4/4) to dark reddish brown (10R 3/
4) mudstone ripups. The pebbles are up to 1 cm in diameter, whereas the
ripups are as much as 7 cm in diameter. Articulated Coelophysis skeletal
material is found throughout this unit. The bones are well preserved but
of lower density than in the principal bonebed. Flattening and distortion
of the fossils is generally less here than in the overlying, more silty
principal bonebed, probably as a result of the less compressible nature of
the sand. Some articulated specimens extend vertically between this bed
and unit 7, which contains the principal bonebed. This probably indi-
cates that although there are two apparent fining-up sequences in the
quarry, they must represent two surges of the same flood event rather
than two widely temporally separated events.
Unit 7: A 49 cm thick unit, and containing the principal Coelophysis
bonebed in its lower 20 cm. The lithology is very similar to unit 5:
generally a pale reddish brown (10R 5/4), muddy silt with dark reddish
brown (10R 3/4) mud ripups. The uppermost 20 cm appear to be re-
worked and are bone rich, but the fossil material is almost completely
disarticulated or loosely associated at best. This unit was not prepared.
Taphonomy
A massive fossil assemblage such as that of the Whitaker Quarry
demands explanation. How did so many animals die, and how did they
come to be buried and fossilized in a single localized stratum? Schwartz
and Gillette (1994) produced an analysis of the taphonomy of the
Whitaker Quarry that provides the starting point for this inquiry.
They reached the following scenario of events: (1) the dinosaurs
FIGURE 26. Microstratigraphic section of the layers exposed in the NMMNH represent carcasses transported to the site by fluvial currents; (2) the
block (C-8-82), restored to their original positions with oldest on bottom. carcasses blocked a small channel and were then buried by silt; and (3)
The densest portion of the bonebed is in the lower half of bed 7 and extends the dinosaurs buried in the quarry died as part of a “regional environmen-
down into bed 6 in places. tal crisis, such as a drought” (Schwartz and Gillette, 1994, p. 1118).
Our data support the conclusion that the dinosaur carcasses that
ultimately produced the Whitaker Quarry bonebed were transported
26
hydraulically to their place of burial. As Schwartz and Gillette (1994, fig.
10) demonstrated, long bones, skulls, limbs and vertebral columns of
Coelophysis in the quarry show preferred orientations indicative of flu-
vial transport and hydraulic sorting. Preferred orientations of the same
elements in the NMMNH block support the observations of Schwartz
and Gillette (1994). Our data show that these bone orientations are
essentially identical to the orientation of the mudclast ripups (Fig. 27).
Nevertheless, we are not convinced that the Coelophysis carcasses
clogged a small channel, as posited by Schwartz and Gillette (1994).
Instead, our microstratigraphic and fossil data (Fig. 26) indicate that a
pond existed at the spot where the bones of the Whitaker Quarry subse-
quently accumulated. The pond represented a topographic depression in
which the dinosaurs and associated carcasses accumulated. Indeed, the
panel diagram of the microstratigraphy of the Whitaker Quarry pre-
sented by Schwartz and Gillette (1994, fig. 6) does not show a
channelform geometry to the bone-bearing deposit. Instead, the diagram
suggests a sheetflood geometry of the bone-bearing bed. Furthermore,
microstratigraphy of the NMMNH block (Fig. 26) supports the conclu-
sion that the dinosaur carcasses were washed into the quarry site by a
single event that also ripped up underlying lacustrine siltstones and even
introduced some of the bones into the uppermost level of the lacustrine
interval (top of unit 5: Fig. 26).
What killed the dinosaurs at the Whitaker Quarry? We don’t know
with any certainty, but several lines of evidence suggest a parsimonious
hypothesis. With respect to Schwartz and Gillette’s (1994) idea of a
“regional environmental crisis” we note that the Rock Point Formation
exposures can be followed for many kilometers around the Whitaker
Quarry, but no other bonebeds occur at the same stratigraphic level. This
suggests that a unique, localized event killed the dinosaurs at the Whitaker
Quarry.
Our examination of nearly all of the articulated skeletons from the
quarry provides some clues as to their death and burial: 1. In all com-
plete, articulated skulls, the lower jaws are clamped shut. 2. No individu-
als are found lying on their backs. Many are on their sides and some are
upright, but none are upside-down. 3. Practically all of the skeletons
exhibit the oposthotonic posture described by Faux and Padian (2007)
that indicates death under hypoxic conditions. These indicators, taken
together with the fact that the dinosaur skeletons are contained in an
overbank flood deposit, give a fairly strong impression that the animals
were overcome, drowned and buried by a flood event, and that they may
have been alive and struggling to remain upright during the process.

FIGURE 27. A, Rose diagram of the alignment of mudstone ripup clasts in

unit 4 of the NMMNH Whitaker Quarry block (Fig. 26). B-C, Large ripups
(B) and carbonaceous particles (C) with reduction halos in unit 6 of Fig. 26.
 27
INVERTEBRATE FOSSILS FROM THE WHITAKER QUARRY
numerous growth rings, and lack radial ribs and rostral spines
The upper 2 cm of unit 4, ~ 65 cm below the main dinosaur (Fig. 28A-D). Growth rings are double, and intra-ring ornamentation
bonebed and ~ 10 cm above the fine sand of Bed 1, contain conchostracans consists of anastomosing longitudinal striae (hachure marks) superim-
and ostracodes. No ripup clasts or other flow structures are present in posed on a finely punctate background. Umbones are located at ~ 35% of
this crustacean-bearing bed or below it. The grain size varies from silty to length and are only very slightly inflated. Thirty-eight of the recovered
muddy. Generally, the conchostracans are present in the muddier areas, specimens were measured using precision digital calipers; lengths varied
and ostracodes are found in the siltier portions, although they are occa- from 0.82 mm to 5.26 mm and averaged 4.08 mm. Measured lengths,
sionally found in very close proximity to one another. A cursory descrip- heights, and descriptive statistics are given in Appendix 2.
tion and estimate of the taphonomic importance of these invertebrates The height-to-length ratio of the sample is 0.68 (descriptive sta-
was presented earlier (Rinehart et al., 2004c). Here we provide thorough tistics, Appendix 2). Relative growth of the Whitaker Quarry
descriptive detail and ecological interpretation based on the presence of conchostracans over the size range of our sample is shown (Fig. 29A),
these crustaceans, which were previously unknown at the Whitaker where Base-10 log transforms of the length and height data are plotted.
Quarry. Growth is approximately isometric as indicated by slope of the regres-
sion line (0.96; ~1). The conchostracans essentially do not change shape
Conchostracans during ontogeny.
As is typical of the Conchostraca, the Whitaker quarry sample
Description appears to be monospecific (Pennak, 1978; Webb, 1979). The
All reasonably well-preserved specimens of the Whitaker Quarry conchostracans from Ghost Ranch have been identified by Heinz Kozur
conchostracans have pellucid, ovate valves with smooth margins, show (pers. comm., 2009) as a new species of Shipingia which is indicative of

FIGURE 28. A-D, Shipingia new sp. from the NMMNH Whitaker Quarry block C-8-82. Mean valve length 4.08 mm.
28
duration of the body of water (pond, pool, puddle). This assumes that
the water body is ephemeral, which is generally the case (Frank, 1988).
To find the age of the oldest individual, we first estimated the height (the
dimension that is the sum of the growth rings) of the largest individual by
probability plotting (Fig. 29B). We assume that the largest individual
would be a 3-sigma outlier. We estimate the height of an individual at the
3-sigma probability level (99.85 %) by extending the height data trendline
to be approximately 4.1 mm. Measured growth rings of the Shipingia in
our sample average 0.113 mm, yielding an approximate age for this large
individual of:
(4.1 mm / 0.113 mm/ring) ? 3 days/ring = 109 days.
Applying this calculation, the pond containing the conchostracans
had a duration of ~ 3.6 months. Webb (1979) agrees that some
conchostracans survive for the duration of their pond and cites an ex-
ample of cyzicids living for many months (Massal, 1953), but notes that
molting rate varies based on temperature and age, and argues effectively
against Tasch’s average 3-day intermolt period. We find, however, based
on Webb’s (1979) molting period data in which the absolute extremes are
1 and 11 days, that Tasch’s method should provide a good order of
magnitude estimate. Thus, the Whitaker Quarry pond probably survived
for several weeks to a very few months – not a few days, and not many
months.
Both the allometry plot (Fig. 29A) and the probability plot (Fig.
29B) show a group of three outliers that are statistically significantly
different from the main population. ANOVA testing shows the groups to
be different at the > 99% confidence level (Appendix 3). It is possible
that these few outliers represent a different species or are a product of
hydraulic sorting or other taphonomic process. We suspect that they are
a second cohort of Shipingia because they have a similar appearance and
show the same height-to-length characteristics (Fig. 29A). According to
Webb (1979), separate generations may inhabit the same pool in the
same season. Some conchostracans require desiccation of the eggs to
prepare them for hatching, whereas others do not. It is therefore possible
that this “second generation” could have been produced by rain increas-
ing the size of the pond and incorporating new, dry, egg-bearing mud
(Massal, 1953) as opposed to their being direct offspring of the older,
larger cohort.
Ostracodes
FIGURE 29. A, Allometric plot of LOG height versus LOG length in the
Whitaker Quarry Shipingia sample. Linear regression curve fit. N = 38. B, We assign the Whitaker Quarry ostracodes to Darwinula sp., a
Normal cumulative probability plot of lengths and heights of the Whitaker taxon with a temporal range of Pennsylvanian to Recent (Benson et al.,
Quarry Shipingia sample. Two component distributions probably represent 1961). Darwinula is the common Chinle Group ostracode, and it is well
two cohorts. N = 38. known from the homotaxial Redonda and Sloan Canyon formations of
a late Norian age (cf. Kozur and Weems, 2005, 2007). The Whitaker the Chinle Group in eastern New Mexico (Kietzke, 1989). The Whitaker
Quarry conchostracans comprise NMMNH specimen numbers P- 41277 Quarry ostracodes comprise specimen numbers NMMNH P-41294, P-
through P-41297, P-41299 through P-41301, P-41401 through P-41409, 41297, and P-42572. Each specimen number comprises numerous indi-
P-41418, and P-42570. viduals.

Discussion Description
Conchostracans typically live only in ephemeral bodies of fresh The ostracodes from block C-8-82 appear to be a monospecific
water (Webb, 1979). Their presence in strata underlying the primary assemblage. The valves are ovate in shape and lack ornamentation. The
bonebed allows us to infer aspects of the depositonal environment prior larger individuals are ~ 0.5 mm long and ~ 0.24 mm high. The right valve
to the deposition of the vertebrate skeletons. Based on comparison with is larger than the left valve, and the anterior end is smaller than the
extant conchostracans, the presence of Shipingia in the Whitaker quarry posterior end (Fig. 30).
probably indicates an ephemeral pond of less than one hectare (10,000 The cosmopolitan distribution and large temporal range of
m2), less than one meter in depth, and Ph between 6 (slightly acidic) and Darwinula make it of little biostratigraphic significance. Extant Darwinula
9.7 (alkaline) (Webb, 1979). In the following paragraphs we document prefer larger, fresh to slightly brackish lakes with sandy to muddy bot-
the data leading us to this inference. toms (Hoff, 1942; Swain and Brown, 1972). Triassic darwinulids, how-
Growth rings show the number of moltings, which occur (at least ever, are believed to have inhabited warm, shallow, calcium carbonate-
in extant taxa) approximately every three days according to Tasch (1973). mineralized lakes in semiarid to arid environments (Neustrueva, 1977).
Conchostracans often survive for the duration of the water body in
Paleoecological Implications of the Invertebrate Fossils
which they live (Frank, 1988). Therefore, Tasch (1973) maintains that if
the age of the oldest individual can be ascertained, it will approximate the The presence of both Shipingia and Darwinula in muddy to silty
 29
sediment just above a bed of fine sandstone indicates the presence of a
pond with a sandy to muddy bottom. The Whitaker Quarry pond was
probably small and shallow (one meter deep or less), possibly slightly
alkaline, and likely ephemeral, with a duration of several weeks to a few
months. The pond was located on the floodplain of a river in a strongly
seasonal, monsoonal environment with extended dry seasons and in-
creasing aridity, as indicated by other studies of this stratigraphic inter-
val (Dubiel et al., 1991; Tanner, 2003) The Whitaker Quarry pond may
have been a small cut-off meander (oxbow) lake. The river would have
been large enough to support phytosaurs and large coelacanth fish, the
latter of which required a perennial water source (Tanner, 2003).

FIGURE 30. Darwinula sp. from the Whitaker Quarry block C-8-82.
30
FISHES FROM THE WHITAKER QUARRY tion of the former.
Schaeffer (1967) described and illustrated a poorly preserved,
Fish specimens from the NMMNH Whitaker Quarry block that “indeterminate” redfieldiid fish skull from the Whitaker Quarry (AMNH
are illustrated in this section were selected based on taxonomic signifi- 5719). He stated that the specimen bore some resemblance to Synorichthys
cance, preservational quality, and state of preparation. in its cheek and opercular pattern and in the ornamentation of the dermal
Two identifiable fish taxa are present: the coelacanth Chinlea elements, but that the scales near the shoulder girdle were probably too
sorenseni and a redfieldiid that we tentatively assign to Synorichthys cf. large to support assignment to Synorichthys. The new fish material from
S. stewarti. Synorichthys is known from the Upper Triassic Redonda block C-8-82 is much more complete and better preserved. We think that
Formation of the Chinle Group (Johnson et al., 2003), which is contem- at least some of it supports a tentative assignment to Synorichthys cf. S.
poraneous with the Rock Point Formation. Dozens of isolated coela- stewarti based on dermal skull bone ornamentation (Schaeffer, 1967).
canth, redfieldiid, and unidentified scales and bones, and a few apparent Final assessment of the redfieldiid material awaits preparation and fur-
fish coprolites were recovered from block C-8-82. All these specimens ther study.
were found in the fining-up overbank flood units just below the tetra- NMMNH P-14302- This specimen is a dermal skull element of a
pod-bearing layer (unit 4 in Figure 26), but not in the underlying pond small redfieldiid fish (Fig. 31A). Identification is based on the ornamen-
facies (unit 3 in Figure 26). Most of the material has not been prepared. tation, and comparison to illustrated and described specimens (Schaeffer,
We assigned the coelacanth remains to Chinlea sorenseni based on 1967; Schaeffer and McDonald, 1978).
comparison of the basisphenoid with illustrations and descriptions of NMMN P-41304- An articulated fin of uncertain asssignment. Its
the basisphenoids of Chinlea sorenseni and Diplurus newarki in Schaeffer overall shape suggests a pectoral or pelvic fin rather than a midline fin
(1967) and Moenkopia wellesi and the extant Latimeria in Hunt (1997). (Fig. 31B).
We also made direct comparison to the holotypic specimen of Quayia NMMNH P-41411 and NMMNH P-41412- Isolated fin rays
zideki (NMMNH P-16888), which is named in Hunt (1997). The possibly pertaining to Synorichthys (Fig. 31C-D). The anteriormost fin
basisphenoids of Chinlea and Diplurus are more similar to each other rays are larger than the more posterior rays, and exhibit the small, ante-
than to those of the other coelacanths, but differences in the well-pre- riorly-projecting barbs seen in specimens of Synorichthys (e.g., Schaeffer,
served antotic processes and the sphenoid condyles support identifica- 1967; Schaeffer and McDonald, 1978)

FIGURE 31. A, NMMNH P-14302, a dermal skull element of a small redfieldiid fish. B, NMMNH P-41304, a nearly complete, articulated fin. C, NMMNH
P-41411, an isolated fin ray possibly pertaining to a Synorichthys. D, NMMNH P-41412, an isolated fin ray possibly pertaining to a Synorichthys.
 31
NMMNH P-41417- A nearly complete specimen of Synorichthys (Fig. 33C) that probably pertains to Chinlea sorenseni. Also see P-
cf. S. stewarti (Fig. 32A). The skull and mandible exhibit ornamentation 41422.
distinctive of the species (Fig. 32B).
NMMNH P-41420- An isolated incomplete basisphenoid of the
coelacanth, Chinlea sorenseni (Fig. 33A).
NMMNH P-41422- A part and counterpart coelacanth scale (Fig
33B). Also see P-42305.
NMMNH P-41441- A pair of part and counterpart slabs contain-
ing a disarticulated, probably partial, skull of Chinlea sorenseni (Fig.
33D). Identification is based on its distinctive basisphenoid (directly
above the “10 cm” inscription on the scale).
NMMNH P-41599- A partial, articulated specimen of Synorichthys
(Fig. 34A).
NMMNH P-42136- A large, incomplete cleithrum of an
osteichthyan (Fig. 34B).
NMMNH P-42305- A single, well-preserved scale of a coelacanth

FIGURE 33. A, NMMNH P-41420, an isolated incomplete basisphenoid of

the coelacanth Chinlea sorenseni shown in dorsal aspect. B, NMMNH P-
41422, a part and counterpart coelacanth scale (mm scale). C, NMMNH P-
42305, a single well-preserved scale of a coelacanth fish (mm scale). D,
NMMNH P-41441, part and counterpart matrix slabs containing a
disarticulated skull of Chinlea sorenseni (cm scale).

FIGURE 32. A, NMMNH P-41417, a nearly complete specimen of

Synorichthys cf. S. stewarti (cm scale). Skull located under the black arrow. FIGURE 34. A, NMMNH P-41599, an incomplete, articulated specimen of
B, The skull and mandible in closeup view show taxonomically distinct Synorichthys (cm scale). B, NMMNH P-42136, a large, incomplete cleithrum
ornamentation. of an osteichthyan (mm scale).
32
weakly laterally compressed, and recurved. It lacks serrations, but there
TETRAPODS OTHER THAN COELOPHYSIS
is a weakly developed carina on the lingual side of the mesial edge. Both
FROM THE WHITAKER QUARRY the labial and lingual surfaces are ornamented with vertical (basal to
Microvertebrate teeth apical) lines. Basally these consist of a recurring sequence of a dark
reddish lines followed by two thinner orange lines. The orange lines tend
We recovered five isolated teeth from the block that we describe to break up and anastamose more apically, but on the lingual surface
separately here (Fig. 35). They are important in documenting taxa that these lines correspond to ridges that extend apically from the base. The
are poorly represented (or not represented) in the block’s macrofauna. overall morphology of the tooth suggests affinities with phytosaurs—it
For purposes of describing these teeth we follow the terminology of appears to be an unserrated, or “type U,” tooth as originally described
Heckert (2004), who provided the most synthetic description of Chinle by Hunt (1994) and followed by Heckert (2004). “Type U” teeth corre-
microvertebrates to date. Taxonomically, Heckert (2004) followed spond to the “premaxillary” teeth of Hungerbuhler (2000). Matrix (nec-
Godefroit and Cuny (1997) in recognizing teeth that are “thecodont, essary to preserve the tooth base) obscures the base, but the tooth
conical and pointed mediolaterally, possess anterior and posterior keels appears to be a largely intact crown, so we interpret it as a type U shed
and are frequently serrated” (Heckert, 2004, p. 18) as archosauriform. phytosaur tooth.
All of the teeth described here possess these features, and are therefore The next largest tooth, NMMNH P-41307, is incomplete, but
identified as archosauriform, with some teeth having affinities to less was probably at least moderately tall, conical, and laterally compressed
inclusive taxa. (Fig. 35D-F). As preserved it is ~5 mm tall and slightly more than 3.5 mm
The largest isolated tooth, NMMNH P-41306, is slightly more long. It is conical, but the mesial and distal edges are not symmetrical,
than 10 mm tall as preserved, and would probably be between 12-13 mm giving the tooth the appearance of being slightly canted, or backswept,
tall if it was complete (Fig. 35A-C). The base is as much as 4 mm long, distally. Both edges bear numerous, relatively coarse (8/mm) denticles
but only slightly more than 2 mm wide. This tooth is moderately tall, that form serrated carinae. The tip appears to have been worn or even

FIGURE 35. Isolated microvertebrates from the Whitaker quarry in the NMMNH block. A-C, Possible phytosaur type U tooth (NMMNH P-41306) in
A, labial, B, lingual and C, mesial views. D-F, Archosauriformes indet. tooth (NMMNH P-41307) in D, labial, E, lingual and F, mesio-distal views. G-I,
Archosauriform tooth that may pertain to a sphenosuchian (NMMNH P-41600) in G, labial, H, lingual and I, mesio-distal views. J-K, Archosauriform
indet. tooth (NMMNH P-42571) in J, labial, K, lingual and L, mesio-distal views. Scales = 10 mm for A-C, 2 mm for D-I and 1 mm for J-L.
 33
broken and then subsequently worn. The mesial edge is also somewhat The holotype of Whitakersaurus bermani was recovered from
worn. The tooth somewhat resembles that of an indeterminate the bonebed, approximately 30 cm below the stratigraphic top of the
archosauriform from the Colorado City Formation of West Texas illus- block and within two cm of its westernmost edge. Whitakersaurus is the
trated by Heckert (2004, fig. 20g-i); it is identified as Archosauriformes only Chinle sphenodontian known from more than a single dentulous
indet. element. Heckert et al. (2008) considered Whitakersaurus to represent a
The best preserved microvertebrate tooth from the block, mosaic of both primitive and derived sphenodontian features. For ex-
NMMNH P-41600, is moderately low, leaf-shaped (phylloform), and ample, the teeth are numerous (19 positions), relatively simple conicals,
strongly laterally compressed (Fig. 35G-I). As preserved it is ~4 mm tall, and possess only modest ornamentation, unlike those of derived taxa
as much as ~2.5 mm long, but only ~1 mm thick. We interpret this tooth such as Planocephalosaurus and Clevosaurus. More derived character-
as either a left upper or right lower tooth. The mesial edge is slightly istics include its entirely acrodont dentition that is heterodont with much
expanded relative to the base and the incomplete distal edge is damaged in larger, weakly striated posterior (distal) teeth than the smaller, smoother
this region. The relatively complete base suggests that this is a shed anterior (mesial) dentition (Fig. 36).
tooth. It is serrated mesially and distally, with numerous, relatively
coarse (8/mm) denticles that are slightly subperpendicular to the tooth’s Postosuchus kirkpatricki Chatterjee, 1985
edge. The bulbous shape (basal expansion) of the tooth is somewhat The rauisuchian Postosuchus kirkpatricki has been known from
similar to those of sphenosuchians, although it is more laterally com- the Whitaker quarry since the 1990s. Thus, Long and Murry (1995, p.
pressed than most sphenosuchian teeth. Harris et al. (2000) used the 236) listed an “articulated skeleton of small individual missing skull,
term phylloform to describe leaf-shaped teeth in the sphenosuchian neck and forelimbs” in the Carnegie Museum of Natural History collec-
Phyllodontosuchus. The phylloform tooth described here matches the tion, as well as a “left ilium (92 mm long) and possibly associated hindlimb
posterior maxillary teeth in the sphenosuchian Hesperosuchus from the of a very small individual” in the Ruth Hall Museum collection (RHMP
Whitaker quarry described and illustrated by Clark et al. (2000, p. 687), GR 1007). Long and Murry (1995, figs. 145-146) illustrated parts of the
as “not recurved, and have convex anterior and posterior carinae,” yet, as CM skeleton and the RHMP left ilium. Novak (2004) described the
illustrated (Clark et al., 2000, figs. 1-2), many of the posterior maxillary skeleton, CM 73372, in her M.S. thesis, indicating that it also includes
teeth appear conical to recurved, as well as phylloform. four cervical vertebrae and a nearly complete left forelimb, but the mate-
The most bulbous of the isolated teeth, NMMNH P-41601, lacks rial has never been fully described in the scientific literature.
the tip but is moderately low to low, conical with a slightly recurved
apical portion, and bulbous. As preserved it is slightly more than 4 mm Shuvosaurus okeeffeae (Nesbitt and Norell, 2006)
tall and was no more than 5 mm tall. It is approximately 3 mm long and
Chatterjee (1993) named Shuvosaurus inexpectatus for a nearly
2 mm wide. The labial surface is somewhat more convex than the lingual
complete skull, left mandible, a dorsal vertebra (holotype), and addi-
and possesses several baso-apical ridges that are more pronounced ba-
tional, less complete cranial material, part of an atlas and a right scapula
sally than apically. The base is very slightly expanded with a well-
(referred specimens) from the Post quarry in the Revueltian Bull Canyon
developed resorbtion pit, indicating that this is also a shed tooth. Both
Formation in West Texas. Chatterjee (1993) regarded Shuvosaurus as a
the mesial and distal carinae bear relatively coarse denticle (6/mm) serra-
coelurosaurian theropod dinosaur close to ornithomimosaurs based on a
tions. Apically the denticles comprising these serrations are perpendicu-
cladistic analysis, whereas Rauhut (1997, 2003) also considered
lar or nearly perpendicular to the tooth margin, except basally where
Shuvosaurus a theropod, albeit a more basal form. Long and Murry
they are more offset and worn. The mesial carina is displaced slightly
(1995) named Chatterjeea elegans for relatively small postcranial mate-
lingually, worn and/or damaged, and lacks the basal portion, yet it is
rial of an archosaur from the Post quarry. Chatterjee (1985) had earlier
sufficiently well preserved to suggest the presence of even coarser den-
considered these specimens to be juvenile postcrania of Postosuchus.
ticles. A few of the basal denticles on the distal carina appear split or
Puzzled by the relative abundance of postcrania and apparent absence of
cleft, although not to the degree seen in some taxa from the slightly older
cranial material of Chatterjeea in the Post quarry, Long and Murry (1995,
(Revueltian) Owl Rock Formation in northern Arizona (Kirby, 1991;
p. 162) suggested that the cranial material named Shuvosaurus might
Heckert, 2001). The tooth is similar to the morphotype H tooth illus-
belong with the postcrania named Chatterjeea. As Long and Murry (1995)
trated by Heckert (2004, fig. 40d-f) but more coarsely denticulated,
acknowledged, this would make Shuvosaurus the senior synonym of
suggesting assignment to Archosauriformes indet.
Chatterjeea.
The smallest tooth recovered, NMMNH P-42571, is incomplete,
Nesbitt and Norell (2006) erected and Nesbitt (2007) described
with much of one of the mesio-distal margins missing (Fig. 35J-L). How-
Effigia okeeffeae for associated cranial and postcranial specimens from
ever, a basal resorbtion pit appears intact, albeit filled with sediment,
the Whitaker quarry. Because the skull of Effigia is nearly indistinguish-
indicating that most of the tooth is present in spite of its superficial
able from that of Shuvosaurus (Fig. 37) and the postcrania are remark-
resemblance to a tooth tip. The tooth is very low, laterally compressed,
ably similar to that of Chatterjeea, Lucas et al. (2007a) argued that the
and conical, forming a low triangle in lateral view. As preserved, the tooth
two taxa are synonymous. The associated postcrania and skull of Effigia
is barely 2 mm tall, but more than 2 mm long and less than 1 mm wide.
demonstrate that the Shuvosaurus skull belongs with a Chatterjeea
The mesial and distal carinae bear relatively coarse serrations (~8/mm)
postcranium, as was postulated previously (Long and Murry, 1995;
that are slightly finer near the tip of the crown. Heckert (2004, fig. 113f-
Hunt et al., 1998; Nesbitt and Norell, 2006; Nesbitt, 2007). Thus,
g) ascribed a somewhat similar tooth to a young phytosaur as a type B
Chatterjeea elegans Long and Murry, 1995 is a junior subjective syn-
tooth, but P-42571 is very low and too strongly laterally compressed to
onym of Shuvosaurus inexpectatus Chatterjee, 1993. Furthermore, the
readily assign to a phytosaur, so we assign it to Archosauriformes indet.
associated postcrania described by Nesbitt and Norell (2006; Nesbitt,
Whitakersaurus bermani Heckert, Lucas, Rinehart and Hunt, 2008 2007) demonstrate that Shuvosaurus is a suchian archosaur as argued by
several authors (Long and Murry, 1995; Hunt et al., 1999; Heckert and
Heckert et al. (2008) named a new sphenodontian, Whitakersaurus Lucas, 2000; Nesbitt et al., 2007), and not a theropod dinosaur, as main-
bermani, for left and right dentaries, an incomplete left? maxilla? and tained by Chatterjee (1997) and others (Rauhut, 1997, 2003; Lehane,
impressions of a probable palatal element from the NMMNH Whitaker 2005). Lucas et al. (2007a) tentatively regarded the Ghost Ranch mate-
quarry block (Fig. 36). This is the most complete small vertebrate recov- rial as a distinct species, S. okeeffeae, distinguished from S. inexpectatus
ered from the Whitaker quarry and one of the youngest Chinle Group by the difference in the maxilla originally used to differentially diagnose
sphenodontians.
34

FIGURE 36. NMMNH P-43125, Whitakersaurus bermani, A, C, E, G, SEM micrographs and B, D, F, H, interpretive sketches (B,D,F,H) of the holotype.
A-B, Anterior portion of left dentary in occlusal view, C-D, posterior teeth of left dentary in oblique occlusal view; E-F, left(?) maxilla(?) fragment in
labial(?) view, G-H, incomplete right dentary in lateral view.

Effigia from Shuvosaurus. S. inexpectatus has a well-documented early referred to this fossil as an unusual archosauriform, but made no more
Revueltian distribution in Texas-New Mexico (Hunt, 2001), whereas S. precise identification. Hunt et al. (2002) identified it as Vancleavea campi,
okeeffeae is well documented only from the Apachean Whitaker quarry a taxon previously known from an incomplete postcranium that includes
in New Mexico. very specialized and distinctive osteoderms (e.g., Long and Murry, 1995;
A 40 mm-long, indeterminate reptilian scapula and coracoid, Hunt et al., 2005). A complete description of RHM GR 138 was recently
NMMNH P-42554, may be assignable to Shuvosaurus (Fig. 38A). published (Nesbitt et al., 2009). The NMMNH Coelophysis block from
the Whitaker quarry also contains osteoderms and apparent bones of
Vancleavea campi Long and Murry, 1995 Vancleavea campi (Fig. 39).
Downs and Davidge (1997) reported an articulated skeleton of an NMMNH P-55351 is a largely unprepared, partial, disarticulated
armored reptile in the Ruth Hall Museum collection (RHM GR 138) that skeleton and dermal armor of the archosauromorph Vancleavea campi
they identified as a choristodere. Subsequently, Small and Downs (2002) that was removed from the block. Several dorsal and a few appendicular
 35

FIGURE 37. Comparison of the skulls of “Effigia” (=Shuvosaurus) and

Shuvosaurus. A, Drawing of left side of holotype skull of “Effigia” okeeffeae,
AMNH FR 30587, as preserved (after Nesbitt, 2007, fig. 4). B,
Reconstruction of skull of “Effigia” okeeffeae in left lateral view (after
Nesbitt and Norell, 2006, fig. 1). C, Reconstruction of skull of Shuvosaurus
inexpectatus in left lateral view (after Lehane, 2005, fig. 5). Diagonal lines
represent reconstructed portions. Abbreviations: a, angular; aof, antorbital
fenestra; ar, articular; bo, basioccipital; bs, basisphenoid; de, dentary; en,
external nares; f, frontal; j, jugal; l, lacrimal; ltf, lateral temporal fenestra;
mf, mandibular fenestra; mx, maxilla; n, nasal; o, orbit; op, opisthotic; p,
parietal; pfr, prefrontal; pmx, premaxilla; po, postorbital; ps, parasphenoid;
q, quadrate; qj, quadratojugal; sq, squamosal; stf, supratemporal fenestra;
su, surangular.

dermal ossicles have been prepared and are diagnostic (Fig. 39). The
specimen originally rested near the west edge of the block below the main
bonebed in an area that was disrupted by numerous east-west trending
joints.
Hesperosuchus agilis Colbert, 1952
Clark et al. (2001) described an incomplete, articulated skeleton in
the Carnegie Museum of Natural History collection (CM 29894) of the
crocodylomorph Hesperosuchus agilis from the Whitaker quarry. This
skeleton is one of the most complete and well-preserved of a basal FIGURE 38. A, NMMNH P-42554, indeterminate reptile scapula. B,
crocodylomorph and provided new anatomical information about an NMMNH P-55352, the incomplete skeleton of an unidentified reptile (scale
otherwise poorly known taxon (Colbert, 1952; Long and Murry, 1995). in cm). C, NMMNH P-57654, Hesperosuchus, associated osteoderms in
The NMMNH Coelophysis block from the Whitaker quarry also con- ventral view.
tains some osteoderms of Hesperosuchus. Redondasaurus bermani Hunt and Lucas, 1993
NMMNH P-57654 consists of at least five rhombic to rectangu-
lar osteoderms assignable to Hesperosuchus which are preserved on the Hunt and Lucas (1993b, p. 193) named the phytosaur
surface of C-8-82-J1 (Fig. 38C). Excluding the anterior point, the largest Redondasaurus bermani for CM 69727, a skull and lower jaw from the
of these measures 17 x 15 mm, and the smallest measures 12 x 8 mm. The Whitaker quarry (Fig. 40). Ballew (1989, p. 331) had earlier identified
osteoderms are exposed in ventral view, but where they are broken the this skull as Pseudopalatus buceros, and Long and Murry (1995, p. 58)
dorsal pit sculpturing pattern that supports their referral to assigned it to Arribasuchus buceros (Cope, 1881), thereby synonymiz-
Hesperosuchus, is impressed in the matrix. ing R. bermani and A. buceros. Spielmann et al. (2006, p. 508) reiterated
36
lepidosauromorph reptile Drepanosaurus. This assignment was based
on: a slight mediolateral expansion at the dorsal end of the tall, narrow
scapula; the angle of the scapular blade with respect to the acromion; a
coracoid that is dorsoventrally taller than its craniocaudal length; fusion
between the scapula and coracoid; and a glenoid fossa that is craniocaudally
compressed and bordered dorsoventrally by a distinct ridge. All these
features are unique to Drepanosaurus, within the Drepanosauromorpha.
Given the isolated nature of the specimen, Renesto et al. (2009) refrained
from assigning this specimen to a species of Drepanosaurus.
GR 1113 was the first described material of a drepanosauromorph
from the Whitaker Quarry. Subsequently, additional drepanosauromorph
fossils have been illustrated and described by Renesto et al. (2010;
summarized below) and various other specimens remain to be described
(JAS, pers. obs., 2009; A. Downs, pers. commun. to LFR, 2009).
Drepanosauromorph indet.
NMMNH P-57651 is an isolated, partially articulated hindlimb
of a drepanosaurosauromorph (NMMNH P-57651; Fig. 41B). Portions
of the fibula, tibia, astragalus, calcaneum and various elements from
digits 1 through 4 are present. Both the astragalus and calcaneum are
present, are not fused, and only slightly disarticulated from the tibia and
fibula. The ungual phalanx possesses a recurved claw and large insertion
site for the flexor musculature that is characteristic of
FIGURE 39. Osteoderms of the armored reptile, Vancleavea, NMMNH P- drepanosauromorphs. The clawed ungual phalanx and the morphology
55351. of the astragalus and calcaneum confirm the drepanosaurid identity of
this specimen. This specimen was originally described in detail by Renesto
morphological features that distinguish Redondasaurus from et al. (2010).
Pseudopalatus (= Arribasuchus), as did Hungerbühler (2002).
Indeterminate reptile
Redondasaurus thus remains a distinct phytosaur genus and an index
fossil of the Apachean land-vertebrate faunachron (Lucas, 1998; Lucas NMMNH P-55352 is an unidentified non-Coelophysis specimen.
and Tanner, 2007; Lucas et al., 2007b). It comprises an incomplete shoulder girdle, five vertebrae, and a partially
NMMNH P-44920 comprises a nearly complete, slightly disar- exposed limb bone, possibly a humerus (Fig. 38B).
ticulated, juvenile phytosaur skull and lower jaws (Fig. 41A). The skull
probably pertains to Redondasaurus. It resides on the partially pre- Eucoelophysis sp.
pared surface of jacket C-8-82-J1 together with P-57653, a Coelophysis A single specimen, TMP84-63-33, that we assign to the
bauri skull, disarticulated cervical vertebrae and ribs, P-57654, several dinosauromorph Eucoelophysis, is one of 34 specimens on the Royal
Hesperosuchus osteoderms (Fig. 38C), and numerous uncataloged shoul- Tyrrell Museum block. All other specimens on the RTM block appar-
der, forelimb, and foot (manus and pes) bones. The snout tip is missing ently represent Coelophysis bauri. The RTM block, which is part of C-
off the east edge of the jacket. There is no sign of the snout tip “bulb,” 1-82, has been assigned RTM number TMP 2006.00.30. Each specimen
but we estimate that only 10 to 20 mm of the snout are missing, based on on the block bears an individual specimen number.
the length of the preserved portion and tooth count. As preserved, the Our assignment of TMP84-63-33 to Eucoelophysis is based prin-
specimen is 212 mm long from the broken snout tip to the squamosal tip, cipally on comparison of its proximal femur morphology to that of the
and 197 mm from the broken snout tip to the quadrate. The mandible tip holotype of Eucoelophysis baldwini, NMMNH P-22298. The proximal
is similarly broken and missing; this element measures 195 mm to the femur end shape and femoral sulcus, lesser (anterior) trochanter, and
retroarticular process tip, and the mandibular fenestra measures 51.3 mm anterolateral trochanter are particularly informative and will be discussed
by 7.3 mm. There are 28, mostly empty, maxillary alveoli. The center-to- below.
center spacing of the tooth sockets varies from 3.7 to 5.5 mm from the TMP84-63-33 comprises a partial presacral series, a complete
posterior to the anterior end of the series. sacrum and pelvic girdle, a partial caudal series, a partial shoulder girdle
The skull is obliquely sheared, so as to be exposed in both palatal and forelimb, and complete hindlimbs and peds exposed in left lateral
and left lateral views. The palate, including the maxillae, tapers from 28.5 view (Fig. 42A). A data sheet with complete metrics is in the RTM
mm near the posterior end of the tooth row to 9.8 mm at the broken section of Appendix 1.
snout tip. Only the ventral portion of the orbit border is exposed. The Sixteen presacral vertebrae are present. One through 13 (numbers
radius of curvature of this orbit segment is large, indicating a relatively proceeding anteriorly from the sacrum) are clearly dorsal vertebrae, whereas
large orbit. The orbits are oriented dorsally. The antorbital fenestra is number 14 is covered by TMP84-63-32. It is unfortunate that number
25.6 mm long anteroposteriorly and 7.4 mm dorsoventrally. The lateral 14 is covered, because this is the cervical-dorsal transition region and it is
temporal fenestra is large and measures 30 mm dorsoventrally and 20 not possible to ascertain whether this specimen has 13 dorsals, as does
mm anteroposteriorly. Skull height at the quadrate is approximately 36 Coelophysis bauri. Assuming that there are 13 dorsal vertebrae, the dor-
mm. sal series is 360 mm long. Fifteen and 16 are cervical vertebrae; they are
Drepanosaurus sp. longer than the dorsal vertebrae and possess pleurocoels. Number 15 is
30.7 mm long, and number 16 is partially covered anteriorly and
Harris and Downs (2002) originally described an isolated, three- unmeasurable.
dimensionally preserved pectoral girdle (GR 1113) in the Ghost Ranch The 75.5 mm-long sacrum comprises 4 vertebrae (Fig. 42B). S3
Whitaker Quarry block. After re-examination of GR 1113, Renesto et al. and s4 are not fused. The sacral neural spines have not been exposed.
(2010) concluded that this pectoral girdle pertains to the Seven caudal vertebrae are present; six of which are measurable. The
 37

FIGURE 40. A-E, CM 69727, Redondasaurus bermani, holotype skull in A, dorsolateral, B, dorsal and C, occipital views, D, close up of posterior skull
in dorsal view, E, holotype mandible in right lateral view.
38

FIGURE 41. A, NMMNH P-44920, Redondasaurus, a very small, nearly complete, juvenile phytosaur skull in palatal and left lateral view. The lower jaws
are visible under the skull. B, NMMNH P-57651, the incomplete hindlimb and pes of a small drepanosauromorph. Scale in cm.
 39

FIGURE 42. A-C, E, RTM specimen TMP84-63-33, Eucoelophysis sp., an incomplete, articulated skeleton, A, Specimen in left lateral view, specimen tag
number 33 lies on the left ilium, and the scale bar lies between the left and right femora, B, Close up of pelvis and sacrum in left lateral view, C, Stereo pair
of the articular surface of the femur head in proximal view, showing the femoral sulcus, E, Stereo pair of the left femur head in anterior view, with the
prominent, crest-like lesser trochanter and the anterolateral trochanter. D, F, NMMNH P-22298, Eucoelophysis baldwini, holotype, D, Articular surface
of the femur head in proximal view, lighting conditions and angle account for much of the apparent difference in the depth of the sulcus, F, Left femur head
in anterior view.
40
caudal vertebrae in Figure 42A that appear to continue past caudal num- trochanter is slender, but prominent and crest-like with a sharp
ber seven actually belong to a Coelophysis specimen underlying TMP84- anteromedial edge. It is very distinctly different from the blocky lesser
63-33. trochanter of Coelophysis bauri and bears great similarity to Raath’s
The left scapula is present, but is covered proximally (Fig. 42A) (1969, 1990) gracile morph of Coelophysis rhodesiensis. A small, sharp,
whereas the coracoid is either missing or is still covered, and the glenoid distinct, crest-like anterolateral trochanter is located immediately ante-
fossa is not visible. The distal scapula is less rounded and the middle rior to the greater trochanter on the anterior surface of the femur head as
shaft portion is distinctly wider than in Coelophysis bauri. The left is seen in Eucoelophysis baldwini (Fig. 42E-F). This trochanter is unique,
humerus is disarticulated from the shoulder girdle and lies subparallel to and we consider it to be the single most diagnostic feature in our assign-
the anterior dorsal vertebral series (Fig. 42A). ment of TMP84-63-33 to Eucoelophysis. The fourth trochanter consti-
The sacrum is complete and articulated (Fig. 42B). The anterior tutes a narrow ridge approximately one-third the length of the femur that
portion of the ilium is fragmented but otherwise is complete. There is a terminates at about midlength of the femur.
well-developed supra-acetabular shelf, and a deeply sculpted The tibiae and fibulae are not fused to the astragalocalcaneum. The
posteroventral arch is evident. The pubes and ischia are complete, though tibia shows an appressed tibia surface (a wide, shallow sulcus to accom-
the puboischiatic plate is mostly covered by the femur.. The pubes are modate the fibular shaft) as in Eucoelophysis baldwini (Sullivan and
long and gently curved, whereas the ischia are shorter and straight. The Lucas, 1999).
pubes show a small, rounded, forward-projecting pubic boot. The ischia The ankle of TMP84-63-33 consists of fused astragalus and cal-
possess a smaller, posteriorly-projecting, bulb-like terminus. caneum that possess a short ascending processes both anteriorly and
The femur-tibia length ratios of TMP84-63-33 are virtually iden- posteriorly, and two additional tarsals that are very much more robust
tical (well within the 95% confidence interval) to those of Coelophysis than any observed in Coelophysis bauri. The medial tarsal measures 11.7
bauri. The femora are bowed anteriorly and show a mediolaterally trend- mm mediolaterally and >8 mm (partially covered) anteroposteriorly. It is
ing sulcus on their articular surfaces (Fig 42C) that is similar to that of 3.6 mm thick. The lateral tarsal measures 12.2 mm mediolaterally, 7.2
Eucoelophysis baldwini (Fig. 42D). The femur head is rounded and offset mm anteroposteriorly, and is 2.8 mm thick. Both are of subrectangular,
medially (Fig. 42E). The Eucoelophysis baldwini holotype femur head tabular form.
exhibits a somewhat more angular character in its offset, and the articular The feet are essentially indistinguishable from those of Coelophysis
surface is less rounded (Fig. 42F). This difference could be of taxonomic bauri. They show an attenuated first digit, and the fifth metatarsal is
significance or could be an artifact of preservation. We judge that it is reduced to a thin splint of bone. Phalangeal counts are identical to C.
probably due to the weathered condition of the holotype. The lesser bauri.
 41
DESCRIPTION AND BIOMETRICS
OF THE NMMNH COELOPHYSIS MATERIAL

Descriptions of the Coelophysis Material from Block C-8-82

Here we describe and illustrate the fossil material in NMMNH
block C-8-82. To the extent possible, every bone on the prepared block
surface has been measured. Descriptions proceed in order of specimen
number, and their biometric measurements are presented in Appendix 1
in the same order as they are described here.
Block C-8-82 contains no complete skeletons. Even the most
complete individuals lack a portion of the caudal series and/or some
distal limb elements. Portions of approximately seven adults or large
juveniles and more than 20 moderate sized to very small juveniles are
represented. Additionally, there are a great many isolated bones, articu-
lated but incomplete series, and disarticulated series.
Specimens that have not been removed from the block are de-
scribed as they appear on the prepared surface. Because the block was
prepared from the bottom-up (see the Preparation section of this mono-
graph), a specimen described as lying on its right side on the prepared
surface would have been lying on its left side in the quarry, and top
(superior) view represents a bottom (inferior) view in the sense of the
originally upright bonebed. The most disarticulated and weathered mate-
rial is closest to the original (upright) surface of the block and appears to
have been reworked. Most of the bones lie flat on the bedding planes,
although there are numerous exceptions. Most of the articulated speci-
mens are lying on their sides, but a few are exposed in ventral view, and
none are exposed in dorsal view. This suggests that at least some of these
animals may have been alive and struggling to stay upright as they were
buried or possibly that they naturally float in the upright position.

NMMNH P-41416
NMMNH P-41416 comprises a tibia and nearly complete fibula
and an articulated astragalocalcaneum (Fig. 43) which were removed
from the original upright surface of the block before it was turned over.

NMMNH P-41419
This specimen encompasses a disarticulated, incomplete skull
that was removed from the block’s original upper surface before it was
turned over. It includes portions of the palate and ventral elements of the
braincase, the basal skull, skull roof, nasals, maxillae, and premaxillae
(Fig. 44). The skull is approximately 200 mm long as preserved.

NMMNH P-42200
This specimen is apparently a small juvenile based on its 124.6
mm femur length and it is completely articulated (Fig. 45). The greater
part of the skeleton includes the skull, plus the cervical, dorsal, and sacral
vertebrae, seven proximal caudal vertebrae, incomplete shoulder girdle
and forelimbs, complete pelvic girdle and hindlimbs except for the distal
ischia and some parts of the left pes. The presacral portion is exposed in
left lateral view and the sacrum and hindlimbs are rotated into ventral
view.
Skull – The skull is 123 mm long and almost perfectly preserved
in three dimensions (Fig. 45). The exposed (left) side of the skull is in
unit 6 (see Microstratigraphy section) which contains the best preserved
specimens in the block, whereas the unexposed (right) side is in dark
reddish brown siltstone of unit 7 in which preservation is generally
poorer. It is therefore likely that the unexposed side of the skull is also
less well preserved. The posteriormost portion, behind and ventral to
the supratemporal fenestrae, is slightly crushed, and the neck joins the
skull at a probably unnaturally acute angle of approximately 90o.
The orbit has been elongated 2.6 mm posterodorsally by slight
displacement of the postorbital. A complete sclerotic ring is present; it
and the orbit will be thoroughly described in a later section. FIGURE 43. NMMNH P-41416, Coelophysis bauri tibia and incomplete
fibula from the top of the originally upright block surface.
42

FIGURE 44. NMMNH P-41419, an incomplete disarticulated Coelophysis bauri skull from the top of the originally upright block surface. The quadrates
are at the right; nasals and skull roof bones are spread out to the left. The premaxilla is at the far left beneath the scale.

The mandibles are closed against the maxillae and premaxillae. The broken, but the perforation appears to form a very rounded rhombus,
dentition is completely intact, with the maxillary teeth overlapping the measuring ~6.7 mm anteroposteriorly and ~7.7 mm mediolaterally. It is
dentary teeth. The maxilla has 19 teeth and 23 alveoli in a space of 58 located ~ 3 mm distal (ventral) and medial to the obturator foramen.
mm. There are 14 teeth in the anterior 35 mm of the dentary toothrow; The hindlimbs are fairly well preserved but are intersected by
the remainder are covered by the maxillary teeth. Three teeth are pre- numerous joints passing through this area of the block. Near these joints,
served in the four alveoli of the premaxilla. some bone fragments are missing, apparently having fallen into the joints.
Postcrania – All 10 cervical vertebrae are well preserved. The Part of the right pes is disrupted and missing because of intersecting
atlas is partially covered by the posterior end of the skull. The entire joints. The hindlimbs are folded under the animal in the classic “death
series of cervical ribs are in articulation. These ribs are long and taper to pose” of large extant birds as described by Carpenter (2003) and found
a very small diameter (0.4 mm) distally and overlap as many as five by Faux and Padian (2007) to be associated with death under hypoxic
vertebrae near the center of the cervical series. and/or asphyxiation conditions.
The 13 dorsal vertebrae are well preserved and exposed except for The seven most proximal caudal vertebrae are articulated in a 127
dorsals three through five, which are covered by the incomplete shoulder mm-long series. Chevrons begin at the posterior end of caudal number
girdle and left humerus, and numbers 12 and 13, which are fractured by one. The tips of most of the chevrons were lost during excavation of the
joints, but are in place. The neural arches of the dorsal vertebrae are block.
almost completely covered by ribs and matrix, but the vertebral centra
are visible. Partial ribs are in articulation on most dorsal vertebrae. Five NMMNH P-42351
sacral vertebrae are present and are thoroughly fused. The incomplete This large adult specimen lies across the block surface in a near
shoulder girdle does not preserve a furcula. east-west orientation and is exposed in ventral view (Fig. 46A). Judging
The pelvis is complete except for the distal ischia, which were from its femur length of 181.5 mm, it is the largest individual on the
oriented vertically and were probably lost during the removal of a “field” prepared block surface, although femora as long as 240 mm have been
jacket directly above the specimen (Fig. 45). Part of the ischia may be reported in the Ghost Ranch material at the State Museum of Pennsylva-
present in this jacket, whereas some fragments are preserved in the nia. The specimen essentially comprises the dorsal and sacral series,
museum collections. (See the section on Preparation for details concern- incomplete pelvis, and some limb bones. The neck, with the exception of
ing the removal of jackets from block C-8-82.) The ilia are present but a fragment of cervical vertebra 10, is off the east edge of the block, and all
mostly covered. As they are exposed in ventral aspect, their posteroventral of the tail was removed in a plaster jacket and is unprepared. The 13
arches are prominent. The acetabulum is subcircular with a mean diam- dorsal vertebrae are well preserved and exposed in ventral aspect so that
eter of 23 mm. their neural arches are inaccessible. The dorsal ribs are present, but not
The pubes are complete except for some missing fragments in well preserved. The left scapula and coracoid are well preserved and lie
areas where they are intersected by joints in the block. Fortunately, across dorsal vertebra three, which is farther back than usually observed
impressions in the matrix retain the outline of these missing portions. in our specimens. Typically, the coracoid lies across dorsal vertebra
The left pubis is exposed so as to reveal the obturator foramen and the number one. All that remains of the manus is a single metacarpal.
larger pubic “ceratosaur” fenestra, which was not described by Colbert The pelvis is slightly dorsoventrally compressed. Because the
(1989), but was illustrated by Tykoski and Rowe (2004, fig. 3.7) and specimen is exposed in ventral view, the ilia are mostly covered except
previously by Rowe and Gauthier (1990, fig. 5.7). The presence of this for the posteroventral arches. The distal ends of both the pubes and
fenestra has been considered a diagnostic feature of the Coelophysidae ischia are missing. Portions of the distal pubes were lost to the adjoining
(Tykoski and Rowe, 2004). The medial border of the pubic fenestra is block, whereas portions of the distal ischia were lost to a large fracture
 43

FIGURE 45. An overview of NMMNH P-42200, a nearly complete Coelophysis bauri juvenile of the 1 yr size class. Inset, The skull of Coelophysis bauri
NMMNH P-42200, which is essentially three-dimensional and contains a complete sclerotic ring.
44

FIGURE 46. A, The largest adult Coelophysis bauri on the NMMNH block, NMMNH P-42351 is exposed in ventral aspect. Its femoral length places it
in the three-year-old size class. B, NMMNH P-42352 comprises the shoulder girdle through a nearly complete caudal series of a small juvenile Coelophysis
bauri. The left hindlimb lies tightly folded above the pelvis. One forelimb is parallel to the left edge of the scale, and the gastral basket lies to the right of
the scale.
 45
joint. Five sacral vertebrae are present; S1 through S4 are solidly fused, and articulated except for some of the manus elements, which may be
whereas S4 and S5 are not fused. The left hindlimb is well preserved, and present but buried.
as in nearly all the dinosaurs in the block, is folded up under the body. The complete pelvic girdle is exposed in right lateral view. The
The metatarsals are in place, but some of the phalanges are displaced. sacrum is visible and comprises five vertebrae. The first four sacral
The right hindlimb consists only of the femur and the proximal ends of vertebrae are firmly fused, and the fifth is partially fused; it appears
the tibia and fibula with the remainder being lost along the southern edge fused dorsally, but ventrally there is a well-defined separation between
of the block. The partial tibia and fibula were removed, revealing a promi- S4 and S5. Both femora are in place and the left is articulated to the tibia
nent lateral process on the distal end of the femur. and fibula, but both pedes are missing off the west side of the block.

NMMNH P-42352 NMMNH P-42354

NMMNH P-42352 is preserved lying in a nearly east-west direc- NMMNH P-42354 comprises an incomplete cervical series, com-
tion with the distal caudals off the west edge of the block (Fig. 46B). Its plete dorsal series, incomplete shoulder girdle, left forelimb and manus,
femur length (135 mm) places it in the small juvenile class. Most of the and part of the gastral basket (Fig. 48). The dorsal series length of 199
animal is exposed in left lateral aspect. An incomplete shoulder girdle and mm places it among the smallest juveniles on the block. This specimen
left forelimb that includes the manus, one or two poorly preserved cervi- was removed from the block with some of the underlying matrix.
cal vertebrae, a nearly complete dorsal vertebral series, a complete sacrum
and pelvis, a hindlimb with pes, and a proximal to mid-caudal series of 28 NMMNH P-42576
vertebrae are present. The gastral basket is nearly complete, the poste- This very small juvenile specimen is lacking the skull. However,
rior half being best preserved. Some fragmentary material near the front skull NMMNH P-50533 may belong to it based on size and relative
of the gastral basket may represent stomach contents, but could be position on the block. The skeleton lies on its right side and is partially
disarticulated fragments of the anterior gastralia. Chevrons are present covered by NMMNH P-42577 (Fig. 49A). An incomplete cervical ver-
on all the caudal vertebrae beginning at the posterior end of caudal num- tebral series consisting of 11 through 13 is present and is bent sharply
ber one. Caudal vertebrae 13-15 have been displaced into a small fault, posteriorly over the shoulder girdle. The shoulder girdle is complete and
but are prepared in lateral view. includes a furcula that is positioned over the scapulae-coracoids but
A coprolite is present in the vent area midway between the ischia rotated somewhat away from the acromion processes of the scapulae.
and the proximal caudals. The coprolite contains numerous small, flat Both forelimbs are present with the left manus and and carpus being
bone fragments together with hundreds of very small, rod-like elements. particularly well exposed and well preserved (Fig. 49B). The dorsal
At first impression they appear to be fish scale fragments, but micro- vertebral series, sacrum, and pelvis are present but mostly covered. The
scopic examination could not positively confirm this. The coprolite is hindlimbs and feet are present, but are folded up under the animal and
illustrated and further described below. only partially visible.

NMMNH P-42353 NMMNH P-42577

NMMNH P- 42353 is preserved on the southwest corner of the This specimen represents a very small, nearly complete presacral
block. A skull length of 185 mm and a femur length of 152 mm suggest skeleton exposed in left lateral aspect (Fig. 50A). The skull is missing its
that it represents one of the larger juveniles or a small adult. The skeleton snout and is exposed in dorsal view (Fig. 50B). A large joint in the block
lies on its left side with the body and neck arched toward one another in interrupts the otherwise complete cervical series. The cervical series is
a complete loop with the skull directly over the ilia (Fig. 47A). This bent sharply posteriorly and dorsally to the vertebral column in the
opisthotonic posture, again, probably indicates death under hypoxic classic opisthotonic death pose. The shoulder girdle is complete and the
conditions (Faux and Padian, 2007). The hindlimbs are flexed, but are not furcula is in articulation. The forelimbs and manus, and the eight
pulled up closely under the body. anteriormost dorsal vertebrae are also present (Fig. 50C). The gastral
Skull – The skull is flattened obliquely, exposing the skull roof basket is complete, articulated, and seen in ventral view (Fig. 50D). This
and right side (Fig. 47B). Approximately 2 mm of the premaxillae were specimen overlies the posterior portion of NMMNH P-42576.
lost along the south margin of the block. A total of five sclerotic ossicles
are present in and near the orbit. Two of these are articulated and remain NMMNH P-42578
in the orbit. Three other articulated ossicles were removed from the area
NMMNH P-42578 comprises an incomplete dorsal vertebral se-
directly below the orbit. The ossicles show the same overall shape,
ries, shoulder girdle, gastral basket, and the forelimbs and manus of a
imbrication, and sigmoid cross section as those in the sclerotic ring of P-
very small individual (Fig. 51A-C). Based on their position relative to
24400.
the shoulder girdle, the vertebrae are probably dorsals three through
Two hyoids extend anteroventrally from beneath the lower jaw.
eight. A folded right hindlimb with a perfectly preserved pes of this
The exposed portions are approximately 64 mm long as preserved, and
specimen was removed from the block (Fig. 51D).
are complete distally, but are covered by the lower jaw proximally.
Postcrania – The cervical and dorsal vertebral series are present,
but are interrupted and offset by a few cm by a small dip-slip fault. The NMMNH P-42579
fault exhibits slickensides and the formation of gypsiferous slickenfibers. This specimen is a large skull exposed in ventral view (Fig. 52A).
Two cervical and four dorsal vertebrae have been dragged into the fault The lower jaws are tightly closed, and the maxillary teeth project out of
and are covered. Only the first two proximal caudal vertebrae are in the prepared block surface. No hyoids are exposed. Only the area com-
place; the remainder of the tail is continued off the southwest corner of prising the premaxillae to the anterior portion of the antorbital fenestra
the block. on the right side is visible. The remainder of the skull is covered by
The shoulder girdle is intact, and except for slight mediolateral NMMNH P-42353. See also NMMNH P-44553, which is an incom-
compression is well preserved. A small portion of the right coracoid has plete cervical series that probably belongs to the same individual as
been lost. The right ramus of an apparently complete furcula lies across NMMNH P-42579.
the anterior margin of the right coracoid, but the left ramus is mostly
covered by the left scapula and coracoid. The furcula will be described NMMNH P-42580
and discussed in a later section. The forelimbs are essentially complete NMMNH P-42580 represents a very small complete pelvis,
46

FIGURE 47. A, An overview of Coelophysis bauri, NMMNH P-42353, with its skull-over-hip opisthotonic posture. This specimen preserves an articulated
furcula as well. B, Close-up view of the skull of NMMNH P-42353, Coelophysis bauri, showing its nearly complete hyoids, and two articulated sclerotic
ossicles in the posterodorsal quarter of the orbit.
 47

FIGURE 48. NMMNH P-42354 is the incomplete presacral series of a very small juvenile Coelophysis bauri. The right arm and manus parallel the anterior
dorsals.
sacrum, and incomplete hindlimbs in right lateral view (Fig. 52B). The fragments, including an incomplete premaxilla with one tooth, and a
limb bones are broken and separated by a very large joint in the block, maxilla or jaw fragment with two teeth. These fragments are assignable to
but are nonetheless complete and measurable. The 126 mm-long femur a small juvenile Coelophysis.
and 80 mm-long sacrum place it in the smallest juvenile size class.
NMMNH P-44552
NMMNH P-44551 NMMNH P-44552 is a nearly complete postcranial skeleton com-
P-44551 comprises a skull in right lateral and dorsal views and an prising an incomplete dorsal vertebral series, incomplete pelvis and sacrum,
incomplete cervical series of five vertebrae (Fig. 53A). The mandibles are part of both hindlimbs, an incomplete pes, and a complete caudal series
not attached to the skull. The exposed cervicals represent three through (Fig. 54). The anterior dorsal series extends off the block, but based on
seven. The first three are covered, and the posterior portion of the series position, proximity, and size this specimen is almost certainly the
is off the block edge. The neck is kinked near the base of the skull so that postcrania of the skull and partial neck, NMMNH P-44551. The speci-
the cervical vertebrae and their ribs are seen in left lateral view. men is exposed in right lateral view and is bisected by a joint in the block
The maxilla, antorbital fenestra, and part of the skull roof are well that disrupts the pelvis and hindlimbs. The complete tail extends in a
exposed. The premaxillae are off the block edge. A nearby pair of man- dorsally curved loop.
dibles, P-50529, may belong to this skull based on their size and proxim- A large body of probable coprolite was present in and under
ity to NMMNH P-44551 (Fig. 53B). A small forearm and manus, prob- (recalling that the block is now upside down) the vent area between the
ably belonging to this animal, lie across the skull. The forearm lies over ischia and the proximal caudals (Fig. 54). Most of the coprolite was
the anterior portion of the snout, and the hand covers the orbit. The removed from the block for micropreparation. Bone fragments are densely
proximal humerus and shoulder girdle are lost to a joint in the block. packed in the coprolite, and although mostly unidentifiable, they include
A large probable regurgitalite fills part of the oral cavity as well as hand and wrist bones assignable to a small juvenile Coelophysis (Rinehart
covering the posterior portion of the skull. Recalling that the block has et al., 2005a, b). This is particularly significant in that the apparently
been turned over, this material would have originally been under the associated skull, NMMNH P-44551, has regurgitalite with tooth-bear-
skull. The apparent regurgitated material, which is more completely ing premaxilla and maxilla fragments of a small Coelophysis in its mouth
described in the section on Diet and Cannibalism, contains many bone area. It appears that this specimen’s last meal was part of a small juvenile
Coelophysis, which further supports cannibalism in the species.
48

FIGURE 49. NMMNH P-42576, a small juvenile Coelophysis bauri specimen is incompletely covered by NMMNH P-42577. A, Overview with the
gastralia and arms in the upper center of the photo; a furcula lies atop the shoulder girdle (upper right quarter of photo). B, The extraordinarily well-
preserved left wrist of this specimen includes all five carpals.
 49

FIGURE 50. NMMNH P-42577, a small juvenile Coelophysis bauri overlying NMMNH P-42576. A, Overview with the gastralia and forelimbs just above
the scale. B, Skull in dorsal view. C, Left wrist and incomplete manus. D, Complete, articulated gastral basket.
50

FIGURE 51. NMMNH P-42578, a very small incomplete juvenile Coelophysis bauri. A, Complete forelimb including manus at left. B, Complete, but
disarticulated gastral basket. C, Close up of complete right manus. D, Right pes in plantar view.
 51

FIGURE 52. A, The skull and lower jaws of a large Coelophysis bauri, NMMNH P-42579, in ventral view. B, Coelophysis bauri pelvis, sacrum, and
incomplete hind limbs of a very small individual, NMMNH P-42580, in right lateral view.
52

FIGURE 53. The skull and incomplete cervical series of NMMNH P-44551, Coelophysis bauri. A, Lateral view of the postcranial skeleton, NMMNH P-
44552. A forearm lies over the snout, parallel to the toothrow, and the manus covers the right orbit. The “mouth” area is filled with regurgitalite containing
teeth and bone fragments of a small juvenile Coelophysis. B, A pair of detached lower jaws, NMMNH P-50529, in the lower right corner of A, may belong
to this specimen.
 53

FIGURE 54. Coelophysis bauri, NMMNH P-44552, Overview of nearly complete individual, with its anterior dorsal series extending off the block, its
complete caudal series arranged in a loop and a regurgitalite and coprolite preserved in situ. Inset, Vent area between the ischia and proximal caudals where
a large body of coprolite, mostly removed for micropreparation, was located. Abbreviations: cd1, caudal vertebra 1; is, ischium; m, maxilla; man,
mandible; pu, pubis.
54
NMMNH P-50529 (Fig. 57B). Ninety-three mm of jaw anterior to the
NMMNH P-44553 mandibular fenestra are preserved; the posterior portion of the specimen
NMMNH P-44553 consists of three large, articulated mid-cervi- extends off the northwest edge of the block. The right mandible is ex-
cal vertebrae (Fig. 55A) with the anterior portion of the series covered by posed in lateral view, and the left is in medial view. Twenty six tooth
NMMNH P-42353, whereas the posterior portion extends off the west positions are preserved, only one is represented by an empty alveolus. A
edge of the block. It is very likely that the cervicals belong to skull complete tooth replacement cycle, beginning with an empty alveolus in
NMMNH P-42579 based on proximity, alignment with the skull, and position seven and ending with a full-sized replacement tooth in position
size. This specimen is exposed in left lateral and ventral view. At one 23, is present and well exposed in the dentary. The jaws may belong to
point along its ventral margin 10 parallel cervical ribs are evident, indicat- skull NMMNH P-44551 based on their size and proximity to the cra-
ing that the rib length in this area is equal to at least five times that of the nium, although they are rotated 180º and moved well out of articular
cervical vertebrae. position.

NMMNH P-44554 NMMNH P-50530

This specimen consists of an incomplete shoulder girdle, cervical This specimen comprises the snout tip of a very large skull ex-
vertebrae approximately eight through twelve, and a single dorsal verte- posed in left lateral view, (Fig. 57B). Some oblique shearing has occurred
bra of a large individual. A fragment of a large furcula lies on the left so that part of the lingual side of the right premaxilla and maxilla are
coracoid. The specimen is badly disrupted by a north-south trending visible. The premaxillae are slightly detached, displaced anteriorly one or
joint in the block. NMMNH P-44555, an incomplete skull described two mm, and rotated slightly anteroventrally with respect to the rest of
below, is probably part of this same individual. the snout. The premaxilla measures 15.4 mm from its anterior tip to the
anterior margin of the external nares. The entire skull posterior to the
NMMNH P-44555 anterior antorbital fenestra is extended off the northwest edge of the
block. There are no apparent associated postcrania.
This 195 mm-long skull lies near the center of the southern edge of
the block (Fig. 55B). The skull is obliquely flattened so as to expose the
NMMNH P-50531
palate and right side. It is generally poorly preserved and is missing most
of its roof and the distal portions of the quadrates. The incomplete NMMNH P-50531 is an incomplete distal caudal vertebral series
shoulder girdle and cervical series, NMMNH P-44554, almost certainly of approximately 19 vertebrae exposed in right lateral view, with a total
belong to this skull based on position, proximity, and size. length of 415 mm. The series lies approximately west to east across the
block with the proximal end of the series extending off the east edge of
NMMNH P-44556 the block and partially covered by NMMNH P-44552.
This specimen is a large proximal caudal vertebral series. Approxi-
mately the first 16 caudals and their chevrons are present. At least three NMMNH P-50532
vertebrae are covered by NMMNH P-50536 and P-42351. The total NMMNH P-50532 consists of an articulated incomplete distal
length of the series, as preserved, is 620 mm. caudal series of approximately 15 vertebrae (Fig. 58A), having a total
preserved length of 490 mm.
NMMNH P-44802
NMMNH P-50533
This specimen, possibly complete except for the greater portion
of the distal tail, is located in C-8-82 J-3, one of the 11 small jacketed This specimen is an incomplete, totally disarticulated skull of a
blocks removed from the main block during preparation. The caudal very small individual (Fig. 58B). There are several tiny teeth in the
vertebral series ends after number five at the edge of the block. In its exposed posterior portion of the maxilla. Almost none of the bones are
current state of preparation only the pelvis, sacrum, proximal caudal complete or well enough exposed to measure. The premaxilla measures
vertebrae, hindlimbs, and pedes are exposed (Fig. 56A). The hindlimbs 9.5 mm from its anterior tip to the anterior margin of the external nares.
are tightly folded up under the animal, and it is exposed in ventral aspect. The skull may belong to the postcranial skeleton NMMNH P-42576,
The femur length, 119.8 mm, places it in the smallest size class (prob- based on their position relative to one another and being of the appropri-
ably one year). The posterior portions of the sacrum, ilia, and proximal ate size.
ischia, are perfectly preserved and exposed three dimensionally, which is
rarely encountered. The “vent area” of this specimen is further described NMMNH P-50534
in the section on Growth, Allometry, and Population Dynamics that This specimen represents the incomplete skull of a very small
relates to reproduction. individual (Fig. 59A). The skull table up to and including the prefrontals
and the posterior portion of the nasals are preserved and exposed in
NMMNH P-46615 dorsal view. As preserved, this incomplete skull is 55 mm long.
NMMNH P-46615 is a complete, disarticulated and unassociated
furcula (Fig. 56B). It is described and illustrated in the section on Furcu- NMMNH P-50535
lae. NMMNH P-50535 comprises an incomplete pelvis, hindlimbs
and pedes of a small individual (Fig. 59B). Its 149-mm-long femur places
NMMNH P-50528 it in the uppermost range of the small juvenile size class.
This specimen consists of seven articulated proximal caudal ver-
tebrae with some chevrons from a very small individual (Fig. 57A). They NMMNH P-50536
do not appear to be associated with any other specimens on the block. NMMNH P-50536 comprises a large left tibia, astragalo-
The vertebrae are exposed in ventral view and, as preserved, have a total calcaneum, and pes (Fig. 60A). An associated femur is buried under the
length of 185 mm. tibia.

NMMNH P-50529 NMMNH P-50537

The incomplete, articulated jaws of a small individual comprise NMMNH P-50537 represents a complete left hindlimb and pes,
 55

FIGURE 55. A, Three articulated Coelophysis bauri cervical vertebrae, NMMNH P-44553, probably belong to the skull of NMMNH P-42579. B,
Coelophysis bauri skull in right lateral and ventral aspect, NMMNH P-44555 shows the posterior palate and probably is part of the same individual as the
cervical and shoulder material, NMMNH P-44554.
56

FIGURE 56. A, NMMNH P-44802, the pelvis, sacrum and hind limbs of a small Coelophysis bauri showing an undistorted vent area and the obturator
foramen and “ceratosaur” foramen in the pubes. B, NMMNH P-46615, a disarticulated Coelophysis bauri furcula.
 57

FIGURE 57. A, NMMNH P-50528, a 185-mm-long series of Coelophysis bauri proximal caudal vertebrae from a small individual. B, NMMNH P-50529,
a pair of lower jaws of Coelophysis bauri, that probably pertain to the skull of NMMNH P-44551 and show a complete tooth replacement cycle. NMMNH
P-50530 is the snout portion of a very large C. bauri skull.
58

FIGURE 58. A, NMMNH P-50532, a distal Coelophysis bauri caudal series, ~ 0.5 m long. B, A very small disarticulated skull of Coelophysis bauri,
NMMNH P-50533, may pertain to the postcranial skeleton, NMMNH P-42576.
 59

FIGURE 59. A, NMMNH P-50534 represents the skull table of a small Coelophysis bauri in dorsal view. B, NMMNH P-50535, the incomplete pelvis and
hindlimbs of a small Coelophysis bauri specimen.
60

FIGURE 60. A, Coelophysis bauri, NMMNH P-50536, a distal hindlimb and pes. B, A very small incomplete vertebral column, pelvis, and hindlimb of
Coelophysis bauri comprise NMMNH P-50537.
 61
incomplete pelvis, and part of both the posterior dorsal and proximal NMMNH P-55344
caudal series. All the elements except the hindlimb are very poorly pre- This specimen is a large left femur that is covered distally by
served (Fig. 60B). The femur measures 123 mm in length, indicating that matrix (Fig. 62B).
it belongs to the smallest juvenile size class.
NMMNH P-55345
NMMNH P-55336
This specimen consists of the poorly exposed hindlimbs of a
This specimen is a very small, well preserved, isolated pelvis in medium-sized individual (Fig. 62C). The limbs enter the block surface at
right lateral view (Fig. 61A). an approximate 30º angle and extend almost completely under NMMNH
P-50536.
NMMNH P-55337
This specimen is the incomplete hindlimbs and pedes of a small NMMNH P-55346
individual (Fig. 61B). This specimen consists of three large, articulated distal caudal
vertebrae (Fig. 63A).
NMMNH P-55338
This specimen comprises a poorly preserved, incomplete poste- NMMNH P-55348
rior cervical vertebral series of seven through 10 together with associated NMMNH P-55348 is a small, completely disarticulated incom-
ribs. The vertebral series is bent sharply backward, apparently in the plete skull and a few associated, disarticulated cervical vertebrae and
death pose often seen in the Ghost Ranch Coelophysis. cervical ribs. Based on its size and position it could belong to NMMNH
P-42352.
NMMNH P-55339
This is a completely disarticulated, incomplete pes associated NMMNH P-55349
with an incomplete anterior caudal vertebral series of approximately 10 An incomplete distal caudal series including approximately 12
vertebrae. kinked and poorly preserved vertebrae comprise NMMNH P-55349.

NMMNH P-55340 NMMNH P-55350

This specimen consists of an incomplete tibia and fibula with This specimen represents three articulated distal caudal vertebrae,
their distal ends lost off the east edge of the block (Fig. 61C). two of of which are complete, and the third is represented by an impres-
sion in the matrix (Fig. 63B).
NMMNH P-55341
This specimen is the knee region of a hindlimb. Only the distal NMMNH P-57652
femur and proximal tibia-fibula are present (Fig. 62A). The balance of the This disarticulated, incomplete skull is mostly covered by
specimen is missing off the west edge of the block. NMMNH P-50536. Only small portions of the skull bones and some
teeth are exposed.
NMMNH P-55342
An incomplete shoulder girdle associated with six disarticulated NMMNH P-57653
cervical vertebrae and cervical ribs that represent a very small individual. This slightly disarticulated, incomplete, midsized skull and lower
jaws and several disarticulated cervical vertebrae and their ribs are pre-
NMMNH P-55343 served in jacket C-8-82-J1 together with a juvenile phytosaur skull,
NMMNH P-55343 comprises two articulated cervical vertebrae. NMMNH P-44920, and Hesperosuchus osteoderms, NMMNH P-
The remainder of the cervical series is either covered, lost, or on an 57654. The skull is exposed in ventral view, with the snout tip off the
adjoining block. west edge of the jacket. An associated regurgitalite covers the posterior
portion of the skull. The ichnofossil measures 55 x 39 mm and contains
numerous small bone fragments, the largest of which measures 5 x 6 mm.
62

FIGURE 61. A, NMMNH P-55336, a right pelvis of Coelophysis bauri in lateral view. B, Coelophysis bauri, NMMNH P-55337, small hindlimbs and both
pedes. C, NMMNH P-55340, a small Coelophysis bauri incomplete tibia and fibula.
 63

FIGURE 62. A, NMMNH P-55341, Coelophysis bauri, a distal femur and proximal tibia. B, NMMNH P-55344, Coelophysis bauri, a large proximal femur.
C, NMMNH P-55345, Coelophysis bauri, incomplete hindlimbs.
64

FIGURE 63. A, NMMNH P-55346, Coelophysis bauri, incomplete caudal series. B, NMMNH P-55350, Coelophysis bauri, incomplete caudal series.
 65
POSTCRANIAL OSTEOLOGY OF COELOPHYSIS BAURI anterior dorsal vertebrae have shorter centra than more anterior cervical
centra (e.g., cervical 10 is shorter than cervical 9, which is shorter than
None of the Coelophysis skeletons in the NMMNH Whitaker cervical 8). Given that the cervical series is articulated, it is difficult to
quarry block (C-8-82) are complete. This necessitates a composite oste- assess the morphology of the anterior/posterior articular surfaces of the
ology, using various specimens. Significant redundancy of elements al- centra. The posterior two centra, preceeding the cervical/dorsal transi-
lows a thorough description of each element of the postcranium and tion are demonstrably amphicoelous (e.g., NMMNH P-42200), indicat-
description of intraspecific variation within various functional complexes. ing that all the cervical centra are likewise amphicoelous. The centra are
Descriptions of each element include NMMNH specimen numbers as elongate, approximately three to four times as long (anteroposteriorly)
references for the element or morphology being described, whereas addi- as tall (dorsoventrally) in the anterior cervicals (Figs. 65C, 66). The
tional specimens on the NMMNH block may possess the same element elongation of the centra is reduced in more posterior cervicals. An ellip-
or morphology, the described specimen is the best example. tical pleurocoel with an anteroposteriorly oriented long axis is present on
the lateral surface of the posterior half of the centrum; the pleurocoel is
Skull less pronounced in cervicals 9 and 10, and is absent in the dorsals. The
The skull of Coelophysis bauri is currently being studied by R. body of the centrum is arched and slightly waisted, giving the vertebrae
Reiz and others (pers. commun. to LFR) and therefore its description is an hourglass-shape in ventral view. A single small, but distinct, ventral
not undertaken here. However, we have illustrated and labeled the best keel is present. The prezygapophyses are triangular and extend beyond
example of the skull from the NMMNH Whitaker Quary block (Fig. 64). the anterior margin of the centrum with the articular facet oriented
dorsomedially. The prezygapophyses overlap the posterior neural arch
Vertebral Series of the preceding cervical. The neural spine is short, consisting of just a
small ridge on the neural arch. The postzygapophyses are also triangular,
The description of the vertebral series of Coelophysis is based
though more elongate and narrower than the prezygapophyses.
solely on complete or nearly complete series; isolated vertebrae were not
The cervical ribs are articulated to the vertebrae via a subcircular
used.
parapophyseal facet. The cervical rib bifurcates into anterior and poste-
Cervical Vertebrae and Ribs rior processes. The anterior process of the cervical rib extends anteriorly
The two cervical series used for the description are those of: to the level of the anterior tip of the prezygapophyses. The posterior
NMMNH P-42200, a nearly complete individual exposed in lateral view cervical ribs are very elongated, extending a maximum posterior distance
(Fig. 65A), and NMMNH P-42353, an anterior half of an individual with equivalent to the length of four or five vertebrae, resulting in as many as
a nearly complete cervical series exposed in ventral view that is over- five ribs overlapping each other at a particular level.
lapped by a caudal series of another individual (Fig. 65B)
There are 10 vertebrae in the cervical series in Coelophysis bauri, Dorsal Vertebrae
including the atlas and axis, only eight of which are visible in NMMNH Two dorsal vertebral series are used for this description: NMMNH
P-42200. The transition from cervical to dorsal vertebrae is identified by P-42351 exposed in ventral view (Fig. 67B) and NMMNH P-42576
centrum length, following Colbert (1989). The last two cervicals and exposed in lateral view.

FIGURE 64. NMMNH P-42200, outline drawing of juvenile skull in left lateral view, with elements labeled. Abbreviations: a, articular; d, dentary; emf,
external mandibular fenestra; f, frontal; j, jugal; l, lacrimal; m, maxilla; n, nasals; o, orbit; p, parietal; pa, prearticular; pf, prefrontal; pm, premaxilla; po,
postorbital; pp?, postparietal?; qj, quadratojugal; sa, surangular; sq?, squamosal?.
66

FIGURE 65. Coelophysis bauri, examples of cervical vertebrae. A, NMMNH P-42200, complete cervical series of a juvenile individual in left lateral view.
B, NMMNH P-42353, partial cervical series of an incomplete individual in right lateral view. C, NMMNH P-42577, incomplete cervical series in left
lateral view. Cervical and dorsal vertebrae are numbered; arrow highlights overlapping cervical ribs. Scale bars are 5 cm.
 67

FIGURE 66. Coelophysis bauri, example of cervical vertebrae. Line drawing of NMMNH P-42577, incomplete cervical series in left lateral view (same
specimen as Fig. 62C). Abbreviations: c8-c10, numbered cervical vertebrae; cr, overlapping cervical ribs; d1-d2, numbered dorsal vertebrae; sc,
scapulacoracoid.
68

FIGURE 67. Coelophysis bauri, examples of dorsal vertebrae. A, NMMNH P-42200, complete dorsal series of a juvenile individual in ventrolateral view.
B, NMMNH P-42351, complete dorsal series of an incomplete adult individual. Scale bar is 5 cm.
 69
The dorsal series includes 13 vertebrae. The dorsal centra are not articulate along the ventral midline but rather they form an imbri-
weakly amphicoelous and slightly waisted. The pre- and cated, “zig-zag” pattern (Fig. 71C). This pattern is observed in
postzygapohpyses are very small compared to the cervicals. The articu- archosauriforms (Claessens, 2004, fig. 9B) and, indeed, in all theropod
lar facets of the prezygapophyses face dorsally, whereas the facets of dinosaurs.
the postzygapophyses are oriented ventrolaterally. The capitular facets
are subcircular in cross-section. Thin laminae extend from the anterior, Pectoral Girdle
posterior, anteroventral and posteroventral margins of the tubercular The pectoral girdle is described on the basis of three specimens:
facet. The tubercular facets are sub-triangular and flare laterally. The NMMNH P-42351, P-42353 and NMMNH P-42577 (left lateral view)
transverse process is more posteriorly positioned in the more posterior (Fig. 72B). NMMNH P-42351 is an incomplete skeleton, all elements
dorsal vertebrae (Fig. 67A). The neural spine is rectangular, with a con- anterior to the pectoral girdle, the entire caudal series, the right pectoral
cave posterior margin, and extends along the entire length of the neural girdle and forelimb and the distal right hindlimb (tibia, fibula and pes) are
arch. missing or covered (Fig. 72A). NMMNH P-42353 is an incomplete
skeleton with a fragmentary skull, a complete cervical vertebral series, a
Sacral Vertebrae partial mid-dorsal vertebral series and an articulated pectoral girdle (Figs.
Three sacral series were used for the description: NMMNH P- 72C-D, 73A-B).
42351, sacral series in ventral view (Fig. 68B); NMMNH P-42200, a
Scapula
sacral series in ventral view of a nearly complete individual (Figs. 68A,
69) and NMMNH P-42576, a complete sacral series in dorsal view. The scapula is expanded ventrally, with a constricted scapular
Five vertebrae are present in the sacral series. The sacral centra blade that expands dorsally and terminates with a semicircular tip. The
are proportionally shorter than the cervical or dorsal centra and are suture with the coracoid extends anteriorly from the elliptical (long axis
weakly amphicoelous. The sacral ribs are fused to the ilium. The sacral oriented dorsoventrally) glenoid. The suture is not always visible in all
neural spines are similar to those of the dorsals. In one specimen, NMMNH specimens. In anteroposterior view the dorsal tip of the scapular blade is
P-42576, all the sacral neural spines are fused into a single blade of bone. angled medially, whereas the glenoid is angled posteroventrally. The
This feature is a sexually dimorphic character in C. bauri as is the fusion acromion process is on the anterior margin of the scapula and is
of sacral 4 and 5. The state of fusion of sacral 4 and 5 and of the sacral mediolaterally thin, anteroposteriorly broad and semicircular.
neural spines partially defines the sexual dimorphs which are discussed
below. Coracoid
The coracoid is subcircular and has a foramen anterior to the
Caudal Vertebrae and Chevrons glenoid. The coracoid has a posterior glenoid process. Just anterior to the
The three caudal series used for the description are: NMMNH P- subtriangular postglenoid process is a pyramidal process that extends
44552, a complete series in left lateral view (Fig. 70B), NMMNH P- laterally.
42352, an incomplete skeleton with a nearly complete series (Fig. 70A)
and NMMNH P-42200, a nearly complete skeleton with a very well Furcula
preserved anterior portion of the series. The furcula has an open V-shape outline and articulates with the
There are approximately 41 vertebrae in the caudal series. The acromion process of the scapula. See the Furcula section below for addi-
centra are weakly amphicoelous. Prezygapophyses in the anterior caudals tional description and reconstruction of the pectoral girdle
extend slightly beyond the anterior margin of the centra and are triangu-
lar, with a mediodorsally facing articular facet. Posterior caudals have Forelimb
expanded prezygapophyses that overlap the posterior quarter of the The forelimb of Coelophysis bauri is currently being studied by S.
preceding vertebrae (contra Gay, 2001). The anterior 17 caudals have Nesbitt (pers. commun. to LFR). We illustrate the best examples of the
transverse processes, which are subrectangular with expanded, curved forelimbs in the NMMNH Whitaker Quary block (Figs. 74-75).
lateral tips. These transverse processes incorporate two buttressed sup-
ports. The neural spines are rectangular and offset posteriorly; only the Pelvic Girdle
17 anterior caudals have distinct neural spines. The postzygapophyses
are very small and triangular. The description of the pelvic girdle is based on three specimens:
NMMNH P-42200 (ventrolateral view), P-42351 (ventral view) (Figs.
Chevrons occur in the proximal caudals until the junction of the
27th and 28th. Anterior chevrons are curved, whereas more posterior 76B, 77) and P-55336 (Fig. 76A). NMMNH P-42200 and P-42351 are
discussed above. NMMNH P-55336 is a nearly complete, isolated, right
chevrons are more linear with angled ventral tips. The articular facets of
the chevrons are subtriangular. Chevrons begin on caudal 1 in all morphs, pelvic girdle, with a caudal series overlapping the iliac blade. NMMNH
P-44802 provided the best basis for a complete pelvic girdle reconstruc-
so this is not a dimorphic character. The anteriormost two chevrons
(between caudals 1 and 2 and 2 and 3) are smaller than subsequent tion (Fig. 78).
chevrons.
Ilium
Gastralia The iliac blade is subrectangular, with a long posteroventral arch,
Description of the gastralia of Coelophysis bauri is based on which is a U-shaped groove on the ventral surface of the iliac blade that
extends from the acetabulum to the posterior end. Dorsal to the midlength
NMMNH P-42577 (Fig. 71A). The gastralia of C. bauri are composed
level of the obturator foramen there is a subtriangular projection. The
of four units: two lateral gastralia and two medial gastralia. A medial
acetabulum is circular. Ischial peduncles are positioned ventrally and
gastralium has a triangular, medial end with a posteriorly-directed pro-
cess. A lateral gastralium is a long, curved element that tapers at its ends. form the most restricted portion of the vent opening.
There is a large overlap between the medial and lateral gastralium; ap-
proximately half the length of both elements overlap (Fig. 71B). The Pubis
diameter of an individual gastralium is approximately half, or less, that of The pubis extends anteroventrally with a slight posterior projec-
the associated ribs. tion at its ventral tip. The pubis is rod-like, with a constant thickness
Minimally there are eight gastralia in Coelophysis bauri; they do throughout its length. A subcircular, ventral process on the pubis con-
70

FIGURE 68. Coelophysis bauri, examples of sacral vertebrae. A, NMMNH P-42200, sacral series, partially obscured by pelvic girdle, demonstrating fused
s4-s5. B, NMMNH P-42351, sacral series, partially obscured by pelvic girdle, demonstrating unfused s4-s5. Scale bar is 5 cm.
 71

FIGURE 69. Coelophysis bauri, example of sacral vertebrae. Line drawing of NMMNH P-42200, sacral series, partially obscured by pelvic girdle,
demonstrating fused s4-s5 (same specimen as Fig. 68A). Abbreviations: ca1, caudal vertebra one; il, ilium; s3-s5, numbered sacral vertebrae.
72

FIGURE 70. Coelophysis bauri, examples of caudal vertebrae and chevrons. A, NMMNH P-42352, nearly complete caudal series in left lateral view. B,
NMMNH P-44552, complete caudal series looped over itself in left lateral view. Scale bar is 5 cm. Caudal vertebrae are numbered.
 73

FIGURE 71. Coelophysis bauri, examples of gastralia. A, NMMNH P-42577, gastral basket in ventral view. Scale in cm. B-C, line drawings of B, isolated
gastralium and C, imbricated gastralia, based on NMMNH P-42577. Abbreviations: lg, lateral gastralium; mg, medial gastralium.
74

FIGURE 72. Coelophysis bauri, examples of pectoral girdles. A, NMMNH P-42351, large isolated scapulacoracoid in left lateral view. B, NMMNH P-
42577, articulated scapulacoracoids in left lateral view. C-D, NMMNH P-42353, articulated pectoral girdle with fucula in C, left lateral and D, ventral
views. Scales in cm.
 75

FIGURE 73. Coelophysis bauri, example of pectoral girdles. A-B, Line drawings of NMMNH P-42353, articulated pectoral girdle with furcula in A, left
lateral and B, ventral views (same specimen as Fig. 72C-D). Abbreviations: cora, coracoid; fur, furcula; scap, scapula.
76

FIGURE 74. Coelophysis bauri, examples of forelimbs. A, NMMNH P-42353, articulated forelimbs with incomplete mani in anterior view and NMMNH
P-42358, isolated forelimb. B, NMMNH P-42577, articulated forelimbs in anterior view. Abbreviations: cora, coracoid; h, humerus; lc, left carpals; lh,
left humerus; lr, left radius; lu, left ulna; mc, metacarpals; ph, articulated phalanges; r, radius; rh, right humerus; rr, right radius; ru, right ulna; sc,
scapulacoracoid; u, ulna.
 77

FIGURE 75. Coelophysis bauri, examples of forelimbs. A, NMMNH P-42353, articulated forelimbs with incomplete mani in anterior view and NMMNH
P-42358, isolated forelimb (same specimens as Fig. 74A). B, NMMNH P-42577, articulated forelimbs in anterior view (same specimen as Fig. 74B).
Abbreviations: cora, coracoid; fur, furcula; h, humerus; lc, left carpals; lh, left humerus; lr, left radius; lu, left ulna; mc, metacarpals; ph, articulated
phalanges; r, radius; rh, right humerus; rr, right radius; ru, right ulna; sc, scapulacoracoid; u, ulna.
78

FIGURE 76. Coelophysis bauri, examples of pelvic girdles. A, NMMNH P-55336, isolated right pelvic girdle in right lateral view. B, NMMNH P-42351,
articulated pelvic girdle in ventral view, anterior is to the left. Abbreviations: il, ilium; is, ischium; lf, left femur; p, pubis.
 79

FIGURE 77 Coelophysis bauri, example of pelvic girdle. Line drawing of NMMNH P-42351, articulated pelvic girdle in ventral view, anterior is to the left
(same specimen as Fig. 76B). Abbreviations: lf, left femur; lp, left pubis; lil, left ilium; lis, left ischium; ril, right ilium; ris, right ischium; rp, right pubis;
s1, s4, s5, numbered sacral vertebrae.
80

FIGURE 78. Coelophysis bauri, NMMNH P-44802, pelvic girdle restored in A, left lateral and B, in ischial view as preserved.

tains the elliptical obturator foramen. An accessory foramen (the

Femur
“ceratosaur foramen”) is medial to and larger than the obturator foramen.
This accessory foramen was not described or discussed by Colbert (1989). The femur possesses a hooked head, with a prominent dorsal/
A paper-thin sheet of bone connects the distal pubes along their length. ventral-oriented groove running down its posterior surface (not illus-
trated in Colbert, 1989, fig. 80). The anterior trochanter is triangular,
Ischium with a lateral groove that wraps around the lateral margin of the femoral
shaft. The fourth trochanter is thin and blade-like and is thinner and less
The ischium projects posteriorly, thins along its length and has a extensive than that illustrated by Colbert (1989). Distally, the tibial
small posterior projection on its ventral tip. The ischia are elliptical in
(medial) condyle is subrectangular with the long axis oriented
cross section and fused along their length distally. A distinct, thin anteroposteriorly and larger than the rounded fibular (lateral) condyle.
puboischiadic plate extends from the proximal, posterior pubes to the
The fibular condyle process is triangular and projects from the fibular
proximal, anterior ischium and underlies the acetabulum. condyle at an acute angle. A groove is present in some specimens be-
Hindlimb tween the two distal condyles, though this feature is more prominent in
larger individuals (compare NMMNH P-42351, a large individual, to
Discription of the hindlimbs is based on four specimens: NMMNH NMMNH P-42200, a smaller, juvenile individual).
P-42200 (Figs. 79A, D, 80A-B, D-E, 81); P-42351 (Fig. 79C); P-42352
(Fig. 80C); P-55344 (Fig. 79B). NMMNH P-42200, P-42351 and P- Tibia
42352 are discussed above; NMMNH P-55344 is a proximal femur The tibia has a subtriangular proximal end with a hooked anterior
fragment exposed in anterior view.
margin. The proximal end of the tibia is approximately twice the width of
 81

FIGURE 79. Coelophysis bauri, examples of hindlimb and femoral morphology. A, D, NMMNH P-42200, A, nearly complete left hindlimb in ventrolateral
view and D, articulated knee joint in articular view. B, NMMNH P-55344, a proximal left femur fragment in anterior view. C, NMMNH P-42351, right
femur in lateral view. Abbreviation: at, anterior trochanter. Scales in cm.
82

FIGURE 80. Coelophysis bauri, examples of pedes and hindlimb variation. A-B, D-E, NMMNH P-42200, A, nearly complete left pes in medial view, B,
fused tibia and fibula in anterior view, D, fused metatarsals two and three and E, closeup of ankle. C, NMMNH P-42352, unfused tibia and fibula in anterior
view. Abbreviations: f, fibula; mt, metatarsal; t, tibia. Scales in cm.
 83

FIGURE 81. Coelophysis bauri, examples of pedes and hindlimbs. A-C, NMMNH P-42200, line drawings of A, nearly complete left pes in medial view
(same specimen as Fig. 80A), B, nearly complete left hindlimb in ventrolateral view (same specimen as Fig. 79A), C, articulated knee joint in articular view
(same specimen as Fig. 79D). Abbreviations: lf, left femur; lfi, left fibula; lt, left tibia; mt, mt2, mt3, metatarsals; ph, phalanges.
84
the tibial shaft. The tibia and fibula are demonstrably longer than the
Pes
femur. In some specimens the proximal portion of the tibia is fused to the
fibula (NMMNH P-42200), though this is variable (NMMNH P-42352). Overall, the pes is very similar to the illustrations of Colbert
The distal end of the tibia is not visible due to its fusion with the astraga- (1989, figs. 84-86), although a few key differences can be noted. There is
lus/calcaneum. only one tarsal, though in NMMNH P-42200 it is more polygonal than,
and not as flat as that illustrated by Colbert (1989, fig. 84A). There is
Fibula fusion, in at least some specimens, between the 2nd and 3rd metatarsal.
The 3rd metatarsal is the thickest, whereas the 2nd and 4th are approxi-
The head of the fibula is semicircular in shape and the shaft is thin mately the same thickness. The unguals are longer and more strongly
and tapers distally. As in the tibia, the distal fibula is not visible due to
curved than those illustrated by Colbert (1989, fig. 85).
fusion with the astragalus/calcaneum.
 85
VISION CHARACTERISTICS OF COELOPHYSIS BAURI (reptiles + birds) and bony fish, although the evolutionary origin of the
ossicles is not considered to be the same in the two groups (Walls, 1942).
Methodology Walls (1942) cites embryological evidence that the sclerotic ossicles of
the Sauropsida are homologous with the dermal circumorbital bones of
In this section we analyze the vision characteristics of Coelophysis
fishes, but because the sclerotic ossicles coexist with circumorbital bones
bauri on the basis of NMMNH P-42200, a specimen that preserves the
in fishes, they apparently have a different origin in that clade. The scle-
first documented complete sclerotic ring of the species (Fig. 82). Skull,
rotic rings of Silurian and Devonian armored fish (e.g., Acanthodii and
orbit, and sclerotic ring morphology allow us to estimate the relative
Placodermi) have four ossicles (Long, 1995). In modern fishes the scle-
importance of vision, eye morphology, night vision capability and bin-
rotic ring generally comprises only two ossicles, which in fast swimming
ocular overlap of the visual fields. These factors were compared to extant
species (e.g., tuna and swordfish) are co-ossified to form a stiff support
lizard and bird outgroups that phylogenetically bracket dinosaurs and to
against the impact of the water.
Archaeopteryx.
One purpose of the sauropsid sclerotic ring is to maintain the
Background shape of a nonspheroid eye against its internal pressure (Walls, 1942). In
general, these eyes may be divided into two segments: a corneal segment
Little fossil evidence is available concerning the attributes of dino- through which light enters the eye is in front of the lens, and an orbital
saur eyes, and their vision characteristics are infrequently studied. Of- segment, which is behind the lens and is contained within the skull (Fig.
ten, the skeletal elements of fossil specimens, including orbit size and 83). Typically, the radius of curvature of the corneal segment is smaller
shape, are described in great detail, but with only a mention that a scle- than that of the orbital segment. This arrangement creates a circular
rotic ring is “present.” Descriptions of fossil sclerotic rings exist, for depression, the corneal-orbital sulcus, around the periphery of the cor-
example, Galton’s (1974) thorough description of the sclerotic ring of nea where it joins the orbital segment. The sclerotic ring reinforces the
Hypsilophodon foxii, but such descriptions are rare in the literature. sulcus and prevents it from being pushed out by the internal pressure of
Sclerotic rings are relatively fragile structures and frequently fail to fos- the eye.
silize. Undoubtedly, many are lost to the processes of taphonomy and The sauropsid sclerotic ring also functions in accommodation
preparation. (Welty, 1963), which is the ability of the eye to adjust its focus to
A single specimen in the NMMNH’s Whitaker Quarry block C- various distances. Fish eyes accommodate by moving the lens to the
8-82 preserves the first complete sclerotic ring documented in Coelophysis front or rear along the optical axis. Mammals possess a ring of muscles
bauri (Fig. 82). Of the seven incomplete to complete Coelophysis skulls that apply radial stretching force to the periphery of the lens. Birds and
that are exposed on the block only two, NMMNH P-42200 and P- reptiles uniquely share an arrangement in which rings of striated, ciliary
42353, preserve part or all of a sclerotic ring. Additionally, incomplete muscles (Bruke’s and Crampton’s muscles, respectively: m. cornealis
sclerotic rings are present in MCZ 4327, RHMP GR 144 and RTM posterior and m. cornealis anterior) circumferentially squeeze the pe-
TMP 84-63-29. Colbert’s (1989) preliminary description was based on riphery of the lens to provide accommodation. Crampton’s muscles
the MCZ specimen. originate on the sclerotic ring, and Bruke’s muscles originate on the sclera
The presence of the sclerotic ring in the near-perfect, three-dimen- near the ring (Martin, 1985). This lens-squeezing type of accommoda-
sional skull of P-42200 (Fig. 82) permits us to reconstruct the eye mor- tion produces changes in the internal pressure of the eye that are resisted
phology of Coelophysis more accurately than is usual for a fossil species. by the sclerotic ring.
The eye is a physical device that admits, refracts, and focuses light on a No mammals or modern amphibians have sclerotic rings, and some
sensory panel. As such, its vision characteristics are, to a large degree, extant fish and reptiles lack them, notably the elasmobranchs, snakes,
determined by shape (cf. Schmitz, 2009). According to Walls (1942), one and crocodylians. Some Mesozoic crocodylians possessed sclerotic rings,
function of the sclerotic ring is to maintain the shape of the eye against its but these are absent in Cenozoic crocodylians (Bellairs, 1970). Some
internal pressure (some other functions may play a part in accommoda- very primitive tetrapods possessed sclerotic rings, but their purpose is
tion). Therefore, the sclerotic ring not only reflects the shape of the eye, less clear, because they presumably retained the fish-type of accommo-
but to a large extent it defines the shape of the eye. The inside diameter dation (Clack, 2002). Numerous Paleozoic amphibians, including
of the sclerotic ring fixes the cornea size and therefore determines the temnospondyls (Hunt et al., 1992), and microsaurs (Carroll and Gaskill,
maximum quantity of light that may be admitted. The soft structures of 1978), possessed sclerotic rings. A compilation of data and references on
the eye, particularly the iris, imposes further limits on light admittance. sclerotic rings in various fossil taxa (Table 2) demonstrates their broad
The cross-sectional shape of the ring is, in general, associated with one of distribution among extinct tetrapods.
three basic bird and reptile eye types, each of which has a known range Sclerotic rings of extant birds typically have 10 to 18 ossicles,
of acuity and sensitivity in extant animals (Walls, 1942). Thus, we use with a modal number of 15. Lacertilian rings typically have 11 to 16
careful observations of sclerotic ring morphology and its relationship to ossicles, with a modal number of 14. Extremes of as many as 40 and as
the orbit, and comparison to the morphology of extant animals with few as 6 exist, but are rare (Underwood, 1970). Sclerotic ossicles are
known vision characteristics to infer the vision characteristics of designated (+) if they overlap both neighboring ossicles, (-) if they are
Coelophysis. overlapped by both neighboring ossicles, and (i) if they are imbricated. In
Review of Sclerotic Rings and Sauropsid Eyes sauropsids there is nearly always a (+) ossicle in the ventral, or slightly
temporo-ventral position. This ossicle is numbered “one” and counting
The detailed anatomy and function of sclerotic rings and sauropsid proceeds temporo-dorsally (counter clockwise on the left side: Fig. 84B).
eyes is a topic not generally discussed by paleontologists. Therefore, we Ossicle patterns vary, but a few patterns are prevalent. In lizards,
present a brief review of these subjects as background to our comparison there is the usual mid-ventral (+), and most commonly a mid-dorsal (+)
of the Coelophysis data to those of the outgroups and to clarify the basis (-) (+) sequence (11 lizard families), and less commonly (3 lizard fami-
for our interpretation of the results. lies) a mid-dorsal (+) ossicle. A complementary pair of (-) ossicles is
Sclerotic rings are thin, bony, annuli composed of several articu- almost always found in the mid-nasal and mid-temporal positions of the
lated ossicles set in the sclera surrounding the cornea of the eye. The ring. The most common bird pattern is of a mid-dorsal and mid-ventral
stiffness of the ring varies greatly between taxa. The smaller, thinner pair of (+) ossicles and a mid-nasal and mid-temporal (-) pair of ossicles
rings (e.g., those of passerine birds and small lizards) are quite flexible in (Underwood, 1984).
life. The shape and number of ossicles also varies considerably between In some sauropsids the sclerotic ring is asymmetric. The ossicles
taxa. Possession of sclerotic rings is a primitive character in the Sauropsida on the nasal side of the ring are larger and thicker in owls (Strigiformes),
86

FIGURE 82. A, The skull of NMMNH P-42200, 123 mm in length, representing a large juvenile Coelophysis bauri (cm scale). B, Close-up stereo pair of
the orbit and sclerotic ring of NMMNH P-42200 (cm scale).
 87

TABLE 2. Sclerotic rings known in some representative fossil tetrapods.

FIGURE 83. A typical sauropsid eye in longitudinal section (along the bulbar
axis), showing the corneal and orbital segments. The optical axis length and
diameter reflect the globose eye type. The sclerotic ring maintains the eye’s
shape against its internal pressure (arrows).

hawks and eagles (Falconiformes) and nocturnal or crepuscular night-

hawks (Caprimulgiformes) (Walls, 1942), which are generally aerial preda-
tors of insects and occasionally small birds (Walls, 1942; Terres, 1991).
These birds are fast-moving predators that engage fast-moving, highly
maneuverable prey. It is not obvious, however, why this shared behavior
would produce or require an asymmetric sclerotic ring. It may have a
function in allowing for a high degree of binocular overlap in the visual
fields or in the very rapid accommodation of the eye as prey are engaged.

Eye size, shape and movement

Exceptions exist, but in general, the eyes of both lizards and birds
are large and essentially fill the orbit. This observation, corroborated by
our dissection data, makes orbit size a general indicator of eye size in
sauropsids, and we treat it as such here.
In lizards a cartilaginous interorbital septum separates the eye-
balls in the center of the skull (Bellairs, 1970). Eye movement capability
in lizards ranges from completely fixed in less visually-oriented species
(e.g., Heloderma, the Gila Monster) to extremely mobile (e.g., Chame-
leon has 180° horizontal, 90° vertical motion range). Most lizards have ~
40° of motion range. With few exceptions lizard eye movement is inde-
pendent in the two visual fields. Crocodylians, like amphibians, seem to
have practically no capability to move their eyes within the orbits (Walls,
1942).
Avian eyes are generally larger than those of lizards and essen-
tially meet in the center of the skull. Usually a thin, perforate, bony
septum separates them, but not always. As examples, some owls that
weigh less than 0.5 kg have larger eyes than an adult human, and the
ostrich has the largest eyes of any land vertebrate, with an optical axis
(axis bulbi) length of 50 mm.
With respect to visual acuity, one of the most important biometric
parameters of the eye is the posterior nodal distance (Martin, 1993),
because it determines the area of the retina over which the image is
projected. This is the distance from the optical node at or near the back
of the lens to the focal point (retina). The photoreceptors on the retina
are of finite size; therefore, increasing the projected image size increases
resolution of details. The practical limit of eye size occurs when imper-
fections in the refractory elements, the cornea and lens, begin to domi-
nate the process and negate the advantage of any further increase in
bulbar axis length or retinal area (Barlow, 1986).
In most birds and in some lizards the eyes fit tightly in the orbit
and have practically no mobility. Owl eyes, for example, are rigidly fixed
88
TABLE 3. Measurements of selected sclerotic ossicles from Coelophysis
bauri (NMMNH P-42200); two falconine birds, a Bald Eagle, a Red Tailed
Hawk; and a caprimulgid, the Common Nighthawk. The corneal apertures
(IDs) are circular, but the rings are asymmetric, so outside diameter (OD) is
given as the average (ODave) of nasio-temporal and dorso-ventral diameters.
Asymmetry is principally reflected in the difference in the length and
height of nasal and temporal ossicles, the nasal ossicles always being the
larger of the two. Quantities C-C’, F1, and F2 define sigmoid flexure.

FIGURE 84. A, The sclerotic ring of NMMNH P-42200 as preserved.

Three ossicles are shown in cross section. Ossicle number one shows sigmoid
flexure data designations (Table 3). B, A parsimonious reconstruction of the
sclerotic ring of Coelophysis bauri based on NMMNH P-42200.

by their tight fit in the skull, but the immobility is compensated by

extreme neck flexibility (owls can rotate their heads 270°, hawks ~ 90°).
Exceptions exist in the diving birds (e.g., penguins, gulls, cormorants)
and bucerotids (hornbills), but are relatively rare (Walls, 1942). Birds in
general, and especially long-necked birds, employ their very flexible
necks in gaze stabilization and in pursuit movements (i.e., target track-
ing) (Wallman and Letelier, 1993).
Walls (1942) described three variants of sauropsid eyes based on
shape (Fig. 85) that have been widely accepted and used (Walls, 1942;
Underwood, 1970; Sillman, 1973; Martin, 1985). The three variants are
defined by the relationship of bulbar axis length to orbital segment diam-
eter. In both reptiles and birds, the most common eye type is designated
“flat” because the bulbar axis is shorter than the orbital segment diameter
(Fig. 85A). This eye type is very common in the diurnal lacertilians
(principally Iguanidae, Agamidae, Teiidae, Scincidae, Lacertidae, Anguidae
and Varanidae) (Underwood, 1970), and in the Passeriformes,
Columbiformes, Galliformes, Gruiformes, Anseriformes, and most other
bird orders. The associated sclerotic ring is approximately flat, but usu-
ally shows a very slightly concave, convex, or sigmoid cross section (Fig.
86). In the second eye type, the “globose” eye, the bulbar axis is approxi-
mately the same length as the diameter of the orbital segment (Fig. 85B).
This produces a larger image on the retina and more acute vision. This
eye type is characteristic of the Falconiformes, a few passerine birds
(e.g., the crow family), and some lizards. Especially in the falconine
birds, the sclerotic ring of this eye type is strongly sigmoid in cross
section and asymmetric, being substantially heavier on the nasal side axis that is longer than the orbital segment diameter and sacrifices acuity
(Fig. 86). The eyes of the lizard families Pygopodidae, Gekkonidae, and for light sensitivity, as is required by the nocturnal or crepuscular owls.
Xantusiidae show the nocturnal adaptations of a reduced sulcus and slit The sclerotic ring of this eye type is a long, robust, tubular structure (Fig.
pupil, and are more globose than those of diurnal lizards. An approxi- 86). Considerably less of a continuum exists between the globose and
mate continuum of shapes exists between the flat and globose eye vari- tube eye types than exists between the flat and globose variants. A few
ants in both lizards and birds. fossorial lizards (e.g., Amphisbaena) possess a tube-shaped eye, but
The third eye type, “tubular” (Fig. 85C) is almost strictly limited their degenerate eyes essentially lack a sclerotic ring and are certainly not
to the Strigiformes and a few eagles. The tube-shaped eye has a bulbar comparable to the owl eye.
 89
displaced under number 1. There is no evidence of a (+) (-) (+) series in
the ring (common in extant lizards), but the dorso-temporal area where
this series would most likely occur is slightly disarticulated, so it may
have existed. The most parsimonious interpretation is that there are 20
ossicles with a mid-dorsal and mid-ventral (+) pair, a mid-nasal and mid-
temporal (-) pair and all other ossicles (i). We find, therefore, that Colbert’s
(1989) interpretation of the number, distribution, and relationship of the
sclerotic ossicles in C. bauri was essentially correct. In the following
paragraphs, we describe details of the sclerotic ossicles and ring mor-
phology that were unavailable prior to this study.
As preserved, the inside diameter (ID) of the ring is 13.8 mm
naso-temporally, and 11.6 mm dorso-ventrally. The outside diameter
(OD) is 25.5 mm naso-temporally and 22.9 mm dorso-ventrally. The
orbit of P-42200 (after repositioning the postorbital) measures 26.8 mm
dorso-ventrally and 34 mm antero-posteriorly. The included area (treated
as an ellipse) is 726 mm2. The area of the sclerotic ring, excluding the
corneal aperture, is 281 mm2. The ring thus occludes 38% of the orbit
area.
Measurements of the best-exposed individual ossicles were made
using digital calipers (Table 3). They demonstrate that the ring is asym-
metric and the ossicles on the nasal side of the ring (12 through 20) are
larger than the corresponding ossicles on the temporal side (2 through
10). The temporal ossicles also appear thinner based on their apparent
opacity and the delicacy of their edges. The sigmoid flexure of the os-
sicles was quantified by applying a straightedge to the best-exposed
ossicles and measuring from the straightedge down to the ossicle surface.
This was accomplished using precision calipers under a microscope. The
cross section of ossicle number one (Fig. 84A) shows the designators
used to quantify flexure. The area enclosed between the straightedge and
the ossicle approximates an arc and its chord. The chord length is desig-
nated C-C’, and the arc height is designated F1. Quantity F2 describes the
extent of the downward flex at the outer circumference of the ossicle.
Ossicle length, height, and sigmoid flexure data for selected sclerotic
ossicles of P-42200 can be compared to those metrics for an eagle, a
FIGURE 85. The three types of sauropsid eyes. A, The flat eye of a swan. B, hawk, and a nighthawk (Table 3). These three birds represent typical
The globose eye of a hawk. C, The tubular eye of an owl. A-C, to the same extant taxa that possess asymmetric, sigmoid cross-section sclerotic rings
scale. After Sillman (1973) and Walls (1942). generally similar to the sclerotic ring of Coelophysis bauri.

The Sclerotic Ring of Coelophysis bauri Reconstruction of the Sclerotic Ring and Orbit

Colbert (1989) described an incomplete sclerotic ring of Coelophysis We based guidelines for reconstructing the sclerotic ring of
in MCZ 4327. Five or six articulated ossicles were found in the dorsal Coelophysis bauri on comparisons to 30 extant lizards and birds
part of the orbit, and three were displaced in the temporal area of the (crocodylians lack sclerotic rings). We also made comparisons to com-
orbit. Based on this specimen he speculated that the complete ring would plete articulated or reconstructed rings of other dinosaurs (Brown, 1912;
comprise 20 ossicles with a pair of dorsal and ventral ossicles that over- Parks, 1928; Russell, 1939; Romer, 1956; Galton, 1974). In general, the
lap both their neighbors, designated (+), a pair of nasal and temporal ossicles were arranged so as to eliminate substantial notches around the
ossicles that are overlapped by both their neighbors, designated (-), and outer circumference and to allow for an overlap of about half the ossicle
four imbricated (i) ossicles between each of these. Colbert cited the width at the widest point (Brown, 1912; Galton, 1974) (Fig. 84B).
existence of such a ring in Struthiomimus samueli (Parks, 1928). The temporo-dorsal region of the ring in P-42200 appears to have
Specimen P-42200 has an essentially complete left ring in situ suffered the same type and degree of taphonomic distortion as the pos-
(Fig. 82). This is the first complete sclerotic ring known for the species. torbital. When the temporo-dorsal portion of the ring is adjusted in the
The sclerotic ring in P-42200 shows slight taphonomic distortion in same direction and amount required to correctly reposition the postor-
which the temporal portion of the ring and the postorbital are displaced bital (2.6 mm naso-ventrally), the corneal aperture of the ring becomes
temporo-dorsally ~ 2.6 mm, and the postorbital is rotated ~ 25° clock- circular. Almost without exception in the taxa used for comparison the
wise. Still, its excellent preservation allows us to test and elaborate upon corneal aperture is circular. In only two cases of extremely small, fragile
Colbert’s (1989) hypotheses. lizard sclerotic rings did this aperture depart from circularity. These
Individual ossicles are rounded trapezoids in shape (Fig 84A), cases appeared to be due to desiccation of tissue remaining on the deli-
except the nasal and temporal ossicles, which appear more ovoid and cate structures.
elongate. The bases of the trapezoidal ossicles form the corneal aperture Eighteen ossicles are visible “face-on” in the sclerotic ring of
of the ring. The ossicles in the most undisturbed areas of the ring overlap Coelophysis bauri (NMMNH P-42200) (Fig. 84A). We posit the exist-
the adjoining ossicle by approximately one half of its width. In the naso- ence of two more because their edges can be seen under numbers 12 and
dorsal area, one ossicle is displaced out of the ring, and one appears to be 19-20. It is conceivable that these edges represent exfoliated flakes of
displaced under ossicle 12. In the naso-ventral area, one or two ossicles bone from other ossicles, but we take them at face value and reconstruct
appear to be displaced under numbers 19 and 20. Ossicle 20 is somewhat the ring with 20 ossicles (Fig. 84B).
90

FIGURE 86. Stereo pair showing the sclerotic rings typically associated with the three eye types. Nasal direction is to the left. The large sclerotic ring of
the Great Horned Owl, Bubo virginianus (Strigiformes) associated with the tube type eye. The sclerotic ring of Haliaeetus leucocephalus, the Bald Eagle
(Falconiformes), sigmoid in cross section, asymmetric, and more heavily constructed on the nasal side, is associated with the globose eye type. Two nearly
flat rings associated with the flat eye type; the larger ring is from Columba fasciata, the Band-tailed Pigeon (Columbiformes), the smaller is from the
Common European Starling, Sturnus vulgaris (Passeriformes).

Outgroups
Projecting to adult size
The eyes of juvenile Coelophysis are disproportionately larger Adult extant lizards and birds (Appendix 4) were chosen as
than those of adults (Colbert, 1989; Chure, 1998). NMMNH P-42200 outgroups for morphometric comparison to the sclerotic ring, orbit, and
(Fig. 82) is a juvenile with a skull length of 123 mm. Direct comparison skull metrics of Coelophysis. Lizards, as lepidosaurs, and birds phyloge-
of P-42200 to adult animals would create inaccuracies because of its netically bracket dinosaurs (e.g., Gauthier et al., 1988; Witmer, 1997). As
disproportionately large eyes. We alleviated this problem by scaling the the only extant archosaurs, crocodylians are frequently used as an
sclerotic ring of P-42200 to adult size (Reiss, 1989). outgroup in dinosaur studies, but extant crocodylians lack sclerotic rings
Choosing the neotype of Coelophysis bauri, AMNH 7224, as a and are thus irrelevant to this study. The outgroup members were chosen
typical adult that agrees well with average skull length of fully adult to represent a wide range of taxa with diverse habitats, sizes, and feeding
Coelophysis, 207 mm (Rinehart, 2001), we scaled the sclerotic ring of P- strategies.
42200 to fit this skull. The skull length and orbit dimension that were The lizard outgroup includes nine species from five families. In
used to generate the scaling equation are presented in Table 4. Linear brief, the outgroup comprises: a fast runner and able swimmer, the agamid
scaling was used because during our dissection of bird eyes we observed Green Water Dragon (Manthey and Schuster, 1996); two Tegus, fast-
that the eye generally filled the orbit throughout growth, and our allom- running, predatory teiids (Ditmars, 1967; Zug, 1993); one large and one
etric data showed that the sclerotic ring diameter was approximately small species of the diurnal, carnivorous Varanidae, a Crocodile Monitor,
isometric with respect to the orbit diameter (Fig. 87). Bird dissections and a spotted Tree Monitor, respectively (Coborn, 1997); three iguanids,
included growth series of robins, roadrunners, and some doves (all were including the slow-moving, terrestrial, herbivorous Chuckwalla and the
failed rehab birds collected from rehabilitation centers in New Mexico). arboreal Green Iguana, which is generally herbivorous, but apparently
The same dissections and observations of the orbits and sclerotic rings of relishes baby birds (Hatfield, 1993); and the fast, pugnacious crotaphytid,
skeletonized lizards also showed that orbit length much more accurately the Collared Lizard (Degenhardt et al., 1996), a predator on insects,
reflected eye size than orbit height. This is because the dorsal portion of lizards and, occasionally snakes.
the orbit is generally sculpted (to varying degrees) back into the skull The bird outgroup includes a very diverse assortment of 21 spe-
roof, so the orbit height is often larger than the actual eye diameter. We cies from 11 orders. Represented are four hawks of various sizes and two
thus use orbit length throughout as the preferred orbit size metric. Our eagles, one large and one small owl, three cranes ranging in size from a
adult skull, then, has an orbit height of 33.3 mm and length of 30 mm. Whooping Crane to a small rail, a ratite, a woodpecker, a roadrunner, a
Proportionate scaling of the sclerotic ring yielded an average outside goose, a pelican, a nighthawk, a pigeon, a starling and a crow. The basal
diameter of 24.8 mm and inside diameter of 13 mm. We used these bird Archaeopteryx lithographica was also included in the study. Mea-
dimensions for comparison to the outgroups in addition to the P-42200 surements of the sclerotic ring and skull were made from a high resolution
“as preserved” dimensions. “As preserved” data have the taphonomic cast (NMMNH C-4600) of the Berlin specimen. Two Archaeopteryx
distortion removed, but are not scaled to adult size. specimens preserve a skull and sclerotic ring. The Berlin specimen was
 91
TABLE 4. Skull and orbit length, and orbit height of various undeformed
Coelophysis bauri skulls. These data were used by Rinehart (2001) to generate linear regression line, which expresses the “average” value of eye volume
a scaling equation for projecting the orbit of NMMNH P-42200 to adult across the skull volume range of these sauropsids. The slope of the line,
size. 0.85, is the somewhat negative allometric constant that indicates rela-
tively larger eyes in the smaller species. The residual values, which are
the distances of individual data points above or below the “predicted” or
“average” value indicated by the regression line, show the extent that a
particular species varies above or below the average eye investment of
these sauropsids. Although the actual pertinent numerical values here are
the root mean squares residuals, to judge whether individual data points
are more or less above or below average their simple vertical or diagonal
distance from the line may be compared by eye. With respect to the
“average” line, the lizards generally (seven out of nine lizards) fall below
average. The exceptions are the tegu, Tupinambus teguixin, and the chuck-
walla, Sauromalus obesus. The nonraptorial birds divide nearly equally
above and below the average line. As mentioned above, the pelican,
Pelecanus erythrorhynchos, probably rates very low because its ex-
tremely long beak artificially inflates its skull volume. The highest re-
sidual value seen among the 11 nonraptorial birds is that of the ratite,
Rhea amaracana. All except one (Cooper’s hawk, Accipiter cooperii) of
the six Falconiformes (hawks and eagles) show relatively high positive
residuals, as do the owls and the nighthawk. The owls probably show
artificially low positive residuals because their eyes are tubular and project
out of the skull, so the spherical model used would yield a low value.
Archaeopteryx has a relatively high positive residual, so it had much of
its skull volume invested in its eyes. Significantly, in Coelophysis, both
the “as preserved” juvenile specimen (P-42200) and the adult show
positive residual values that rank between the falconine and the
nonraptorial birds with the uppermost values; it also had a very large
skull volume invested in its eyes.

Morphological assessment of the sclerotic ring

We compared the sclerotic ring morphology of Coelophysis to
those of the outgroups in three distinct ways. Firstly, an “on-axis” com-
parison was made by plotting the LOG transforms of the area of the
sclerotic ring (projected on a plane in the case of non-flat rings) versus
the orbit area (Fig. 89A). The ring area was calculated as the difference
between two concentric circles, one circle being defined by the inside
FIGURE 87. Allometric plot of bird sclerotic ring inside diameter (ID) diameter (ID) and the other by the outside diameter (OD). In asymmetric
versus orbit diameter demonstrates an isometric relationship. rings the average OD was used. Orbit area was modeled as the area of an
ellipse, where the orbit length and height were the major and minor
chosen over the Eichstatt specimen because it is an adult and its sclerotic diameters. In this comparison Coelophysis groups with the hawks and
ring is better preserved. Again, choosing the adult specimen was impor- eagles (contained within the ellipse in the plot), including both the juve-
tant because there is variation between the relative sizes of the skull and nile and the adult data points. Archaeopteryx falls into the overlap area
the orbit during the ontogeny of Archaeopteryx (Houck et al., 1990). of nonraptorial birds and lizards.
Skull width, a parameter needed for volume calculations, could not be Secondly, we compared the sclerotic ring cross-sectional shapes
measured directly from the cast. Skull width was estimated using a skull (perpendicular to the bulbar axis) (Fig. 89B). The rings judged to be
height-to-width ratio derived from reconstructions of the Eichstatt speci- convex or concave were only very slightly so. These rings and the flat
men (Wellnhofer, 1974), and the Berlin specimen (Buhler, 1985). In both rings all belong to animals of the flat eye type. The sigmoid group,
reconstructions the skull width is a few percent less than the height. We characteristic of the globose eye type, contains Coelophysis, the hawks
thus estimated Archaeopteryx’s skull width at 96% of the measured and eagles, a few lizards and nonraptorial birds (heavy ellipse in Fig.
height. 89B). Because there is a continuum of shapes between the flat and
Comparison of Coelophysis to the Outgroups globose eye variants, the location of the boundary between these types
is not sharply defined and is necessarily subjective. Archaeopteryx, with
To estimate the relative importance of vision to the various study its slightly concave ring, probably had the flat eye type.
animals we plotted the allometry of eye volume (both eyes) versus skull Finally, the asymmetry of Coelophysis’ sclerotic ring is compared
volume (Fig. 88). A first approximation was achieved by modeling the to that of several falconine birds (Fig. 89C). Clearly, the asymmetry of
eyes as simple spheres of mean orbit diameter and the skull as an ellipti- Coelophysis sclerotic ring is comparable to that of hawks and eagles. The
cal cone. The basal skull height and width were treated as the major and Coelophysis sclerotic ring is both sigmoid in cross section and asym-
minor diameters of the cone base. The skull length (including the beak in metrical and therefore similar to the Falconiformes and Caprimulgiformes,
birds) was taken as the cone height. These are admittedly crude esti- but dissimilar to the Strigiformes, which are asymmetric but tubular in
mates, especially in the case of long-beaked birds such as the pelican, but cross section (Fig. 86A, Table 3). No other bird groups have significantly
they were applied equally on all study animals, so the relative errors asymmetric sclerotic rings (Walls, 1942). The sclerotic ring of Archaeop-
throughout should be approximately nullified. teryx was judged to be symmetric based on the reconstruction of Buhler
The plot (Fig. 88) shows interesting groupings with respect to the (1984) and on the Berlin specimen cast measurements. The sclerotic ring
92

FIGURE 88. Allometric plot of estimated skull volume versus eye volume.
The linear regression represents the allometric “average.” The residual
values (distance above or below the line) show the relative “eye investment”
of the various groups. Seven of the nine lizards fall below the line.
Nonraptorial birds are scattered almost evenly above and below the average.
Five of the six Falconiformes, the owls, and the nighthawk are above the
average. Also plotted are Archaeopteryx (filled square) and Coelophysis
(large and small filled circles represent adult and juvenile, respectively).

of the Eichstatt Archaeopteryx specimen appears to be asymmetrical,

but Wellnhofer (1974) indicated that the circumferential edges of the
temporal area ossicles are broken.
In the lizard outgroup only the fast-running predatory agamid
Physignathous cocincinus, the Green Water Dragon, has a notably asym-
metrical sclerotic ring. Its ring is also sigmoid in cross section, as in the
raptorial birds. The rings of two other predators, Varanus salvadori, and
a teiid, Tupinambus teguixin, are also somewhat asymmetrical, but less
than in P. cocincinus. Within the constraints of the outgroups used here,
the asymmetrical sclerotic ring is only associated with fast, agile preda-
tors.

Night vision estimate

Good night vision may be enhanced by possession of large eyes.
In most small birds and lizards the eyes are too small for nocturnal use.
However, possession of large eyes does not assure good night vision. FIGURE 89. Plots of sclerotic ring and orbit metrics. A, Allomtery of orbit
The cornea must also be large enough to admit adequate light (Walls, area versus sclerotic ring area; a comparison of “on-axis” morphology.
1942; Martin, 1985). Simply scaling up a diurnal eye only spreads the Coelophysis groups with the hawks and eagles (inside the ellipse).
faint image over a larger area on the retina; the image remains faint. To Archaeopteryx is in the overlap area of nonraptorial birds and lizards (filled
increase the image brightness on the retina the aperture size of the eye square). B, Same axes as A, except not LOG transformed, but groupings
must be increased with respect to overall eye size. Therefore, the rela- show cross sectional sclerotic ring shape, a comparison of morphology
tionship of orbit size (length) (~ eye size) and maximum aperture size (~ perpendicular to the bulbar axis. Coelophysis (large filled circle, adult, and
sclerotic ring inside diameter, ID) was chosen to assess night vision. small filled circle, juvenile) groups with hawks and eagles plus a few
Additional night vision factors, which are not accessible by our methods, nonraptorial birds and lizards in the heavy ellipse. Archaeopteryx (filled
include the density of rod type photoreceptors on the retina and lens square) is well within the convex-concave-flat sclerotic ring group strongly
position within the eye (the lens being farther back in the night-adapted indicating a flat eye type. C, Allometry plot showing the asymmetry of the
sclerotic rings of some falconine birds compared to Coelophysis bauri.
 93
eye). of flesh reconstructions apparently show preorbital grooves in C.
An allometric plot of LOG sclerotic ring ID versus LOG orbit rhodesiensis that are similar to those we describe here in Coelophysis
length shows separate curve fits for noctournal birds, Falconiformes, bauri.
nonraptorial birds and lizards (Fig. 90). The uppermost and lowermost We applied Snell’s Law to determine whether light arriving at the
curves indicating the best and poorest night vision. The nocturnal spe- cornea of Coelophysis by way of the preorbital groove would be admit-
cies, owls and the nighthawk, in the uppermost curve have the best night ted into the eye or reflected away. Snell’s Law defines incidence, trans-
vision. The Falconiformes fall below the noctournal birds and above the mission, and reflection relationships for electromagnetic radiation (in-
nonraptorial birds. The lizards show the poorest night vision. Archaeop- cluding light) as it transitions boundaries between various media, and
teryx had absolutely small eyes, but probably had relatively good night calculates the “Brewster’s angle,” at which incident waves reflect off a
vision, though not as keen as the truly nocturnal group, but equivalent to surface (such as water) as opposed to entering it. Snell’s equation was
the falconine birds. Coelophysis’ position between the nonraptorial birds solved for the minimum angle of incidence encountered at the cornea of a
and the lizards probably indicates relatively poor night vision. globose eye with a 40° visual field overlap. We assumed a tangent to the
corneal surface to calculate the angle of incidence and a corneal refractive
Binocular overlap measurement index of 1.38 (Hecht, 1987). The calculation showed that the light would
The almost-perfect three-dimensional preservation of the enter the cornea rather than reflect away. All indications are, then, that
NMMNH P-42200 Coelophysis skull (Fig. 82) indicates that its preor- Coelophysis, with a globose, hawk-like eye could utilize the full 40° of
bital skull morphology enhanced binocular overlap of the visual fields. overlap in its visual fields, because light arriving from even the most
The preorbital process of the lacrimal bows gently inward approxi- oblique angle permitted by the shape of the pre-orbital skull would enter
mately 2.4 mm over its 15 mm length, thus bringing the visual fields more the eye.
forward. A wide, shallow preorbital groove (typically termed the “cheek A bar graph (Fig. 92) compares the visual field overlap in the
groove”) extends anteriorly from the preorbital process of the lacrimal, outgroup animals and crocodylians (Walls, 1942) to Coelophysis. In each
across the antorbital fenestra, to the area just dorsal and posterior to the outgroup the short bar shows the least and the long bar the greatest visual
external nares. This groove is a forward continuation of the inwardly field overlap for the group. In the case of Coelophysis the short bar
bowed lacrimal and clearly promotes binocular overlap of the visual indicates the result if the globose eye is fitted to the adult skull as
fields. illustrated by Colbert (1989) whereas the long bar shows the globose eye
We estimated the binocular overlap in Coelophysis by a simple fitted to the skull of NMMNH P-42200. Again, Coelophysis, with 33°
ray-tracing method (Fig. 91). A straightedge positioned down the center to 40° of overlap, groups best with the hawks and eagles.
of the preorbital groove from the preorbital process of the lacrimal to the Discussion
bridge of the nose above and behind the nostrils subtends an angle of 13°
with the centerline of the skull. The absolute minimum visual field over- Paul (1988) advocated that dinosaurs were visually-oriented ani-
lap then, assuming a very low-profile flat eye that is just able to see over mals, and theropods were particularly large-eyed and possessed acute
the lacrimal, is 26°. If the orbit is fitted with a globose eye with a bulbar vision. Paul also concluded that stereopsis developed in tyrannosaurs
axis equal to its diameter, as is indicated by the preceding morphometric for the purpose of improving the accuracy of head strikes. Tyrannosaurids
analysis, the eye bulges slightly out of the orbit and produces a visual apparently acquired more front-facing eyes throughout their evolution,
field overlap of 40°. presumably enhancing their stereopsis (Currie, 1997). Chure (1998) points
The binocular overlap measured on the juvenile skull of NMMNH out some acute vision and stereopsis implications of orbit morphology
P-42200 may change as the animal grows. Using the reconstruction of an in theropods, but cautions that these hypotheses are speculative and
adult skull by Colbert (1989), the overlap is reduced to < 33°. Similar practically untestable. Our approach of comparing sclerotic ring, orbit,
results are found in the skull of the closely-related ceratosaur, Coelophysis and skull metrics to phylogenetically bracketing outgroups with known
rhodesiensis (Tykoski, 1998), if it is fitted with globose eyes. Using the vision characteristics reduces the speculative quality of the results
skull reconstructions of Raath (1977), ~ 32° to ~ 38° of visual field (Rinehart et al., 2004d). Similar vision assessments, using sclerotic ring
overlap result, depending on eye size. Raath’s (1977, pl. 5) photographs and orbit metrics, have recently been applied to birds (Hall, 2004) and to
ichthyosaurs (Schmitz, 2004).

Eye investment
The comparison of eye volume to skull volume (Fig. 88) shows an
overall trend of reduction in relative eye size as skull size increases. This
allometric effect encompasses all the study animals and may reflect the
point of diminishing returns reached when imperfections in the refrac-
tory elements begin to dominate visual acuity and larger eyes produce no
additional benefit. In interpreting Figure 88 it should be borne in mind
that all of the outgroup animals are highly visual. The average eye invest-
ment among these animals is therefore much greater than if marginally

FIGURE 90. An allometry plot of sclerotic ring inside diameter versus orbit FIGURE 91. A sketch of the skull roof of NMMNH P-42200 shows the
diameter (length) for nocturnal birds, Falconiformes, nonraptorial birds and binocular visual field overlap if the skull is fitted with globose eyes as
lizards. Coelophysis ranks between the nonraptorial birds and lizards. indicated by the morphology study.
94
visual or poorly visual animals had been included. the Tawny Owl’s threshold is two orders of magnitude lower, ~ 10-6
Coelophysis groups between the nonraptorial and falconine birds. cdm2 (Martin, 1982). Clearly, tremendous dynamic range is required for
Within this group only the American White Pelican and Rhea have larger the nocturnal eye to function in starlight and yet not be completely
skulls than Coelophysis. Of these, only the Rhea has a larger skull vol- overwhelmed by full sunlight.
ume invested in its eyes. By any measure, Coelophysis bauri was large- In reptiles a large fraction of this range is purchased by the action
eyed. Coelophysis kayentakatae, an Early Jurassic ceratosaur that is of the pupil. The round pupil can only adjust its aperture over a limited
very closely related to Coelophysis bauri (Tykoski, 1998), was also range without requiring impossibly large muscle contractions. The slit
large-eyed. Adequate measurements for its inclusion in this study were pupil, however, can contract its aperture to an extremely small area, and
not available, but Rowe (1989) described its eyes as strikingly large and in some cases, until it is completely closed. Slit pupils, then, are a night
crowded in the orbit. vision adaptation that has nothing to do with seeing in the dark. Rather,
Paul (2002) assessed Archaeopteryx’s vision to be excellent based they allow the rod-rich, night-adapted eye to function in daylight with-
on large orbit size and near filling of the orbit by the sclerotic ring. Our out being overwhelmed (Walls, 1942). Few, if any birds have slit pupils.
work agrees, indicating a skull volume investment in the eyes of Archae- Most extant diurnal species of reptiles that have slit pupils (e.g., some
opteryx that is greater than in most nonraptorial birds and equal to that of snakes) are considered secondarily diurnal.
some raptors. As stated earlier, large eyes alone do not guarantee good night
vision. To increase the image brightness on the retina the aperture size of
Sclerotic ring morphology the eye must be increased with respect to overall eye size. This is why
The ossicle pattern in Coelophysis appears to be the same as the we choose the relationship between orbit length (~ eye size) and maxi-
prevalent bird pattern. Taphonomic disruption of the ring, however, mum aperture size to evaluate possible night vision capability. The
prevents precise definition of the pattern. The morphological assess- Common Nighthawk’s eyes are among the smallest in the outgroups, yet
ment of sclerotic rings shown in Figure 89 allies Coelophysis’ eyes very it is a nocturnal predator that engages insects on the wing, although there
convincingly with those of the falconine birds. In particular, Coelophysis may be some close-in echolocation involved in the capture (Terres, 1991).
falls squarely in the center of the hawk and eagle grouping (Fig. 89A). Our estimate of Coelophysis’ night vision potential ranks it no
These plots, together with the ring asymmetry comparisons (Table 3), better than most lizards and lower than the lowest nonraptorial birds. It
leave little doubt that Coelophysis had a hawk-like, globose eye. By appears unlikely that it was active at night. Coelophysis’ relatively poor
virtue of its long posterior nodal distance, this eye would have had very night vision probably indicates a round, as opposed to slit pupil. Ar-
high acuity. chaeopteryx ranks second highest among the nonraptorial birds, raising
In Figure 89A the Archaeopteryx data point falls near the overlap the question as to whether it was crepuscular.
area between lizards and nonraptorial birds. Its data point position in
Figure 89A, together with its slightly concave (nearly flat) sclerotic ring Binocular vision and stereopsis
(Fig. 89B), strongly indicate a flat eye. Buhler (1985) notes that morpho- Binocular overlap of the visual fields does not prove the existence
logically the optic lobes of Archaeopteryx have an intermediate position of depth perception by stereopsis (Pettigrew, 1986; Molnar and Farlow,
between modern birds and crocodylians. It is an interesting parallel that 1990; Molnar, 1991; Chure, 1998). It is generally known that stereopsis
our results with respect to eye volume investment and sclerotic ring exists in raptorial birds, and that this character is associated with their
morphology, always place Archaeopteryx in the overlap area of the ex- large binocular overlap, but exceptions do exist in the nonraptorial birds.
tant bird and lizard outgroups. Konishi and Pettigrew (1975) investigated vision in the Oilbird (Steatornis
caripensis), an unusual frugivorous caprimulgid. This bird has the typi-
Night vision cal front-facing eyes of the predatory caprimulgids but feeds entirely on
Natural luminance, measured in candles/m2 (cdm-2), varies by over fruit. Konishi and Pettigrew showed experimentally that the images from
seven orders of magnitude between unobstructed sunlight (~ 105 cdm-2), the Oilbird’s visual fields, with their large binocular overlap, were not
and dimmest starlight (~ 3.5·10-3 cdm-2). Thick clouds reduce these levels processed by the visual wulst (equivalent to the visual cortex of mam-
by a factor of ten. Full moonlight yields ~ 0.5·10-1 cdm-2 (Martin, 1985). mals) to produce a stereoscopic image. Rather, the front-facing eyes
The absolute minimum visual threshold of the pigeon is ~ 10-4 cdm-2, and seem to be an additional night vision adaptation that reduces the aberra-
tions inherent in an eye with a large aperture by bringing the optical axes
more nearly into line with the visual interest area.
Distance determination requires image samples that are separated
in space or in time. In stereopsis the samples are acquired simulta-
neously but separated in space by the distance between the eyes. Dis-
tance judgment based on motion parallax is made by sampling the image,
then moving the eyes and sampling again. In this case, the images are
separated in time (McFadden, 1993). Some sauropsids use the motion
parallax method. Many side-eyed birds and lizards that possess little
binocular overlap (e.g., Charadriiformes, Teiidae) show head-bobbing
behavior. This behavior samples the visual fields from different angles
over time and determines distance from parallax information. Even among
animals that have similar vision characteristics and use the same distance
measuring method (e.g., cats and owls) there are substantial differences
in how the images are processed to obtain the distance information (Casini
et al., 1993).
Reconstructing the skull of Coelophysis with globose eyes that fit
FIGURE 92. Bar graph of binocular visual field overlap of Coelophysis, and the orbits yields up to 40° of binocular overlap (Fig. 91). We have shown
the highest and lowest overlap in each of the outgroups. In Coelophysis, the that Coelophysis was almost surely not crepuscular or nocturnal, so
short bar indicates the approximate overlap in an adult skull as reconstructed there is no reason to suspect this overlap to be a nocturnal adaptation, as
by Colbert (1989); the long bar represents NMMNH P-42200. Coelophysis in the Oilbird. Further, it would seem maladaptive to sacrifice up to 40°
groups best with the hawks and eagles.
 95
of peripheral vision without also acquiring the benefit of depth percep- asymmetric sclerotic ring with a sigmoid cross section occurred in any
tion. This would appear to be especially true in an animal that was not animal other than a fast-running predaceous lizard or an aerial predator.
the top predator in its environment and must have been vulnerable to Its presence in Coelophysis therefore seems to correlate the animal with
predation itself. It is therefore reasonable to conclude that, because other fast, agile, visually-oriented predators.
Coelophysis had the same degree of visual field overlap as hawks and Eye movement was probably greatly restricted in Coelophysis,
eagles, it had a similar degree of stereopsis. but this could have been compensated for by the flexibility of its long
neck. Gaze stabilization and pursuit movements of the eyes would have
Eye motion been accomplished by neck movement, as they are in extant long-necked
Although the motion range of Coelophysis’ eyes cannot be deter- birds.
mined from our assessment, the implication is that it was probably low. In assessments related to visual acuity, Coelophysis allies with
The eyes of Coelophysis were not lizard-like, so there is no indication of birds instead of lizards. With respect to the amount of skull volume
a typical lizard-like motion range. The eyes probably filled the orbit as in invested in its eyes, Coelophysis ranks between the falconine and the
the falconine birds, and probably in similar fashion had practically no nonraptorial birds. Based on its relatively poor night vision, Coelophysis
range of motion. This would not have created much of a disadvantage for probably had round pupils. In the three comparisons of sclerotic ring and
Coelophysis, because its long flexible neck could have given it bird-like orbit morphology to the outgroups (1, on-optical axis, 2, perpendicular
head mobility. An assessment of neck mobility in theropod dinosaurs is to optical axis, and 3, symmetry) and in the measurement of binocular
in progress (Samman, 2004) and may provide insight into the likelihood overlap, Coelophysis groups with the Falconiformes.
of target tracking and gaze stabilization by neck motion in small theropods. The results of this study make a strong case for convergence
between the eye morphologies of Coelophysis and the falconine birds.
Conclusions Coelophysis was built to run (Carrano, 1999, this study). Surely, it
would have used its running capability in pursuit of prey. The same
Because at least one purpose of the sauropsid sclerotic ring is to
characters that equip hawks and eagles to spot, dive on, and capture
define and maintain the shape of the eye, which is strongly associated small, fast-moving prey would have been ideal for fast-running
with vision characteristics, sclerotic ring morphology and its relation-
Coelophysis in pursuit of its small, fast-moving prey. These characters
ship to the orbit can reveal much about vision characteristics in sauropsids. include: acute vision produced by large, globose eyes with a long poste-
Careful preparation and analysis of well-preserved fossil sclerotic rings
rior nodal distance; a high power of accommodation to keep prey in
can likewise provide strong indications of vision characteristics in extinct focus as distance rapidly changes; and a high degree of binocular overlap
species.
in the visual fields to allow distance determination by stereopsis. Our
The purpose of asymmetry in the sclerotic ring shape is unclear, comparative analyses of the skull, orbit, and sclerotic ring morphology
but its association among the outgroup animals appears significant. Of
strongly indicates that these characters were present in Coelophysis.
the 31 animals in our outgroups, there were no instances in which an
96
THE FURCULA OF COELOPHYSIS BAURI NMMNH P-42353
In this section we describe the furcula and reconstruct the shoul- The furcula of NMMNH P-42353 (Fig. 93A) is preserved in
der girdle of Coelophysis bauri. Birds and numerous dinosaurs possess anterior aspect and is broken, but apparently complete. It is broken just
furculae, and they are important documentation of the bird-dinosaur to the left of center, where it turns sharply and the left ramus extends
phylogenetic connection. The furcula of C. bauri is the oldest known, under the left scapula. The right ramus, which appears undistorted, lies
and thus represents an interesting and important feature. Furcula were atop the right coracoid, and the tip extends to near the acromion process
first positively associated with C. bauri and formally described by of the scapula. In dorso-ventral aspect it presents an obtuse, boomerang-
Rinehart et al. (2007). For the sake of completeness we briefly redescribe like shape, and in lateral view it bows slightly anteriorly, so that it
them here. In doing so, we draw heavily on the aforementioned work. closely hugs the anterior margin of the coracoid. Distally it tapers to a
Possession of clavicles is plesiomorphic in the Archosauria (Bryant blunt point. The center of the furcula crosses cervical vertebra number 10
and Russell, 1993). It is generally accepted that furculae are formed by near its posterior end, and crosses at least cervical rib pairs nine and 10;
midline fusion of the clavicles (Gauthier, 1986; Ostrom, 1976), although it lacks a hypocleideum.
Bryant and Russell (1993) have pointed out that cladistic analysis casts
some doubt on the homology of the dinosaurian furcula with the
archosaurian clavicle pair. Furculae have previously been well docu-
mented in all species of coelophysids except Coelophysis bauri. Tykoski
et al. (1998, 2002) described furculae in C. kayentakatae and C.
rhodesiensis. The furcula of C. rhodesiensis had been previously illus-
trated by Raath (1977, pl. 10) but misidentified as hyoids, because
furculae were unknown in the Dinosauria at the time, and one of the
furcula had been found near a skull (Tykoski, 1998). The pectoral ele-
ment originally described as a clavicle in Segisaurus (Camp, 1936; Bryant
and Russell, 1993) was reidentified by Rauhut (2003) and Carrano et al.
(2005) as a furcula. Padian (1997) illustrated and described “clavicles” in
C. bauri but our data show that these elements are, in fact, furculae.
Downs (2000) described a furcula from the Ruth Hall Museum of
Paleontology’s Whitaker Quarry block, C-9-82. He correctly speculated
that it belonged to Coelophysis bauri, but was unwilling to positively
establish the connection because it was an isolated element not in articu-
lation with any of the C. bauri in the block, and because there were other
animals in the block to which it could have belonged. This furcula is
identical to the specimens we have found in articulation in C. bauri, so
the documentation presented here confirms Down’s work (Rinehart et
al., 2007).
Description of the Furcula
We describe four essentially complete furculae and a fragment of a
fifth. Three of the furculae are in articulation in juvenile Coelophysis
bauri specimens (one or two year size class), and one is disarticulated
and not associated with any specimen but is judged to belong to a juve-
nile by its size. The fragment is large in diameter, is associated with a
large shoulder girdle, and thus is judged to belong to an adult. Measure-
ments of four C. bauri furculae are given in Table 5.
TABLE 5. Measurements (mm) of C. bauri furculae from NMMNH block
C-8-82. Length was measured along the curve of the furcula, not the straight-
line distance between the tips. Diameters were measured near the center, but
do not include the hypocleideum if present; C-C = craniocaudal, D-V =
dorsoventral. Notations ap = as preserved, e = estimated, N/A = not available.
Associated humerus length is included as an overall size indicator for the
specimens. The best preserved of the humeri was measured in each case.

FIGURE 93. Coelophysis bauri furculae (cm scale). A, NMMNH P-42353,

an articulated furcula in cranial view. B, NMMNH P-42577, an articulated
furcula with hypocleideum, left lateral view. C, NMMNH P-42576, an
associated furcula with hypocleideum in left lateral view. The furcula is
rotated ~ 90º clockwise. Abbreviations: (l), left, (r), right; acr, acromion
process of scapula; cor, coracoid; cr, cervical rib; cv, cervical vertebra; ep,
epicleideal facet; fu, furcula; g, glenoid; hc, hypocleideum; hum, humerus;
sc, scapula.
 97
NMMNH P-42577
This specimen has a complete furcula (Fig. 93B) except for the
distal end of the left ramus, which is missing approximately 3 mm of its
length. The estimate of the missing segment is based on the length and
diameter proportions of the furculae that have complete ends. Visible in
left lateral view, this furcula emerges from beneath the right coracoid,
continues across the posterior end of the 10th cervical vertebra, and its
broken end rests on the acromion process of the left scapula at the level
of the anterior end of dorsal vertebra number one. This furcula has a
hypocleideum that is approximately 2 mm long, 1 mm tall, and pointed.

NMMNH P-42576
In NMMNH P-42576 the midpoint of the furcula is in close
association where the hypocleideum lies between the coracoids. The
furcula is visible in anterior view (Fig. 93C); the left ramus is broken and
missing approximately 5 mm of length, based on symmetry about the
hypocleideum. The hypocleideum is approximately 2 mm long, ~0.75
mm tall, and distinctly pointed. The furcula is rotated approximately 90
degrees, so the right end is above the right coracoid, and the left end rests
on the left scapula near the coracoid suture and above the glenoid.

NMMNH P-46615
This specimen is a disarticulated furcula seen in posterior view
(Fig. 94A). It is well isolated from the articulated C. bauri specimens.
The element is broken into four pieces, with little or no offset between
the segments. With the bone reconstructed and the offset corrected, it
subtends an angle of ~ 120o compared to 140o for the furcula of C.
kayentakatae, and 115o to 140o for the same measurement in C.
rhodesiensis (Tykoski et al., 2002). The hypocleideum is not present.
Oval epicleidial facets near the distal ends of the rami are ~ 1.7 mm long
and ~ 1.0 mm wide and are longitudinally grooved, as in C. kayentakatae,
C. rhodesiensis (Tykoski et al., 2002) and Allosaurus (Chure and Madsen,
1996).

NMMNH P-44554
The shoulder girdle of NMMNH P-44554, an adult Coelophysis
bauri, is damaged by a large joint in the block. The left distal portion of
a furcula lies on the lateral surface of the left coracoid (Fig. 94B). The
fragment is 12.2 mm long, is approximately round, and has a 3 mm
diameter. The fragment is visible in caudal view, so the epicleidial facet, a
1 mm wide groove, is apparent. This fragment represents the only evi-
dence of a furcula in an adult C. bauri. Although it seems extremely
improbable that fusion of the clavicles would be lost during ontogeny,
this element is a distal fragment, so it does not positively demonstrate
the presence of a furcula in an adult specimen.
Allometry
Allometric plots of the furculae of juvenile Coelophysis bauri
show the relative growth of the major and minor diameters of the ellipti-
cal cross section of the furculae with respect to length (Fig. 95A) over a
limited size range (no complete adult furculae were found). The minor
diameter (allometric growth constant ~ 1.2) grows at a higher rate than
the major diameter (allometric growth constant ~ 0.4). Therefore, as
growth proceeds, the diameters become more nearly equal, and the struc-
ture becomes more rounded, an observation borne out by the incomplete
adult specimen just described. The growth of the furcula length and
FIGURE 94. Coelophysis bauri furculae (cm scale). A, NMMNH P-46615,
associated scapulocoracoid lengths with respect to the humerus length is an isolated furcula in caudal view showing epicleideal facets. B, NMMNH P-
shown in Figure 95B. The allometric constants are very close to one 44554, a fragment of the distal left ramus of a furcula lies on the lateral
(~1.1) in both the furcula and scapulocoracoid, indicating that their growth surface of the left coracoid of an adult specimen. The groove-like epicleideal
is in near isometry with that of the humerus. facet is evident. C, A reconstruction of the furcula of C. bauri. Based on
The number of data points in the allometric plots is low (3 to 5), specimens NMMNH P-42353 and NMMNH P-46615. Scales in cm.
and more data points would strengthen confidence in the accuracy of the Notations same as Fig. 93.
regressions. However, scatter about the regression lines is very low,
98
Reconstruction of the Shoulder Girdle
Knowledge of the furcula morphology should contribute greatly
to the accurate reconstruction of the shoulder girdle in Coelophysis.
However, an understanding of the shape and curvature of the
scapulocoracoid is also essential to the reconstruction. Most of the
Coelophysis bauri scapulocoracoids lie on bedding planes and have been
flattened postdepositionally. One juvenile C. bauri specimen, NMMNH
P-42578, preserves the right scapulocoracoid in a position that is per-
pendicular to the bedding, and it appears to maintain its original radius of
curvature (Fig. 96A). The coracoid appears flat, but the scapula forms an
arc that must have conformed to the curvature of the side of the animal.

FIGURE 95. Allometric plots of Coelophysis bauri furculae. Log10 transforms

of measurements in mm are plotted. A, Dorsoventral diameter (squares) and
craniocaudal diameter (circles) as a function of furcula length. B, Furcula
length (circles) and scapulocoracoid length (squares) as a function of humerus
length show near-isometric growth.

indicating a small standard error of the estimate and strengthening confi-

dence in the curve fits (Glantz and Slinker, 2001).
Discussion
Two of the four approximately complete furculae show a
hypocleideum. The presence of a hypocleideum is variable in the furcula
of some other theropod dinosaurs. Chure and Madsen (1996) described
furculae in Allosaurus. These furculae had previously been identified as
gastralia (Madsen, 1976), but this was before non-avian furculae were
known. They also show variation in the presence of the hypocleideum
(their “ventromedian process”). The hypocleidea in Coelophysis bauri
furculae are more similar to the relatively small, sharply-pointed pro-
cesses of Allosaurus furculae than to the apparently homologous but
relatively wider feature seen in C. kayentakatae and described by Tykoski
(2002) as an “apical ridge.”
All four of the approximately complete furculae we have de-
scribed taper only slightly over the distance from the midline to 2 or 3
mm from their ends, and from that point they begin to taper quickly to a
blunt point. They are elliptical in cross section at the midline and become
nearly round distally. Major (dorsoventral) and minor (anteroposterior) FIGURE 96. Reconstructed shoulder girdle of Coelophysis bauri. A, The
diameters, measured near the midline, but not including the hypocleidea, scapulocoracoid of NMMNH P-42578 in caudal view. B, Lateral view of
are given in Table 5. Several breaks and damaged areas expose the furculae reconstructed shoulder girdle. C, Cranial view of shoulder girdle with elements
in cross section and show them to be hollow near the ends of the rami, in correct spatial relationship. Abbreviations: acr, acromion; AH, arc
but fractures near the midline show a solid central area. height; AL, arc length; C, scapula chord length; COR, coracoid length; cor,
coracoid; ep, epicleideal facet; fu, furcula; g, glenoid; hc, hypocleideum; R,
radius of curvature; sc, scapula.
 99
In P-42578 the scapula has an arc length of 68 mm, a chord length of 64 shoulder girdle with the elements in proper spatial relationship (Fig.
mm, and radius of curvature of 57 mm. The associated coracoid is bro- 96B-C).
ken, and its estimated length, based on the intact left coracoid, is 14 mm. Discussion
P-42578 does not preserve a furcula.
We have now established the morphology of the scapulocoracoid Previous to Rinehart et al. (2007), all coelophysid dinosaurs ex-
and furcula of Coelophysis bauri, the position of the acromion process cept Coelophysis bauri had been shown to possess furculae. Before that
on the scapula and the epicleidial facets of the furcula that articulate to publication, the oldest well-documented furculae were those in the Early
them. The least laterally compressed of the NMMNH specimens (P- Jurassic coelophysids C. kayentakatae and C. rhodesiensis. The confir-
42353: Fig. 93A), as well as documentation in other taxa, for example mation of furculae in Apachean-aged C. bauri documents the appearance
Segisaurus (Carrano et al., 2005, fig. 5), show that the furcula closely of these elements in the Late Triassic and shows that furculae are
conforms to the anterior margin of the coracoid. This fact, together with synapomorphic in the Coelophysidae, and therefore possibly more ba-
the length and angle of the furculae, determines the spacing and angles of sally in theropod dinosaurs.
the shoulder girdle elements and allows us to reconstruct the complete
100
GROWTH, ALLOMETRY AND POPULATION DYNAMICS Probability plotting
Introduction The first step in statistically generating growth data is to test for
size (= age) classes in the metric data, which should be expressed as
Mass death assemblages such as that of the Whitaker Quarry at subpopulations, so that the total population is multimodal. Harding
Ghost Ranch provide an excellent opportunity to analyze ontogenetic (1949) showed that probability plots could be used to resolve the com-
growth, variation, allometry, population dynamics and paleoecology ponent modes of multimodal distributions and that this method was
(Brinkman et al., 2007). It is relatively rare in the realm of vertebrate useful in biology and paleontology. This method, and extensions and
paleontology for enough individuals to be present in a single sample to variants of it, have been used to define size classes and produce growth
support solid statistical analysis. The large population sample of and population data for extant wild animal populations (e.g., Semlitsch,
Coelophysis bauri at Ghost Ranch invites the use of statistical methods 1980), and we use it here to test for statistical size classes in the Ghost
that require large populations for maximum accuracy. Properly applied, Ranch Coelophysis data. Statistically distinct size classes, which repre-
statistical methods can yield information on growth, demographics, on- sent the component modes of the multimodal distribution of sizes that
togenetic life stages, sexual dimorphism, and other quantities and quali- make up a population, plausibly correspond to age groups (Hunt, 1967;
ties (e.g., Stoddart, 1980; Oxnard, 1987; Bennett, 1995; Carrano, 1999; Raup and Stanley, 1971; Andrews, 1982; Bennett, 1995; Brinkman et al.,
Rinehart and Lucas, 2001; Rinehart et al., 2002a, b, 2005c; Lucas and 2007) and may thus be used to generate growth curves and age distribu-
Rinehart, 2005; Brinkman et al., 2007). We thus attempt here to extract tions.
growth, allometry, variation, age and mass distribution, and survivorship Probability plotting is a simple, powerful, graphic method of
information from our database of the Ghost Ranch population of comparing a data set to a statistical distribution function (King, 1971). A
Coelophysis bauri. This study is intended to reflect the characteristics of probability plot shows the probability (x axis) that a random observed
this single population, so data from other Coelophysis occurrences (e.g., variable will be less than or equal to a given value (y axis). Probability
Spielmann et al., 2007) are not included in the analysis. calculations are simple, and the plots are easily made by hand, but here
Missing data are a problem in statistical morphometric analyses we used KaleidaGraph (1997) graphic analysis software. The probabil-
(Chapman et al., 2008), because even in a locality with extraordinary ity data are plotted against a scale that is related to a specific distribution
preservation, such as the Whitaker Quarry, very few specimens are function (e.g., normal [Gaussian], log normal, extreme value, etc.). Straight
complete. Generally, we employ a linear regression method to project line fits of the data on a specific probability scale indicate that the data
the missing values. In some few instances, pairwise deletion is used, but are distributed according to the function that is related to that scale.
this method reduces the size of the sample. Chapman et al. (2008) showed Departures from a straight line fit indicate that the data are skewed,
that reduced major axis regressions were slightly more accurate than truncated, multi-modal, or do not fit the distribution function. Each of
linear regressions as a solution to the missing data problem (a few per- these deviations produces a unique, characteristic shape on the plot that
cent improvement in some cases), but linear regressions performed very may be further explored to reveal information about the nature of the
well for most of the analyses presented here. We only employed scaling distribution and the process or circumstances that produced it ( Kock
in the case of a few caudal series lengths; in this case we projected the full and Link, 1970; King, 1971).
length of some incomplete caudal series by scaling based on the (only)
two complete caudal series available. Size/Age Classes and Growth in Coelophysis bauri

Probability Plotting: Extracting Femur length was chosen as the best metric to test for the pres-
Size/Age Classes from Metric Data ence of size/age classes because of the large sample size available (N = 70
in our database) and because our observations suggest that femora are, in
If a breeding season exists in a group of animals, each yearly general, among the sturdiest and least distorted bones in the quarry. Care
hatching or birthing (cohort) should constitute a statistical size class that was taken to insure that only one femur (the best preserved) from each
remains distinct from older and younger cohorts for as long as yearly individual was included in the analysis. If size/age classes were present in
growth surpasses individual variation within the cohorts (Brinkman et the data, they would be expressed as component modes that made up an
al., 2007), possibly throughout its lifetime. If the population is killed all overall multimodal distribution of sizes.
at once, a “snapshot” of its yearly age groups may be preserved in the A normal (Gaussian) probability plot of 70 Coelophysis bauri
fossil record. Thus, in catastrophic death assemblages as in the Ghost femur lengths has several “plateaus” that vary from approximately hori-
Ranch Coelophysis population, the age groups, if present, should be zontal to somewhat inclined (Fig. 97A). These indicate that the overall
readily apparent, whereas in attritional assemblages they will be less distribution of lengths comprises several component modes (King, 1971).
evident, but may still be resolvable. In animals that have indeterminate If a smooth curve were drawn through the data points it would show that
growth, the age groups may remain statistically separable for the lifespan the plateaus are connected by steep jogs that contain concave-up to
of the animal if enough members of the age group survive and if the concave-down inflections (proceeding left to right). These inflection
growth rate in the older animals is high enough. In animals that have points, which are marked by horizontal bars in Figure 97, indicate the
determinate (asymptotic) growth, however, the size groups representing approximate “boundaries” of the component distributions (Harding, 1949;
younger animals tend to “catch up” to those of the older animals that are Peck, 1987; Knoop and Owen, 1994). Theoretically, the distributions
approaching their growth asymptote. Ultimately, the size classes over- extend infinitely in either direction from the mean, but given a finite data
lap and become statistically inseparable near the asymptote because set, the inflection points provide a very good approximation of each
individual variation is larger than the yearly growth increments and sample component mode. Only the concave-up to concave-down inflections are
size is smaller among the older animals. In this case, only the asymptote used to show the component mode boundaries; concave-down to con-
itself may be resolved, although it may be possible to determine the age cave-up inflections (in the center of the plateaus) indicate the approxi-
at which the asymptote is approached. mate mean of the components. Note that there is a large jog between the
Other factors may create or accentuate the yearly size/age groups fourth and sixth component modes, which contains no data points. This
and produce statistically separable classes. These include, among others, suggests the existence of a size/age class in which no members are present.
seasonal abundance of food, seasonal weather patterns (e.g., tropical wet We separated the overall distribution (Fig. 97A) into its compo-
and dry seasons), estivation, and, after sexual maturity, the periodic nent modes at the inflection points and replotted each component (Fig.
alternation of energy resources between growth and reproduction 97B). The straight line curve-fits to the components indicate that they
(Andrews, 1982; Zug, 1993). Such size groups may be evident and have been correctly resolved and that they are distributed normally (D.S.
resolvable even in attritional assemblages (Brinkman et al., 2007). Peck, pers. comm.). The uppermost components that contain only two
 101
TABLE 6. The mean femur length, standard deviation, coefficient of
variation (CV), and number of individuals in each size/age class of the Ghost
Ranch population of Coelophysis bauri.

within the resolved cohorts is generally small, and decreases incremen-

tally with age from ~ 7% in the one year age class to probably < 1% in the
oldest groups.
Analysis of variance (ANOVA) tests were applied to determine if
there was a statistically significant difference between all of the resolved
size groups (Zar, 1999). Holm’s multiple comparison test was employed
because it is the most accurate test where: (1) the data are normally
distributed, (2) the groups are of unequal size, and (3) there are more than
three or four groups to compare (Glantz, 2005; Glantz and Slinker,
2001). All of the resolved component size groups show statistically
significant differences from each other at the 95% confidence level (P <
0.05).

Femur length growth curve

The age and femur length data from Table 6 are plotted to show
growth of this element in Coelophysis (Fig. 98). We used 20 mm as the
zero-age hatchling femur length based on Colbert’s (1989) estimate. This
estimate seems reasonable in light of the ~32 mm femur length of an
embryonic Oviraptor (Fig. 97A; Norell et al., 1994), a somewhat larger
FIGURE 97. A, Probability plot of Coelophysis bauri femur lengths (N = animal than Coelophysis, but still a small theropod. The reduced CV seen
70) showing horizontal bars at the inflection points. B, Overall distribution in the larger, older size/age classes may indicate that these groups are
of femur lengths has been separated into its component modes, which we approaching an asymptote, but the sample size is too small to inspire
interpret as size (= age) classes in the data. confidence in this observation.
or three members are, of course, very rough approximations. The mean In general, the growth curve shows rapid growth in the juveniles
value and standard deviation (Sigma) of each component may be read that is more bird-like than reptilian in its character based on growth of
directly off the plot. The mean value occurs at the 50% probability crocodylians (McIlhenny, 1935; Neill, 1971; Chabreck and Joanen, 1979)
point; +Sigma and -Sigma occur at the 84% and 16% probability points, and birds (e.g., Latimer, 1927). The curve flexes sharply toward the
respectively (King, 1971). For maximum accuracy, we used the coeffi- horizontal after the first year. Sexual maturity is usually reached above
cients of the linear curve fits (Fig. 97B) provided by the graphic analysis this “knee” in the growth curve (Zug, 1993). Visually, it is difficult to
software. Mean values, standard deviations, the population of each re- interpret whether the curve is approaching an asymptote, which would
solved size group, and the coefficient of variation (CV) are given for each indicate determinate growth or not (indeterminate growth). Padian (2004)
component mode (size/age class) (Table 6). We have thus resolved the examined the femur of Coelophysis (UCMP 129618) histologically and,
component modes (distributions) that make up the overall distribution although LAGs could not be resolved, he determined from the bone
of femur lengths, and for each of them we provide the three factors that structure that growth would have been rapid as seen in large modern
uniquely define a distribution: shape (here, the normal, or Gaussian bell birds such as the rhea and turkey vulture. This is much higher than the
curve), center (mean value in a symmetrical distribution), and width or rates he observed in the nearly contemporaneous Early Jurassic ornithis-
variation (standard deviation). The coefficient of variation (CV) pro- chian, Scutellosaurus. Chinsamy’s (1994) femur length growth data, based
vided (Table 6) is a dimensionless quantity that equals the standard on her histology study of Coelophysis rhodesiensis, are included in Fig-
deviation divided by the mean value and multiplied by 100 so as to ure 98 for comparison to the statistically-derived curve. Indeed, when
express it as a percent of the mean. the fact that C. rhodesiensis is a smaller animal than C. bauri is taken into
account, the curves are virtually identical.
CV = (std. dev./mean) * 100
This quantity is of greater utility than standard deviation when compar- Growth and allometry of large skeletal complexes
ing variation in animals of different sizes, because the size units have Here we used a database of measurements of the large skeletal
been removed and variation is expressed as a percent of overall size complexes (i.e., skull, cervical series, dorsal series, sacrum, caudal series,
(Simpson et al., 1960; Van Valen, 2005). The coefficient of variation and limb lengths) of 18 of the most complete specimens (Appendix 5) to
102
(Fig. 100) and the large skeletal element growth data are tabulated (Table
7). We chose age increments so that age essentially doubles with each
data point (the last data point representing all animals 7 years and older).
Growth of the Neck and Skull

Skull and neck morphs

Two skull and neck morphs of Coelophysis bauri were identified
by Colbert (1989, 1990) based largely on AMNH specimens 7223 and
7224, which represent the “gracile” and “robust” morphs, respectively
(Paul, 1988). Summarizing Colbert’s (1989) description of the two mor-
phs: the gracile morph (AMNH 7223) has a longer skull and neck,
shorter forelimbs, and fused sacral neural spines, whereas the robust
morph (AMNH 7224) has a shorter skull and neck, longer forelimbs,
and unfused sacral neural spines. The data set of 15 specimens compiled
for this study shows agreement with Colbert (1989), indicating that two
skull and neck length morphs are present, and quantifies growth in these
elements using dorsal vertebral series length as a metric of overall size
FIGURE 98. Growth in Coelophysis bauri based on femur length (upper
(Fig. 101). The uppermost data points in each of the four series in Figure
curve) with comparative data from Chinsamy’s (1994) histological growth
101 represent AMNH specimens 7223 (longer skull and neck) and 7224
study of C. rhodesiensis (lower curve).
(shorter skull and neck). The skull length regression lines are close to-
define growth for these body components. These specimens comprise gether and nearly parallel, but do they represent two real, distinct mor-
the most complete ontogenetic series that we were able to assemble. We phs or do the apparently separate lines result from random scatter in the
chose dorsal series length as the “independent” variable (x axis). We then data? The regression lines are linear and therefore lend themselves to
plotted the lengths of all the major skeletal component elements versus tests for statistically significant difference (Glantz, 2005). ANOVA lin-
dorsal series length (Fig. 99A-B). The various slopes and curves of the ear regression testing was applied to the two data sets representing the
regression lines show clearly that the components grow at different rates hypothetical long and short skull morphs. The tests showed that the
and that some growth trajectories accelerate and others decelerate during data are better described by two regressions than by one to a highly
ontogeny. Femur length is included in the plots as a method of assigning significant level (P < 0.01), demonstrating that the dimorphs are statisti-
age increments to the growth data. We note also that the long and short cally different with greater than 99% confidence. The ANOVA test sum-
neck and skull morphs (= robust and gracile morphs) proposed by Colbert mary is shown in Appendix 7.
(1989) are supported by our data (although his long and short arm mor- The hypothetical long and short neck data follow curved lines so
phs are not). We will define and discuss these morphs below. Here we they are unsuitable for linear regression testing. We attempted to linear-
use regressions that fit the average neck and skull lengths without respect ize the data points by LOG-transforming the data, thus allowing the
to their morphological type. application of the ANOVA tests, but the LOG-transformed data fit non-
The allometry of the large skeletal elements with respect to dorsal linear (polynomial) curves as well, indicating complex allometry
series length is shown in two plots of LOG –transformed data (Fig. 99C- (McKinney and McNamara, 1991), and disallowing the test. These two
D). The allometric growth constants are included in the plots as the curves, however, appear even more distinctly separate in the plot (Fig.
coefficient of x in the linear regression curve fits. Allometric growth 101) than the skull data. We therefore conclude that the apparent long
constants >1 indicate positive allometry in which the dependent variable and short neck morphs are also statistically significant groups.
grows at a higher rate than the independent variable. An allometric con-
stant = 1 indicates isometric growth whereas allometric constants < 1 Discussion
show negative allometry in which the dependent variable grows at a It is interesting that the long and short skull and neck morph curve
lower rate than the independent variable (Huxley, 1932; Gould, 1966). fit lines do not converge, even in the region of the smallest (youngest)
Allometry thus expresses shape change throughout ontogenetic growth. specimens. Therefore, if these morphs represent sexual dimorphs, as
Note that we fit a polynomial regression to the cervical series data, Colbert (1989, 1990) suggested, the dimorphism is already established at
because they have not been fully linearized by the LOG transformation. approximately one year of age. Given that the skull and neck samples are
The regression indicates complex allometry in which the growth con- odd numbered (nine each, but not the same nine specimens), the data
stant continually changes (in this case, decreases) throughout ontogeny points split, as nearly as possible, equally between the two morphs: four
(McKinney and McNamara, 1991). to five in each case for the necks and skulls. This even split is convincing
support for the idea of sexual dimorphism in skull and neck length.
Total length growth
Perhaps the most conventional and informative way to look at Skull growth and allometry
growth is with respect to total length. There are, however, very few Smith (1997) published a preliminary allometry study of
specimens that are complete enough to show total length, so generating Coelophysis skulls and later an analysis of facial variation in the species
such a curve is problematic. Our solution is to use the large skeletal (Smith and Merrill, 2006). Other workers are producing a detailed study
element growth data and to sum the lengths of the axial body compo- of the growth and allometry of the skull of Coelophysis bauri (R. Reisz,
nents, which can be read off of the plots (Fig. 99A-B) for each age class. pers. comm.), so we offer only a cursory look at growth of two gross
In Figure 99A-B we extended a vertical line from the femur length regres- features: skull height, and orbit diameter.
sion line through all the other regression lines at the mean femur length From a database of 25 skulls, we plot skull height at the prefrontal
that represents each age class (Table 7). Thus, all the regression lines are and at the quadrate versus skull length (Fig. 102A). It is obvious from the
“tied together” at these temporal landmarks and we can readily obtain plot that scatter in the data is high and that growth in skull height pro-
the lengths of each large skeletal element at each age class from the plots. ceeds at a much reduced rate compared to length. The high degree of
Total length growth data, obtained by the described method, are plotted scatter in the data (relatively low R2 statistic) may result in part from
 103

FIGURE 99. Metrics of the growth of Coelophysis bauri relative to length of the dorsal vertebral series. A, Growth of the large body elements of
Coelophysis bauri based on dorsal series length with yearly size increments marked. B, Growth of the caudal series of Coelophysis bauri with yearly
increments marked. C, Allometry of the large body elements of Coelophysis bauri with respect to dorsal series length. D, Allometry of the caudal series
of Coelophysis bauri.

taphonomic processes, as the skull bones are extremely thin, very fragile Orbit growth and allometry
and poorly sutured. The skulls are, therefore, more easily sheared, flat- We show growth of the antero-posterior (A-P) orbit diameter of
tened, or otherwise distorted than most of the other bones. Coelophysis bauri relative to skull length in a bivariate plot based on 23
An allometry plot of the log-transformed skull height data (Fig. specimens with skull length varying over a range of 88 to 277 mm (Fig.
102B) shows strong negative allometry in skull height with allometric 102C). We chose A-P diameter as our preferred metric here, because it
growth constants of 0.51 and 0.57 at the prefrontal and quadrate, respec- more accurately reflects the size of the eyeball than does the dorso-
tively. We find that the portion of the skull between the prefrontals and ventral diameter. This fact became apparent during dissections and study
the posteriormost part of the skull grows in strong negative allometry. of bird and lizard eyes, and was discussed in the earlier section on Vision
The skull is therefore proportionately very tall in juveniles, and its in Coelophysis. With respect to orbit growth, we see that while the skull
height grows relatively little throughout ontogeny. length increases by a factor of >3 in our sample, the orbit diameter
increases by a factor of 1.6.
104
In most vertebrates the size of the orbit increases at a much re- TABLE 7. Summary of total lengths, body proportions, mass and allometric
duced rate relative to the rest of the body throughout ontogeny (i.e., in constants for the 1, 2, 4 and 7+ years size classes of Coelophysis bauri.
negative allometry) (Walls, 1942). We show an allometry plot of
Coelophysis bauri orbit diameter relative to skull length that exemplifies
this character (Fig. 102D). The allometry of the A-P orbit diameter is
extremely negative, with an allometric growth constant of 0.38, even
more negative than the allometry seen in the height of the posterior
portion of the skull, which contains it.
Growth and Allometry of the Sacrum and Pelvis
The sacrum lengths in our sample of 37 individuals range from 70
mm to 148 mm. We have shown growth and allometry of the sacrum
relative to other major body segments in Figure 99A. Here, we show
growth of the pelvic elements relative to sacrum length (Fig. 103A).
Femur length is included in the plot as a metric of overall size that relates
easily to the age groups. A plot of LOG-transformed data (Fig. 103B)
shows the allometric relationship of the pelvic elements to sacrum length.
With respect to sacrum length, the lengths of the pubis, ilium, and is-
chium grow with allometric constants of 0.78, 0.85 and 0.93, respec-
tively. Thus, the pelvic elements grow in moderate to slight negative
allometry. Scatter in the data is relatively high, and reduces confidence in
the constants. The scatter could be a combination of taphonomic distor-
tion, individual variation, and sexually dimorphic characters, but we are
unable to resolve two distinct groups that appear to be dimorphs in these
pelvic element data.
The sacrum of Coelophysis bauri comprises five vertebrae. Fu- pendent of size and that exactly 50% of the population falls into each
sion in various parts of the sacrum is variable, and Colbert (1989, 1990) class. We conclude that this is very strong evidence that the state of s4-
considered fusion of the sacral neural spines, or lack of it, to be a prob- s5 fusion is a sexually dimorphic character. It is interesting that this
able sexual dimorphic character in the species. He associated the fused dimorphic character is present in even the smallest individuals, as was
neural spine condition with his “robust” morph (AMNH 7224) and the evident in the skull/neck length dimorphism. Only a single specimen
unfused condition with the “gracile” morph (AMNH 7223). During the (AMNH 7228) in which the s4-s5 fusion state was observable also fell
course of preparation and subsequent measurement of the NMMNH into one of the long or short skull and neck divisions. In this case, the long
specimens, we noticed that in some specimens sacral vertebrae s4 and s5 skull and neck (gracile morph) were associated with the unfused state of
were solidly fused and, in others, they were not. We recorded this state s4-s5.
of fusion during data collection at other museums and eventually amassed
a sample of 19 individuals of varying sizes in which the state of s4-s5 Morphology and growth of the vent area
fusion was observable. The sample showed nine specimens in the unfused A single specimen (NMMNH P-44802) preserves the vent area in
condition and 10 fused, as near an even split as is possible in the odd- apparently undistorted condition. This small juvenile is exposed in ven-
numbered sample. tral aspect so that the vent fenestra, enclosed by the sacrum, left and
We show binary plots that relate the state of s4-s5 fusion to right ilia, and left and right ischia, is exposed and measurable (Fig. 104).
sacrum length and to femur length as measures of individual size (Fig. Measurements of the internal boundaries of the vent fenestra show that
103C-D). The plots show that the state of fusion is completely inde- the most restrictive dimension is between the ischial peduncles. There

FIGURE 100. Growth of Coelophysis bauri; total length as a function of FIGURE 101. Bivariate plot showing the growth of the long and short skull
age. and neck morphs of Coelophysis bauri based on dorsal series length.
 105

FIGURE 102. Metrics describing the growth of the skull in Coelophysis bauri, A, Skull height as a function of skull length in Coelophysis bauri. B, The
allometric relation of skull height and length. C, Orbit antero-posterior diameter versus skull length. D, A plot of orbit diameter allometry showing extreme
negative allometry.

are only three specimens in our study database in which preservation femur, to distinguish various morphs (Colbert, 1989, 1990; Raath, 1977,
and exposure allowed measurement between the ischial peduncles. This 1990; Gay, 2005) as well as taxonomic details in coelophysids, espe-
very small sample size is unfortunate and calls into question the accu- cially cladistic analyses (Rowe and Gauthier, 1990; Tykoski and Rowe,
racy of the data, but having recognized these limitations, we show the 2004) and theropod functional morphology (Hutchinson, 2001a, b).
minimum inside diameter of the vent fenestra (between the ischial pe- Because of its paleobiological implications related to the mass of the
duncles) versus femur length as a measure of overall size (Fig. 103E). We animal, we treat the femoral midshaft separately.
project the regression line to the maximum femur length seen in the
largest individuals (240 mm) and observe that the corresponding mini- Forelimbs
mum vent fenestra diameter is ~ 60 mm (Fig. 103E). Whether sexual Growth in the various forelimb and manus elements versus the
dimorphism is present in the vent area is, of course, undetectable using dorsal series length is shown together with femur length to easily estab-
these data because of sample size. An allometry plot of the same data lish relationship to age groups (Fig. 105A). Growth in the forelimb is
shows that allometry is significantly positive in the vent diameter, with nonlinear and accelerates with age as documented by the strongly con-
an allometric growth constant of ~1.8 with respect to femur length (Fig. cave-up curve fit to the data. No sexual dimorphism is evident in any of
103F). the elements as single regression lines provide good fits to the data with
Growth and Allometry in the Limbs low scatter. The allometric growth constants for the forelimb and manus,
the humerus, radius, and hand, measured at the third (longest) digit, are
In order to quantify the limb proportions, we provide plots of 1.18, 1.2, 1.05 and 1.3, respectively. Moderate positive allometry is
overall forelimb and hindlimb lengths relative to the dorsal vertebral therefore evident in the forelimb and manus and except for the radius,
series in Figures 105-106. We address the hindlimb in much greater detail which appears to grow in near isometry (Fig. 105B). The forelimb and
because of the long history of researchers’ focus on it, especially the manus are thus relatively short in juveniles, and grow relatively more
than the dorsal series length.
106

FIGURE 103. Metrics describing pelvic growth and morphology in Coelophysis bauri. A, Comparative growth of the sacrum and pelvis based on sacrum
length. B, Allometry of the pelvic elements based on sacrum length. C, A binary plot of fusion between sacral vertebrae s4-s5 based using sacrum length
as a measure of overall size. D, A binary plot of fusion between sacral vertebrae s4-s5 based using femur length as a measure of overall size. E, A plot of the
inside diameter of the vent foramen measured at the ischial peduncles with the curve fit line extended to the length of the largest femora (240 mm). F, An
allometry plot of vent foramen diameter showing positive allometry.
 107
Hindlimbs
Growth of the hindlimb and pes elements with respect to the
dorsal series length is shown (Fig. 106A). Negative allometry is evident
in all elements of the hindlimb (Fig. 106B). The hindlimbs of juveniles are
relatively long and grow at a slower rate than the dorsal series. There is
no evidence of sexual dimorphism in the relative lengths of the femur and
tibia of Coelophysis. A plot of femur length versus tibia length after Gay
(2005) shows two distinct groups, which he assessed to represent sexual
dimorphs (Fig. 107A). When the sample size was increased (this study)
the two groups blended into a single continuous distribution (Fig. 107B)
and dimorphism was no longer evident. The allometric relationship of
tibia length and femur length is weakly negative (Fig. 107C). With an
allometric growth constant of 0.95, the tibia grows very slightly less
aggressively than the femur.

Proximal femur details

Femora are often well preserved and specifics of their morphol-
ogy have been used to diagnose taxa (e.g., Sullivan and Lucas, 1996;
Tykoski and Rowe, 2004; Irmis et al., 2007; Nesbitt et al., 2009) and
demonstrate sexual dimorphism (Raath, 1977, 1990) in various
dinosauromorphs and primitive theropods. Additionally, they are ex-
ceedingly important in soft-tissue reconstructions (Hutchinson, 2001a,
b).
In Coelophysis bauri the lesser (or anterior) trochanter (for the
iliofemoralis muscle) comprises a blocky, nearly rectangular, basal struc-
ture with a smaller pointed process extending proximally from its dorsal
surface. We show growth of some proximal femur features with respect
to femur length (Fig. 108A). In this figure, “lesser trochanter width”
refers to the width of the blocky, basal structure and “lesser trochanter
point width” refers to the basal width of the proximal pointed process of
the blocky structure. The width of the femur head (e.g., the mediolateral
width) data show an interesting pattern (Fig. 108A). As was seen in the
skull length data (Fig. 102), there appear to be two distinct lines of data
points. We separated and plotted these apparently distinct lines (Fig.
108B) and applied ANOVA testing to show that they are statistically
significantly separate distributions (P<0.001). The ANOVA test details
are given in Appendix 8. The 17 data points separated into groups of
eight and nine between the two regression lines, indicating that they
probably represent a sexually dimorphic character.
The femur head height, measured from the notch in the ventral
side of the offset femur head to its proximal surface, and the lesser
trochanter width grow in near isometry with allometric constants of 0.98
and 1.09, respectively. Femur head width and the lesser trochanter point
width grow in positive allometry, with constants of 1.26 and 1.28, re-
spectively.
Finally, we recreate the plot of femur head width as a function of
lesser trochanter width identical to that of Raath (1977, 1990), in which
he demonstrated sexual dimorphism in Coelophysis rhodesiensis (Fig.
108D). He diagnosed the sexual dimorphism on the basis of slender
(gracile) versus blocky (robust) shapes of the lesser trochanters. In the
C. bauri data, we see no separation in the distribution of the lesser
trochanter measurements that would indicate the type of dimorphism
described in C. rhodesiensis.

Femoral midshaft
An allometric plot of the proximal femur features (Fig. 108C) also
FIGURE 104. A, Vent area of NMMNH P-44802. The specimen is inverted includes average midshaft diameter data. Midshaft diameter of the femur
and viewed in postero-ventral aspect; scale is in cm. ACT, acetabulum; is an important metric in that it shows whether the limb bone strength
CD1, first caudal vertebra; FM, femur; IS, ischium; ISP, ischial peduncle of (determined by cross-sectional area) increases so that constant stress is
the ilium; PB, pubis; PVA, postero-ventral arch of the ilium; S4, S5, sacral maintained in the bone as the animal grows. The strength of a bone is
vertebrae; ?, unknown bone fragment in the vent fenestra – probably part proportional to its cross sectional area, and area is proportional to the
of the (broken) right PVA. B, Drawing of the vent area of P-44802. square of a linear dimension, L, which can be loosely correlated as, A =
L2. However, volume, which is proportional to the weight (stress) ap-
108

FIGURE 105. Growth in the forelimb elements of Coelophysis bauri as a function of dorsal series length. A, Forelimb growth plotted relative to dorsal series
length. B, Allometry plot showing positive allometry of the forelimb elements.

plied to the bone, is directly related to length cubed as V = L3. Based on Conversely, they possessed relatively short and tall skulls, short torsos
these relationships, McGowan (1999) has shown that the diameter of (dorsal and sacral series) and short forelimbs.
the bone must increase as the square root of the volume; that is,
Age/Mass Distribution
SQRT L3 = L3/2 = L1.5
We plot the number of individuals in each resolved age class (Table
for the stress level to remain constant throughout growth. Thus, the 6) as a histogram to show the age distribution of the Ghost Ranch
allometric constant of the midshaft diameter must equal at least 1.5 for Coelophysis population (Fig. 110). The age distribution shows a nearly
the strength of the limbs to keep up with the increasing weight of the “hyperbolic” character in which attrition is very high in the juveniles.
animal. Yet, if the survivors reach about three years of age, expectation of yearly
McMahon (1975) and Alexander (1977) demonstrated that the survival and thus a relatively long life is greatly improved. The power
limb bone diameters of the cursorial ungulates (e.g., antelopes) generally curve fit to the data indicates that the population dies off at a rate
maintain an allometric growth factor near 1.5. Furthermore, they showed approximately equal to the age in years to the -1.6 power (x -1.6).
that the limb bones of most animals do not exhibit this relationship. In a We applied the empirical formula of Christiansen and Farina
wide variety of animals (shrews to elephants) the limb bone diameters (2004) to calculate the mass of the 70 specimens for which we have
grow approximately isometrically with respect to length. These animals femur length data. They relate mass in kg to femur length in mm by the
accommodate the increasing stress on their limb bones by modifying following equation:
their behavior as they grow. Larger animals tend to keep more feet on the
ground during locomotion (reduction of duty cycle) and to hold the limbs LOG10 mass = -6.288 + 3.222 * LOG10 femur length
straighter so they are more column-like (graviportal stance) thus reduc- We show the resulting mass versus femur length data in kg and
ing shear stress (Biewener, 1983). The midshaft diameter allometric pounds (Fig. 111A), and the mass distribution of the population as a
growth constant of Coelophysis’ femur is ~1.5 (Fig. 108C). We conclude histogram (Fig. 111B). A plot of mass as percent of population (Fig.
that this is strong additional evidence that Coelophysis was cursorial. 111C) demonstrates that, whereas, the one-year-old size class accounts
for 40% of the population, the largest adults account for only about 2%
Growth and Allometry of the Entire Body
of the population. There are about 20 “babies” for each of the largest
We describe growth in the large skeletal elements of Coelophysis adults in the Whitaker Quarry population. This observation lends sup-
based on total length by plotting the length of each element at each age as port to Paul’s (1994) observations that “Dinosaur populations may
a function of total length (the sum of all the axial skeleton segments) (Fig. have consisted mostly of posthatchling juveniles that were not depen-
109A). In this plot, the four “columns” of data points represent ages 1, dent on their parents…” and that dinosaurs were probably r-strategists.
2, 4, and 7+ years. Whole body allometry, based on total length, is shown Finally, we show mass as a function of age (Fig. 111D); the curve-fit
in a natural log-transformed data plot (Fig. 109B). The allometric growth exhibits the typical sigmoid expression of the logistic equation that de-
constants are seen as the x value coefficients of the curve fit equations scribes growth in terms of mass as a function of age (Ricklefs, 1983).
and are tabulated in Table 7. We show mass as a function of length where a power curve with
Negative allometry is seen in the hindlimbs, cervical series and an exponent of 3.33 provides an excellent fit to the data (Fig. 111E).
caudal series, and positive allometry is seen in the skull, dorsal series, Finally, in a plot of natural log-transformed data we show mass allom-
sacrum and forelimbs. Thus we reconstruct yearling Coelophysis speci- etry (Fig. 111F). The allometric constant, 3.33, is shown as the x value
mens as having relatively long necks, tails, and legs as well as large feet. coefficient (and as the exponent of x in Fig. 111E), and is tabulated in
 109

FIGURE 106. Growth of the hindlimb in Coelophysis bauri. A, Growth in

the hindlimb and pedal elements of Coelophysis bauri as a function of dorsal
series length. B, Allometry plot of hindlimb and pedal elements showing
negative allometry.

Table 7 together with mass at each age increment. As mass is propor-

tional to length cubed (L3) and the allometric constant is the exponent of
a power curve that fits the data, an allometric constant of 3 would
indicate isometric mass increase (Gould, 1966). The allometric constant
of 3.33 indicates moderate positive allometry in growth in terms of mass
and therefore the adults would have a bulkier build (i.e., not just larger,
but more robust) than the juveniles.
Growth and Proportion Summary
Combining size-class and proportional and dimorphic growth data FIGURE 107. Growth of the femur in Coelophysis bauri relative to other
allows for a reconstruction of Coelophysis bauri at various growth stages. elements. A, Femur length versus tibia length in Coelophysis bauri after Gay
The growth stages highlighted are 1-year (Fig. 112A), 2-year (Fig. 112B), (2005) showing apparent sexual dimorphism. B, Femur length versus tibia
4-year (Fig. 113) and 7-year (Fig. 114) for both long- and short-necked length with an increased sample size (this study) showing no dimorphism.
morphs. C, Allometry plot showing very slight negative allometry in tibia growth
with respect to the femur.
Survivorship
Survivorship curves show, at any given age or size (x-axis), the
110

FIGURE 108. Details of growth in the femur of Coelophysis bauri. A, Plot of details of growth in the proximal femur of Coelophysis bauri. B, A plot of
femur length versus proximal femur width shows dimorphic qualities. C, Allometry of the proximal femur features plus midshaft diameter based on femur
length. D, A plot of femur head width versus lesser trochanter width shows no dimorphic characters.
 111

FIGURE 110. The age distribution of the Ghost Ranch population of

Coelophysis bauri with a curve fit showing exponential decrease with age.

FIGURE 109. Growth of the large body elements of Coelophysis bauri. A,

Growth of large body elements based on total length. B, Allometry plot of
the large body elements showing positive allometry in the dorsal series,
sacrum, forelimb and skull, and negative allometry in the caudal series,
cervical series, and hindlimb.
112

FIGURE 111. Estimates of body mass, mass distribution and mass allometry of Coelophysis bauri. A, Mass versus femur length of 70 specimens of
Coelophysis bauri based on Christiansen and Farina (2004). B, A histogram showing the mass distribution of Coelophysis bauri. C, A plot of mass
distribution shown as percent of population shows that there are ~20 babies for each of the largest adults. D, A plot of mass as a function of age shows the
logistic character of the relationship. E, A bivariate plot showing the approximately cubic relationships of mass to body length. F, An allometry plot of body
mass in C. bauri. The allometric constant of 3.3 (the coefficient of x in the curve fit equation) shows positive allometry. An allometric constant of 3.0
would indicate an isometric relationship because mass varies as the cube of body length.
 113

FIGURE 112. Skeletal reconstructions of two Coelophysis bauri morphs at the hypothesized A, “1-year” and B, “2-year” growth stages. Some elements
modified from figures in Colbert (1989) and Carpenter (2002).
114

FIGURE 113. Skeletal reconstructions of two Coelophysis bauri morphs at the hypothesized “4-year” growth stage. Some elements modified from figures
in Colbert (1989) and Carpenter (2002).
 115

FIGURE 114. A, Skeletal reconstructions of two Coelophysis bauri morphs at the hypothesized “7-year” growth stage. Some elements modified from
figures in Colbert (1989) and Carpenter (2002). B, Averages of the gracile and robust morphs for one, two, four and seven year stages scaled to equal lengths
to demonstrate allometric relationships.
116

FIGURE 115. Estimates of survivorship in the Ghost Ranch population of Coelophysis bauri. A, Survivorship in Coelophysis bauri based on femur length.
B, A survivorship curve with age increments noted. C, Survivorship as a function of age in Coelophysis bauri. D, A semi-log plot of survivors versus age
shows the concave up curve characteristic of Type III survivorship (Deevey, 1947).
surviving fraction of an arbitrary cohort (y-axis) (Pearl and Miner, 1935; Indeed, we feel that this sample is one of the best population samples
Deevey Jr., 1947). Here survivorship for the Ghost Ranch population of available in the dinosaurian literature. As we see individuals of varying
Coelophysis bauri is based on femur length (Fig. 115A). The tabulated ages throughout many of the blocks we do not think that the depositional
survivorship data are shown in Appendix 6. The relationship of femur disaster preferentially included juveniles (or, conversely, allowed adults
length to age is nonlinear (Fig. 107) (the relationship approximates a to escape). Furthermore, although it could be argued that complete juve-
power curve), so we add age increments to the same data (Fig. 115B) to niles are more likely to be found in a single block than are larger, complete
make the age relationship clearer, and then plot survivorship as a func- adults, our techniques rely on measuring all preserved major elements of
tion of age (Fig. 115C). These curves match well the age-mass curve, all specimens, so that incomplete skeletons are just as well, if not better,
although this is due at least in part to their dependence on a single represented. Thus, we hypothesize that the age group distribution we
element, the femur. see in the population reflects the true population structure, not any sort
When the logarithms of the number of survivors is plotted against of taphonomic bias.
age, the rate of mortality with respect to age is shown (Deevey Jr., 1947) As was noted previously, the implication of the Type III survi-
(Fig. 115D). A straight line fit to the data would indicate that death vorship curve is that mortality is very high in the juveniles, but if a few
occurs at a constant rate throughout the animal’s lifetime. The concave- years of age are attained, the expectation of a long life improves greatly.
up curve that fits the Coelophysis data shows a continuously decreasing Therefore, in the life of Coelophysis it was good to be big, yet another
rate (decreasing slope of the line) of mortality and approximates the interesting implication of the survivorship data may be seen in the data
Type III curve of Pearl and Miner (1935) and Deevey (1947) in quality points representing the smallest individuals (Fig. 115A-B). The first five
but not in the severity of its hyperbolic nature. or six data points proceed in a concave-down fashion, indicating a very
low rate of mortality for these smallest individuals. There is then an
Discussion inflection point in the curve and mortality increases very rapidly. We
We feel that the ~70 individuals we use in these synthetic studies hypothesize that these smallest individuals may have been benefiting
are a random, and therefore a representative sample of the population. from parental care, which is well documented in several studies (Carpen-
We assume that individuals are randomly distributed throughout the ter et al., 1994; Horner, 2000; Chiappe and Dingus, 2001). The sharp
quarry, the blocks examined represent an essentially random sample of negative flexure at approximately 120 mm femur length could represent
the quarried area, and that the quarry itself appears to represent a single, a size threshold below which the young are protected. Perhaps in the life
“catastrophic” event that caught all members of the Coelophysis flock. of Coelophysis there were also benefits in being very small.
 117
PALEOBIOLOGY OF COELOPHYSIS BAURI blage of Coelophysis bauri is that they gathered for purposes related to
Paleoenvironment of the Whitaker Quarry Site reproduction. There are no hatchlings in the quarry and no eggs are
known from either inside or outside the body cavities, and therefore, if
The Ghost Ranch population of Coelophysis lived in the equato- the reproduction hypothesis is correct, it would probably have been
rial tropics of Late Triassic New Mexico at about 5o north latitude (Fig. early in the reproductive cycle. In any case, some circumstantial evi-
116A). Their mass grave is located in the uppermost sediments depos- dence supports this premise. We note that the smallest size class in our
ited in the Chinle basin by the vast Chinle river system. The Chinle river growth analysis averages more than 1.5 m in length (Table 7) and is
system existed in a large, subsiding, back-arc basin with highlands prima- therefore much larger than the largest hypothetical hatchling size (dis-
rily to the south and southeast that shed sediments into the northwest- cussed and documented below) for C. bauri. It appears likely that these
draining basin (Lupe and Silberling, 1985; Riggs et al., 1996; Dickinson smallest animals were approximately one year old and that a new breed-
and Gehrels, 2008) (Fig. 116B). ing season was imminent. We also point out the sharp downturn in the
The microstratigraphy and fauna of block C-8-82 has revealed survivorship curve at ~ 120 mm femur length, suggests that only the very
much about the paleoenvironment of the Whitaker Quarry site. The large smallest individuals were receiving parental protection or care (Fig. 115B).
group of Coelophysis and other animals were probably killed, and were An end to parental care would seem to be necessary before a new breed-
certainly buried, by a flood. Just below the overbank flood deposit that ing and “child rearing” event could occur. Some type of ecological sepa-
contains the Coelophysis bonebed, there is a muddy bed that is rich in ration of the babies from the larger juveniles and adults could also explain
ostracodes and conchostracans. These factors indicate that the quarry the population structure observed in the Whitaker Quarry, but there is
site was a generally small, probably ephemeral pond or oxbow lake lying no supporting evidence for this.
in a topographic low on the floodplain of a river. The aquatic fauna of the Paul (1988) did not believe that the large number of individuals in
Whitaker Quarry, including the small redfieldiid Synoricthys, the large the Whitaker Quarry could represent a single group. However, we see no
coelacanth Chinlea and phytosaur material were found in the flood de- evidence that the quarry represents anything other than a single occur-
posit rather than in the underlying pond deposit. There is no indication rence of a mass killing. That is, there is no evidence to support the
that the pond was large enough or permanent enough to support a popu- hypothesis of an attritional assemblage. The microstratigraphy of the
lation of fish, and we assume that the fish, and possibly the phytosaur, quarry indicates that the deposit was laid down by an overbank flood.
were swept out of the river and deposited together with the terrestrial No transport, scavenging, or weathering of the fossil material is evident,
animals that were caught in the flood. and the articulated skeletons cut across all bedding and geological unit
Phytosaur remains are incorporated in the main tetrapod bonebed boundaries of the flood deposit. Also, there is no differential preserva-
and are well represented in the Whitaker Quarry. The type skull of tion of the fossils as would be expected in an attritional assemblage. We
Redondasaurus bermani was located in a CMNH Whitaker Quarry block thus conclude that a single very large group of Coelophysis, along with
(Hunt and Lucas, 1993b). AMNH block I contains an incomplete articu- some other animals was overcome, killed, and buried in a single flood
lated phytosaur tail, hip, and ribs. A large phytosaur vertebra is located event. We believe that the isolated elements, which all occur in the upper-
among the Coelophysis skeletons in AMNH Block IX, and a small, most strata, represent reworking.
nearly complete juvenile phytosaur skull (NMMNH P- 44920) was Coelophysis’ reproductive strategy was probably inclined in the
found in the NMMNH block C-8-82. Postosuchus material is known “r” direction, producing large numbers of offspring that may have re-
from the CMNH block C-4-81. Both the phytosaurs and the Postosuchus ceived some degree of parental care up to a size threshold above which
were much larger carnivorous reptiles than Coelophysis. They lived in their numbers declined very rapidly. The vast majority of the population
the same general environment, and likely preyed on the latter. comprised babies and juveniles with only a few surviving to achieve large
size. Coelophysis grew very quickly in the first year and may have
Lifestyle Implications of the Paleobiological Data approached a growth asymptote of ~3 m length in seven or eight years.
In brief overview, we find that Coelophysis bauri was a visually- The young had a suite of typical juvenile tetrapod characters, including
oriented predator, probably possessing acute, diurnally optimized, ste- short snouts, large eyes, and large feet. Sexual maturity probably oc-
reoscopic vision, and was an obligate biped capable of fast, agile running curred at two or three years of age.
in pursuit of prey or to escape from predators. The teeth of Coelophysis Sexual dimorphism was present in Coelophysis bauri. The long
show that it was carnivorous, and some preserved gut contents indicate skull/long neck morph termed “gracile” and the short skull/short neck
that it was probably cannibalistic. One specimen contains gut contents morph termed “robust” represent the two sexes (Paul, 1988). We find
that were probably of a piscine origin, but identification is difficult. that the longer, “gracile” morph and the shorter, “robust” morph prob-
Upon viewing the various exhibited Coelophysis blocks, people ably represented the females and males, respectively. The degree of
often ask, “Why were there so many together?” The species was obvi- sexual dimorphism in terms of size was relatively small, and this has
ously gregarious, assembling in large numbers (hundreds at least) for some interesting implications that will be discussed below.
some purpose, which may have included a seasonal food source or repro- In the following sections, we expand on some of these lifestyle
ductive activity, as these seem to be the most probable stimuli to attract and reproduction issues, and offer substantiating data from our recent
such a large gathering of predators. It is noteworthy that both Coelophysis work and the work of others.
bauri and C. rhodesiensis are known from mass death assemblages (e.g.,
Raath, 1977, 1990). The tentative identification of fish remains in the gut Cursorality
contents of NMMNH P-42352 led us to speculate that the Ghost Ranch Cursoriality in mammals and dinosaurs exists as one end member
Coelophysis may have gathered to exploit a “fish run,” such as the salmon of a locomotor continuum ranging from cursorial to graviportal (Carrano,
spawning events that attract large numbers of bears to Alaskan streams. 1999). Small bipedal ornithischians and small theropod dinosaurs were
We attempted to determine if the redfieldid or coelacanth fish found in generally cursorial and were among the fastest running dinosaurs (Coombs,
the deposit were possibly anadromous or catadromous and therefore 1978; Alexander, 1989). However, even the most cursorial dinosaurs did
may have been staging a spawning event that would have drawn the not achieve the same level of specialized cursorial adaptation as some
dinosaurs to the river. However, after some research, and discussion with mammals (e.g., giraffids) (Carrano, 1999). Although cursorial animals, by
biologists, no skeletal features were identified that would be preserved as definition, are adapted for running, theropod trackways show that these
an indication of such behavior in the fish. animals generally walked and seldom ran (Farlow et al., 2000). Running
In our view, a more likely hypothesis to explain the large assem- is an exhausting behavior and presumably was reserved for prey pursuit,
118

FIGURE 116. Paleogeographic maps of Late Triassic time. A, Mollewide globe showing the Pangean supercontinent. Asterisk marks the approximate
position of north-central New Mexico. B, Southwestern United States during the Late Triassic showing the Chinle river system draining to the northwest.
Asterisk marks the approximate location of the Whitaker Quarry. Both maps used with permission of Dr. Ronald Blakey, Northern Arizona University,
Flagstaff, AZ.
 119
escape from predation, and other situations of alarm or necessity. though adults of C. bauri were likely proportionately more robust than
Colbert (1989, 1990) considered Coelophysis bauri to be adapted juveniles (see Growth section), the femur was able to handle the in-
for running, and Carrano (1999), using principal component analysis, creased stresses associated with this greater mass. Throughout growth
categorized the Coelophysoidea as cursorial. Indeed, one cannot examine the strength of the femur increases as a power function with an exponent
the hindlimb of Coelophysis with its short femur, long tibia and pes, of ~ 1.5, producing a concave-up power curve that approximates the
reduced number of functional toes, and its strong, simple ankle that has increasing stress applied to the bone by the increasing mass of the ani-
clearly sacrificed flexibility for strength without thinking that this animal mal. Among most mammals (shrews to elephants), the femur diameter
must have been a runner. has an allometric constant of 1 to 1.1, although a constant of 1.5 is
Coombs (1978) enumerated the skeletal characters that define required for bone strength to keep pace with greater stress levels brought
cursoriality. We show that Coelophysis bauri exhibits all these morpho- on by increased size (McGowan, 1999). Most mammals must modify
logical indicators (Table 8). Only in the single instance of “tibia-fibula their behavior as they grow to prevent overstressing their limb bones.
reduced to single elements” is the expression partial. We used the diam- Today, only the cursorial ungulates (e.g., antelopes) have a femur diam-
eters of C. bauri tibia and fibula to calculate the relative cross-sectional eter allometric constant of ~1.5 (Scott, 1985). Therefore, we consider
area of these elements and thus develop a measure of strength of the bone Coelophysis’ femur diameter allometric constant of ~ 1.5 to be additional
(McGowan, 1999). We found that the tibia has ~ 29 times greater cross- strong evidence of cursoriality in the species.
sectional area than the fibula and presumably supports that much more
weight (~ 96% of the weight applied to the limb). Therefore, even though Diet and Cannibalism
the tibia and fibula are not coalesced into a single element, as in ungulate
mammals, they show considerable tendency in the direction of acting as Carnivory
a single element. This is even more apparent when the fusion of the tibia- The laterally compressed, recurved, finely-serrated teeth of
fibula to the astragalocalcaneum is considered, as this simplifies the ankle Coelophysis indicate a carnivorous lifestyle (Fig. 117). Both the anterior
joint to a single proximal element (the astragalocalcaneum) articulating and posterior edges of the maxillary teeth are serrated, with the anterior
with the proximal tarsals. edge showing slightly smaller, denser denticles. Farlow et al. (1991)
We found additional evidence for cursoriality in the femur diam- noted that Coelophysis’ denticle count (8 to 12 denticles/mm on the
eter allometry of Coelophysis (Fig. 108C). In the present allometry study, anterior edge and 8 to 10/mm on the posterior edge) is among the highest
the femur diameter allometric growth constant is ~ 1.5. This constant of carnivorous dinosaurs. Our counts (~8 to 11/mm anterior, 8 to 9/mm
indicates moderately strong positive allometry in the femur diameter, posterior) agree with Farlow et al. (1991) and appear to be independent
and thus shows increasing strength, which in turn suggests that even of tooth size.

TABLE 8. Cursorial characters as defined by Coombs (1978) and their Coprolites and cololites
expression in Coelophysis bauri. At least three specimens of Coelophysis bauri from block C-8-82
have fossil fecal material directly associated with articulated skeletal
material. These specimens apparently include material retained within
the posterior intestine (cololites), as well as evacuated material (coproli-
tes) (Hunt, 1992; Hunt et al., 1998). The cololites occur between the
ischia and the proximal caudal vertebrae, and the “true” coprolites are
posteroventral to this area. Most of the coprolite material is formless
and was apparently somewhat mixed with still-wet mud and silt at or
near the time of death.
The skeleton of NMMNH P-44801 has a small amount of associ-
ated coprolite material that contains very small, sparse, unidentifiable
bone fragments. This coprolite material was removed from the block to
facilitate preparation of underlying bones.
The cololite and/or coprolite material associated with NMMNH
P-42352 is an ovoid mass located between the proximal caudals and the
ischia (Fig. 118A). It consists of small, thin, bone fragments in a densely-

FIGURE 117. The recurved, laterally compressed, serrated teeth of

Coelophysis bauri (NMMNH P-50529).
120

FIGURE 118. A coprolite/cololite associated with NMMNH P-42352. A, In overview, the coprolite/cololite lying between the proximal caudals and ischia
(cm scale). B, Close-up of the coprolite/cololite showing numerous small, bony, rod-like inclusions (cm scale). C, Photomicrograph of the rod-like
inclusions (fish scale material?) (mm scale). D, Photomicrograph of the surface of a fish scale from the Whitaker Quarry showing wavy, microscopic, rod-
like structure that may be the source of the rod-like structures in the coprolite.

packed matrix of small (~1 mm long, by ~0.1 mm diameter), rod-shaped concluded that the “cannibalism evidence” was a taphonomic artifact and
material (Fig. 118B-C). These very small rod-like structures may be the not behaviorly-based.
remains of digested fish scales. A fish scale with a rod-like surface texture In 2005, new evidence of cannibalism was discovered in the
from block C-8-82 is figured for comparison (Fig. 118D). The coprolitic NMMNH Coelophysis block (Rinehart et al., 2005a, b). A large body of
material located in the proximal tail area of NMMNH P-44552 is copi- coprolite was found in the area between the proximal caudals and the
ous, formless, and rich in bone fragments. It contains hand and wrist ischia of NMMNH P-44552 (Fig. 119A). This coprolitic mass measured
bones apparently pertaining to a small juvenile Coelophysis and is de- ~ 10 cm long, ~ 6 cm wide, and ~ 1 cm thick. Most of the coprolite, which
scribed below in the section on cannibalism. contained hundreds of unidentifiable bone fragments in a pale reddish
brown granular matrix, was removed from the block for micropreparation.
Cannibalism Carpal and manual bones indistinguishable from those of a small juvenile
Colbert (1989) observed what he believed to be juvenile Coleophysis were prepared from it (Fig. 119B-C).
Coelophysis bones within the body cavities of AMNH specimens 7223 Nesbitt et al. (2006) further prepared AMNH 7223 and 7224 to
and 7224. Based on this, he attributed cannibalism to the species. This investigate the cannibalism question. They agreed with Gay (2002) that
opinion was reiterated in popular (e.g., Lucas, 1994; Colbert, 1995) and in the case of AMNH 7223 the adult specimen simply overlies the
scientific literature (e.g., Paul, 1988; Fastovsky and Smith, 2004). Gay supposed stomach contents and showed that the actual body cavity
(2002) observed that, at least in AMNH 7223, both left and right ribs of contents of AMNH 7224 were bones of a crocodylomorph, not a juve-
the adult lie across some of the “juvenile remains” and thus indicate that nile Coelophysis. They suggested that, based on their new evidence,
the small bones are in fact under the larger specimen and not inside it. He Coelophysis was not cannibalistic. They further stated that the carpal
 121

FIGURE 119. Coprolite with Coelophysis bones associated with NMMNH P-44552. A, In overview located between the proximal caudals and ischia. Most
of the coprolite has been removed for micropreparation (cm scale). B, An ulnare and other bone fragments in the coprolite (mm scale). C, Coelophysis
wrist elements from the coprolite (mm scale). D, An ulnare and a phalanx fragment (arrows) from the coprolite compared to the manus and wrist of
Coelophysis bauri (cm scale). E, Regurgitalite in situ in the mouth area of NMMNH P-44551, probably the skull and neck of P-44552 (cm scale). F, Close-
up of the regurgitalite showing small Coelophysis premaxillary (circled, upper left) and jaw teeth (circled, near center).
122
and manual elements of the NMMNH coprolite specimen, which they indicates the males in species where the males are larger (Fairbairn, 1997).
had never examined, “…cannot be considered as evidence of cannibalism This mistake was compounded by a misunderstanding of the terms “ro-
in Coelophysis as the purported digested material is taxonomically unin- bust,” which Rinehart et al. (2001) took to mean the longer, larger ani-
formative.” We agree that most bones of the manus, particularly phalan- mals and “gracile” to mean shorter, smaller animals. In any case, the
ges, are very similar across taxa. However, the distal metacarpals and current, larger sample size of this study (70 specimens) no longer sup-
phalanges of coelophysoids are distinct in that they have asymmetric ports this interpretation.
and offset condyles that facilitate the grasping closure of the hand. The Sexual dimorphism in Coelophysis is present in the skull and neck
similarity is remarkable when the manual and carpal bones from the lengths, even in the smallest (~ one-year-old) size class (Fig. 101); the
coprolite are directly compared to those of Coelophysis bauri (Fig. 119D). long skull/long neck morph termed “gracile,” and the short skull/short
We further disagree with the statement of Nesbitt et al. (2006) that their neck morph, “robust.” Sexual dimorphism is also present in the sacrum
findings suggest that Coelophysis was not cannibalistic. Their findings and probably also in the femoral head of Coelophysis (Fig. 108B). We
do demonstrate conclusively that the gut contents of AMNH 7223 and speculate that the longer, gracile morphs (e.g., AMNH 7228) were fe-
7224 are not juvenile Coelophysis, but they have no bearing on the more male based on the unfused condition of sacral vertebral centra s4 and s5.
general question of whether Coelophysis was cannibalistic. This condition may have allowed more flexibility in the sacrum for the
During the time between the publication of the NMMNH copro- purpose of egg laying. Conversely, the fused condition of s4-s5 appears
lite carpal and manual bones (Rinehart et al., 2005a, b) and the Nesbitt et in the shorter, robust morph (e.g., NMMNH P-42200). We summarize
al. (2006) article, preparation of the NMMNH block continued and the sexual dimorphism characters of Coelophysis bauri in Table 9.
additional evidence of cannibalism was found. Regurgitalite material was As noted above, Colbert (1989) also considered the state of fusion
revealed in and around the mouth area of P-44551, a skull and incomplete of the sacral neural spines to be a sexually dimorphic character and noted
neck, which is presumed to belong to the postcranial skeleton P-44552 fusion in the gracile morph (e.g., AMNH 7223) and no fusion in the
(Fig. 119E). The regurgitalite material contains numerous bone fragments, robust morph (e.g., AMNH 7224). Although it would seem that fused
generally much larger than those seen in the coprolite of NMMNH P- sacral neural spines would stiffen the sacrum, counter to the idea of a
44552, in a dark reddish brown matrix that is less granular than the more flexable sacrum in the gracile morph, we have noted that in the few
coprolite matrix and darker in color than the surrounding siltstone. Among specimens where both areas are visible, that when the sacral vertebral
the bone fragments are two laterally compressed finely serrated teeth in centra s4 and s5 are not fused, the neural spines of s4-s5 are also unfused.
a jaw fragment and a premaxilla fragment with one complete, slightly We can calculate a sexual dimorphism index (SDI = SSD, or Sexual
curved, conical, premaxillary tooth with faint, fine longitudinal grooving Size Difference of Fairbairn, 1997) for Coelophysis bauri based on the
(Fig. 119F). The bone fragments are too incomplete for identification, total lengths of the long neck and skull morphs versus the short neck and
but the teeth are morphologically identical (including the denticle count) skull morphs. The SDI is arbitrarily designated positive (+) when the
to those of a small Coelophysis. We note that no other animals from the females are larger and negative (–) when the males are larger.
quarry show similar elements. We conclude that the most parsimonious
SDI = (mean size larger sex/ mean size smaller sex) - 1
interpretation of the small, apparent Coelophysis teeth, manual, and
carpal elements in the coprolite and regurgitalite almost certainly demon- Our data show that the mean skull length of the ~ 4 year adult size
strate cannibalism in the species. class is ~ 210 mm for the robust morph and ~ 290 mm for the gracile
morph (Table 7). The mean neck lengths are, respectively, 420 mm and
Sexual Dimorphism and Reproduction
500 mm. The total lengths of these morphs are 2611 mm and 2771 mm
and the SDI = + 0.061. This is a relatively small SDI, generally more on
Sexual dimorphism par with those of birds than with reptiles and mammals, which typically
Sexual dimorphism has been suggested in Coelophysis bauri by have SDIs that are 10 (reptiles) to 100 (mammals) times larger (Fairbairn,
previous workers, including Paul (1988), Colbert (1989, 1990), Rinehart 1997.
et al. (2001), and Gay (2005). Both Paul and Colbert used principally In modern ethology and sociobiology an interesting suite of traits
AMNH 7223 and AMNH 7224 to define the sexual dimorphs. Paul is often associated with a low SDI, including pair-bonding, monogamy, a
(1988) first illustrated the dimorphs and applied the terms “robust” for
the shorter, stockier animal (AMNH 7224), and “gracile” for the longer, TABLE 9. The sexually dimorphic characters of Coelophysis bauri as
more slender animal (AMNH 7223). Colbert (1989) hypothesized that enumerated by Paul (1988), Colbert (1989), and this study.
differences in the skull, neck, arm lengths, and in the degree of fusion of
the sacral neural spines were indicators of sexual dimorphism. Our sta-
tistical study supports Colbert’s specimen-based indicators of dimor-
phism in skull and neck proportions, but fails to support his arm length
dimorphism. This lack of support could be due to an inadequate sample
size, as complete arms are relatively few. Therefore, we do not deny the
possibility of dimorphism in the arm length of Coelophysis bauri, we
only state that our statistical approach is unable to support it. Gay
(2005) mistakenly diagnosed sexual dimorphism in the ratio of femur to
tibia lengths in Coelophysis bauri. In a plot of femur length versus tibia
length his sample of 10 individuals divided neatly into two groups that
he assumed to be dimorphs (Fig. 107A). The study of a larger sample of
24 presented here shows a continuous, monotonic relationship (Fig.
107B) with no evidence of dimorphism in the femur – tibia relationship.
In a preliminary study of Coelophysis size metrics Rinehart et al.
(2001), using a sample of ~ 35 specimens, mistook two of the larger
statistical size classes of femora for sexual dimorphs and speculated that
the robust morphs were male based on their Sexual Dimorphism Index
(SDI) and on greater variance in the larger class. SDI is usually higher in
species where the males are the larger sex, and higher variance usually
 123
high degree of paternal care of juveniles, a high degree of mate selective- brates) versus those that produce few offspring and care for them assidu-
ness in males, and low male-male aggression (Huxley, 1966). We do not ously (e.g., humans). Based on the age distribution and survivorship data
have sufficient evidence to assign these traits to Coelophysis. We do, presented here Coelophysis bauri occupied an intermediate position, but
however, point out that they are generally associated with a low degree was disposed in the r direction.
of sexual dimorphism, and some of them may have been present in
Coelophysis. Reproductive effort
We judge reproductive effort in terms of egg, clutch, and hatchling
Reproductive strategy size compared to adult female body size. Our method was to assemble a
We consider the very large proportion of small juvenile Coelophysis pertinent database from outgroups of birds and reptiles compiled in
bauri in the age/size distributions (Fig 110), and the ~Type III survivor- several sources (Amadon, 1943; Ferguson, 1970; Andrews and Rand,
ship curve (Fig. 115), to be strong evidence of an r-directed reproductive 1974; Visser and Beintema, 1991; Dzialowski and Sotherland, 2004),
strategy in Coelophysis bauri. First, we will briefly review the r and K develop allometric regressions to fit the data, and then to compare
reproductive strategies, then we will use comparisons to bird and reptile Coelophysis to the results. The database includes 18 birds representing
outgroups to form hypotheses regarding the reproductive effort of C. five orders, and 23 reptiles representing crocodylians, lizards, turtles,
bauri in terms of probable egg size, clutch size, and hatchling weight. and snakes (Table 10). The reproductive effort metrics are summarized
in Table 11.
r versus K First we look at the allometry of hatchling weight versus adult
The notion of r and K selection or reproductive strategy arises weight. Wherever possible in the data used here, the weight of the actual
from Verhulst’s (1838) logistic differential equation, female that produced the hatchlings is used. We see negative allometric
constants for both birds (0.60) and reptiles (0.50) (Fig. 121A). These are
close to the allometric constants of a study by Blueweiss et al. (1978).
which describes the time rate of growth of a population (N) in terms of Using a different (and larger) database, they arrived at allometric con-
r and K. MacArthur and Wilson (1967) fully expanded on the factors r, stants of 0.69 and 0.42 for birds and reptiles, respectively. Using our
which represents unopposed intrinsic exponential population growth, allometric equations, we solved for the hatchling mass that Coelophysis
and K, which represents the actual carrying capacity of the environment. would have had if it had bird-like or reptile-like reproductive effort. We
“Unopposed growth” here means that growth is not influenced by any chose a hypothetical “breeding size” range of 15 to 24 kg for Coelophysis,
negative environmental factors, such as competition for food, space or which represents the calculated mass of the 4-7 year size classes (Table
resource limitations, storms or droughts, predation, disease, etc. The 7). These size classes are clearly above the “bend” of the growth curve,
“carrying capacity” of the environment includes all such negative influ- a position that probably indicates sexual maturity (Zug, 1993). The bird
ences and thus brings into balance the tendency toward runaway popu- equation yields a hatchling mass range of 236 to 316 g, and the reptile
lation growth. equation yields a range of 32 to 41 g.
The influences of r and K may be seen in the sigmoid shape of a Next, we examine egg weight and clutch weight versus female
graph of the logistic function. The graph shows a stretched S shape that weight (Fig. 121B). Again, in both reptiles and birds the allometric rela-
begins in a concave-up fashion and then transitions through an inflection tionships are negative and are generally close to those generated by
point to a concave-down shape (Fig. 120). In the lower, concave-up Blueweiss et al. (1978). Rarely, some birds lay as many as 14 eggs/clutch
portion of the curve, r dominates the expression, and the curve begins to (some gulls), whereas others lay as few as two (some partridges), but the
show exponential growth. In the upper, concave-down portion of the vast majority lay three to six eggs/clutch (Altman and Dittmer, 1964).
curve the restraining force of K has become the dominant factor, so We examine reptile clutch size below. The bird equations (Fig. 121B)
population growth levels off and begins to approach an asymptote that yield egg and clutch masses of 424 to 587 g and 1751 to 2397 g, respec-
represents the maximum population that the resources of the environ- tively, whereas the reptile equations produce 32 to 40 g and 608 to 822
ment can support. g, respectively, over the Coelophysis mass range of 15 to 24 kg.
The “end member” reproductive strategies, r, and K, then, imply It is interesting to note that the bird egg size regression line is very
populations that, respectively, produce extremely large numbers of off- close to the reptile clutch size line, so for the same size female, a single
spring and provide them little or no parental care (e.g., some inverte- bird egg can be nearly the size of the entire clutch of reptile eggs (Fig.
121B). Birds lay fewer, larger eggs; reptiles lay more, but smaller eggs.
This fits well the idea that birds are more K-selected and reptiles more r-
selected. We posit that Coelophysis was a more r-selected animal and
thus fits the reptile reproductive model better.
We will now show that Coelophysis could not have laid eggs of the
size required of it using the “bird model” (i.e., smaller clutches of larger
eggs and more parental care). We collected and used data to dertermine
egg diameter versus egg mass across reptile species from several sources
(Ferguson, 1970; Gaulke and Horn, 2004; Horn and King, 2004; Visser,
2004). Crocodylians and varanid lizards were used to bracket the size
range of Coelophysis (Table 12). We show allometric regressions for the
major and minor diameters of eggs versus egg mass (Fig. 121C). Using
these equations to solve for the minor egg diameter of Coelophysis we
calculate a range of 31 to 33.5 mm for a “reptile model” Coelophysis egg,
and 65 to 86 mm for the “bird model” Coelophysis egg. In the Growth,
Allometry, and Population section above (Fig. 103E) we estimated that
the maximum vent diameter (distance between the ischial peduncles) of
the largest Coelophysis, excluding soft tissue, was 61 mm. Clearly the 65
to 86 mm (largest) diameter egg would not have passed through the vent
FIGURE 120. Graph of a logistic function showing initial exponential buildup, of even the largest Coelophysis, even without accounting for soft tissue.
followed by decreasing slope and final asymptote approach. Finally, we estimate the number of eggs per clutch of Coelophysis
 124

TABLE 10. Reproductive effort database, including 23 reptiles and 18 birds.

 125

TABLE 11. Summary tabulation of reproductive effort allometrics for birds and reptiles, with calculated values for Coelophysis bauri based on the bird and
reptile “models.” Coelophysis fits the “reptile model.”

FIGURE 121. Allometric plots showing reproductive effort in birds and reptiles. A, Hatchling weight as a function of adult female weight. B, Egg and clutch
weight as a function of adult female weight. C, Major and minor diameters of reptile eggs as a function of egg mass. D, Number of reptile eggs per clutch
as a function of adult female weight.
126
TABLE 12. Mass, and major and minor diameters of various crocodylian and varanid lizard eggs.

bauri. Using only reptile data (Table 10) we plot the number of eggs/ to their respective collections: Carl Mehling, American Museum of Natural
clutch versus adult female weight (Fig. 121D). We fit a power curve to History, New York, NY; David Berman and Amy Henrici, Carnegie
the LOG-scaled plot and find that although scatter is high in the data, the Museum of Natural History, Pittsburgh, PA; Michael Ryan, Gary Jack-
number of eggs/clutch increases as the adult mass to the 0.14 power son, and David Chapman, Cleveland Museum of Natural History, Cleve-
(allometric constant = 0.14). Using the curve fit equation to solve for land, OH; Alex Downs, Ruth Hall Museum of Paleontology, Ghost
eggs/clutch in Coelophysis weighing between 15 and 24 kg, we get 24 to Ranch, NM; Farish Jenkins and Bill Amaral, Museum of Comparative
26 eggs/clutch. This is a much greater number than the typical 3 to 6 eggs/ Zoology, Harvard University, Cambridge, MA; and Brandon Strilisky
clutch of most birds (Table 10) (Altman and Dittmer, 1964), but com- and Jackie Wolf, Royal Tyrrell Museum, Drumheller, Alberta, Canada.
pares very well to the average of the reptile eggs/clutch, which is 23. Tom Giermakowski at the Museum of Southwestern Biology, Univer-
Recalling that in the population study, we found approximately 20 one- sity of New Mexico, Albuquerque, NM, provided access to their herpe-
year-olds for each of the very largest animals (7+ year-olds), and assum- tology collection for the purpose of comparative studies of sclerotic
ing that the breeding population would include most of the animals above rings and other morphology. Andrew Johnson, also of the Museum of
the three or four-year-old classes, 24 to 26 eggs per breeding female Southwestern Biology, granted access to the bird collection for similar
seems a reasonable number. purposes and, entirely above and beyond the call of duty, his staff
We conclude that Coelophysis’ reproductive effort was compa- collected “failed rehab” birds from around the state for us to dissect in
rable to that of modern reptiles of similar size. The reproductive strategy comparative anatomy studies.
was r-directed, as is now supported by several lines of evidence, not K- Robert Sullivan provided some measurements of specimens on
directed, as it is in birds. Even though Coelophysis had numerous bird- the State Museum of Pennsylvania’s Coelophysis block. Dale Zelinski,
like characters, including hollow bones, furculae, and overall appearance, who prepared the Cleveland Museum of Natural History Coelophysis
its reproductive characters best fit the “reptile model.” block, provided a bone map of that block. Although the Museum of
Northern Arizona’s Coelophysis block was under study by other work-
ACKNOWLEDGMENTS ers and unavailable to us, David Gillette provided some measurements of
First, we thank David S Berman of the Carnegie Museum, Pitts- specimens from that block. David Smith provided measurements of some
burgh, PA and the volunteer crew that he assembled for excavating the specimens that were used in the early stages of this work, although they
Whitaker Quarry in 1982-83. We also thank the Ghost Ranch staff and were not used in the final analysis because the authors had remeasured
administration for allowing quarry block C-8-82 to be donated to the the same material using a different protocol.
New Mexico Museum of Natural History. Michael Pierce, Sarah Hartshorne, Hillary Jenkins, and several
We especially thank David S Berman and Gregory S. Paul, who volunteers assisted in the rejacketing, rocker construction, and rolling the
reviewed this monograph and improved it through their comments and block. Robert Ungnade took the photos of the roll.
suggestions. Matthew Carrano and Kristian Remes reviewed Rinehart et The Janet Stearns Upjohn Memorial Trust paid part of LFR’s
al. (2007), regarding the furcula of Coelophysis bauri, and improved it travel expenses for data collection trips to AMNH, CMNH, CLMNH,
with their comments. Hans Larsson and Ryosuke Motani reviewed an and RTM. ABH’s work, including travel expenses for data collection
earlier version of the sclerotic ring and vision section; their comments and trips to AMNH and MCZ was funded by University Research Council
suggestions improved its content. Robert Sullivan provided efficient grants from the Office of Graduate Studies at Appalachian State Univer-
associate editorship of the manuscript, handling and evaluating the re- sity. Inexpensive housing for ABH and LFR in Manhattan, used during
view and revision. AMNH studies, was provided by Appalachian State University.
We thank the following institutions and staff members for access
 Rinehart et al., 2009, The Paleobiology of Coelophysis bauri. New Mexico Museum of Natural History and Science, Bulletin 45.

127
REFERENCES

Alexander, R.M., 1977, Allometry of the limbs of antelopes (Bovidae): Carrano, M.T., 1999, What, if anything, is a cursor? Categories versus
Journal of Zoology, London, v. 183, p. 125-146. continua for determining locomotor habit in mammals and dinosaurs:
Alexander, R.M., 1989, Dynamics of dinosaurs and other extinct giants: Journal of the Zoological Society of London, v. 247, p. 29-42.
New York, Columbia University Press, 167 p. Carrano, M.T., Hutchinson, J.R. and Sampson, S.D., 2005, New informa-
Altman, P.L. and Dittmer, D.S., 1964, Biology data book, Biological hand- tion on Segisaurus halli, a small theropod from the Early Jurassic of
books: Washington, D.C., Federation of American Societies for Ex- Arizona: Journal of Vertebrate Paleontology, v. 25, p. 835-849.
perimental Biology, 633 p. Carroll, R.L. and Gaskill, P., 1978, The Order Microsauria: Memoirs of the
Amadon, D., 1943, Bird weights and egg weights: The Auk, v. 60, no. 2, p. American Philosophical Society, v. 126, 211 p.
221-234. Casini, G., Fontanesi, G. and Bagnoli, P., 1993, Binocular processing in
Andrews, R.M., 1982, Patterns of growth in reptiles; in Gans, C. and Pough, frontal-eyed birds; in Zeigler, H.P. and Bischof, H.-J., eds., Vision, brain,
F.H., eds., Biology of the Reptilia Vol. 13 Physiology D: London, Aca- and behavior in birds: The MIT Press, Cambridge, Massachusetts, p.
demic Press, p. 273-321. 159-168.
Andrews, R. and Rand, A.S., 1974, Reproductive effort in anoline lizards: Chabreck, R.H. and Joanen, T., 1979, Growth rates of American alligators
Ecology, v. 55, p. 1317-1327. in Louisiana: Herpetologica, v. 35, p. 51-57.
Ballew, K.L., 1989, A phylogenetic analysis of Phytosauria from the Late Chapman, R., Sadleir, R.W., Dodson, P. and Makovicky, P., 2008, Handling
Triassic of the western United States; in Lucas, S.G. and Hunt, A.P., eds., missing data in paleontological matrices: Approaches for exploratory
Dawn of the age of dinosaurs in the American Southwest: Albuquerque, multivariate analyses in morphometrics: Journal of Vertebrate Paleon-
New Mexico Museum of Natural History, p. 309-339. tology, v. 28, supplement to no. 3, p. 62A.
Barlow, H.B. 1986. Why can’t the eye see better?; in Pettigrew, J.D., Chatterjee, S., 1985, Postosuchus, a new thecodontian reptile from the
Sanderson, K. and Levick, W.R., eds., Visual neuroscience: Cambridge Triassic of Texas and the origin of tyrannosaurs: Philosophical Trans-
University Press, Cambridge, United Kingdom, p. 3-18. actions of the Royal Society of London, Series B, v. 309, p. 395-460.
Baumel, J.J., King, A.S., Breazile, J.E., Evans, H.E., and Vanden Berge, J.C., Chatterjee, S., 1993, Shuvosaurus, a new theropod: National Geographic
2005, Handbook of avian anatomy: Nomina anatomica avium: Cam- Research and Exploration, v. 9, p. 274-285.
bridge, Harvard University Nuttall Orithological Club, 779 p. Chiappe, L. M. and Dingus, L., 2001, Walking on eggs: Scribner, New York,
Bellairs, A. 1970. The life of reptiles, Volume 2. Universe Books, New 219 p.
York, 590 p. Chiappe, L.M., Norell, M. and Clark, J., 2001, A new skull of Gobipteryx
Bennett, S.C., 1995, A statistical study of Ramphorynchus from the minuta (Aves: Enantiornithes) from the Cretaceous of the Gobi Desert.
Solnhofen Limestone of Germany: Year-classes of a single large species: American Museum Novitates, no. 3346, p. 1-15.
Journal of Paleontology, v. 69, p. 569-580. Chinsamy, A., 1994, Dinosaur bone histology: Implications and inferences;
Benson, R.H., Berdan, J.M., Bold, W.A.V.D., Hanai, T., Hessland, I., Howe, in Rosenberg, G.D. and Wolberg, D.L., eds., Dino Fest: The Paleonto-
H.V., Kesling, R.V., Levinson, S.A., Reyment, R.A., Moore, R.C., Scott, logical Society Special Publication: The Paleontological Society, Knox-
H.W., Shaver, R.H., Sohn, I.G., Stover, L.E., Swain, F.M. and Sylvester- ville, p. 213-227.
Bradley, P.C., 1961, Crustacea - Ostracoda: Systematic descriptions; in Christiansen, P. and Farina, R.A., 2004, Mass prediction in theropod dino-
Moore, R.C. and Pitrat, C.W., eds., Treatise on Invertebrate Paleontol- saurs: Historical Biology, v. 16, p. 85-92.
ogy Part Q Arthropoda 3: University of Kansas Press and The Geologi- Chure, D.J. 1998. On the orbit of theropod dinosaurs: Gaia, v. 15, p. 233-
cal Society of America, Lawrence, p. Q99-Q414. 240.
Biewener, A.A., 1983, Allometry of quadrupedal locomotion: The scaling Chure, D.J. and Madsen, J.H., 1996, On the presence of furculae in some
of duty factor, bone curvature and limb orientation to body size: Journal non-maniraptoran theropods: Journal of Vertebrate Paleontology, v.
of Experimental Biology, v. 105, p. 147-171. 16, p. 573-577.
Blueweiss, L., Fox, H., Kudzma, V., Nakashima, D., Peters, R., and Sams, S., Clack, J.A., 2002, Gaining ground: The origin and evolution of early tetra-
1978, Relationships between body size and some life history param- pods: Indiana University Press, Bloomington, Indiana, 369 p.
eters: Oecologia, v. 37, p. 257-272. Claessens, L.P.A.M., 2004, Dinosaur gastralia: Origin, morphology, and
Brinkman, D.B., Eberth, D.A. and Currie, P.J., 2007, From bonebeds to function: Journal of Vertebrate Paleontology, v. 24, p. 89-106.
paleobiology: Applications of bonebed data; in Rogers, R.R., Eberth, Clark, J.M., Norell, M.A. and Rowe, T., 2002, Cranial anatomy of Citipati
D.A., and Fiorillo, A.R., eds., Bonebeds: University of Chicago Press, osmolskae (Theropoda, Oviraptorosauria), and a reinterpretation of
Chicago, p. 221-263. the holotype of Oviraptor philoceratops: American Museum Novitates,
Brown, B., 1912, A crested dinosaur from the Edmonton Cretaceous: Ameri- no. 3364, p. 1-24.
can Museum of Natural History, Bulletin 31, p. 131-136. Clark, J.M., Sues, H.-D. and Berman, D.S., 2001, A new specimen of
Brown, B., 1917, A complete skeleton of the horned dinosaur Monoclonius, Hesperosuchus agilis from the Upper Triassic of New Mexico and the
and description of a second skeleton showing skin impressions: Bulletin interrelationships of basal crocodylomorph archosaurs: Journal of Ver-
of the American Museum of Natural History, v. 37, p. 281-306. tebrate Paleontology, v. 20, p. 683-704.
Brown, B., and Schlaikjer, E.M., 1940, The structure and relationships of Coborn, J., 1997, Savanna monitors: T. F .H. Publications, Inc., Neptune
Protoceratops: Annals of the New York Academy of Sciences, v. 40, p. City, New Jersey, 63 p.
133-266. Colbert, E.H., 1947, The little dinosaurs of Ghost Ranch: Natural History,
Bryant, H.N. and Russell, A.P., 1993, The occurrence of clavicles within v. 56, p. 392-399, 427-428.
Dinosauria: Implications for the homology of the avian furcula and the Colbert, E.H., 1952, A pseudosuchian reptile from Arizona: Bulletin of the
utility of negative evidence: Journal of Vertebrate Paleontology, v. 13, American Museum of Natural History, v. 99, p. 561-592.
p. 171-184. Colbert, E. H., 1964, The Triassic dinosaur genera Podokesaurus and
Camp, C.L., 1936, A new type of small bipedal dinosaur from the Navajo Coelophysis: American Museum Novitates: no. 2168, p. 1-12.
Sandstone of Arizona: University of California Publications in Geologi- Colbert, E.H., 1989, The Triassic dinosaur Coelophysis: Museum of North-
cal Sciences, v. 24, p. 39-56. ern Arizona, Bulletin 57, 160 p.
Carpenter, K., 2002, Forelimb biomechanics of nonavian theropod dino- Colbert, E.H., 1990, Variation in Coelophysis bauri; in Carpenter, K. and
saurs in predation: Senckenbergiana Lethaea, v. 82, p. 59-76. Currie, P.J., eds., Dinosaur systematics, approaches and perspectives:
Carpenter, K., Hirsch, K.F. and Horner, J.R., 1994, Dinosaur eggs and Cambridge University Press, Cambridge, p. 81-90.
babies: Cambridge University Press, New York, 372 p. Colbert, E.H., Charig, A.J., Dodson, P., Gillete, D.D., Ostrom, J.H. and
128
Weishampel, D., 1992, Coelurus bauri Cope, 1887 (currently Coelophysis Serie B (Geologie und Paläontologie), no. 118, p. 1-39.
bauri; Reptilia, Saurischia): Proposed replacement of the lectotype by a Gaulke, M. and Horn, H.-G., 2004, Varanus salvator; in Pianka, E.R., King,
neotype: Bulletin of Zoological Nomenclature, v. 49, p. 276-279. D.R. and King, R.A., eds., Varanoid lizards of the world: Indiana Univer-
Cope, E.D., 1881, Belodon in New Mexico: American Naturalist, v. 15, p. sity Press, Bloomington, p. 244-271.
922-923. Gauthier, J., 1986, Saurischian monophyly and the origin of birds; in Padian,
Cope, E.D., 1887a, The dinosaurian genus Coelurus: American Naturalist, K., ed., The origin of birds and the evolution of flight: Memoirs of the
v. 21, p. 367-369. California Academy of Sciences: California Academy of Sciences, San
Cope, E.D., 1887b, A contribution to the history of the Vertebrata of the Francisco, p. 1-55.
Trias of North America: Proceedings of the American Philosophical Gauthier, J., Kluge, A.G. and Rowe, T., 1988, Amniote phylogeny and the
Society, v. 24, p. 221-227. importance of fossils: Cladistics, v. 4, p. 105-209.
Cope, E.D., 1889, On a new genus of Triassic Dinosauria: American Natu- Gay, R.J., 2002, The myth of cannibalism in Coelophysis bauri: Journal of
ralist, v. 23, p. 626. Vertebrate Paleontology, v. 22, supplement to no. 3, p. 57A.
Currie, P.J. 1997. Theropoda; in Currie, P.J. and Padian, K., eds., Encyclo- Gay, R., 2005, Sexual dimorphism in the Early Jurassic theropod dinosaur
pedia of Dinosaurs: Academic Press, San Diego, p. 731-736. Dilophosaurus and a comparison with other related forms; in Carpen-
Deevey Jr., E.S., 1947, Life tables for natural populations of animals: Quar- ter, K., ed., The carnivorous dinosaurs: Indiana University Press,
terly Review of Biology, v. 22, p. 283-314. Bloomington, p. 277-283.
Degenhardt, W.G., Painter, C.W. and Price, A.H., 1996, Amphibians and Glantz, S.A., 2005, Primer of biostatistics: McGraw-Hill, Inc., New York,
reptiles of New Mexico: University of New Mexico Press, Albuquerque, 520 p.
431 p. Glantz, S.A. and Slinker, B.K., 2001, Primer of applied regression and
Dickinson, W.R. and Gehrels, G.E., 2008, U-Pb ages of detrital zircons in analysis of variance: McGraw-Hill, Inc., New York, 949 p.
relation to paleogeography: Triassic paleodrainage networks and sedi- Godefroit, P. and Cuny, G., 1997, Archosauriform teeth from the Upper
ment dispersal across southwest Laurentia: Journal of Sedimentary Re- Triassic of Saint-Nicolas-de-Port: Palaeovertebrata, v. 26, p. 1-34.
search, v. 78, p. 745-764. Goldstein, B.S., Trinh, T., Hargrave, E. and Demko, T.M., 1996, Geologic
Ditmars, R.L., 1967, Reptiles of the world: Macmillian Company, New and stratigraphic setting of the Coelophysis quarry, Ghost Ranch, north-
York, 321 p. ern New Mexico: A reassessment: Geological Society of America Ab-
Downs, A., 2000, Coelophysis bauri and Syntarsus rhodesiensis compared, stracts with Programs, v. 28, p. 9.
with comments on the preparation and preservation of fossils from the Gould, S.J., 1966, Allometry and size in ontogeny and phylogeny: Biologi-
Ghost Ranch Coelophysis Quarry: New Mexico Museum of Natural His- cal Reviews, v. 41, p. 587-640.
tory and Science, Bulletin 17, p. 33-37. Hall, M., 2004, A new approach for infering activity pattern in avians using
Downs, A. and Davidge, J., 1997, A choristodere from the Ghost Ranch the orbit and sclerotic ring. Journal of Vertebrate Paleontology, v. 24,
Coelophysis quarry: Journal of Vertebrate Paleontology, v. 17, supple- supplement to no. 3, p. 67A.
ment to no. 3, p. 43A-44A. Harding, J.P., 1949, The use of probability paper for the graphical analysis
Dubiel, R.F., 1989, Depositional environments of the Upper Triassic Chinle of polymodal frequency distributions: Journal of the Marine Biological
Formation in the eastern San Juan Basin and vicinity, New Mexico: U.S. Association of the United Kingdom, v. 28, p. 141-153.
Geological Survey, Bulletin 1808B, p. 1-22. Hargrave, E.V., Goldstein, B.S., Trinh, T. and Trinh, T., 1996, Paleomag-
Dubiel, R.F., Parrish, J.T., Parrish, J.M. and Good, S.C., 1991, The Pangaean netism of the Chinle Formation at Ghost Ranch, New Mexico: Age
megamonsoon - evidence from the Upper Triassic Chinle Formation, constraints for Coelophysis bauri: Geological Society of America Ab-
Colorado Plateau: Palaios, v. 6, p. 347-370. stracts with Programs, v. 77, p. 72.
Dzialowski, E.M. and Sotherland, P.R., 2004, Maternal effects of egg size Harris, J.D. and Downs, A., 2002, A drepanosaurid pectoral girdle from the
on emu Dromaius novaehollandiae egg composition and hatchling phe- Ghost Ranch (Whitaker) Coelophysis quarry (Chinle Group, Rock Point
notype: Journal of Experimental Biology, v. 207, p. 597-606. Formation, Rhaetian), New Mexico: Journal of Vertebrate Paleontol-
Fairbairn, D.J., 1997, Allometry for sexual size dimorphism: Pattern and ogy, v. 22, p. 70-75.
process in the coevolution of body size in males and females: Annual Harris, J.D., Lucas, S.G., Estep, J.W. and Li, J., 2000, A new and unusual
Reviews of Ecology and Systematics, v. 28, p. 659-687. sphenosuchian (Archosauria: Crocodylomorpha) from the Lower Juras-
Farlow, J.O., Brinkman, D.L., Abler, W.L. and Currie, P.J., 1991, Size, sic Lufeng Formation, People’s Republic of China: Neues Jahrbuch für
shape, and serration density of theropod dinosaur lateral teeth: Modern Geologie und Paläontologie Abhandlungen, v. 214, no. 1, p. 47-68.
Geology, v. 16, p. 161-198. Hatfield, J.W.I., 1993, Green iguana: The ultimate owner’s manual:
Fastovsky, D.E. and Smith, J.B., 2004, Dinosaur paleoecology; in Dunthorpe Press, Portland, Oregon, 656 p.
Weishampel, D.B., Dodson, P. and Osmolska, H., eds., The Dinosauria: Hay, O.P., 1930, Second bibliography and catalogue of the fossil vertebrates
University of California Press, Berkeley, p. 614-626. of North America, v. II: Carnegie Institution of Washington, Publica-
Faux, C.M. and Padian, K., 2007, The oposthotonic posture of vertebrate tion 390, v. 2, 1074 p.
skeletons: Postmortem contraction or death throes?: Paleobiology, v. Hecht, M., Ostrom, J.H., Viohl, G. and Wellnhofer, P., eds., 1985, The
33, p. 201-226. beginnings of birds, Eichstatt: Freunde des Jura-Museums, Eichstatt,
Ferguson, M.W.J., 1970, Reproductive biology and embrylogy of the croco- Germany, 382 p.
dilians; in Gans, C., Billett, F. and Maderson, P.F.A., eds., Biology of the Heckert, A.B., 2001, The microvertebrate record of the Upper Triassic
Reptilia: John Wiley & Sons, New York, p. 239-381. (Carnian) lower Chinle Group, southwestern U.S.A. and the early evolu-
Frank, P.W., 1988, Conchostraca: Paleogeography, Paleoclimatology, Pa- tion of dinosaurs [Ph.D. thesis]: University of New Mexico, Albuquer-
leoecology, v. 62, p. 399-403. que, 465 p.
Galton, P.M., 1973, Redescription of the skull and mandible of Parksosaurus Heckert, A.B., 2004, Late Triassic microvertebrates from the lower Chinle
from the Late Cretaceous with comments on the family Group (Otischalkian-Adamanian: Carnian), southwestern U.S.A.: New
Hypsilophodontidae (Ornithischia): Life Science Contributions of the Mexico Museum of Natural History and Science, Bulletin 27, 170 p.
Royal Ontario Museum, v. 89, p. 1-21. Heckert, A.B., Lucas, S.G., Rinehart, L.F. and Hunt, A.P., 2008, A new genus
Galton, P.M., 1974, The ornithischian dinosaur Hypsilophodon from the and species of sphenodontian from the Ghost Ranch Coelophysis quarry
Wealden of the Isle of Wight: Bulletin of the British Museum (Natural (Upper Triassic: Apachean), Rock Point Formation, New Mexico, USA:
History) Geology, v. 25, p. 1-152. Palaeontology, v. 51, p. 827-845.
Galton, P.M., 1985, Cranial anatomy of the prosauropod dinosaur Sellosaurus Hodos, W. 1993. The visual capabilities of birds; in Zeigler, H.P. and Bischof,
gracilis from the Middle Stubensandstein (Upper Triassic) of H.-J., eds., Vision, brain, and behavior in birds: The MIT Press, Cam-
Nordwürttemberg, West Germany: Stuttgarter Beiträge zur Naturkunde bridge, Massachusetts, p. 63-76.
 129
Hoff, C.C., 1942, The ostracodes of Illinois, their biology and taxonomy: p.
Illinois Biological Monographs, v. 14, 196 p. Huxley, J.S., 1966, Introduction to a discussion on ritualization of behavior
Horn, H.-G. and King, D.R., 2004, Varanus giganteus; in Pianka, E.R., in animals and man: Philosophical Transactions of the Royal Society of
King, D.R. and King, R.A., eds., Varanoid lizards of the world: Indiana London B, v. 251, p. 249-271.
University Press, Bloomington, p. 335-354. Irmis, R.B., Nesbitt, S.J., Padian, K., Smith, N.D., Woody, D. and Downs, A.,
Horner, J.R., 2000, Dinosaur reproduction and parenting: Annual Review of 2007, A Late Triassic dinosauromorph assemblage from New Mexico
Earth and Planetary Sciences, v. 28, p. 19-45. and the rise of dinosaurs: Science, v. 317, p. 358-361.
Houck, M.A., Gauthier, J.A. and Strauss, R.E., 1990, Allometric scaling of Johnson, S.C., Lucas, S.G. and Hunt, A.P., 2002, Macro-fish fauna of the
the earliest fossil bird, Archaeopteryx lithographica: Science, v. 247, p. Upper Triassic (Apachean) Redonda Formation, eastern New Mexico:
195-198. New Mexico Museum of Natural History, Bulletin 21, p. 107-114.
Hungerbühler, A., 2000, Heterodonty in the European phytosaur Nicrosaurus KaleidaGraph, 1997, Graphic analysis software, Version 3.09, Synergy Soft-
kapffi and its implications for the taxonomic utility and functional ware.
morphology of phytosaur dentitions: Journal of Vertebrate Paleontol- Kietzke, K., 1989, Calcareous microfossils from the Triassic of the south-
ogy, v. 20, p. 31-48. western United States; in Lucas, S.G. and Hunt, A.P., eds., Dawn of the
Hungerbühler, A., 2002, The late Triassic phytosaur Mystriosuchus westphali, age of dinosaurs in the American Southwest: New Mexico Museum of
with a revision of the genus: Palaeontology, v. 45, p. 377-418. Natural History, Albuquerque, p. 223-232.
Hunt, A.P., 1992, Late Pennsylvanian coprolites from the Kinney Brick King, J.R., 1971, Probability charts for decision making: Industrial Press
Quarry, central New Mexico, with notes on the classification and utility Inc., New York, 290 p.
of coprolites: New Mexico Bureau of Mines and Mineral Resources, Kirby, R.A., 1991, A vertebrate fauna from the Upper Triassic Owl Rock
Bulletin 138, p. 221-229. Member of the Chinle Formation in northern Arizona [M.S. thesis]:
Hunt, A.P., 1994, Vertebrate paleontology and biostratigraphy of the Bull Northern Arizona University, Flagstaff, 476 p.
Canyon Formation (Chinle Group, Upper Triassic), east-central New Knoop, P.A. and Owen, R.M., 1994, Computer software for identifying
Mexico with revisions of the families Metoposauridae (Amphibia: compositional subpopulations in marine sediment geochemical data us-
Temnospondyli) and Parasuchidae (Reptilia: Archosauria) [Ph.D. the- ing threshold value analysis: Marine Georesources and Geotechnology,
sis]: University of New Mexico, Albuquerque, 404 p. v. 12, p. 11-24.
Hunt, A.P., 1997, A new coelacanth (Osteichthyes: Actinista) from the Kock, J.G.S. and Link, R.F., 1970, Statistical analysis of geological data:
continental Upper Triassic of New Mexico: New Mexico Museum of Dover Publications, New York, 438 p.
Natural History, Bulletin 11, p. 25-27. Konishi, M. and Pettigrew, J.D., 1975, Some observations on the visual
Hunt, A.P. and Lucas, S.G., 1989, Late Triassic vertebrate localities in New system of the oilbird (Steatornis caripensis): National Geographic Soci-
Mexico; in Lucas, S.G. and Hunt, A.P., eds., Dawn of the age of dinosaurs ety Research Reports, v. 16, p. 439-449.
in the American Southwest: New Mexico Museum of Natural History, Kozur, H.W. and Weems, R.E., 2005, Conchostracan evidence for a late
Albuquerque, p. 71-101. Rhaetian to early Hettangian age for the CAMP volcanic event in the
Hunt, A.P. and Lucas, S.G., 1991, Rioarribasaurus, a new name for a Late Newark Supergroup, and a Sevatian (late Norian) age for the immedi-
Triassic dinosaur from New Mexico (USA): Paläontologische Zeitschrift, ately underlying beds: Hallesches Jahrbuch für Geowissenschaften. Reihe
v. 65, p. 191-198. B: Geologie, Paläontologie, Mineralogie, v. 27, p. 21-51.
Hunt, A.P. and Lucas, S.G., 1993a, Triassic vertebrate paleontology and Kozur, H.W. and Weems, R.E., 2007, Upper Triassic conchostracan bios-
biochronology of New Mexico: New Mexico Museum of Natural His- tratigraphy of the continental basins of eastern North America: Its
tory and Science, Bulletin 2, p. 49-60. importance for correlating Newark Supergroup events with the Ger-
Hunt, A.P. and Lucas, S.G., 1993b, A new phytosaur (Reptilia: Archosauria) manic Basin and the international geologic time scale: New Mexico
genus from the uppermost Triassic of the western United States and its Museum of Natural History and Science, Bulletin 41, p. 137-188.
biochronological significance: New Mexico Museum of Natural History Larson, P. and Rigby, J.K., 2005, Furcula of Tyrannosaurus rex; in Carpen-
and Science, Bulletin 3, p. 193-196. ter, K., ed., The carnivorous dinosaurs: Indiana University Press,
Hunt, A.P., Heckert, A.B., Lucas, S.G. and Downs, A., 2002, The distribu- Bloomington, p. 247-255.
tion of the enigmatic reptile Vancleavea in the Upper Triassic Chinle Latimer, H.B., 1927, Postnatal growth of the chicken skeleton: American
Group of the western United States: New Mexico Museum of Natural Journal of Anatomy, v. 40, p. 1-57.
History and Science, Bulletin 21, p. 269-273. Lehane, J., 2005, Anatomy and relationships of Shuvosaurus, a basal
Hunt, A.P., Lucas, S.G. and Berman, D.S., 1992, The Late Pennsylvanian theropod from the Triassic of Texas [M.S. thesis]: Texas Tech Univer-
amphibian fauna of the Kinney Quarry, central New Mexico: New Mexico sity, Lubbock, 92 p.
Bureau of Mines and Mineral Resources, Bulletin 138, p. 211-220. Litwin, R.J., 1986, The palynostratigraphy and age of the Chinle and
Hunt, A.P., Lucas, S.G., Heckert, A.B., Sullivan, R.M. and Lockley, M.G., Moenave formations, southwestern USA [Ph.D. thesis]: The Pennsyl-
1998, Late Triassic dinosaurs from the western United States: Geobios, vania State University, College Park, 265 p.
v. 31, p. 511-531. Litwin, R.J., Traverse, A. and Ash, S.R., 1991, Preliminary palynological
Hunt, A.P., Lucas, S.G. and Lockley, M.G., 1998, Taxonomy and strati- zonations of the Chinle Formation, southwestern U.S.A., and its corre-
graphic and facies significance of vertebrate coprolites of the Upper lation to the Newark Supergroup (eastern U.S.A.): Review of
Triassic Chinle Group, western United States: Ichnos, v. 5, p. 225-234. Palaeobotany and Palynology, v. 68, p. 269-287.
Hunt, A.P., Lucas, S.G. and Spielmann, J.A., 2005, The holotype specimen Long, J.A., 1995, The rise of fishes: The Johns Hopkins University Press,
of Vancleavea campi from Petrified Forest National Park, Arizona, Baltimore, 223 p.
with notes on the taxonomy and distribution of the taxon: New Mexico Long, R.A. and Murry, P.A., 1995, Late Triassic (Carnian and Norian)
Museum of Natural History and Science, Bulletin 29, p. 59-66. tetrapods from the southwestern United States: New Mexico Museum of
Hunt, A.S., 1967, Growth, variation, and instar development of an agnostid Natural History and Science, Bulletin 4, 254 p.
trilobite: Journal of Paleontology, v. 41, p. 203-208. Lucas, S.G., 1993, The Chinle Group: Revised stratigraphy and biochronology
Hutchinson, J.R., 2001a, The evolution of femoral osteology and soft of Upper Triassic strata in the western United States: Museum of North-
tissues on the line to extant birds (Neornithes): Zoological Journal of ern Arizona, Bulletin 59, p. 27-50.
the Linnean Society, v. 131, p. 169-197. Lucas, S.G., 1994, Dinosaurs: The Textbook: Wm. C. Brown Publishers,
Hutchinson, J.R., 2001b, The evolution of pelvic osteology and soft tissues Dubuque, 290 p.
on the line to extant birds (Neornithes): Zoological Journal of the Lucas, S.G., 1997, Upper Triassic Chinle Group, western United States: A
Linnean Society, v. 131, p. 123-168. nonmarine standard for Late Triassic time; in Dickins, J. M., Yang, Z.,
Huxley, J.S., 1932, Problems of relative growth: Dial Press, New York, 276 Yin, H., Lucas, S. G. and Acharyya, S. K., eds., Late Palaeozoic and early
130
Mesozoic circum-Pacific events and their global correlation: Cambridge lishing House, Boston, 117 p.
Univerity Press, Cambridge, p. 209-228. McKinney, M.L. and McNamara, K.J., 1991, Heterochrony, the evolution
Lucas, S.G., 1998, Global Triassic tetrapod biostratigraphy and of ontogeny: Plenum Press, New York, 437 p.
biochronology: Palaeogeography, Palaeoclimatology, Palaeoecology, v. McMahon, T.A., 1975, Allometry and biomechanics: Limb bones in adult
143, p. 347-384. ungulates: American Naturalist, v. 109, p. 547-563.
Lucas, S.G. and Hunt, A.P., 1992, Triassic stratigraphy and paleontology, Molnar, R.E., 1991, The cranial morphology of Tyrannosaurus rex:
Chama Basin and adjacent areas, north-central New Mexico: New Mexico Palaeontographica Abteilung, A, v. 217, p. 137-176.
Geological Society, Guidebook 43, p. 151-167. Molnar, R.E. and Farlow, J.O., 1990, Carnosaur paleobiology; in Weishampel,
Lucas, S.G. and Rinehart, L.F., 2005, Nonmarine bivalves from the Lower D.B., Dodson, P. and Osmolska, H., eds., The Dinosauria: University of
Permian (Wolfcampian) of the Chama Basin, New Mexico: New Mexico California Press, Berkeley, California, p. 210-224.
Geological Society, Guidebook 56, p. 283-287. Neill, W.T., 1971, The last of the ruling reptiles; alligators, crocodiles, and
Lucas, S.G. and Tanner, L.H., 2007, Tetrapod biostratigraphy and their kin: Columbia University Press, New York, 486 p.
biochronology of the Triassic-Jurassic transition on the southern Colo- Nesbitt, S.J., 2007, The anatomy of Effigia okeeffeae (Archosauria, Suchia),
rado Plateau, USA: Palaeogeography, Palaeoclimatology, Palaeoecology, theropod-like convergence, and the distribution of related taxa: Bulletin
v. 244, p. 242-256. of the American Museum of Natural History, v. 302, 84 p.
Lucas, S.G., Heckert, A.B. and Anderson, O.J., 1997, The Triassic System Nesbitt, S.J., Irmis, R.B., Parker, W.G., Smith, N.D., Turner, A.H. and
on the Fort Wingate quadrangle, west-central New Mexico: New Mexico Rowe, T., 2009, Hindlimb osteology and distribution of basal
Geology, v. 19, p. 33-42. dinosauromorphs from the Late Triassic of North America: Journal of
Lucas, S.G., Heckert, A.B. and Estep, J.W., 1999, Correlation of Triassic Vertebrate Paleontology, v. 29, p. 498-516.
strata across the Rio Grande rift, north-central New Mexico: New Mexico Nesbitt, S.J. and Norell, M.A., 2006, Extreme convergence in the body plan
Geological Society, Guidebook 50, p. 305-310. of an early suchian (Archosauria) and ornithomimosaurs (Theropoda):
Lucas, S.G., Heckert, A.B., Estep, J.W., and Anderson, O.J., 1997, Stratigra- Proceedings of the Royal Society B, v. 273, p. 1045-1048.
phy of the Upper Triassic Chinle Group, Four Corners Region: New Nesbitt, S.J., Stocker, M.R., Small, B.J. and Downs, A., 2009, The osteology
Mexico Geological Society, Guidebook 48, p. 81-107. and relationships of Vancleavea campi (Reptilia: Archosauriformes):
Lucas, S.G., Hunt, A.P., Heckert, A.B. and Spielmann, J.A., 2007b, Global Zoological Journal of the Linnean Society, v. 157, p. 814-864.
Triassic tetrapod biostratigraphy and biochronology: 2007 status: New Nesbitt, S.J., Turner, A.H., Erickson, G.M. and Norell, M.A., 2006, Prey
Mexico Museum of Natural History and Science, Bulletin 41, p. 229- choice and cannibalistic behavior in the theropod Coelophysis: Biology
240. Letters, v. 2, p. 611-614.
Lucas, S.G., Spielmann, J.A. and Hunt, A.P., 2007a, Taxonomy of Neustrueva, I.Y., 1977, Ostracode biofaces in Paleozoic and Mesozoic lake
Shuvosaurus, a Late Triassic archosaur from the Chinle Group, Ameri- sediments of the USSR; in Loffler, H. and Danielopol, D.L., eds., As-
can Southwest: New Mexico Museum of Natural History and Science, pects of ecology and zoogeography of recent and fossil Ostracoda: Junk,
Bulletin 41, p. 259-261. The Hague, p. 451-458.
Lucas, S.G., Zeigler, K.E. and Heckert, A.B., 2003, Invertebrate paleontol- Norell, M.A., Clark, J.M., Demberelyin, D., Rhinchen, B., Chiappe, L.M.,
ogy of the Upper Triassic Snyder quarry, Chinle Group, Chama basin, Davidson, A.R., McKenna, M.A., Altangerel, P. and Novacek, M.J.,
New Mexico: New Mexico Museum of Natural History and Science, 1994, A theropod dinosaur embryo and the affinities of the Flaming
Bulletin 24, p. 63-65. Cliffs dinosaur eggs: Science, v. 266, p. 779-782.
Lupe, R. and Silberling, N.J., 1985, Genetic relationship between Lower O’Sullivan, R.B., 1974, The Upper Triassic Chinle Formation in north-
Mesozoic strata of the Colorado Plateau and marine strata of the Great central New Mexico: New Mexico Geological Society, Guidebook 25, p.
Western Basin: Significance for accretionary history of Cordilleran 171-178.
lithotectonic terranes; in Howell, D. G., ed., Tectonostratigraphic ter- Ostrom, J.H., 1961, Cranial morphology of the hadrosaurian dinosaurs of
ranes of the Circum-Pacific region: Circum-Pacific Council for Energy North America. Bulletin of the American Museum of Natural History, v.
and Mineral Resources, p. 263-271. 122, p. 35-186.
MacArthur, R.H. and Wilson, E.O., 1967, The theory of island biogeogra- Ostrom, J.H., 1976, Archaeopteryx and the origin of birds: Biological Jour-
phy: Princeton University Press, Princeton, 203 p. nal of the Linnean Society, v. 8, p. 91-182.
Madsen, J.J.H., 1976, Allosaurus fragilis: a revised osteology: Utah Geo- Oxnard, C. E., 1987, Fossils, teeth and sex: New perspectives on human
logical Survey, 109 p. evolution: Hong Kong University Press, Hong Kong, 280 p.
Madsen, J.H., Jr. and Welles, S.P., 2000, Ceratosaurus (Dinosauria, Padian, K., 1986, On the type material of Coelophysis Cope (Saurischia:
Theropoda) a revised osteology: Utah Geological Survey Miscellaneous Theropoda) and a new specimen from the Petrified Forest of Arizona
Publication, v. 2, p. 1-89. (Late Triassic: Chinle Formation); in Padian, K., ed., The beginning of
Manthey, U. and Schuster, N., 1996, Agamid lizards: T.F.H. Publications, the age of dinosaurs: Faunal change across the Triassic-Jurassic bound-
Inc., Neptune City, New Jersey, 189 p. ary: Cambridge University Press, Cambridge, p. 45-60.
Marks, L. S., 1951, Mechanical engineer’s handbook: McGraw-Hill Book Padian, K., 1997, Pectoral girdle; in Currie, P. J. and Padian, K., eds.,
Company, Inc., New York, 2236 p. Encyclopedia of dinosaurs: Academic Press, San Diego, p. 530-536.
Martin, G.R., 1982, An owl’s eye: Schematic optics and visual performance Padian, K., Horner, J.R., and d’Ricqlès, A.J., 2004, Growth in small dino-
in Strix aluco: Journal of Comparative Physiology, v. 145, p. 341-349. saurs and pterosaurs: The evolution of archosaurian growth strategies:
Martin, G.R., 1985, Eye; in King, A.S. and McLelland, J., eds., Form and Journal of Vertebrate Paleontology, v. 23, p. 555-571.
function in birds, Vol. 3: Academic Press, London, p. 311-373. Parks, W.A., 1928, Struthiomimus samueli, a new species of Ornithomimidae
Martin, G.R., 1993, Producing the image; in Zeigler, H.P. and Bischof, H.- from the Belly River Formation of Alberta: University of Toronto
J., eds., Vision, brain, and behavior in birds: MIT Press, Cambridge, p. 5- Studies: Geological Series, v. 26, p. 1-24.
24. Paul, G.S., 1988, Predatory dinosaurs of the world: Simon and Schuster, New
Massal, L., 1953, Premiere note sur le milieu et la croissance des estheries: York, 464 p.
Bulletin Societe Science Naturelle Tunis, v. 6, p. 181-190. Paul, G.S., 1994, Dinosaur reproduction in the fast lane: Implications for
McGowan, C., 1999, A practical guide to vertebrate mechanics: Cambridge size, success, and extinction; in Carpenter, C., Hirsch, K.F. and Horner,
University Press, New York, 301 p. J.R., eds., Dinosaur eggs and babies: Cambridge University Press, New
McFadden, S.A., 1993, Constructing the three-dimensional image; in Zeigler, York, p. 244-255.
H.P. and Bischof, H.-J., eds., Vision, brain, and behavior in birds: The Paul, G.S., 2002, Dinosaurs of the air: Johns Hopkins University Press,
MIT Press, Cambridge, p. 47-62 Baltimore, 460 p.
McIlhenny, E.A., 1935, The alligator’s life history: The Christopher Pub- Pearl, R. and Miner, J.R., 1935, Experimental studies on the duration of
 131
life. XIV. The comparative mortality of certain lower organisms: Quar- Rinehart, L.F., Lucas, S.G., Heckert, A.B. and Hunt, A.P., 2004c, Ostracodes
terly Review of Biology, v. 10, p. 60-79. and conchostracans from the Upper Triassic Whitaker Quarry, Rock
Peck, D.S., 1987, Accelerated testing handbook: Technology Associates, Point Formation, Chinle Group, north-central New Mexico: Geological
Portola, 136 p. Society of America Abstracts with Programs, v. 36, no. 4, p. 29.
Pennak, R.W., 1978, Freshwater Invertebrates of the United States: Wiley, Rinehart, L.F., Lucas, S.G., Heckert, A.B. and Hunt, A.P., 2004d, Vision
New York, 803 p. characteristics of Coelophysis bauri based on sclerotic ring, orbit, and
Pettigrew, J.D., 1986, The evolution of binocular vision; in Pettigrew, J.D., skull morphology: Journal of Vertebrate Paleontology, v. 24, supple-
Sanderson, K. and Levick, W.R., eds., Visual neuroscience: Cambridge ment to no. 3, 104A.
University Press, Cambridge, p. 208-222 Rinehart, L.F., Rasnitsyn, A.P., Lucas, S.G. and Heckert, A.B., 2005c, Instar
Pierce, M.J., Rinehart, L.F., Heckert, A.B. and Hunt, A.P., 2004, Rotation sizes and growth in the Middle Permian monuran Dasyleptus brongniarti
of an Upper Triassic Ghost Ranch, New Mexico Whitaker Quarry block: (Insecta: Machilida: Dasyleptidae): New Mexico Museum of Natural
Turning over a fragile 12,0000 lb rock: Journal of Vertebrate Paleontol- History and Science, Bulletin 30, p. 270-272.
ogy, v. 24, supplement to no. 3, p. 101A. Rinehart, L.F., Lucas, S.G. and Hunt, A.P., 2007, Furculae in the Late
Raath, M.A., 1977, The anatomy of the Triassic theropod Syntarsus Triassic dinosaur Coelophysis bauri: Paläontologische Zeitschrift, v.
rhodesiensis (Saurischia: Podokosauridae) and a consideration of its bi- 81, p. 174-180.
ology [Ph.D. thesis]: Rhodes University, Salisbury, Rhodesia, 233 p. Romer, A.S., 1956, Osteology of the reptiles: University of Chicago Press,
Raath, M.A., 1990, Morphological variation in small theropods and its Chicago, 772 p.
meaning in systematics: Evidence from Syntarsus rhodesiensis; in Car- Rowe, T., 1989, A new species of theropod dinosaur Syntarsus from the
penter, K. and Currie, P.J., eds., Dinosaur systematics, approaches and Early Jurassic Kayenta Formation of Arizona: Journal of Vertebrate
perspectives: Cambridge University Press, Cambridge, p. 91-105. Paleontology, v. 9, p. 125-136.
Rauhut, O.W.M., 2003, The interrelationships and evolution of basal Rowe, T. and Gauthier, J., 1990, Ceratosauria; in Weishampel, D.B., Dodson,
theropod dinosaurs: The Paleontological Association, London: Special P. and Osmolska, H., eds., The Dinosauria: University of California
Papers in Palaeontology, no. 69, p. 1-213. Press, Berkeley, p. 151-168.
Raup, D.M. and Stanley, S.M. 1971, Principles of paleontology: W. H. Russel, L.S., 1939, Sclerotic ring in the skull of Lambeosaurus: Bulletin of
Freeman and Company, San Francisco, 481 p. the Geological Society of America, v. 50, p. 1966.
Reiss, M.J., 1989, The allometry of growth and reproduction: Cambridge Samman, T., 2004, Assessing cervical range of motion in theropods: Journal
University Press, Cambridge, 182 p. of Vertebrate Paleontology, v. 24, supplement to no. 3, 107A.
Renesto, S., Spielmann, J.A., Lucas, S.G. and Spagnoli, G.T., 2010, The Schaeffer, B., 1967, Late Triassic fishes from the western United States:
taxonomy and paleobiology of the Late Triassic (Carnian-Norian: Bulletin of the American Museum of Natural History, v. 135, p. 285-
Adamanian-Apachean) drepanosaurs (Diapsida: Archosauromorpha: 342.
Drepanosauromorpha): New Mexico Museum of Natural History and Schaeffer, B. and McDonald, N. G., 1978, Redfieldiid fishes from the Trias-
Science, Bulletin 46, 80 p. sic-Liassic Newark Supergroup of eastern North America: Bulletin of the
Ricklefs, R.E., 1983, Avian postnatal development; in Farner, D.S., King, American Museum of Natural History, v. 159, p. 4.
J.R. and Parkes, K.C., eds., Avian biology: Academic Press, New York, p. Schmitz, L., 2004, Possibility of thick sclera in ichthyosaurs and its impli-
1-83. cations for visual capability: Journal of Vertebrate Paleontology, v. 24,
Riggs, N.R., Gehrels, G.E. and Dickinson, W.R., 1996, Detrital zircon link supplement to no. 3, 110A.
between headwaters and terminus of the Upper Triassic Chinle-Dockum Schwartz, L., 2009, Quantitative estimates of visual performance features
paleoriver system: Science, v. 273, p. 97-100. in fossil birds: Journal of Morphology, v. 70, p. 759-773.
Rinehart, L.F., 2001, Preliminary statistical analysis defining the juvenile, Schwartz, H.L. and Gillette, D.D., 1994, Geology and taphonomy of the
robust and gracile forms of the Triassic dinosaur Coelophysis: Journal of Coelophysis quarry, Upper Triassic Chinle Fromation, Ghost Ranch,
Vertebrate Paleontology, v. 21, supplement to no. 3, 93A. New Mexico: Journal of Paleontology, v. 68, p. 1118-1130.
Rinehart, L.F. and Lucas, S.G., 2001, A statistical analysis of a growth series Semlitsch, O., 1980, Geographic and local variation in population param-
of the Permian nectridean Diplocaulus magnicornis showing two-stage eters of the slimy salamander Plethodon glutinosus: Herpetologica, v.
ontogeny: Journal of Vertebrate Paleontology, v. 21, p. 803-806. 36, p. 6-16.
Rinehart, L.F., Heckert, A.B., Lucas, S.G. and Hunt, A.P., 2004a, The scle- Senter, P., 2006, Scapular orientation in theropods and basal birds, and the
rotic ring of the Late Triassic dinosaur Coelophysis: New Mexico Geol- origin of flapping flight: Acta Palaeontologica Polonica, v. 51, p. 305-
ogy, v. 26, p. 64. 313.
Rinehart, L.F., Heckert, A.B., Lucas, S.G. and Hunt, A.P., 2004b, Sereno, P.C. and Novas, F.E., 1993, The skull and neck of the basal theropod
Microstratigraphy and fauna of a single block from the Late Triassic Herrerasaurus ischigualastensis: Journal of Vertebrate Paleontology, v.
Whitaker Quarry (Coelophysis) Quarry (Rock Point: Chinle Group), 13, p. 451-476.
north-central New Mexico: New Mexico Geology, v. 26, p. 74. Sillman, A.J., 1973, Avian vision; in Farner, D.S. and King, J.R., eds., Avian
Rinehart, L.F., Hunt, A.P., Lucas, S.G. and Heckert, A.B., 2005a, Coprolites biology, Vol. 3.: Academic Press, New York, p. 349-387.
and cololites from the Late Triassic dinosaur Coelophysis bauri, Whitaker Simpson, G.G., Roe, A. and Lewontin, R.C., 1960, Quantitative zoology:
Quarry, Rio Arriba County, New Mexico: New Mexico Geology, v. 27, p. Harcourt, Brace and World, Inc., New York, 440 p.
53. Small, B.J., 1998, The occurrence of Aetosaurus in the Chinle Formation
Rinehart, L.F., Hunt, A.P., Lucas, S.G., Heckert, A.B. and Smith, J.A., 2005b, (Late Triassic, USA) and its biostratigraphic significance: Neues Jahrbuch
New evidence of cannibalism in the Late Triassic (Apachean) dinosaur, für Geologie und Paläontologie Monatshefte, v. 1998, p. 285-296.
Coelophysis bauri (Theropoda: Ceratosauria): Journal of Vertebrate Small, B. and Downs, A., 2002, A unusual archosauriform from Ghost Ranch,
Paleontology, v. 25, supplement to no. 3, p. 105A. New Mexico: Journal of Vertebrate Paleontology, v. 22, supplement to
Rinehart, L.F., Lucas, S.G. and Heckert, A.B., 2002a, Probability plots no. 3, p. 108A.
distinguish morphological groupings in biometric data: Analysis of Smith, D.K., 1997, Cranial allometry in Coelophysis: preliminary results; in
Coelophysis and comparison to other Diapsida: New Mexico Geology, v. Southwest Paleontology Symposium, Mesa Southwest Museum, p. 41-
24, p. 62. 48.
Rinehart, L.F., Lucas, S.G. and Heckert, A.B., 2002b, Stasis in bivalve growth Smith, D.K. and Merrill, A., 2006, Facial variation in Coelophysis bauri and
and population ecology: Age distribution, growth curves, and biomass of the status of Megapnosaurus (Syntarsus): New Mexico Museum of Natural
a population of Revueltian (Upper Triassic: early-mid Norian) unionids History and Science, Bulletin 37, p. 592-598.
from West Texas: Geological Society of America, Abstracts with Pro- Spielmann, J.A., Lucas, S.G. and Hunt, A.P., 2006, The vertebrate macrofauna
grams, v. 34, p. 354. of the Upper Triassic (Apachean) Redonda Formation, east-central
132
New Mexico: New Mexico Museum of Natural History and Science, Underwood, G.L., 1984, Scleral ossicles of lizards: An exercise in character
Bulletin 37, p. 502-509. analysis: Symposium of the Zoological Society of London, v. 52, p.
Spielmann, J.A., Lucas, S.G., Rinehart, L.F., Hunt, A.P., Heckert, A.B. and 483-502.
Sullivan, R.M., 2007, Oldest records of the Late Triassic theropod dino- Van Valen, L., 2005, The statistics of variation; in Hallgrimsson, B. and
saur Coelophysis bauri: New Mexico Museum of Natural History and Hall, B.K., eds., Variation: A central concept in biology: Elsevier Aca-
Science, Bulletin 41, p. 384-401. demic Press, Amsterdam, p. 29-47.
Spielmann, J.A., Rinehart, L.F., Lucas, S.G., Berman, D.S, Henrici, A.C. and Verhulst, P.F., 1838, Notice sur la loi que la population dans son
Harris, S.K., 2010, Taxonomic revision Sphenacodon, a Late Pennsyl- accroissement: Correspondance Mathematique et Physique, v. 10, p.
vanian-Early Permian eupelycosaur (Sphenacodontidae: 113-121.
Sphenacodontinae) from New Mexico: New Mexico Museum of Natural Visser, G., 2004, Varanus flavescens; in Pianka, E.R., King, D.R. and King,
History, Bulletin 47, in press. R.A., eds., Varanoid lizards of the world: Indiana University Press,
Stewart, J.H., Poole, F.G. and Wilson, R.F., 1972, Stratigraphy and origin of Bloomington, p. 179-183.
the Chinle Formation and related Upper Triassic strata of the Colorado Visser, G.H. and Beintema, A.J., 1991, Reproductive characteristics of meadow
Plateau region: U.S. Geological Survey, Professional Paper 690, 336 p. birds and other European waders: Wader Study Group, Bulletin 61, p. 6-
Stoddart, D.M., 1980, Methods of sex-identification of subfossil bones: 11.
Journal of the Zoological Society of London, v. 192, p. 543-546. Von Huene, F., 1911, Kurze Mitteilung über Perm, Trias und Jura in New
Sullivan, R.M., Lucas, S.G., Heckert, A.B. and Hunt, A.P., 1996, The type Mexico: Neues Jahrbuch für Mineralogie Beilage-Band, v. 32, p. 730-
locality of Coelophysis, a Late Triassic dinosaur from north-central 739.
New Mexico (U.S.A): Paläontologische Zeitschrift, v. 70, p. 245-255. Von Huene, F., 1915, On reptiles of the New Mexican Triassic in the Cope
Swain, F.M. and Brown, P.M., 1972, Lower Cretaceous, ?Jurassic, and Trias- collection: Bulletin of the American Museum of Natural History, v. 34,
sic Ostracoda from the Atlantic costal region: U. S. Geological Survey, p. 485-507.
Professional Paper 795, p. 1-55. Wallman, J., and Letelier, J.-C., 1993, Eye movements, head movements,
Tanner, L.H., 2000, Palustrine-lacustrine and alluvial facies of the (Norian) and gaze stabilization in birds; in Zeigler, H.P. and Bischof, H.-J., eds.,
Owl Rock Formation (Chinle Group) Four Corners region, southwestern Vision, brain, and behavior in birds: The MIT Press, Cambridge, p. 245-
U.S.A: Implications for Late Triassic paleoclimate: Journal of Sedimen- 264.
tary Research, v. 70, p. 1280-1289. Walls, G.L., 1942, The vertebrate eye and its adaptive radiation: The
Tanner, L.H., 2003, Pedogenic features of the Chinle Group, Four Corners Cranbrook Press, Bloomfield Hills, 785 p.
region: Evidence of Late Triassic aridification: New Mexico Geological Webb, J.A., 1979, A reapprasial of the palaeoecology of conchostracans
Society, Guidebook 54, p. 269-280. (Crustacea: Branchiopoda): Neues Jahrbuch für Geologie und
Tasch, P., 1969, Branchiopoda; in Moore, R.C., ed., Treatise on inverte- Paläontologie Abhandlungen, v. 158, p. 259-275.
brate paleontology Part R: The Geological Society of America and The Wellnhofer, P. 1974. Das funfte Skelettexemplar von Archaeopteryx.
University of Kansas, Lawerence, p. R128-R191. Palaeontographica Abteilung A, v. 147, p. 169-215.
Tasch, P., 1973, Paleobiology of the invertebrates: John Wiley and Sons, Welty, J.C., 1963, The life of birds: Alfred A. Knopf, New York, 546 p.
Inc., New York, 946 p. Witmer, L.M., 1997, The evolution of the antorbital cavity of archosaurs:
Terres, J.K., 1991, The Audubon Society encyclopedia of North American A study in soft-tissue reconstruction in the fossil record with an analysis
birds: Alfred A. Knopf, Inc., New York, 1109 p. of the function of pneumaticity: Journal of Vertebrate Paleontology,
Trinh, T.A., Goldstein, B.S. and Demko, T.M., 1996, Sedimentology, stratig- Memoir 3, 73 p.
raphy, and paleontology of the Chinle Formation, Ghost Ranch, New Yates, A.M. and Vasconcelos, C.C., 2005, Furcula-like clavicles in the
Mexico: Paleoenvironmental interpretation: Geological Society of prosauropod dinosaur Massospondylus: Journal of Vertebrate Paleontol-
America Abstracts with Programs, v. 77, p. 118-119. ogy, v. 25, p. 466-468.
Tykoski, R.S., 1998, The osteology of Syntarsus kayentakatae and its Zar, J.H., 1999, Biostatistical analysis: Pearson Education, Singapore, 663
implications for Ceratosaurid phylogeny [M.S. thesis]: University of p.
Texas at Austin, Austin, 217 p. Zeigler, K.E. and Geissman, J.W., 2009, Examining new possibilities for the
Tykoski, R.S. and Rowe, T., 2004, Ceratosauria; in Weishampel, D.B., age of Upper Triassic strata in northern New Mexico using
Dodson, P. and Osmolska, H., eds., The Dinosauria: University of Cali- magnetostratigraphy: New Mexico Geology, v. 31, p. 42.
fornia Press, Berkeley, p. 47-70. Zeigler, K.E., Kelley, S. and Geissmann, J.W., 2008, Revisions to strati-
Tykoski, R.S., Forster, C.A., Rowe, T., Sampson, S.D. and Munyikwa, D., graphic nomenclature of the Upper Triassic Chinle Group in New Mexico:
2002, A furcula in the coelophysid theropod Syntarsus: Journal of Verte- New insights from geologic mapping, sedimentology, and
brate Paleontology, v. 22, p. 728-733. magnetostratigraphic/paleomagnetic data: Rocky Mountain Geology, v.
Underwood, G.L., 1970, The eye; in Gans, C. and Parsons, T.S., eds., Biol- 43, p. 121-141.
ogy of the Reptilia, Vol. 2: Morphology B: Academic Press, London, p. Zug, G.R. 1993. Herpetology: An introductory biology of amphibians and
1-93. reptiles: Academic Press, Inc., New York, 527 p.
 Rinehart et al., 2009, The Paleobiology of Coelophysis bauri. New Mexico Museum of Natural History and Science, Bulletin 45.

133
APPENDIX 1

All 121 datasheets for Coelophysis bauri osteological measurements used in this study from various institutions with Whitaker Quarry
blocks.
SKULL AND MANDIBLE MEASUREMENT DEFINITIONS FOR THE DATASHEETS:

A-B anterior tip of premaxilla to anterior naris boundary

A-C anterior tip of premaxilla to anterior antorbital fenestra boundary

A-D anterior tip of premaxilla to anterior orbit boundary

A-E anterior tip of premaxilla to anterior lateral temporal fenestra boundary

A-F anterior tip of premaxilla to anterior supratemproal fenestra boundary

A-G anterior tip of premaxilla to posteriormost part of skull (i.e., max length)

PA-PB anterior tip of premaxilla to anterior internal naris boundary

PA-PC anterior tip of premaxilla to anterior palatine fenestra boundary

PA-PD anterior tip of premaxilla to anteriormost portion of the basisphenoid

PA-PE anterior tip of premaxilla to anterior quadrate surface (upper jaw joint)

MA-MB anterior tip of mandible to anterior mandibular fenestra boundary

MA-MC anterior tip of mandible to posterior tip of retroarticular process (max length)
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
 255
APPENDIX 2

Measurements and descriptive statistics of valves of Shipingia n. sp. from NMMNH Ghost Ranch Whitaker Quarry block C-8-82.
256
APPENDIX 3

Test for statistically significant difference between the heights of two proposed cohorts of Shipingia from the Whitaker Quarry. Holm’s
multiple comparison is used because it is the most accurate ANOVA test where the populations of the groups are unequal (Glantz, 2005). The heights
of the two groups are significantly different at the 99% probability level.
 257
APPENDIX 4

All study animals and metrics. Dimensions are in mm, areas in mm2, and volumes in mm3.
 258

APPENDIX 5

Large element measurements of 18 specimens comprising a growth series.

 259
APPENDIX 6

Tabulated survival data.

260
APPENDIX 7

Compare Two Regressions – Coelophysis bauri skull length versus dorsal series length: are there two significantly distinct regression lines
or a single line?

APPENDIX 8

Compare Two Regressions – Coelophysis bauri femur head width versus femur length: are there two significantly distinct regression lines
or a single line?