Bulletin 49

New Mexico Museum of Natural History & Science

A Division of the

DEPARTMENT OF CULTURAL AFFAIRS

Carboniferous-Permian transition in
Cañon del Cobre, northern New Mexico

edited by
Spencer G. Lucas, Jörg W. Schneider
and Justin A. Spielmann

Albuquerque, 2010
Carboniferous-Permian
J.A.
Spielmann,
and
J.W.
Schneider,
S.G.,
Lucas,
49
Bulletin
NMMNH&S
Mexico
New
northern
Cobre,
del
Cañon
in
transition
 Bulletin 49

New Mexico Museum of Natural History & Science

A Division of the
DEPARTMENT OF CULTURAL AFFAIRS

Carboniferous-Permian transition in
Cañon del Cobre, northern New Mexico

edited by
Spencer G. Lucas, Jörg W. Schneider
and Justin A. Spielmann

Albuquerque, 2010
 Bulletin 49

New Mexico Museum of Natural History & Science

A Division of the
DEPARTMENT OF CULTURAL AFFAIRS

Carboniferous-Permian transition in
Cañon del Cobre, northern New Mexico

by
Spencer G. Lucas, Jörg W. Schneider
and Justin A. Spielmann

New Mexico Museum of Natural History & Science

Printed with the support of the

Vortman Fund

Albuquerque, 2010
 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: Cast of holotype skeleton and flesh restoration of the diadectomorph Limnoscelis
paludis from the Upper Pennsylvanian of Cañon del Cobre, New Mexico. Graphics and illustration by
Mark A. Klingler, Carnegie Museum of Natural History. See Bermanet al., this volume.

Original Printing

ISSN:1524-4156

Available from the New Mexico Museum of Natural History and Science, 1801 Mountain Road NW,
Albuquerque, NM87104; Telephone (505) 841-2800; Fax (505)841-2866; www.nmnaturalhistory.org

NMMNH Bulletinsonline 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

GuillermoAlvarado Asociación Costarricense de Geotecnica, San José, Costa Rica

AndreaArcucci 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
Brent Breithaupt Laramie, WY, 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 Colorado at Denver, Denver, CO, USA
of
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 Studidell’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
 NEWMEXICO MUSEUM OF NATURAL HISTORYAND SCIENCE BULLETINS
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.
45. The Paleobiology of Coelophysis bauri (Cope) from the Upper Triassic (Apachean) Whitaker quarry, New Mexico, with
detailed analysis of a single quarry block, 2009. by Larry F. Rinehart, Spencer G. Lucas, Andrew B. Heckert, Justin A.
Spielmann and Matthew D. Celeskey, 260 pp.
46. The taxonomy and paleobiology of the Late Triassic (Carnian-Norian: Adamanian-Apachean) drepanosaurs
(Diapsida: Archosauromorpha: Drepanosauromorpha, 2010. by Silvio Renesto, Justin A. Spielmann, Spencer G. Lucas and
Giorgio Tarditi Spagnoli, 81 pp.
47. Ichnology of the Upper Triassic (Apachean) Redonda Formation, east-central New Mexico, 2010. by Spencer G. Lucas,
Justin A. Spielmann, Hendrik Klein and Allan J Lerner, 75 pp.
48. New Smithian (Early Triassic) ammonoids from Crittenden Springs, Elko County, Nevada: Implications for
taxonomy, biostratigraphy and biogeography, 2010. by James F. Jenks, Arnaud Brayard, Thomas Brühwiler and Hugo Bucher,
41 pp.
 Lucas et al., eds., 2010, Carb-Permian transition in Cañon del Cobre. New Mexico Museum of Natural History and Science Bulletin 49.

i
TABLE OF CONTENTS

Introduction

Carboniferous-Permian transition in Cañon del Cobre, northern New Mexico:

An overview.......................................................Spencer G. Lucas, Jörg W. Schneider and JustinA. Spielmann 1

One hundred and fifty years of geological and paleontological research in Cañon del Cobre, northern New
Mexico..................................................................Spencer G. Lucas, Susan K. Harris and Justin A. Spielmann 7

Geology

Summary of geology of Cañon del Cobre, Rio Arriba County, New Mexico......................Spencer G. Lucas,
Karl Krainer and JustinA. Spielmann 15

Sedimentology of the Pennsylvanian-Permian Cutler Group and Lower Permian Abo Formation, northern
New Mexico..................................................................................................Karl Krainer and Spencer Lucas 25

Ichnology

Pennsylvanian lake-margin ichnoassemblage from Cañon del Cobre, Rio Arriba County,
New Mexico...........................................................................................Spencer G. Lucas and Allan JLerner 37

Euramerican Late Pennsylvanian/Early Permian arthropleurid/tetrapod associations - implications for the

habitat and paleobiology of the largest terrestrial arthropod...........Jörg W. Schneider, Spencer G. Lucas,
Ralf Werneburg and Ronny Rößler 49

Palynology/Paleobotany

Palynological investigation of the Upper Pennsylvanian (Carboniferous) El Cobre Canyon Formation, Cutler
Group, Cañon del Cobre, RioArriba County, New Mexico, U.S.A..........John Uttingand Spencer G. Lucas 71

Late Pennsylvanian floras in western equatorial Pangea, Cañon del Cobre,

New Mexico..................................WilliamA. Dimichele, Dan S. Chaney, Hans Kerp and Spencer G. Lucas 75

Vertebrate Paleontology

Vertebrate paleontology, biostratigraphy and biochronology of the Pennsylvanian-Permian Cutler Group,

Cañon del Cobre, northern New Mexico.............…...............Spencer G. Lucas, Susan K. Harris, Justin A.
Spielmann, Larry F. Rinehart, David SBerman,Amy C. Henriciand Karl Krainer 115

Dissorophoid record from the Upper Pennsylvanian of Cañon del Cobre, New Mexico............Ralf Werneburg,
David SBerman and Spencer G. Lucas 125

First Pennsylvanian Eryops (Temnospondyli) and its Permian recordfrom New Mexico............RalfWerneburg,
Spencer G. Lucas, Jörg W. Schneider and Larry F. Rinehart 129
ii
A partial skeleton of Ophiacodon navajovicus (Eupelycosauria: Ophiacodontidae) from the Upper
Pennsylvanian ofCañondel Cobre, New Mexico....Susan K. Harris,SpencerG.Lucas and JustinA. Spielmann 137

Re-evalutaion of Ruthiromia elcobriensis (Eupelycosauria: Ophiacodontidae?) from the Lower Permian

(Seymouran?) of Cañon del Cobre, north-central New Mexico....JustinA. Spielmann and Spencer G. Lucas 151

Redescription of the cranial anatomy of Sphenacodon ferox Marsh (Eupelycosauria: Sphenacodontidae)

from the Late Pennsylvanian-Early Permian ofNew Mexico....................JustinA. Spielmann,Larry F.Rinehart,
Spencer G. Lucas, David SBerman, Amy C. Henrici and Susan K. Harris 159

Redescription of the skull of Limnoscelis paludis Williston (Diadectomorpha:

Limnoscelidae)…..............................................................David SBerman, Robert R. Reisz and Diane Scott 185

Description of the postcranial skeleton of Limnoscelis (Diadectomorpha: Limnoscelidae) from the Upper
Pennsylvanian of northern New Mexico and central Colorado…….....................................Natalia K. Kennedy 211

Typothorax coccinarum (Archosauria: Stagonolepididae) from the Upper Triassic (Revueltian) Petrified
Forest Formation, El Puertocito, Cañon del Cobre, RioArriba County, New Mexico...........JustinA.Spielmann
and Spencer G. Lucas 221
 Lucas et al., eds., 2010, Carb-Permian transition in Cañon del Cobre. New Mexico Museum of Natural History and Science Bulletin 49.

1
CARBONIFEROUS-PERMIAN TRANSITION IN CAÑON DEL COBRE,
NORTHERN NEWMEXICO: ANOVERVIEW

SPENCER G. LUCAS1, JÖRGW. SCHNEIDER2 AND JUSTINA. SPIELMANN1

1New Mexico Museum of Natural History and Science, 1801 Mountain Road N.W., Albuquerque, New Mexico 87104;
2TU Bergakademie Freiberg, Institut für Geologie, B.v.Cotta-Strasse 2, D-09596 Freiberg, Germany

Abstract—Cañon del Cobre is a large box canyon in southeastern Rio Arriba County, New Mexico, along the
southeastern border of the Colorado Plateau. The canyon exposes an ~400 m thick section of Upper Pennsylva
nian-Lower Permian fluvial red beds (El Cobre Canyon and Arroyo del Agua formations of the Cutler Group) that
accumulated under wet-subhumid to semiarid climatic conditions. These beds yield palynomorphs, megafossil
plants, trace fossils and vertebrate fossils. The paleofloral and paleofaunal components are the main focus of the
articles in this volume. The red-bed succession across the Carboniferous-Permian boundary in Cañon del Cobre
documents increasing aridity that caused a slight drop in paleofloral diversity and a change in composition to
greater abudance of conifers in the paleoflora. These changes, however, are not mirrored by tetrapod assemblages,
which do not change substantially across the Carboniferous-Permian boundary.

INTRODUCTION
Cañon del Cobre is a large box canyon in southeastern Rio Arriba
County, New Mexico, along the southeastern border of the Colorado
Plateau. The main portion of the canyon is a complexly faulted valley
with extensive exposures of the El Cobre Canyon and overlying Arroyo
del Agua formations of the Carboniferous-Permian Cutler Group (Lucas
and Krainer, 2005; Lucas et al., 2005b; Kempter et al., 2007). These
Cutler Group strata were deposited by syntectonic fluvial systems in
relatively upland and inland settings along the southern flank of the
Uncompahgre uplift of the ancestral Rocky Mountains (Fig.1).
This volume documents geological and paleontological research in
Cañon del Cobre, most of which took place during the last decade as a
collaborative project coordinated by the New Mexico Museum of Natu
ral History and Science. This research uncovered numerous new fossils
in the canyon, especially invertebrate traces, palynomorphs, megafossil
plants and vertebrates. Some of these fossils have provided the basis for
reasonably certain (but not exact) placement of the Carboniferous-Per
mian boundary in the red-bed section exposed at Cañon del Cobre. Also,
these fossils and their enclosing sediments document relatively little
biotic change across the system boundary. In this article, we provide an
introduction to and overview of the articles in this volume and summa
rize some of the trends across the Carboniferous-Permian boundary
evident in Cañon del Cobre.
HISTORY OF STUDY
In this volume, Lucas, Harris and Spielmann recount the his-
tory of geological and paleontological studies in Canon del Cobre, which
began with J.S. Newberry in 1859, who discovered Triassic plant fossils
in the copper mines in the canyon (Newberry, 1876). David Baldwin
first discovered Carboniferous-Permian tetrapod fossils in Cañon del
Cobre in the late 1870s, and this led to subsequent searches for Carbon
iferous-Permian tetrapod fossils. S.W. Williston and E.C. Case, working
in 1911, first recognized that the oldest tetrapod fossils from Cañon del
Cobre are of Carboniferous age (Williston and Case, 1912), and subse
quent research has supported this conclusion (e.g., Vaughn, 1963; Fracasso,
1980; Hunt and Lucas, 1992; Lucas et al., 2005b) (Fig. 2). However,
most of the bedrock strata in Cañon del Cobre were assigned a Triassic
age until 1961, when C. Smith, A. Budding and C.W. Pitratassigned them
to the Permian Cutler Formation. Collecting and research in Cañon del
Cobre has now identified invertebrate and vertebrate traces,
palynomorphs, plants and tetrapod bones and teeth of Carboniferous
age, traces and tetrapods of Early Permian age and Late Triassic tetrapod
and plant fossils. Much of this volume focuses on documenting these
fossils.
GEOLOGY
Lucas, Spielmann and Krainer here review the geology of Cañon
del Cobre. The canyon is traversed and bounded by several north-north
east-trending faults, and its eastern wall is along a faulted, north-north
east-trending anticline, structures that were activated by Neogene exten
sional tectonism of the nearby Rio Grande rift. The oldest rocks that
floor Cañon del Cobre are siliciclastic red beds of the El Cobre Canyon
Formation of the Cutler Group (Fig. 2). An extensive unconformity
separates Cutler Group strata from overlying Upper Triassic strata of
the Chinle Group around the rim of Cañon del Cobre (in this volume,
Spielmann and Lucas document aetosaur fossils collected from Chinle
strata near the mouth of Cañon del Cobre). Cenozoic bedrock of Eocene
and Oligocene age is locally exposed around the rim of Cañon del Cobre.
Quaternary deposits – quartzite erratic boulders, alluvial terrace depos
its, colluvium and alluvium – crop out in and around Cañon del Cobre.
In this volume, Krainer and Lucas review the sedimentology of
Cutler Group strata in Cañon del Cobre, confirming observations of
earlier studies by Eberth (see especially Eberth and Miall, 1991). Thus,
the El Cobre Canyon and Arroyo del Agua formations are red beds that
formed in river systems under semi-arid to arid climatic conditions. Sand
stones are dominantly arkoses to litharenites, and rare sublitharenites.
Sandstone sheets represent deposits of a braided river system; the inter
calated sandstone beds and lenses are interpreted as sheet splays and
minor channel fills that formed during overbank flooding. Siltstone and
mudstone units represent floodplain deposits. The exposed El Cobre
Canyon Formation is about 230 m thick (total thickness ~ 640 m) and is
mostly siltstone and sandstone deposited in an ephemeral braided stream
environment. The overlying Early Permian Arroyo del Agua Formation
is as much as 160 m thick, and also consists mostly of siltstone and
sandstone, but was deposited on laterally extensive floodplains that
developed between relatively stable channels.
ICHNOLOGY
Prior to the last decade, no trace fossils were reported from Cañon
del Cobre. Recent discoveries have highlighted two localized records in
the canyon—a lake margin invertebrate and vertebrate ichnoassemblage
and a site with large Diplichnites, which are the trackways of the largest
known terrestrial arthropod, Arthropleura.
In Cañon del Cobre, rare lacustrine strata (thinly laminated silt
stones) of the El Cobre Canyon Formation represent shallow, localized
floodplain lakes. In this volume, Lucas and Lerner document the first
ichnofossil assemblage known from such a Pennsylvanian intermontane
lake deposit, and it consists of arthropod trackways, invertebrate graz
ing and feeding traces, tetrapod footprints and fish swimming trails. This
2
FIGURE 1. Principal uplifts (gray), basins and faults associated with the late
Paleozoic ancestral Rocky Mountains orogeny and location of Cañon del
Cobre in New Mexico (after Woodward et al., 1999). Faults are: A =
Albuquerque fault, E = Estancia fault, J = Joyita fault, P = Peñasco fault, PP
= Pecos-Picuris fault, TC = Tijeras-Cañoncito fault and TP = Tusas-Picuris
fault.
ichnofossil assemblage is assigned to the Scoyenia ichnofacies and is
composed of members of the Diplichnites and Mermia ichnoguilds, con
sistent with a shallow lacustrine setting. The relatively low ichnodiversity
of the Cañon del Cobre ichnofossil assemblage indicates a short-lived
water body colonized by a mobile invertebrate epifauna of detritus feed
ers and predators that provided a food source for tetrapod predators.
In this volume, Schneider, Lucas, Werneburg and Rössler re
view the record of the giant arthropod Arthropleura, which was a com
mon member of the late Paleozoic continental biota of paleo-equatorial
biomes for more than 35 million years, from the middle Mississippian to
the Early Permian. In Late Pennsylvanian red beds in Cañon del Cobre,
trackways of Arthropleura are present in strata that also yield body
fossils of the amphibian Eryops (Lucas et al., 2005a). A review of the
Arthropleura tracksite from Cañon del Cobre, as well as other tracksites
of this animal and arthropleurid/eryopid associations, supports the con
clusion that Arthropleura was well adapted to alluvial environments of
ever wet humid to seasonally dry and semihumid climates. Preferred
habitats of semi-adult and adult Arthropleura were open, vegetated, river
landscapes, where these giant arthropods co-existed with semi-aquatic
eryopid amphibians and terrestrial pelycosaurs.
PALEOBOTANY
Smith et al. (1961) first reported Carboniferous plants from Cañon
del Cobre. Their locality, which is in the middle part of the El Cobre
Canyon Formation, also yields palynomorphs reported on by Utting
and Lucas in this volume. These are mostly miospores of low thermal
maturity and indicate a Late Pennsylvanian (Carboniferous) age not older
than Stephanian B. Fracasso (1980; also see Hunt and Lucas, 1992) also
reported a Pennsylvanian megafossil plant locality in the floor of Cañon
del Cobre.
During the last decade, an extensive search for fossil plants of
Carboniferous-Permian age in Cañon del Cobre identified localities at
five stratigraphic levels in the middle of the El Cobre Canyon Formation.
In this volume, DiMichele, Chaney, Kerp and Lucas document these
floras, which are compositionally typical of the Late Pennsylvanian,
specifically of the Virgilian (Gzhelian). Conspicuous elements are
Alethopteris zeilleri, Macroneuropteris scheuchzeri, Danaeites emersonii,
Sphenophyllum verticillatum, S. oblongifolium, and S. angustifolium,
Annularia carinata, Asterophyllites equisetiformis, Sigillaria brardii, and
other, less common forms typical of wetland floras, such as
Pseudomariopteris cordato-ovata. Walchian conifers and Taeniopteris
are rare. These floras mostly appear to have been growing on bars or
along the margins of braided streams, resulting in frequent burial of plants,
particularly calamite stems, in upright position, or the preservation of
tangled masses of sphenophyll stems and leaves, largely in siltstones.
Finer-grained deposits, seemingly representing small lakes, were strongly
dominated by pteridosperms.
VERTEBRATE PALEONTOLOGY
In the late 1870s, David Baldwin discovered fossil vertebrates in
the Carboniferous-Permian red beds of Cañon del Cobre, and since then
collecting has yielded unique assemblages of fossil vertebrates that are
among the oldest tetrapod fossils from North America. In this volume,
Lucas, Harris, Spielmann, Rinehart, Berman and Henrici review
and update Lucas et al. (2005b), which presented a detailed vertebrate
biostratigraphy in Cañon del Cobre constructed from all localities that
can be placed into precise lithostratigraphic position (Fig. 2). This bio
stratigraphy identifies three, stratigraphically-successive vertebrate as
semblages that support recognition of three time-successive land verte
brate faunachrons (LVFs) across the Pennsylvanian-Permian boundary:
Cobrean, Coyotean and Seymouran (also see Lucas, 2006). The Penn
sylvanian-Permian boundary is in the Coyotean LVF.
The other articles on vertebrate paleontology in this volume focus
on describing new fossils, documenting new anatomy and interpreting
these fossils within a taxonomic and/or biostratigraphic context. Thus,
Werneburg, Berman and Lucas describe the snout region of a large
dissorophoid skull from the Late Pennsylvanian interval of the El Cobre
Canyon Formation. Werneburg, Lucas, Schneider and Rinehart docu
ment a well-preserved, partial skull of the temnospondyl amphibian
Eryops from the Pennsylvanian strata of the El Cobre Canyon Forma
tion. This Pennsylvanian Eryops is compared to Permian Eryops, includ
ing specimens from the Early Permian of New Mexico, but no species
assignment is possible.
The most famous vertebrate fossil from Cañon del Cobre is the
type specimen of Limnoscelis paludis, discovered by David Baldwin in
1879 and first described by Williston (1911). Long considered pivotal to
understanding reptile (amniote) origins, Limnoscelis is now considered a
diadectomorph. The original type specimen from Cañon del Cobre re
mains the best preserved and most complete fossil known of the taxon.
In this volume, Berman provides a detailed redescription of the skull
that corrects errors in previous descriptions and adds substantially to
knowledge of the cranial anatomy of Limnoscelis. Kennedy provides a
companion piece that redescribes the postcranial anatomy of Limnoscelis,
largely based on the superb skeleton that David Baldwin collected in
Cañon del Cobre.
A recently discovered partial skeleton of the pelycosaurian-grade
synapsid Ophiacodon navajovicus, the most complete yet found, from
the Late Pennsylvanian interval of the El Cobre Canyon Formation, is
described by Harris, Lucas and Spielmann. Its stratigraphic distribu
tion in New Mexico, Utah and Colorado indicate that O. navajovicus
 3

FIGURE 2. Composite lithostratigraphic section in Cañon del Cobre, showing distribution of fossil localities (numbers are localities of the New Mexico
Museum of Natural History and Science) and taxa.
4
characterizes the Cobrean LVF.
Ruthiromia elcobriensis is an eupelycosaur from the Permian strata
of the Cañon del Cobre known only from the holotype, a partial skel
eton. The type locality of R. elcobriensis has never been rediscovered.
However, given available information, Spielmann and Lucas conclude
it is in the lower part of the Arroyo del Agua Formation. Their reconsid
eration of its anatomy places R. elcobriensis outside of the Varanopsidae
and tentatively within the Ophiacodontidae.
Sphenacodon is a pelycosaurian-grade synapsid, best known from
the Early Permian of Rio Arriba County, northern New Mexico. Of the
two species (S. ferox and S. ferocior), S. ferox is known from compara
tively little skull material. Here, Spielmann, Rinehart, Lucas, Berman,
Henrici and Harris describe a newly collected, nearly complete skull of
S. ferox that demonstrates numerous cranial differences between the two
species of Sphenacodon. These features, and metric analysis, support
the concept of two distinct species of Sphenacodon. S. ferox has a tem
poral range from the Late Pennsylvanian (Coyotean LVF) through the
Early Permian (Seymouran LVF), whereas S. ferocior is restricted to the
Coyotean LVF, but does span the Pennsylvanian-Permian boundary.
CARBONIFEROUS-PERMIAN TRANSITION
Megafossil plants, palynomorphs and fossil vertebrates indicate
that the Pennsylvanian-Permian boundary (younger than the Virgilian
Wolfcampian boundary) is stratigraphically high in the El Cobre Canyon
Formation in Cañon del Cobre, though the exact position of the bound
ary is not certain (Fig. 2). Indeed, without biostratigraphic data based on
phylogenetic lineages of successive zone-species, as described by
Werneburg and Schneider (2006) for amphibians, for insects as reported
by Schneider and Werneburg (2006) and supported by isotopic ages
(Roscher and Schneider 2005, Lützner et al. 2007), determining the exact
placement of the Pennsylvanian/Permian boundary in nonmarine set
tings, such as Cañon del Cobre, will remain difficult.
The fossil record of the Carboniferous-Permian transition in Cañon
del Cobre is more extensive for the Pennsylvanian portion of the section
than for the Early Permian (Fig. 2). Fortunately, a diverse Early Permian
record of plants and vertebrates is known from the upper part of the El
Cobre Canyon Formation at Arroyo del Agua, about 25 km southwest of
Cañon del Cobre (e.g., DiMichele and Chaney, 2005; Lucas et al., 2005c).
Using the Arroyo del Agua record to enhance the Early Permian record,
the following trends (or lack of trends) are evident across the Carbonifer
ous-Permian boundary at Cañon del Cobre (Fig. 3):
1. There is a change in the architecture of the fluvially-deposited
red beds from braided systems of the Pennsylvanian to more
laterally extensive floodplains (with mature caliches) of the
Early Permian. Eberth and Miall (1991) attributed this change
to increasing aridity. The presence of a thick section of eolianites
at the base of the Yeso Group (DeChelly Sandstone) immedi
ately above the Arroyo del Agua Formation in the Arroyo del
Agua area is evidence of even increased aridity by approxi
mately the beginning of Leonardian time.
2. The paleoflora declines in diversity from the Pennsylvanian to
the Permian (though the species pool for both time periods is
substantially unchanged), a trend also seen in the co-eval north
Texas section (DiMichele and Chaney, 2005) (Fig. 3).
3. As shown by Roscher and Schneider (2006), the generally
increasing aridity during the Permian reflects a cyclical
sequence of wet and dry phases in which each subsequent
phase progressively becomes dryer. Therefore, the paleoflora
of the red beds in Cañon del Cobre is very similar in composi
tion to the Late Pennsylvanian gray facies (particularly of the
FIGURE 3. Summary of key trends (and non-trends) during the Carboniferous
Permian transition recorded in the Cutler Group strata of Cañon del Cobre
and nearby outcrops in Rio Arriba County, New Mexico.
Virgilian/Gzhelian). During the Late Pennsylvanian-Early Per
mian, each successive paleoflora in the Euramerican paleo
equatorial belt was increasingly dominated by mesophilous
and xerophilous elements, especially by an increase in conifers
(“walchians”) (Fig. 3). The rarity of walchians in the red
beds of Cañon del Cobre points to a relatively wet-subhumid
to perhaps dry-subhumid climate, most similar to the red
beds of the European Stephanian B (Roscher and Schneider,
2005, 2006). Nevertheless, it should be taken into account that
regional orographic situations can modify this picture. Thus,
for example, at the southern border of the Late Pennsylvanian
Hercynian low mountain belt, in the Mauretanides of
Morocco, in addition to the typical hygrophilous Stephanian
paleofloral elements, the mesophilous to xerophilous walchians
and callipterids are very common as early as the Stephanian B
gray facies of the Moroccan Souss basin (Hmich et al. 2006).
Such orographic effects, however, apparently did not influ
ence paleofloral composition across the Carboniferous-Per
mian transition in northern New Mexico.
4. The terrestrial tetrapod assemblages remain substantially
unchanged from Late Pennsylvanian to Early Permian - they
represent what can be termed the Coyotean chronofauna domi
nated by eryopid temnospondyls, diadectomorphs and
sphenacodontid pelycosaurs (Lucas, 2006) (Fig. 3). In con
trast, for the environment of aquatic animals the cyclical in
crease in aridity resulted in a directed development of lake and
river types with associated aquatic invertebrate and vertebrate
assemblages. The aquatic lake faunas of the European Penn
sylvanian/Permian transition thus are characterized by a strong
decline in diversity caused by climate changes and tectonic
reorganization of basins and river systems (Franconian vol
cano-tectonic event: Schneider et al. 1995, 2000). This signal
has so far not been reported from the Ouachita-Appalachian
region, and is also not evident in the ancestral Rocky Moun
tain basin of northern New Mexico.
The red-bed succession across the Carboniferous-Permian bound
ary in Cañon del Cobre thus documents increasing aridity that caused a
slight drop in paleofloral diversity and a change in composition to a more
conifer-dominated paleoflora. These changes, however, are not mirrored
by tetrapod assemblages, which do not change substantially across the
Carboniferous-Permian boundary.
ACKNOWLEDGMENTS
We are grateful to numerous colleagues and volunteers for collabo
ration in the field in Cañon del Cobre. The U. S. Forest Service permitted
our fieldwork in Cañon del Cobre. William DiMichele and Karl Krainer
provided helpful comments on an earlier draft of the manuscript.

REFERENCES
DiMichele, W.A. and Chaney, D.S., 2005, Pennsylvanian-Permian fossil
floras from the Cutler Group, Cañon del Cobre and Arroyo del Agua
areas, in northern New Mexico: New Mexico Museum of Natural His-
tory and Science, Bulletin 31, p. 26-33.
Eberth, D.A. and Miall, A.D., 1991, Stratigraphy, sedimentology and evolu
tion of a vertebrate-bearing, braided to anastomosed fluvial system,
Cutler Formation (Permian-Pennsylvanian), north-central New Mexico:
Sedimentary Geology, v. 72, p. 225-252.
Fracasso, M.A., 1980, Age of the Permo-Carboniferous Cutler Formation
vertebrate fauna from El Cobre Canyon, New Mexico: Journal of Pale
ontology, v. 54, p. 1237-1244.
Hunt, A.P. and Lucas, S.G., 1992, The paleoflora of the lower Cutler Forma
tion (Pennsylvanian, Desmoinesian?) in El Cobre Canyon, New Mexico,
and its biochronological significance: New Mexico Geological Society,
Guidebook 43, p. 145-150.
Hmich, D., Schneider, J.W., Saber, H., Voigt, S. and El Wartiti, M., 2006,
New continental Carboniferous and Permian faunas of Morocco – im-
plications for biostratigraphy, palaeobiogeography and palaeoclimate;
in Lucas, S.G., Cassinis, G. and Schneider J.W., eds., Non-marine Per
mian biostratigraphy and biochronology: Geological Society, London,
Special Publication 265, p. 297-324.
Kempter, K., Zeigler, K., Koning, D. and Lucas, S., 2007, Preliminary
geologic map of the Canjilon SE quadrangle, Rio Arriba County, New
Mexico: New Mexico Bureau of Geology and Mineral Resources, Open
file Geological Map 150, 1:24,000.
Lucas, S.G., 2006, Global Permian tetrapod biostratigraphy and
biochronology; in Lucas, S.G., Cassinis, G. and Schneider, J.W., eds.,
Non-marine Permian biostratigraphy and biochronology: Geological
Society, London, Special Publication 265, p. 65-93.
Lucas, S.G. and Krainer, K., 2005, Stratigraphy and correlation of the
Permo-Carboniferous Cutler Group, Chama Basin, New Mexico: New
Mexico Geological Society, Guidebook 56, p. 145-159.
Lucas, S.G., Lerner, A.J, Hannibal, J.T., Hunt, A.P. and Schneider, J.W.,
2005a, Trackway of a giant Arthropleura from the Upper Pennsylva
nian of El Cobre Canyon, New Mexico: New Mexico Geological Society,
Guidebook 56, p. 279-282.
Lucas, S.G., Harris, S.K., Spielmann, J.A., Berman, D.S and Henrici, A.C.,
2005b, Vertebrate biostratigraphy and biochronology of the Pennsylva
nian-Permian Cutler Group, El Cobre Canyon, northern New Mexico:
New Mexico Museum of Natural History and Science, Bulletin 31, p.
128-139.
Lucas, S.G., Harris, S.K., Spielmann, J.A., Berman, D.S, Henrici, A.C.,
Heckert, A.B., Zeigler, K.E. and Rinehart, L.F., 2005c, Early Permian
vertebrate assemblage and its biostratigraphic significance, Arroyo del
Agua, Rio Arriba County, New Mexico: New Mexico Geological Society,
Guidebook 56, p. 288-296.
5
Lützner, H., Littmann, S., Mädler, J., Romer, R.L. and Schneider, J.W.,
2007, Stratigraphic and radiometric age data for the continental
Permocarboniferous reference-section Thüringer-Wald, Germany: Pro
ceedings XVth International Congress on Carboniferous and Permian
Stratigraphy 2003, Utrecht, p. 161-174.
Newberry, J.S., 1876, Geological report; in Macomb, J.N., ed., Report of the
exploring expedition from Santa Fe…in 1859: U.S. Army Engineer
Department, Washington, D.C., p. 9-118.
Roscher, M. and Schneider, J.W., 2005, An annotated correlation chart for
continental Late Pennsylvanian and Permian basins and the marine
scale: New Mexico Museum of Natural History and Science, Bulletin 30,
p. 282-291
Roscher, M. and Schneider, J.W., 2006, Early Pennsylvanian to Late Per
mian climatic development of central Europe in a regional and global
context; in Lucas, S.G., Cassinis, G. and Schneider, J.W., eds., Nonma
rine Permian chronology and correlation: Geological Society, London,
Special Publication 265, p. 95-136.
Schneider, J. W. and Werneburg, R., 2006, Insect biostratigraphy of the
European Late Carboniferous and Early Permian; in Lucas, S.G., Cassinis,
G. and Schneider J.W., eds., Non-marine Permian biostratigraphy and
biochronology: Geological Society, London, Special Publication 265, p.
325-336.
Schneider, J. W., Rössler, R. and Gaitzsch, B., 1995, Time lines of the late
Variscan volcanism - holostratigraphic synthesis: Zentralblatt für
Geologie und Paläontologie Teil I, v. 5/6, 477-490.
Schneider, J.W., Hampe, O. and Soler-Gijón, R., 2000, The Late Carbonif
erous and Permian: Aquatic vertebrate zonation in southern Spain and
German basins.-IGCP 328. Final Report: Courier Forschung-Institut
Senckenberg, p. 543-561.
Smith, C. T., Budding, A. J. and Pitrat, C. W., 1961, Geology of the south
eastern part of the Chama Basin: New Mexico Bureau of Mines and
Mineral Resources, Bulletin 75, 57 p
Vaughn, P.P., 1963, The age and locality of the late Paleozoic vertebrates
from El Cobre Canyon, Rio Arriba County, New Mexico: Journal of
Paleontology, v. 37, p. 283-296.
Werneburg, R. and Schneider, J.W., 2006, Amphibian biostratigraphy of the
European Permo-Carboniferous; in Lucas, S.G., Cassinis, G. and Schneider
J.W., eds., Non-marine Permian biostratigraphy and biochronology:
Geological Society, London, Special Publication 265, p. 201-215
Williston, S.W., 1911, A new family of reptiles from the Permian of New
Mexico: American Journal of Science, v. 31, p. 378-398
Williston, S.W. and Case, E.C., 1912, The Permo-Carboniferous of north
ern New Mexico: Journal of Geology, v. 20, p. 1-12.
Woodward, L. A., Anderson, O. J. and Lucas, S. G., 1999, Late Paleozoic
right-slip faults in the ancestral Rocky Mountains: New Mexico Geo
logical Society, Guidebook 50, p. 149-153.
6
 Lucas et al., eds., 2010, Carb-Permian transition in Cañon del Cobre. New Mexico Museum of Natural History and Science Bulletin 49.

7
ONE HUNDRED AND FIFTYYEARS OF GEOLOGICALAND PALEONTOLOGICAL RESEARCH
IN CAÑON DELCOBRE, NORTHERN NEWMEXICO

SPENCER G. LUCAS, SUSANK. HARRIS AND JUSTINA. SPIELMANN

New Mexico Museum of Natural History and Science, 1801 Mountain Road NW, Albuquerque, NM 87104-1375

Abstract—Scientific study of the geology and paleontology in Cañon del Cobre, northern New Mexico, began
with J.S. Newberry in 1859, who discovered Triassic plant fossils in the copper mines in the canyon. David
Baldwin first discovered Carboniferous-Permian tetrapod fossils in Cañon del Cobre in the late 1870s, and this led
to subsequent searches for Carboniferous-Permian tetrapod fossils. S.W. Williston and E.C. Case, in 1911, first
recognized that the oldest tetrapod fossils from Cañon del Cobre are of Carboniferous age, and subsequent research
has supported this conclusion. However, most of the bedrock strata in Cañon del Cobre were assigned a Triassic
age until 1961, when C. Smith, A. Budding and C.W. Pitrat assigned them to the Permian Cutler Formation.
Collecting and research in Cañon del Cobre has now identified invertebrate and vertebrate traces, palynomorphs,
plants and tetrapod bones and teeth of Carboniferous age, traces and tetrapods of Early Permian age and Late
Triassic tetrapod and plant fossils. Cañon del Cobre provides rare and important documentation of the Carbonif
erous-Permian transition, especially of tetrapods, in a nonmarine red-bed succession.

“From Abiquiu our entrance into the Red Beds was made in
the famous cañon known as El Cobre…” – Williston and
Case (1912, p. 4-5).
INTRODUCTION
Cañon del Cobre is a large box canyon located northwest of Abiquiu
in Rio Arriba County, New Mexico (Fig. 1). Also known by its English
name El Cobre Canyon, “copper canyon” (cobre is Spanish for copper)
is so named because of the copper mines located high on its eastern walls.
The Spanish first developed these mines, perhaps as early as the 1600s.
Newberry (1876), who briefly visited the canyon in 1859, provided the
first geological and paleontological information about Cañon del Cobre.
In the 150 years since Newberry’s visit, a variety of geologists and
paleontologists have studied Cañon del Cobre, and today it is a famous
locality for Carboniferous, Permian and Triassic fossils. Here, we review
the history of geological and paleontological research in Cañon del Cobre.
NEWBERRY (1859)
In 1859, John Strong Newberry (1822-1892) participated as the
geologist for the San Juan Exploring Expedition led by Captain J.N.
Macomb (often referred to as the “Macomb Expedition”). Newberry’s
(1876) report on the expedition contains the first descriptions of the
geology and paleontology of the Chama River basin, which includes the
drainage of Cañon del Cobre. FIGURE 1. Index map showing the location of Cañon del Cobre in northern
On 17 July 1859, Newberry (1876, p. 68-69) visited some of the New Mexico (modified from Ash, 1974).
copper mines (Minas de Pedro) developed in the Upper Triassic strata
exposed high on the eastern wall of Cañon del Cobre. Here, Newberry sandstones…here and there they showed large patches of
collected the first Triassic plant fossils with leaves discovered in the white, interstratified with green and brown shales…this
American West (Triassic fossil wood had been discovered in the 1840s). formation contains much saline matter…These cliffs [of
Newberry (1876) described these plants (Fig. 2), and his collection and Cañon del Cobre] are composed at base of the saliferous
additional material was published on by Fontaine and Knowlton (1890). sandstones and interstratified marls, some 250 feet in thick
Newberry (1876) believed that the red beds that floor Cañon del ness; above these blood-red marls and calcareous sand
Cobre and are exposed along most of its walls are of Triassic age: stones, 200 feet thick; the whole crowned by coarse yellow
sandstones having a thickness of about 150 feet…
This [Cañon del Cobre] is excavated in the Triassic series,
(Newberry, 1876, p. 68).
and its sides exhibit bands of brilliant color, red, orange,
blue, white, etc., as vivid as could be drawn from an artist’s Newberry (1876) correctly assigned a Triassic age to the fossil plants he
color-box…Ascending the arroyo, toward its head, we found collected from strata now assigned to the Shinarump Formation of the
the strata rising rapidly toward the north, and cut through Chinle Group (Lucas et al., 2005a). However, Newberry (1876) assigned
so as to expose the saliferous sandstones – the lowest the underlying red beds, Pennsylvanian-Permian strata of the Cutler
member of the Trias – in sections of at least 200 feet. The Group, a Triassic age (Fig. 3), and he also incorrectly assigned a Creta
group here consists of thick-bedded, chocolate ceous age to the sandstones above the plant-bearing horizon (Poleo For
8
FIGURE 2. Plate 5 of Newberry (1876), illustrations of some of the Upper
Triassic plants he collected in the Shinarump Formation of the Chinle
Group along the eastern rim of Cañon del Cobre. 1, 1a and 2, Zamites
occidentalis; 3, Otozamites macombi; and 4, 5, Pagiophyllum newberryi.
See Ash (1974) for further information.
FIGURE 3. Evolution of stratigraphic nomenclature and age assignments
applied to Pennsylvanian, Permian and Triassic strata exposed in Cañon del
Cobre.
mation of the Upper Triassic Chinle Group). Thus, Newberry (1876, p.
69) concluded “we have, therefore, in these plants evidence of the Trias
sic age of all the variegated gypsiferous rocks of northern New Mexico;
for the Lower Cretaceous sandstones immediately overlie the plant-beds
of the Cobre” (italics from Newberry, 1876).
BALDWIN (1877-1880)
David Baldwin was a fossil collector employed by O.C. Marsh
(1831-1899), the famous Yale University paleontologist, from 1876-
1880, to collect vertebrate fossils in northern New Mexico. According to
Williston (1911a, p. 378), Baldwin collected Carboniferous-Permian ver
tebrate fossils for Marsh from Rio Arriba County from November 1877
to December 1880, believing the fossils to be of Triassic age. Romer
(1960) states that Baldwin first discovered Carboniferous fossils in Cañon
del Cobre in 1877.
Marsh, however, did nothing with the bones Baldwin sent him
from Cañon del Cobre. In 1880, Marsh and Baldwin ended their working
relationship – Marsh was generally dissatisfied with the quality of the
fossils Baldwin was sending him, and Baldwin was not being compen
sated adequately or regularly by Marsh.
In 1880, Baldwin thus switched allegiance to Marsh’s scientific
rival, Philadelphia paleontologist Edward Drinker Cope (1840-1897).
However, Baldwin did no further collecting in Cañon del Cobre while
employed by Cope. Thus, Baldwin can be credited with discovering
Carboniferous vertebrate fossils in Cañon del Cobre, though it took
decades for his discovery to be recognized.
WILLISTON, CASE AND HUENE (1911)
Baldwin’s collection of Permian vertebrates from Rio Arriba
County, New Mexico, languished for nearly 30 years in the collections of
the Yale Peabody Museum. Then, about 1911, Samuel Wendell Williston
(1851-1918), a University of Chicago paleontologist, unpacked the
Baldwin collection, revealing the remarkable skeleton of a reptile from
Cañon del Cobre that Williston (1911a) named Limnoscelis paludis (Fig.
4). Williston (1911a, b, 1912) regarded Limnoscelis as a primitive reptile
(cotylosaur), and the exceptional specimen collected by Baldwin in Cañon
del Cobre has played a prominent role in subsequent discussions of
reptilian (amniote) origins (e.g., Romer, 1946; Carroll, 1969a, b, 1982;
Panchen, 1972; Lombard and Bolt, 1979; Heaton, 1980; Kemp, 1980;
Reisz and Heaton, 1980; Fracasso, 1987).
Indeed, Williston was so impressed by the Baldwin collection
that he resolved to visit Cañon del Cobre (and other nearby localities in
Rio Arriba County) to collect more fossils. Thus, in the summer of 1911,
Williston was joined by Ermine Cowles Case (1871-1953), a paleontol
ogy professor at the University of Michigan, Friedrich Freiherr von
Huene (1875-1969), a German paleontologist from Tübingen, and Paul
C. Miller, the fossil preparator and assistant curator working for Williston,
in an expedition to the red beds of Rio Arriba County (Williston and
Case, 1912; Case et al., 1913; Shor, 1971; Lucas, 2005).
No exact chronology of the expedition is available to us, but the
first stop was Cañon del Cobre, described by Williston (in Short, 1971,
p. 221) as “the most dreary and forsaken region I think I was ever in.”
Both a lack of water in the canyon and flash flooding of the canyon floor
proved a challenge to the expedition. Fieldwork in the canyon (Figs. 5-6)
proved relatively disappointing, with few and fragmentary vertebrate
fossils recovered. In Cañon del Cobre, the Williston team collected verte
brate fossils from the “lowermost beds” at “various places in the basin”
FIGURE 4. Restoration of the skeleton of Limnoscelis paludis (original
skeleton about 2 m long). From Williston (1912, fig. 32).
 9

FIGURE 5. Selected photographs of the 1911 Williston expedition to northern New Mexico (from a photo album of E.C. Case donated to NMMNH by
Jiri Zidek). A-C, Wagon (A), team (B) and camp kitchen (C). D-E, S.W. Williston. F, Paul Miller.
10
FIGURE 6. Selected photographs of the 1911 Williston expedition to northern New Mexico (from a photo album of E.C. Case donated to NMMNH by
Jiri Zidek). A-C, First camp in the Cañon (at the Puertocito). D, Paul Miller (left) and E.C. Case (right).
(Williston and Case, 1913, p. 2), and also explored strata in the north
wall “somewhat below the middle of the Permo-Carboniferous strata of
El Cobre canyon” (Case et al., 1913, p. 17), where a nearly complete
skeleton of the diadectid Diasparactus zenos was found (Case et al.,
1913). Williston and Case (1912), however, did provide a more detailed
description of the geology and stratigraphy of Cañon del Cobre than had
Newberry. As they noted, “El Cobre Canyon or basin is formed by the
erosion of an unsymmetrical dome-shaped anticline more or less faulted
on the northeastern and southeastern sides, the brim formed everywhere
by the massive sandstones of basal Upper Triassic age, the strata sloping
in all directions, but chiefly east and west” (Williston and Case, 1912, p.
5-6).
Williston and Case also presented a stratigraphic section of the
western wall of Cañon del Cobre, assigning the strata to the Permo
Carboniferous, Lower Trias? and Upper Trias (Fig.2). Their work thus
represented a major advance in our understanding of the paleontology
and age of strata exposed in Cañon del Cobre. Also, based on the work,
Williston (1911b, fig. 1) published what can be considered the first
reconstruction of a Carboniferous landscape in Cañon del Cobre (Fig. 7).
ROMER (1926-1952)
Although several descriptions of fossil vertebrates from Cañon
del Cobre were based on specimens collected by Baldwin and the
Williston expedition, their fieldwork did not inspire many followers.
The reputation of Cañon del Cobre as a remote, nearly inaccessible locale
whose main passageway, Arroyo del Cobre, is subject to flash floods
(one of which nearly drowned members of the Williston party of 1911),
discouraged extensive collecting efforts in the canyon for decades.
Harvard University paleontologistAlfred Sherwood Romer (1894
1973) noted that he collected Pennsylvanian-Permian vertebrate fossils
in Cañon del Cobre “in 1926, 1931, 1946 and 1952, and other Harvard
workers went there on three other occasions” (Romer, 1960, p. 48).
Although Romer reported fossils in the canyon to be rare, he was able to
describe the varanopseid Aerosaurus greenleeorum (Romer, 1937) from
a “double handful of bones and bone fragments, not articulated but pre
sumably from a single individual” (Romer and Price, 1940, p. 277).
Descriptions of the edaphosaurid Nitosaurus jacksonorum (Romer, 1937)
and the ophiacodontid Baldwinonus trux (Romer and Price, 1940) were
based on specimens collected by Baldwin. Romer (1950) also published
a photograph taken by Williston in 1911 with a horizontal line indicating
the vertebrate-producing horizon on the north wall of Cañon del Cobre,
which provides the only precise information we have of where the
Williston expedition collected its fossils (Fig. 8). Soon after, Langston
(1953) provided a useful review of the Paleozoic vertebrate fauna of
Cañon del Cobre, but he did not collect there.
SMITH, BUDDING AND PITRAT (1950S)
During the 1950s, the geological studies conducted by field classes
from the New Mexico Institute of Mining and Mineral Technology
(Socorro) led to the published geological mapping of Smith et al. (1961).
Thus, they mapped the geology of the U.S. Geological Survey Canjilon
SE 7.5-minute topographic map, within which Cañon del Cobre is lo
cated. Most significantly, Smith et al. (1961; also see Budding et al.,
1960) applied the lithostratigraphic name Cutler Formation to the Cañon

FIGURE 7. “A permocarboniferous [sic] landscape, with two figures of Eryops upon the land, each of about seven feet in length, and Limnoscelis, a
cotylosaur of the same length, in the water” (Williston, 1911b, p. 58, fig. 1).
del Cobre section, assigning most of the bedrock exposed in the canyon
to this unit, which they regarded as Early Permian in age (Fig. 3). They
assigned overlying strata, which rim the canyon walls and contain the
copper mines and fossil plants first collected by Newberry, to the “lower
sandstone member” of the Chinle Formation (Fig. 3). They did not apply
the formal stratigraphic subdivisions of the Chinle introduced by Wood
and Northrop (1946) in the Rio Puerco River Valley some 50 km to the
southwest. Near the northern terminus of Cañon del Cobre, Smith et al.
(1961, p. 5, pl. 6) located “thin-bedded carbonaceous, gypsiferous and
micaceous siltstone…exposed along a small fault” with “a well pre
served flora, consisting predominantly of Alethopteris serlii…” They
judged these strata to be of Pennsylvanian age and mapped them as a unit
separate from the Cutler Formation (Fig. 3).
VAUGHN (1963) AND LEWIS AND OLSEN (1965)
University of California at Los Angeles paleontologist Peter P.
Vaughn (1963) reported disarticulated, fragmentary vertebrate fossils
that he collected from two sites on the floor of Cañon del Cobre. How
ever, he deemed the specimens sufficiently diagnostic to recognize three
additional taxa in the Cañon del Cobre vertebrate assemblage: the
dissorophid Platyhystrix rugosus, based on a large number of neural
spine fragments with surface ornamentation consistent with that of P.
rugosus from the Arroyo del Agua locality; the temnospondyl
Chenoprosopus cf. C. milleri, recognized from a cranial fragment that
consists of the dermal roofing and palatal elements of the left narial
region; and the edaphosaurid Edaphosaurus cf. E. novomexicanus, based
on several incomplete neural spines bearing tubercles on both sides, with
a diameter of 8 mm, as in E. novomexicanus. In addition, Vaughn col
11
lected several incomplete vertebrae referable to Limnoscelis. Vaughn’s
(1963) principal conclusion was to reaffirm a Carboniferous age for the
stratigraphically lowest vertebrate fossils in Cañon del Cobre.
During a brief trip to the canyon in 1965, A. Lewis and S. Olsen of
Harvard University collected a partial skeleton of a pelycosaur that was
later described by Eberth and Brinkman (1983) as a new genus and
species of varanopsid pelycosaur, Ruthiromia elcobriensis (also see
Spielmann and Lucas, this volume). Although the precise locality is not
known, field notes made by Lewis and Olsen (Eberth and Brinkman,
1983) indicate that the material was collected from the western wall of
the canyon, strongly suggesting a stratigraphically high position in the
Arroyo del Agua Formation of Lucas and Krainer (2005), and thus a
younger age for this specimen than that of fossils collected from the
classic sites on the canyon floor.
FRACASSO (1978)
Fracasso (1980) reported vertebrate fossils collected from five
sites on the canyon floor of the Cañon del Cobre collected during the
summer of 1978. The most significant find of this trip, made by one of us
(SGL), as a member of the Fracasso reconnaissance team, was a partial
jaw and cranial fragments of the primitive diadectomorph Desmatodon
aff. D. hollandi, indicative of a Late Pennsylvanian age for the strata
exposed on the canyon floor. It was described, but not illustrated, by
Fracasso (1980) and Berman and Sumida (1995; but see Lucas et al.,
2005c, fig. 5f, i-j for an illustration). Fracasso (1980) also reported a
megafossil plant locality from the canyon floor that the late S.H. Mamay
of the Smithsonian National Museum of Natural History deemed to be
of Late Pennsylvanian age. Fracasso (1987) subsequently presented an
12
FIGURE 8. Photograph by S.W. Williston (1911) looking approximately
northward of part of the western wall of Cañon del Cobre (view is primarily
of N½ secs. 17-18 and S1/4 secs. 7-8, T24N, R06E). The horizontal line
marks the vertebrate-producing horizon in the El Cobre Canyon Formation;
the “+” indicates the base of the Upper Triassic. From Romer (1950).
analysis of the fluvial sedimentology of the lower part of the Cutler
Group exposed in Cañon del Cobre.
BERMAN (1980s)
Collecting efforts of the last quarter century in Cañon del Cobre
began with work by joint field parties from the Carnegie Museum of
Natural History and the University of Toronto in the early 1980s led by
David S Berman from the former institution. As a result of this field
work, three additional taxa were recognized in the Cañon del Cobre
vertebrate-fossil assemblage, and include the new genus and species of
trematopid temnospondyl, Anconastes vesperus, described by Berman
et al. (1987b) on the basis of two partial skeletons; a small, as yet
undescribed embolomere known from a single multispecimen occurrence;
and the lungfish Gnathorhiza preserved in an aestivation burrow cast. In
addition to these specimens, collected at three different intervals in the El
Cobre Canyon Formation section, a rich concentration of bones high in
the section in the overlying Arroyo del Agua Formation on the eastern
wall of the canyon was also discovered. Berman (1993) and Eberth and
Berman (1993) noted that at least three taxa, including Platyhystrix,
Diadectes and Sphenacodon ferox, are well represented at this site, which
yields disarticulated, though well preserved bones, on which fine details
are preserved.
In conjunction with paleontological studies, Eberth (1987; Eberth
and Miall, 1991; Eberth and Berman, 1983, 1993) studied the fluvial
sedimentology of the Cutler Group strata exposed in Cañon del Cobre.
This work presented the first analysis of the depositional context of the
fossils found in the Permo-Carboniferous Cutler Group strata in Cañon
del Cobre.
NEWMEXICO MUSEUM OF NATURALHISTORY (1991-2007)
Based on the field studies undertaken during the late 1980s, Lucas
and Hunt (1992; also see Lucas et al., 2005b) presented a revised Triassic
stratigraphy in the Chama basin. Currently, based on this work, the
Upper Triassic section exposed along the rim of the Cañon del Cobre is
assigned to the (ascending order) Shinarump, Salitral and Poleo forma
tions of the Chinle Group (Fig. 3).
New Mexico Museum of Natural History field efforts in the
upper Paleozoic strata of Cañon del Cobre began in the summer of 1991
with the collection of plant megafossils from two localities just below
the lowest tetrapod-bearing beds. Hunt and Lucas (1992) re-examined
the fossil plant locality reported by Fracasso (1980) and determined the
FIGURE 9. NMMNH fieldwork in Cañon del Cobre in 2005. A, Central mess
tent of base camp, manned by Chester Lawrence. B, Larry Rinehart
excavating Eryops skull. C, Joseph Hannibal at giant Diplichnites trackway.
D, Joshua Smith (above) and Matt Celeskey at Sphenacodon quarry.
paleoflora, as well as the fossil-vertebrate fauna from “megasequence
one” of Eberth (1987) to be Pennsylvanian and suggested an age as old as
Desmoinesian. Although the paleobotany of Cañon del Cobre remains an
active research pursuit of the NMMNH and its associates (see DiMichele
and Chaney, 2005), recent paleontological investigations (since 2000)
have included the discovery and collection of numerous fossil verte
brates at several localities.
A combined field effort of the New Mexico Museum of Natural
History, National Museum of Natural History (Smithsonian Institution)
and Carnegie Museum of Natural History in Cañon del Cobre began in
2000 (Fig. 9). Particularly significant fieldwork took place in 2000-2007,
which was devoted to prospecting and collecting any and all fossil locali
ties in the canyon and studying the stratigraphy and sedimentology of
the Cutler Group strata in the canyon. Overlapping and in collaboration
with the NMMNH, the New Mexico Bureau of Geology mapped the
geology of the Canjilon SE quadrangle (Kempter et al., 2007). Also, K.
Krainer of the University of Innsbruck (Austria) collaborated on the
stratigraphy and sedimentology of the Cutler Group strata exposed in
Cañon del Cobre, and J. Hannibal of the Cleveland Museum of Natural
History and J. Schneider of Freiberg University (Germany) also worked
on some aspects of Cutler Group paleontology. Initial results of the
NMMNH project were published in 2005 (DiMichele and Chaney, 2005;
Hannibal et al., 2005; Lucas and Krainer, 2005; Lucas et al., 2005b, c, d).
Particularly significant results of the NMMNH collaborative re
search in Cañon del Cobre include: (1) lithostratigraphic refinement by
recognition of two formations of the Cutler Group in the Chama basin,
the older El Cobre Canyon Formation overlain by the Arroyo del Agua
Formation and studies of their petrography and sedimentology (Lucas
and Krainer, 2005; Kempter et al., 2007); (2) stratigraphic organization
of the fossil vertebrate localities in the Cutler Group strata in Cañon del
Cobre, allowing for the recognition of three temporally successive as
semblages that provided part of the basis for establishment of the Cobrean,

Coyotean and Seymouran land-vertebrate faunachrons (Lucas, 2006;
Lucas et al., 2005c, d); (3) discovery and collection of a substantial
pteridosperm-dominated paleoflora (and associated palynomorphs) from
the El Cobre Canyon Formation (DiMichele and Chaney, 2005); and (4)
study of the ichnology of the El Cobre Canyon Formation, including
discovery of the walking traces of the giant Carboniferous arthropod
Arthropleura (Lucas et al., 2005c).
REFERENCES
Ash, S.R., 1974, Upper Triassic plants of Cañon del Cobre, New Mexico:
New Mexico Geological Society, Guidebook 25, p. 179-184.
Berman, D.S, 1993, Lower Permian vertebrate localities of New Mexico
and their assemblages: New Mexico Museum of Natural History and
Science, Bulletin 2, p. 11-21.
Berman, D.S, Reisz, R.R. and Eberth, D.A., 1987a, Seymouria sanjuanensis
(Amphibia, Batrachosauria) from the Lower Permian Cutler Formation
of north-central New Mexico and the occurrence of sexual dimorphism
in that genus questioned: Canadian Journal of Earth Science, v. 24, p.
1769-1784.
Berman, D.S, Reisz, R.R. and Eberth, D.A., 1987b, A new genus and species
of trematopid amphibian from the Late Pennsylvanian of north-central
New Mexico: Journal of Vertebrate Paleontology, v. 7, p. 252-269.
Berman, D.S and Sumida, S., 1995, New cranial material of the rare diadectid
Desmatodon hesperis (Diadectomorpha) from the Late Pennsylvanian
of central Colorado: Annals of Carnegie Museum, v. 64, p. 315-336.
Budding, A.J., Pitrat, C.W. and Smith, C.I., 1960, Geology of the southern
part of the Chama basin: New Mexico Geological Society, Guidebook 11,
p. 78-92.
Carroll, R.L., 1969a, Problems of the origin of the reptiles: Biological
Reviews, v. 44, p. 393-432.
Carroll, R.L., 1969b, Origin of reptiles; in Gans, C., Bellairs, A. d’A. and
Parsons, T.S., eds., Biology of the Reptilia, vol. 1 Morphology A: Aca
demic Press, New York, p. 7-44.
Carroll, R.L., 1982, Early evolution of reptiles: Annual Review of Ecology
and Systematics, v. 13, p. 87-109.
Case, E.C. and Williston, S.W., 1913, Description of a nearly complete
skeleton of Diasparactus zenos Case: Carnegie Institution of Washing
ton, Publication no. 181, p. 17-35.
Case, E.C., Williston, S.W. and Mehl, M.G., 1913, Permo-Carboniferous
vertebrates from New Mexico Carnegie Institution of Washington, Pub
lication no. 181, p. 1-81.
DiMichele, W.A. and Chaney, D.S., 2005, Pennsylvanian-Permian fossil
floras from the Cutler Group, Cañon del Cobre and Arroyo del Agua
areas, in northern New Mexico: New Mexico Museum of Natural His-
tory and Science, Bulletin 31, p. 26-33.
Eberth, D.A., 1987, Stratigraphy, sedimentology, and paleoecology of Cut
ler Formation redbeds (Permo-Pennsylvanian) in north-central New
Mexico [Ph.D. dissertation]: University of Toronto, Toronto, 264 p.
Eberth, D.A. and Berman, D.S, 1983, Sedimentology and paleontology of
Lower Permian fluvial redbeds of north-central New Mexico – prelimi
nary report: New Mexico Geology, v. 5, p. 21-25.
Eberth, D.A. and Berman, D.S, 1993, Stratigraphy, sedimentology and ver
tebrate paleontology of the Cutler Formation redbeds (Pennsylvanian
Permian) of north-central New Mexico: New Mexico Museum of Natu
ral History and Science, Bulletin 2, p. 33-48.
Eberth, D.A. and Brinkman, D., 1983, Ruthiromia elcobriensis, a new
pelycosaur from El Cobre Canyon, New Mexico: Breviora, v. 474, p. 1
26.
Eberth, D.A. and Miall, A.D., 1991, Stratigraphy, sedimentology and evolu
tion of a vertebrate-bearing, braided to anastomosed fluvial system,
Cutler Formation (Permian-Pennsylvanian), north-central New Mexico:
Sedimentary Geology, v. 72, p. 225-252.
Fracasso, M.A., 1980, Age of the Permo-Carboniferous Cutler Formation
13
ACKNOWLEDGMENTS
We are grateful to numerous NMMNH volunteers as well as
colleagues from various institutions who collaborated with us in our
work in Cañon del Cobre. Special thanks go to David Berman and Amy
Henrici for sharing their knowledge of the paleontology of the canyon.
David Berman, William DiMichele, Amy Henrici and Karl Krainer pro
vided helpful reviews of the manuscript.
vertebrate fauna from El Cobre Canyon, New Mexico: Journal of Pale
ontology, v. 54, p. 1237-1244.
Fracasso, M.A., 1983, Cranial osteology, functional morphology, system
atics and paleoenvironment of Limnoscelis paludis Williston [Ph.D.
dissertation]: Yale University, New Haven, 624 p.
Fracasso, M.A., 1987a, Braincase of Limnoscelis paludis Williston: Postilla,
v. 201, p. 1-22.
Fracasso, M.A., 1987b, Fluvial deposition and paleoenvironment of the
Cutler Formation red beds, El Cobre Canyon, New Mexico: New Mexico
Geology, v. 9, p. 14-18.
Hannibal, J.T., Lucas, S.G., Lerner, A.J and Chaney, D.S., 2005, An euryp
terid (Adelophthalmus sp.) from a plant-rich lacustrine facies of Upper
Pennsylvanian strata in El Cobre Canyon, New Mexico: New Mexico
Museum of Natural History and Science, Bulletin 31, p. 34-38.
Heaton, M.J., 1980, The Cotylosauria: A reconsideration of a group of
archaic tetrapods; in Panchen, A.L., ed., The terrestrial environment
and origin of land vertebrates: Academic Press, New York, p. 497-551.
Hunt, A.P. and Lucas, S.G., 1992, The paleoflora of the lower Cutler Forma
tion (Pennsylvanian, Desmoinesian?) in El Cobre Canyon, New Mexico,
and its biochronological significance: New Mexico Geological Society,
Guidebook 43, p. 145-150.
Kemp, T.S., 1980, Origin of the mammal-like reptiles: Nature, v. 283, p.
378-380.
Kempter, K., Zeigler, K., Koning, D. and Lucas, S., 2007, Preliminary
geologic map of the Canjilon SE quadrangle, Rio Arriba County, New
Mexico: New Mexico Bureau of Geology and Mineral Resources, Open
file Geological Map 150, 1:24,000.
Langston, W., Jr., 1953, Permian amphibians from New Mexico: Univer
sity of California, Publications in Geological Sciences, v. 29, p. 349
416.
Lombard, R.E. and Bolt, J.R., 1979, Evolution of the tetrapod ear: An
analysis and reinterpretation: Biological Journal of the Linnean Society,
v. 11, p. 19-76.
Lucas, S.G., 2005, J.S. Newberry and the geology of the Chama basin: New
Mexico Geological Society, Guidebook 56, p. 53.
Lucas, S.G., 2006, Global Permian tetrapod biostratigraphy and
biochronology; in Lucas, S.G., Cassinis, G. and Schneider, J.W., eds.,
Non-marine Permian biostratigraphy and biochronology: Geological
Society, London, Special Publication 265, p. 65-93.
Lucas, S.G. and Hunt, A.P., 1992, Triassic stratigraphy and paleontology,
Chama basin and adjacent areas, north-central New Mexico: New Mexico
Geological Society, Guidebook 43, p. 151-167.
Lucas, S.G. and Krainer, K., 2005, Stratigraphy and correlation of the
Permo-Carboniferous Cutler Group, Chama Basin, New Mexico: New
Mexico Geological Society, Guidebook 56, p. 145-159.
Lucas, S.G., Zeigler, K.E., Heckert, A.B. and Hunt, A.P., 2005a, Review of
Upper Triassic stratigraphy and biostratigraphy in the Chama basin,
northern New Mexico: New Mexico Geological Society, Guidebook 56,
p. 170-181.
Lucas, S.G., Harris, S.K., Spielmann, J.A., Berman, D.S and Henrici, A.C.,
2005c, Vertebrate biostratigraphy and biochronology of the Pennsylva
nian-Permian Cutler Group, El Cobre Canyon, northern New Mexico:
New Mexico Museum of Natural History and Science, Bulletin 31, p.
128-139.
14
Lucas, S.G., Harris, S.K., Spielmann, J.A., Berman, D.S, Henrici, A.C.,
Heckert, A.B., Zeigler, K.E. and Rinehart, L.F., 2005d, Early Permian
vertebrate assemblage and its biostratigraphic significance, Arroyo del
Agua, Rio Arriba County, New Mexico: New Mexico Geological Society,
Guidebook 56, p. 288-296.
Lucas, S.G., Lerner, A.J, Hannibal, J.T., Hunt, A.P. and Schneider, J.W.,
2005b, Trackway of a giant Arthropleura from the Upper Pennsylva
nian of El Cobre Canyon, New Mexico: New Mexico Geological Society,
Guidebook 56, p. 279-282.
Newberry, J.S., 1876, Geological report; in Macomb, J.N., ed., Report of the
exploring expedition from Santa Fe…in 1859: U.S. Army Engineer
Department, Washington, D.C., p. 9-118.
Panchen, A.L., 1972, The interrelationships of the earliest tetrapods; in
Joysey, K.A. and Kemp, T.S., eds., Studies in vertebrate evolution: Win
chester Press, New York, p. 65-87.
Reisz, R.R. and Heaton, M.J., 1980, Origin of mammal-like reptiles: A reply
to T.S. Kemp: Nature, v. 288, p. 193.
Romer, A.S., 1937, New genera and species of pelycosaurian reptiles: New
England Zoological Club, Proceedings, v. 16, p. 89-96.
Romer, A.S., 1946, The primitive reptile Limnoscelis restudied: American
Journal of Science, v. 244, p. 149-188.
Romer, A.S., 1950, The upper Paleozoic Abo Formation and its vertebrate
fauna; in Colbert, E.H. and Northrop, S.A., eds., Guidebook for the
fourth field conference of the Society of Vertebrate Paleontology in
northwestern New Mexico: American Museum of Natural History, New
York, p. 48-55.
Romer, A.S., 1960, The vertebrate fauna of the New Mexico Permian: New
Mexico Geological Society, Guidebook 11, p. 48-54.
Romer, A.S. and Price, L.I., 1940, Review of the Pelycosauria: Geological
Society of America, Special Paper No. 28, 538 p.
Shor, E.N., 1971, Fossils and flies: The life of a complete scientist, Samuel
Wendell Williston (1851-1918): University of Oklahoma Press, Norman,
285 p.
Smith, C. T., Budding, A. J. and Pitrat, C. W., 1961, Geology of the south
eastern part of the Chama basin: New Mexico Bureau of Mines and
Mineral Resources, Bulletin 75, 57 p
Vaughn, P.P., 1963, The age and locality of the late Paleozoic vertebrates
from El Cobre Canyon, Rio Arriba County, New Mexico: Journal of
Paleontology, v. 37, p. 283-296.
Williston, S.W., 1911a, A new family of reptiles from the Permian of New
Mexico: American Journal of Science, v. 31, p. 378-398
Williston, S.W., 1911b, American Permian vertebrates: University of Chi
cago Press, Chicago, 145 p.
Williston, S.W., 1912, Restoration of Limnoscelis, a cotylosaur reptile
from New Mexico: American Journal of Science, v. 34, p. 457-468.
Williston, S.W., 1914, The osteology of some American Permian verte
brates: Journal of Geology, v. 22, p. 408-410.
Williston, S.W. and Case, E.C., 1912, The Permo-Carboniferous of north
ern New Mexico: Journal of Geology, v. 20, p. 1-12.
 Lucas et al., eds., 2010, Carb-Permian transition in Cañon del Cobre. New Mexico Museum of Natural History and Science Bulletin 49.

15
SUMMARY OF GEOLOGY OF CAÑON DELCOBRE,
RIOARRIBA COUNTY, NEWMEXICO

SPENCERG. LUCAS1, JUSTINA. SPIELMANN1 AND KARLKRAINER2

1New Mexico Museum of Natural History and Science, 1801 Mountain Road NW, Albuquerque, NM 87104-1375;
2 Institute of Geology and Paleontology, Innsbruck University, Innrain 52, Innsbruck, A-6020 AUSTRIA

Abstract—Cañon del Cobre is a large box canyon in southeastern Rio Arriba County, New Mexico, along the
southeastern border of the Colorado Plateau, that exposes strata of Pennsylvanian, Permian, Triassic and Cenozoic
age. The oldest rocks here are siliciclastic red beds of the El Cobre Canyon Formation of the Cutler Group, strata
of Late Pennsylvanian-Early Permian (Virgilian-Wolfcampian) age. The exposed El Cobre Canyon Formation is
about 230 m thick (total thickness ~ 640 m) and is mostly siltstone and sandstone deposited in an ephemeral
braided stream environment. The overlying Lower Permian Arroyo del Agua Formation of the Cutler Group is as
much as 160 m thick, and also consists mostly of siltstone and sandstone, but was deposited on laterally extensive
floodplains that developed between relatively stable channels. An extensive unconformity separates Cutler Group
strata from overlying Late Triassic strata of the Chinle Group around the rim of Cañon del Cobre. Chinle Group
strata are of fluvial origin and are assigned (in ascending order) to the Shinarump (up to 25 m of quartzose
sandstone, conglomeratic sandstone and siliceous, extraformational conglomerate), Salitral (up to 7 m of variegated
mudstone) and Poleo (up to 30 m of sandstone, conglomeratic sandstone and conglomerate) formations. The
Shinarump Formation hosts stratiform copper deposits along the eastern wall of the Cañon del Cobre that the
Spanish began mining during the 1600s.
Cenozoic strata locally exposed around the rim of Cañon del Cobre include the Eocene El Rito Formation (up
to 50 m of red-bed siltstones, sandstones and conglomerates) and Oligocene Ritito Conglomerate (up to 60 m of
conglomerate and arkosic sandstone). Quaternary deposits – quartzite erratic boulders, alluvial terrace deposits,
colluvium and alluvium – crop out in and around Cañon del Cobre. Cañon del Cobre is traversed and bounded by
several north-northeast-trending faults, and its eastern wall is along a faulted, north-northeast-trending anticline.
These structures were accomodated by Neogene extensional tectonism of the nearby Rio Grande rift.

INTRODUCTION
Cañon del Cobre (also known as El Cobre Canyon or as Copper
Canyon) is a large box canyon located in southeastern Rio Arriba County,
New Mexico (Fig. 1). This volume focuses on the geology and paleontol
ogy of the Carboniferous-Permian transition as recorded in Cutler Group
red beds exposed in Cañon del Cobre. Here, we present a brief summary
of the outcrop geology in and immediately around the canyon (Fig.2). In
this volume, Lucas et al. (2010) review the history of geological and
paleontological studies in Cañon del Cobre. Discussion of the Pennsyl
vanian, Permian and Triassic rocks is based primarily on our own work,
whereas observations on the Cenozoic geology and structure are summa
rized primarily from Kempter et al. (2007).
PHYSIOGRAPHY
Located in the southern Rocky Mountains along the southeastern
border of the Colorado Plateau, Cañon del Cobre extends from a narrow
set of water gaps at El Puertocito at its southern end to its northeastern
rim over a distance of about 10 km (Fig. 2). Throughout most of its
course, the canyon is 2-3 km wide, and its floor elevation ranges from
about 6300 to 7400 feet (1920 to 2255 m) above sea level. Maximum
depth of the canyon is about 305 m. The canyon has an almost north
south orientation for the southern 5 km, then bends to the northeast over
its northern 5 km. FIGURE 1. Index map showing location of Cañon del Cobre in Rio Arriba
The floor of the canyon is occupied by the Arroyo del Cobre, an County, northern New Mexico (after Ash, 1974).
intermittent tributary of the Rio Chama. A single, four-wheel drive dirt
road enters Cañon del Cobre through El Puertocito and extends through unnamed and unmarked mines (uranium or copper prospects) are lo
the southern half of the canyon. Old Spanish copper mines (first de cated in the floor of the canyon. An active artesian well is also present in
scribed by Newberry, 1876) – Las Minas Jimmie and Las Minas de the canyon (NW¼ SW¼ sec. 19, T24N, R06E), as are various seasonal
Pedro – are located high on the canyon’s eastern rim. Various other springs.
16

FIGURE 2. Geologic map of Cañon del Cobre (modified from Kempter et al., 2007). Square grid is km-spaced UTM (Universal Transverse Mercator) grid.
 17
STRATIGRAPHY
We identify three groups of sedimentary deposits in Cañon del
Cobre: (1) Carboniferous-Permian Cutler Group; (2) Late Triassic Chinle
Group; and (3) Cenozoic strata, colluvium and alluvium (Fig. 3). Des
cription of the Cutler Group is based in part on Lucas and Krainer
(2005a, b), Chinle Group strata are described largely based on Lucas and
Hunt (1992), and Kempter et al. (2007) is our primary source for data on
the Cenozoic rocks. Jurassic strata (Entrada Formation) are also exposed
just SE of Cañon del Cobre (Fig. 2), but are not discussed here.

Carboniferous and Permian Systems

Cutler Group
The floor and canyon walls of Cañon del Cobre expose extensive
bedrock badlands, slopes and escarpments developed in an ~400 m thick
section of Cutler Group siliciclastic red beds (Smith et al., 1961; Lucas
and Krainer, 2005a, b; Kempter et al., 2007). These strata are of fluvial
origin (Eberth and Berman, 1983, 1993; Fracasso, 1987; Eberth and
Miall, 1991) and have been divided into a lower, El Cobre Canyon For
mation and an upper, Arroyo del Agua Formation (Lucas and Krainer,
2005a, b) (Figs. 3-6).

El Cobre Canyon Formation

In Cañon del Cobre, a total thickness of about 230 m of El Cobre
Canyon Formation is exposed, and it is mostly siltstone and sandstone
(Figs. 3-6; Appendix 1). Minor rock types are conglomerate/conglomer
atic sandstone, sandy shale and calcrete. These rocks are characteristi
cally “brown” (pale reddish brown) and are readily distinguished from
the overlying Arroyo del Agua Formation based on color and lithology.
Siltstones of the El Cobre Canyon Formation contain numerous rhizoliths,
and sandstones are typically coarse grained, arkosic, micaceous, trough
crossbedded and form laterally extensive, multistoried bodies that erode
to cliffs and ledges. Local shale beds represent lacustrine deposits. Con
glomerates characteristically have extraformational clasts of quartzite,
granite and gneiss (Kempter et al., 2007) or calcrete rip-ups (Appendix
1). The El Cobre Canyon Formation is interpreted to represent deposits
of a shallow, ephemeral braided stream environment according to Eberth
and Miall (1991).
The El Cobre Canyon Formation is best exposed in the floor of
Cañon del Cobre and in the Rio Puerco Valley near Arroyo del Agua,
about 40 km to the southwest. Its lower contact is not exposed, but
subsurface data published by Smith et al. (1961) indicate that ~410 m of
El Cobre Canyon Formation strata are present below the outcrop base in
Cañon del Cobre, and that the El Cobre Canyon Formation rests on
Proterozoic basement (Smith et al., 1961; Eberth and Miall, 1991). The
upper contact of the El Cobre Canyon Formation appears to be con
formable at the base of the first “orange” siltstone slope of the Arroyo
del Agua Formation. The El Cobre Canyon Formation approximately
corresponds to megasequence 1 of Eberth and Miall (1991).

Arroyo del Agua Formation

Conformably overlying the El Cobre Canyon, the Arroyo del
Agua Formation in Cañon del Cobre is as much as 120 m thick (Figs. 3
6) and mostly consists of sandy siltstone and sandstone. Minor rock
types are calcrete and conglomerate/conglomeratic sandstone. Siltstones
of the Arroyo del Agua Formation are characteristically “orange” (mod
erate reddish brown), locally contain rhizoliths and may contain abun
dant calcrete nodules. They underlie long slopes between thin sheets of
sandstone that are coarse grained, arkosic, trough crossbedded and, lo
cally, multistoried. Conglomerates are not as common in the Arroyo del
Agua Formation as in the underlying El Cobre Canyon Formation, and
most are intraformational (composed of calcrete clasts). However, in the
upper part of the Arroyo del Agua Formation extraformational conglom
erates (primarily composed of quartzite clasts) are present (this is FIGURE 3. Summary of stratigraphic units exposed in Cañon del Cobre.
18
FIGURE 4. View of part of western wall of Cañon del Cobre (principally secs. 24-25, T24N, R5E) with stratigraphic units labeled.
megasequence 3 of Eberth and Miall, 1991).
The Arroyo del Agua Formation is well exposed along the walls of
Cañon del Cobre and in the Rio Puerco Valley from Arroyo del Agua east
to Abiquiu Reservoir. At all outcrops in the Chama Basin, the Arroyo del
Agua Formation conformably overlies the El Cobre Canyon Formation.
Gradation and interfingering characterizes this contact, so it is appar
ently conformable. Throughout Cañon del Cobre, the Chinle Group rests
disconformably on the Arroyo del Agua Formation (Lucas and Hunt,
1992; Lucas et al., 2003, 2005a; Kempter et al., 2007). (Figs. 4-5)
The Arroyo del Agua Formation approximately encompasses
megasequences 2 and 3 of Eberth and Miall (1991). The facies shows
some significant differences compared to the El Cobre Canyon Forma
tion. The Arroyo del Agua Formation is characterized by locally abun
dant, major sandstone ribbons and u-shaped channel fills, separated by
thick intervals of siltstone containing abundant calcrete nodules. When
compared to El Cobre Canyon Formation deposition, the fluvial style
had changed to laterally extensive floodplains that developed between
relatively stable channels that included local, anastomosed channels. The
climate has been interpreted as more arid than during deposition of the El
Cobre Canyon Formation (Eberth and Miall 1991). However, the change
in the character of the fluvial deposits between the El Cobre Canyon
Formation and the Arroyo del Agua Formation could have been caused
by a shift in base level as the result of erosion and accumulation. In
Cañon del Cobre, all but two of the known Paleozoic vertebrate fossil
localities are in the El Cobre Canyon Formation (Lucas et al., 2005b). All
known Paleozoic fossil plant and invertebrate trace fossil sites are also in
the El Cobre Canyon Formation (Fig. 4).
Triassic System
Chinle Group
Late Triassic strata of the Chinle Group (Lucas, 1993) are ex
posed across the Chama River basin and have recently been reviewed by
Lucas and Hunt (1992) and Lucas et al. (2003, 2005a), who provide
detailed descriptions and discussions. Chinle strata crop out along the
upper rimrock (periphery) of the Cañon del Cobre and rest with pro
found unconformity on the Arroyo del Agua Formation of the Cutler
Group (Figs. 4, 5, 7).
Shinarump Formation
The base of the Chinle section is the Shinarump Formation (for
merly Agua Zarca Member of the Chinle Formation: Wood and Northrop,
1946). The Shinarump is up to 25 m thick at Cañon del Cobre and
consists of fluvially deposited quartzose and feldspathic litharenite sand
stone, conglomeratic sandstone and siliceous (quartz, chert and/or quartz
ite), extraformational conglomerate. Colors are mostly greenish gray and
grayish yellow, and trough crossbeds are the dominant bedform. The
Shinarump Formation forms a light-colored bench or cliff at the base of
the Chinle Group. Its basal contact is usually a sharp, erosional scour on
underlying finer grained strata of the Arroyo del Agua Formation of the
Cutler Group. Locally, there are as much as 2 m of color-mottled strata
between the Shinarump and Arroyo del Agua red beds – the “mottled
strata” of Stewart et al. (1972) or Zuni Mountains Formation of Heckert
and Lucas (2003) (Fig. 7). Local copper mineralization in the Shinarump
Formation was mined at Las Minas Jimmie and Las Minas de Pedro,
beginning during the 1600s (R. Eveleth, written commun., 2009).
Salitral Formation
The Shinarump Formation at Cañon del Cobre is sharply overlain
by mudstone of the Salitral Formation. As much as 7 m thick, the Salitral
is color mottled (grayish red, purple, yellowish gray and light gray)
mudstone and sandy mudstone. It forms a notch or slope between the
more resistant Shinarump Formation (below) and the Poleo Formation
(above).
Poleo Formation
The youngest Late Triassic strata exposed around the periphery
of Cañon del Cobre belong to the Poleo Formation. As much as 30 m
thick locally, the Poleo Formation is fluvially-deposited sandstone, con
glomeratic sandstone and conglomerate. Sandstones are mostly yellow
ish gray micaceous litharenites and are typically crossbedded or laminar.
Conglomerates have mixed populations of intrabasinal (siltstone and
nodular calcrete) and extrabasinal (chert and quartzite) clasts. The Poleo
forms most of the cliffs that rim Cañon del Cobre (Fig. 4). Its base is an
erosional unconformity on the Salitral Formation. Younger Chinle Group
strata (Petrified Forest and Rock Point formations) are exposed south of
El Puertocito, just outside the southern periphery of Cañon del Cobre
(Kempter et al., 2007).
Cenozoic Units
The oldest Cenozoic strata exposed around the rim of Cañon del
 19

FIGURE 5. Measured stratigraphic section of Cutler Group strata in Cañon del Cobre, which is the type section of the El Cobre Canyon Formation of Lucas
and Krainer (2005a). See Appendix 1 for description of section.
20

FIGURE 6. Selected outcrops of the El Cobre Canyon Formation of the Cutler Group in Cañon del Cobre. A, C, Characteristic crossbedded sandstones. B,
Light-colored, thinly laminated siltstone of localized lacustrine deposit. D, Extabasinal matrix- to clast-supported conglomerate of (primarily) quartzite
pebbles. E, Rhizoliths. F, Bioturbation (cf. Planolites) in channel-margin sandstone.
 21

FIGURE 8. Structural cross-section from northwest-to-southeast across

Cañon del Cobre. Modified from Kempter et al. (2007).

Cobre belong to the Eocene El Rito Formation. These are red-bed silt
stones, sandstones and conglomerates up to 50 m thick (Smith et al.,
1961; Kempter et al., 2007). They are fluvial deposits of late Laramide
(Eocene) age deposited at the northern end of the Eocene Galisteo-El
Rito basin (Logsdon, 1981; Chapin and Cather, 1981; Lucas, 1984).
The next Cenozoic unit exposed around the rim of the Cañon del
Cobre is the Ritito Conglomerate (formerly the lower member of the
Abiquiu Formation) (Kempter et al., 2007; Maldonado and Kelley, 2009).
These strata are gray to pink conglomerates and arkosic sandstones up to
60 m thick (Kempter et al., 2007).
Quaternary deposits in the Cañon del Cobre include valley-floor
alluvium, alluvial terrace deposits, high-level gravel deposits, colluvium,
landslides and quartzite erratic boulders, well described by Kempter et
al. (2007). Younger alluvium (late Pleistocene to Holocene) consists of
sands, silts and conglomerates that lie at or within 2 m above modern
drainage bottoms. Older alluvium found in the bottom of Cañon del
Cobre is composed of sand with minor gravel and silty-clayey beds
preserved as 0.2 to 2.0-m-thick terrace deposits on top of sandstone
benches. Late Pliocene-middle Pleistocene gravels may form relatively
thin veneers on high-level erosion surfaces west of the western escarp
ment of Cañon del Cobre (D. Koning, written commun., 2009). Collu
vium is mostly poorly sorted talus and rock debris, often as hillslope
deposits up to 8 m thick, and is of likely late Pleistocene to Holocene age.
Landslide deposits on steep slopes are thicker, some more than 20 m
thick, and composed of unsorted, chaotic debris that was emplaced dur
ing single, mass-wasting events, such as slump or block slides (Kempter
et al., 2007). Kempter et al. (2007) suggested a catastrophic flood event
emplaced Proterozoic quartzite boulders over 2 m in diameter, particu
larly on Poleo Formation dipslopes northeast of Las Minas Jimmie.

STRUCTURE
Cañon del Cobre is located on the southeastern edge of the Colo
rado Plateau, where the Plateau adjoins the Abiquiu embayment of the
Rio Grande rift. Westward-dipping strata on the west side of the canyon,
and eastward-dipping strata on its east side give Cañon del Cobre the
overall appearance of a large, breached anticlinal structure.
Cañon del Cobre is traversed and bounded by several east-north
east-trending, high angle normal faults, and its eastern wall is along a
faulted, north-northeast-trending anticline that plunges to the southwest
(Smith et al., 1961; Kempter et al., 2007) (Figs. 2, 8). The faults in Cañon
del Cobre are down to the north, east or west and most have less than 50
m of offset. These structures are part of the broad boundary between the
eastern Colorado Plateau and the Rio Grande rift. Most of them formed
during the Neogene extensional tectonism associated with the develop
ment of the rift.
ACKNOWLEDGMENTS
FIGURE 7. Measured stratigraphic section of the Chinle Group strata exposed
We thank several colleagues and volunteers for field assistance in
at Minas de Pedro on the eastern rim of Cañon del Cobre. See Appendix 2
for description of section. Modified from Lucas and Hunt (1992).
studying the geology of Cañon del Cobre. Dan Koning, John Nelson and
Joerg Schneider provided helpful reviews of the manuscript.
22
REFERENCES
Ash, S.R., 1974, Upper Triassic plants of Cañon del Cobre, New Mexico:
New Mexico Geological Society, Guidebook 25, p. 179-184.
Chapin, C. E. and Cather, S. M., 1981, Eocene tectonics and sedimentation
in the Colorado Plateau-Rocky Mountain area: Arizona Geological Soci
ety Digest, v. 14, p. 173-198.
Eberth, D.A. and Berman, D.S, 1983, Sedimentology and paleontology of
Lower Permian fluvial redbeds of north-central New Mexico – prelimi
nary report: New Mexico Geology, v. 5, p. 21-25.
Eberth, D.A. and Berman, D.S, 1993, Stratigraphy, sedimentology and ver
tebrate paleontology of the Cutler Formation redbeds (Pennsylvanian
Permian) of north-central New Mexico: New Mexico Museum of Natu
ral History and Science, Bulletin 2, p. 33-48.
Eberth, D.A. and Miall, A.D., 1991, Stratigraphy, sedimentology and evolu
tion of a vertebrate-bearing, braided to anastomosed fluvial system,
Cutler Formation (Permian-Pennsylvanian), north-central New Mexico:
Sedimentary Geology, v. 72, p. 225-252.
Fracasso, M.A., 1987, Fluvial deposition and paleoenvironment of the
Cutler Formation red beds, El Cobre Canyon, New Mexico: New Mexico
Geology, v. 9, p. 14-18.
Heckert, A. B. and Lucas, S. G., 2003, Triassic stratigraphy in the Zuni
Mountains, west-central New Mexico: New Mexico Geological Society,
Guidebook 54, p. 245-262.
Kempter, K., Zeigler, K., Koning, D. and Lucas, S., 2007, Preliminary
geologic map of the Canjilon SE quadrangle, Rio Arriba County, New
Mexico: New Mexico Bureau of Geology and Mineral Resources, Open
file Geological Map 150, 1:24,000.
Logsdon, M. J., 1981, A preliminary basin analysis of the El Rito Formation
(Eocene), north-central New Mexico: Geological Society of America
Bulletin, pt. I, v. 92, p. 968-975.
Lucas, S.G., 1984, Correlation of Eocene rocks of the northern Rio Grande
rift and adjacent areas: Implications for Laramide tectonics: New Mexico
Geological Society, Guidebook 35, p. 123-128.
Lucas, S.G., 1993, The Chinle Group: Revised stratigraphy and biochronology
of Upper Triassic strata in the western United States: Museum of North
ern Arizona, Bulletin 59, p. 27-50.
Lucas, S.G. and Hunt, A.P., 1992, Triassic stratigraphy and paleontology,
Chama basin and adjacent areas, north-central New Mexico: New Mexico
Geological Society, Guidebook 43, p. 151-167.
APPENDIX1
Cutler Group Measured Section, Cañon de Cobre
Measured along the western floor and wall of El Cobre Canyon in
secs. 25-26, T24N, R5E. Base at UTM Zone 13, 378369E, 4016319
(NAD27) and top at 376985E, 4016278N. Strata dip 8° to S40°W. This
is section in Figure 5; also see Lucas and Krainer (2005a). Thicknesses
are in meters.
Late Triassic:
Chinle Group:
Shinarump Formation:
85.Conglomeratic sandstone; very pale orange (10YR8/2) and
moderate yellowish brown (10YR5/4); quartzose; coarse
grained; not calcareous; clasts are chert, jasper and quartzite
pebbles; trough crossbedded. Not measured.
unconformity
Permo-Pennsylvanian:
Cutler Group:
Lucas, S.G. and Krainer, K., 2005a, Stratigraphy and correlation of the
Permo-Carboniferous Cutler Group, Chama basin, New Mexico: New
Mexico Geological Society, Guidebook 56, p. 145-159.
Lucas, S.G. and Krainer, K., 2005b, Cutler Group (Permo-Carboniferous)
stratigraphy, Chama Basin, New Mexico: New Mexico Museum of Natu
ral History and Science, Bulletin 31, p. 90-100.
Lucas, S. G., Harris, S. K. and Spielmann, J. A., 2010, One hundred and fifty
years of geological and paleontological research in Cañon del Cobre,
northern New Mexico: New Mexico Museum of Natural History and
Science Bulletin, this volume.
Lucas, S.G., Harris, S.K., Spielmann, J.A., Berman, D.S and Henrici, A.C.,
2005b, Vertebrate biostratigraphy and biochronology of the Pennsylva
nian-Permian Cutler Group, El Cobre Canyon, northern New Mexico:
New Mexico Museum of Natural History and Science, Bulletin 31, p.
128-139.
Lucas, S.G., Zeigler, K.E., Heckert, A.B. and Hunt, A.P., 2003, Upper Tri
assic stratigraphy and biostratigraphy, Chama basin, north-central New
Mexico: New Mexico Museum of Natural History and Science, Bulletin
24, p. 15-39.
Lucas, S.G., Zeigler, K.E., Heckert, A.B. and Hunt, A.P., 2005a, Review of
Upper Triassic stratigraphy and biostratigraphy in the Chama basin,
northern New Mexico: New Mexico Geological Society, Guidebook 56,
p. 170-181.
Maldonado, F. and Kelley, S.A., 2009, Revision to the stratigraphic nomen
clature of the Abiquiu Formation, Abiquiu and contiguous areas, north
central New Mexico: New Mexico Geology, v. 31, p. 3-8.
Newberry, J.S., 1876, Geological report; in Macomb, J.N., ed., Report of the
exploring expedition from Santa Fe…in 1859: U.S. Army Engineer
Department, Washington, D.C., p. 9-118.
Smith, C.T., Budding, A.J. and Pitrat, C.W., 1961, Geology of the southeast
ern part of the Chama basin: New Mexico Bureau of Mines and Mineral
Resources, Bulletin 75, 57 p.
Stewart, J.H., Poole, F.G. and Wilson, R.F., 1972, Stratigraphy and origin of
the Chinle Formation and related Upper Triassic strata of the Colorado
Plateau region: U.S. Geological Survey, Professional Paper 690, 336 p.
Wood, G.H. and Northrop, S.A., 1946, Geology of the Nacimiento Moun
tains, San Pedro Mountain, and adjacent plateaus of Sandoval and Rio
Arriba counties, New Mexico: U.S. Geological Survey, Oil and Gas Inves
tigations Map OM-57.
Arroyo del Agua Formation:
84.Sandy siltstone; mottled pale red (10R6/2) and light greenish
gray (5GY8/1); not calcareous; blocky. 1.5
83.Sandy siltstone; same as unit 75. 6.0
82. Sandstone; yellowish gray (5Y8/1) and moderate reddish brown
(10R4/6); arkosic; coarse grained; not calcareous; trough
crossbedded; bench. 3.4
81.Sandy siltstone; moderate reddish brown (10R4/6); micaceous;
not calcareous. 1.3
80.Sandy siltstone; same as unit 75. 6.0
79.Sandstone; same as unit 76.3.3
78.Calcrete pebble conglomerate; pale yellowish brown (10YR6/
2) and light olive gray (5Y6/1); bench. 0.7
77.Sandy siltstone; same as unit 75. 4.5
76.Sandstone; moderate reddish brown (10R4/6) and pale reddish
brown (10R5/4); arkosic; coarse grained; not calcareous; trough
crossbedded; multistoried. 2.6
75.Sandy siltstone; moderate reddish brown (10R4/6); slightly
calcareous; some calcrete nodules; slope. 27.0
74.Sandstone; moderate reddish brown (10R4/6); arkosic; coarse
grained; trough crossbeded. 1.5

73.Muddy siltstone; moderate reddish brown (10R4/6); calcrete
nodules. 14.0
72.Conglomeratic sandstone; pale reddish brown (10R5/4) and
moderate reddish brown (10R4/6); arkosic; coarse grained;
conglomerate is siltstone and calcrete rip-ups; trough
crossbedded; scour base; multistoried bench. 5.7
71.Muddy siltstone; same as unit 73. 5.2
70. Sandstone; yellowish gray (5Y8/1) and moderate reddish brown
(10R4/6); arkosic; coarse grained; calcareous; trough
crossbedded; multistoried. 2.8
69.Sandstone; moderate reddish brown (10R4/6); arkosic;
micaceous; coarse grained; slightly calcareous; in ledges (0.3-m
thick) separated by blocky siltstone like unit 68; some calcrete.
6.0
68.Sandy siltstone; same as unit 60; many calcrete nodules. 7.0
67.Sandstone; same as unit 63. 2.0
66.Sandy siltstone; same as unit 60. 6.0
65.Conglomeratic sandstone; sandstone is light greenish gray
(5GY8/1) and moderate reddish brown (10R4/6); arkosic, coarse
grained and calcareous; clasts are calcrete pellets; trough
crossbedded. 1.2
64.Sandy siltstone; same as unit 60. 1.6
63.Sandstone; pale reddish brown (10R5/4) and moderate reddish
brown (10R4/6); arkosic; coarse grained; trough crossbedded;
scour base with 3 m of relief. 6.0
62.Sandy siltstone; same as unit 60. 0.8
61.Sandstone; same as unit 61. 0.7
60.Sandy siltstone; moderate reddish brown (10R4/6); not
calcareous; blocky; calcrete nodules. 4.6
59.Sandy siltstone; same as unit 54; abundant calcrete. 4.0
58.Sandstone; moderate reddish brown (10R4/6); arkosic; coarse
to very coarse grained; calcareous; some calcrete pebbles; trough
crossbedded; scour base locally cuts down to unit 55. 3.0
57.Sandy siltstone; grayish red (10R4/2); calcareous; blocky. 1.0
56.Sandy siltstone; same as unit 54.6.2
55.Sandstone; pebbly at base; moderate reddish brown (10R4/6)
and pale red (10R6/2); arkosic; coarse to very coarse grained;
calcareous; pebbles are siltstone and calcrete rip ups; trough
crossbedded; bench. 1.7
54.Sandy siltstone; moderate reddish brown (10R4/6); calcare
ous; blocky. 8.7
El Cobre Canyon Formation:
53.Sandstone; pale reddish brown (10R5/4); fine grained; arkosic;
very calcareous; laminar and bioturbated. 1.7
52.Sandy siltstone; same as unit 44. 1.6
51.Sandstone; same as unit 47. 0.6
50.Sandy siltstone; same as unit 44. 0.2
49.Sandstone; same as unit 47. 0.3
48.Sandy siltstone; same as unit 44.1.0
47. Sandstone; pale reddish brown (10R5/4) and grayish red (10R4/
2); fine grained; arkosic; calcareous; massive. 0.8
46.Sandy siltstone; same as unit 44. 1.5
45.Sandstone; pale reddish brown (10R5/4); arkosic; fine grained;
calcareous; bioturbated. 0.6
44.Sandy siltstone; pale reddish brown (10R5/4); calcareous;
calcrete; blocky. 2.0
43.Conglomeratic sandstone; pale reddish brown (10R5/4); coarse
to very coarse grained; arkosic; clasts are quartzite and calcrete
pebbles; very calcareous; trough crossbedded; scour base; bench.
2.7
42.Sandy siltstone; same as unit 44 except laminar. 1.8
41.Sandy siltstone; same as unit 44. 7.3
40.Sandstone; same as unit 38. 0.6
39.Sandy siltstone; same as unit 44.3.8
23
38.Sandstone; pinkish gray (5YR8/1) and light brownish gray
(5YR6/1); arkosic; coarse grained; very calcareous; ripple
laminated. 0.3
37.Sandy siltstone; pale reddish brown (10R5/4); abundant
calcrete nodules; blocky; slope. 2.7
36.Sandstone; same as unit 38; thin bioturbated ledge. 1.8
35.Sandy siltstone; same as unit 37; some thin sandstone lenses
like unit 38. 3.2
34.Sandstone; same as unit 32. 0.5
33.Sandy siltstone; same as unit 31.1.0
32.Sandstone; pinkish gray (5YR8/1) and light olive gray (5Y6/1);
arkosic; calcareous; coarse grained; bioturbated. 1.0
31.Sandy siltstone; pale reddish brown (10R5/4); micaceous;
numerouscalcrete nodules; blocky slope; a few lenses of
sandstone like unit 32.4.0
30.Sandstone; yellowish gray (5Y8/1) and light olive gray (5Y6/
1); arkosic; coarse grained; calcareous; trough crossbedded. 3.5
29.Sandstone; same as unit 27, bench. 2.9
28.Sandy siltstone; pale red (10R6/2) and pale reddish brown
(10R5/4); laminar; calcareous. 0.8
27.Sandstone; yellowish gray (5Y8/1); arkosic; coarse grained;
very calcareous; trough crossbedded; bench. 2.9
26.Sandy shale; light olive gray (5Y6/1); not calcareous; plant
debris. 2.5
25.Conglomeratic sandstone; yellowish gray (5Y8/1) and moder
ate orange pink (10R7/4); arkosic; coarse grained; very
calcareous; clasts are quartzite up to 6 cm diameter; trough
crossbedded. 1.9
24.Sandy siltstone; same as unit 18.1.6
23.Calcrete ledge; light olive gray (5Y6/1); some associated calcrete
pebble conglomerate. 1.3
22.Sandy siltstone; same as unit 18; blocky; slope. 3.1
21.Calcrete ledge; same as unit 23.0.1
20.Sandy siltstone; same as unit 18. 0.7
19.Sandstone; pale red (10R6/2); arkosic; fine grained; very
calcareous; massive; cuesta. 0.6
18.Sandy siltstone; pale reddish brown (10R5/4); micaceous; not
calcareous; numerous rhizoliths. 7.5
17.Conglomerate; light olive gray (5Y6/1); calcrete pebbles in
coarse-grained arkosic sandstone matrix; trough crossbedded;
bench. 1.6
16.Sandy siltstone; same as unit 18.3.3
15.Sandstone; same as unit 6. 0.2
14.Sandy siltstone; same as unit 7.1.9
13.Sandstone; yellowish gray (5Y8/1) and moderate orange pink
(10R7/4); arkosic; coarse grained; calcareous; trough
crossbedded; multistoried bench. 2.4
12.Sandy siltstone; same as unit 7 except trough crossbedded. 2.2
11.Sandy siltstone; same as unit 7.5.0
10.Sandstone; pale reddish brown (10R5/4); arkosic; coarse
grained; not calcareous; trough crossbedded; top surface has
desiccation tracks. 1.2
9. Sandy siltstone; same as unit 7.2.2
8. Sandstone; pale reddish brown (10R5/4); arkosic; medium
grained; very calcareous; some calcrete rip-ups; trough
crossbedded; multistoried bench. 2.0
7. Sandy siltstone; grayish red (10R4/2) and pale reddish brown
(10R5/4); micaceous; not calcareous; abundant rhizoliths. 1.6
6. Sandstone; light olive gray (5Y6/1); arkosic; micaceous;
medium grained; calcareous; ripple laminated. 0.4
5. Conglomerate; light olive gray (5Y6/1); calcrete pebbles and a
sparse matrix of arkosic sand grains; bench. 0.2
4. Sandy siltstone; pale reddish brown (10R5/4) and grayish red
(10R4/2); micaceous; not calcareous; laminar. 5.5
3. Conglomeratic sandstone; same as unit 1.3.2
24
2. Sandy siltstone; pale reddish brown (10R5/4); micaceous;
calcareous; blocky. 7.7
1. Conglomeratic sandstone; pale reddish brown (10R5/4); arkosic;
micaceous; very calcareous; coarse grained; clasts are gray
quartzite; trough crossbedded; multistoried. 0.9
APPENDIX2
Chinle Group Section, El Cobre Canyon
Measured in the SE¼ SW¼ sec. 18 (unsurveyed), T24N, R06E,
Rio Arriba County. Modified Minas de Pedro section of Lucas and Hunt
(1992, p. 172). See Figure 7 for graphic illustration of section. Thick
nesses are in meters.
Chinle Group:
Poleo Formation:
15.Sandstone with conglomeratic beds: sandstone is light olive
brown (5Y 5/6), very fine grained, subangular-subrounded,
moderately sorted, noncalcareous, micaceous litharenitic;
conglomerate is light greenish gray (5GY 8/1) and composed of
siltstone and nodular calcrete clasts; laminar and planar
crossbedded with conglomerate concentrated at scour base.
11.3
14.Conglomerate: matrix is pale olive (10Y 6/2), very fine-coarse
grained, subangular quartzose sandstone; clasts are brown,
white and red chert up to 1 cm in diameter; trough crossbedded.
1.6
13.Sandstone: yellowish gray (5Y 7/2); very fine-fine grained;
subrounded; moderately sorted; micaceous subarkosic; very
calcareous; laminar. 2.2
12.Sandstone: yellowish gray (5Y 7/2); very fine-fine grained;
subangular-subrounded; moderately sorted; micaceous
quartzarenitic; very calcareous; trough crossbedded;
multistoried with pebbly beds at scour bases. 10.3
11.Sandstone: pale greenish yellow (10Y 8/2), weather moderate
yellowish brown (10Y 5/4); fine grained; subrounded; well
sorted; quartzarenitic; noncalcareous; trough crossbedded. 2.0
10.Sandstone:grayish orange (10R 7/4); very fine grained;
subrounded; well sorted; arkosic litharenite; calcareous; some
cover. 1.3
9. Conglomeratic sandstone: sandstone is dusky yellow (5Y 6/
4), very fine-fine grained, subangular, moderately sorted,
micaceous litharenitic, calcareous; pebbles are light gray, black,
white quartzite and petrified wood; trough crossbedded; scour
base with 0.3-0.6 m of relief. 3.0
unconformity (Tr-4)
Salitral Formation:
8. Sandy mudstone and muddy sandstone: sandy mudstone is
mottled grayish red purple (5RP4/2) and yellowish gray (5Y
7/2), bentonitic and very calcareous; muddy sandstone is
medium light gray (N6), fine grained, subangular, poorly sorted,
lithic wacke; trough crossbedded ledges with mudstone
partings. 5.5
Shinarump Formation:
7. Sandstone: pale greenish yellow (10Y 8/2); fine-medium grained;
subrounded; moderately sorted; quartzose; slightly
calcareous; trough crossbedded; some siliceous pebbles. 3.0
6. Sandstone: yellowish gray (5Y 7/2); very fine-medium grained;
subrounded; poorly sorted; feldspathic litharenitic;
noncalcareous; trough crossbedded; contains some siliceous
pebbles; upper half of unit thinner bedded with some
mudstone partings. 3.1
5. Sandy mudstone: pale reddish purple (5RP6/2); noncalcareous.
1.5
4. Sandstone: yellowish gray (5Y 7/2); fine-medium grained;
subrounded; moderately sorted; micaceous feldspathic
litharenitic; calcareous; trough crossbedded; some dispersed
siliceous pebbles. 2.8
3. Conglomerate: matrix is pale green (5G 7/2), fine-medium
grained, subangular, poorly sorted quartzose sandstone; clasts
are greenish black (5G 2/1) chert and white quartzite up to 15
cm in diameter; trough crossbedded. 0.5
unconformity (Tr-3)
Mottled strata (Zuni Mountains Formation):
2. Siltstone: mottled grayish yellow green (5GY 7/2) and dark
reddish brown (10R 3/4); bioturbated and massive. 1.2-1.4
Cutler Group:
Arroyo del Agua Formation:
1. Sandstone: moderate reddish brown (10R4/6); very fine-fine
grained; subangular; moderately sorted; micaceous lithic wacke;
noncalcareous; trough crossbedded. 2.0+
 Lucas et al., eds., 2010, Carb-Permian transition in Cañon del Cobre. New Mexico Museum of Natural History and Science Bulletin 49.

25
SEDIMENTOLOGY OF THE PENNSYLVANIAN-PERMIAN CUTLER GROUP
AND LOWER PERMIANABO FORMATION, NORTHERN NEWMEXICO

KARL KRAINER1 AND SPENCERG. LUCAS2

of Geology and Paleontology, Innsbruck University, Innrain 52, Innsbruck, A-6020 AUSTRIA;
1Institute

2 New Mexico Museum of Natural History and Science, 1801 Mountain Road NW, Albuquerque, NM 87104-1375

Abstract—The Cutler Group at Cañon del Cobre, northern New Mexico, is divided into the El Cobre Canyon and
Arroyo del Agua formations, both composed of red beds that formed in a braided river system under semi-arid to
arid climatic conditions. Sandstones are dominantly arkoses to lithic arenites, and rarely sublitharenites. Towards
the south the Cutler red beds grade into the Abo Formation, which is well exposed along the Jemez River. Red beds
of the Abo Formation at the Jemez River and Gilman sections do not show any significant differences in facies and
mineralogical composition when compared to the Cutler red beds. Sandstone sheets represent deposits of a braided
river system; the intercalated sandstone beds and lenses are interpreted as sheet splays and minor channel fills
which formed during overbank flooding. Siltstone and mudstone units, which are the dominant lithofacies of the
Abo Formation, represent floodplain deposits. Whereas the Cutler Group includes red beds of Pennsylvanian age,
the Abo Formation is restricted to the Early Permian. As the red beds of the Cutler Group and Abo Formation are
very similar, a sharp boundary cannot be drawn, so we consider the Abo Formation to be part of the Cutler Group.

INTRODUCTION
Siliciclastic red beds of Late Pennsylvanian and Early Permian age
are exposed both north and south of the Jemez Mountains volcanic
edifice in northern New Mexico. North of the Jemez Mountains, in the
Chama River basin, these strata are assigned to the El Cobre Canyon and
the Arroyo del Agua formations of the Upper Pennsylvanian-Lower
Permian Cutler Group. However, to the south, the red beds are assigned
to the Lower Permian Abo Formation. A particularly significant outcrop
of these strata is in Cañon del Cobre (El Cobre Canyon) of Rio Arriba
County, where the type section of the El Cobre Canyon Formation is
overlain by a characteristic section of the Arroyo del Agua Formation.
The nonmarine red beds of the El Cobre Canyon and Arroyo del Agua
formations at Cañon del Cobre are particularly well known for their
fossil vertebrates, and also yield significant trace fossils, plant megafossils
and palynomorphs (see Lucas et al., this volume). Here, we present an
analysis of Abo-Cutler sedimentology in northern New Mexico to deter
mine depositional environments and to aid in the regional correlation of
these strata.
CUTLER GROUP, CAÑON DELCOBRE

Introduction
Red beds between the underlying Rico Formation and overlying
Triassic strata in southern Colorado, southeastern Utah, northeastern
Arizona and northwestern New Mexico were originally termed Cutler
Formation, named after Cutler Creek, west of Ouray, Colorado (see Jicha
and Lochman-Balk, 1958). Due to strong lateral and vertical facies changes
within the “Cutler Formation,” the succession was elevated to group
rank and divided into several formations of distinctive lateral and vertical
extent (Baars, 2000). In southeastern Utah, the Cutler Group FIGURE 1. Index map of part of northern New Mexico showing location of
unconformably overlies the Upper Pennsylvanian (Missourian) Honaker the measured sections at Cañon del Cobre (El Cobre Canyon), the Jemez
Trail Formation and is divided into Elephant Canyon Formation, Halgaito River and Gilman.
Shale, Cedar Mesa Sandstone, Organ Rock Shale and White Rim Sand
stone (e.g., Baars, 2003). In the Chama basin of northern New Mexico, includes strata of the El Cobre Canyon and Arroyo del Agua formations.
the Cutler Group consists of the El Cobre Canyon and overlying Arroyo Fracasso (1987) discussed the sedimentary environment of the
del Agua formations (Lucas and Krainer, 2005a, b). The most completely Cutler Formation at El Cobre Canyon, suggesting deposition of the red
exposed Cutler Group section in this area is at Cañon del Cobre (El beds in a single-channellow sinuosity stream system in an upper alluvial
Cobre Canyon), near Abiquiu (Fig. 1). At Cañon del Cobre (Fig. 1), we plain near a distal alluvial fan and braided-stream complex under semiarid
measured two sections: section A (Fig. 2) is 100 m thick and comprises climatic conditions. Fracasso (1987) did not subdivide the sedimentary
the El Cobre Canyon Formation, and section B (Fig. 3) is 255 m thick and succession.
26
FIGURE 2. Measured section of the El Cobre Canyon Formation (section
A) at Cañon del Cobre. See Appendix for map coordinates of section.
A detailed facies analysis of the Cutler Group at El Cobre Canyon
was presented by Eberth and Miall (1991) and Eberth and Berman (1993).
The El Cobre Canyon Formation corresponds to megasequence 1 of
Eberth and Miall (1991), and the Arroyo del Agua Formation corre
sponds to megasequences 2 and 3 of Eberth and Miall (1991), although
the boundary between El Cobre Canyon and Arroyo del Agua formations
does not correspond to the boundary between megasequence 1 and 2 but
lies slightly higher, within the lower part of megasequence 2 (Lucas and
Krainer, 2005a, b). Here, we base our interpretation of deposition of the
El Cobre Canyon Formation primarily on data gathered from the two
sections we measured (Fig. 2-3).
Lithofacies
Except for the basal 11 m of section A, the sediments are brown
ish-red (pale reddish brown) colored in the El Cobre Canyon Formation
and orange red (moderate reddish brown) in the Arroyo del Agua Forma
tion. Lithofacies are classified using the facies codes of Miall (1996). All
lithofacies are listed (Table 1) and were described in detail by Eberth and
Miall (1991).
The basal 11 m of section A (El Cobre Canyon Formation: Fig. 2)
is composed of horizontally-laminated greenish siltstone to fine-grained
sandstone (lithofacies Sh), overlain by greenish to dark gray shale with
plant fossils (lithofacies Fl), and two horizontally-laminated brownish
sandstone beds 0.3 and 0.2 m thick (Sh), separated by 0.6 m of greenish
gray shale (Fl). The sandstone is overlain by gray shale containing plant
fossils, grading into greenish siltstone (Fl) that is sharply overlain by
coarse-grained, red arkosic sandstone.
Conglomerate and sandstone occur as individual beds intercalated
in mudstone-siltstone or as thicker horizons that extend laterally over
more than 100 m, forming sandstone sheets. In the measured sections,
sandstone sheets are up to 10.5 m thick. Thicker sandstone sheets dis
play a fining-upward trend with a thin conglomerate at the base (Gcm,
Gt), grading into large-scale to small-scale trough cross-bedded sand
stone (St). Locally, medium- to fine-grained, horizontally-laminated sand
stone (Sh) and massive sandstone (Sm), and, rarely, ripple-cross-lami
nated sandstone (Sr), form the tops of sandstone sheets. Eberth and
Miall (1991) distinguished major and minor sandstone sheets, which
they described in detail. Locally, conglomerate and trough cross-bedded
sandstone are erosively cut into the underlying mudstone-siltstone, form
ing a relief up to 6.3 m deep and representing sandstone ribbons (see
Eberth and Miall, 1991 for details).
Conglomerate occurs as lithofacies Gcm (massive, clast supported;
lithofacies Gm of Eberth and Miall, 1991) and Gt (trough cross-bedded).
Maximum clast size is mostly 5 to 6 cm, rarely up to 15 cm. The
dominant clast type is carbonate derived from reworking of caliche hori
zons and caliche nodules within red mudstone-siltstone. Subordinately,
extraformational clasts are present. Conglomerate occurs at the base of
sheet sandstones and ribbon sandstones and as single conglomerate beds
intercalated in red mudstone-siltstone. Conglomerates are erosively cut
into underlying red mudstone-siltstone.
Trough cross-bedding (lithofacies St) is the most common sedi
mentary structure of sandstone. Large-scale trough cross-bedding sets
thicker than 1 m are rare and represent lateral accretion deposits (lithofacies
Sla of Eberth and Miall, 1991). Medium-scale (> 50 cm) and small-scale
(< 50 cm) trough cross-bedded sandstone represents channel-fill depos
its (Sti and Stii of Eberth and Miall, 1991). Coarse sandstone may be
pebbly.
Lithofacies Se contains mudstone-siltstone rip-up clasts and rep
resents erosional scours. Sandstone with low-angle cross-bedding is very
rare. Horizontally laminated sandstone (Sh) and massive sandstone (Sm)
occur as individual sandstone beds up to 0.5 m thick intercalated in
mudstone-siltstone and in the upper part of sandstone sheets associated
with trough cross-bedded sandstone. Ripple cross-lamination (Sr) is
very rare. Medium- to fine-grained sandstone occasionally is bioturbated
(principally Planolites). Rarely, desiccation cracks are observed in sand
stone.
Mudstone-siltstone is more abundant in the Arroyo del Agua
Formation than in the El Cobre Canyon Formation (Fig. 3). Mudstone
siltstone intervals are up to 7.5 m thick in the Arroyo del Agua Forma
tion, and up to 27 m thick in its uppermost strata. The color is brownish
red in the El Cobre Canyon Formation and orange-red in the Arroyo del
Agua Formation. Mudstone-siltstone is mostly massive (lithofacies Fsm),
rarely laminated (Fl) and locally contains caliche nodules that are concen
trated in thin layers (lithofacies Pc). A few mudstone-siltstone horizons
contain rhizoliths (lithofacies Fr), and rare desiccation cracks are ob
served. Mudstone-siltstone is summarized under lithofacies Flin Eberth
and Miall (1991).
Besides pedogenic limestone (caliche, lithofacies Lp), Eberth and
Miall (1991) distinguished other limestone types such as limestone nod
ules of diagenetic origin (Ld), lacustrine marls (Lm) and limestone of
travertine/tufa origin (Lt).
 27

FIGURE 3. Type section of the El Cobre Canyon Formation at Cañon del Cobre. See Appendix for map coordinates of section.
28
Table 1. Lithofacies.
Petrography
Sandstones of the Cutler Group are mostly medium- to coarse
grained, moderately to poorly sorted, angular arkoses to lithic arenites,
and rarely sublitharenites (see details in Lucas and Krainer, 2005a, b; also
see Table 2).
Depositional Environments
Our data are consistent with and support the interpretations of
depositional environments presented by Eberth and Miall (1991) and
Eberth and Berman (1993). They recognized eight architectural elements
in Cutler Group strata, including major sandstone sheets and ribbons,
minor sandstone sheets and lenses, stacked sheets and lenses, U-shaped
mixed-fill, and two types of floodplain deposits. The dominant architec
tural element composed of coarse clastics is major sandstone sheets that
are interpreted as braided, fluvial channel deposits. Major sandstone
ribbons are less common and interpreted to represent anastomosed flu
vial channels. The remaining architectural elements are interpreted as
overbank deposits, including crevasse channel and splay deposits, and
floodplain deposits.
During deposition of the El Cobre Canyon Formation (approxi
mately correlating to megasequence 1 of Eberth and Miall, 1991), broad
and shallow channels of a braided river system (major sandstone sheets)
developed under semiarid climatic conditions. Overbank areas were fre
quently inundated by sheet floods and splays, causing deposition of
minor sandstone sheets and lenses, and stacked sheets and lenses. Small
ponds filled with fine-grained lacustrine sediments were common (for
example, the pond deposit at the base of section A).
During deposition of theArroyo del Agua Formation (megasequence
2 of Eberth and Miall, 1991), channels became narrower and between the
channels extensive floodplains developed on which locally pedogenic
carbonate horizons (caliche) formed. According to Eberth and Miall
(1991), renewed uplift caused development of a poorly confined shallow
braided river system and deposition of coarse clastic sediments
TABLE 2. Mineralogic composition of Cutler Group sandstones. For
acronyms (column heads), see captions to Figures 9 and 10.
(megasequence 3, uppermost part of the Arroyo del Agua Formation).
Climate changed from semi-arid to increasingly arid conditions during
sedimentation of the Cutler Group.
ABO FORMATION (JEMEZRIVERAND GILMAN)
Introduction
Along the Jemez River, the measured section of the Abo Forma
tion is 143 m thick (Fig. 4). The base is not exposed, and Abo red beds are
overlain by eolian sandstone of the “Meseta Blanca Sandstone Member
of the Yeso Formation” (DeChelly Formation of the Yeso Group: Lucas
et al., 2005). At Gilman, the Abo Formation is completely exposed, and
the measured thickness is 131 m (Figs. 5-6). The Abo Formation is
underlain by the Guadalupe Box Formation (Krainer et al., 2005) and
overlain by a few meters of Yeso red beds and lower Pleistocene Bandelier
Tuff.
At both localities, the Abo can be divided into lower and upper
members. The lower member is dominated by mudstone and siltstone
that constitute 69% of the Abo section at the Jemez River and 72% at
Gilman. In the upper member, mudstone and siltstone constitute much
less of the Abo section— 25% at the Jemez River and 41% at Gilman.
The lower member measures 83 m at the Jemez River and 90 m at
Gilman, and the upper member is 60 m at the Jemez River and 41 m at
Gilman. The lower member probably corresponds to the upper part of
the El Cobre Canyon Formation, and the upper member to the Arroyo
del Agua Formation of the Cutler Group (Lucas and Krainer, 2005a, b),
and, respectively, to the Scholle Member and Cañon de Espinoso Mem
ber of the Abo type section in central New Mexico (Lucas et al., 2005).
Lithofacies
Lithofacies were classified according to the scheme proposed by
Miall (1978, 1981, 1996) using the facies codes of Miall (1996). In
general, in the Abo Formation the same lithofacies types as in the Cutler
Group are present (lithofacies types are listed in Table 1).
Lithofacies Gcm and Gt at the Jemez River contain intraclasts,
mostly carbonate clasts that are derived from caliche. Subordinate mud
stone clasts (rip-up clasts) are present. At Gilman, rare (unit 31 of
section A and unit 14 of section B) extraformational clasts including rare
granite and abundant well rounded quartzite clasts occur. Lithofacies Se
is rare and occurs as thin lenticular beds containing abundant red mud
stone rip-up clasts.
Trough cross-bedding (St) is the most common sedimentary struc
ture, occurring at different scales (Fig. 7). Large-scale trough-crossbedding
(St-l), which represents lateral accretion units, is rare and occurs in the
lower part of sandstone sheets, grading upward into medium- to small
scale trough cross-bedded sandstones (St-m, St-s) and horizontally-lami
nated sandstone (Sh). Sandstone displaying large-scale trough
crossbedding may be pebbly, containing individual clasts up to 5 cm in
diameter.
Medium- to small-scale trough cross-bedding (St-m, St-s) is the
most common sedimentary structure of sandstone sheets, forming stacked
cosets of multistory channel fill successions with fining-upward trends
and internal erosion surfaces. Low-angle cross-bedding (Sl) is rare and
occurs as thin beds associated with medium-scale trough cross-bedding
 29

FIGURE 4. Measured section through the Abo Formation at Jemez River. See Appendix for map coordinates of section.
30
FIGURE 5. Measured section through the Abo Formation (lower part) at
Gilman. See Appendix for map coordinates of section. Unit 25 = unit 1 in
section B (Fig. 6).
and horizontal lamination in the upper part of sandstone sheets.
Sandstone with horizontal lamination (Sh) occurs as thin beds
intercalated in mudstone or in the upper, finer-grained part of sandstone
sheets. Massive to crudely stratified sandstone (Sm) occurs as interca
lated beds up to 0.3 m thick within mudstone or associated with trough
cross-bedded (St-m, St-s) and horizontally laminated sandstone (Sh) in
the upper part of sandstone sheets. Asymmetric current ripples with
ripple cross-lamination are rare at the Jemez River where they occur in
units 30 and 32 in the fine-grained upper part of sandstone sheets asso
ciated with horizontally-laminated sandstone. Ripples were not observed
at Gilman.
Mudstone-siltstone units are up to 13.4 m thick, and partly cov
ered. The color is red to purple, and the beds are massive (Fsm) or rarely
laminated (Fl). At the Jemez River desiccation cracks occur in mudstone
of unit 18 (Fm). The red mudstone of unit 31 at the Jemez River is red
greenish mottled, containing rhizoliths (lithofacies Fr; Fig. 8). Large cal
cified root structures are also present in coarse, pebbly sandstone (unit
28) at the Jemez River.
Bioturbated sandstone is present at Gilman and occurs as a 0.7-m
thick, dark red, fine-grained sandstone unit intercalated in purple-red
mudstone in the lower member (unit 26 of section A) and in medium- to
fine-grained sandstone on top of sandstone sheets (units 2, 21, 30) of
section B. Soft sediment deformation (convolute bedding) was observed
in one fine-grained sandstone bed on top of a fining upward sequence
(unit 20) at the Jemez River.
Many of the mudstone-siltstone units contain abundant small
carbonate nodules (caliche nodules). Pedogenic carbonate horizons (cali
che, lithofacies P) occur in the lower member of both sections. At the
Jemez River a pedogenic horizon 1.2 m thick composed of pedogenic
limestone nodules in red mudstone occurs near the base (unit 2). Another
pedogenic horizon (unit 14) is 1.4 m thick and composed of dispersed
pedogenic limestone nodules in red mudstone, overlain by a nodular
caliche bed. At Gilman, two thin pedogenic carbonate horizons 0.3 and
0.5 m thick occur in the lower member in the red to purple mudstone of
unit 1 in section B. Another pedogenic carbonate horizon (nodular cali
che) 0.3 m thick is intercalated in red siltstone in the upper part of the
lower member (unit 18 of section B).
Architectural Elements
According to Miall (1985, 1992, 1996), architectural elements are
characterized by a distinctive facies assemblage, internal geometry, exter
nal form and, partly, by a distinct vertical profile. Miall proposed eight
basic types. Each type represents a particular process occurring within
a depositional system. Within the Abo Formation we distinguished three
architectural elements—sandstone sheets, intercalated sandstone beds
and lenses and siltstone-mudstone.
Sandstone sheets
The most characteristic and distinct facies assemblages in the Abo
Formation are sandstone sheets (Fig. 8), which correspond to the archi
tectural element CH (channel). Sandstone sheets are characterized by
width-to-depth ratios greater than 15, whereas sandstone ribbons also
corresponding to the element CH, have ratios smaller than 15.
Within the Abo Formation sandstone sheets form prominent, re
sistantledges that can be traced laterally over long (commonly more than
100 m, up to several hundred meters) distances. The base is generally
erosive, and lithofacies commonly show an upward decrease in bed thick
ness and grain size. Common lithofacies are Gcm and Gt at the base,
grading into multistoried trough cross-bedded sandstone (St), which in
the upper part may be associated with lithofacies Sh, Sm, Sl and Sr.
Rarely, fine-grained sandstone in the uppermost part is bioturbated.
Synsedimentary deformation structures are rarely observed in fine-grained
sandstone in the upper part. Sandstone sheets are very similar to the
“major sandstone sheets” that Eberth and Miall (1991) described from

FIGURE 6. Measured section through the Abo Formation (upper part) at
Gilman. See Appendix for map coordinates of section. Unit 1 = unit 25 in
section A (Fig. 5).
31
the Cutler red beds. Sandstone ribbons have not been observed in the
studied sections.
Abundance of internal scours and erosional surfaces, presence of
conglomeratic lags at the base, variation in grain-size, and poorly- to
well-developed fining-upward sequences are indicative of ephemeral flu
vial regimes (e.g., Picard and High, 1973; Tunbridge, 1981, 1984; Miall,
1996). Sheet sandstones are interpreted as deposits of broad, shallow
channels of a probable braided river system. The lack of lateral accretion
deposits indicates that the channels were shallow, so that large macroform
bars could not develop.
The sandstone sheets are very similar to the “major sandstone
sheets” of the Cutler Group described in detail by Eberth and Miall
(1991). These authors compared the sheet sandstones to those of the
modern Platte River (model 9 of Miall, 1985).
Intercalated sandstone beds and lenses
Thin intercalated sandstone beds and lenses are mostly 10 to 30
cm, rarely up to 50 cm thick. They occur as single sandstone beds of
lithofacies St-s, St-m, Sh, Sl and Sm. Stacked sandstone units are up to 1
m thick. The sandstone beds occur as tabular or lens-shaped bodies; the
base may be erosive. Rarely, sandstone sheets are bioturbated. This
architectural element is similar to the “minor sandstone sheets and lenses”
that Eberth and Miall (1991) described from the Cutler red beds.
Intercalated, laterally restricted sheet sands and lenses are charac
teristic of overbank flooding, particularly in ephemeral systems (e.g.,
Williams, 1971; Picard and High, 1973; Tunbridge, 1981). The tabular
sandstone beds probably originated as sheet splays. Basal erosional
surfaces with mudstone rip-up clasts and reworked caliche clasts indi
cate high-energy conditions of rapid flooding and reworking of mudstone
and caliche beds on the floodplain. Lack of internal erosion and reactiva
tion surfaces within the intercalated sandstone beds is typical of sheet
like, non-channelized flow. Sandstone lenses represent minor channel
fills and may represent the feeder channels (crevasse channels) of the
sheet splays.
Siltstone-mudstone
Siltstone and mudstone units are the dominant lithofacies of the
Abo Formation, occurring as lithofacies Fsm, Fm, Fl and Fr (Fig. 8).
Rarely, pedogenic limestone is intercalated.
The siltstone-mudstone facies that forms sheet-like units extend
ing laterally over at least hundreds of meters belongs to element FF
(floodplain fine) of Miall (1996) and is interpreted as floodplain depos
its formed by settling from sheetfloods. The presence of lamination
(lithofacies Fl) indicates deposition from suspension. However, stratifi
cation is mostly absent (lithofacies Fsm), probably due to the presence
of organisms (bioturbation or pedoturbation). Desiccation cracks indi
cate periodic drying out; long-term drying out led to the formation of
calcic paleosols (pedogenic limestone)
Sedimentary Petrography
From both sections, petrographic thin sections of sandstones were
studied under the microscope (Jemez River 6, Gilman 14). Sandstones of
both sections are very similar in petrographic composition and textural
properties (see Table 2).
Coarse-grained sandstones are moderately to poorly sorted; finer
grained sandstones commonly are moderately to well sorted. Most grains
are angular to subangular, but subrounded to rounded grains also occur. In
particular, sedimentary rock fragments (reworked caliche grains) are
rounded to well rounded. In some sandstones, the original rounding is
diagenetically overprinted as grains are partly (randomly) replaced by
carbonate cement.
At the Jemez River all studied sandstones are calcite cemented
(Fig. 9), whereas at Gilman both carbonate-cemented sandstones and
sandstones in which carbonate cement is rare to absent occur (Fig. 10).
32
FIGURE 7. Trough crossbedded sandstone (lithofacies St-m), the most
common lithofacies type of the sandstone facies within the Abo Formation
(Gilman).
Monocrystalline quartz is a common constituent of all sandstones. Most
monocrystalline quartz grains are slightly undulose, and subordinately
nonundulatory quartz grains are present.
Polycrystalline quartz grains are mostly composed of only a few,
large internal crystals of more equant shape with little or no intercrystalline
suturing (Fig. 9). These grains are nonundulose to slightly undulose.
Stretched metamorphic rock fragments composed of abundant small,
elongate, intensely sutured and strongly undulose crystals are rarely
observed.
Detrital feldspars are abundant (Figs. 9 and 10). Most of the
feldspars are slightly to strongly altered, although fresh grains occur, too.
Potassium feldspars dominate. They are mostly untwinned. Microcline
is present throughout both sections, except in the lower part of the
Gilman section. Perthitic feldspars are common. Plagioclase showing
polysynthetic albite twinning is rare.
The most abundant type of rock fragment consists of quartz and
potassium feldspar and is of magmatic origin (Figs. 9A-C, E, 10A).
Rarely, magmatic rock fragments composed of several feldspar crystals
occur. Fine-grained schistose metamorphic (phyllitic) rock fragments
composed of elongate quartz and mica occur in very small quantities.
Metamorphic rock fragments composed of several mica crystal units are
very rare.
Some sandstone beds contain sedimentary rock fragments of re
crystallized micritic carbonate, commonly with smallangular quartz grains
(Figs. 9A and 10H). A few carbonate grains display fissure cracks filled
with calcite. Some carbonate grains display dark brown rims.
Detrital micas include brown and greenish biotite, muscovite (Fig.
10F) and rare chlorite. They are more common in fine-grained sandstone
and rare to almost absent in coarse-grained sandstone. Common acces
sory minerals (heavy minerals) are zircon, tourmaline and apatite. Gar
net is present in a few samples from the lower part of both sections.
The sandstones contain very small amounts of matrix or are free
of matrix, resulting in high original porosity and permeability. Bending of
detrital micas indicates that original porosity was reduced by mechanical
compaction. Remaining pore space was filled with cement minerals such
as quartz, calcite and dolomite.
Quartz cement occurs as authigenic overgrowths on detrital mono
and polycrystalline quartz grains. At Gilman, some of the sandstones
contain abundant quartz grains with authigenic quartz overgrowths (Figs.
10D, G). These sandstones also contain small amounts of matrix, whereas
carbonate cement is rare, occurring as small patches. Most of the sand
stones of both sections are characterized by high amounts of calcite
FIGURE 8. Sandstone sheets (interpreted as deposits of broad shallow
channels of a braided river system), intercalated in siltstone-mudstone
(floodplain deposits); Abo Formation at Jemez River. Rhizoliths are
developed in the lower siltstone-mudstone.
cement, which is coarse, blocky and commonly poikilotopic (Figs. 9,
10A-B). Calcite cement randomly replaces detrital grains, particularly
feldspars, and (subordinately) also quartz. In the upper part of the Jemez
Riversection, subhedral dolomite rhombs replace calcite, and also replace
feldspar and quartz subordinately.
Authigenic feldspar overgrowths were rarely observed in the Gilman
section. Replacement textures indicate that feldspar and quartz
overgrowths were formed first, followed by calcite cement and finally by
dolomite, which replaced all other cements and detrital grains.
Most of the grains (most mono- and polycrystalline quartz grains,
detritalfeldspars and magmatic rock fragments) are derived from granitic
source rocks. Only a small percentage of the detrital grains (some mono
and polycrystalline quartz grains and metamorphic rock fragments) are
derived from metamorphic source rocks. Individual layers contain high
amounts of sedimentary rock fragments, particularly reworked pedogenic
carbonate grains. The amount of individual grain types varies consider
ably within one section, depending mainly on the grain size of the sedi
ment. The coarser the sediment the higher is the amount of granitic rock
fragments and polycrystalline quartz, and the lower is the amount of
detrital feldspars, monocrystalline quartz and micas.
The average petrographic composition of the Abo Formation at
the Jemez River and Gilman is very similar to that of the Cutler Group at
Cañon del Cobre and Arroyo del Agua (Table 2). Sandstones are domi
nantly arkoses and lithic arenites, subordinately subarkoses. Minor dif
ferences exist in the amount of granitic rock fragments, carbonate cement,
and matrix. Slight differences also exist compared to the Abo type section
in central New Mexico, where the amount of monocrystalline quartz is
higher, the amount of detrital feldspars is slightly higher and the amount
of carbonate cement lower than at the Jemez River and at Gilman. These
differences are mainly caused by grain size and petrography of the source
rocks.
The main source rock type is granite, from which the majority of
mono- and polycrystalline quartz, almost all detritalfeldspars and micas,
and all granitic rock fragments were derived. Metamorphic rocks pro
vided a few mono- and polycrystalline quartz, and metamorphic rock
fragments. Most sedimentary rock fragments were derived from the re
working of caliche; subordinately, reworked mudstone and siltstone grains
are present. At Gilman, coarse sandstone beds contain up to 58% carbon
ate grains (reworked caliche).
DISCUSSION
A comparison of the facies and mineralogical composition demon
strates that no significant differences exist between the red beds of the
 33

FIGURE 9. Thin section photographs (all taken under polarized light) of sandstones from the Abo Formation at Jemez River. Width of each photograph
is 6.3 mm. A, Coarse-grained arkosic sandstone composed of mono- (Qm) and polycrystalline quartz (Qp), detrital feldspars (F) including microcline (M),
sedimentary rock fragments (dark micritic carbonate grains - C) and rare granitic rock fragments (G), cemented by calcite. Sample JE 2. B, C, D, Arkosic
sandstone with mono- (Qm), polycrystalline quartz (Qp), perthitic feldspars (F), microcline (M) and granitic rock fragments (G) cemented by coarse,
poikilotopic calcite. B = sample JE 4, C and D = sample JE 6. E, Sandstone containing mono- (Qm) and polycrystalline quartz (Qp), detrital feldspars (F),
sedimentary rock fragments (dark micritic carbonate grains - C) and rare granitic rock fragments (G), cemented by calcite. Sample JE 2. F, H, Coarse
grained, arkosic sandstone, poorly sorted, angular-subangular, containing mono- (Qm), polycrystalline quartz (Qp) and detrital feldspars (F) cemented by
calcite. F = sample JE 5, H = sample JE 3. G, Arkosic sandstone containing abundant detrital feldspars, mostly perthitic potassium feldspars, microcline,
and detrital quartz grains. Sample JR 7.
34

FIGURE 10. Thin section photographs (all taken under polarized light) of sandstones from the Abo Formation at Gilman. Width of each photograph is
6.3 mm, except for F and G it is 3.2 mm. A, B, Coarse-grained, calcite-cemented arkosic sandstone containing mono- (Qm) and polycrystalline quartz
(Qp), detrital feldspars (F) including microcline (M) and a few granitic rock fragments (G). Feldspars are altered and partly replaced by calcite cement. A
= sample GI 13, B = sample GI 15b. C, Arkosic sandstone containing abundant feldspars (F) including microcline (M), mono- and polycrystalline quartz
and rare granitic rock fragments. Detrital grains are cemented by calcite. Sample GI 15a. D, Arkosic sandstone containing abundant monocrystalline quartz
grains, subordinate polycrystalline quartz grains, detrital feldspars and rare detrital micas (muscovite; yellow, lower right). Many quartz grains display
authigenic overgrowths (arrows). Sample GI 14. E, Arkosic sandstone, densely packed, composed of mono- and polycrystalline quartz, detrital feldspars,
some microcline and rare granitic rock fragments. Sample GI 23. F, Arkosic sandstone containing a large detrital muscovite grain. Sample GI 19. G, Arkosic
sandstone with quartz grains displaying well-developed authigenic quartz overgrowths (arrows). Qm monocrystalline, Qp polycrystalline quartz, F detrital
feldspars, Mi detrital mica. Sample GI 16. H, Coarse-grained sandstone containing abundant well rounded sedimentary rock fragments (micritic carbonate
grains), subordinate granitic rock fragments (G) and detrital feldspars (F). Detrital grains are cemented by calcite. Sample GI 17.

Cutler Group and the Abo Formation in northern New Mexico. Differ
ences exist only in age. Whereas the Cutler Group also includes red beds
of Pennsylvanian age, the Abo Formation is restricted to the Early Per
mian and either overlies Late Pennsylvanian marine strata (e.g. Guadalupe
Box Formation, Oso Ridge Member of Bursum Formation) or directly
overlies Precambrian basement rocks, as in the southern Jemez Moun
tains east of Cuba (e.g., Woodward, 1996; Krainer et al., 2005).
Due to the similar facies and composition it is impossible to draw
a boundary between the Permian part of the Cutler Group and the Abo
Formation in northwestern New Mexico. Indeed, Baars (1983) stated
that the arkosic red beds of the Abo Formation are similar in physical
aspect and age to the lower Cutler beds of the Defiance uplift and Monu
ment Valley region. We therefore include theAbo Formation in the Cutler
Group.
In the Paradox Basin, Campbell (1980) interpreted the red beds of
the Cutler Group as alluvial fan deposits that were shed southward from
the Uncompahgre uplift and southwestward from the San Luis uplift.
Like the Abo Formation in New Mexico, the Cutler Group also displays
a facies change from proximal to distal fluvial facies from the Uncompahgre
uplift towards the south and southwest (see also Eberth and Miall, 1991;
Huffman and Condon, 1993). For both the Cutler Group and Abo For
mation of northern New Mexico, the Uncompahgre uplift composed of
Precambrian rocks was the main source area; both display a very similar
mineralogical composition. Local sources for the Abo Formation were
the Zuni and Peñasco uplifts. Both the Zuni and the Peñasco uplifts
were already buried during mid- to late Wolfcampian time (Woodward,
1987).
In the northern Nacimiento Mountains, Wood and Northrop (1946)
arbitrarily separated the Abo from the Cutler red beds along the border
between Rio Arriba and Sandoval counties. Woodward (1987) called all
of these red beds Abo and reported thicknesses of 76 m at the southern
end of the Nacimiento Mountains to as much as 884 m along the northern
side of San Pedro Mountain.
REFERENCES
Baars, D.L., 1983, The Colorado Plateau. A geologic history: University of
New Mexico Press, Albuquerque, 279 p.
Baars, D.L., 2000, The Colorado Plateau. A geologic history: University of
New Mexico Press, Albuquerque, 254 p. (revised and updated edition)
Baars, D.L., 2003, Geology of Canyonlands National Park, Utah; in Sprinkel,
D.A., Chidsey, T.C. Jr. and Anderson, P.B., eds, Geology of Utah´s Parks
and Monuments (second edition): Utah Geological Association Publica
tion 28, p. 61-83.
Campbell, J.A., 1980, Lower Permian depositional systems and Wolfcampian
paleogeography, Uncompahgre basin, eastern Utah and southwestern
Colorado; in Fouch, T.D. and Magathan, E.R., eds., Paleozoic paleo
geography of the west-central United States: Society of Paleontologists
and Mineralogists Rocky Mountain Section, Rocky Mountain Paleo
geography Symposium 1, p. 327-340.
Eberth, D.A, and Berman, D.S., 1993, Stratigraphy, sedimentology and
vertebrate paleoecology of the Cutler Formation red beds (Pennsylva
nian-Permian) of north-central New Mexico: New Mexico Museum of
Natural History and Science, Bulletin 2, p. 33-47.
Eberth, D.A. and Miall, A.D., 1991, Stratigraphy, sedimentology and evolu
tion of a vertebrate-bearing, braided to anastomosed fluvial system,
Cutler Formation (Permian-Pennsylvanian), north-central New Mexico:
Sedimentary Geology, v. 72, p. 225-252.
Fracasso, M.A., 1987, Fluvial deposition and paleoenvironment of the
Cutler Formation red beds, El Cobre Canyon, New Mexico: New Mexico
Geology, v. 9, p.14-25.
Huffman, A.C. and Condon, S.M., 1993, Stratigraphy, structure, and paleo
geography of Pennsylvanian and Permian rocks, San Juan Basin and
adjacent areas, Utah, Colorado, Arizona, and New Mexico: U.S. Geologi
cal Survey, Bulletin 1808, p. 1-44.
35
According to Kues and Giles (2004), the Wolfcampian of northern
and central New Mexico is represented by nonmarine red beds of the
Abo Formation and equivalent units. Towards the south, in southern
New Mexico (Sierra and Doña Ana counties), the Abo red beds interfinger
with marine carbonate rocks of the Hueco Group. To the west, Abo red
beds grade into Cutler red beds in the subsurface northwest of a line that
runs roughly from Cuba to Gallup.
At many places in central New Mexico, red beds of the Abo
Formation disconformably overlie the Bursum Formation. Locally, as in
the hills east of Socorro, the Abo Formation base is a conglomerate that
erosively overlies the Bursum Formation. Limestone clasts in this con
glomerate indicate that, locally, substantial parts of the Bursum Forma
tion were eroded prior to deposition of the Abo red beds, probably
forming deeply incised valleys (Krainer and Lucas, 2009). This
unconformity is also recognized at other locations and indicates a major
tectonic pulse near the Virgilian/Wolfcampian boundary (Krainer and
Lucas, 2009).
There is much evidence that the unconformity on top of the Bursum
Formation resulted from a major tectonic pulse of the ancestral Rocky
Mountain deformation, causing a significant rejuvenation of basement
uplifts that resulted in increased siliciclastic influx and deposition of
nonmarine red beds of the Abo Formation in central New Mexico. This
unconformity near the Virgilian/Wolfcampian boundary is not developed
within the Cutler Group, in the upper part of the El Cobre Canyon
Formation. This underlines the fact that tectonic activity of the ancestral
Rocky Mountain deformation occurred at different times at different
places during the Late Pennsylvanian-Early Permian.
ACKNOWLEDGMENTS
Jess Hunley, Peter Reser and Justin Spielmann provided assis
tance in the field. The U.S. Forest Service granted access to land. Felix
Heller prepared thin sections. We thank Giuseppe Cassinis (Pavia) and
Jörg W. Schneider (Freiberg) for reviewing the manuscript and helpful
suggestions.
Jicha, H. L. Jr. and Lochman-Balk, C., 1958, Lexicon of New Mexico
geologic names: Precambrian through Paleozoic: New Mexico Bureau of
Mines and Mineral Resources, Bulletin 61, 137 p.
Krainer, K. and Lucas, S.G., 2009, Cyclic sedimentation of the Upper
Carboniferous Bursum Formation, central New Mexico: Tectonics ver
sus glacioeustasy: New Mexico Geological Society, Guidebook 60, p.
167-182.
Krainer, K., Vachard, D. and Lucas, S.G., 2005, Lithostratigraphy and bio
stratigraphy of the Pennsylvanian-Permian transition in the Jemez
Mountains, north-central New Mexico: New Mexico Museum of Natu
ral History and Science, Bulletin 31, p. 74-89.
Kues, B.S. and Giles, K.A., 2004, The late Paleozoic ancestral Rocky Moun
tains system in New Mexico; in Mack, G.H. and Giles, K.A., eds., The
geology of New Mexico, a geologic history: New Mexico Geological
Society, Special Publication 11, p. 95-136.
Lucas, S.G. and Krainer, K., 2005a, Cutler Group (Permo-Carboniferous)
stratigraphy, Chama Basin, New Mexico: New Mexico Museum of Natu
ral History and Science, Bulletin 31, p. 90-100.
Lucas, S.G. and Krainer, K., 2005b, Stratigraphy and correlation of the
Permo-Carboniferous Cutler Group, Chama Basin, New Mexico: New
Mexico Geological Society, Guidebook 56, p. 145-159.
Lucas, S.G., Krainer, K. and Colpitts, R.M., 2005, Abo-Yeso (Lower Per
mian) stratigraphy in central New Mexico: New Mexico Museum of
Natural History and Science, Bulletin 31, p. 101-117.
Miall, A.D., 1978, Lithofacies types and vertical profile models in braided
rivers: A summary; in Miall A.D., ed., Fluvial sedimentology: Canadian
Society of Petroleum Geologists, Memoir 5, p. 597-604.
Miall, A.D., 1981, Analysis of fluvial depositional systems: AAPG Educa
tion Course Note Series 20, 75 p.
36
Miall, A. D., 1985, Architectural-element analysis: A new method of facies
analysis applied to fluvial deposits: Earth Science Reviews, v. 22, p. 261
308.
Miall, A.D., 1992, Alluvial deposits; in Walker, R.G. and James, N.P., eds.,
Facies models. Response to sea level change: Geological Association of
Canada, Special Publication, p. 119-142.
Miall, A.D., 1996, The geology of fluvial deposits. Berlin, Springer, 582 p.
Picard, M. D. and High, L.R., 1973, Sedimentary structures of ephemeral
streams: Developments in Sedimentology, no. 17, 223 p.
Tunbridge, I. P., 1981, Sandy high-energy flood sedimentation – some
criteria for recognition, with an example from the Devonian of SW
England: Sedimentary Geology, v. 28, p. 79-96.
Tunbridge, I. P., 1984, Facies models for a sandy ephemeral stream and clay
playa complex: The Middle Devonian Trentishoe Formation of North
Devon, UK: Sedimentology, v. 31, p. 697-716.
Williams, G. E., 1971, Flood deposits of the sandbed ephemeral streams of
central Australia: Sedimentology, v. 17, p. 1-40.
Wood, G.H. and Northrop, S.A., 1946, Geology of the Nacimiento Moun
tains, San Pedro Mountain, and adjacent plateaus in parts of Sandoval
and Rio Arriba Counties, New Mexico: U.S. Geological Survey, Oil and
Gas Investigations Map OM-57.
Woodward, L., 1987, Geology and mineral resources of Sierra Nacimiento
and vicinity, New Mexico: New Mexico Bureau of Mines and Mineral
Resources, Memoir 42, 84 p.
Woodward, L., 1996, Paleotectonics of the late Paleozoic Peñasco uplift,
Nacimiento region, northern New Mexico: New Mexico Geological So
ciety, Guidebook 47, p. 107-113.

APPENDIX

Location of measured stratigraphic sections

All coordinates reported here are UTM’s in zone 13, datum NAD 27.

Cañon del Cobre: base of section at 378369E, 4016319N; top of section at 376985E, 4016278N.

Cañon del Cobre A: section at 382564E, 4021118N and immediate vicinity.

Gilman: base of segment A at 341134E, 3952490N, top at 341224E, 3552720N; base of segment B at 341320E, 3953055N,
top at 341432E, 3953124N.

Jemez River: base of section at 343754E, 3952554; top at 344598E, 3952583N.

 Lucas et al., eds., 2010, Carb-Permian transition in Cañon del Cobre. New Mexico Museum of Natural History and Science Bulletin 49.

37
PENNSYLVANIANLAKE-MARGINICHNOASSEMBLAGE
FROM CAÑON DELCOBRE, RIO ARRIBA COUNTY, NEWMEXICO

SPENCER G. LUCAS AND ALLAN JLERNER

New Mexico Museum of Natural History and Science, 1801 Mountain Road N. W., Albuquerque, NM 87104

Abstract—In Cañon del Cobre of Rio Arriba County, northern New Mexico, red-bed siliciclastic sediments of the
Late Pennsylvanian-Early Permian Cutler Group were deposited by synorogenic fluvial systems in upland and
inland settings along the flank of the Uncompahgre uplift of the ancestral Rocky Mountains. Rare lacustrine strata
(thinly laminated siltstones) of the El Cobre Canyon Formation of the Cutler Group represent shallow, localized
floodplain lakes of the late Paleozoic Pangean tropics. We document the first ichnofossil assemblage known from
such an intermontane lake deposit, in strata of Late Pennsylvanian age. This ichnofossil assemblage consists of
arthropod trackways (Diplichnites x Diplopodichnus, Diplichnites gouldi, cf. Paleohelcura tridactyla, Protichnites
isp.), invertebrate grazing and feeding traces (Gordia indianensis and Helminthoidichnites tenuis), tetrapod foot
prints (Batrachichnus salamandroides and aff. Amphisauropus isp.) and fish swimming trails (Undichna britannica).
The numerically dominant ichnofossils are Helminthoidichnites and Diplichnites; the other ichnotaxa are repre
sented by few specimens. The Cañon del Cobre ichnofossil assemblage is readily assigned to the Scoyenia
ichnofacies and is composed of members of the Diplichnites and Mermia ichnoguilds, consistent with a shallow
lacustrine setting. The relatively low ichnodiversity of the Cañon del Cobre ichnofossil assemblage indicates a
short-lived water body colonized by a mobile invertebrate epifauna of detritus feeders and predators that provided
a food source for tetrapod predators.

INTRODUCTION
Cañon del Cobre (El Cobre Canyon) is a large box canyon in Rio
Arriba County, northern New Mexico (Fig. 1) that exposes extensive
bedrock badlands, slopes and escarpments developed in an ~400 m thick
section of Pennsylvanian-Permian Cutler Group siliciclastic red beds
(Smith et al., 1961; Eberth and Miall, 1991; Eberth and Berman, 1993;
Lucas and Krainer, 2005a, b). These strata are primarily of fluvial origin
divided into a lower, El Cobre Canyon Formation, and an upper, Arroyo
del Agua Formation. Fossils described from these strata, most of them
from the El Cobre Canyon Formation, include palynomorphs, megafossil
plants and vertebrates. Indeed, Cañon del Cobre is most famous
paleontologically for its vertebrate fossil assemblages that span the Penn
sylvanian-Permian transition (Vaughn, 1963; Fracasso, 1980; Lucas et
al., 2005a; Lucas, 2006).
The fluvial strata in Cañon del Cobre also contain ichnofossils,
mostly a low diversity assemblage (dominated by Palaeophycus) in chan
nel-margin deposits. One of these deposits also includes trackways of
the giant millipede-like Arthropleura, assigned to the ichnogenus
Diplichnites (Lucas et al., 2005a; Schneider et al., this volume). Also,
among the fluvial red beds of the El Cobre Canyon Formation in Cañon
del Cobre are some rare and localized lacustrine deposits that yield all the
fossil leaves known from these strata. One of these lake deposits (Figs.
2-3) also yields the ichnofossil assemblage described here. This assem
FIGURE 1. Location of Cañon del Cobre in northern New Mexico (after
blage is a rare example of a Carboniferous ichnoassemblage of a clastic
Ash, 1974).
lake margin in the ancestral Rocky Mountain basins of western North
America. extraformational clasts of quartzite, granite and gneiss. The El Cobre
Institutional abbreviation: NMMNH = New Mexico Museum Canyon Formation approximately corresponds to megasequence 1 of
of Natural History and Science, Albuquerque.
Eberth and Miall (1991).
GEOLOGY The ichnoassemblage described here is from NMMNH locality
5874 in the floor of Cañon del Cobre. This outcrop encompasses an
In Cañon del Cobre, a total thickness of about 230 m of El Cobre abandoned copper mine developed in a localized deposit that we inter
Canyon Formation is exposed, and it is mostly siltstone and sandstone. pret as representative of a clastic lake (Figs. 2-3). The section begins
Minor rock types are conglomerate/conglomeratic sandstone, shale and with a sandstone bench that is mostly laminated and contains an exten
calcrete. Siltstones of the El Cobre Canyon Formation contain numerous sive paleoflora on at least two bedding planes. Strata above this sand
rhizoliths, and sandstones are typically coarse grained, arkosic, mica stone bench are a 4.7-m-thick unit of thinly laminated purple and green
ceous, trough crossbedded and form multistoried bodies that erode to siltstone/shale that contains some isolated limestone (micrite) concre
thick cliffs and benches. Conglomerates characteristically have tions. A crossbedded sandstone bench caps this succession. The trace
38
FIGURE 2. Photographs of NMMNH locality 5874. A, Overview of the
locality, which is just above a mine adit, and below a prominent, bench
forming crossbedded sandstone (see Fig. 3). The locality is at the base of a
slope formed by thinly laminated siltstone. B, The lowest bed, marked with
5874, is the trace-fossil-bearing bed, immediately below thinly-laminated
siltstone. Scale in cm.
fossils reported here (NMMNH locality 5874) occur in the upper few
mm of the sandstone bench, just below the siltstone/shale (Figs. 2-3).
Deposition of the El Cobre Canyon Formation took place in an
ephemeral braided stream environment (Eberth and Miall, 1991; Krainer
and Lucas, this volume). Localized (less than 100 m on strike) deposits
of thinly laminated siltstone/shale are rare in this depositional system,
and the strata that encompass locality 5874 (Figs. 2-3) are interpreted by
us as the deposits of a clastic lake because: (1) they have a limited extent
and lenticular geometry, pinching out laterally into overbank floodplain
mudrock; (2) the lithofacies of the deposit, especially the thinly lami
nated siltstone/shale interval, are those often associated with clastic lakes
(e.g., Picard and High, 1981); and (3) the stratigraphic succession is from
laminar sandstone to thinly laminated siltstone/shale with micritic lime
stone nodules, readily interpreted as an upward-fining succession from
lake margin sandstone to lake bottom siltstone/shale (e.g., Picard and
High, 1981).
ARTHROPOD TRACKWAYS
The arthropod trackway terminology that we use conforms to
that of Minter et al. (2007a), which was modified from that of Anderson
(1975), Trewin (1994), and Braddy (2001).
FIGURE 3. Measured stratigraphic section at locality 5874.
Ichnogenus Diplichnites Dawson, 1873
Diplichnites gouldi Gevers, in Gevers et al., 1971
Fig. 4B
Referred Specimens: NMMNH P-56250 (Fig. 4B), P-56262,
P-56269,
Characteristics: P-56250 shows a single trackway preserved in
part and counterpart. Symmetrical track rows consist of closely-spaced
tracks that are 1 mm or slightly less in width and rounded to elongate in
shape. Elongate tracks are oblique to the midline. The internal width
between track rows is 3 mm, and the external width is 5 mm. There are no
medial impressions. The preservation is too indistinct to accurately de
termine the number of tracks in a series. Trackway length is 236 mm, and
the track course is mostly straight with slight curves.
Specimen P-56262 is preserved in part and counterpart. Sym
metrical track rows consist of closely-spaced, elongate tracks that are 1
mm wide or slightly less and oblique to the midline. The internal width
between track rows is 4 mm, and the external width is 7 mm. There are no
medial impressions, and the trackway is short and only 12 mm long.
Indistinct undertracks of a similar trackway are also seen on the slab
surface.
Specimen P-56269 shows a single trackway preserved in positive
hyporelief. Symmetrical track rows consist of closely-spaced tracks that
are 1 mm wide or slightly less and rounded to elongate in shape. Elongate
tracks are oblique to the midline, and the number of tracks in a series is
six. The internal width between track rows is 3 mm, and the external
width is 6 mm. The trackway length is 23 mm, and there are no medial
impressions.
Discussion: Specimens are assigned to Diplichnites gouldi based
 39

FIGURE 4. Diplichnites x Diplopodichnus (A, NMMNH P-56252) and Diplichnites gouldi (B, NMMNH P-56250) from locality 5874.

on their close similarity to the emended diagnosis of the ichnospecies by aged individuals of one type of animal.
al.,
and
Buatois
widths (e.g.,
exhibitedetdifferent
2007a).
morphology,
These
al. which
Trewinfew
(1998a). may
Some
Cañon and
Cobre
indicate
morphotypes
and del
McNamara, that of were
Smith
trackways
occurrences
a1995;they
relatively wide al.,
etare range
all
produced
Diplichnites 2003; by
ofMinter
ofgouldi
similar et
size
same-
external
have Diplichnites is generally thought to have been produced by myri
apods or arthropleurids (Weber and Braddy, 2004), although crustaceans
and euthycarcinoids are also possible track makers (Minteret al., 2007a).
D. gouldi ranges from the Ordovician to Permian and is found in a variety
of settings, including transitional terrestrial to subaqueous freshwater
40
environments (Buatois et al., 1998a).
This is the second occurrence of Diplichnites from Cañon del
Cobre. An arthropleurid trackway assigned to D. cuithensis has previ
ously been reported (Hunt et al., 2004; Lucas et al., 2005b). This track
way was the first report of this ichnospecies from the American West
and is one of the largest examples of D. cuithensis known (see Schneider
et al., this volume).
Diplichnites x Diplopodichnus
Fig. 4A
Referred specimen: NMMNH P-56252 (Fig. 4A).
Characteristics: P-56252 is preserved in part and counterpart.
The slab surface shows three similar trackways with external widths
between 5 and 7 mm. Each trackway shows two track rows that consist
of small rounded tracks that, in places, become continuous ridges in
which individual tracks are no longer distinguishable. Two of the trackways
overcross and have mostly straight courses with some gentle curves.
Both of these trackways have similar lengths of between 105 and 115
mm. The third trackway shows a curved course, with one abrupt change
in the opposite direction. It is the longest trackway on the slab, with a
length of 185 mm. The internal margin between inner track rows on all
three trackways displays either ridges or furrows along some of the
course. Small grazing trails (cf. Helminthoidichnites) occur in association
with the trackways, together with impressions of sparse, probable plant
fragments.
Discussion: The trackways on this slab show characteristics that
intergrade between Diplichnites and Diplopodichnus Brady, 1947. Minter
et al. (2007b) discussed the different trackway morphologies that result
from the varying moisture content of “soft” substrates (soup, soft and
firmgrounds), and made specific reference to the classic example of the
intergradation of Diplichnites, Diplopodichnus and Dendroidichnites.
Rather than arguing for synonomy Minter et al. recommended retention
of each ichnogenus when seen in isolation but not for when intergrada
tion occurs with other ichnotaxa. They recommended that these
intergrading examples should be regarded as hybrids. Accordingly, we
note P-56252 as a hybrid form (Diplichnites x Diplopodichnus), which
resulted from the track maker having traversed a moist substrate that
varied in its consistency.
Diplopodichnus is a facies-crossing ichnotaxon that ranges in age
from the Ordovician to Permian (Buatois et al., 1998a). Previous reports
of Diplopodichnus include those from a Carboniferous lacustrine facies
in Argentina (Aceñolaza and Buatois, 1991, 1993). Davis et al. (2007)
have recently reported experimentally producing traces similar to
Diplopodichnus and Dendroidichnites in substrates approximating tran
sitional subaqueous-subaerial conditions using an African giant black
millipede.
Ichnogenus Paleohelcura Gilmore, 1926
cf. Paleohelcura tridactyla Gilmore, 1926
Fig. 5B
Referred Specimens: NMMNH P-56640 (Fig. 5B), P-56641,
P-56642, P-56643, P-56644.
Characteristics: P-56640 shows a partial trackway through which
some of the slab surface has flaked off and left a few remaining series of
tracks contained in two track rows. The best preserved row contains
three complete series consisting of three tracks each. Two of these series
are linear in arrangement, and one is triangular. There is a single track at
one end of the track row, which appears to have been part of an incom
plete series that flaked off. The series widths are 8-9 mm, measured from
inner to outer tracks. Inter-series distances are between 2 and 3 mm.
Tracks are round, except for one in the triangular group, which is bifid.
The tracks are typically 1 mm in diameter. The stride is 9 mm, but pace
cannot be accurately determined due to the incomplete preservation. The
opposite track row shows a single series of tracks that are preserved at
one end. This series shows three round tracks with 1 mm diameters that
are arranged into a linear group. It also has a series width of 9 mm,
measured from inner to outer track. It constitutes a set with the series
that occurs on the end of the opposite track row. The external width
between track rows is 30 mm, as measured by the single set. The internal
width is 15 mm. A medial impression is not seen. The series appear to be
obliquely arranged with staggered symmetry, although both cannot be
confidently determined due to the limits of preservation. Preservation is
in positive hyporelief.
P-56641 consists of a part and counterpart that shows multiple
circular tracks distributed across the surface in a confused (“over-tracked”)
arrangement. Linear groups of three tracks are recognizable within this
overall distribution, but none can be placed confidently into a recogniz
able trackway pattern. Typical width of these groups of three tracks is 9
mm, which is the same as that seen for the series in P-56640. The
remaining three specimens (P-56642, P-56643, P-56644) show surfaces
with a few circular tracks that are sometimes arranged in groups of three.
Single tracks in isolation and groups of two also occur. Medial impres-
sions are absent, and preservation is in positive hyporelief.
Discussion: These specimens are poorly preserved and are most
likely undertracks. As best seen on P-56640, the appearance of these
mostly circular tracks and their oblique arrangement into linear groups of
three is most typical of Paleohelcura tridactyla, to which these speci
mens are provisionally referred. There is one group of tracks on P-56640
that resembles Octopodichnus by showing a faintly impressed fourth
imprint.
These tracks show some superficial resemblance to the circular
surface expression of vertical burrows such as Treptichnus or Arenicolites.
However, a track preserved on P-56641 occurs on the very edge of part
and counterpart slabs, and examination of both slabs on-edge revealed no
indication of vertical burrowing beneath the surface imprints. Indeed,
vertical burrows of any sort are absent within the Cañon del Cobre
ichnoassemblage.
Another indication that these specimens are locomotion traces
rather than burrows is the presence of a bifid track on P-56640.
Neoichnological experiments (Brady, 1947; Sadler, 1993) have shown
that scorpions, which possess bifid extremities, are capable of producing
Paleohelcura trackways. While it is likely that the Cañon del Cobre
specimens were produced by scorpionids or arachnids (Davis et al.,
2007), it is possible, too, that they were made by eurypterids.
Paleohelcura is similar to Palmichnium, which is generally attributed to
eurypterid producers. The main feature distinguishing these two
ichnogenera is different gait patterns (Trewin and McNamara, 1995).
Unfortunately, the gait pattern of the referred Cañon del Cobre material
cannot be determined due to poor preservation.
Eurypterids and scorpions are known to have co-existed during
the middle to late Paleozoic (Selden, 1985). Late Carboniferous euryp
terids were mostly amphibious and largely represented by the genus
Adelophthalmus (Selden and Nudds, 2004), a form typically found with
freshwater and terrestrial animals and plants (Plotnick, 1997). Terrestrial
excursions by eurypterids were unlikely to be the result of food scaveng
ing, as they lacked suitable adaptations for terrestrial feeding (Braddy,
2004). A more likely, albeit speculative, explanation is that these excur
sions were driven by environmental factors, such as the drying up of a
formerly inhabited body of water (Donovan, 2001) or to mate (Braddy,
1999). A eurypterid prosoma assigned to Adelophthalmus sp. was dis
covered at Cañon del Cobre from an areally restricted lacustrine deposit
(Hannibal et al., 2005). It comes from a succession that is stratigraphically
above locality 5874, establishing the presence of these animals in the
Cañon del Cobre biota and indicating their potential as track makers.
The Late Carboniferous to Early Permian Carrizo Arroyo
Lagerstätte in central New Mexico also yields a significant Adelophthalmus
record, due in part to the unusually large number of specimens (Kues and
Kietzke, 1981). These eurypterids inhabited shallow ponds or enclosed
lagoons on the deltaic plain adjacent to the ancestral Cañon del Cobre

FIGURE 5. Protichnites ichnosp. (A, NMMNH P-56248) and cf. Paleohelcura tridactyla (B, NMMNH P-56640) from locality 5874.
area (Kues and Kietzke, 1981). The presumed leg span of the Carrizo
Arroyo eurypterids during locomotion was about 100 mm (Briggs and
Rolfe, 1983). In comparison, the partial trackway seen on specimen P
56640 indicates a leg span of about 30 mm, which would indicate a
juvenile, if the Cañon del Cobre track producer were a eurypterid.
Paleohelcura is primarily known from Permian eolian deposits in
the American Southwest, including the Coconino Sandstone of Arizona
(Gilmore, 1926, 1927; Brady, 1947; Brand and Kramer, 1996), the De
Chelly Sandstone of Arizona (Sadler, 1993), and the Lyons Sandstone of
Colorado (Toepelman and Rodeck, 1936). There is also a record from the
eolian Jurassic Navajo Sandstone of Colorado (Faul and Roberts, 1951).
There is a single prior Pennsylvanian record, which comes from a tidal
flat setting in the McAlester Formation of Oklahoma (Lucas et al., 2004).
There are no previous North American records of Paleohelcura from
lacustrine settings.
41
Ichnogenus Protichnites Owen, 1852
Protichnites isp.
Fig. 5A
Referred Specimen: NMMNH P-56248 (Fig. 5A).
Characteristics: The specimen is preserved in part and counter
part. The slab surfaces show a single trackway consisting of two rows of
variably-shaped tracks that are separated by a medial impression. One
track row on the part is missing approximately half of its length from the
slab surface. The counterpart track row that corresponds to the missing
section is obscured by a thin covering of matrix. It probably appears
there as a result of differential splitting that occurred along a plane of
weakness created by a medial impression between the track rows. The
more complete, and therefore longer, track row shows one end that is
narrow, with a width that varies between 1 and 2 mm. The tracks vary in
42
size and are irregularly rounded to ellipsoidal in shape. The larger tracks
measure 2 mm at their longest axis, and the smaller tracks are 1 mm or
less. The tracks seen along this section of the row are generally disorga
nized but sometimes form groups of two or three that are oblique to the
midline. Individual series cannot be determined. The track row gently
curves midway through the course and widens to 3 mm. The tracks along
this section are more regularly rounded and crowded into irregular group
ings of 2-3 that are oblique to the midline.
The shorter, incomplete track row consists of tracks that are
similar to those seen in the same section of the opposite row. They are
irregularly arranged, although some occur in groups of two and three that
vary in their orientation to the midline. Individual series cannot be deter
mined. The track row varies in width between 1and 2 mm, and it termi
nates where the course is seen to curve. The external width between the
two track rows averages 7 mm. The internal width averages 3 mm. The
length of the longer track row is 45 mm, and that of the shorter track row
is 22 mm. A centrally-placed medial impression is present in the portion
of the trackway where both track rows are seen. It is less than 1 mm
wide, and a small amount of splay appears along its course. A secondary
impression branches from the centrally placed one at the point where the
trackway course turns and the shorter track row terminates. This some
what thinner, secondary impression is faintly impressed onto the cen
tralized one where the branching occurs. It is laterally offset to a high
degree and makes contact with some of the inner tracks in the longer track
row. It then becomes discontinuous and curves away from the track row
and back towards the midline.
Discussion: This short trackway is referred to Protichnites based
on its simple rounded tracks shown in similar series on either side of a
medial impression (after Keighley and Pickerill, 1998). The Cañon del
Cobre specimen most resembles P. multinotatus Owen in its track ar
rangement, but the material is not well enough preserved to make an
ichnospecific assignment, so it is left in open nomenclature. The trace
producer’s terminal appendage may have had to have been lifted when it
was turned, as indicated by the secondary medial impression. Walker
(1985) referred to this as “offsetting,” which reflects a degree of lateral
inflexibility in the tail. Owen (1852) similarly noted in his original de
scription of P. multinotatus that the medial impression was formed by an
appendage that continued to incline to one side after the body had begun
to turn in another direction.
Possible producers of Protichnites in terrestrial freshwater envi
ronments include small crustaceans or euthycarcinoids (Savage, 1971;
Smith et al., 2003). North American Carboniferous records of Protichnites
are primarily from the Mabou and Cumberland groups of eastern Canada,
where it occurs in various floodplain-sheetflood depositional settings
including ephemeral ponds (Keighley and Pickerill, 1998, 2003).
Protichnites has the distinction of having been the first fossilized arthro
pod trackway that was given a Linnaean name and formal description
(MacNaughton and Hagadorn, 2006). This is the first record of
Protichnites from New Mexico.
GRAZING AND FEEDING TRACES
Ichnogenus Gordia Emmons, 1844
Gordia indianaensis Miller, 1889
Fig. 6B
Referred Specimen: NMMNH P-56268 (Fig. 6B).
Characteristics: This single specimen consists of a horizontal,
unbranched, burrow with a diameter of 1 mm. The burrow fill is struc
tureless and identical to the surrounding slab matrix. The burrow course
is mostly straight but contains a tight turn that over-crosses the straight
segment. Total burrow length is 80 mm, and preservation is in part and
counterpart.
Discussion: This specimen displays self-overcrossing, so we sepa
rate it from the similar burrow Helminthoidichnites, which also appears
in the Cañon del Cobre ichnofauna. This specimen most closely re
sembles Gordia indianaensis, which is characterized by having sharply
angled turns (Buatois et al., 1998b). Records of Gordia indianaensis
have primarily come from Carboniferous lacustrine to marginal-marine
settings (Buatois et al., 1998b). There is a single Early Permian report
from a playa lake ichnoassemblage at Castle Peak, Texas (Minter et al.,
2007a).
Ichnogenus Helminthoidichnites Fitch, 1850
Helminthoidichnites tenuis Fitch, 1850
Fig. 6A, C
Referred specimens: NMMNH P-56251, P-56265, P-56271,
P-56273 (Fig. 6C), P-56274, P-56283 (Fig. 6A).
Characteristics: Slab surfaces show multiple, unbranched, hori
zontal burrows preserved in positive hyporelief. These simple burrows
are thin and have diameters of 1 mm or less. Burrow courses are varied
from straight to curving, although circular courses are occasionally seen.
Overcrossing among different specimens commonly occurs, although
self-overcrossing does not. Burrow fill is structureless and identical to
the surrounding slab matrix. Burrow lengths are relatively short, although
rare examples extend to 180 mm.
Discussion: These simple grazing traces are readily assigned to
Helminthoidichnites tenuis based on their simple morphology, generally
straight to curving courses, and lack of self-overcrossing. Since the early
1990’s there has been an increase in reports of H. tenuis from lacustrine
settings (Metz, 1992, 1996; Pickerill, 1992; Buatois and Mángano, 1993;
Aceñolaza and Buatois, 1993; Buatois et al., 1996, 1997; Keighley and
Pickerill, 1997, Minter et al., 2007a), although some of these are now
regarded as Helminthopsis following a review of that ichnogenus by Han
and Pickerill (1995).
TETRAPOD TRACES
Ichnogenus Batrachichnus Woodworth, 1900
Batrachichnus salamandroides Geinitz, 1861
Fig. 7A
Referred Specimens: NMMNH P-56261, P-56284, P-56635,
P-56636 (Fig. 7A).
Characteristics: Slab surfaces show similar, small, tracks of a
quadruped that are preserved in positive hyporelief. P-56636 is the best
preserved specimen and has two track sets, each containing a manus and
pes. One set has a semi-plantigrade manus with four thin, curved digits
that increase in length from digit I to III, with digit IV being the longest.
The length of this manus is 10 mm, and the width is 8 mm. The accom
panying pes slightly oversteps the manus’ posterior and is slightly far
ther from the midline. This pes is plantigrade and somewhat deeply
impressed, which obscures other details. The second track set show a
manus that is semi-plantigrade and contains four, thin, straight digits that
increase in length from digits I to IV, with digit IV being the longest. The
length of this manus is 10 mm, and the width is 7 mm. The accompanying
pes is set closely posterior to the manus. It is semi-plantigrade, and its
details are somewhat obscured. It is slightly out of line with the accom
panying manus, being closer to the midline. The outer width between
opposite, corresponding tracks is 27 mm. There is no medial impression
between the track sets.
Discussion: Batrachichnus is a commonly occurring Carbonifer
ous-Early Permian track that is attributed to small temnospondyl am
phibian producers. The wide range of extramorphological variation that
occurs in the preservation of Batrachichnus tracks has been well docu
mented (e.g., Haubold, 1996, 2000; Melchor and Sarjeant, 2004; Lucas,
2005). The elongate digit impressions seen in the Cañon del Cobre speci
mens are characteristic of a type of extramorphologic variation referred
to as “Gracilichnium” preservation. This morphotype is typically asso
ciated with depositional environments that had substrates with high
moisture content (e.g., tidal flats, fluvial and lacustrine settings).

FIGURE 6. Helminthoidichnites tenuis (A, NMMNH P-56283, C, NMMNH P-56273) and Gordia indianensis (B, NMMNH P-56268) from locality
5874.
Ichnogenus Amphisauropus Haubold, 1970
aff. Amphisauropus isp.
Fig. 7B
Referred Specimens: NMMNH P-56637 (Fig. 7B).
Characteristics: The slab shows two manus tracks, preserved in
negative epirelief, and a partial pes track, preserved in positive epirelief.
One manus is plantigrade, with a broad, rounded posterior that is 30 mm
wide. This track shows three digits that increase in length from digits I to
III; with digit I being 13 mm, digit II being 18 mm, and digit III being 30
mm long. These digits are tapered and directed inward, and may partially
include drag marks in their length. The overall length of this track is 70
mm. A single, partially preserved pes with a rounded heel is seen directly
43
posterior to the manus. It is 30 mm wide and 20 mm long. The second
manus also shows a rounded posterior that is 30 mm wide. The digits are
not distinguishable; although three anterior drag marks are seen. The
longest one is 40 mm long and is on line with the outside margin of the
track. Directly posterior to this manus is the faint outline of an undertrack
that is probably from the accompanying pes. The external width be
tween both pes tracks is 150 mm.
Discussion: The best preserved tracks on this slab show a broad,
inwardly directed manus and an apparently forward directed pes, which
are characteristic of Amphisauropus (Haubold, 1970, 1971). We cau
tiously leave this single specimen in open nomenclature due to its rela
tively poor preservation.
Amphisauropus is a commonly occurring and widely distributed
44
FIGURE 7. Batrachichnus salamandroides (A, NMMNH P-56636), aff. Amphisauropus ichnosp. (B, NMMNH P-56637) and Undichna britannica (C,
NMMNH P-56263) from locality 5874.
track in the Early Permian interval of the European Rotliegend (Lucas et
al., 2001; Voigt, 2005). The first United States record of Amphisauropus
came from sheetflood deposits in the Early Permian Abo Formation of
central New Mexico (Lerner et al., 2001; Lucas et al., 2001, 2009). There
have since followed other U.S. reports, all of which are from the Early
Permian (Wolfcampian-Leonardian). These reports are from the
Hennessey Formation of central Oklahoma (Lucas and Suneson, 2002),
the Clear Fork Formation of north-central Texas (Lucas et al., 2003), the
Yeso Group of central New Mexico (Smith et al. 2005; Lucas et al.,
2005c) and the Choza Formation of north-central Texas (Minter et al.,
2007a). Amphisauropus has also been reported from Early Permian red
bed sequences in eastern Canada (Mossman and Place, 1989; Van Allen
and Calder, 2003; Van Allen et al., 2005). The presumed producers of
Amphisauropus were seymouriamorph amphibians.
Ichnogenus Undichna Anderson, 1976
Undichna britannica Higgs, 1988
Fig. 7C
Referred Specimens: NMMNH P-56263 (Fig. 7C), P-56275.
Characteristics: P-56254 is a slab that shows several sets of
thin, paired trails on its surface. Preservation is in positive hyporelief.
Two sets of trails are relatively well preserved compared to the others.
The first set shows a pair of thin trails with similar courses, each of
which consists of a single wave-like curve. The longer trail crosses over
the shorter one once. The trails are of different lengths, with the longer
being 70 mm and the shorter being 50 mm. The trails are separated by a
distance of 6 mm at their farthest extent. The second set shows a pair of
trails with similar courses, each of which consists of a single wave-like
curve. They overcross once midway through their course. The trails are
of equal length, which is 65 mm. The trails are separated by a distance of
16 mm at their farthest extent. The wavelengths and amplitudes of both
sets of trails cannot accurately be determined because each set abruptly
ends at the edge of the slab, leaving these traces partially preserved. The
slab also contains two similar sets of thin, paired trails that are of rela
tively short length, which is about 20 mm. The courses of these trails are
gently curved and roughly parallel, and they are separated by a distance
of 6 mm at their greatest extant.
Discussion: These simple, intertwined sinusoidal trails are as
signed to Undichna britannica. They are somewhat similar in appear

ance to some of the invertebrate feeding burrows commonly seen at this
locality. The Undichna traces, however, appear more sharply incised
and fluidly produced than the invertebrate traces, which lends support
to their interpretation as locomotion trails produced by fishes. The sev
eral pairs of trails seen on P-56255 are not aligned to any one direction
along the slab surface, which might otherwise suggest that they were tool
marks produced by currents.
A review of Undichna, with examples from the Early Permian of
New Mexico, was recently provided by Minter and Braddy (2006b).
The majority of Undichna occurrences are from the Carboniferous, in
cluding several from various lacustrine depositional settings (Walter,
1982; Turek, 1989, 1996; Melchor et al., 1993; Buatois and Mángano,
1993, 1994; Keighley and Pickerill, 1997; Higgs, 1998; Cameron et al.,
1998; Pazos, 2002; Netto et al., 2003).
DISCUSSION
The ichnofossil assemblage described here consists of arthropod
trackways (Diplichnites, Diplopodichnus, cf. Paleohelcura, Protichnites),
grazing and feeding traces (Gordia, Helminthoidichnites), tetrapod tracks
(Batrachichnus, aff. Amphisauropus) and fish-swimming trails
(Undichna). Arthropod walking traces and grazing/feeding traces domi
nate the assemblage; all other kinds of traces are rare, being represented
by one or a few specimens. This assemblage is readily assigned to the
Scoyenia ichnofacies (cf. Buatois and Mángano, 2007), though note that
it lacks horizontal, meniscate back-filled traces.
The Cañon del Cobre ichnofossil assemblage consists of ichnotaxa
characteristic of both the Diplichnites and Mermia ichnoguilds of Buatois
REFERENCES
Aceñolaza, F. G. and Buatois, L. A., 1991, Tazas fósiles de la Formación
Patquía en el Bordo Atravesado, Sierra de Famatina, La Rioja: Acta
Geológica Lilloana, v. 15, p. 19-29.
Aceñolaza, F. G. and Buatois, L. A. 1993, Nonmarine perigondwanic trace
fossils from the late Paleozoic of Argentina: Ichnos, v. 2, p. 183-201.
Anderson, A. M., 1975, The ‘trilobite’ trackways in the Table Mountain
Group (Ordovician) of South Africa: Palaeontologia Africana, v. 18, p.
35-45.
Anderson, A. M., 1976, Fish trails from the Early Permian of South Africa:
Palaeontology, v. 19, p. 397-409.
Ash, S.R., 1974, Upper Triassic plants of Cañon del Cobre, New Mexico:
New Mexico Geological Society, Guidebook 25, p. 179-184.
Braddy, S. J., 1995, A new arthropod trackway and associated invertebrate
ichnofauna from the Lower Permian Hueco Formation of the Robledo
Mountains, southern New Mexico: New Mexico Museum of Natural
History and Science, Bulletin 6, p. 101-105.
Braddy, S. J., 1999. Eurypterid palaeoecology: Palaeobiological, ichnological
and comparative evidence for a ‘mass-moult-mate’ hypothesis:
Palaeogeography, Palaeoclimatology, Palaeoecology, v. 172, p. 115
132.
Braddy, S. J., 2001, Trackways-arthropod locomotion; in Briggs, D.E.G.
and Crowther, P. R., eds., Palaeobiology II, p. 389-393.
Braddy, S. J., 2004, Ichnological evidence for the arthropod invasion of
land: Fossils and Strata, v. 51, p. 136-140.
Brady, L. F., 1947, Invertebrate tracks from the Coconino Sandstone of
northern Arizona: Journal of Paleontology, v. 21, p. 466-472.
Brand, L. R. and Kramer, J., 1996, Underprints of vertebrate trackways in
the Permian Coconino Sandstone in Arizona: Ichnos, v. 4, p. 225-230.
Briggs, D. E. G., and Rolfe, I., 1983, A giant arthropod trackway from the
Lower Mississippian of Pennsylvania: Journal of Paleontology, v. 57, p.
377-390.
Buatois, L. A. and Mángano, M. G., 1993, Trace fossils from a Carbonifer
ous turbidtic lake: Implications for the recognition of additional nonma
rine ichnofacies: Ichnos, v. 2, p. 237-258.
45
et al. (1998c). Thus, the Diplichnites ichnoguild is represented by surface
to shallow subsurface walking traces of arthropod predators and detritus
feeders (Diplichnites, Diplopodichnus, Paleohelcura, Protichnites). The
Mermia ichnoguild comprises surface to shallow subsurface traces of
detritus feeders (Helminthoidichnites, Gordia). One (Diplichnites
ichnoguild) typically represents lake margins, whereas the other (Mermia
ichnoguild) represents lake bottoms, and their co-occurrence in a clastic
lake margin is not surprising. Both ichnoguilds are represented by transi
tory structures made by mobile predators and detritus feeders.
Ichnofossil assemblages of clastic freshwater lakes of Carbonifer
ous age are not common (Parker and Gierlowski-Kordesch, 2007), and
we know of no such record from the ancestral Rocky Mountains of the
western United States. The Cañon del Cobre ichnofossil assemblage thus
adds to a limited fossil record, and shows successful colonization of a
clastic lake margin in a relatively inland and upland setting of the ances
tral Rocky Mountain foreland. The presence of mobile detritus feeders
and arthropod predators confirms a pattern already well documented in
other Carboniferous terrestrial ichnofaunas (Buatois et al., 1998c; Parker
and Gierlowski-Kordesch, 2007; Buatois and Mángano, 2007). The Cañon
del Cobre ichnofossil assemblage thus provides evidence of a community
of detritus feeding and predatory arthropods, likely preyed upon by
tetrapod predators, on the Late Pennsylvanian braided river plains of the
ancestral Rocky Mountain foreland.
ACKNOWLEDGMENTS
Dan Chaney and Josh Smith provided valuable field assistance
collecting NMMNH locality 5874. Elizabeth Gierlowski-Kordesch, Joerg
Schneider and Nicholas J. Minter reviewed the manuscript.
Buatois, L. A. and Mángano, M. G., 1994, Pistas de peces en el Carbonifero
de la cuenca Paganzo (Argentina): Su significado estragráfico y
paleoambiental: Ameghiniana, v. 31, p. 33-40.
Buatois, L. A. and Mángano, M. G., 2007, Invertebrate ichnology of conti
nental freshwater environments; in Miller, W., III, ed., Trace fossils:
Concepts, problems, prospects: Amsterdam, Elsevier, p. 285-323.
Buatois, L. A., Jalfin, G. and Aceñolaza, F. G., 1997, Permian nonmarine
invertebrate trace fossils from the southern Patagonia, Argentina:
Ichnologic signatures of substrate consolidation and colonization se
quences: Journal of Paleontology, v. 71, p. 324-336.
Buatois, L. A., Mángano, M. G., Maples, C. G. and Lanier, W. P., 1998a,
Taxonomic reassessment of the ichnogenus Beaconichnus and addi
tional examples from the Carboniferous of Kansas, U.S.A.: Ichnos, v. 5,
p. 287-302.
Buatois, L. A., Mángano, M. G., Maples, C. G. and Lanier, W. P., 1998b.
Ichnology of an Upper Carboniferous fluvio-estuarine paleovalley: The
Tonganoxie Sandstone, Buildex Quarry, Eastern Kansas, USA: Journal
of Paleontology, v. 72, p. 152-180.
Buatois, L. A., Mángano, M. G., Wu, X. and Zhang, G., 1996, Trace fossils
from Jurassic lacustrine turbites of the Anyao Formation (central China)
and their environmental and evolutionary significance: Ichnos, v.4, p.
287-303.
Buatois, LA., Mángano, M. G., Genise, J. F. and Taylor, T. N., 1998c, The
ichnologic record of the invertebrate invasion of nonmarine ecosys
tems: Evolutionary trends in ecospace utilization, environmental ex
pansion, and behavioral complexity: Palaios, v. 13, p. 217-240.
Cameron, B., Wood, D. and van Dommelen, R., 1998, Vertebrate and inver
tebrate trace fossils from the Horton Bluff Formation (Lower Carbonif
erous) of Nova Scotia: Mining Matters for Nova Scotia, v. 98, p. 6.
Davis, R. B., Minter, N. J. and Braddy, S. J., 2007. The neoichnology of
terrestrial arthropods: Palaeogeography, Palaeoclimatology, Palaeo
ecology, v. 255, p. 284-307.
Dawson, J. W., 1873, Impressions and footprints of aquatic animals and
imitative markings on Carboniferous rocks: American Journal of Sci
46
ence and Arts, v. 105, p. 16-24.
Donovan, S. K., 2001, Fossils explained 37, eurypterids: Geology Today, v.
17, p. 195-199.
Eberth, D. S. and Berman D. S, 1993, Stratigraphy, sedimentology and
vertebrate paleontology of the Cutler Formation redbeds (Pennsylva
nian-Permian) of north-central New Mexico: New Mexico Museum of
Natural History and Science, Bulletin 2, p. 33-48.
Eberth, D. A., and Miall, A. D., 1991, Stratigraphy, sedimentology and
vertebrate paleoecology of the Cutler Formation redbeds (Pennsylva
nian-Permian) of north-central New Mexico: Sedimentary Geology, v.
72, p. 225-252.
Emmons, E., 1844, The Taconic System: Based on observations in New
York, Massachusetts, Maine, Vermont and Rhode Island. Caroll and
Cook, Albany, 68 p.
Faul, H., and Roberts, W. A., 1951, New fossil footprints from the Na
vajo(?) Sandstone of Colorado: Journal of Paleontology, v. 25, p. 266
274.
Fitch, A., 1850, A historical, topographical and agricultural survey of the
County of Washington. Part 2-5: Transactions of the New York Agri
cultural Society, v. 9, p. 753-944.
Eberth, D. S. and Miall, A. D., 1991, Stratigraphy, sedimentology and evo
lution of a vertebrate-bearing, braided to anastomosed fluvial system,
Cutler Formation (Permian-Pennsylvanian), north-central New Mexico:
Sedimentary Geology, v. 72, p. 225-252.
Fracasso, M. A., 1980, Age of the Permo-Carboniferous Cutler Formation
vertebrate fauna from El Cobre Canyon, New Mexico: Journal of Pale
ontology, v. 54, p. 1237-1244.
Geinitz, H. B., 1861. Dyas, I, 130 p.
Gevers, T. W., Frakes, L. A., Edwards, N. and Marzolf, J. E., 1971, Trace
fossils in the Lower Beacon sediments (Devonian), Darwin Mountains,
southern Victoria Land, Antarctica: Journal of Paleontology, v. 45, p.
81-94.
Gilmore, C. W., 1926, Fossil footprints from the Grand Canyon: Smithsonian
Miscellaneous Collections, v. 77, p. 1-41.
Gilmore, C. W., 1927, Fossil footprints from the Grand Canyon: Smithsonian
Miscellaneous Collections, v. 80, p. 1-78.
Han, Y. and Pickerill, R. K., 1995, Taxonomic review of the ichnogenus
Helminthopsis Heer 1877 with a statistical analysis of selected
ichnospecies: Ichnos, v. 4, p. 83-118.
Hannibal, J. T., Lucas, S. G., Lerner, A. J and Chaney, D. S., 2005, An
eurypterid (Adelophthalmus sp.) from a plant-rich lacustrine facies of
Upper Pennsylvanian strata in El Cobre Canyon, New Mexico: New
Mexico Museum of Natural History and Science, Bulletin 31, p. 34-38.
Haubold, H., 1970, Vesuch einer Revision der Amphibien-Fährten de Karbon
und Perm: Freiberger Forschungs Hefte, C 260, p. 83-117.
Haubold, H., 1971, Ichnia amphibiorum et reptiliorum fossilium: Encyclo
pedia of Paleoherpetology, v. 18, 124 p.
Haubold, H., 1996, Ichnotaxonomie und Klassifikation von
Tetrapodenfährten aus dem Perm: Hallesches Jahrbuch
Geowissenschaften, v. B18 p. 23-88.
Haubold, H., 2000, Tetrapodenfährten aus dem Perm-Kenntnisstand und
Progress 2000: Hallesches Jahrbuch Geowissenschaften, v. B22 p. 1-16.
Higgs, R., 1988, Fish trails in the Upper Carboniferous of south-west En
gland: Palaeontology, v. 31, 255-272.
Hunt, A. P., Lucas, S. G. and Huber, P., 1990, Early Permian footprint fauna
from the Sangre de Cristo Formation of northeastern New Mexico: New
Mexico Geological Society, Guidebook 41, p. 291-303.
Hunt, A. P., Lucas, S. G., Lerner, A. J and Hannibal, J. T., 2004, The giant
Arthropleura trackway Diplichnites cuithensis from the Cutler Group
(Upper Pennsylvanian) of New Mexico: Geological Society of America,
Abstracts with Programs, v. 36, no. 5, p. 66.
Keighley, D. G. and Pickerill, R. K., 1997, Systematic ichnology of the
Mabou and Cumberland groups (Carboniferous) of western Cape Breton
Island, eastern Canada, 1: Burrows, pits, trails, and coprolites: Atlantic
Geology, v. 33, p. 181-215.
Keighley, D. G. and Pickerill, R. K., 1998, Systematic ichnology of the
Mabou and Cumberland groups (Carboniferous) of western Cape Breton
Island, eastern Canada, 2: Surface markings: Atlantic Geology, v. 34, p.
83-112.
Keighley, D. G. and Pickerill, R. K., 2003, Ichnocoenoses from the Carbon
iferous of eastern Canada and their implications for the recognition of
ichnofacies in nonmarine strata: Atlantic Geology, v. 39, p. 1-22.
Kues, B. S. and Kietzke, K., 1981, A large assemblage of a new eurypterid
from the Red Tanks Member, Madera Formation (Late Pennsylvanian
Early Permian) of New Mexico: Journal of Paleontology, v. 55, p. 709
729.
Lerner, A. J, Lucas, S. G. and Haubold, H., 2001, The footprint ichnogenera
Amphisauropus and Varanopus in the Lower Permian of New Mexico:
Journal of Vertebrate Paleontology, v. 21, no. 3, p. 72A.
Lucas, S. G., 2005, Tetrapod ichnofacies and ichnotaxonomy: Quo vadis?
Ichnos, v. 12, p. 1-6
Lucas, S. G., 2006, Global Permian tetrapod biostratigraphy and
biochronology; in Lucas, S.G., Cassinis, G. and Schneider, J.W., eds.,
Non-marine Permian biostratigraphy and biochronology. Geological
Society, London, Special Publication 265, p. 65-93.
Lucas, S. G. and Krainer, K., 2005a, Stratigraphy and correlation of the
Permo-Carboniferous Cutler Group, Chama Basin, New Mexico: New
Mexico Geological Society, Guidebook 56, p. 145-159.
Lucas, S. G. and Krainer, K., 2005b, Cutler Group (Permo-Carboniferous)
stratigraphy, Chama basin, New Mexico: New Mexico Museum of Natu
ral History and Science, Bulletin 31, p. 90-100.
Lucas, S. G. and Suneson, N. H., 2002, Amphibian and reptile tracks from
the Hennessey Formation (Leonardian, Permian), Oklahoma County,
Oklahoma: Oklahoma Geology Notes, v. 62, no. 2., p. 56-62.
Lucas, S. G., Lerner, A. J, Brunner, M. and Shipman, P., 2004, Middle
Pennsylvanian ichnofauna from eastern Oklahoma, USA: Ichnos, v. 11,
p. 45-55.
Lucas, S. G., Lerner, A. J, Hannibal, J. T., Hunt, A. P. and Schneider, J. W.,
2005a, Trackway of a giant Arthropleura from the Upper Pennsylva
nian of El Cobre Canyon, New Mexico: New Mexico Geological Society,
Guidebook 56, p. 279-282.
Lucas, S. G., Minter, N. J., Spielmann, J. A., Hunt, A. P. and Braddy, S.J.,
2005b, Early Permian ichnofossil assemblage from the Fra Cristobal
Mountains, southern New Mexico: New Mexico Museum of Natural
History and Science, Bulletin 31, p. 140-150.
Lucas, S. G., Lerner, A. J, and Haubold, H., 2001, First record of
Amphisauropus and Varanopus in the Lower Permian Abo Formation,
central New Mexico: Hallesches Jarhbuch für Geowissenschaften, v. B23,
p. 69-78.
Lucas, S. G., Spielmann, J. A. and Lerner, A. J, 2009, The Abo Pass tracksite:
A Lower Permian tetrapod footprint assemblage from central New
Mexico: New Mexico Geological Society, Guidebook 60, p. 285-290.
Lucas, S. G., Lerner, A. J, Nelson, Hunt, A. P., and DiMichele, W. A., 2003.
Early Permian (Leonardian) tetrapod footprints from Lake Kemp, north
central Texas: Geological Society of America, Abstracts with Programs,
v. 35, no. 6, p. 499.
Lucas, S. G., Smith, J. A., and Hunt, A. P., 2005c. Tetrapod tracks from the
Lower Permian Yeso Group: New Mexico Museum of Natural History
and Science, Bulletin 31, p. 119-124.
MacNaughton, R. B., and Hagadorn, J. W., 2006. Report on plaster casts of
arthropod-produced trace fossils (Protichnites) figured in W.E. Logan’s
Geology of Canada (1863), and recently copied from material in the
Amherst College Museum of Natural History, Amherst, MA: Report
001-RBM-2006/1, Geological Survey of Canada (Calgary).
Melchor, R. N. and Sarjeant, W. A. S., 2004, Small amphibian and reptile
footprints from the Permian Carapacha basin, Argentina: Ichnos, v. 11,
p. 57-78.
Melchor, R. N., Bazán, J., and Fernández, M. A., 1993. Asociación de trazas
fósiles de la facies peliticas de la Formación Agua Escondida (Carbonifero
Superior?), sureste de Mendoza, Argentina: Primera Reunión Argentina
de Icnologia, Resúmenes y Conferencias Invitadas, p. 15.
Metz, R., 1992. Trace fossils from the Lower Jurassic nonmarine Towaco
Formation, New Jersey: Northeastern Geology, v. 14, p. 29-34.
Metz. R, 1996. Newark Basin ichnology: the Late Triassic Perkasie Mem
ber of the Passaic Formation, Sanatoga, Pennsylvania: Northeastern
Geology and Environmental Sciences, v. 18, p. 118-129.

Miller, S.A., 1889. North American geology and palaeontology for the use
of amateurs, students and scientists. Western Methodist Book Concern,
Cincinnati, Ohio. 664 p.
Minter, N.J., and Braddy, S.J., 2006a. Walking and jumping with Palaeozoic
apterygote insects: Palaeontology, v. 49, p. 827-835.
Minter, N. J., and Braddy, S. J., 2006b. The fish and amphibian swimming
traces Undichna and Lunichnium, with examples from the Lower Per
mian of New Mexico, USA: Palaeontology, v. 49, p. 1123-1142.
Minter, N. J., Braddy, S. J., and Davis, R. B., 2007b. Between a rock and a
hard place: Arthropod trackways and ichnotaxonomy: Lethaia, v. 40, p.
365-375.
Minter, N. J., Krainer, K., Lucas, S. G., Braddy, S.J., and Hunt, A. P., 2007a.
Palaeoecology of an Early Permian playa lake trace fossil assemblage
from Castle Peak, Texas, USA: Palaeogeography, Palaeoclimatology,
Palaeoecology, v. 246, p. 390-423.
Mossman, D. J., and Place, C. H., 1989, Early Permian fossil vertebrate
footprints and their stratigraphic setting in megacycle II red beds, Prim
Point, Prince Edward Island: Canadian Journal of Earth Science, v. 26, p.
591-605.
Netto, R. G., Balistieri, P., Paz, C. P., and Lavina, E., 2003. Upper Carbon
iferous-Lower Permian rhythmites from Itararé Group (Paraná Basin,
Brazil): Are they all varvites?: III Simposio Argentino del Paleozoico
Superior y Reunión Anual de Grupo de Trabajo Argentino del Proyecto
IGGP 471, La Plata, p. 4.
Owen, R., 1852. Description of the impressions and footprints of the
Protichnites from the Potsdam sandstone of Canada: Quarterly Journal
of the Geological Society [London], v. 8, p. 214-225.
Parker, L. E. and Gierlowski-Kordesch, E. H., 2007, Paleozoic lake faunas:
Establishing aquatic life on land: Paleogeography, Palaeoclimatology,
Palaeoecology, v. 249, p. 160-179.
Pazos, P. J., 2002. Palaeoenvironmental framework of the glacial-postgla
cial transition (Late Paleozoic) in the Paganzo-Calingasta Basin (south
ern South America) and the Great Karoo-Kalahari Basin (southern Af
rica): Ichnological implications: Gondwana Research, v. 5, p. 619-640.
Picard, M. D. and High, L. R., Jr., 1981, Physical stratigraphy of ancient
lacustrine deposits: SEPM Special Publication, no. 31, p. 233-259.
Pickerill, R. K., 1992. Carboniferous nonmarine invertebrate ichnocoenoses
from southern New Brunswick, eastern Canada: Ichnos, v. 2, p. 21-35.
Plotnick, R. E., 1997. Eurypterida; in Shabica, C., and Hay, A., eds.,
Richardson’s Guide to the Fossil Fauna of Mazon Creek, p. 1-308.
Pollard, J. E., and Walker, E., 1984. Reassessment of sediments and trace
fossils from the Old Red Sandstone (Lower Devonia) of Dunure, Scot
land, described by John Smith (1909): Geobios, v. 17, p. 567-576.
Sadler, C. J., 1993. Arthropod trace fossils from the Permian De Chelly
Sandstone, northeastern Arizona: Journal of Paleontology, v. 67, p.
240-249.
Savage, N. M., 1971. A varvite ichnocoenosis from the Dwyka Series of
Natal: Lethaia, v. 4, p. 217-233.
Selden, P. A., 1985. Eurypterid respiration: Philosophical Transactions of
the Royal Society, London, B, v. 309, p. 219-226.
Selden, P. A. and Nudds, J., 2004. Evolution of fossil ecosystems: Manson
Publishing Ltd, p. 1-160.
47
Smith, A., Braddy, S.J., Marriott, S. B., and Briggs, D. E. G., 2003. Arthro
pod trackways from the Early Devonian of South Wales: a functional
analysis of producers and their behaviour: Geological Magazine, v. 140,
p. 63-72.
Smith, C. T., Budding, A. J. and Pitrat, C. W., 1961, Geology of the south
eastern part of the Chama Basin: New Mexico Bureau of Mines and
Mineral Resources, Bulletin 75, 57 p.
Smith, J., Lucas, S. G., Hunt, A. P., and Schneider, J., 2005. First report of
tetrapod tracks from the Meseta Blanca Member of the Lower Permian
Yeso Formation, central New Mexico: Journal of Vertebrate Paleontol
ogy, v. 25, no. 3, p. 116.
Toepelman, W. C., and Rodeck, H. G., 1936. Footprints in late Paleozoic
red beds near Boulder, Colorado: Journal of Paleontology, v. 10, p. 660
662.
Trewin, N. H., 1994. A draft system for the identification and description of
arthropod trackways: Palaeontology, v. 37, p. 811-823.
Trewin, N. H., and McNamara, K. J., 1995. Arthropods invade the land:
Trace fossils and palaeoenvironments of the Tumblagooda Sandstone
(?late Silurian) of Kalbarri, Western Australia: Transactions of the Royal
Society of Edinburgh: Earth Sciences, v. 85, p. 177-210.
Turek, V., 1989. Fish and amphibian trace fossils from Westphalian sedi
ments of Bohemia: Palaeontology, v. 32, p. 623-643.
Turek, V., 1996. Fish trace fossil interpreted as a food gathering swimming
trail from the Upper Carboniferous (Westphalian) of Bohemia: Casopis
Narodniho Muzea Rada Prirodovedna, v. 165, p. 5-8.
Van Allen, H. E. K., and Calder, J. H., 2003. Brule; the earliest record of
gregarious tetrapod behaviour in the world’s only known walchian coni
fer fossil forest: Geological Society of America, Abstracts with Pro
grams, v. 35, no. 3, p. 25.
Van Allen, H. E. K., Calder, J. H. and Hunt, A. P., 2005, The trackway record
of a tetrapod community in a walchian conifer forest from the Permo
Carboniferous of Nova Scotia: New Mexico Museum of Natural History
and Science, Bulletin 30, p. 322-332.
Vaughn, P. P., 1963, The age and locality of the Paleozoic vertebrates from
El Cobre Canyon, Rio Arriba County, New Mexico: Journal of Paleon
tology, v. 37, p. 283-296.
Voigt, S., 2005. Die Tetrapodenichnofauna des kontinentalen Oberkarbon
und Perm im Thüringer Wald – Ichnotaxonomie, Paläoökologie und
Biostratigraphie. Cuvillier Verlag, Göttingen, Germany, 179 p.
Walker, E. F., 1985. Arthropod ichnofauna of the Old Red Sandstone at
Dunure and Montrose, Scotland: Transactions of the Royal Society of
Edinburgh: Earth Sciences, v. 76, p. 287-297.
Walter, H., 1982. Neue Arthropodenfährthen aus den Oberhöfer Schichten
(Rotliegendes, Thüringer Wald) mit Bemerkungen über Ichnia limnisch
terrestricher Tuffite innerhalb der varistischen Molasse: Freiberger
Forschungschefte, C 382, p. 87-100.
Weber, B, and Braddy, S. J., 2004. A marginal marine ichnofauna from the
Blaiklock Glacier Group (?Lower Ordovician) of the Shackleton Range,
Antarctica: Transactions of the Royal Society of Edinburgh: Earth Sci
ences, v. 94, p. 1-20.
Woodworth, J. B., 1900. Vertebrate footprints on Carboniferous shales of
Plainville, Massachusetts: Bulletin of the Geological Society of America,
v. 11, p. 449-454.
48

|
 Lucas et al., eds., 2010, Carb-Permian transition in Cañon del Cobre. New Mexico Museum of Natural History and Science Bulletin 49.

49
EURAMERICAN LATE PENNSYLVANIAN / EARLY PERMIAN
ARTHROPLEURID/TETRAPOD ASSOCIATIONS – IMPLICATIONS FOR THE
HABITAT AND PALEOBIOLOGY OF THE LARGEST TERRESTRIALARTHROPOD

JÖRGW. SCHNEIDER1, SPENCERG. LUCAS2, RALFWERNEBURG3 AND RONNYRÖßLER4

1TU Bergakademie Freiberg, Institut für Geologie, B.v.Cotta-Strasse 2, D-09596 Freiberg, Germany;

2 New Mexico Museum of Natural History and Science, 1801 Mountain Road N.W., Albuquerque, New Mexico 87104;
3 Naturhistorisches Museum Schloss Bertholdsburg, Burgstrasse 6, D-98553 Schleusingen, Germany;

4 Museum für Naturkunde, Moritzstrasse 20, D-09111 Chemnitz, Germany

Abstract—The giant arthropod Arthropleura was a common member of the late Paleozoic continental biota of
paleo-equatorial biomes for more than 35 million years, from the Early Carboniferous late Visean (FOD; Middle
Mississippian, Asbian/Brigantian) up to the Early Permian lower Rotliegend (LOD = ?LAD; Asselian). In Upper
Pennsylvanian red beds in Cañon del Cobre of northern New Mexico, trackways of Arthropleura are present in
strata that also yield body fossils of the amphibian Eryops. We review the Arthropleura tracksite from Cañon del
Cobre, New Mexico, as well as other tracksites of this animal and arthropleurid/eryopid associations in order to
better interpret the paleoenvironmental preference and the paleobiology of Arthropleura. This review supports
the conclusion that Arthropleura was well adapted to alluvial environments of ever wet humid to seasonally dry
and semihumid climates. Preferred habitats of semi-adult and adult Arthropleura were open, vegetated, river
landscapes. They co-occurred in these habitats with semi-aquatic eryopid amphibians and terrestrial pelycosaurs.

INTRODUCTION ments in order to better interpret the paleoenvironmental preference and

Since the first description of Arthropleura (von Meyer, 1853;
Jordan and v. Meyer, 1854; cf. Kraus and Brauckmann, 2003), most
remains of this enigmatic giant arthropod were sampled by plant collec
tors and palaeobotanists in plant-rich roof shales of coal seams from
mine dumps. This led to the assumption that this animal preferably lived
in the tight hygrophilous to hydrophilous vegetation of Carboniferous
coal-forming tropical forests (e.g., Rolfe and Ingham, 1967; Rolfe, 1969,
1980). However, an interesting observation, which would have modified
this picture very early, has been overlooked: Guthörl (1940) reported
from his detailed paleontological core investigation of the deep drilling
Hangard No. 38 the discovery of Arthropleura remains at six levels in the
depth interval between 50 to 350 m in the Westphalian D Saarbrücken
Group of the Saar-Nahe basin. Remarkably, those finds do not come
from roof shales of coal seams but from a 235-m-thick sequence of
alluvial plain deposits without any indication of coal-forming forests and
mires. Obviously, Arthropleura body remains are as common in alluvial
plain deposits as they are in deposits of peat environments.
In contrast to body remains, which could be transported over
large distances, tracks of animals are absolutely autochthonous. Since the
first discovery of Arthropleura tracks in the Westphalian B (Bashkirian/
Moscovian) of Nova Scotia, Canada (Ferguson, 1966), at least 11 more
tracksites became known in North America and Europe, ranging from the
Visean of Scotland (Pearson, 1992) up to the Stephanian/Lower Rotliegend
transition of Germany (Early Permian, Asselian: Walter and Gaitzsch,
1988) as well as the assumed Early Permian of Nova Scotia (Ryan,
1986). These track records considerably widen our knowledge of the
locomotion and paleoenvironments of Arthropleura (Briggs et al., 1979,
1984; Wilson, 2003). The most recent description of an Arthropleura
tracksite reports the co-occurrence with the eryopid amphibian track
Limnopus in the Upper Pennsylvanian of Kentucky (Martino and Greb,
2009). It provides additional strong evidence of the co-existence of
arthropleurids and eryopid amphibians in alluvial environments outside
of coal-forming forests and swamps, which was discussed earlier by
Schneider and Barthel (1997), and by Schneider and Werneburg (1998).
Here, the Arthropleura tracksite from Cañon del Cobre, New Mexico
(Lucas et al., 2005c) (Fig. 1) will be discussed in comparison with other
tracksites of this animal as well as further arthropleurid/eryopid environ
the paleobiology of Arthropleura.
Institutional abbreviations: FG – Technische Universität
Bergakademie Freiberg, Paleontological Collection, Germany; MfNC–
Museum für Naturkunde, Chemnitz, Germany; MHNA – Museé
d´Histoire naturelle, Autun, France; MTB - private collection M. Thiele
Bourcier, Kirkel-Neuhäusel, Germany; NHMS - Naturhistorisches Mu
seum Schloss Bertholdsburg, Schleusingen, Germany; NMMNH – New
Mexico Museum of Natural History and Science, Albuquerque, New
Mexico, USA; SSB – Sammlungen Schloß Burgk, Feital, Germany.
LITHO- AND BIOFACIES OF THE DIPLICHNITES
CUITHENSIS OCCURRENCE IN THE EL COBRE CANYON
FORMATION, LATEPENNSYLVANIAN/EARLYPERMIAN,
CHAMA BASIN, NEWMEXICO
Nonmarine siliciclastic red beds at the base of the Phanerozoic
section across most of the Chama Basin of northern New Mexico are
assigned to the Pennsylvanian-Permian Cutler Group. These strata are
divided into two mappable lithostratigraphic units, the El Cobre Canyon
and overlying Arroyo del Agua formations (Krainer and Lucas, 2001;
Lucas and Krainer, 2005). The El Cobre Canyon Formation is up to 500
m of brown siltstone, sandstone and extraformational conglomerate of an
ephemeral braided and anastomosed stream environment (Eberth and
Miall, 1991) that overlies Proterozoic basement in the subsurface and is
conformably overlain by the Arroyo del Agua Formation. Siltstone beds
of the El Cobre Canyon Formation contain numerous rhizoliths (Fig.
2E-F) and comprise relatively thin, slope-forming units between
multistoried sandstone beds that are arkosic, micaceous, coarse grained
and trough crossbedded (Fig. 2A-C). Scoyenia traces are typical of hori
zontal bedded, mica-rich siltstones (Fig. 2 D). Complete bioturbation of
the Planolites montanus type is rare.
The El Cobre Canyon Formation approximately corresponds to
megasequence 1 of Eberth and Miall (1991). According to Eberth and
Miall (1991), during deposition of megasequence 1 the climate was semi
arid, major channels were broad and shallow, interfluve areas were fre
quently inundated by sheet floods and splays, and floodbasin ponds
were relatively common. These sediments are devoid of coal, but do
contain leaf floras dominated by seed ferns (DiMichele and Chaney,
2005).
50

FIGURE 1. Location map of Cañon del Cobre in northern New Mexico and stratigraphy and fossil locality distribution (numbers are fossil localities of the
New Mexico Museum of Natural History and Science) of the Upper Pennsylvanian and Early Permian strata in Cañon del Cobre (after Lucas et al., 2005b).
 51

FIGURE 2. Facies pattern and facies-typical fossils of wet red beds of the El Cobre Canyon Formation (Late Pennsylvanian/Early Permian) in Cañon del
Cobre. A, Typical alluvial plain multistoried sandstones and pedogenically-overprinted siltstones, which make up the whole profile of the formation. B,
One of the rare, horizontal-bedded, shallow lacustrine mudstone and fine sandstone horizons, which is erosively overlain by trough-cross bedded channel
sandstones; at channel bottoms are large mudstone intraclasts (arrows). C, Interbedding of lenticular sandstones of minor channels with overbank
siltstones, covered by a major channel sandstone. Siltstones altered to calcic soil with bright red to whitish-bluish color mottling and horizons and layers
of carbonate nodules. D, Fine sandstone with cf. Scoyenia burrows. E, Densely arranged vertical rhizoliths in siltstone. F, Horizontal rhizoliths weathered
out from siltstone.
52
The age of the El Cobre Canyon Formation is Late Pennsylva
nian-Early Permian (Lucas et al., 2005b; Lucas, 2006). Fossils from the
lower part of the formation in the floor of Cañon del Cobre indicate a
Late Pennsylvanian age: palynomorphs (Utting and Lucas, 2010, this
volume), megafossil plants (Alethopteris flora: Smith et al., 1961; Fracasso,
1980; Hunt and Lucas, 1992; DiMichele and Chaney, 2005; DiMichele
et al., 2010, this volume) and fossil vertebrates such as Desmatodon and
Limnoscelis (Fracasso, 1980; Lucas et al., 2005b).
In general, the red-bed sequences of the El Cobre Canyon Forma
tion belong to the type called “wet red beds” (Schneider et al., 2006;
Roscher and Schneider, 2006). Wet red beds are characterized by the
following litho- and biofacies markers, which are observable infine clastics
of overbank deposits in channel-sandstone-dominated alluvial plains as
well as in fine, clastic-dominated floodplain deposits:
1. Horizontal, planar to weakly flaser bedded, fine sandy to clayey,
often mica rich, red-brown to red siltstones with common Scoyenia bur
rows and/or completely bioturbated by Planolites montanus burrows
(Scoyenia ichnofacies). In places, laminated siltstone and claystone of
shallow pools and lakes (Fig. 2B) preserve traces of the Mermia
ichnoguild.
2. Bed-scale, pedogenically overprinted, in situ micro-brecciation
(mm- to cm-scale), pedogenic slickensides (vertisols) and color mottling.
3. In places, mm-thin, mainly horizontal arranged and branched
root systems, preserved as red clay inundations or leached whitish.
4. Calcic soils of different maturity, ranging from mm- to dm-sized
calcareous nodules and rhizoconcretions up to massive calcrete horizons
(Fig. 2C, E, F).
5. Rarely, dm-thick lacustrine micritic limestones with ostracods,
gastropods and characeans.
6. Channel sandstones and conglomerates that may be leached to
bright reddish and whitish-greenish because of paleo-groundwater flow.
7. Rare desiccation cracks and raindrop imprints.
In contrast to the semiarid to arid dry red beds of playa and
sabkha environments, wet red beds are indicative of a semihumid to
semiarid climate with seasonal wet and dry conditions. Evaporation is
higher than precipitation, allowing the formation of calcic soils. Seasonal
high groundwater levels allow for sparse vegetation only, which is merely
documented by the root structures. Above ground (surface) macro-re
mains are mainly rare or missing. In the paleo-equatorial belt, these kind
of red beds were first widespread in the Westphalian D after the
Westphalian C/D wet phase A of Roscher and Schneider (2006) and can
be observed in all the succeeding dry phases up to the intensive aridization
during the late Early Permian (Kungurian: cf. Roscher and Schneider,
2006, fig. 15a-b; Schneider et al., 2006), when they are increasingly
replaced by playa red beds.
The El Cobre arthropleurid trackways (Lucas et al., 2005c) occur
in the lower part of the exposed section of the El Cobre Canyon Forma
tion on the top surface of a 13- to 15-cm-thick bed of mainly fine- to
medium-, slightly coarse-grained, small scale trough-cross bedded to ripple
laminated, micaceous, arkosic channel sandstone at NMMNH locality
6037 (Figs. 3-4). The channel sandstone is erosively cut into small scale,
trough-cross bedded, dark reddish-brown-violet, fine-sandy siltstone and
it is overlain by grayish-red, plant-bearing muddy siltstone beds with
dm-wide fine sandstone lenses. The siltstone beds are indistinctly hori
zontal bedded; common cm-size hematitic nodules as well as roots point
to a low pedogenic overprint. They can be regarded as an immature
ferralatic paleosol.
Two trails with the typical features of Diplichnites cuithensis
were found (Hunt et al., 2004; Lucas et al., 2005c), both running nearly
to the west about 5 m distant from one another. The best preserved track
is stored now in the New Mexico Museum of Natural History and
Science (NMMNH P-45287). It consists of two track rows of mostly
crescentic imprints oriented perpendicular to the midline (Fig. 4B; for
further details see Lucas et al., 2005c). Imprints are 5-7 cm wide and up
to 1.5 to 2 cm deep (Fig. 4C-E). The external width of the trackway is 32
38 cm. Based on the body width/length ratio of 3.6 to 4.4 (Kraus, 1993;
Schneider and Werneburg, 1998), an Arthropleura of 1.37 to 1.67 m body
length was the trail maker.
The track-bearing sandstone horizon is interpreted as the result of
lateral avulsion on the floodplain of minor side channels of a major
channel in a braided river system. Larger conglomeratic channels are
exposed 1.95 m above the track-bearing horizon (Fig. 3), which could
easily be regarded as lateral equivalents of the minor channels. A part of
the tracks is preserved as more than 2 cm deep, impressed, crescent
shaped undertracks, well visible where the higher parts of the sandstone
with the surface, on which the animal walked, were weathered away (Fig.
4C-D). The remarkable depth of the tracks as well as the soft sediment
deformation with indications of plastic to fluid behavior of the river sand
during trackmaking indicate that the animal moved on wet sand (Fig. 4B
E). Identical observations were made on several other Arthropleura
tracksites (see below). The sandstone is strongly carbonate cemented,
which may indicate high primary porosity as a prerequisite for fluidiza
tion. Bright red to whitish colors and greenish patches (Fig. 4B-D) result
from leaching by paleo-groundwater flow, which is responsible for the
synsedimentary to early diagenetic carbonate cementation as well.
About 15 m above the Arthropleura tracksite (NMMNH locality
6037), another NMMNH locality (6121) contained a well preserved
Eryops skull together with bones of an embolomere, of the dissorophid
Platyhystrix and pelycosaurs in muddy sandstones (Lucas et al., 2005a;
Werneburg et al., 2010, this volume).
A rare ichnofossil assemblage of the nonmarine Late Carbonifer
ous was recently discovered at NNMNH locality 5874 about 35 m
below the Arthropleura tracksite (Lucas and Lerner, 2010, this volume).
The assemblage comes from a clastic lake-margin facies and consists of
arthropod trackways (Diplichnites x Diplopodichnus, Diplichnites gouldi,
cf. Paleohelcura tridactyla, Protichnites isp.), invertebrate grazing and
feeding traces (Gordia indianensis and Helminthoidichnites tenuis), tet
rapod footprints (Batrachichnus salamandroides and aff. Amphisauropus
isp.) and fish swim traces (Undichna isp.) The numerically dominant
ichnofossils are Helminthoidichnites and Diplichnites; the otherichnotaxa
are represented by a few specimens. Overall, the wet red beds of the El
Cobre Canyon Formation belong to the Scoyenia ichnofacies (cf. Buatois
and Mángano, 2007) and it is composed of members of the Diplichnites
and Mermia ichnoguilds, typical of floodplains and shallow lacustrine
settings (Buatois et al., 1998; Buatois and Mángano, 2009).
LITHO- AND BIOFACIES OF THEARTHROPLEURA OCCUR
RENCE IN THE MANEBACHFORMATION, EARLYASSELIAN,
THURINGIAN FOREST BASIN, GERMANY
The Thuringian Forest Basin (formerly SW-Saale basin), an ap
proximately 40 to 60 km wide NW-SE oriented depression, is largely
exposed in the horst structure of the Thuringian Forest Mountains in
East Germany. The basin is situated at the northern border of the Variscan
orogen above deeply-eroded Variscan methamorphites and granites. The
Gzhelian/Asselian (Late Stephanian) to Guadalupian/Lopingian basin
fill of about 6000 m thickness is subdivided into nine formations (Lützner
et al., 2007). The coal-bearing (in the type area up to 180 m thick)
completely gray Manebach Formation was deposited in a low-relief
landscape by forested swamps, local lakes, and fine-clastic-dominated
fluvial deposits rich in organic matter. Volcanic rocks in the Manebach
Formation are restricted to mm to cm thick ash layers within lacustrine
black shales. This formation is famous for its characteristic and well
investigated Euramerican Stephanian/Lower Rotliegend (Gzhelian/
Asselian) flora (e.g., Barthel, 2001, 2003-2008).
Based on both the lithofacies and fossil content the following sub
environments can be distinguished in the Manebach Formation (Figs.
5A, 6; for details see Schneider, 1996; Werneburg, 1997; Barthel, 2001,
2003-2008; Lützner, 2001):
 53

FIGURE 3. Sketch of the Arthropleura tracksite at NMMNH locality 6037 (compare Fig. 4) in alluvial plain deposits of the El Cobre Canyon Formation
(Late Pennsylvanian/Early Permian) in Cañon del Cobre. 1, 30 cm thick (base not exposed), silty fine sandstone, dark reddish-brown-violet, mica rich,
trough-cross bedded, intercalated with up-to-10-cm-thick lenticular fine sandstone. 2, 13-15 cm thick, bright red to whitish, fine- to medium-grained, small
scale trough-cross bedded to ripple laminated, micaceous arkosic channel sandstone; at the top is the up-to-38-cm-wide Diplichnites cuithensis trail,
NMMNH P-45287, extending W-E. 3, 45 cm thick, silty fine sandstone, mica rich, grayish-red-brown, small scale trough-cross bedded, and two
intercalated, 5-cm-thick, discontinuous layers of fine-grained sandstone lenses. 4, 40 cm thick, silty fine sandstone, dark brown, structureless/massive. 5,
10 cm thick, silty fine sandstone, greenish leached (immature soil). 6, 100 cm thick, fine sandy siltstone, dark brown, at mm-scale indistinctly horizontal
bedded, containing plant fragments as well as cm-size hematitic nodules, some with about 1 mm wide central root channels (tubes) (immature ferralitic soil).
7, up to 45 cm thick, pebbly coarse channel sandstone, whitish, trough cross-bedded. 8, 70 cm thick (top eroded), fine conglomeratic coarse channel
sandstone, dark brown, mainly rounded to sub-rounded quartz pebbles.

1. Medium- to coarse-grained, pebbly, trough cross-bedded chan homogenized by bioturbation (Pelecypodichnus).

nel sandstones; common are stem and strobili remains of Calamites gi
gas, twigs of meso- to xerophilous conifers (“walchians“) and skeletal
remains of the eryopid amphibian Onchiodon (Fig. 5A) as well as iso
lated bones of a pelycosaur (?Haptodus) (Werneburg, 2007). In the im
mediate neighborhood of channels as well as in point bar sandstones and
channel sandstones, C. gigas has been found buried in an upright posi
tion (Fig. 5E). At the Manebach localities, this unique succulent calamite
forms (nearly) monotypic stands with about 1 m distance between the
stems (Barthel and Rößler, 1996; Barthel, 2001).
2. Fine- to coarse-grained sheet sands generated during flood events
as overbank and crevasse splay deposits; autochthonous Calamites gi
gas stands are present as well as allochthonous remains of meso- to
xerophilous plants from different growth sites above the groundwater
level, such as sand bars along river courses and from drier, elevated areas
inside the basin and along the basin borders (callipterids, different coni
fers such as walchians, the coniferophyte Dicranophyllum and
Odontopteris lingulata, etc.).
3. Fine, sandy siltstones to clayey floodplain siltstones deposited
during waning stages of flood events with layers of species-rich
parautochthonous (often well-preserved large fern fronds), hygro- to
mesophilous plant remains, representative of the fern-pteridosperm
calamite vegetation of floodplain areas outside the peat-forming mires;
Arthropleura remains are not rare.
4. Laminated claystones and siltstones of floodplain pools be
tween the channels, with layers of Anthraconaia (Fig. 5D) in places
5. Rooted siltstones and claystones of very wet floodplains, and
in places pure hydromorphic to subhydric cordaitalean root horizons of
coal-forming forest mires. In the roof shales of seams, the autochthonous
swamp forest communities are preserved; most common are Psaronius
ferns with their fronds (Pecopteris, Scolecopteris) and pteridosperms,
like Odontopteris schlotheimii, Dicksonites pluckenetii, Taeniopteris
jejunata, different neuropterids and others, the hygrophilous Calamites
multiramis and C. undulates as well as the coal-forming cordaitaleans
(Barthel, 2001); insect remains are present (mostly cockroaches).
6. Above the coal-seam-containing part of the profile there appear
lacustrine, carbon-rich siltstones and claystones with the typical Early
Permian palaeoniscid-xenacanthid-fish association of smaller lakes
(Schneider et al., 2000). Branchiosaur amphibians are very rare; common
lacustrine invertebrates are ostracods, and, in layers, conchostracans; the
terrestrial biota is represented by diverse plant fragments and blattid
insects (most common, as in many Euramerican lake sediments, is
Anthracoblattina).
The Manebach locality has been sampled by private plant collec
tors and palaeobotanists for about 300(!) years (Barthel and Rößler,
1995). Collecting has focused on the plant-rich roof shales of the coal
seams. Arthropleura and tetrapod skeletal remains were not discovered
before the first paleontological research in the fluvial deposits of this
profile commenced (Werneburg, 1987; Schneider and Werneburg, 1998).
Obviously, they are restricted to those fluvial deposits and their deposi
tional environments. From reconstructed leg length and the size of
54

FIGURE 4. Overview of the Arthropleura tracksite, NMMNH locality 6037, and the preservation of Diplichnites cuithensis in alluvial plain deposits of
the El Cobre Canyon Formation (Late Pennsylvanian/Early Permian) in Cañon del Cobre. A, D. cuithensis trail on top of a sandstone horizon originated
from channel avulsion; both the trail-bearing, bright red to whitish channel sandstone and the dark reddish-brown, silty fine sandstone below are trough
cross-bedded. B, Up to 38 cm wide D. cuithensis trail (NMMNH P-45287), which is weathered down to different levels of undertracks; the black star marks
the track shown in figs. D and E, double arrow indicates the inside/outside orientation of the trail and tracks in figs. D and E. C, Deepest undertracks in
front; tracks become shallower and sharper in outline in the direction of the hammer. D, Closeup of the upper track row in B; the transition from simple
crescent undertracks (right) to increasingly shallow but more detailed tracks (left) is not caused just by the varying depth of modern erosion of the trail but
also by the primary depth of imprints, which is 1.5 cm at the track marked by the black star and more than 2 cm (measured from the level of the modern
erosion surface) for the track above the scale bar; the depth of undertracks obviously depends on changing wetness of the channel sand. E, Closeup of tracks
in B (for orientation see black star); the form of the imprints indicates fluid behavior of the sand during trackmaking. F, Thin section of the track-bearing
sandstone; grain size ranges from dominantly medium to fine sand with some coarse grains, mainly angular, mostly quartz followed by plagioclase, rare
lithoclasts, strongly carbonate cemented; strong cementation point on high primary porosity and water fill, which was responsible for the fluid behavior
of the still uncemented sand during trackmaking.
 55

FIGURE 5. Exposure and fossil content of the Manebach Formation (Early Permian) at the type locality south of Manebach village, slope at the B4 road,
Thuringian Mountains, Germany. A, Excavation site 2005; numbers 2 to 8 refer to the meter-scale in Figure 6; remains of the eryopid Onchiodon
thuringiensis Werneburg, 2007 were found in the pebbly channel sandstones at 5; Arthropleura remains come from plant-bearing overbank siltstones
between 7 and 8 as well as from loose blocks. B, ca. 30-cm-long skull of Onchiodon thuringiensis; NHMS-WP 2140a. C, Cardiocarpus fructifications
washed together in fluvial fine sandstone; NHMS-WP 3310. D, Freshwater pelecypod Anthraconaia in floodplain pool siltstones; NHMS-WP 3350. E,
Trunk base of Calamites gigas buried upright in fluvial sandstone; NHMS-WP 3407.

paratergits (“pleurites”), body lengths of 0.85 m to 2.25 m have been
calculated for the individuals from Manebach. Based on the above char
acterized litho- and biofacies types, the biotope of Arthropleura and
Onchiodon during Manebach time could be described as follows.
The Manebach Formation belongs to the Late Gzhelian/Early
Asselian wet phase C of Roscher and Schneider (2006), which is repre
sented, e.g., in the Saar-Nahe basin of western Germany, by the
Breitenbach to Altenglan formations and in the French Massif Central by
the Molloy and Igornay formations (Roscher and Schneider, 2006, fig.
15 a-b). Red sediment colors are almost completely absent; only in the
alluvial fan conglomerates of the basin border facies do violet-colored
coarse clastics appear in places. Characteristically, coarse channel clastics
are of whitish-yellowish color, which is interpreted as the result of
leaching by paleo-groundwater flows. This is supported by the presence
of pale gray leached, primary dark violet to reddish brown rhyolite clasts
of the fan and channel conglomerates (Lützner, 2001; Lützner et al.,
2007). Lamination as an indication of seasonality is not really well ex
pressed in the lake deposits. Therefore, nearly year-around high precipi
tation as well as high groundwater levels can be inferred. In this way, the
Manebach Formation is climatically very close to the Westphalian C
(Bolsovian) and early Westphalian D, from which most Arthropleura
remains were discovered in Europe. Both Arthropleura and Onchiodon
lived outside the swamp areas in a river landscape that was dominated by
Calamites gigas stands along the river banks as well as by fern-pteri
dosperm-calamite vegetation on floodplain areas between river courses.
LITHO- AND BIOFACIES OF THEARTHROPLEURA OCCUR
RENCE IN THE DÖHLEN FORMATION, EARLYASSELIAN,
DÖHLEN BASIN, GERMANY
This Late Pennsylvanian to Early Permian basin forms a small
half graben of 22 km by 6 km with its primary extent in the Elbe zone,
which is part of the NW-SE striking Elbe lineament (Schneider, 1994).
The basement and the border of the basin are formed by different Variscan
rock units such as the Meißen Intrusive Complex, metamorphosed and
folded early Paleozoic sediments as well as the Ore Mountain (Erzgebirge)
gneisses. Basin topology and the high number of clastic dikes (formed by
paleo-earthquakes) indicate that basin development and tectonic activity
along the Elbe zone were coeval. Preserved basin fill amounts to about
800 m of thickness and is subdivided into four fining-up megacycles or
formations. In contrast to other Permocarboniferous basins, pyroclastic
rocks dominate, comprising up to 50% of the deposits of this basin
(Schneider and Hoffmann, 2001). Generally, sedimentation and facies
pattern are governed by higher subsidence along the main fault at the
southwestern graben border as well as by strong volcano-tectonic activ
ity, including strong seismicity.
The Döhlen Formation (as much as 100 m thick) comprises two,
approximately 50-m-thick fining-up mesocycles. The mainly red-col
56
ored basal conglomerates are overlain by gray sandstones and siltstones,
locally with seams of carbonaceous shale. The second mesocycle con
sists of grayish to yellowish-white fluvial pebbly arkoses, pyroclastic
rocks, and five seams of carbonaceous shale and coal beds up to 6 m, and
locally 12 m, thick. As in the Manebach Formation of the Thuringian
Forest basin, red beds are completely missing in the coal-bearing upper
mesocycles. Primary dark violet rhyolite clasts of channel conglomerates
are leached to pale gray. The presence of subaerial to subaquatically
deposited air fall pyroclastics and fluvially-reworked pyroclastics be
tween the coal seams indicates that nearly continuous peat formation
occasionally was interrupted by strong ash falls. These ash falls are
responsible for in situ preservation of plant communities at their growth
sites (Figs. 7F, 8; Barthel, 1976; Schneider and Barthel, 1997; Rößler and
Barthel, 1998). Upright-standing (up to 3 m tall) calamite trunks at the
tops of seams indicate that they were buried during catastrophic ash-fall
events.
Because of the absence of surface outcrops, the lithofacies and
biofacies pattern of the Döhlen Formation are known only from subsur
face coal mining and well cores in the coal-seam-containing facies in front
of the alluvial fans that frame the northeastern basin border. Generally,
facies architectures are governed by volcanotectonic subsidence and re
gional ash falls. Characteristic are vertical and lateral changes between
fluvial tuffitic pebbly arkoses, fine-grained ash tuffs and coarse-grained
crystal tuffs, shallow lacustrine tuffitic pelites and marlstones as well as
dirty coals and pure coals. Peat formation was interrupted several times
by ash falls on the swamps, which form basin wide, cm-to-dm-thick
pyroclastic layers (tonsteins) within the coal seams as well as the sharp
upper boundaries of the main seams. The ash falls have buried plant
communities in situ, often as monospecific stands (sinuses) (Figs. 7F, 8),
and very often the trunks of calamites in upright position.
Based on lithofacies and fossil content the following sub-environ
ments could be distinguished in the Döhlen Formation (for details see
Schneider and Barthel, 1997; Rößler and Barthel, 1998; Schneider and
Hoffmann, 2001):
1. Well-drained channel arkoses, sand flats of coarse grained tuffs/
tuffites and elevations inside the mire bear monospecific, in situ stands
of the xerophytic Calamites gigas, rarely together with single C. undulatus.
2. Inside the mire, pure stands of the hydrophilous to hygrophil
ous, peat-forming, rhizome-bearing Calamites multiramis (=Calamites
striata in anatomical preservation) are present on wet organic substrates
(mostly buried in upright position).
3. Hygrophilous forest swamp communities in permanently wa
terlogged areas with broad-leafed, coal-forming cordaitaleans and their
typical roots (Fig. 7G-H), psaroneaceous tree ferns, Calamites multiramis,
Sphenophyllum oblongifolium and rare Nemejcopteris feminaeformis as
well as other ferns; fluent lateral transitions into pure cordaitalean swamps
as well as into C. multiramis stands.
4. Together with 3, locally monotonous scrambler communities
(sinuses) of S. oblongifolium and N. feminaeformis in places with
Botryopteris, Dactylotheca, Taeniopteris and Calamites multiramis.
5. Mesophilous pteridosperm-fern-communities (open pioneer
communities) of Oligocarpia leptophylla, Senftenbergia, Dactylotheca,
different pecopterid ferns and pteridosperms like Dicksonites,
Neuropteris and Barthelopteris, and rarely, Autunia and Subsigillaria
brardii, on well-drained elevated drier areas inside and at the borders of
the swamps; the border facies of the swamp is characterized by shallow
but wide fluvial channels filled with reworked pyroclastics.
6. Mesophilous to xerophilous conifers on slopes and hills sur
rounding the basin and possibly on well-drained elevations and swells
inside the basin, too.
The articulated ventral part with sternites and legs of a recon
structed, approximately 0.85-m-long Arthropleura was found in a 6-cm
thick, slightly fluvially-reworked pyroclastic horizon. This horizon is
situated at the top of a 0.6-m- to 0.8-m-thick, medium to coarse, air fall
crystal tuff that overlies with sharp contact and load casts the third coal
seam in the mining field Bannewitz of the Döhlen basin. Both pyroclas
tic horizons contain the above-characterized mesophilous pteridosperm
fern community (sub-environment 5). Compared to the mining field
Gittersee, 3.5 km away, the discovery horizon in the mining field
Bannewitz was situated on an elevation generated by the ash falls, as
shown by Schneider and Barthel (1997, fig. 3). Immediately overlying
strata are 3-4 m of sandy tuffites containing tight Calamites gigas stands.
The occurrence of eryopids in the Döhlen Formation is demon
strated by the discovery of Limnopus tracks in pyroclastics between the
first and second coal seams of the Döhlen Formation (Hausse, 1910, and
later discoveries). The track-bearing horizons are cm-thick, silty to fine
sandy pyroclastics with desiccation cracks. Additionally, in sediments
just above the uppermost or first seam, which is about 6.5 m above the
track-bearing horizon, the famous concentration of six skeletons of
Haptodus (Pantelosaurus) saxonicus was discovered (Hausse, 1902; v.
Huene 1925). Thus, Arthropleura, eryopids and haptodonts (pelyco
saurs) can be regarded as quasi-contemporaneous in the above-character
ized drier environments with hygro- to mesophilous plant communities
of the coal-forming Döhlen Formation.
COMPARISON OF PARAUTOCHTHONOUS ARTHROPLEURA
BODY REMAIN OCCURENCES AND AUTOCHTHONOUS
TRAIL OCCURRENCES
The oldest true Arthropleura body fragments (legs, paratergits),
which mark the oldest occurrence of this animal, are known from the
Visean (late Asbian to early Brigantian, Middle Mississippian) in the
Hainichen basin of East Germany (Rößler and Schneider, 1997). Leg
lengths of about 6 cm (Fig. 9A) indicate individuals of 18 cm body width
and 0.8 m body length; lengths of up to 1 m are indicated by the size of
the paratergits (Fig. 9B) (calculation based on Rolfe, 1969; Schneider and
Werneburg, 1998), which is in good agreement with the body sizes calcu
lated from tracks from the Visean (see below). The Arthropleura remains
occur together with other terrestrial arthropods such as arachnids (e.g.,
the phalangiotarbid Bornatarbus mayasi and the trigonotarbid
Aphantomartus areolatus) and scorpionids as well as shark egg capsules
of the Fayolia type. They are preserved in fluvial sandstones of an
alluvial braid plain/floodplain environment with upright, buried stumps
of arborescent lycopsids and sphenopsids of the Archaeocalamites
radiatus type. These trees colonized channel banks, mires and open
forested areas, the latter containing higher amounts of mesophilous plants
(Schneider et al., 2005).
The next oldest record of Arthropleura body remains comes from
the Westphalian A (Langsettian, Middle Bashkirian, Early Pennsylva
nian) of the Ostrava-Karvina basin in the Czech Republic (eho and
ehoova, 1972). In younger strata, the finds increase in abundance,
with a maximum in the Westphalian C (Bolsovian) and D (Moscovian,
Middle Pennsylvanian), which is caused simply by widespread coal
mining in deposits of this time slice in both Europe and North America
(Hannibal, 1997). Guthörl (1936) counted 105 Arthropleura fragments
in the Westphalian B and C (Duckmantian and Bolsovian) of the Saar
basin (Guthörl, 1936, fig. 1 and p.189-195), which led him to the conclu
sion that this arthropod “is the most common animal fossil of the
Saarbrücken beds.”
The vast majority of these finds come from roof shales of coal
seams, stored on the mine dumps. (Since the 1980s, a German private
collector, Michael Thiele-Boucier [personal comm.], has sampled more
than 100 Arthropleura remains from coal mine dumps of the Westphalian
C and D in the Saar basin). Roof shales represent the silting-up of swamps
by fluvial deposits in intracontinental basins and by tidalites in paralic
basins (Gastaldo et al., 2004). Floras and faunas of roof shales are,
therefore, a nearly indistinguishable admixture of terrestrial and aquatic
organisms of swamps, floodplain environments and shallow marine en
 57

FIGURE 6. Sedimentology and fossil content of the Manebach Formation (Early Permian) at the type locality south of Manebach village, slope at the B4
road, Thuringian Mountains, Germany; excavation site 2005, compare Figure 5A. Channel lag deposits at profile-meter 5 contain calamite trunks as well
as the Onchiodon remains – skull about 30 cm long with the lower jaw as well as femur (Fem) and ilium (Il) (from Gebhardt et al., 1995; completed after
the 2005 excavation by Werneburg).
58

FIGURE 7. Arthropleura and selected fossils from the Döhlen Formation (Early Permian) of the Döhlen basin in Saxony, Germany; (D from the Saar basin,
Germany, see below). A-B, Arthropleura in partially fluvially redeposited pyroclastics above the 3rd coal seam in the mining field Bannewitz-North of the
Döhlen basin, ca. 24 cm long portion of ventral side showing the well preserved articulated limbs as well as the elements of the ventral exoskeleton, E1
– 1st limb segment, E8 – 8th (last) limb segment with the terminal claw, Sp – movable articulated unpaired spines at each limb segment, R – rosette plate,
PM – ventral part of the flexible pleural membrane on which the sclerites S and K are fixed, S – sternites, K – K-plate, B – non-sclerotised B-element (?a
kind of tracheal pocket); apodema of the limbs in black; FG Thümmel collection. C, Closeup from A, intensely wrinkled B-element and the strongly
sclerotised undeformed sternite and K-plate; sternite and K-plate show on the outer (ventral) surface fine dimples with increasing density towards the
posterior borders of these sclerites. D, Ventral sclerites; note the fine dimples on the outer surface of the sternite and the K-plate as in C and the smooth
surface of the B-element apart from some dimples at the base only; early Westphalian D, pit 4 at Göttelborn, Saar basin, Germany; MTB 1215. E,
Limnopus, tracks of an eryopid amphibian in pyroclastics between the 1st and 2nd coal seams of the Döhlen Formation; SSB. F, Apical portion of a Calamites
multiramis shoot with Annularia spinulosa leaves, buried in situ by volcanic ash fall, pyroclastics above the 2nd coal seam, Döhlen Formation, Zauckerode;
FG 15/98. G, Root horizon in fine-grained volcanic ashes with three-dimensionally preserved cordaitalean roots, Döhlen Formation, Zauckerode; FG 609.
H, Cordaites principalis, pyroclastics above the 3rd coal seam, Döhlen Formation; private collection Hertl, Freital, Germany.
 59

FIGURE 8. Reconstruction of the floral associations buried in situ above the 3rd coal seam of the Döhlen Formation (Early Permian) in the Döhlen basin
based on plant collections from the mining fields Gittersee and Bannewitz (from Schneider and Barthel, 1997). The Arthropleura ventral side, shown in
Figure 7A-C, comes from the mining field Bannewitz. Floras of the Gittersee type belong to the hygro- to hydrophile peat-forming communities of poorly
drained environments close to or below the groundwater level. The Bannewitz flora represents the pioneering mesophile associations on well-drained
mineral stands above groundwater level. The open vegetation of comparable environments on sandy substrates of alluvial plains and deltaic settings seems
to have been the preferred habitat of semi-adult and adult Arthropleura.

vironments. Additionally, roof shale samples from mine dumps lack
sedimentological context. They are not suited to answer the question of
the environmental preference of Arthropleura. As shown by Guthörl
(1940) for the deep drilling Hangard No. 38 (see above), Arthropleura
finds are as common in coal-seam-free alluvial plain deposits as they are
in roof shales. This fits with the still scattered but nearly continuous
records of Arthropleura tracks from the Namurian up to the Early Per
mian of Europe and North America in alluvial environments outside
swamp areas. Those occurrences will be characterized in the following
sections based on literature data and personal observations.
Strathclyde Group, Visean (Asbian, Mississippian), Fife,
Scotland (Pearson, 1992)
The sediments of the Strathclyde Group of Fife are interpreted as
deltaic in origin with occasional marine incursions. Nonmarine facies
include channel sandstones and sheet-flood deposits, thin coals, seatearths,
lagoonal mudstones and laminated siltstones and shales as well as red or
black ironstones and dolomites with nonmarine bivalves. The track
bearing part of the section consists of meter-thick dark, organic-rich and
often finely laminated mudstones and siltstones interspersed with ~ 1.5-
m-thick, white, well-sorted, medium-grained channel-fill sandstones. The
sandstones display multi-storied trough-cross bedded units with undu
lating surfaces and, in places, current ripples. Parts of the track-bearing
sandstone bed are spotted with small mineralized rootlets; also present
are poorly-preserved Stigmaria root traces. Minimally, 13 individual
trails occur, ranging in width between 18 and 30 (?46) cm; most of them
have a relatively narrow range between 21-24 cm. The longest preserved
trail is 5 m, forming an S-bend. Some trails cross one another; at one place
the sediment is pitted by deeply impressed, crescent-shaped and ran
domly oriented marks, which may represent hundreds of individual foot
falls. Generally, individual tracks often appear crescent shaped and deeply
impressed (1-2 cm estimated depth from the photographs in Pearson,
1992), comparable to the El Cobre trails. Not excluding the possibility
that the trails are made by two or three individuals only, this mass
occurrence of Arthropleura trails did not originate by chance. Animals of
0.80 m to 1.32 m body length (maximally ?2.11 m for the uncertain, 46
cm wide trail) were the trail makers.
Namurian Limestone Coal Group (Pendleian, Early
Serpukhovian), Arran, Scotland (Briggs et al., 1979)
In general, the Limestone Coal Group was deposited in a proximal
deltaic environment. The trail-bearing horizon is situated 6 mm below
the top of a 6-m-thick white sandstone unit, which formed the roof of
coal seams. The trail is preserved on the surface of a bedding plane of
heterogeneous sandstone ranging in grain size from quartz arenite to fine
sandstone and containing discontinuous layers of shale. Bedding struc
tures indicate deposition in a variable low-velocity flow regime in a
gradually silted-up fluvial channel. Rootstructures suggest that the veg
etation was penecontemporaneous with the formation of the trail. A
subaqueous origin of the trail is regarded as unlikely because of the
clarity of the imprints in such a coarse lithology. The width of the
trackway is 36 cm, suggesting a 1.6-m-long individual as the trackmaker.
60
FIGURE 9. Oldest known Arthropleura remains, from the Visean (late Asbian to early Brigantian, Middle Mississippian) in the Hainichen basin, Borna near
Chemnitz, Germany (Rößler and Schneider, 1997). A, Counterpart of a nearly complete, 7.6 cm long limb, R – rosette plate, E1 – 1st limb segment, E8 –
8th (last) limb segment; MfNC F 439b. B, Counterpart of a fragmentary paratergite (“pleurite”) with the typical sculpture of A. armata from the
Pennsylvanian; MfNC F 1125.
Little River Formation, Cumberland Group, ?Late Namurian to
basal Westphalian, (?Kinderscoutian/Langsettian, Bashkirian,
Early Pennsylvanian), Lower Cove, Nova Scotia, Canada (Calder
et al., 2005)
The Little River Formation of the Lower Cumberland Group was
deposited in the Cumberland sub-basin as part of the late Paleozoic
Maritimes Basin complex of Atlantic Canada (Falcon-Lang et al., 2006).
The formation consist of a wet red bed-succession dominated by
mudrocks, which exhibit pervasive mottling from root traces and local
pedogenic carbonate, with channel sandstone bodies typically 3-6 m
thick. It represents the deposits of a well-drained alluvial plain dissected
by shallow rivers characterized by flashy flow under a seasonally dry
climate. Coals seams and brackish, marine-bivalve-bearing limestone beds,
which are present in the overlying Joggins Formation, are missing. A 23
cm wide Diplichnites cuithensis is reported from inside a body of stacked
channels by Calder et al. (2005). The tracemaker may have had a length
of about 1 m. This is the stratigraphically oldest record of Arthropleura
in wet red beds of seasonally dry environments.
Joggins Formation, Cumberland Group, Westphalian A
(Langsettian; Bashkirian, Early Pennsylvanian), Joggins, Nova
Scotia, Canada (Ferguson, 1966, 1975; Briggs et al., 1979;
Falcon-Lang et al., 2006)
The Early Pennsylvanian Joggins Formation, containing the fa
mous Joggins paleoecosystem, overlies the Little River Formation (see
above). At that time, the Cumberland sub-basin was connected to the
open sea during sea-level highstands, as indicated by brackish incursions,
and was more restricted and intra-continental during sea-levellowstands
(Falcon-Lang et al., 2006). The latter authors revised the fossil content of
the Joggins Formation in a precise facies context, recognizing three main
communities: open water brackish sea fossil assemblages (OW), poorly
drained coastal plain assemblages (pPDF) and the 31% of the Joggins
Formation comprising well-drained alluvial plain red-bed assemblages
(WDF). Arthropleura trails were reported from the last two assem
blages. In the “grey mudstone with channel bodies facies” of the pPDF,
a few D. cuithensis trails (Falcon-Lang et al., 2006) and Dromillopus
microsaur trackways (Cotton et. al., 1995) occur on top of some channel
bodies.
The trails reported by Briggs et al. (1979) occur in the “red mud
stone with channel bodies facies” of the WDF (Falcon-Lang, et al., 2006).
This facies comprises red mudstone successions with scattered pedogenic
carbonate nodules, sandstone sheets, and small, ribbon-like channel sand
stone bodies. They contain common upright calamiteans, and a few
upright lycopsids with attached Stigmaria roots. Plant assemblages are
dominated by cordaitaleans (Dadoxylon, Mesoxylon, Cordaites, Artisia)
as well as a few lycopsids and calamiteans. The trails occur in sheet
sands thickening into a channel sand bed. Ranging in total width from 20
to 26 cm, they indicate an animal about 0.88 to 1.15 m long. The preser
vation of the largest trail, described as a paired series of regularly-spaced,
oval depressions, elongate normal to the axis, resemble the preservation
of the El Cobre trackways in wet sandstone. The smaller trail shows
individual imprints, apparently arranged in closely spaced diagonals.
These differences in preservation are explained by decreasing water con
tent of the sediment, i.e., increased cohesiveness when the smaller trail
was formed.
Tynemouth Creek Formation, Westphalian A to ?B (Langsettian -
?Duckmantian; Bashkirian/?Moscovian, Pennsylvanian), New
Brunswick, Canada (Briggs et al., 1984)
The Tynemouth Creek Formation is regarded as partially coeval
to the Joggins Formation of the late Paleozoic Maritimes Basin complex
of Atlantic Canada (see above and Falcon-Lang, 2006). The formation is
built up mainly of red siltstones, red and gray sandstones, and coarse

conglomerates in an overall coarsening upward sequence (Briggs et al.,
1984). Rare freshwater limestones are locally present. The sedimentary
environment is interpreted as a relatively dry alluvial fan environment
characterized by periodic sheetfloods across an otherwise quiescent area
of relatively slow deposition. Red/green mottled paleosols with slicken
sides and carbonate nodules are interpreted as vertisols of a seasonally
dry climate (Falcon-Lang, 2006). The trail-bearing section consists of
red, fine, slightly silty, tabular sandstones, interbedded with red and
green siltstones. The up-to-2-m-thick sandstones are dominantly mas
sive, but include planar and cross-laminated units. Upright buried calam
ite stems, up to 10 cm in diameter, are common.
The best-preserved trails occur at Gardner Creek in the top of a
40-cm-thick, fine sandstone with numerous in situ calamitaleans rooted
in the 12-cm-thick siltstone horizon below. The fine sandstone grades up
into a few mm of siltstone, and it is on this surface that the Arthropleura
trail is preserved. The sandstones are interpreted as deposits of major
sheet floods. The trail-bearing siltstone accumulated during the waning
phase of the flood, and the trail was produced after subaerial emergence,
when the Arthropleura was walking along a sinuous course through the
calamite forest. The width of the trail in the straight part is 29.5 cm,
suggesting an animal about 1.30 m long. Two further trails are mentioned
in the top of channel sandbodies beneath overbank sediments at approxi
mately the same stratigraphic level as the Gardner Creek trails. Their
width of 27 cm and 30 cm implies similar large animals.
Cumberland Group, Westphalian B-C (Duckmantian-Bolsovian;
Moscovian, Middle Pennsylvanian), Smith Point and Pugwash,
Cumberland County, Nova Scotia, Canada (Ryan, 1986)
The Westphalian Cumberland Group of Cumberland County con
sists, according to Ryan (1986), of continental gray to red-gray calcare
ous mud-chip conglomerate, coarse- to fine-grained sandstone, siltstone
and mudstone with minor amounts of coal, limestone and shale. Sand
stones and conglomerates are trough cross-stratified and were deposited
in low sinuosity streams. At Smith Point, 25 trails were found on two
different surfaces that are separated by a 1.3-m-thick, cross-stratified,
medium-grained sandstone. Crossover of trails is common; up to 300
degrees of bending is observed. The average width of the trails on the
upper bedding surface is 37 cm, and it is 32 cm on the lower bedding
surface, which gives a body length of 1.63 and 1.40 m, respectively, for
the producer. The tracks are closely spaced, about 24 per meter; maxi
mum depth of the tracks is 3 cm. At the Pugwash locality, 12 trails
preserved mainly as undertracks were found. The trail width is approxi
mately 38 to 41 cm; depth of tracks varies between 2 and 3 cm. The
length of the trail-producing Arthropleura is calculated as about 1.8 m.
At both localities, the tracks occur in pebbly arkosic sandstones
that are overlain by thin veneers of mudstone, preserved on some of the
trail surfaces. The depositional environment is interpreted as mid-chan
nel dunes prograding during flood stages, which were subsequently sub
aerially exposed as channel bars during a dry period, in which the trails
were produced.
Stull Shale Member, Kanwaka Formation, Shawnee Group,
Virgilian (Gzhelian, Late Pennsylvanian), Waverly, Eastern
Kansas (Mángano et al., 2002)
The Stull Shale Member of the Forest City basin in eastern Kan
sas represents tidal flat deposits along a microtidal shoreline directly
connected with the open sea. The arthropleurid tracks occur at the top of
a 30-cm-thick, medium- to fine-grained channelized sandstone body 66
m wide, which cuts erosively into an intertidal runoff channel. Both
channel bodies are encased by progressively shallowing intertidal mixed
to mudflat deposits. The track-bearing sandstone is interpreted as a
channel fill in a coastal fluvial system. The width of the trackways is
23.3 to 30.2 cm, which indicates body lengths of about 1 to 1.3 m. The
preservation style resembles very closely the El Cobre trackways with
61
closely spaced, elongated and sigmoidal imprints oriented normal to the
axis of the trackway. Details of the imprints are not preserved because of
the superimposition of imprints and soft-sediment deformation, which
is visible in figure 32C of Mángano et al. (2002). Obviously, the sandy
sediment was wet and plastically deformable when the animal was walk
ing on it, as was the case with the El Cobre trackway.
Lower Conemaugh Formation, Virgilian (Gzhelian, Late
Pennsylvanian), Boyd County, Kentucky (Martino and Greb,
2009)
The lower Conemaugh Formation of the Appalachian basin in
Kentucky is regarded as an equivalent of the Glenshaw Formation of the
Conemaugh Group in neighboring West Virginia, which consists pre
dominantly of coastal plain fluvial sandstones and red to olive mudrocks
deposited in a tropical flood-basin setting. Martino and Greb (2009)
reported new direct evidence for the co-occurrence of Arthropleura with
eryopid amphibians, indicated by the tracks of both on one-and-the
same bedding surface from the lower Conemaugh Formation, 6.5 m be
low the marine Ames Limestone. The roof shales of the Ames Limestone
can be directly correlated with the early Stephanian of Central Europe
based on fossil insects (Schneider and Werneburg, 2006). The tracks
were found in float from a roadcut.
The trackways were derived from a 4-m-thick interbedded sand
stone and shale interval 4.3 m below the base of the Harlem coal. The
tracks are preserved as convex hyporeliefs in fine to very fine sandstone
that is interbedded with silty shale. They were mainly produced on top
of the silty shale. In one case (specimen PC-6 of Martino and Greb,
2009, p. 142, fig. 4), an Arthropleura trail is preserved as undertracks in
the sandstone together with a Limnopus manus cast (?undertrack). The
track-bearing sandstones occur as three broad, thin sheets and a small
channel fill. This, together with graded bedding, parallel and current
ripple cross lamination and desiccation cracks leads to the interpretation
of crevasse splay deposits. Besides sparse root traces, no plant remains
were reported. The Pittsburgh red shale, which directly underlies the
trackway interval, is indicative of the climate conditions during deposi
tion. These red beds consist of regionally-developed calcic vertisols,
which classify them as wet red beds in a seasonal dry/wet semihumid to
semiarid climate (see above and Greb et al., 2006). Open woodland
vegetation is assumed. Based on the width of the Arthropleura tracks of
20 and 30 cm, we infer a body length of the tracemaker of 0.88 m for one
of the tracks and 1.32 m for two other tracks. The Limnopus traces
consist of 75 to 80 mm long pes and 53? to 57 mm wide manus imprints.
An additional tracksite from the Conemaugh Group, Casselman Forma
tion, in Cambria County, Pennsylvania, was briefly reported by Marks
et al. (1998). The 15 cm wide by 3 m long trackway resembles D.
cuithensis as far as can be determined from the published photographs.
Montceau-les-Mines, latest Stephanian C/earliest Autunian
(Asselian, Cisuralian, Early Permian), French Massif Central
(Briggs, 1986; Briggs and Almond, 1994)
The Blanzy-Montceau basin is one of the typical Late Pennsyl
vanian/Early Permian continental coal basins of the French Massif Cen
tral. It is famous for its fossiliferous nodules similar to Mazon Creek in
Illinois. Arthropleurid trails were found in a unit of alternating pale gray
mudstone and fine sandstones at the level of the first coal seam of the
Assise de Montceau (Langiaux, 1984). The trails are preserved at the top
of whitish-gray, silty, fine sandstone that contains numerous plant frag
ments (Fig. 10C; Briggs, 1986). The width of the trail of 10.8 cm indi
cates a length of the producing arthropleurid of 0.47 m. About 15 m
above the trail-bearing level, three-dimensionally preserved tiny
arthropleurids of 2.7 to 4.5 cm length have been found in sideritic nod
ules (Fig. 10A-B; Secretan, 1980; Almond, 1985; Briggs and Almond,
1994). The width of these dwarf arthropleurids ranges from 1.1 to 1.7
cm. They differ from the smallest true Arthropleura (A. moyseyi Calman,
62

FIGURE 10. The tiny arthropleurids of Montceau-les Mines, France, and trails of juvenile arthropleurids from Flechting High, Germany. A-B, Two
approximately 1.6 to 1.7 cm wide, three-dimensional arthropleurids preserved in sideritic nodules, Assise de Montceau, latest Stephanian C/earliest
Autunian (Asselian, Cisuralian, Early Permian), Montceau-les-Mines, French Massif Central; A, 3.7 cm long natural cast of the ventral (inner) side of the
lateral paratergits and the dorsal side of the body as well as two incomplete ventral molds of syntergites that are about half of the length of the specimen;
anteriormost plate of the ventral mold of a collum or cephalic shield (cf. Briggs and Almond, 1994; Kraus et al., 2003); MHNA 002122b. B, 3.6 cm long
natural cast of the ventral (inner) side of the lateral paratergits and the dorsal side of the body as well as on the right distal side of the specimen imprints
and counterprints of closely arranged limbs; note the strongly tapered posterior region of the trunk; MHNA 002123b. C-D, D. cuithensis trails of juvenile
Arthropleura preserved as convex hyporelief in silty-clayey pyroclastics, Eiche Member, Altmark Subgroup, Stephanian ?C (Late Pennsylvanian/Early
Permian, Ghzelian/Asselian), Flechting High, Germany (see Walter and Gaitzsch, 1988); lower trail 2.0 to 2.2 cm wide; note the faint striae (especially in
the bent part of the trail) made by the movable spines of the limbs during forward stroke; this trail was made in wet plastic mud; upper trail 3.2 cm wide;
tracks have a lesser width despite the larger trail width and show more distinct limb end claw imprints of repeated foot falls; this trail was possibly made in
a somewhat drier substrate; note the very common small cf. Batrachichnus tetrapod tracks; FG 607.

1915) of about 6 cm body length by a distally very strong, tapering body
and possible diplopody (Briggs and Almond, 1994) as well as very long
and thin, spike-like sculpture cones on the paratergits (JWS, personal
observation on silicon rubber casts). They may belong to Arthropleura
or represent a new minute form of arthropleurid — for discussion see
Briggs and Almond (1994) and below.
The depositional environment is described as a fault-bounded
graben with rapid lateral and vertical transitions of alluvial fan deposits,
debris flows and braided river clastics interfingering with palustrine (lo
cally up to 40 m thick coal seams!) and lacustrine sediments. Siltstones,
intercalated into the first coal seam have yielded a diverse tetrapod track
association (Langiaux, 1984; Gand, 1994), produced mainly by
temnospondyl amphibians (e.g., Batrachichnus, Limnopus) and pelyco
saurs (Dimetropus). The fossiliferous nodules and track-bearing Assise
de Montceau is commonly attributed to the Stephanian B/C based on
macroflora. Spiloblattinid cockroach wings indicate a latest Stephanian
C/earliest Autunian, i.e., Early Permian (Asselian) age (Schneider and
Werneburg, in prep.). A transitional Stephanian/Autunian age had been
advocated by Doubinger (1994) based on palynomorphs, but was never
taken into account.
Eiche Member, Altmark Subgroup, Stephanian ?C (Late
Pennsylvanian/Early Permian, Gzhelian/Asselian), Flechting
High, northeast Germany (Walter and Gaitzsch, 1988)
The Eiche Member is exposed at the top of subeffusive basaltic
volcanites situated near the base of the Late Pennsylvanian/Early Per
mian Altmark Volcanite Complex of northeast Germany (Gaitzsch et al.,
1995). It consists mainly of about 10 m of light-gray fluvial sandstones
at the base followed by several meters offine ash tuffs covered by debris
flow fanglomerates with an ignimbrite at the top. The track-bearing unit
is composed of (?water laid) grayish-green air fall tuffs as well as fluvially
reworked and lacustrine, redeposited fine ash tuffs. Shallow channels
with trough-cross bedding have a lateral width of only about one meter.
Sedimentation took place under low energy conditions from suspension
as indicated by the silty grain size and normal grading. Raindrop marks,
shallow wave ripples and rare “anthracosian” (?Anthraconaia) imprints
as well as Pelecypodichnus traces indicate the presence of shallow pools.
The track-bearing layers are horizontally bedded at a 1 to 2 cm scale.
Tens of arthropleurid trails per square meter are preserved on the bed
ding planes, in places as undertracks. Crossover and bending is common
(Fig. 10D).
Because of their small size, ranging from 2.0 to 4.0 cm width, the
trails have been described as Diplichnites minimus Walter and Gaitzsch,
1988. Apart from bundles of striae situated in front of the single tracks in
both track rows of some trails (Fig. 10D, lower track) no real differences
exist when compared to D. cuithensis. The striae result from the spines
on the limb segments during forward stroke. Therefore, D. minimus
should be regarded as a junior subjective synonym of D. cuithensis.
Based on the width of the trails, a length of 8.8 to 17.6 cm of the
trailmakers can be calculated. This corresponds well to the smallest
known, unquestionable Arthropleura (A. moyseyi Calman, 1915) with a
body length of about 6 cm. There are no indications of the strong poste
rior tapering of the body as observed for the tiny “juvenile” arthropleurids
from Montceau-les-Mines (see Briggs and Almond, 1994, figs. 1 and 2;
here Fig. 10D). As discussed by Briggs and Almond (1994), the posterior
tapering of the trunk and the closely-spaced rear limbs on either side
should result in a trail in which the imprints occupy over 50% of the total
width. The trail width of these small forms should be about 1.6 cm,
corresponding to the largest width in the middle part of the body. This is
very close to the smallest trails of 2 cm width from the Eiche Member.
But, the trails from this member do not show any indication of a trailmaker
with a posteriorly tapering body – they display the normal picture of
two rows of tracks, as do the larger D. cuithensis. It is therefore very
unlikely that the minute arthropleurids from Montceau-les-Mines be
63
long to the same Arthropleura species as the D. cuithensis trail makers.
The only way to produce trails of D. cuithensis type by the tiny
Montceau-les-Mines arthropleurids would be if the posteriormost limbs
were held clear of the ground (Briggs and Almond, 1994, p.132).
The Eiche Member D. cuithensis occur on different bedding planes.
In places, normal surface trails (epichnia) are crossed by undertracks of
trails produced later after a new layer of sediment was laid down. They
are commonly associated with small cf. Batrachichnus trails of 3 to 7 mm
track length (Fig. 10D). The wide range of preservation forms of the
arthropod and tetrapod trails indicate different amounts of wetness of
the substrate, ranging from possibly thin water cover over muddy to
nearly stiff. Sparse remains of macroflora, mainly calamite trunks (up to
5 cm diameter) and Annularia leaves as well as rare fern leaves, indicate
scattered vegetation in surrounding areas of small ponds. The high fre
quency of arthropleurid trails (laterally and vertically) and in places the
common association with trails of small tetrapods demonstrates that this
environment was settled by juveniles of Arthropleura. The lack of larger
trails may indicate a different biotope preference of juveniles and adults.
Pictou Group, Upper Red Beds, Early Permian (Cisuralian), Cape
John, Pictou County, Nova Scotia, Canada (Ryan, 1986)
The Early Permian Upper Red Beds, now called the Cape John
Formation (Ryan et al., 1991), of the Pictou Group are composed of
fluvial cycles of red to gray conglomerates and arkosic sandstones as well
as siltstones and mudstones. The trail-bearing pebbly arkosic sandstones
are interpreted as anastomosing river deposits, with sediment transport
and accumulation primarily as the result of dune progradation at channel
bottoms. At Cape John, on the northern coast of Nova Scotia, three trails
were found, and two of them cross each other. The width of the trails is
on average 47 cm, giving a body length of the arthropleurid of about 2.07
m. They are preserved as two parallel rows of tracks, each 13 cm wide,
elongate and nearly normal to the trail axis with 28 tracks per meter. The
depth of the tracks is about 1.3 cm.
In the Cape John Formation as well, about 8 km southwest of the
Cape John Arthropleura tracksite, a spectacular tetrapod tracksite to
gether with in-situ stumps of a walchian conifer forest is exposed at
Brule Harbour, Colchester County (Van Allen et al., 2005). The tetrapod
tracks represent a typical cosmopolitan Euramerican association with
strong European affinity, consisting of Limnopus, Amphisauropus,
Dimetropus, Varanopus and Dromopus and others. Absolutely domi
nant is Limnopus, followed closely by Amphisauropus. Provisionally
determined invertebrate traces, as Gordia, Beaconichnus, and
Onisciodichnus, are typical representatives of the shallow freshwater
Mermia ichnoguild of the Scoyenia ichnofacies, an interpretation that is
strongly supported by the co-occurrence with the ostracod Carbonita
and leaiaid conchostracans. Besides Paleohelcura, no arthropod walking
trails are reported. Most tetrapod tracks are preserved in thin, silty
sandstone beds commonly draped by very thin mud layers. Desiccation
cracks are typical. The depositional environment of the Brule Harbour
tetrapod tracksite is described as an abandoned distributary channel
hollow that was repeatedly flooded and exposed. Red beds of Prince
Edward Island, adjacent and contiguous to Brule Harbour, have yielded
skeletal remains of eryopid and brachyopoid amphibians and the
reptiliomorphs Seymouria and Diadectes as well as of different pelyco
saurs (Van Allen et al., 2005).
THE HABITAT PREFERENCE OFARTHROPLEURA
Based on the foregoing, Arthropleura trail-bearing paleoenviron
ments can be characterized as follows:
1. Arthropleurid trails of dm width are preferably preserved both
in nearshore transitional terrestrial/marine deltaic and tidal flat sediments
as well as in intracontinental river braid plains on sandflats and on bed
ding planes at the top of channel sandstones. Smaller trails up to about 5
cm wide are preserved in silty mud of temporary pools.
64
2. As a prerequisite for the preservation of tracks, the substrate
must be wet (of course) when the animal was walking on it; some of the
trails show indications of plastic deformation/fluidization of the sand
and mud caused by deeply impressed tracks (Fig. 4D-E).
3. Vegetation of the tracksites is scattered, consisting mainly of
calamitaleans, and during the Westphalian, additionally of arborescent
lycopsids, as indicated by their upright buried stumps and Stigmaria
rhizomes/rootlets.
4. Associated with the track-bearing sandstones in coastal marine
environments are gray lagoonal and fluvial overbank mudstones and silt
stones, more or less rooted. Additionally, tracks are present since the
middle Westphalian in alluvial plain and floodplain red siltstones and
mudstones of wet red-bed facies type in continental settings.
5. In rare cases only, coal seams or swamp deposits are associated
with the Arthropleura track-bearing facies architectures.
6. The most common tetrapods in those settings are terrestrially
adapted eryopid amphibians (Figs. 5B, 7E) and pelycosaurid reptiles.
The first question to be answered is: do the trail occurrences
reflect the biotope preference of the trail makers or is it biased by sam
pling or by preservation potential? Of course, the discovery of large
trails of 20 cm and more width and several meters long, and the chance to
recognize them as trails, obviously depends on the dimensions of the
exposed bedding planes. They are largest in areas of high recenterosional
rates such as at modern sea coasts and in sandstone quarries. There, the
fine clastics covering track-bearing sandstones are quickly eroded on
large surfaces or – in the case of the quarries – removed. This is clearly
shown, for example, in the case of the tracksites along modern coasts of
Scotland, by the exceptional exposures along the coast of Nova Scotia
and by the channel sandstones exposed in the modern, non-vegetated and
seasonally dry river bed of Cañon del Cobre as well. The apparent
absence of these large trails in Central Europe, where Arthropleura body
remains are very common, is simply caused by the lack of such large
exposure surfaces in upper Paleozoic fluvial sandstone units.
The preservation potential for animal tracks (invertebrates and
vertebrates as well) is generally highest in fine clastics of river flood
plains and playas (e.g., Haubold and Katzung, 1978; Haubold and Lucas,
2001; Voigt, 2005; Lucas, 2005). Tetrapod tracks are very common on
floodplain mudflats and around temporary pools on floodplains that
became dry after floods. But, those tetrapod track-bearing sediments
have not delivered any associated arthropleurid tracks on the same bed
ding planes. Even so, they are broadly contemporaneous with Arthropleura
tracksites, as in the case of Cañon del Cobre (Lucas and Lerner, 2010)
and the case of the spectacular tetrapod tracksite of Brule Harbor, Nova
Scotia (Van Allen et al., 2005). The only two exceptions are the Limnopus
tracks together with 20 to 30 cm wide Arthropleura trails in the Upper
Pennsylvanian of Kentucky (Martino and Greb, 2009; see above) as well
as the minute Batrachichnus tracks together with 2-4 cm wide
Arthropleura trails in the Upper Pennsylvanian/Lower Permian of the
Flechting High, Germany (Walter and Gaitzsch, 1988; see above).
As shown before, the vast majority of Arthropleura trails are
preserved in sandy substrates. Additionally, those trails are observed
(depending on exposure conditions) mostly as mass occurrences (e.g., up
to 25 trails at Smith Point, Nova Scotia – see above). Besides the above
cited 10 cm wide Montceau-les-Mines trail and the small Flechting High
trails, the width of the trails has a range of 18 to 47 cm, which indicates
body lengths ranging from about 0.80 to 2 m. The variety of body lengths
for particular sites, calculated from the width of the trails (e.g., Fife,
Scotland: 18-30 cm; Kentucky 20-30 cm; see above), indicates animals of
different growth stages as tracemakers (despite variations in trail width
caused by locomotory behavior).
This leads to the conclusion that arthropleurids avoided muddy
substrates, except for early juveniles of up to about 0.15 m body length,
as at Flechting High (Fig. 10D). The preservation in sandy substrates
begins by about 0.8 m body length. This may be caused simply by the
fact that only animals of this size were heavy enough to leave tracks in
wet sands. Alternatively, the absence of tracks produced by arthropleurids
larger than 0.18 m in the mass occurrence at Flechting High may indicate
different habitat preferences of juveniles and adults. Nevertheless, it
could be stated that arthropleurids, starting no later than 0.8 m body
length, walked preferably on sandy substrates. This confirms Mángano
et al. (2002), who described the depositional environment of D. cuithensis
as “typical subaerial, commonly exposed fluvial bars, silted channels,
and desiccated sheetflood deposits.” More generally, the preferred habi
tat of adult arthropleurids could be characterized as loosely vegetated
sandy areas in open river landscapes under a year-round wet climate in
coastal environments to seasonally wet, semihumid climates in continen
tal settings. As indicated by skeletal remains and tracks, the most com
mon tetrapods in these habitats since the start of the Late Pennsylvanian
were eryopid amphibians and pelycosaurid reptiles (haptodonts).
IMPLICATIONS FOR THE
PALEOBIOLOGY OFARTHROPLEURA
The most frequently discussed issues concerning the anatomy
and physiology of Arthropleura concern the up-to-now-missing struc
tures of the head, especially the mandibles and the missing respiratory
organs as well as the arrangement of the limbs (Brauckmann et al., 1997;
Kraus and Brauckmann, 2003). Kraus and Brauckmann (2003) and Kraus
(2005) discussed the phylogenetic relationships of Arthropleura as well
as the anatomy and biology of this animal based on the reinvestigation of
the tiny Montceau-les-Mines arthropleurids, the holotype of A. moyseyi,
the large, well-preserved, so-called Maybach-specimen (Guthörl 1935;
Hahn et al., 1986), and the Arthropleura ventral side from the Döhlen
basin (Schneider and Barthel, 1997) as well as several new body frag
ments, including a presumed cephalic shield and presumed head cap
sules, from the Saar basin. Some of their observations and inferences as
well as those of other authors will be discussed here with regard to our
conclusions on the paleobiology of Arthropleura.
As shown by Schneider and Barthel (1997), based on a well
preserved, 42-cm-long part of a ventral side with about 25 limb seg
ments, there are no indications that succeeding pairs of legs form groups
of two pairs. In contrast, the legs had an almost regular arrangement
(their pl. 7; here Fig. 7A-B). Whether or not one pair of legs corresponds
to one of the dorsal syntergites still remains open (cf. Schneider and
Barthel, 1979, fig. 5). Diplopody, as shown by the tiny Montceau-les
Mines specimens (Briggs and Almond, 1994), can be excluded for
Arthropleura of more than 80 cm body length. A change from diplopody
in early juvenile stages to a regular arrangement of single limb pairs in late
juvenile and adult stages can also be excluded. Therefore, the inferences
of Briggs et al. (1979) regarding the interpretation of Arthropleura traces
still appear to be correct. However, the assumption of Kraus (2005) that
the tiny Montceau-les-Mines arthropleurids are early juveniles of the
giant Arthropleura as well as his conclusions on phylogenetic relation
ships to the Penicillata clade of Diplopoda are not well supported.
The respiratory structures of Arthropleura remain unknown. Rolfe
and Ingham (1967) and Rolfe (1985) argued that the K- and B-plates may
have covered a respiratory chamber. Schneider and Barthel (1997) have
shown that the K-plate was as strongly sclerotised as is the sternite plate
(Fig. 7A-C). Unlike the K-plates and the sternites, the so called B
“plates” are often wrinkled (Schneider and Barthel, 1997, pl. 8, figs. 1-3;
here Fig. 7C). Possibly, this B-element, which is situated close to the
base of each walking limb and linked to apodema, had a respiratory
function as a kind of tracheal pocket. The “spongy internal structures”
(Kraus and Brauckmann, 2003, p. 47) of K-plates are fine dimples on the
outer surface, which give the inner surface a verrucose appearance. Not
only the K-plate shows this structures but the sternites do, too (Fig.
7D), both dorsally and ventrally. Similar pits are situated on parts of the
posterior surfaces of the limbs. Remarkably, the B-elements show only
very weak pits on their median ends (Fig. 7D). A semiaquatic life and
“plastron-breathing via series of paired ventral K-plates” (ibid, p. 48)
 65

FIGURE 11. Reconstruction of the Arthropleura habitat in well-drained areas of alluvial environments with calamitaleans stands (here Calamites gigas)
under ever wet humid to seasonally dry semihumid climate conditions; in the background an eryopid amphibian, often associated with arthropleurids as
indicated by tracks and/or skeletal remains of both animals (modified after Schneider and Werneburg, 1998, fig. 15).

can be excluded because of the nature of the K-plates and the definite
subaerial habitats indicated by the above discussed trails of late juveniles
and adults. However, it cannot be excluded that a part of the small
Flechting High trails was produced under thin water cover, but this is
irrelevant to the mode of respiration.
Kraus and Brauckmann (2003) interpreted Arthropleura as a thin
skinned, caterpillar-like animal, stabilized by means of antagonistic hy
draulics of body fluids only. This interpretation is based on fern leaves
and other plant fragments impressed from below through paraterga (ibid,
fig. 6). This interpretation has to be handled with care, because most if
not all Arthropleura remains are exuvians. They exhibit unquestionably
thin cuticles – but this could be simply a result of resorption processes
during ecdysis. The common preservation of arthropleurid trails with up
to 3 cm deep impressions of the limbs as well as the commonly deep
undertracks in wet sand indicate a heavy animal with a well sclerotised
exoskeleton similar to functionally comparable modern counterparts such
as some chilopods (scolopenders).
A herbivorous diet of Arthropleura seemed to be demonstrated by
the description of gut contents by Rolfe and Ingham (1967). Kraus
(2005, p. 18-20) has reinvestigated this specimen with the result that
“the wooden parts were accidentally fossilized together with the
Arthropleura specimen and do not constitute any gut content.” How
ever, assuming that Arthropleura was a penicillate diplodod, Kraus (2005,
p. 20) concluded in a perfectly circular argument that Arthropleura was
herbivorous, as are (nearly) all diplopods. The gigantism of arthropleurids
should therefore be “an evolutionary answer to the availability of an
unusually plentiful food niche: masses of spores and pteridosperm pol
len, perhaps also prothallia.” However, feeding on spores and pollen
appears unlikely for an animal of this size, even if they are concentrated
in lacustrine deposits such as cannel coals. If Arthropleura had used this
nutrient source, lakeshores should have been most frequented by this
animal. Again, this is very improbable because Arthropleura remains are
completely missing in lake deposits thus far (see Hannibal, 1997, for the
supposed occurrence in the lacustrine black shales of Linton, Ohio).
Alternatively, fructifications, megasporophylls and cm-size seeds
(ovules) such as Cardiocarpus and Trigonocarpus could have been an
energy-rich food for large arthropods. The question is whether or not
this food source was available year-round. The production of fructifica
tions and seeds could have been very seasonal as the climate became
increasingly seasonal after the late Westphalian (Late Moscovian). Spo
radic mass occurrences of gymnosperm seeds such as Samaropsis and
Cardiocarpus in distinct layers (as in the Manebach Arthropleura local
ity – see Barthel, 2001; here Fig. 5C), are interpreted here as the result of
seasonal seed production and taphonomic effects (washed together).
Assumed carnivory and interpretation of Arthropleura as a chilopod in
Barthel and Schneider (1997) is simply an incorrect citation by Kraus
(2005, p. 20). Barthel and Schneider (1997, p. 198) solely compared
functionally the construction of the ventral side of Arthropleura with
chilopods but did not do so in terms of any phylogenetic relationship.
Nevertheless, it could not be excluded that Arthropleura was predatory
66
and could overcome tetrapods as large as itself, as is the case with mod
ern predatory chilopods such as Scolopendra. It can be speculated only
that amphibians such as the eryopids, which lived in the same environ
ment, could have easily been preyed upon by this giant arthropod. How
ever, the problem of diet remains open as long as mouthparts or true gut
contents of Arthropleura have not been found.
CONCLUSIONS
Well adapted to alluvial environments of every wet humid to
seasonal dry semihumid climates, Arthropleura was a common member
of the late Paleozoic continental biota of the paleo-equatorial biomes for
more than 35 Ma, from the Early Carboniferous Late Visean (FOD;
Middle Mississippian, Asbian/Brigantian) up to the Early Permian Lower
FIGURE 12. Stratigraphic distribution of the arthropleurid track Diplichnites
cuithensis (1-6) in different environments (gr – gray facies, wr – wet red
beds) and the occurrence of Arthropleura body remains (a - ?d). Occurrences
of D. cuithensis: 1 - Visean (Asbian, Mississippian), Fife, Scotland (Pearson,
1992); 2 - Namurian (Pendleian, Early Serpukhovian), Arran, Scotland
(Briggs et al., 1979); 3 - ?Late Namurian to basal Westphalian
(?Kinderscoutian/Langsettian, Bashkirian, Early Pennsylvanian), Nova
Scotia (Calder et al., 2005); 4 - Westphalian A (Langsettian; Bashkirian,
Early Pennsylvanian), Nova Scotia (Ferguson, 1966, 1975; Briggs et al.,
1979; Falcon-Lang et al., 2006); Westphalian A to ?B (Langsettian
?Duckmantian; Bashkirian/?Moscovian, Pennsylvanian), New Brunswick
(Briggs et al., 1984); Westphalian B-C (Duckmantian-Bolsovian; Moscovian,
Middle Pennsylvanian), Nova Scotia (Ryan, 1986); 5 - Virgilian (Gzhelian,
Late Pennsylvanian), eastern Kansas (Mángano et al., 2002); Virgilian
(Gzhelian, Late Pennsylvanian), Kentucky (Martino and Greb, 2009);
Virgilian (Late Pennsylvanian), New Mexico (Lucas et al., 2005c); 6 - latest
Stephanian C/earliest Autunian (Asselian, Cisuralian, Early Permian), French
Massif Central (Briggs, 1986; Briggs and Almond 1994); Stephanian ?C
(Late Pennsylvanian/Early Permian, Gzhelian/Asselian), Northeast Germany
(Walter and Gaitzsch, 1988); Early Permian (Cisuralian), Cape John, Pictou
County, Nova Scotia, Canada (Ryan, 1986). Occurrences of Arthropleura
body remains: a - Visean (late Asbian to early Brigantian, Middle Mississippian),
Hainichen basin, Germany (Rössler and Schneider, 1997); b - Westphalian A
(Langsettian, Middle Bashkirian, Early Pennsylvanian), Ostrava-Karvina
basin, Czech Republic (eho and ehoova, 1972); c – Westphalian B
(Bashkirian, Pennsylvanian) to Lower Rotliegend (Early Asselian, Cisuralian)
of the European basins (e.g., Hahn et al., 1986; for the youngest occurrences
in the Permian (see Schneider and Barthel, 1997; Schneider and Werneburg,
1998); in North America (Hannibal, 1997): Lower Conemaugh, ?Missourian,
Kasimovian, Ohio; Desmoinesian, Moscovian, Pennsylvania; Desmoinesian,
Moscovian, Mazon Creek, Illinois; Springhill, Atokan, Bashkirian/
Moscovian, Nova Scotia; d – upper Lower Rotliegend (Sakmarian, Cisuralian),
Döhlen basin, Germany, large chitinous cuticle fragments that may belong
to Arthropleura (Schneider and Barthel, 1997).
Rotliegend (LOD = ?LAD; Asselian) (Rössler and Schneider, 1997;
Schneider and Barthel, 1997; Schneider and Werneburg, 1998). Trails are
as common in the gray facies as in the facies of wet red beds. Preferred
habitats of semi-adults and adults were open-vegetated river landscapes.
They co-occurred in these habitats since the Late Pennsylvanian with
semi-aquatic eryopid amphibians and terrestrial pelycosaur reptilians.
But, the large size of 1 m to more than 2 m length in adult stages of
Arthropleura protected it for a long time against natural enemies. Why
did it suddenly become extinct in the Early Permian?
The changes from arborescent lycopsid- to tree fern- and
cordaitalean-dominated wetlands during the Pennsylvanian through the
Early Permian (Kerp, 2000; DiMichele et al., 2001, 2009; Falcon-Lang,
2004, Pfefferkorn, 2008) had no influence on Arthropleura as shown by
the occurrences of (parautochthonous) body remains in the cordaitalean
dominated mires of the Early Permian in the Döhlen basin and Thuringian
Forest basin of Germany (Schneider and Barthel, 1997; Schneider and
Werneburg, 1998). Obviously, Arthropleura was not conditioned to any
particular kind of mire and wetland vegetation.
It thus seems more likely that environmental changes driven by
broadscale climate developments may have been one of the causes of the
disappearance of Arthropleura. The last known occurrence of
Arthropleura is in the “wet phase C” of Roscher and Schneider (2006)
around the Gzhelian/Asselian transition. As shown by them, during the
cyclically increasing aridization towards the Late Permian, each wet
phase was drier than the preceding one. Wet phase C represents the last
extensive gray facies with coal formation in most of the European basins
(Roscher and Schneider, 2006, fig. 51a-b). The wet red beds of the fore
going dry phase of the Stephanian B to early C (Early Gzhelian) as well
as of the succeeding dry phase around the Middle Asselian are character
ized by common calcic soils and calcretes. Arthropleura had been adapted
to these seasonally dry environments since the Early Westphalian (see
above, Little River Formation, Nova Scotia).

But, the intensity of seasonality increased as is indicated by in
tensely laminated (varved) lake sediments, which are very typical of all
perennial lakes in the European basins after wet phase C (Clausing and
Boy, 2000; Roscher and Schneider, 2006). Intensified seasonal dryness
might have contributed to the demise of Arthropleura. Local populations
may only have survived up to wet phase D in the Sakmarian, if the up to
2 cm size cuticle fragments, described from the Döhlen basin by Schneider
and Barthel (1997, p. 200, pl. 9, figs. 1 and 2), belong to this animal.
Possibly, the Döhlen basin during wet phase D was one of the “wet
spots,” shrinking refuges of a formerly widespread wetland biota
(DiMichele et al., 2006). Increasingly critical prey-predator relation
ships between Arthropleura and terrestrially-adapted amphibians and
reptiles may have been an additional trigger to the demise of this giant
animal, which was the largest terrestrial arthropod of earth history.
ACKNOWLEDGMENTS
For collaboration in the fieldwork in Cañon del Cobre we thank
Larry Rinehart and Josh Smith (both Albuquerque), at the Manebach
locality in Germany Berthold Lugert (Manebach), Georg Sommer
(Schlesuingen), Silke Sekora (Freiberg) and Harald Lützner (Jena) as well
as at Flechting High, Birgit Gaitzsch (Freiberg). Manfred Barthel (Ber
lin) gave valuable information on floras of the Thuringian Forest basin
REFERENCES
Almond, J. E., 1985, Les arthropleurides du Stephanian de Montceau-les
Mines, France: Bulletin de la Societe d’Histoire Naturelle Autun, v. 115,
p. 59-60.
Barthel, M., 1976, Die Rotliegendflora Sachsens: Abhandlungen Staatliches
Museum für Mineralogie und Geologie Dresden, v. 24, p. 1-190.
Barthel, M., 2001, Pflanzengruppen und Vegetationseinheiten der Manebach
Formation: Beiträge zur Geologie von Thüringen, Neue Folge, v. 8, p.
93-123.
Barthel, M., 2003, Die Rotliegendflora des Thüringer Waldes. Teil 1:
Einführung und Keilblattpflanzen (Sphenophyllales): Veröffentlichungen
Naturhistorisches Museum Schleusingen, v. 18, p. 3-16.
Barthel, M., 2004, Die Rotliegendflora des Thüringer Waldes. Teil 2:
Calamiten und Lepidophyten: Veröffentlichungen Naturhistorisches
Museum Schleusingen, v. 19, p. 19-48.
Barthel, M., 2005, Die Rotliegendflora des Thüringer Waldes. Teil 3: Farne:
Veröffentlichungen Naturhistorisches Museum Schleusingen, v. 20, p.
27-56.
Barthel, M., 2006, Die Rotliegendflora des Thüringer Waldes. Teil 4:
Farnsamer und Farnlaub unbekannter taxonomischer Stellung:
Veröffentlichungen Naturhistorisches Museum Schleusingen, v. 21, p.
33-72.
Barthel, M., 2007, Die Rotliegendflora des Thüringer Waldes Teil 5:
Gingkophyten, Coniferophyten: Veröffentlichungen Naturhistorisches
Museum Schleusingen, v. 22, p. 17-43.
Barthel, M., 2008, Die Rotliegendflora des Thüringer Waldes Teil 6: Wurzeln
und fertile Organe. Algen und Bakterien. Vegetation: Veröffentlichungen
Naturhistorisches Museum Schleusingen, v. 23, p. 39-62.
Barthel, M. and Rößler, R., 1995, “Eine ganz unbekanndte Frucht.” 300
Jahre paläobotanisches Sammeln und Beobachten in Manebach:
Veröffentlichungen Naturhistorisches Museum Schleusingen, v. 10, p.
49-56.
Barthel, M. and Rößler, R., 1996, Paläontologische Fundschichten im
Rotliegend von Manebach (Thür. Wald) mit Calamites gigas
(Sphenophyta): Veröffentlichungen Naturhistorisches Museum
Schleusingen, v. 11, p. 3-21.
Brauckmann, C., Gröning, E. and Thiele-Bourcier, M., 1997, Kopf- und
Schwanz-Region von Arthropleura armata Jordan, 1854 (Arthropoda;
Ober-Karbon): Geologica et Palaeontologica, v. 31, p. 179-192.
Briggs, D. E. G., 1986, Walking trails of the giant arthropod Arthropleura:
67
and the Döhlen basin and provided the photograph in Figure 7G. Michael
Magnus (Freiberg) made the interpretation and photographs of thin sec
tions (Fig. 4F). Thanks to Jean-Michel Pacaud (Paris) for casts of the
Montceau-les-Mines arthropleurids and to Dominique Chabard (Autun)
for the permission to investigate the originals in the collection of the
Autun Museum. Cordial thanks to Thilo Thümmel (Paulsdorf), who
provided the Döhlen Arthropleura specimen on permanent loan to the
TU Bergakademie Freiberg. We thank Carsten Brauckmann (Clausthal
Zellerfeld) and Otto Kraus (Hamburg) for long discussions on the
paleobiology of arthropleurids; for discussions on sedimentology,
Sebastian Voigt, Olaf Elicki, Birgit Gaitzsch (all Freiberg), Ute Gebhardt
(Karlsruhe) and Berit Legler (London), and for discussions on paleocli
matology as well as for drawing Figure 12, Marco Roscher (Oslo).
Cristopher J. Cleal (Cardiff) gave information on the Fife locality in
Scotland, and Jan Fischer (Freiberg) helped with literature. Thanks to
Dieter Waloszek (Ulm) for critical hints concerning the anatomy and life
reconstruction of Arthropleura. Photographs for Figure 5C-D were de
livered by Silke Sekora (Freiberg), for Figure 7E by Harald Walter
(Freiberg) and for Figure 7D by Michael Thiele-Bourcier (Kirkel
Neuhäusel). The reviewers, Bill DiMichele and Carsten Brauckmann,
provided helpful comments. This publication is part of the German
Science Foundation projects of J.W.S. and R.W., grants DFG SCHN408/
12 and DFG WE 2833/3.
Bulletin de la Société d´Histoire Naturelle d´Autun, v. 117, p. 141-148.
Briggs, D. E. G. and Almond, J. E., 1994, The arthropleurids from the
Stephanian (Late Carboniferous) of Montceau-les-Mines (Massif Cen
tral - France); in Poplin, C. and Heyler, D., eds., Quand le Massif Central
était sous l’équateur: Un écosystème Carbonifère à Montceau-les-Mines:
Mémoires de la Section des Sciences, v. 12, p. 127-135.
Briggs, D. E. G., Plint, A. G. and Pickerill, R. K., 1984, Arthropleura trails
from the Westphalian of eastern Canada: Palaeontology, v. 27, p. 843
855.
Briggs, D. E. G., Rolfe, W. D. I. and Brannan, J., 1979, A giant myriapod
trail from the Namurian of Arran, Scotland: Palaeontology, v. 22, p.
273-291.
Buatois, L. A. and Mángano, M. G., 2007, Invertebrate ichnology of conti
nental freshwater environments; in Miller, W., III, ed., Trace fossils:
Concepts, problems, prospects: Amsterdam, Elsevier, p. 285-323.
Buatois, L. A. and Mángano, M. G., 2009, Applications of ichnology in
lacustrine sequence stratigraphy: Potential and limitations:
Palaeogeography, Palaeoclimatology, Palaeoecology, v. 272, p. 127
142.
Buatois, L. A., Mángano, M. G., Genise, J. F. and Taylor, T. N., 1998, The
ichnologic record of the invertebrate invasion of nonmarine ecosys
tems: Evolutionary trends in ecospace utilization, environmental ex
pansion, and behavioral complexity: Palaios, v. 13, p. 217-240.
Calder, J. H., Rygel, M. C., Ryan, R. J., Falcon-Lang, H. J. and Herbert, B.
L., 2005, Stratigraphy and sedimentology of Early Pennsylvanian red
beds at Lower Cove, Nova Scotia, Canada: The Little River Formation
with redefinition of the Joggins Formation: Atlantic Geology, v. 41, p.
143-167.
Calman, W. T., 1915, On Arthropleura moyseyin. sp., from the Coal
Measures of Derbyshire: Geological Magazine, v. 51, p. 541-544.
Clausing, A. and Boy, J. A., 2000, Lamination and primary production in
fossil lakes: Relationship to palaeoclimate in the Carboniferous-Per
mian transition; in Hart, M.B., ed., Climates: Past and present. Geologi
cal Society, London, Special Publication 181, p. 5-16.
Cotton, W. D., Hunt, A. P. and Cotton, J. E., 1995, Paleozoic vertebrate
tracksites in eastern North America: New Mexico Museum of Natural
History and Science, Bulletin 6, p. 189-211.
DiMichele, W.A. and Chaney, D.S., 2005, Pennsylvanian-Permian fossil
floras from the Cutler Group, Cañon del Cobre and Arroyo del Agua
68
areas, in northern New Mexico: New Mexico Museum of Natural His-
tory and Science, Bulletin 31, p. 27-33.
DiMichele, W. A., Pfefferkorn, H. W. and Gastaldo, R. A., 2001, Response
of Late Carboniferous and Early Permian plant communities to climate
change: Annual Review of Earth and Planetary Sciences, v. 29, p. 461-
487.
DiMichele, W. A., Chaney, D. S., Kerp, H. and Lucas, S. G., 2010, Late
Pennsylvanian floras in western equatorial Pangaea, Cañon del Cobre,
New Mexico: New Mexico Museum of Natural History and Science,
Bulletin, this volume.
DiMichele, W.A., Tabor, N.J., Chaney, D.S. and Nelson, W.J., 2006, From
wetlands to wet spots: Environmental tracking and the fate of Carbon
iferous elements in Early Permian tropical floras: Geological Society of
America, Special Paper 399, p. 223-248.
DiMichele, W.A., Montañez, I. P., Poulsen, C. J. and Tabor, N. J., 2009,
Climate and vegetational regime shifts in the late Paleozoic ice age
earth: Geobiology, v. 7, p. 200-226.
Doubinger, J., 1994, Spores et pollen du basin Carbonifère et Permien de
Blanzy-Montceau (Massif Central – France); in Poplin, C. and Heyler,
D., eds., Quand le Massif Central était sous l’équateur: Un écosystème
Carbonifère à Montceau-les-Mines: Mémoires de la Section des Sciences,
v. 12, p. 61-72.
Eberth, D. A. and Miall, A. D., 1991, Stratigraphy, sedimentology and
evolution of a vertebrate-bearing, braided to anastomosed fluvial sys
tem, Cutler Formation (Permian-Pennsylvanian), north-central New
Mexico: Sedimentary Geology, v. 72, p. 225-252.
Falcon-Lang, H. J., 2004, Pennsylvanian tropical rain forests respond to
glacial-interglacial rhythms: Geology, v. 32, p. 689-692.
Falcon-Lang, H. J., 2006, Vegetation ecology of Early Pennsylvanian allu
vial fan and piedmont environments in southern New Brunswick, Canada:
Palaeogeography, Palaeoclimatology, Palaeoecology, v. 233, p. 34-50.
Falcon-Lang, H. J., Benton, M. J., Braddy, S. J. and Davies, S.J., 2006, The
Pennsylvanian tropical biome reconstructed from the Joggins Forma
tion of Nova Scotia: Journal of the Geological Society of London, v.
163, p. 561-576.
Ferguson, L., 1966, The recovery of some large track-bearing slabs from
Joggins, Nova Scotia: Maritime Sediments and Atlantic Geology, v. 2, p.
128-130.
Ferguson, L., 1975, The Joggins section: Maritime Sediments and Atlantic
Geology, v. 11, p. 69-76.
Fracasso, M. A., 1980, Age of the Permo-Carboniferous Cutler Formation
vertebrate fauna from El Cobre Canyon, New Mexico: Journal of Pale
ontology, v. 54, p. 1237-1244.
Gaitzsch, B., Ellenberg, J., Lützner, H. and Benek, R., 1995, Flechtinger
Scholle; in Plein, E., ed., Stratigraphie von Deutschland I. Norddeutsches
Rotliegendbecken. Rotliegend-Monographie Teil II: Courier
Forschungsinstitut Senckenberg, v. 183, p. 84-96.
Gand, G., 1994, Les traces de vertébrés tétrapodes du Carbonifère superieur
et du Permien du bassin de Montceau-les-Mines (Massif Central – France);
in Poplin, C. and Heyler, D., eds., Quand le Massif Central était sous
l’équateur: Un écosystème Carbonifère à Montceau-les-Mines. Mémoires
de la Section des Sciences, v. 12, p. 269-281.
Gastaldo, R.A., Stevanovic-Walls, I., Ware, W.N. and Greb, S.F., 2004,
Community heterogeneity of Early Pennsylvanian peat mires: Geology,
v. 32, p. 693-696.
Gebhardt, U., Gaitzsch, B., Schneider, J. and Rössler, R., 1995, Stratigraphy
and facies of the Middle European continental Carboniferous and Per
mian. Guide to Excursion A5: 13. International Congress on Carbonifer
ous-Permian, Krakow 1995, Supplement, 48 p.
Greb, S. F., DiMichele, W. A. and Gastaldo, R. A., 2006, Evolution and
importance of wetlands in earth history; in Greb S.F. and DiMichele,
W.A., eds., Wetlands through time: Geological Society of America, Spe
cial Paper 399, p. 1-40.
Guthörl, P., 1935, Entdeckung und Bergung des größten, bis jetzt bekannten
Exemplars von Arthropleura armata Jordan & v. Meyer in Grube
Maybach-Saar: Zeitschrift der Deutschen Geologischen Gesellschaft, v.
87, p. 686-692.
Guthörl, P., 1936, Arthropleura, der Riesengliederfüßler des Obercarbons
und seine Verbreitung in den europäischen Steinkohlenbecken: Glückauf,
v. 72, p. 965-975.
Guthörl, P., 1940, Zur Arthropoden-Fauna des Karbons und Perms 12.
Insekten- und Arthropleura-Reste aus der Tiefbohrung 38 (Hangard) bei
Neunkirchen-Saar: Paläontologische Zeitschrift, v. 22, p. 109-119.
Hahn, G., Hahn, R. and Brauckmann, C., 1986, Zur Kenntnis von
Arthropleura (Myriapoda; Ober-Karbon): Geologica et Palaeontologica,
v. 20, p. 125-137.
Hannibal, J.T., 1997, Remains of Arthropleura, a gigantic myriapod arthro
pod, from the Pennsylvanian of Ohio and Pennsylvania: Kirtlandia, v.
50, p. 1-9.
Haubold, H. and Katzung, G., 1978, Palaeoecology and palaeoenvironments
of tetrapod footprints from the Rotliegend (Lower Permian) of Central
Europe: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 23, p.
307-323.
Haubold, H. and Lucas, S. G., 2001, Early Permian tetrapod tracks - preser
vation, taxonomy, and Euramerian distribution: Natura Bresciana, v. 25,
p. 347-354.
Hausse, R., 1902, Ein Massengrab von Sauriern im Unter-Rotliegenden des
Döhlener Kohlenbecken im Plauenschen Grunde bei Dresden: Jahrbuch
für das Berg- und Hüttenwesen in Sachsen, v. 1902, p. 25-50.
Hausse, R., ed., 1910, Fossile Tierfährten im Unterrotliegenden des
Steinkohlenbecken im Plauenschen Grunde (des Döhlener Beckens) bei
Dresden: Jahrbuch für das Berg- und Hüttenwesen in Sachsen, v. 1910, p.
3-19.
Huene, v., F., 1925, Ein neuer Pelycosaurier aus der unteren Permformation
Sachsens: Geologische und Paläontologische Abhandlungen, N.F., v. 14,
p. 5-52.
Hunt, A. P. and Lucas, S. G., 1992, The paleoflora of the lower Cutler
Formation (Pennsylvanian, Desmoinesian?) in El Cobre Canyon, New
Mexico, and its biochronological significance: New Mexico Geological
Society, Guidebook 43, p. 145-150.
Hunt, A. P., Lucas, S. G., Lerner, A. J and Hannibal, J. T., 2004, The giant
Arthropleura trackway Diplichnites cuithensis from the Cutler Group
(Upper Pennsylvanian) of New Mexico: Geological Society of America,
Abstracts with Programs, v. 36, no. 5, p. 66.
Jordan, F. W. H. and von Meyer, H., 1854, Ueber die Crustaceen der
Steinkohlenformation von Saarbrücken: Palaeontographica, v.4, p. 1
15.
Kerp, H., 2000, The modernization of landscapes during the late Paleozoic
early Mesozoic: Paleontological Society Papers, v. 6, p. 79-113.
Krainer, K. and Lucas, S. G., 2001, Pennsylvanian-Permian depositional
break in the Cutler Formation, El Cobre Canyon, New Mexico: Geologi
cal Society of America, Abstracts with Programs, v. 33, no. 5, p. A5-A6.
Kraus, O., 2005, On the structure and biology of Arthropleura species
(Atelocerata, Diplopoda; Upper Carboniferous/Lower Permian):
Verhandlungen des naturwissenschaftlichen Vereins Hamburg, Neue Folge,
v. 41, p. 5-23.
Kraus, O. and Brauckmann, C. 2003, Fossil giants and surviving dwarfs.
Arthropleurida and Pselaphognatha (Atelocerata, Diplopoda): Charac
ters, phylogenetic relationships and construction: Verhandlungen des
naturwissenschaftlichen Vereins Hamburg, Neue Folge, v. 40, p. 5-50.
Kraus, W., 1993, Eine 2 m lange semiflexible Rekonstruktion von
Arthropleura, dem größten bekannten Myriapoden aus dem Oberkarbon
Mitteleuropas: Der Präparator, v. 39, p. 141-158.
Langiaux, J., 1984, Flores et faunes des formations supérieures du Stéphanien
de Blanz-Montceau (Massif Central Francais) stratigraphie et
paléoécologie: La Physiophile, suppl. no. 100, 270 p.
Lucas, S. G., 2005, Tetrapod ichnofacies and ichnotaxonomy: Quo vadis?
Ichnos, v. 12, p. 1-6.
Lucas, S. G., 2006, Global Permian tetrapod biostratigraphy and
biochronology; in Lucas, S. G., Cassinis, G. and Schneider, J. W., eds.,
Non-marine Permian biostratigraphy and biochronology. Geological
Society, London, Special Publication 265, p. 65-93.
Lucas, S. G. and Krainer, K., 2005, Cutler Group (Permo-Carboniferous)
stratigraphy, Chama basin, New Mexico: New Mexico Museum of Natu
ral History and Science, Bulletin, v. 31, p. 90-100.
Lucas, S. G. and Lerner, A. J., 2010, Pennsylvanian lake-margin

ichnoassemblage from Cañon del Cobre, Rio Arriba County, New Mexico:
New Mexico Museum of Natural History and Science Bulletin, this vol
ume.
Lucas, S. G., Rinehart, L. F. and Hunt, A. P., 2005a, Eryops (Amphibia,
Temnospondyli) from the Upper Pennsylvanian of El Cobre Canyon,
New Mexico: New Mexico Museum of Natural History and Science,
Bulletin 31, p. 118-120.
Lucas, S. G., Harris, S. K., Spielmann, J. A., Berman, D. S. and Henrici, A. C.,
2005b, Vertebrate biostratigraphy and biochronology of the Pennsylva
nian-Permian Cutler Group, El Cobre Canyon, northern New Mexico:
New Mexico Museum of Natural History and Science, Bulletin 31, p.
128-139.
Lucas, S. G., Lerner, A. J., Hannibal, J. T., Hunt, A. P. and Schneider, J. W.,
2005c, Trackway of a giant Arthropleura from the Upper Pennsylva
nian of El Cobre Canyon, New Mexico: New Mexico Geological Society,
Guidebook 56, p. 279-282.
Lützner, H., 2001, Sedimentologie der Manebach-Formation in den
fossilführenden Aufschlüssen bei Manebach: Beiträge zur Geologie von
Thüringen, N.F., v. 8, p. 67-91.
Lützner, H., Littmann, S., Mädler, J., Romer, R. L. and Schneider, J. W.,
2007, Stratigraphic and radiometric age data for the continental
Permocarboniferous reference-section Thüringer-Wald, Germany:
Procedings XVth International Congress on Carboniferous and Permian
Stratigraphy, 2003, p. 161-174.
Mángano, M. G., Buatois, L. A., West, R. R. and Maples, C. G., 2002,
Ichnology of a Pennsylvanian equatorial tidal flat - the Stull Shale
Member at Waverly, eastern Kansas: Kansas Geological Survey, Bulletin
245, p. 1-133.
Marks, W. J., Marks, R. I. and Pompa, A. M., 1998, Problematic tracks in
the Casselman Formation of Cambria County: Pennsylvania Geology, v.
29, no. 2-3, p. 2-6.
Martino, L. R. and Greb, S. G., 2009, Walking trails of the giant terrestrial
arthropod Arthropleura from the Upper Carboniferous of Kentucky:
Journal of Paleontology, v. 83, p. 140-146.
Menning, M., Aleseev, A. S., Chuvashov, B. I., Favydov, V.I., Devuyst, F.
X., Forke, H. C., Grunt, T. A., Hance, L., Heckel, P. H., Izokh, Jin, Y.
G., Jones, P., Kotlyar, G. V., Kozur, H. W., Nemyrovska, T.I., Schneider,
J. W., Wang, X.-D., Weddige, K., Weyer, D. and Work, D. M., 2006,
Global time scale and regional stratigraphic reference scales of Central
and West Europe, East Europe, Tethys, South China, and North America
as used in the Devonian-Carboniferous-Permian Correlation Chart 2003
(DCP 2003): Palaeogeography, Palaeoclimatology, Palaeoecology, v.
240, p. 318-372.
Montañez, I. P., Tabor, N. J., Niemeier, D., DiMichele, W.A., Frank, T. D.,
Fielding, C. R. and Isbell, J. L., 2007, CO2-forced climate and vegetation
instability during late Paleozoic deglaciations: Science, v. 315, p. 87-91.
Pearson, P. N., 1992, Walking traces of the giant myriapod Arthropleura
from the Strathclyde Group (Lower Carboniferous) of Fife: Scottish
Journal of Geology, v. 28, p. 127-133.
Pfefferkorn, H.W., Gastaldo, R.A., DiMichele, W.A. and Phillips, T.L.,
2008, Pennsylvanian tropical floras of the United States as a record of
changing climate: Geological Society of America, Special Paper 441, p.
305-316.
eho, F. and ehoova, M., 1972, Makrofauna uhlonosného karbonu
eskoslovenské asti hornoslezské pánve: Profil, v. 1972, p. 1-130.
Rolfe, W. D. I., 1969, Arthropleurida; in Moore, R. C., ed., Treatise on
invertebrate paleontology: Geological Society of America, University
of Kansas Press., Part R, Arthropoda, v.4, p. 607-620.
Rolfe, W.D.I., 1980, Early invertebrate terrestrial faunas; in Panchen, A.L.,
ed., The terrestrial environment and the origin of land vertebrates: New
York, Academic Press, p. 117-157.
Rolfe, W. D. I., 1985, Aspects of the Carboniferous terrestrial arthropod
community: Comptes Rendus IX Congrès International de Stratigraphie
et de Géologie de Carbonifère, p. 303-316.
Rolfe, W. D. I. and Ingham, J. K., 1967, Limb structure, affinity and diet of
the Carboniferous “centipede” Arthropleura: Scottish Journal of Geol
ogy, v. 3, p. 118-124.
Roscher, M. and Schneider, J. W., 2006, Early Pennsylvanian to Late
69
Permian climatic development of central Europe in a regional and glo
bal context; in Lucas, S.G., Cassinis, G. and Schneider, J.W., eds., Non
marine Permian chronology and correlation: Geological Society, Lon
don, Special Publication 265, p. 95-136.
Rößler, R. and Barthel, M., 1998, Rotliegend taphocoenoses preservation
favoured by rhyolithic explosive volcanism: Freiberger Forschungshefte,
v. C474, p. 59-101.
Rößler, R. and Schneider, J., 1997, Eine bemerkenswerte Paläobiocoenose
im Unterkarbon Mitteleuropas - Fossilführung und Paläoenvironment
der Hainichen-Subgruppe (Erzgebirge-Becken): Veröffentlichungen des
Museum für Naturkunde Chemnitz, v. 20, p. 5-44.
Ryan, R. J., 1986, Fossil myriapod trails in the Permo-Carboniferous strata
of northern Nova Scotia, Canada: Maritime Sediments and Atlantic
Geology, v. 22, p. 156-161.
Ryan, R. J., Boehner, R. C. and Calder, J. H., 1991, Lithostratigraphic
revision of the Upper Carboniferous to Lower Permian strata in the
Cumberland basin, Nova Scotia and the regional implications for the
Maritime Basin in Atlantic Canada: Bulletin of the Canadian Society of
Petroleum Geology, v. 39, p. 289-314.
Schneider, J., 1994, Environment, biotas and taphonomy of the lacustrine
Niederhäslich Limestone, Döhlen Basin, Germany: Transactions of the
Royal Society of Edinburgh, Earth Science, v. 84, p. 453-464.
Schneider, J., 1996. Biostratigraphie des kontinentalen Oberkarbon und
Perm im Thüringer Wald, SW-Saale-Senke - Stand und Probleme: Beiträge
zur Geologie von Thüringen, N.F., v. 3, p. 121-151.
Schneider, J. and Barthel, M., 1997, Eine Taphocoenose mit Arthropleura
(Arthropoda) aus dem Rotliegend (Unterperm) des Döhlen-Becken (Elbe
Zone, Sachsen): Freiberger Forschungsheft, v. C466, p. 183-223.
Schneider, J. W. and Hoffman, U., 2001, Jungpaläozoikum der Döhlener
Senke; in Alexovsky, W., Schneider, J. W., Tröger, K. A. and Wolf, L.,
eds., Geologische Karte des Freistaates Sachsen 1:25.000, Erläuterungen
zu Blatt 4948 Dresden, p. 15-40.
Schneider, J. W. and Werneburg, R., 1998, Arthropleura und Diplopoda
(Arthropoda) aus dem Unter-Rotliegend (Unter-Perm, Assel) des
Thüringer Waldes (Südwest-Saale-Senke): Veröffentlichungen
Naturhistorisches Museum Schleusingen, v. 13, p. 19-36.
Schneider, J. W. and Werneburg, R., 2006, Insect biostratigraphy of the
European Late Carboniferous and Early Permian; in Lucas, S.G., Cassinis,
G. and Schneider J.W. eds., Non-marine Permian biostratigraphy and
biochronology. Geological Society, London, Special Publication 265, p.
325-336.
Schneider, J. W., Hampe, O. and Soler-Gijon, R., 2000, The Late Carbonif
erous and Permian aquatic vertebrate zonation in Southern Spain and
German basins. IGCP 328. Final Report: Courier Forschungsinstitut
Senckenberg, p. 543-561.
Schneider, J. W., Hoth, K., Gaitzsch, B. G., Berger, H.-J., Steinborn, H.,
Walter, H. and Zeidler, M., 2005, Carboniferous stratigraphy and devel
opment of the Erzgebirge Basin, East Germany: Zeitschrift der Deutschen
Gesellschaft für Geowissenschaften, v. 156, p. 431-466.
Schneider, J. W., Körner, F., Roscher, M. and Kroner, U., 2006, Permian
climate development in the northern peri-Tethys area – the Lodève
basin, French Massif Central, compared in a European and global con
text: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 240, p.
161-183.
Secretan, S., 1980. Les arthropodes du Stéphanien de Montceau-les-Mines:
Bulletin Trimestriel de la Société d’Histoire Naturelle et des Amis du
Muséum d’Autun, v. 94, p. 23-35.
Smith, C. T., Budding, A. J. and Pitrat, C. W., 1961, Geology of the south
eastern part of the Chama Basin: New Mexico Bureau of Mines and
Mineral Resources, Bulletin 75, 57 p.
Van Allen, H. E. K., Calder, J. H. and Hunt, A. P., 2005, The trackway record
of a tetrapod community in a walchian conifer forest from the Permo
Carboniferous of Nova Scotia: New Mexico Museum of Natural History
and Science, Bulletin 30, p. 322-332.
Voigt, S., 2005, Die Tetrapodenichnofauna des kontinentalen Oberkarbon
und Perm im Thüringer Wald - Ichnotaxonomie, Paläoökologie und
Biostratigraphie: Cuvillier Verlag, Göttingen, 305 p.
Walter, H. and Gaitzsch, B., 1988, Beiträge zur Ichnologie limnisch
70
terrestrischer Sedimentationsräume. Teil II: Diplichnites minimus n.
ichnosp. aus dem Permosiles des Flechtinger Höhenzuges: Freiberger
Forschungsheft, v. C427, p. 73-84.
Werneburg, R., 1997, Ein See-Profil aus dem Unter-Rotliegend (Unter
Perm) von Manebach (Thüringer Wald): Veröffentlichungen
Naturhistorisches Museum Schleusingen, v. 12, p. 63-67.
Werneburg, R., 1987, Schädelrest eines sehr großwüchsigen Eryopiden (Am
phibia) aus dem Unterrotliegenden (Unterperm) des Thüringer Waldes:
Veröffentlichungen Naturhistorisches Museum Schleusingen, v. 2, p. 52
56.
Werneburg, R., 2007, Der “Manebacher Saurier” – ein neuer großer Eryopide
(Onchiodon) aus dem Rotliegend (Unter-Perm) des Thüringer Waldes:
Veröffentlichungen Naturhistorisches Museum Schleusingen, v. 22, p. 3
40.
Werneburg, R., Lucas, S. G., Schneider, J. W. and Rinehart, L.F., 2010, First
Pennsylvanian Eryops (Temnospondyli) and its Permian records from
New Mexico: New Mexico Museum of Natural History and Science,
Bulletin, this volume.
Wilson, H. M., 2003, Functional morphology of locomotion in the giant
Paleozoic millipede Arthropleura: Insights from trace fossils and kine
matics of locomotion in extant millipedes: Geological Society of America
Abstracts with Programs, v. 35, no. 6, p. 538.
 Lucas et al., eds., 2010, Carb-Permian transition in Cañon del Cobre. New Mexico Museum of Natural History and Science Bulletin 49.
PALYNOLOGICAL INVESTIGATION OF THE UPPER PENNSYLVANIAN
(CARBONIFEROUS) ELCOBRE CANYONFORMATION, CUTLERGROUP,
CAÑON DEL COBRE, RIO ARRIBA COUNTY, NEWMEXICO, U.S.A.
JOHNUTTING1 AND SPENCER G. LUCAS2
1GeologicalSurvey of Canada, Natural Resources Canada, 3303 33rd St. NW, Calgary, Alberta, T2L 2A7, Canada;
2 New Mexico Museum of Natural History, 1801 Mountain Road NW, Albuquerque, New Mexico 87104-1375, U.S.A.
Abstract—Palynomorphs from the middle part of the El Cobre Canyon Formation in Cañon del Cobre of Rio
Arriba County, New Mexico, come from a 2 m interval of lacustrine shale of Late Pennsylvanian (Carboniferous)
age, not older than Stephanian B. The thermal maturity is low.
INTRODUCTION
The Pennsylvanian (Carboniferous)-Permian Cutler Group red
beds in the Chama basin of northern New Mexico (Fig. 1) are fluvial
strata that were deposited along the flanks of the Uncompaghre uplift,
one of the basement-cored positive areas of the ancestral Rocky Moun
tains. The Cutler Group was deposited in western Pangea in the equato
rial tropics under a relatively wet, but seasonally dry paleoclimate
(Schneider et al., this volume). It yields a diversity of nonmarine fossils,
including vertebrates and plants. DiMichele and Chaney (2005) sug
gested that the latter were typical of wetland floodplains of the Late
Pennsylvanian (Carboniferous). The Cutler Group strata are divided
into two mappable lithostratigraphic units, the El Cobre Canyon and
overlying Arroyo del Agua formations (Lucas and Krainer, 2005). The El
Cobre Canyon Formation is up to 500 m thick, but only 213 m are
exposed in the study area, where they comprise brown siltstone, sand
stone and extraformational conglomerate of an ephemeral braided stream
environment. The conglomerate overlies Proterozoic basement in the
subsurface and is conformably overlain by 143 m of the Arroyo del Agua
Formation.
The aim of the study was to determine the age and thermal matu
rity of the El Cobre Canyon Formation based on palynomorphs col
lected in the Cañon del Cobre of Rio Arriba County, New Mexico.
FIGURE 1. Location map of Cañon del Cobre in Rio Arriba County, New
Mexico.
71
LOCALITIES
In the following text, the abbreviation GSC refers to sample num
bers of the Geological Survey of Canada, Calgary, and NMMNH refers
to sample numbers of the New Mexico Museum of Natural History,
Albuquerque. For the location of the latter and further details on the
paleofloras the reader is referred to Hunt and Lucas (1992) and DiMichele
and Chaney (2005). Four samples were collected for palynomorphs
from lacustrine shale in the middle part of the El Cobre Canyon Forma
tion in Cañon del Cobre.
From lower to upper the samples are: GSC, C-430711 (100 m),
NMMNH locality 7827 at UTM zone 13,377751E, 4016075N (NAD
27): GSC, C-430708 (135 m), NMMNH locality 4563, at UTM zone
13,382564E, 4021118N (NAD27):GSC, C-430709 (136 m) and GSC,C
430710 (137 m), NMMNH locality 7826 at UTM zone 13, 378840E,
4016443N (NAD27). Samples GSC, C-430708, GSC,C-430709, GSC,C
430710 from a 2 m shale bed were productive and contained well-pre
served miospores, whereas a stratigraphically lower sample (GSC, C
430711) contained only woody and coaly fragments. The good preserva
tion of miospores in the productive samples suggests conditions at the
sediment/ water interface were neither severely oxic nor anoxic. The
Thermal Alteration Index of the samples is TAI 2, approximately equiva
lent to a vitrinite reflectance of Ro 0.55%, which is in the “oil window”
(Utting et al., 1989).
IDENTIFICATIONS
Because the three productive samples come from a 2 m shale bed
in the middle part of the El Cobre Canyon Formation, identifications are
given as a single composite list. Selected specimens are illustrated on
Figure 2.
Triletes
Anacanthotriletes paucispinosus Ravn 1986, Apiculatisporis
variocorneus Sullivan 1964, Calamospora parva Guennel 1958,
Colatisporites sp. cf. C. decorus Bharadwaj and Venkatachala comb. nov.
emend. Williams 1973, Convolutispora jugosa Smith and Butterworth
1967, C. varicosa Butterworth and Williams 1958 (abundant),
Cyclogranisporites aureus (Loose) Potonié and Kremp 1955,
Dictyotriletes sp., Lophotriletes commissuralis (Kosanke) Potonié and
Kremp 1955, Microreticulatisporites nobilis (Wicher) Knox 1950,
Pilosisporites williamsii Ravn 1986, Punctatisporites glaber (Naumova)
Playford 1962, Reticulatisporites carnosus (Knox) Neves 1964,
Reticulatisporites reticulatus (Ibrahim) Ibrahim 1933, Savitrisporites nux
(Butterworth and Williams) Sullivan emend. Smith and Butterworth 1967,
and Verrucisporites morulatus (Knox) emend. Smith and Butterworth
1967.
Monoletes
Latosporites globosus (Schemel) Potonié and Kremp 1955.

Saccites
Monosaccites: Florinites pumicosus (Ibrahim) Schopf, Wilson
and Bentall 1944, Florinites similis Kosanke 1950, Florinites sp.,
Potonieisporites bhardwajii Remy and Remy, Potonieisporites elegans
(Wilson and Kosanke) Wilson and Venkatachala 1964, and Potonieisporites
bhardwajii/novicus complex.
Bisaccites (taeniate): Illinites unicus Kosanke emend.
Helby1966, and Protohaploxypinus sp.
Bisaccites (non-taeniate): Pityosporites sp.
Plicates
Schopfipollenites ellipsoides (Ibrahim) Potonié and Kremp 1954.
No evidence was found of any marine palynomorphs such as
acanthomorph acritarchs or scolecodonts.
AGE
Although age determination based on a few productive samples
from a single 2 m thick unit may be hazardous, it is nevertheless possible
to make some general observations. The El Cobre Canyon assemblage
contains common trilete spores, but also contains monosaccate, taeniate
bisaccate and non-taeniate bisaccate pollen. Based on palynological events
in western Europe and North America (Clayton et al., 1977; Owens,
1996) the appearance of taeniate pollen indicates an age no older than
Stephanian B (mid-Missourian). The assemblage is similar to that of the
73
Red Tanks Member of the Bursum Formation of Carrizo Arroyo, New
Mexico, dated as Virgilian by Utting et al. (2004). Both units contain
trilete and monolete spores, monosaccate pollen, taeniate bisaccate and
non-taeniate bisaccate pollen. For example, taxa that occur in both the El
Cobre Canyon Formation and the Red Tanks Member include
Colatisporites sp. cf. C. decorus, Florinites pumicosus, Florinites similis,
Illinites unicus, Schopfipollenites ellipsoides, Potonieisporites elegans
and Potonieisporites bhardwajii/novicus complex. However, apparently
lacking in the El Cobre Canyon assemblage are Alisporites zapfei,
Grandispora sp. A and Grandispora sp. B (illustrated but not described
by Utting et al., 2004), Vesicaspora schaubergeri, Platysaccus saarensi,
Striatoabieites richteri, S. elongatus, Striatopodocarpites sp.,
Limitisporites monstruosus and Vittatina costabilis. Based on the occur
rence of Vittatina costabilis the El Cobre Canyon assemblage is slightly
older than the Red Tanks assemblage dated as Virgilian.
Precise comparison of the El Cobre Canyon assemblage with the
spore zones of Western Europe (Clayton et al. 1977; Owens, 1996) and
the North Sea (McLean et al., 2005) is difficult because many of the taxa
on which these zones are based are apparently not present in the New
Mexico material, although further work may indicate further taxa in
common. However, they may prove to have different vertical distribu
tions. For example, Savitrisporites nux, common in the El Cobre Canyon
Formation, has its range top in the late Bolsovian (Westphalian C = late
Atokan). It is not clear why there is such a discrepancy in the range top
of Savitrisporites nux, although it could well be due to climatic differ
ences.

REFERENCES

Clayton, G., Coquel, R., Doubinger, J., Gueinn, K.J., Loboziak, S., Owens, B.
and Streel, M., 1977, Carboniferous miospores of western Europe: Illus
tration and zonation: Report of Commission Internationale de Microflore
du Paléozoïque/Working Group on Carboniferous Stratigraphical Pa
lynology; Mededelingen Rocks Geologische Dienst, v. 29, 71 p.
DiMichele, W.A. and Chaney, D.C., 2005, Pennsylvanian-Permian fossil
floras from the Cutler Group, Cañon del Cobre and Arroyo del Agua
areas, in northern New Mexico: New Mexico Museum of Natural His-
tory and Science, Bulletin 31, p. 26-33.
Hunt, A.P. and Lucas, S.G., 1992, The paleoflora of the lower Cutler Forma
tion (Pennsylvanian,Desmoinesian?) in El Cobre Canyon, New Mexico,
and its biochronological significance: New Mexico Geological Society,
Guidebook 43, p. 145-150.
Lucas, S. G. and Krainer, K., 2005, Stratigraphy and correlation of the
Permo-Carboniferous Cutler Group, Chama basin, New Mexico: New
Mexico Geological Society, Guidebook 56, p. 145-159.
McLean, D, Owens, B. and Neves, R., 2005, Carboniferous miospore bios
tratigraphy of the North Sea; in Collinson, J. D., Evans, D. J. and
Holliday, N. S., eds., Occasional Publications Series of the Yorkshire
Geological Society, v. 7, p. 13-24.
Owens, B., 1996, Chapter 18D. Upper Carboniferous spores and pollen; in
Jansonius, J. and McGregor, D.C., eds., Palynology principles and appli
cations: American Association of Stratigraphic Palynologists Founda
tion, v. 2, p. 597-606.
Smith, A. H. V. and Butterworth, M. A., 1967, Miospores in the coal seams
of the Carboniferous of Great Britain: Palaeontology, Special Paper 1,
324 p.
Utting, J., Hartkopf-Fröder, C., Lucas, S. G. and Traverse, A., 2004, Pa
lynological investigation of the Upper Pennsylvanian Red Tanks Mem
ber, Bursum Formation, Carrizo Arroyo, New Mexico, U.S.A: New Mexico
Museum of Natural History and Science, Bulletin 25, p. 89-96.
Utting, J., Goodarzi, F., Dougherty, B. J. and Henderson, C. M., 1989,
Thermal maturity of Carboniferous and Permian rocks of the Sverdrup
Basin, Canadian Arctic Archipelago: Geological Survey of Canada, Paper
89-19, 20 p.

FIGURE 2. Selected palynomorphs from the El Cobre Canyon Formation, Cañon del Cobre, New Mexico. The magnification of all illustrated specimens
is x 500, unless otherwise stated. In the explanation of figures, the species name is followed by the GSC locality number (preceded by the letter C), the slide
number, stage co-ordinates (in mm), and the GSC type number. All specimens are in the collections of the Geological Survey of Canada, 601 Booth St.,
Ottawa, Ontario, Canada. The stage co-ordinates (in mm) are from a Sensor Control Display digital readout with the zero point of both X and Y axes set
at the upper right hand corner of a 76 x 25 mm glass microscope slide (label on left hand side). 1. Anacanthotriletes paucispinosus C-430710,3,
3,41.1x13.5,GSC 133030. 2. Apiculatisporis variocorneus C-430710,3,-38.1x15.8, GSC 133031. 3. Convolutispora jugosa C-430710,3,-3,27.7x18.9,
GSC 133032. 4. Convolutispora varicosa C-430710,3,-20.7x6.9, GSC 133033. 5. Cyclogranisporites aureus C-430710,3,-30.0x9.2, GSC 133034. 6.
Lophotriletes commissuralis C-430708,3, -39.4x8.4, GSC 133035.7. Microreticulatisporites nobilis C-439710,3,-38.3x8.4, GSC 133036. 8. Pilosisporites
williamsii C-439708,3,-15.1x7.3, GSC 133037.9. Punctatisporites glaber C-430708,3,-17x5.5, GSC133038. 10. Pustulatisporites pustulatus C-430710,3,
28.7x9.3, GSC 133039.11. Raistrickia aculeata C-430708,3, -28.7x7.2, GSC 133040.12. Reticulatisporites carnosus C-430710,3, -23.2x9.25, GSC
133041. 13. Reticulatisporites reticulatus C-430710,4,-18.3x15.9,GSC 133042. 14. Dictyotriletes sp. C-430708,4,-25.9x4.7 (x250), GSC 133043. 15.
Savitrisporites nux C-430708,3,-24.0x7.09,GSC 133044. 16.Verrucosisporites morulatus C-430708,4,-17.9x7.9, GSC 133045. 17. Schopfipollenites
ellipsoides C-43010,4,-40.4x12.1,(x250),GSC 133046. 18. Latosporites globosus C-430708,3,-29.2x7.4, GSC 133047. 19. Florinites pumicosus C
430708,(4)-24.9x4.6,(x 250), GSC 133048. 20. Florinites similis C-430710,4,-19.7x15.1,(x250), GSC 133049. 21. Florinites sp. C-430710,4,-15.7x14.7,
(x250) GSC 133050. 22. Potonieisporites elegans C-430710,3,-13.9x18.6, GSC 133051. 23. Pityosporites sp. C-430710, 3,-14.7x18.5, GSC 133052. 24.
Pityosporites sp. C-430710,4,-36.5x13.7,(x250), GSC 133053. 25. Platysaccus sp. C-430708,3,-44.6x6.6, GSC 133054. 26. Protohaploxypinus sp. C
430710,3,-12.9x4.8,(x250), GSC 133055. 27. Illinites unicus C-430710, 3,-13.2x6.39, GSC 133056.
74
 Lucas et al., eds., 2010, Carb-Permian transition in Cañon del Cobre. New Mexico Museum of Natural History and Science Bulletin 49.

75
LATE PENNSYLVANIANFLORASIN WESTERN
EQUATORIALPANGEA, CAÑON DELCOBRE, NEWMEXICO

WILLIAMA. DIMICHELE1, DANS. CHANEY1, HANS KERP2 AND SPENCERG. LUCAS3

1 Department of Paleobiology, NMNH Smithsonian Institution, Washington, DC 20560;

2 Forschungsstelle für Paläobotanik, Institut für Geologie und Paläontologie, Westfälische Wilhelms-Universität, Münster, GERMANY;
3 New Mexico Museum of Natural History and Science, 1801 Mountain Road NW, Albuquerque, NM 87104

Abstract—Fossil plants were collected from five stratigraphic levels in Cañon del Cobre, New Mexico. These
floras are compositionally typical of the Late Pennsylvanian, particularly of the later part, the Virgilian (Gzhelian).
Conspicuous elements are Alethopteris zeilleri, Macroneuropteris scheuchzeri, Danaeites emersonii, Sphenophyllum
verticillatum, Sphenophyllum oblongifolium, and Sphenophyllum angustifolium, Annularia carinata, Asterophyllites
equisetiformis, Sigillaria brardii, and other, less common forms typical of wetland floras, such as Pseudomariopteris
cordato-ovata. Walchian conifers and Taeniopteris were encountered rarely. For the most part, these floras appear
to have been growing on bars or along the margins of braided streams, resulting in frequent burial of plants,
particularly calamite stems, in upright position, or preservation of tangled masses of sphenophyll stems and
leaves, largely in siltstones. Finer-grained deposits, seemingly representing small lakes, were strongly dominated
by pteridosperms.

Paleozoic equatorial region in North America and Europe (Euramerica).

INTRODUCTION
The Pennsylvanian-Permian transition in the Pangean low lati
tudes records a general change from wetland floras dominated by primi
tive seed plants, marattialean tree ferns, and lycopsids, to floras of sea
sonally dry environments, dominated by derived seed plants, including
conifers, callipterids, and other groups not represented or poorly repre
sented in the wetlands (Potonié, 1893; Kerp and Fichter, 1985). These
two floras can be considered, broadly, to be distinct biomes (DiMichele
et al., 2005). These distinctions reflect different environmental condi
tions in the lowlands caused by differences in ambient climate (Montañez
et al., 2007; DiMichele et al., 2008, 2009a). Oscillations between the two
biomes began during the Middle Pennsylvanian (e.g., Falcon-Lang et al.,
2009), with both wetland and seasonally dry floras becoming common,
at different times, in basinal lowlands during the Late Pennsylvanian
(DiMichele and Aronson, 1992). The seasonally dry flora came to pre
dominate in the Permian equatorial lowlands.
The Pennsylvanian Subperiod has long been recognized as a time
of major ice accumulations in at least the southern polar regions, resulting
in a pattern of regular glacial-interglacial oscillations (Wanless and Weller,
1932; Wanless and Shepard, 1936). Recent work suggests that ice vol
ume declined significantly during the late Middle Pennsylvanian (Field
ing et al., 2008). The Late Pennsylvanian appears to have been a period
of tropical warmth, especially the early part when derived, xeromorphic
floras first become common in lowlands (Cridland and Morris, 1963;
Feldman et al., 2005). The tropical regions appear to have become gener
ally wetter in the later Late Pennsylvanian, at least in central and western
Pangaea. Ice again became prominentator just before the Pennsylvanian
Permian transition (Fielding et al., 2008; Rygel et al., 2008). These changes
in ice volume, possibly attributable to changes in atmospheric CO2
(Montañez et al., 2007), may have been major drivers of the climate
changes that affected terrestrial edaphic conditions and, thence, vegeta
tional distributions during the Pennsylvanian-Permian transition.
Because of the distinct environmental preferences of the two prin
cipal floras/biomes, the change from one to the other is not locally or
regionally useful as a biostratigraphic marker. Temporal internal changes
within each biome may record useful evolutionary trends. The periodic
(or episodic) cycling between these biomes records areal and temporal
changes in environmental conditions that favor one of these biomes over
the other.
Basins preserving this transition are strung out along the late
In most of them, the exact Carboniferous-Permian systemic boundary
cannot be located precisely due to the lack of marine beds that provide
the indices needed to correlate any particular section to the global stan
dards and with equivalent sections in the other basins. However, each
basin records the same basic pattern of transition from dominance of the
wetland flora to dominance, or at least commonness, of the flora typical
of subhumid to semi-arid climates. From west-to-east, the transition
generally is marked by increasingly longer persistence of wetlands, marked
by coal beds and the relative proportions of the two biomes.
In New Mexico, this transition includes the typical change from
predominance of wetland floras to more xeromorphic floras typical of
seasonal drought. There are, however, no indicators of persistent wet
climatic intervals, such as coals (lacking), and a variety of geological
indicators suggest that periods of low moisture availablility were the
most common conditions.
This paper focuses on floras from Cañon del Cobre, in north
central New Mexico (Fig. 1). It details the composition of these floras,
their environmental and paleoecological significance, and their biostrati
graphic implications. Plant fossils have been known from Cañon del
Cobre since the late 1800s (Fontaine, 1890), and have been recognized
explicitly as of Pennsylvanian age since the 1960s (see Hunt and Lucas,
1992). For the most part, the plants have served as somewhat ambiguous
biostratigraphic markers, in some conflict with vertebrate age estimates
(Hunt and Lucas, 1992), suggesting ages from late Middle Pennsylvanian
(Desmoinesian), through earliest Permian. The purported Middle Penn
sylvanian age has been reassessed (DiMichele and Chaney, 2005) and
will be reconsidered here in more detail.
GEOLOGY
The upper Paleozoic rocks of Cañon del Cobre encompass the
Upper Pennsylvanian to Lower Permian El Cobre Canyon Formation
and the Lower Permian Arroyo del Agua Formation, together making up
the Cutler Group (Lucas and Krainer, 2005). Eberth and Miall (1991)
described the El Cobre Canyon Formation as consisting of sediments
deposited in shallow, ephemeral braided streams. Similarly, Lucas et al.
(2005) described these rocks as having formed from a “low sinuosity
river system with locally vegetated interfluves and floodplain surfaces.”
They described trace fossils, including Scoyenia burrows and trackways
of arthropleurid arthropods, the latter consistent with open vegetation
on river floodplains. Also common are ferric soils, immature calicisols,
76

FIGURE 1. Index map and stratigraphic column. United States with the state of New Mexico shaded. Cañon del Cobre area enlarged to show spatial and
elevational relationship between various plant localities labeled A-E, oldest to youngest. Stratigraphic column with stratigraphic position of plant localities
(names to left of rock column and associated locality numbers, both USNM and NMMNH, to the right).

and calcareous rhizoconcretions in paleosols, in combination indicative
of strongly seasonal climates with periodic moisture limitation (Lucas et
al., 2005).
Our own observations suggest that, in the El Cobre Canyon For
mation, paleosols commonly have petrocalcic concretions, and that the
calcic content increases, on average, upward through the section. In some
places, the carbonate nodules are intergrown into thick beds of nodular
pedogenic carbonate, several decimeters thick. These observations sug
gest that climate was strongly seasonal with evapotranspiration exceed
ing rainfall for much of the year, at least during the times of paleosol
formation. The climatic signatures of these paleosols often contrasts
strongly with that inferred from the megafossil flora preserved in chan
nel-fill deposits cut into them. Such contrasts suggest longer-term pat
terns of climatic oscillation, with plant preservation occurring differen
tially during the wetter periods.
AGE OF THE EL COBRE CANYON FORMATION
Biochronological age estimates of the El Cobre Canyon Formation
range from late Middle Pennsylvanian (Desmoinesian) through Early
Permian (Wolfcampian) depending on the fossil group. Palynology
(Utting, 2001), based on limited sampling, provides a very broad range of
possible ages, ranging from late Middle to Late Pennsylvanian. How
ever, as the author notes, the limited numbers of samples require that
these estimates be treated with caution. Plants are quite sensitive to
facies-environmental variables, so attribution of age based on floristic
data can be misleading if there are strong environmental signals in the
rock sequence from which the samples were drawn. Often, environment
will be the major controlling variable, and can greatly confound interpre
tations of age (see, for example, the debate of the age of the Dunkard
Formation of the Appalachian region: Barlow, 1975).
Macrofossil plant remains from the El Cobre Canyon Formation
have been interpreted to indicate an age as old as late Desmoinesian (late
Middle Pennsylvanian) and as young as Virgilian (late Late Pennsylva
nian) (Hunt and Lucas, 1992, and reports of other investigators noted
therein). The most recent analysis (DiMichele and Chaney, 2005) esti
mated the age to be late Late Pennsylvanian, probably later Virgilian. The
flora is broadly consistent with European (e.g., Wagner, 1985; Doubinger
et al., 1995; Laveine, 1989) and American (Blake et al., 2002; DiMichele

et al., 2004, 2005) floras of this age in containing such long-ranging,
mainly Late Pennsylvanian species as Alethopteris zeilleri,
Pseudomariopteris cordato-ovata, Sphenophyllum oblongifolium,
Annularia carinata, and Sigillaria brardii. It also contains more typi
cally later Late Pennsylvanian species, such as, Sphenophyllum
angustifolium, Sphenophyllum verticillatum, and Odontopteris lingulata
(Wagner, 1984; Castro, 2005). As such, the flora falls broadly into
Wagner’s (1984) Sphenophyllum angustifolium or Autunia conferta
biozones, which correlate with the Stephanian C. However, it should be
understood that the stratigraphy of Late Pennsylvanian terrestrial de
posits is in the process of revision at this time, particularly as delimited
by plant fossils (Doubinger et al., 1995; Wagner, 1998). Given the rela
tively small, and depauperate flora described from Cañon del Cobre, a
well-constrained age determination cannot be made with confidence, al
though the flora, to the degree presently known, can be placed in the Late
Pennsylvanian, probably the Virgilian in American terminology (Blake et
al., 2002), likely equivalent to the Gzehlian and within the Stephanian B
or C in international terminology (Wagner, 1984, 1988; Heckel and
Clayton, 2006).
The vertebrate chronology of the Cutler Group is discussed by
Lucas et al. (2005). They note the occurrence of Desmatodon, Limnoscelis
and a primitive Eryops in the lower El Cobre Canyon Formation, which
is consistent with a latest Pennsylvanian to earliest Permian age, and in
conformance with earlier interpretations of the El Cobre Canyon Forma
tion vertebrates as of latest Pennsylvanian age (Fracasso, 1980; Lucas,
2002). Vertebrates from the Arroyo del Agua Formation, better known in
nearby Arroyo del Agua, indicate an unambiguous Early Permian age
(Langston, 1953; Lucas, 2002: Lucas and Krainer, 2005; Lucas et al.,
2005).
MATERIALS AND METHODS
Fossil plants were collected from 21 sites representing five strati
graphic horizons from the El Cobre Canyon Formation exposed within
Cañon del Cobre (Figs. 1-3). Collections were made in as unbiased a
manner as possible, meaning that all hand samples that showed any
indications of fossil plant debris were collected. The collection sizes
from any individual excavation varied from more than 100 specimens to
only a few, with most sites yielding between 25 and 95 hand samples.
All collections are housed in the Paleobotanical Collections of the
National Museum of Natural History, Smithsonian Institution in Wash
ington, DC. Each collecting locality has been given a unique locality
number. All photographed specimens have been given unique collection
numbers.
Specimens were identified using a hand-lens and microscope. Cu
ticles were not prepared. During the process of identification, each col
lection was quantified to establish the relative abundances of the compo
nent species. This was carried out using the quadrat method of Pfefferkorn
et al. (1975), in which each hand specimen is treated as a sampling
quadrat, and taxa are identified as having occurred once if present on a
quadrat/hand-sample, regardless of how large the remains or how numer
ous (e.g., one pinnule counts the same as more than one, if all occur on the
same quadrat). Both sides of each specimen were considered. Counter
parts were not counted twice. This kind of analysis produces a fre
quency distribution in which the relative abundance of a taxon is re
ported as a percentage of the sample quadrats on which it was identified,
i.e., all taxa could, hypothetically, occur at 100% abundance. This method
tends to reduce the relative abundance of the dominant forms and raise
the abundance of the rare forms, compared to point-count methods or
counts of individual plant parts (Wing and DiMichele, 1995). This hand
sample/quadrat method was chosen for several reasons. It can be applied
easily to museum collections (the basis of this study), assuming those
collections were made in an unbiased manner (meaning that an attempt
was made while collecting to preserve the proportions of specimens as
they were found in the excavation, not concentrating “better” material or
77
particular taxa). A quantification can be accomplished rapidly, even with
large collections. It overcomes problems of “what to count,” which is
especially problematic in Pennsylvanian collections, where some taxa
are known mainly as stems of various sizes, others as large leaves, others
mainly as isolated pinnules of larger fronds; if sample size is not too
large, thus the “hand sample” prescription, the potential bias of counting
one large stem or leaf fragment as equal to one tiny pinnule is minimized,
due both to the large number of samples and to the relatively small size
of a “quadrat.” In addition, empirical studies show that this method
preserves the basic dominance-diversity structure of the assemblage
(Lamboy and Lesnikowska, 1988; DiMichele et al., 1991).
Specimens were photographed digitally, mostly under natural light.
All photographs were processed in Adobe Photoshop.
FLORA
Lycopsids
Lycopsid stem remains identified as Sigillaria brardii (Fig. 4)
consist of leaf scars with rounded upper and lower edges, sharply pointed
lateral angles, and distinct foliar parichnos below a small leaf trace. Ligule
pits are not well marked. The scars are not part of well-developed leaf
cushions. Scars are separated by interareas with vertically oriented lines
that weakly converged around the scar areas.
Sigillaria remains were rare in the fossil assemblages; only two
stem specimens were found. These remains were associated with cones
identified as Sigillariostrobus and ribbon-like axes with central thicken
ings (midveins?) that are interpreted either as leaves or rootlets of sigillarian
stigmaria (Fig. 5). We lean toward the identification of these structures as
leaves, given the overall parautochthonous nature of this flora, the sedi
mentary environment (active stream channels or lake deposits) and their
relative abundance.
The presence of S. brardii is an indication of the persistence of
wetter areas within these channel settings, though sigillarians, or at least
some of them, appear to have been tolerant of somewhat drier conditions
than those typical of lepidodendrids (Phillips and DiMichele, 1992;
Pfefferkorn and Wang, 2009).
Calamitales
The calamitalean sphenopsids are represented by stem, foliage
and reproductive remains (Figs. 6-10). Stem remains consist mainly of
Calamites sp., consisting of typically ribbed stems lacking branch scars
at the nodes (Figs. 6, 7; see Fig. 6.4). These kinds of specimens generally
are described as “pith casts,” though we have challenged this conven
tional wisdom (DiMichele et al., 2009b), arguing instead that many ca
lamite stems preserved in compression preservation are, in fact, stem
casts or compressed stems, comparable to modern Equisetum. Following
Rössler and Noll (2006), we see the Calamitales as a much more diverse
group structurally and taxonomically than has been appreciated, at least
in recent work.
Calamite foliage is attributable to Annularia carinata (Kerp, 1984;
Kerp and Fichter, 1985) (Figs. 8, 9.2), possibly Annularia spicata (Fig.
9.1) and Asterophyllites equisetiformis.
Annularia carinata is characterized by leaves of unequal length,
forming ovoid whorls of varying diameter. The individual leaves are
spatulate in form and have mucronate tips (these specimens may also be
referred to Annularia mucronata; the taxonomy of Annularia is in need
of revision). Populations of this species frequently are confused for
mixed assemblages of Annularia sphenophylloides (smaller whorls) and
Annularia stellata (larger whorls). Each of these latter species has sym
metrical whorls formed by leaves of approximately equal length. Leaves
of A. sphenophylloides are spatulate and mucronate. Large mucronate,
spatulate leaves often have been attributed to A. stellata, but this appears
to be incorrect given that the type of A. stellata has lanceolate leaves.
Three forms of calamite cones occur in the El Cobre assemblage;
78

FIGURE 2. Field photos. 1, Copper mine locality with arrows indicating the east-west extent of the outcrop. 2, Closeup of the finely-laminated sediment
of USNM 42293 at the Copper Mine locality from which sparse plant remains were collected. 3, Arroyo wall with the myriapod trackway-producing
sandstone bed low along right side of wash. Arrow indicates finely laminated red micaceous beds from which fossil plants (USNM 42125 and 42126) were
recovered. 4, Closeup of plant-bearing beds above trackway-producing sandstone. 5, Large sandstone ridge into which “uranium” adits were dug. The right
arrow indicates position of the north adit (USNM 42296); left arrow indicates location of south adit (USNM 42297 and 422598). 6, Thick sandstone ledge
at the south adit; there is a small clay lens from which USNM 42297 was collected. Note quarter sheet of newspaper to the left of lense for scale.

FIGURE 3. Field photo. Copper mine outcrop, view from east to west. The
finely laminated beds, seen in Figure 2.2, are at the distant end of the out
crop in the image. In the near ground are specimens, from the sandier more
fluvial facies, laid out on a fallen sandstone block.
all are rare. The three include Palaeostachya sp. (Fig. 10), which is
locally common, Calamostachys sp., which is rare, and Macrostachya
sp. (Figs. 9.3, 9.4), represented by two specimens.
Sphenophylls
Three species of Sphenophyllum were identified in the collections
from Cañon del Cobre plus one additional potentially distinct form.
Sphenophyllum angustifolium (Figs. 11.1, 11.2, 11.4, 11.5, 11.7)
consists of whorls of six narrow leaves, each with four to six veins,
terminating in long pointed teeth. Two veins enter the base of a leaf and
fork near the base, rarely again prior to termination. The edges of the
leaves also can be deeply lacerate. Two pairs of leaves project at sub-90o
angles to the stem, and two additional leaves, one on each side, project at
nearly a 90o angle to the stem, or are slightly deflected toward the base of
the stem.
Sphenophyllum verticillatum (Figs. 11.3, 11.6, 12.3, 12.4, 12.5)
leaf whorls are nearly round, containing six leaves. The leaves are wedge
shaped and broaden greatly from a narrow base. They have arcuate distal
margins, with rounded, but mucronate, teeth. Generally, two veins enter
the base of the leaf and fork two to three times before terminating in a
terminal tooth. Because of the rounded distal leaf margin, the veins termi
nate at slightly different distances from the base. Along the margin of
larger leaves there can be as many as 20 veins. R.H. Wagner suggests
comparing the specimens illustrated in Figures 11.3 and 11.6 with
Parasphenophyllum Asama (Asama, 1970), in that the teeth are some
what elongate and bluntly pointed and the distal margins are not clearly
set off from the lateral margins.
Sphenophyllum oblongifolium (Figs. 12.1, 12.6, 13.1, 13.2, 13.4,
13.5) is characterized by whorls consisting of six leaves. Two opposite
pairs of leaves project at sub-90o angles to the stem. Two additional
leaves form a basal bib, directed roughly parallel to the stem. Leaves have
eight to ten subparallel veins, forking generally one to two times before
terminating in small, pointed teeth at the leaf margin. The distal margins
of the leaves can be lacerate, in addition to toothed.
A single isolated specimen is herein identified as Sphenophyllum
(Parasphenophyllum) cf. thonii (Fig. 13.3). The small leaf is approxi
mately 5 mm wide and has a distinctly rounded distal margin with small,
shallow, blunt teeth. Two veins enter the base and may branch as often as
five to six times before terminating at the margin. Veins terminate all
79
along the lateral margins and curved distal margin of the leaf, so some
traverse a relatively short, arched path. Along the leaf margin, the vein
density appears to be about 35.
Pteridosperms
The most commonly encountered pteridosperm foliar remains are
Alethopteris zeilleri and Macroneuropteris scheuchzeri. All other pteri
dosperm foliage is rare, and includes Odontopteris cf. lingulata,
Odontopteris cf. brardii, Pseudomariopteris cordato-ovata, and isolated
pinnules, possibly of neuropteroid affinity.
A few specimens in the collection were identified tentatively as
Odontopteris cf. lingulata (Figs. 14.1, 14.4 & 14.5, perhaps Figures 14.1,
14.2 and along the edge of Fig. 25.5) based on pinnule shape and vena
tion. Odontopteris lingulata has been designated Mixoneuralingulata by
Wagner and Castro (1998). The specimens from Cañon del Cobre are
quite similar to Macroneuropteris scheuchzeri, and some of the ones
shown here may be. They are large, tongue-shaped with undulatory
sides and a bluntly rounded to broadly pointed apex. The midvein is well
developed, but appears to consist of a bundle of veins rather than a single
vein; it extends about two-thirds the length of the pinnule before splay
ing broadly into secondary veins. The secondaries are relatively widely
spaced and strong, arching steeply before curving to meet the pinnule
margin at about 45o. Lateral veins branch three or more times before
reaching the pinnule margin. Like M. scheuchzeri, O. lingulata may have
smaller, rounded pinnules positioned basipetally to the large, linguloid
forms. These smaller pinnules vary in number from one to many, gener
ally lack midveins, and have a rounded shape. We found only one pos
sible example of this configuration.
Pinnae bearing small, triangular pinnules with broad attachment to
the rachis, multiple veins enter the base of the pinnule, and no obvious
midvein were encounted rarely and always in fragmentary preservation
(Figs. 14.2, 14.6). The identification of such pinnae is difficult and can
not be made with confidence. The specimens in the collection resemble
taxa such as Lescuropteris genuina or Odontopteris brardii.
Alethopteris zeilleri (Figs. 15-17) pinnules are somewhat variable
in shape, degree of confluence with adjacent pinnules, and the degree of
angularity with which the pinnules join the pinna rachis. Generally
straight-sided with bluntly rounded apices, some pinnules may be en
larged in the middle and taper gradually to a broadly triangular apex.
Pinnules may be nearly free to the base and only weakly confluent with
adjacent pinnules, or the confluence may be well developed. The pinnule
surface is strongly vaulted, leading to a sunken midvein, which extends
about two-thirds of the distance from the base to the pinnule apex.
Venation is dense, with veins forking first near the midvein of the pinnule
and generally once more before terminating at the margin, which they
meet nearly at right angles. Pinnule shape and size can be variable within
a pinna. Pinnules decrease in size, become more triangular in shape and
increasingly confluent toward the terminus of a pinna, which ends in an
elongate, narrow, terminal pinnule. Note, R.H. Wagner suggests that
these specimens bear comparison with Alethopteris virginiana Fontaine
and White.
Macroneuropteris scheuchzeri (Figs. 19, 20, possibly Figs. 14.1
and 14.3) pinnules are generally found in isolation. They are straight
sided to slightly falcate, with a thin lamina and little vaulting. The pin
nule base is shallowly cordate, with a single point of attachment to the
pinna rachis. Large pinnules may have two lobes or free smaller, round
pinnules at their bases. These smaller pinnules also may be found in
isolation. The pinnule midvein is strong and extends about 80% of the
distance from the pinnule base to the tip. Lateral veins are fine, dense and
broadly arched, reaching the pinnule margin at an angle of about 30o.
Lateral veins may fork three or more times before reaching the margin. In
many pinnules, small, stiff, apically directed hairs are present, most
commonly on the abaxial surface, and mostly along the midvein or in the
basal third of the pinna; such small, adpressed hairs are diagnostic of this
80
FIGURE 4. Sigillaria brardii. USNM 528655 from loc 42122; 1a, part; 1b, counterpart. Both 2 x.
species, in combination with its shape, size, and venation.
Pseudomariopteris cordato-ovata (Fig. 21, possibly Fig. 33.2)
occurs rarely in the collection. Specimens consist of isolated pinnae
bearing small, rounded to triangular pinnules with acuminate apices, and
constricted, but still broadly attached bases. The pinnules are distinctly
inclined toward the pinna apex. Some pinnules may have basal lobes.
Venation is obscure, probably due to a thick pinnule lamina, which ap
pears to be slightly vaulted, with a thickened or inrolled margin. A midvein
is present, mainly in the lower part of the pinnule. Secondary veins are
widely spaced and thin, arching steeply through the pinnule lamina.
Importantly, the terminal portion of the pinnule ends in a spine-like
prolongation, characteristic of plants in this genus, which is thought be a
vine (Krings and Kerp, 2000; Krings et al., 2003).
A single, isolated pinnule was encounted that may be assignable to
Neurocallipteris cf. planchardii (Fig. 19.3) (see Cleal et al., 1990). The
specimen is auriculate, otherwise straight sided, with a broad, rounded
apex, wide lamina relative to length, and a relatively flat base. The midvein
is thin and extends through two-thirds of the length of the lamina. Lateral
veins are fine, dense, and arch steeply through the lamina, meeting the
margin at an angle of 30o-45o. Clearly, a positive identification is not
possible.
Small isolated pinnules of ambiguous, but likely pteridospermous
affinity were found. These pinnules have a neuropteroid shape, with a
broadly rounded apex, straight to rounded sides, and a tapering ovate
base (Fig. 18). The venation, however, is not neuropteroid, but more
alethopteroid in character. The midvein is straight and extends through
about three-quarters of the lamina, becoming more narrow apically
throughout its length. The dense lateral veins arch upward from the
midvein but shortly become subhorizontal, forking two to three times
before reaching the margin at a high angle, nearly perpendicularly in some
specimens.
Axes most likely attributable to pteridosperms were found at
most collecting localities. The axes varied from striate to striate and
punctuate (Fig. 33.3) to variously smooth. Others were ribbon like and
resembled stigmarian rootlets (Figs. 33.1, 33.4, 33.5), though their affin
ity could not be established. Seeds of indeterminate affinity were com
mon at certain sites. These large, platyspermic, ovoid to cordate seeds
(Figs. 34.1-34.7) are of indeterminate affinity; they may be cordaitalean
or attributable to one of the pteridosperms, though the latter option
seems unlikely. Only a single ribbed, likely pteridosperm seed was found
(Fig. 34.8)
Other Seed Plants
Cordaitalean leaves (Fig. 22) are characterized by broad, flatlami
nae, 2-3 cm wide, that have fine, parallel veins that branch occasionally.
The lamina enlarges gradually from the base, is straight-sided throughout
most of its length, and tapers toward a bluntly pointed apex.
Fragmentary Taeniopteris, similar to T. abnormis (T. multinervis
has been placed in synonymy with T. abnormis by Barthel, 1976; see
also Wagner and Martínez Garcia, 1982), were found rarely (Fig. 23).
 81

FIGURE 5. Lycopsid leaves and cones. 1, Lycopsid leaves, USNM 539039 from loc 42123 x 2. 2, Lycopsid leaf, USNM 539031 from loc 42121 x 1. 3,
Lycopsid cone, USNM 539038 from loc 42122, x 2.

These leaves have wide laminae, nearly 4 cm. Midveins also are wide,
nearly 0.5 cm. Lateral veins depart from the midvein at right angles and
proceed without inflection to the leaf margin. Lateral veins branch imme
diately adjacent to the midvein and may branch one or more times before
reaching the margin.
Walchian conifers (Fig. 24) were preserved only as small frag
ments. Most of these specimens have leaves that appear, probably due
to the nature of their exposure in the rock matrix, as narrow and needle
like, with an upward curving aspect, not adpressed to the stem. How
ever, when exposed by breaks paralleling the outer (abaxial) surfaces,
along the edge of a shoot, these leaves are revealed as 3-4 mm long and
1.0-1.5 mm wide with acuminate apices. They resemble Walchia
pinniformis. A single conifer cone scale was found (Fig. 24.2). Preserva
tion was poor but the scale appears to be multilobed, consistent with a
walchian affinity.
Ferns
Sterile and fertile foliage attributable to Danaeites emersonii is a
common element of the collections (Figs. 25, 26). Sterile foliage is similar
to Polymorphopteris (Acitheca) polymorpha (Zodrow et al., 2006). Pin
nae bear large subopposite pinnules with a constricted but broadly at
tached base, sides that can be slightly convex or acropetally convex and
acropetally weakly concave, with a bluntly pointed tip and vaulted
laminae. Terminal pinnae are small, and rounded, with the pinnules get
ting rapidly smaller basipetally; those adjacent to the terminal pinnule
may have odontopterid venation, with several veins entering the base
without a well developed midvein. The midvein of more proximal pin
nules is sunken and extends two-thirds or less of the distance from the
base to the tip of the pinnule. Venation is “polymorphopterid” (sensu
Wagner 1958b), meaning that lateral veins initially arch steeply upward
and branch close to their point of insertion into the midvein; each deriva
tive vein then tracks at increasing inflexion from the midvein and branches
again, within one-third of the distance to the leaf margin, the final branched
segments extending the long distance to the leaf margin, which they
contact at a high angle of 60o or more, being nearly horizontal with the
rachis in the lower, basipetal portions of a pinnule. Veins can be dense.

FIGURE 6 (Next page). Calamites sp. 1, Calamites sp., 3D cast, 1a and 1b, different views of same specimen, USNM 538814 from loc 42128 x 0.5. 2,
Calamites sp., 3D cast, USNM 528487 from loc 42128, x 1.3, Calamites sp., 3D compacted cast, USNM 536769 from 42994 x 1. 4, Calamites sp., branch
scars, USNM 539022 from loc 42120 x 1.5, Calamites sp., USNM 539023 from loc 42120 x 1.
FIGURE 7 (Following page). Calamites sp. 1, Calamites sp., compression, USNM 536777 from loc 42995 x 0.5. 2, Calamites sp., USNM 536771 from
loc 42294 x 0.5. 3, Calamites sp., USNM 536775 from loc 40621 x 2.4, Calamites sp., USNM 538803 from loc 42128 x 2.5, Calamites sp., ribs not
visible, USNM 536772 from loc 42994 x 1.
FIGURE 8 (Following page). Annularia carinata. 1, Annularia carinata, USNM 536801 from loc 42293 x 2. 2, Annularia carinata, USNM 536770 from
loc 42293 x 2. 3, Annularia carinata, USNM 536768 from loc 42294 x 2. 4, Annularia carinata, USNM 536770 from loc 42294 x 2.
 85

FIGURE 9. Annularia cf. spicata, Annularia carinata and Macrostachya sp. 1, Annularia cf. spicata, USNM 528479 from loc 42128 x 4. 2, Annularia
carinata, USNM 528642 from loc 41878 x 2. 3, Macrostachya sp., USNM 539021 from loc 42120 x 1. 4, Macrostachya sp., USNM 536810 from loc
42298 x 2.
86

FIGURE 10. Palaeostachya sp. 1, Palaeostachya sp., USNM 536778 from loc 42295; shaded polygon in 1a indicates area of enlargement shown in 1b;
1a x 1, 1b x 2. 2, Palaeostsachya sp., USNM 536791 from loc 42295 x 2. 3, Palaeostsachya sp., USNM 536787a from loc 42295 X2. 4, Palaeostachya
sp., USNM 536767 from loc 42994 x 2.5, Palaeostsachya sp., USNM 528460 from loc 42128 x 2. 6, Palaeostachya sp., USNM 528462 from loc 42128
x 2. 7, Palaeostachya sp., USNM 528463 from loc 42128 x 2.
 87

FIGURE 11. Sphenophyllum angustifolium, Sphenophyllum verticillatum. 1, Sphenophyllum angustifolium, USNM 528658 from loc 42120 x 2. 2,
Sphenophyllum angustifolium, USNM 539019 from loc 42120 x 2. 3, Sphenophyllum verticillatum, USNM 539027 from 42121 x 2.4, Sphenophyllum
angustifolium, USNM 539018 from loc 42120 x 2.5, Sphenophyllum angustifolium, USNM 539016 from loc 42120 x 2. 6, Sphenophyllum verticillatum,
USNM 539028 from loc 42121 x 4. 7, Sphenophyllum angustifolium, USNM 539029 from loc 42121 x 2.
88

FIGURE 12. Sphenophyllum. 1, Sphenopyllum oblongifolium, USNM 528637 and Pecopteris unita pinnules from loc 41875 x 2. 2, Sphenophyllum sp.,
stem, USNM 538821 from loc 42126 x 2. 3, Sphenophyllum verticillatum, USNM 528640 from loc 41875 x 1. 4, Sphenophyllum verticillatum, USNM
539017 from loc 42120 x 4, 5, Sphenophyllum oblongifolium, USNM 528636 from loc 41876 x 1.6, Sphenophyllum verticullatum, USNM 528654 from
loc 42121 x 2. 7, Sphenophyllum sp., stem, USNM 538821 from loc 42126 x 1.8, Sphenophyllum sp., stem, USNM 538820 from loc 42126 x 1.
 89

FIGURE 13. Sphenophyllum sp., Sphenophyllum oblongifolium. 1, Sphenophyllum sp., USNM 536773 from loc 40621 x 2. 2, Sphenophyllum sp.,
USNM 536805 from loc 42299 x 1.3, Sphenophyllum cf. thonii, USNM 538602 from loc 42293 x 4. 4, Sphenophyllum oblongifolium, USNM 536773
from loc 40621 x 0.5. 5, Sphenophyllum oblongifolium, USNM 538805 from loc 42128 x 4.
90

FIGURE 14. Odontopteris cf. lingulata/Macroneuropteris scheuchzeri, Odontopteris cf. brardii. 1, Odontopteris cf. lingulata or Macroneuropteris
scheuchzeri, USNM 536794 from loc 42296 x 2. 2, Odontopteris cf. brardii, USNM 528477 from loc 42178 x 4. 3, Odontopteris cf. lingulata or
Macroneuropteris scheuchzeri, USNM 536788 from loc 42295 x 2.4, Odontopteris cf. lingulata, USNM 528466 from loc 42128 x 2.5, Odontopteris cf.
lingulata, USNM 528464 from loc 42128 x 2. 6, Odontopteris cf. brardii, USNM 536781 from loc 42295 x 2.
FIGURE 15 (Facing page). Alethopteris zeilleri. 1, Alethopteris zeilleri, USNM 528644 from loc 41877 x 2. 2, Alethopteris zeilleri, USNM 528657 from
loc 42120 2a with shaded polygon indicating area of enlargement seen in 2b; 2a x 12b x 2. 3, Alethopteris zeilleri, USNM 538842 from loc 42120 x 2.
4, Alethopteris zeilleri, USNM 538809 from loc 42128 x 2.5, Alethopteris zeilleri, USNM 536804 from loc. 42299 x 2. 6, Alethopteris zeilleri, USNM
538810 from loc 42128 x 2. 7, Alethopteris zeilleri, USNM 538898 from loc 42298 x 2.
FIGURE 16 (Following page). Alethopteris zeilleri. 1, Alethopteris zeilleri, USNM 538840 from loc 42120 x 2. 2, Alethopteris zeilleri, USNM 538841 from
loc 42120 x 2. 3, Alethopteris zeilleri, USNM 538844 from loc 42120. 4, Alethopteris zeilleri, USNM 539042 from loc 42123 x 2.5, Alethopteris zeilleri,
USNM 538845 from loc 42120 x 2. 6, Alethopteris zeilleri, USNM 539030 from loc 42121 x 1.7, Alethopteris zeilleri, USNM 539037 from loc 42122
x 2.
94

FIGURE 18. Indeterminate pteridosperm pinnules. All specimens illustrated here are of uncertain affinity. They may be referred to a Neurocallipteris
species such as N. planchardii, an undescribed form of Neuropteris, or even to Alethopteris (though their venation is not like that of A. zeilleri). 1, USNM
536813 from loc 42297 x 2. 2, USNM 536783 from loc 42295 x 2. 3, USNM 536816 from loc 42297 x 2. 4, USNM 536796 from loc 42296 x 4.

Fertile foliage is of the same size and shape as the sterile, but lacks clearly
defined veins, which are obscured by reproductive organs. The sporangia
are arranged circular sori that are closely packed, extending to the margin
of the pinnule in elongate clusters at right angles to the midvein.
A number of forms attributable to Pecopteris accounted for a few
percent of most collections, and rarely are common to abundant. Most
are characterized by small to medium sized pinnules with broad to slightly
constricted basal attachments to the subtending rachis (Figs. 27.1-27.4,
27.6, 28, 30.2). Many of these pecopterid pinnules or partial pinnae
appear to be fragmentary bits of large fronds with lobate pinnule-pinna
architecture, which, where the pattern of lobing and associated venation
could be seen more fully, conform broadly to the concept of Lobatopteris
(Fig. 29) (Wagner, 1958a). Venation in the lobatopterid forms varies in a
regular manner. In some pinnules venation is simple throughout a pin
nule. This grades into pinnules in which some veins are once-branched,
which then grades further into pinnules with more complex vein branch
ing architectures approximating those typical of lobatopterid pecopterids
(Wagner, 1958a). Lobatopterid venation stereotypically is characterized
by veins that branch in a fasciculate candelabra-like manner, each succes
sive forking occurring toward the interior of the fascicle. Terminal pin
nules, where seen, are small, terminating in a gradually tapering, elongate
pinna tip, with pinnules that become progressively smaller in size, with
out fusion. The specimens illustrated in Figure 29 can be compared
tentatively with Lobatopteris (Pecopteris) tenuinervis of Fontaine and
White (1880).
Pinnae of Pecopteris (Ptychocarpus) unita (Figs. 12.1, 27.5) con
sist of small pinnules of rectangular aspect, with a broad basal attach
ment to the rachis and slight confluence between adjacent pinnules, straight
parallel sides, and a blunt, often nearly flat tip. In other instances, these
pinnules may be fused laterally for nearly their full length, leaving just a
free scalloped margin to the pinna. The midvein of a pinnule varies from
acropetally curving at the base, straightening out about one-third from
the base of the pinnule lamina, to straight from the rachis to somewhere
between two-thirds and completely to the pinnule margin. Lateral veins
arch steeply upward in a concave path from their point of midvein
insertion to the margin, often forming what can be described as a fascicle.
In the most extreme cases nearly all lateral veins reach the margin in the
upper 25% of the pinnule.
Nemejcopteris feminaeformis (Figs. 30.1, 30.3, 30.4) is a filicalean
fern of zygopterid affinity. It is easily recognizable by the form of its
pinnules. These are small, broadly attached, narrow, and with serrate
pinnule margins that extend onto a blunt tip. The pinnule midvein is
weak and follows a somewhat irregular path. Lateral veins are high angle
and terminate in the marginal serrations; they may fork one time, about
half-way between midvein and margin.
Sphenopteris biturica was identified from fragmentary pinnae bear
ing small pinnules with sharply pointed, shallow teeth (Figs. 31.1, 31.4
31.8). Larger pinnules may have six or more teeth per side, borne on
larger shallow, somewhat rounded lobes. Pinnules shrink gradually in
size in the vicinity of the pinna apex, creating an elongated appearance,
with the terminal pinnule being small and slightly fused to the nearest
lateral pinnules. Venation is difficult to observe. It appears to be sparse,
with a thin midvein that persists to near the pinnule apex, the lateral
veins terminating in the pinnule teeth. These specimens look similar to
those described as S. biturica by Wagner (1985).
Commonly encountered at some sampling horizons are partial
fronds of sphenopterid aspect, similar to Oligocarpia gutberi or O.
leptophylla (Fig. 32). On the largest specimens, the ultimate pinnae are
borne at nearly rightangles to the next higher order rachis, in alternate and
opposite arrangement. These pinnae shrink in length, gradually approach
ing the apex of the next higher order pinna, creating an overall, somewhat
pointed aspect to the higher-order pinna. Ultimate pinnae reach lengths
of about 2 cm. The pinnules are free up to the apical area of the pinna,
where they fuse laterally, forming an enlarged laminar, apical pinnule,
which is of rounded form. There may be a slight enlargement of the basal
basiscopic pinnule, though this could not be determined with certainly

FIGURE 17 (Previous page). Alethopteris zeilleri. 1, Alethopteris zeilleri, USNM 538827 from loc 42125; 1b with shaded polygon indicating area of
enlargement seen in 1a; 1a x 2, 1b x 0.5. 2, Alethopteris zeilleri, USNM 538824 from loc 42125 x 2. 3, Alethopteris zeilleri, pinna tip, USNM 538832 from
loc 42125 x 4. 4, Alethopteris zeilleri, USNM 538823 from loc 42125. 5, Alethopteris zeilleri, USNM 538822 from loc 4215 x 1.6, Alethopteris zeilleri,
USNM 538830 from loc 42125 x 2. 7, Alethopteris zeilleri, USNM 538831 from loc 42125 x 2.
FIGURE 19 (Facing page). Macroneuropteris scheuchzeri, Neurocallipteris cf. planchardii, Macroneuropteris/Odontopteris lingulata. 1, Macroneuropteris
scheuchzeri, USNM 539034 from loc 41221 x 1. 2, Macroneuropteris scheuchzeri, enlargment of 19-1, x 1.5. 3, Neurocallipteris cf. planchardii, USNM
528461 from loc 42128 x 2.4, Macroneuropteris scheuchzeri, USNM 539025 from loc 42120 x 2.5, Macroneuropteris scheuchzeri, USNM 538843 from
loc 42120 x 1. 6, Macroneuropteris scheuchzeri or Odontopteris lingulata, USNM 536807 from loc 42299 x 2. 7, Macroneuropteris scheuchzeri, USNM
536814 from loc 42297 x 2. 8, Macroneuropteris scheuchzeri, USNM 538812 from loc 42128 x 4.9, Macroneuropteris scheuchzeri, USNM 539033 from
loc 41422 x 4.
95
 97

FIGURE 21. Pseudomariopteris cordato-ovata. 1, Pseudomariopteris cordato-ovata, USNM 528638 from loc 41878 x 2. 2, Pseudomariopteris cordato
ovata, USNM 539040 42123 x 4.

FIGURE 20 (Facing page). Macroneuropteris scheuchzeri. 1, Macroneuropteris scheuchzeri, USNM 538816 from loc 42126 x 1. 2, Macroneuropteris
scheuchzeri, USNM 538818 from loc 42126 x 2. 3, Macroneuropteris scheuchzeri, USNM 538817 from loc 42126 x 2.4, Macroneuropteris scheuchzeri,
USNM 539035A from loc 42125 x 2.5, Macroneuropteris scheuchzeri, USNM 538828 from loc 42125 x 2. 6, Macroneuropteris scheuchzeri, USNM
538829 from loc 42125 x 2. 7, Macroneuropteris scheuchzeri, USNM 538825 from loc 42125 x 1.8, Macroneuropteris scheuchzeri, USNM 528641 from
loc 41878 x 1. 9, Macroneuropteris scheuchzeri, USNM 538826 from loc 42125 x 2.
98

FIGURE 23. Taeniopteris sp. 1a, Taeniopteris sp., USNM 536812a from loc 42298 x 2.1b, Taeniopteris sp., USNM 536812b from loc 42298 x 2. 2,
Taeniopteris sp., USNM 536815 from loc 42297 x 2.
due to the quality of preservation. Pinnules are shallowly lobed. The
pinnule lamina appears to be thin, given its weak carbonization and the
clarity of the veins. The midvein is flexuous and could be argued to
extend nearly to the tip of the pinnule in some instances; in most cases
the midvein forks and becomes indistinguishable from the lateral veins
somewhere between one-half and two-thirds of the distance from the
base of the pinnule. Lateral veins are sparse and mostly extend straight to
the margin at about a 30-45o angle; the basalmost veins, particularly on
the basiscopic side, tend to curve upward strongly as they approach the
pinnule margin. Veins may be unforked or may fork widely, one or more
times before reaching the margin of the pinnule.
FLORISTIC PATTERNS
Flora was sampled at 21 locations from five stratigraphic hori
zons. There are compositional and taphonomic differences among these
sampling levels (Fig. 1; Table 1).
The lowest horizon, designated “Copper Mine,” consists of coarse
siltstone and fine sandstone that appears to have been deposited in an
active channel setting with flashy discharge, perhaps briefly abandoned
parts of a braided stream, where plants were growing on bars in the
stream or along the channel banks. Several examples of upright calamite
stems were noted and collected. These were up to 9 cm in diameter and
appeared to have been buried rapidly, in place. Some flattened calamite
stems were over 25 cm in diameter. Large mats of Sphenophyllum also
were common at one site, suggesting burial of mats or thickets of low
growing habit, close to the site of growth. Calamite remains dominated
most of the collections from this horizon. Sphenophylls, cordaitalean
foliage (and large, possibly cordaitalean seeds), and Oligocarpia each
were common at one of the collecting sites.
The second stratigraphic horizon from the bottom, designated
“Myriopod trackway” because of the large arthropleurid track found at
that level (Lucas et al., 2005), contains a flora previously described by
FIGURE 22. Cordaites sp. 1, Cordaites sp., USNM 536803 from loc 42293; 1a with shaded polygon indicating area of enlargement seen in 1b; 1a x 1, 1b
x 4. 2, Cordaites sp., USNM 528480 from loc 42128 x 1.3, Cordaites sp., USNM 528484 from loc 42128 x 1. 4, Cordaites sp., USNM 528481b from loc
42128 x 1.5, Cordaites sp., USNM 536790 from loc 42295 x 1.
99
Hunt and Lucas (1992). Collections were obtained from two beds of
micaceous, finely to thickly laminated red siltstone with a clay matrix.
The flora from the lower of the two beds, from which 90 hand samples
were collected, contains only Alethopteris zeilleri and Macroneuropteris
scheuchzeri. The specimens are large, though fragmentary, and appear to
have been deposited in standing quiet water; the sediments are laminated
claystones to fine siltstones and the plant fossils are undistorted and
flat-lying on all lamina/bedding surfaces. A considerably smaller collec
tion was obtained from the upper bed in this interval. This collection was
dominantly Macroneuropteris scheuchzeri, with common indeterminate
sphenophyll stems and sparse Annularia carinata whorls. This flora
also appears to have been deposited in a lake or quiet body of water.
Three collections were obtained from the third stratigraphic hori
zon, designated “El Cobre Canyon 3”. Two of these, both very small
voucher collections, come from organic-rich, fine sandstone with
interlaminated layers of organics, consisting of plant material of various
sizes, including comminuted plant debris. The plant material is some
what jumbled and appears to have been deposited rapidly during a pe
riod of waning flow, carried partly in bedload. Most noteworthy is the
occurrence of a single specimen of Walchia. The third collection, consist
ing of nearly 100 hand samples, was obtained from a gray siltstone with
fine-sand stringers. No species dominates this diverse flora. The most
common elements, represented on more than 10% of the hand samples,
include the tree-fern foliage Danaeites, calamite stems, an indeterminate,
small-pinnuled Pecopteris, and Oligocarpia gutberi. Some of the calam
ite remains were preserved in situ, upright, buried in place. As with most
of the previously discussed collections, this flora appears to have been
deposited under slackening flow in an active stream channel deposit,
perhaps in a braided channel with variable, possibly flashy or seasonal
changes in water volume. Walchia is represented in this third collection
by a single specimen.
The fourth collecting horizon, designated “El Cobre Canyon 1,”
100

FIGURE 24. Walchia sp. 1, Walchia sp., USNM 538799 from loc 42127 x 4. 2, Walchia sp., USNM 538806b cone scale from loc 42298 x 4.3, Walchia
sp., USNM 528482a from loc 42128 x 4. 4, Walchia sp., USNM 536627a from loc 42298 x 4.5, Walchia sp., USNM 528482b from loc 42128 x 4, 6,
Walchia sp., USNM 536811 from loc 42298 x 4.

FIGURE 25 (Facing page). Danaeites emersonii. 1, Danaeites emersonii, USNM 538808 from loc 42128 x 2. 2, Danaeites emersonii, USNM 538804 from
loc 42128 x 2. 3, Odontopteris lingulata, USNM 538919 from loc 42296 x 2. 4, Danaeites emersonii, sterile foliage, USNM 536780 from loc 42295 x 1.
5, Danaeites emersonii with tips of Odontopteris cf. lingulata, USNM 536795A from loc 42296 x 1. 6, Danaeites emersonii, USNM 536798 from loc
42296 x 4. 7, Danaeites emersonii, USNM 536784 from loc 42295; 7a x 1, 7b x 4.
101
102

FIGURE 26. Danaeites emersonii fertile foliage. 1, Danaeites emersonii, USNM 536779 from loc 42295 x 2. 2, Danaeites emersonii, USNM 536786 from
loc 42295 x 1. 3, Danaeites emersonii, USNM 536799 from loc 42296 x 3. 4, Danaeites emersonii, USNM 539036 from loc 42122 x 3.
 103

FIGURE 27. Pecopteris sp., Pecopteris unita. 1, Pecopteris sp., USNM 528469 from loc 42128 x 1. 2, Pecopteris sp., USNM 528649 from loc 41877 x
2. 3, Pecopteris sp., USNM 536789 from loc 42295 x 2.4, Pecopteris sp., USNM 528468 from loc 42128 x 2.5, Pecopteris unita, USNM 528637 from
loc 41875; 5a with shaded polygon indicating area of enlargement seen in 5b; 5a x 2, 5b x 8. 6, Pecopteris sp., USNM 528478 from loc 42128 x 3.
104

FIGURE 28. Pecopteris sp. 1, Pecopteris sp., USNM 539015 from loc 42120 x 4. 2, Pecopteris sp., USNM 538811 from loc 42128 x 4. 3, Pecopteris sp.,
large USNM 539032 from loc 42121 x 5.4, Pecopteris sp., USNM 528467 from loc 42128 x 2.5, Pecopteris sp., USNM 528470 from loc 42128 x 2. 6,
Pecopteris sp., USNM 539026 from loc 42120 x 2. 7, Pecopteris sp., USNM 538815 from loc 42128 x 4.
FIGURE 29 (Facing Page). Lobatopteris sp., cf. L. tenuinervis. 1, Lobatopteris sp., USNM 536793 from loc 42295 x 2. 2, Lobatopteris sp., USNM 538837
from loc 42294 x 4.3, Lobatopteris sp., USNM 538807 from loc 42128 x 2.4, Lobatopteris sp., USNM 536792 from loc 42295 x 2.
105
106

FIGURE 30. Nemejcopteris feminaeformis, fertile Pecopteris sp. 1, Nemejcopteris feminaeformis, USNM 536806 from loc 42299 x 4. 2, Fertile Pecopteris
sp., USNM 538802 from loc 42128 x 2. 3, Nemejcopteris feminaeformis, USNM 539012 from loc 42120 x 4.4, Nemejcopteris feminaeformis, USNM
528653 from loc 42120 x 2.
 107

FIGURE 31. Sphenopteris biturica. 1, Sphenopteris biturica, USNM 539061 from loc 42299 x 4. 2, Sphenopteris sp., USNM 536797 from loc 42296 x
2. 3, Sphenopteris sp., USNM 539011 from loc 42120 x 4. 4, Sphenopteris biturica, USNM 538848 from loc 42120 x 2.5, Sphenopteris biturica, USNM
539024 from loc 42120 x 5.6, Sphenopteris biturica, USNM 538847 from loc 42120 x 4. 7, Sphenopteris biturica, USNM 538846 from loc 42120 x 2.
8, Sphenopteris biturica, USNM 539014 from loc 42120 x 2.
108

FIGURE 32. Oligocarpia gutbieri/leptophylla. 1, Oligocarpia gutbieri/leptophylla, USNM 528656 from loc 42120; 1a with shaded polygon indicating
area of enlargement seen in 1b; 1a x 2, 1b x 4. 2, Oligocarpia gutbieri/leptophylla, USNM 536782b from loc 42295 x 2.
 109

FIGURE 33. Miscellaneous unknowns. 1, Smooth axes, possibly lycopsid rootlets, USNM 528472 from loc 42128 x 1. 2, Unidentified foliage, USNM
528639 from loc 41878 x 4. 3, Punctate axis, USNM 538839 from loc 42120 x 2. 4, Ribbon-like axes, probably lycopsid rootlet, USNM 538801 from
loc 42128 x 2.5, Ribbon-like axis, probably lycopsid rootlet, USNM 536785 from loc 42295 x 2.
110
FIGURE 34. Seeds. 1, Seed, USNM 538813 from loc 42128 x 1. 2, Seed, USNM 538833 from loc 42295 x 1. 3, Seed, USNM 538834 from loc 42294 x
1. 4, Seed, USNM 538835 from loc 42294 x1.5, Seed, USNM 538836 from loc 42294 x 1. 6, Seed, USNM 538838 from loc 40621 x 1. 7, Seed, USNM
538800 from loc 40621 x 1. 8, Seed, USNM 539041 from loc 42123 x 1.
TABLE 1. El Cobre Canyon flora: Composition by stratigraphic level.
was the bed on which we focused most of our collecting activity. The
eight separate USNM collections consist of dark gray siltstones and
sandstones with abundant mica, variably iron-rich, poorly bedded, bear
ing poorly sorted, often jumbled plant material. Preservation of the plant
material is variable, although frequently relatively good. The plants ap
pear to have been deposited rapidly, suggesting local transport and depo
sition in active stream channels. One collection (USNM 41875) a
heterolithic bed of thinly bedded finer and coarser siltstones and fine
sandstones, suggesting transition from channel to still-water conditions.
These collections contain a diverse flora typical of a Late Pennsylvanian
wetland. Most are dominated by or rich in Alethopteris zeilleri with
significant amounts of Macroneuropteris scheuchzeri, Pecopteris
(Pecopteris unita and Pecopteris spp.), Sphenophyllum (S. oblongifolium,
S. verticillatum, and S. angustifolium), and Sigillaria brardii, including
its cones and foliage. With the exception of a single, questionable conifer
cone scale, this flora appears to be drawn from a wet substrate environ
ment probably growing under a relatively seasonally wet, perhaps hu
mid-to-subhumid climate regime.
The final, and stratigraphically highest (youngest), collecting ho
rizon is designated “Uranium Mine.” The matrix of all four Uranium
Mine collecting sites consists of micaceous coarse siltstone and fine
sandstone with considerable iron staining. The plant fossils vary from
jumbled in the sediment to flat lying, and appear to have been trans
ported locally and probably deposited rapidly from flood and during
post-flood slack water conditions, much like the other deposits of silt
stone and fine sandstone described above. The assemblage includes large
axes as well as finer plant material. There is no segregation of plant taxa
within the assemblage. Most noteworthy about the composition of these
assemblages is the mixture of the tree fern foliage Danaeites and plants
more typical of seasonally moisture limited conditions, perhaps
subhumid-to-semiarid conditions: Walchia pinniformis, Taeniopteris sp.,
and Odontopteris cf. lingulata. More generally “wet” environment plants
are rare (Alethopteris zeilleri and Macroneuropteris scheuchzeri, calam
ite remains, and sphenophylls).
DISCUSSION
The flora of the El Cobre Canyon Formation is heterogeneous
among the 21 collecting localities from five distinct horizons, but ap

pears to be drawn largely from a common regional species pool, allowing
for the vagaries of sampling and small specimen numbers. It consists
mainly of plants typical of wet substrates, characteristic of Pennsylva
nian wetland vegetation from coal basins of central and eastern North
American through central Europe. Only at the youngest horizon are
plants typical of drier conditions a conspicuous element of the flora.
The floral composition, particularly the combined presence of
Alethopteris zeilleri, Sphenophyllum verticillatum, Sphenophyllum
angustifolium, Annularia carinata, Pseudomariopteris cordato-ovata,
Oligocarpia gutberi/leptophylla, and Sphenopteris biturica strongly sug
gests a later Late Pennsylvanian age, Virgilian in regional terminology,
Stephanian Cin global terrestrial terminology. Sigillaria brardii, Danaeites
emersonii, Nemejcopteris feminaeformis, Odontopteris cf. brardii and
Sphenophyllum oblongifolium are all consistent with a Late Pennsylva
nian (Stephanian) age.
The occurrence of small numbers of conifers (Walchia) and
Taeniopteris is consistent with landscape heterogeneity created by cli
matic seasonality, and with the sedimentary style. The presence of these
plants requires a somewhat more extended discussion.
It is clear that floras dominated by xeromorphic plants are
stratigraphically intercalated throughout the Late Pennsylvanian with
typically wetland floras throughout the Euramerican equatorial region
(Broutin et al., 1990; DiMichele and Aronson, 1992; Kerp, 1996, 2000;
DiMichele et al., 2008). Such stratigraphic, bed- or formational-scale
intercalation likely reflects climate oscillation between seasonally dry,
semi-arid to subhumid climate regimes and humid climates on time scales
of tens of thousands of years. The combination of apparently
parautochthonous, environmentally contrasting mixed floras at some of
the Cañon del Cobre collecting sites requires the existence of a mixture of
physical conditions on the landscape from which the flora was drawn.
In the case of the El Cobre Canyon Formation, there is strong
evidence that most of the plant remains typical of wetter substrates
represent mainly plants growing within and along channels or on the
margins of lakes. For example, there are numerous examples of in situ
calamite stems buried upright, in place. And, in several collections,
Sphenophyllum specimens formed masses, as if buried in place or trans
ported only short distances from stream banks or exposed intra-channel
bars. These deposits also typically have abundant preserved organic
matter and plant remains that cross bedding surfaces, indicating that such
organic material was likely carried in bed load, with periodic, rapid depo
sition. In the few instances of lake-like deposits, the floras were of
monotonously low diversity, and appeared to consist of elements floated
from the margin of the water body into a quiet depositional setting,
largely the seed ferns Alethopteris zeilleri and Macroneuropteris
scheuchzeri.
Evidence of plants growing within the stream channel itself, of
flashy discharge, and rapid burial of upright plants, all suggest seasonally
REFERENCES
Asama, K. 1970. Evolution and classification of Sphenophyllales in
Cathaysia land: Bulletin National Science Museum, v. 13, p. 291-317.
Barlow, J.A., ed., 1975, Proceedings of the First I.C. White Memorial
Symposium, “The Age of the Dunkard” (September 25-29, 1972). West
Virginia Geological and Economic Survey, Morgantown, WV, 352 p.
Barthel, M., 1976, Die Rotliegendflora Sachsens: Abhandlungen Staatliches
Museum für Mineralogie und Geologie (Dresden), v. 24, p. 1-190.
Blake, B.M., Jr., Cross, A.T., Eble, C.F., Gillespie, W.H. and Pfefferkorn,
H.W., 1999, Selected plant megafossils from the Carboniferous of the
Appalachian region, eastern United States: Geographic and stratigraphic
distribution; in Hills, L.V., Henderson, C.M. and Bamber, E.W., eds.,
Carboniferous and Permian of the world: Canadian Society of Petroleum
Geologists Memoir, v. 19, p. 259-335.
111
dry conditions (Fielding et al., 2009). Under such conditions, the land
scape around the channel systems would be expected to be far more
heterogeneous than under more uniformly wet climatic conditions. In
such conditions, it would be possible to have typically “wetland” plants
growing in close proximity to xeromorphic species typical of seasonally
dry landscapes. Prolonged drought increases the microhabitat heteroge
neity of a landscape and encourages local variability, resulting in strong
ecological contrasts between wetter stream corridors and drier inter
fluves.
The paleobotanical patterns at Cañon del Cobre suggest an in
crease in seasonal moisture limitation in the youngest deposits. This is
consistent with patterns identified in other Late Pennsylvanian deposits
from western America (e.g., Mapes and Gastaldo, 1987; Rothwell and
Mapes, 1988; Mamay and Mapes, 1992; DiMichele et al., 2005). Cañon
del Cobre floras are significantly different from younger deposits to the
south in New Mexico. Those from Carrizo Arroyo, which span the
Pennsylvanian-Permian boundary (Lucas and Zeigler, 2004), and thus
are likely slightly younger than El Cobre, are dominated by plants more
typical of seasonally dry settings (Tidwell et al., 1999), including coni
fers, callipterids and the probable seed fern Sphenopteridium
manzanitanum (DiMichele et al., 2004). Floras from the Bursum Forma
tion have been identified still further to the south, in Socorro County, at
the Pennsylvanian-Permian boundary (DiMichele et al., unpublished),
dominated by a seasonally dry flora of callipterids, conifers and
cordaitaleans. This pattern of the progressive appearance of ever drier
floras is one generally identified for the transition between the Pennsyl
vanian and Permian, though it is not certain that it occurs in a time
equivalent manner across the Euramerican portion of Pangaea. What is
certain is that at least two, if not more, distinct biomes, characterized by
different dominant taxa, coexisted in tropical and paratropical latitudes
during the late Paleozoic. The origin and dynamics of the distinct biomes
that developed across the Carboniferous-Permian boundary has been
difficult to interpret for many years. However, information is slowly
emerging that sheds light on this enigmatic period. We have every reason
to believe that further field studies (e.g. Looy, 2007) will continue to
reveal the details of the links between the ecological processes and evo
lutionary origins of the assemblages and their constituent lineages.
ACKNOWLEDGMENTS
The authors thank the Smithsonian Institution for support of
field work in New Mexico. We are grateful to various NMMNH volun
teers and staff, who helped collect the Cañon del Cobre paleofloras. We
thank the late Richard Ballsard who contributed useful comments on
earlier versions of this manuscript. We also wish to thank reviewers
Arden Bashforth, B. Mitchel Blake, and Robert H. Wagner who made
extensive and valuable comments on the paper, and Jack Wittry for
comments on some of the identifications.
Broutin, J., Doubinger, J., Farjanel, G., Freytet, P., Kerp, H., Langiaux, J.,
Lebreton, M.L., Sebban, S. and Satta, S., 1990, Le renouvellement des
flores au passage Carbonifére-Permian: approche stratigraphique,
biologique, sédimentologique: Comptes Rendus de l’Académie des Sci
ences Paris II, v. 311, p. 1563-1569.
Castro, M.P., 2005, La flora estefaniense B de La Magdalena (Léon, España),
un referente europeo. Tomo I: Antecedentes y análisis floristico:
Publicaciones del Instituto Geológico y Minero de España, Cuadernos
del Museo GeoMinero, v.4, p. 1-251.
Cleal, C.J., Shute, C.H. and Zodrow, E.L., 1990, A revised taxonomy for
Palaeozoic neuropterid foliage: Taxon 39: 486-529.
Cridland, A.A. and Morris, J.E., 1963, Taeniopteris, Walchia, and
Dichophyllum in the Pennsylvanian System in Kansas: University of
112
Kansas Science, Bulletin 44, p. 71-85.
DiMichele, W.A. and Aronson, R.B., 1992, The Pennsylvanian-Permian
vegetational transition: A terrestrial analogue to the onshore-offshore
hypothesis: Evolution, v. 46, p. 807-824.
DiMichele, W.A., Kerp, H. and Chaney, D.S., 2004, Tropical floras of the
Late Pennsylvanian-Early Permian transition: Carrizo Arroyo in con
text: New Mexico Museum of Natural History, Bulletin 25, p. 105-110.
DiMichele, W.A. and Chaney, D.S., 2005, Pennsylvanian-Permian fossil
floras from the Cutler Group, Cañon del Cobre and Arroyo del Agua
areas, in northern New Mexico: New Mexico Museum of Natural His-
tory and Science, Bulletin 31, p. 26-33.
DiMichele, W.A., Gastaldo, R.A., and Pfefferkorn, H.W., 2005, Plant
biodiversity partitioning in the Late Carboniferous and Early Permian
and its implications for ecosystem assembly: Proceedings of the Califor
nia Academy of Sciences, v. 56 (supplement 1), p. 32-49.
DiMichele, W.A., Kerp, H., Tabor, N.J., and Looy, C.V., 2008, Revisiting
the so-called “Paleophytic-Mesophytic” transition in equatorial Pangea:
Vegetational integrity and climatic tracking: Palaeoecology,
Palaeclimatology, Palaeogeography, v. 268, p. 152-163.
DiMichele, W.A., Montañez, I.P., Poulsen, C.J. and Tabor, N.J., 2009a,
Climate and vegetational regime shifts in the late Paleozoic ice age
earth: Geobiology, v. 7, p. 200-226.
DiMichele, W.A., Nelson, W.J., Elrick, S. and Ames, P.R., 2009b, Cata
strophically buried Middle Pennsylvanian Sigillaria and calamitean sphe
nopsids from Indiana, USA: What kind of vegetation was this?: Palaios,
v. 24, p. 159-166.
Doubinger, J., Vetter, P., Langiaux, J., Galtier, J., and Broutin, J. 1995, La
flore fossile du basin houiller de Saint-Étienne: Mémoires du Muséum
National d’Histoire Naturelle, v. 164, p. 1-355.
Eberth, D.A. and Miall, A.D., 1991, Stratigraphy, sedimentology and evolu
tion of a vertebrate-bearing, braided to anastomosed fluvial system,
Cutler Formation (Permian-Pennsylvanian), north-central New Mexico:
Sedimentary Geology, v. 72, p. 225-252.
Falcon-Lang, H.J., Nelson, W.J., Elrick, S., Looy, C.V., Ames, P.R. and
DiMichele, W.A., 2009, Incised channel-fills containing conifers indi
cate that seasonally dry vegetation dominated Pennsylvanian tropical
lowlands: Geology, v. 37, p. 923-926.
Feldman, H.R., Franseen, E.K., Joeckel, R.M., and Heckel, P.H., 2005,
Impact of longer-term modest climate shifts on architecture of high
frequency sequences (cyclothems), Pennsylvanian of Midcontinent,
U.S.A.: Journal of Sedimentary Research, v. 75, p. 350-368.
Fielding, C.R., Frank, T.D., Birgenheier, L., Rygel, M.C., Jones, A.T., and
Roberts, J., 2008, Stratigraphic imprint of the late Palaeozoic ice age in
eastern Australia: A record of alternating glacial and nonglacial climate
regime: Journal of the Geological Society, v. 165, p. 129-140.
Fielding, C.R., Allen, J.P., Alexander, J. and Gibling, M.R., 2009, Facies
models for fluvial systems in the seasonal tropics and subtropics: Geol
ogy, v. 37, p. 623-626.
Fontaine, W.M., 1890, Description of the species: U.S. National Museum
Proceedings, v. 13, p. 225-252.
Fontaine, W.M. and White, I.C., 1880, The Permian or Upper Carbonifer
ous flora of West Virginia and S.W. Pennsylvania: Second Geological
Survey of Pennsylvania, Report of Progress PP, p. 1-143.
Fracasso, M.A., 1980, Age of the Permo-Carboniferous Cutler Formation
vertebrate fauna from El Cobre Canyon, New Mexico: Journal of Pale
ontology, v. 54, p. 1237-1244.
Heckel, P.H. and Clayton, G., 2006, The Carboniferous System. Use of the
new official names for the subsystems, series and stages: Geologica Acta,
v.4, p. 403-407.
Hunt, A.P. and Lucas, S.G., 1992, The paleoflora of the Lower Cutler
Formation (Pennsylvanian, Desmoinesian?) in El Cobre Canyon, New
Mexico, and its biochronological significance: New Mexico Geological
Society, Guidebook 43, p. 145-150.
Kerp, H., 1996, Post-Variscan late Paleozoic northern hemisphere gymno
sperms: the onset to the Mesozoic: Review of Palaeobotany and Pa
lynology, v. 90, p. 263-285.
Kerp, H., 2000, The modernization of landscapes during the late Paleozoic
early Mesozoic: Paleontological Society Papers, v. 6, p. 80-113.
Kerp, H. and Fichter, J., 1985, Die Macrofloren des saarpfälzischen
Rotliegenden (?Ober-Karbon – Unter-Perm; SW-Deutschland): Mainzer
geowissschaftliche Mitteilungen, v. 14, p. 159-286.
Kerp, J.H.F., 1984, Aspects of Permian palaeobotany and palynology. V.
On the nature of Asterophyllites dumasii Zeiller, its correlation with
Calamites gigas Brongniart and the problem concerning its sterile foli
age: Review of Palaeobotany and Palynology, v. 41, p. 310-317.
Krings, M. and Kerp, H., 2000, A contribution to the knowledge of the
pteridosperm genera Pseudomariopteris Danzé-Corsin nov. emend. and
Helenopteris nov. gen.: Review of Palaeobotany and Palynology, v. 111,
p. 145-195.
Krings, M., Kerp, H., Taylor, T.N. and Taylor, E.L., 2003, How Paleozoic
vines and lianas got off the ground: On scrambling and climbing Carbon
iferous–Early Permian pteridosperms: Botanical Review, v. 69, p. 204
224.
Lamboy, W. and Lesnikowska, A.D., 1988, Some statistical methods useful
in the analysis of plant paleoecological data: Palaios, v. 3, p. 86-94.
Langston, W., Jr., 1953, Permian amphibians from New Mexico: Univer
sity of California, Publications in Geological Sciences, v. 29, p. 349
416.
Laveine, J.-P. 1989, Guide Paleobotanique dan le Terrain Houiller Sarro
Lorrain: Documents des Houilleres du Bassin de Lorraine, v. 1, p. 1-154.
Looy, C.V., 2007, Extending the range of derived late Paleozoic conifers:
Lebowskia gen. nov. (Majonicaceae): International Journal of Plant
Sciences, v. 168, p. 957-972.
Lucas, S.G., 2002, Tetrapods and the subdivision of Permian time; in Hills,
L.V., Henderson, C.M. and Bamber E.W., eds., Carboniferous and Per
mian of the world: Canadian Society of Petroleum Geologists Memoir,
v. 19, p. 479-491.
Lucas, S.G. and Krainer, K., 2005, Cutler Group (Permo-Carboniferous)
stratigraphy, Chama Basin, New Mexico: New Mexico Museum of Natu
ral History and Science, Bulletin 31, p. 90-100.
Lucas, S.G. and Zeigler, K.E., eds., 2004, Carboniferous-Permian transition
at Carrizo Arroyo, central New Mexico: New Mexico Museum of Natu
ral History and Science, Bulletin 25, 300 p.
Lucas, S.G., Harris, S.K., Spielmann, J.A., Berman, D.S. and Henrici, A.C.,
2005, Vertebrate biostratigraphy and biochronology of the Pennsylva
nian-Permian Cutler Group, El Cobre Canyon, northern New Mexico:
New Mexico Museum of Natural History and Science, Bulletin 31, p.
128-139.
Mamay, S.H. and Mapes, G., 1992, Virgilian plant megafossils from the
Kinney Brick Company Quarry, Manzanita Mountains, New Mexico:
New Mexico Bureau of Mines and Mineral Resources, Bulletin 138, p.
61-85.
Mapes G. and Gastaldo, R.A., 1987, Late Paleozoic non-peat accumulating
floras: University of Tennessee Studies in Geology, v. 2, p. 115–127.
Montañez, I. P., Tabor, N. J., Niemeier, D., DiMichele, W. A., Frank, T. D.,
Fielding, C. R., and Isbell, J. L. 2007. CO2-forced climate and vegetation
instability during late Paleozoic deglaciation: Science, v. 315. p. 87-91.
Pfefferkorn, H.W. and Wang, J., 2009, Stigmariopsis, Stigmaria asiatica,
and the survival of the Sigillaria brardii-ichthyolepis group in the trop
ics of the Late Pennsylvanian and Early Permian: PalaeoWorld, v. 18,
p. 130-135.
Pfefferkorn, H.W., Mustafa, H., and Hass, H., 1975, Quantitative
Characterisierung ober-karboner Abdruckfloren: Neues Jahrbuch für
Geologie und Paläontologie Abhandlung, v. 150, p. 253-269.
Phillips, T.L. and DiMichele, W.A., 1992, Comparative ecology and
life-history biology of arborescent lycopods in Late Carboniferous swamps
of Euramerica: Annals of the Missouri Botanical Garden, v. 79, p. 560-588.
Potonié, R., 1893, Die Flora des Rothliegenden von Thüringen: Abhandlungen
der königlich-preussischen geologischen Landesanstalt N.F., v. 9, p. 1
185.
Rössler, R., and Noll, R., 2006, Sphenopsids of the Permian: Pt. 1, The
largest known anatomically preserved calamite, an exceptional find
from the petrified forest of Chemnitz, Germany: Review of Palaeobotany
and Palynology, v. 140, p. 145–162.
Rothwell, G.W. and Mapes, G., 1988, Vegetation of a Paleozoic conifer
community; in Mapes, G. and Mapes, R.H., eds., Regional geology and
paleontology of upper Paleozoic Hamilton quarry area in southeastern

Kansas: Kansas Geological Survey Guidebook Series, v. 6, p. 213-224.
Rygel, M.C., Fielding, C.R., Frank, T.D., and Birgenheier, L.P., 2008, The
magnitude of late Paleozoic glacioeustatic fluctuations: A synthesis:
Journal of Sedimentary Research, v. 78, p. 500-511.
Tidwell, W.D., Ash, S. R, Kues, B.S., Kietzke, K. K. and Lucas S. G., 1999,
Early Permian megafossils from Carrizo Arroyo, central New Mexico:
New Mexico Geological Society, Guidebook 50, p. 297-304.
Utting, J., 2001, Palynological study of three samples from the Late Car
boniferous Cutler Formation, El Cobre Canyon, Rio Arriba County, New
Mexico submitted by Spencer Lucas, New Mexico Museum of Natural
History and Science (UTM Zone 13 NAD 27): Geological Survey of
Canada, Paleontological Report 1-JU-2001, 3 p.
Wagner, R.H., 1958a, Some Stephanian pecopterids from NW Spain:
Mededelingen van de Geologische Stichting New Series, v. 12, p. 5–23.
Wagner, R.H., 1958b, Pecopteris pseudobucklandi Andrae and its general
affinities: Mededelingen van de Geologische Stichting New Series, v. 12,
p. 25-31.
Wagner, R.H., 1984, Megafloral zones of the Carboniferous: Compte Rendu
Neuvième Congrès International de Stratigraphie et de Géologie du
Carbonifère (Washington and Urbana, 1979), v. 2, p. 109-134.
Wagner, R.H., 1985, Upper Stephanian stratigraphy and paleaeontology of
113
the Puertollano Basin, Ciudad Real, Spain: Annais de Faculdade de
Ciências, Universidade do Porto, v. 64, supplement, p. 171-231.
Wagner, R.H., 1998, Consideraciones sobre los pisos de la Serie Estafaniense:
Mongrafias de la Academia de Ciencias de Zaragoza, v. 13, p. 9-19.
Wagner, R.H. and Castro, M.P., 1998, Neuropteris obtusa, a rare but wide
spread Late Carboniferous pteridosperm: Palaeontology, v. 41, p. 1-22.
Wagner, R.H. and Martínez Garcia, E., 1982, Description of an Early Per
mian flora from Asturias and comments on similar occurrences in the
Iberian Peninsula: Trabajos de Geologia (Universidad de Oviedo), v. 12,
p. 273-287.
Wanless, H.R. and Weller, J.M., 1932, Correlation and extent of Pennsyl
vanian cyclothems: Geological Society of America Bulletin, v. 43, p.
1003-1016.
Wanless, H.R. and Shepard, F.P., 1936, Sea level and climate changes related
to late Paleozoic cycles: Geological Society of America Bulletin, v. 47,
p. 1177-1206.
Wing, S.L. and DiMichele, W.A., 1995, Conflict between local and global
changes in plant diversity through geological time: Palaios, v. 10, p.
551-564.
Zodrow, E.L., Šimnek, Z., Cleal, C.J., Bek, J. and Pšenika, J., 2006,
Taxonomic revision of the Palaeozoic marattialean fern Acitheca
Schimper: Review of Palaeobotany and Palynology, v. 138, p. 239–280.
114
 Lucas et al., eds., 2010, Carb-Permian transition in Cañon del Cobre. New Mexico Museum of Natural History and Science Bulletin 49.

115
VERTEBRATE PALEONTOLOGY, BIOSTRATIGRAPHYAND
BIOCHRONOLOGYOFTHEPENNSYLVANIAN-PERMIAN
CUTLER GROUP, CAÑON DELCOBRE, NORTHERN NEWMEXICO

SPENCER G. LUCAS1, SUSANK. HARRIS1, JUSTINA. SPIELMANN1,

LARRY F. RINEHART1, DAVIDS BERMAN2,AMYC. HENRICI2 AND KARLKRAINER3
1 New Mexico Museum of Natural History and Science, 1801 Mountain Road NW, Albuquerque, NM 87104;

2 Section of Vertebrate Paleontology, Carnegie Museum of Natural History, 4400 Forbes Avenue, Pittsburgh, PA 15213;
3 Institute of Geology and Paleontology, Innsbruck University, Innrain 52, Innsbruck, A-6020 AUSTRIA

Abstracts—For more than a century, fossil vertebrates have been collected from Pennsylvanian-Permian nonma
rine redbeds of the Cutler Group (El Cobre Canyon and Arroyo del Agua formations) in Cañon del Cobre (El Cobre
Canyon), Rio Arriba County, New Mexico. Megafossil plants, palynomorphs and fossil vertebrates indicate that
the Pennsylvanian-Permian (~Virgilian-Wolfcampian) boundary is stratigraphically high in the El Cobre Canyon
Formation in Cañon del Cobre, though the exact position of the boundary is not certain. A detailed vertebrate
biostratigraphy in Cañon del Cobre constructed from all localities that can be placed into precise lithostratigraphic
position identifies three, temporally successive and distinct vertebrate fossil assemblages that represent three land
vertebrate faunachrons (LVFs). The middle and upper assemblages belong to the Coyotean and Seymouran LVFs,
respectively, and the lower assemblage is the characteristic vertebrate fossil assemblage of the Cobrean LVF, an
older, entirely Late Pennsylvanian (Virgilian) LVF. The Pennsylvanian-Permian boundary is in the Coyotean LVF.
The Cañon del Cobre vertebrate fossil record thus provides three, stratigraphically-successive vertebrate assem
blages that support recognition of three time-successive LVFs across the Pennsylvanian-Permian boundary.

INTRODUCTION
Cañon del Cobre (also known as El Cobre Canyon) is a large box
canyon about 8 km northwest of the village of Abiquiu in Rio Arriba
County, north-central New Mexico (Fig. 1). The main portion of the
canyon is a complexly faulted valley with extensive exposures of the El
Cobre Canyon and overlying Arroyo del Agua formations of the Cutler
Group (Lucas and Krainer, 2005a, b; Lucas et al., 2005c, this volume;
Kempter et al., 2007). Sediments of the Chinle Group rest disconform
ably on the Arroyo del Agua Formation along the upper 30 m of the steep
canyon walls (Lucas and Hunt, 1992; Lucas et al., 2005b).
The classic vertebrate-fossil collecting area on the canyon floor, as
well as sites in the canyon walls, have yielded unique assemblages of
fossil vertebrates that are among the oldest tetrapod fossils from North
America. Here, we update the composition of these fossil assemblages as
they were reviewed by Lucas et al. (2005c). Stratigraphic organization of
the Cañon del Cobre vertebrate fossil localities indicates there are three
superposed, biostratigraphically distinct vertebrate fossil assemblages
that correspond to three land vertebrate faunachrons (LVFs) across the
Pennsylvanian-Permian boundary (Fig. 2).
Institutional abbreviations: AMNH = American Museum of
Natural History, New York; CM = Carnegie Museum of Natural His
tory, Pittsburgh; MCZ = Museum of Comparative Zoology, Harvard;
NMMNH = New Mexico Museum of Natural History and Science,
FIGURE 1. Index map showing the location of Cañon del Cobre (after Ash,
Albuquerque; UCLA = University of California, Los Angeles; and YPM 1974).
= Yale Peabody Museum, New Haven.
GEOLOGICAL SETTING have been divided into a lower, El Cobre Canyon Formation and an
upper, Arroyo del Agua Formation (Lucas and Krainer, 2005a, b) (Fig.
Cañon del Cobre is eroded along a series of southwest-northeast 2).
trending normal faults that are downthrown to the northwest (Smith et In Cañon del Cobre, a total thickness of about 230 m of El Cobre
al., 1961, pl. 6; Kempter et al., 2007; Lucas et al., this volume). The floor Canyon Formation is exposed, and composed of mostly siltstone and
and canyon walls expose extensive bedrock badlands, slopes and escarp sandstone (Fig. 2). Minor rock types are conglomerate/conglomeratic
ments developed in an ~400 m thick section of Cutler Group siliciclastic sandstone, sandy shale and calcrete. These rocks are characteristically
red beds (Smith et al., 1961; Lucas and Krainer, 2005a, b; Kempter et al., “brown” (pale reddish brown) and are readily distinguishable on the
2007). These strata of fluvial origin (Eberth and Berman, 1983, 1993; basis of color and lithology from the overlying Arroyo del Agua Forma
Fracasso, 1987; Eberth and Miall, 1991; Krainer and Lucas, this volume) tion. Siltstones of the El Cobre Canyon Formation contain numerous
116

FIGURE 2. Composite lithostratigraphic section in Cañon del Cobre, showing distribution of NMMNH fossil localities (numbered) and vertebrate taxa.

rhizoliths, and sandstones are typically coarse-grained, arkosic, mica
ceous, trough cross-bedded and form multistoried bodies that erode to
thick cliffs and benches. Conglomerates characteristically have
extraformational clasts of quartzite, granite and gneiss.
The El Cobre Canyon Formation is best exposed in the floor of
Cañon del Cobre. Its lower contact is not exposed, and subsurface data
indicate it rests on Proterozoic basement (Smith et al., 1961; Eberth and
Miall, 1991). The upper contact of the El Cobre Canyon Formation
appears to be conformable at the base of the first “orange” siltstone
slope of the Arroyo del Agua Formation. The El Cobre Canyon Forma
tion approximately corresponds to megasequence 1 of Eberth and Miall
(1991). According to Eberth and Miall (1991), sediments of the El Cobre
Canyon Formation were deposited in a shallow, ephemeral braided stream
environment.
Conformably overlying the El Cobre Canyon Formation, the Ar
royo del Agua Formation in Cañon del Cobre is as much as 160 m thick
and consists mostly of siltstone and sandstone (Fig. 2). Minor rock
types are calcrete and conglomerate/conglomeratic sandstone. Siltstones
of the Arroyo del Agua Formation are characteristically “orange” (mod
erate reddish brown) and contain abundant calcrete nodules. They form
relatively thick slopes between thin sheets of sandstone that are coarse
grained, arkosic and trough crossbedded. Conglomerates are not as com
mon in the Arroyo del Agua Formation as in the underlying El Cobre
Canyon Formation, and most are intraformational (composed of calcrete
clasts). However, in the upper part of the Arroyo del Agua Formation
extraformational conglomerates (primarily composed of quartzite clasts)
are present (this is megasequence 3 of Eberth and Miall, 1991).
The Arroyo del Agua Formation is well exposed along the walls of
Cañon del Cobre, and it conformably overlies the El Cobre Canyon
Formation. Gradation and interfingering characterize this contact, so it is
apparently conformable. Throughout Cañon del Cobre, the Upper Tri
assic Chinle Group rests disconformably on the Arroyo del Agua Forma
tion (Lucas and Hunt, 1992; Lucas et al., 2003, 2005b; Kempter et al.,
2007).
The Arroyo del Agua Formation approximately encompasses
megasequences 2 and 3 of Eberth and Miall (1991). The facies contrasts
with those of the El Cobre Canyon Formation. Thus, the Arroyo del
Agua Formation is characterized by locally abundant major sandstone
ribbons and U-shaped channel fills, separated by thick intervals of silt
stone containing abundant calcrete nodules. When compared to El Cobre
Canyon Formation deposition, the fluvial style had changed to laterally
extensive floodplains that developed between relatively stable channels,
and locally anastomosed channels formed. The climate was more arid
than during deposition of the El Cobre Canyon Formation (Eberth and
Miall 1991).
In Cañon del Cobre all but two of the known Paleozoic vertebrate
fossil localities are in the El Cobre Canyon Formation (Fig.2). All Paleo
zoic palynomorph, fossil plant and invertebrate trace fossil sites are also
in the El Cobre Canyon Formation.
VERTEBRATE FOSSILTAXA
The most recent list of fossil taxa comprising the late Paleozoic El
Cobre Canyon vertebrate assemblages was provided by Lucas et al.
(2005c) (Table 1). Here, we review the taxa and add some new records.
Gnathorhiza sp. and Sagenodus sp.
Until very recently, the only record of the dipnoan Gnathorhiza
from strata of the El Cobre Canyon Formation at Cañon del Cobre
pertained to bone preserved within a fossil burrow cast identified as
Gnathorhiza sp. (Berman, 1993). Whereas this estivation chamber was
found lying free in the classic collecting area on the canyon floor, a
lungfish tooth plate (Fig. 3J), collected by one of us (JAS) in the summer
of 2007, was recovered from strata of the north canyon wall. The tooth
plate displays characters diagnostic of the genus Sagenodus but is too
117
incomplete for a species level identification (Schultze and Chorn, 1997).
Both dipnoan specimens were collected from strata of the lower part of
the El Cobre Canyon Formation section that yield the characteristic
tetrapod assemblage of the Cobrean land vertebrate faunachron (LVF)
(Fig.2).
Chenoprosopus cf. C. milleri Mehl, 1913
The cochleosaurid amphibian Chenoprosopus milleri was estab
lished as a new genus and species based on an incomplete skull collected
from the upper part of the El Cobre Canyon Formation in the classic
collecting locality of Arroyo del Agua, Rio Arriba County, New Mexico
(Mehl, 1913). Vaughn (1963) described a cranial fragment, CM 47798
(formerly UCLA VP 1640), that he assigned to Chenoprosopus cf. C.
milleri (Lucas et al., 2005c, fig5C-D). Collected from a site in the floor of
Cañon del Cobre, the specimen consists of the dermal roof and palatal
elements of the left narial region. Vaughn (1963) noted that the posterior
position of the external naris, as well as the pitting of the dermal roof
elements, are especially characteristic of Chenoprosopus.
Eryops sp.
Ubiquitous in Lower Permian tetrapod assemblages, the
temnospondyl amphibian genus Eryops was first established by Cope
(1878) based on a skull and postcranial elements from Texas. Lucas et al.
(2005a, fig. 1A-B) reported a rare Upper Pennsylvanian record of the
genus, the first from the Cañon del Cobre, represented by a nearly com
plete skull (NMMNH P-46379), for which specific assignment awaits
publication of a species-level revision of Eryops (Milner, 1996). In this
volume, Werneburg et al. describe this skull in detail and discuss its
taxonomic position.
Broiliellus novomexicanus (Williston, 1911b)
Williston (1911b, plate 38, fig. 1) described the dissorophidam
phibian Aspidosaurus novomexicanus on the basis of a skull and articu
lated presacral vertebral column (YPM 810) that may have been col
lected near the “type” locality of Limnoscelis paludis in Cañon del Cobre
(Langston, 1953). However, the exact locality is unknown.
Fracasso (1980) reported A. novomexicanus to be endemic to
Cañon del Cobre at the specific level, apparently unaware of Langston’s
(1953) reassignment of the specimen to the genus Broiliellus under the
new combination Broiliellus novomexicanus. In addition, Langston pro
vided a description of a well-preserved skull of B. novomexicanus from
the upper part of the El Cobre Canyon Formation (Lower Permian) in
the nearby locality of Arroyo del Agua. Langston’s reassignment was
followed by Berman and Lucas (2003), despite Carroll’s (1964) reassign
ment to Aspidosaurus (which was followed by others, including Bolt
and DeMar in various articles). The description of a new species of
Aspidosaurus, A. binasser, based on the most complete material of the
genus known to date, allowed Berman and Lucas (2003) to state un
equivocally that “Aspidosaurus obviously does not include A.
novomexicanus.” This exclusion of the species from the genus
Aspidosaurus is followed here, though we acknowledge, as noted by A.
Huttenlocker (written commun., 2009), that the generic assignment of A.
novomexicanus merits further study. In this volume, Werneburg et al.
describe a dissorophoid snout from the type locality of Anconastes
vesperus.
Platyhystrix rugosus (Case, 1910)
Williston (1911a,b) recognized the generic distinctiveness of elon
gate neural spines with small, irregular bosses assigned to the pelycosaur
Ctenosaurus rugosus (Case, 1910) and consequently erected the new
combination, Platyhystrix rugosus, for their reception. Vaughn (1963)
reported the presence of P. rugosus in the Cañon del Cobre vertebrate
assemblage based on a large number of neural spine fragments with sur
face ornamentation consistent with that of P. rugosus from the Arroyo
118
TABLE 1. Fossil vertebrate taxa from the Cutler Group in Cañon del Cobre.
del Agua locality. The largest of these neural spines (CM 47797, for
merly UCLA VP 1638) has a height of 24 cm and an anteroposterior
length of 55 mm (Lucas et al., 2005c, fig.5B). Isolated neural spines with
similar ornamentation have also been collected by NMMNH field par
ties (Lucas et al., 2005c, fig. 7A). The monospecific genus Platyhystrix is
notable among other fossil taxa of the Cañon del Cobre vertebrate assem
blages in that it is the only taxon with a stratigraphic range that spans all
three LVFs represented in the canyon, from the Cobrean through the
Seymouran (Fig.2).
Anconastes vesperus Berman, Reisz and Eberth, 1987
Berman et al. (1987, figs. 5-11) described a new genus and species
of trematopid amphibian, Anconastes vesperus on the basis of an incom
plete, articulated skeleton with skull and mandibles (holotype, CM 41711)
and the right margin of a skull with mandible (paratype, CM 28590).
With the possible exception of Ecolsonia cutlerensis, this is the only
definitive trematopid known from New Mexico, and represents a rare
Upper Pennsylvanian occurrence of this otherwise mainly Lower Per
mian family. The A. vesperus fossils were recovered from near the base
of the locally exposed section of the El Cobre Canyon Formation and
thus represent the stratigraphically lowest vertebrate taxon of the char
acteristic assemblage of the Cobrean LVF (Fig. 2).
Sumida et al. (1996) reported the possible occurrence of this
trematopid genus from the Lower Permian Lower Tambach Formation of
central Germany (Bromacker locality). However, the material, which
consists of a partial skull and postcrania, was subsequently described as
a new trematopid genus and species, Tambachia trogallas (Sumida et al.,
1998).
Embolomeres
A multispecimen occurrence of an undetermined embolomerous
anthracosaur was reported by Berman (1993) from strata of the upper
part of the El Cobre Canyon Formation that yields tetrapods character
istic of the Coyotean LVF (Lucas et al., 2005c). The presence of a much
larger embolomere within the stratigraphically lower Cobrean assem
blage is suggested by a single, isolated embolomerous pleurocentrum
with a diameter, if complete, that would be three times those reported by
Berman (1993) for the smaller embolomere (Lucas et al., 2005c, fig. 7G).
Diasparactus zenos Case, 1910
Case (1910, 1911) described the diadectid species Diasparactus
zenos from material collected in the 1870s by D. Baldwin (precise local
ity unknown) in Cañon del Cobre. The holotype material consists of a
short sequence of articulated dorsal and sacral vertebrae characterized by
the small size of the centra relative to the height and width of the neural
arch and zygapophyses. During the 1911 Williston expedition to Cañon
del Cobre, Case and Williston collected a nearly complete Diasparactus
zenos skeleton that includes the lower portion of a skull, jaws and a well
preserved, articulated postcranial skeleton (Case and Williston, 1913,
figs. 8-22).
A little-known photograph taken by Williston in 1911 in Cañon
del Cobre (Romer, 1950) shows strata of the north wall on which a
superimposed line delineates the vertebrate-producing horizon (see Lucas
et al., this volume). Although Romer did not mention the material Williston
collected from the north wall, it most likely included the D. zenos skel
eton. Case and Williston (1913, p. 17) stated it was collected from
“reddish, clayey sandstone somewhat below the middle of the Permo
Carboniferous strata of El Cobre Canyon,” which is consistent with the
line in the photograph, and obviously not from the canyon floor. Thus,
D. zenos is a likely member of the Coyotean-age tetrapod assemblage
from Cañon del Cobre.
Long considered to be endemic to Cañon del Cobre, D. zenos was
recently recognized from a second locality in the Upper Pennsylvanian
(Virgilian) Ada Formation of Oklahoma (Kissel and Lehman, 2002). The
material includes a left dentary and two dorsal vertebrae that display
several characters diagnostic of Diasparactus.
Desmatodon aff. D. hollandi Case, 1908
The holotype of the type species of Desmatodon, D. hollandi,
was described by Case (1908) on the basis of a tooth-bearing fragment of
a left maxilla from Pitcairn, near Pittsburgh, Pennsylvania. Fracasso (1980)
reported a partial jaw and cranial fragments of the primitive diadectomorph
 119

FIGURE 3. Selected vertebrate fossils from the El Cobre Canyon Formation in Cañon del Cobre. A-D, Sphenacodon sp., NMMNH P-46534. A-D,
complete left femur in A, anterior, B, posterior, C, proximal and D, distal views. E-G, Ophiacodon sp., NMMNH P-58279, distal right femur in E,
anterior, F, posterior and G, distal views. H-I, Edaphosaurus sp., NMMNH P-43101, neural spine fragment in H, anteroposterior and I, lateral views. J,
Sagenodus sp., NMMNH P-58277, incomplete tooth plate in occlusal view. K-M, Desmatodon aff. D. hollandi, YPM 7397. K-M, incomplete maxilla
with natural mold; K, natural mold of “cheek” teeth, L, a pair of anterior teeth, and M, rubber cast of natural mold. Upper left scale applies to A-G, middle
scale applies to H-I, and lower right scale applies to J.
120
Desmatodon aff. D. hollandi (Lucas et al., 2005c, fig. 5F-J) (Fig. 3K-M)
from strata of the canyon floor that yield a Cobrean age assemblage.
These specimens were also described by Berman and Sumida (1995),
who noted that, whereas the size, spacing and cusp development of the
cheek teeth are consistent with those of D. hollandi, the size and mor
phology of the anterior two maxillary teeth more closely match those of
Diadectes lentus. However, they tentatively concurred with Fracasso
(1980) in the assignment of the specimen to Desmatodon aff. D. hollandi.
Limnoscelis paludis Williston, 1911b
Perhaps the best known taxon of the Cañon del Cobre Carbonifer
ous-Permian vertebrate assemblages is Limnoscelis paludis, long consid
ered the oldest reptile, but more recently reassigned to the order
Diadectomorpha (Heaton, 1980). Originally described by Williston
(1911a, b, 1912) and later redescribed by Romer (1946), the holotype
material (YPM 811) consists of an articulated, nearly complete skeleton
(Williston, 1911a, figs. 1-6; 1911b, figs. 3-16; 1912, figs. 10-11, figs. 27-
29; Romer, 1946, figs. 1-4, 5C, 6A, 7A, 8A, 10). Two articulated post
cranial skeletons (MCZ 1947 and MCZ 1948, formerly YPM 819 and
YPM 809, respectively), one of which is complete except for the left and
right manus, were recovered from the same locality (Williston, 1911a,
fig. 7; 1912, figs. 12-14, figs. 24-26, fig. 30). Williston (1912, figs. 1-9,
figs.15-23, fig. 31) reported additional articulated and associated material
with features diagnostic of L. paludis. Vaughn (1963) also collected
Limnoscelis material from Cañon del Cobre that includes one nearly
complete vertebra and several vertebral fragments (Lucas et al. 2005c,
fig. 3E-F).
Fracasso (1980) reported through personal communication with
P. P. Vaughn that material identified as Limnoscelis cf. L. paludis had
recently been discovered in the “Missourian” interval of the Sangre de
Cristo Formation near Howard, Colorado. Consequently, he removed L.
paludis from the list of taxa endemic to the Cañon del Cobre both at the
generic and specific levels. However, the material from Howard, Colo
rado has since been identified as a new limnoscelid species, L. dynatis
(Berman and Sumida, 1990), which is distinguished from L. paludis by
10 characters, most of which suggest L. paludis is much more derived
relative to L. dynatis. Thus, only the Cañon del Cobre locality has yielded
material attributable to the species L. paludis (Berman and Sumida, 1990).
Chamasaurus dolichognathus Williston, 1915
It should be noted that the holotype and only known specimen of
the possible captorhinomorph Chamasaurus dolichognathus described
by Williston (1915) and consisting of the anterior portion of a very small
left mandible, has been lost. Because the identity of this specimen cannot
be confirmed, we have removed this taxon from the Cañon del Cobre
assemblage list (Table 1).
Aerosaurus greenleeorum Romer, 1937
Romer (1937) and Romer and Price (1940, figs. 22,56) described
the varanopid eupelycosaur Aerosaurus greenleeorum from an articu
lated portion of the occipital region of the skull and a fragmentary post
cranial skeleton. A. greenleeorum was long assumed to occur at a second
locality (Arroyo del Agua) outside of Cañon del Cobre (Langston, 1953;
Vaughn, 1963; Fracasso, 1980). However, the Arroyo del Agua material,
which consists of two nearly complete skeletons, was distinguished
specifically from A. greenleeorum as A. wellesi (Langston and Reisz,
1981). Therefore, A. greenleeorum is endemic to Cañon del Cobre at the
specific level.
Nitosaurus jacksonorum Romer, 1937
The validity of the type specimen of the edaphosaurid Nitosaurus
jacksonorum, briefly described by Romer (1937) and redescribed by
Romer and Price (1940, figs. 24, 28, 32, 33, 37, 70), was questioned by
Langston and Reisz (1981) and Reisz (1986). They noted a discrepancy
in size between the very small cranial fragments, which resemble the
eothyridid Oedaleops from the Arroyo del Agua locality, and the much
larger postcranial elements that are consistent in size with the varanopid
Aerosaurus. Also noted as problematic is the completely unassociated
nature of the skeletal elements, strongly suggesting the holotype of
Nitosaurus is a composite. Consequently, we regard Nitosaurus as a
dubious taxon and do not include it in the Cañon del Cobre vertebrate
assemblage list.
Ruthiromia elcobriensis Eberth and Brinkman, 1983
The only vertebrate fossils known to have been recovered from
the western wall of Cañon del Cobre (exact locality unknown) pertain to
the varanopid eupelycosaur Ruthiromia elcobriensis. Collected in 1965
by A. Lewis and S. Olsen, and later described by Eberth and Brinkman
(1983, figs. 2-15), the holotype (MCZ 3150) consists of an incomplete,
articulated vertebral column, pelvis and right hind limb, as well as several
cranial fragments. Spielmann and Lucas (this volume) redescribe the
holotype of R. elcobriensis, assign it to the Ophiacodontidae and suggest
that its holotype came from the Arroyo del Agua Formation, so it is of
Seymouran age (Table 1).
Ophiacodon navajovicus Case, 1907
The numerical dominance of fossils pertaining to the ophiacodontid
Ophiacodon navajovicus in strata of the canyon floor at Cañon del
Cobre led Romer and Price (1940, p. 234) to speculate that “O.
navajovicus was the most common pelycosaur in its area at the time of
deposition of the El Cobre beds.” Its complicated taxonomic history
began with its description by Case (1907, pl. 5, fig. 6; pl. 27, figs. 1-5, 7)
as a species of the genus Dimetrodon, followed by transfer to the genus
Arribasaurus (Williston, 1914, fig. 11, nos. 1-14), culminating in reas
signment to the genus Ophiacodon under the new combination Ophiacodon
navajovicus (Romer and Price, 1940, figs. 22, 25, 30, 33, 34, 36, 38-40).
Although Reisz (1986) noted that known specimens of Ophiacodon
navajovicus are restricted to Cañon del Cobre, Vaughn (1962) reported
the occurrence of Ophiacodon cf. O. navajovicus in the Halgaito Forma
tion of the Cutler Group in San Juan County, Utah. He noted that the
vertebrae are indistinguishable from those of O. navajovicus, and that,
among other characters, the centra display the unusual flattened strip or
double keeled structure characteristic of this species. Another likely
occurrence of O. navajovicus outside of Cañon del Cobre, in the Cutler
Group of southwestern Colorado, was reported by Lewis and Vaughn
(1965). They described and illustrated an articulated string offive small
vertebrae with centra that have a mid-ventral keel characteristic of O.
navajovicus.
In the summer of 2004, a NMMNH field party collected numer
ous associated axial and appendicular elements of a single individual that
exhibit characters diagnostic of Ophiacodon navajovicus. These elements,
some of which were briefly mentioned and illustrated by Lucas et al.
(2005c, fig. 6A-K), are described in this volume by Harris et al.
Baldwinonus trux Romer and Price, 1940
The validity of the ophiacodontid Baldwinonus trux, described by
Romer and Price (1940, fig. 54), was questioned by Reisz (1980) and
Langston and Reisz (1981) based on their assumption that the holotype
material (AMNH 4780) represents two different pelycosaurs, a
sphenacodontid (maxilla and neural spines) and an ophiacodontid (verte
brae). However, restudy and further preparation of the specimen by
Brinkman and Eberth (1986) showed that one of the neural spines fits
together with one of the centra, demonstrating that the material does
indeed represent a single taxon. They concluded that Baldwinonus is a
valid genus, morphologically similar to Stereophallodon, and that both
genera are primitive members of a clade that includes Ophiacodon.
It should be noted that Romer (1952) described a partial maxilla
from the Lower Permian interval of the Dunkard Group of Ohio as

Baldwinonus? dunkardensis. However, because of the very tentative
referral of this specimen to Baldwinonus, the occurrence of the genus
outside of Cañon del Cobre was never widely accepted. Hence, most
workers (Langston, 1953; Vaughn, 1963; Hunt and Lucas, 1992; Eberth
and Berman, 1993) have regarded Baldwinonus as endemic to Cañon del
Cobre.
Edaphosaurus novomexicanus Williston and Case, 1913
Edaphosaurus novomexicanus was described as a new species by
Williston and Case (1913a) on the basis of a skull and incomplete, articu
lated postcranial elements from the nearby locality of Arroyo del Agua.
Its presence in the lower vertebrate assemblage of the canyon floor of
Cañon del Cobre was first recognized by Vaughn (1963) when he col
lected several incomplete neural spines with characteristic bilateral tu
bercles. The longest of these neural spine fragments (CM 47800, for
merly UCLA VP1641) was illustrated by Lucas et al. (2005c, fig. 5E).
Sphenacodon ferox Marsh, 1878
Marsh (1878) described the sphenacodontid Sphenacodon ferox
on the basis of a left dentary collected from the Baldwin bonebed locality
near Arroyo del Agua. Williston (1911b, plate 34, fig. 1) redescribed and
illustrated the holotype specimen in addition to establishing two lecto
types based on left and right maxillae that pertain to the same individual.
A rich concentration of fossil bones, stratigraphically high in the
Arroyo del Agua Formation (Lower Permian, Seymouran LVF) of the
eastern wall of Cañon del Cobre was first reported by Berman (1993)
and Eberth and Berman (1993). Although the bonebed yields several
taxa, including Platyhystrix and Diadectes, a preponderance of material
pertains to Sphenacodon ferox, of which the greatest proportion is disar
ticulated cranial and mandible elements (Lucas et al., 2005c, fig. 4A-G;
Rinehart et al. 2007; Spielmann et al., this volume).
A complete femur diagnostic of the genus Sphenacodon (NMMNH
P-46534, Fig. 3A-D) (the distal portion of the specimen was illustrated
by Lucas et al., 2005c, fig. 7H-I) was collected from strata in the upper
third of the El Cobre Canyon Formation approximately 60 m below the
overlying contact with the Arroyo del Agua Formation (Fig. 2). This site
represents the lowest known interval in the El Cobre Canyon Formation
section to yield Sphenacodon fossils.
Although Fracasso (1980) assumed that Sphenacodon material
collected by Baldwin was from the floor of the canyon, Baldwin’s field
notes regarding the collection of this material, as recounted by Romer and
Price (1940), state that the Sphenacodon fossils were collected from the
“highest horizon up to date in El Cobre.” Therefore, Baldwin’s
Sphenacodon site is significantly higher in the section than the “classic”
sites on the canyon floor, and may be as stratigraphically high as the
Sphenacodon site in the overlying Arroyo del Agua Formation men
tioned above. To date, the oldest known occurrence of the genus
Sphenacodon is from the Virgilian Red Tanks Member of the Bursum
Formation in south-central New Mexico (Harris et al., 2004).
Summary
Three vertebrate-fossil taxa (Anconastes vesperus, Baldwinonus
trux and Ruthiromia elcobriensis) are endemic to Cañon del Cobre at the
genus level, whereas two taxa (Limnoscelis paludis and Aerosaurus
greenleeorum) are endemic at the species level. Five taxa from Cañon del
Cobre, Chamasaurus dolicognathus (sic), Limnoscelis paludis,
Diasparactus xenos (sic), Baldwinonus trux and Nitosaurus jacksonorum,
were listed by Langston (1953) as taxa “unknown elsewhere.” Vaughn
(1963) concurred, listing the same five taxa as “peculiar to El Cobre
Canyon.” However, mostly as a result of putative occurrences of Cañon
del Cobre taxa in other localities, some variation in the list of taxa consid-
ered endemic to Cañon del Cobre is found among different authors.
121
VERTEBRATE BIOSTRATIGRAPHYAND BIOCHRONOLOGY
Vertebrate biostratigraphy and biochronology of the Lower Per
mian was long based on the classic redbed section in north-central Texas
(see review by Lucas, 1998, 2002, 2005, 2006). Because very few Upper
Pennsylvanian vertebrates are known from Texas, understanding of the
biostratigraphic succession of vertebrates across the Pennsylvanian-Per
mian boundary in North America long relied on comparing disparate
Pennsylvanian sites (such as Howard in Colorado or Linton in Ohio) to
the Early Permian record in Texas or elsewhere. Cañon del Cobre is the
only location we know of in North America where substantial assem
blages of vertebrate fossils closely bracket the Pennsylvanian-Permian
boundary (Fig. 2). We base this conclusion on several lines of evidence
that support placement of the Pennsylvanian-Permian boundary in the
upper part of the El Cobre Canyon Formation at Cañon del Cobre:
1. Megafossil plants from the lower part of the exposed section of
El Cobre Canyon Formation in the Cañon del Cobre are an Alethopteris-
dominated paleoflora of Late Pennsylvanian age (Fig.2). Although previ
ous workers disagreed over a precise age in the Pennsylvanian
(Desmoinesian, Missourian or Virgilian: Read in Smith et al., 1961; Mamay
in Fracasso, 1980; Hunt and Lucas, 1992; DiMichele and Chaney, 2005),
in this volume DiMichele et al. argue for a Virgilian age of the extensive
megaflora they collected from the El Cobre Canyon Formation.
2. Utting and Lucas (this volume) assign a Late Pennsylvanian age
to palynomorphs from the El Cobre Canyon Formation at NMMNH
plant localities 4563 and 4564 (Fig. 2).
3. Several tetrapod taxa from the lower part of the Cañon del
Cobre section, including Desmatodon, Limnoscelis and Diasparactus,
are known elsewhere only from Pennsylvanian strata. For example, cor
relation with the Pennsylvanian portion of the Halgaito Formation based
on vertebrate taxa is well supported (Lucas, 2006). Some of the other
taxa from the lower assemblage at Cañon del Cobre, such as Edaphosaurus
novomexicanus, are more primitive than Early Permian close relatives,
suggestive of a Late Pennsylvanian age.
5. At Arroyo del Agua, about 30 km southwest of Cañon del
Cobre, extensive vertebrate fossil assemblages from the upper 80 m of
the El Cobre Canyon Formation have long been (justifiably) correlated to
the lower part of the Texas Wichita Group (primarily Archer City For
mation), which indicates a middle Wolfcampian (Early Permian) age for
the upper part of the El Cobre Canyon Formation (Langston, 1953;
Romer, 1960; Lucas et al., 2005d; Lucas, 2006).
6. A locality about 13 km northeast of Arroyo del Agua yielded six
partial to complete specimens of Seymouria sanjuanensis (Eberth and
Berman, 1983; Berman et al., 1987). This locality is stratigraphically
high in the Arroyo del Agua Formation. The LO (lowest occurrence) of
Seymouria in the Texas section is in the upper Wolfcampian “Nocona”
Formation (= upper Archer City Formation: Lucas, 2006) (Hook, 1989),
so this suggests an age of late Wolfcampian for part of the Arroyo del
Agua Formation.
It is thus reasonable to conclude that the Virgilian-Wolfcampian
boundary is stratigraphically high in the El Cobre Canyon Formation in
Cañon del Cobre, though the exact position of the boundary is not certain
(Fig. 4). A detailed vertebrate biostratigraphy in Cañon del Cobre can be
constructed from all localities that can be placed into precise
lithostratigraphic position (Fig. 2) and suggests the presence of three
temporally successive and distinct vertebrate fossil assemblages (Fig. 4):
1. A lower assemblage of the canyon floor and lower canyon walls,
which occurs through a stratigraphic thickness of about 150 m. This
assemblage yields the characteristically Pennsylvanian tetrapod fossils
in association with Pennsylvanian plants and palynomorphs. This is the
characteristic vertebrate assemblage of the Cobrean LVF of Lucas (2006).
2. A middle assemblage from the LO of Sphenacodon to about the
top of the El Cobre Canyon Formation, a stratigraphic interval about 60
m thick. By correlation to the Arroyo del Agua bonebeds, this assem
122
FIGURE 4. Summary of vertebrate biostratigraphy and biochronology across
the Pennsylvanian-Permian boundary in Cañon del Cobre.
blage is of Coyotean age and most likely of early-middle Wolfcampian
age, and encompasses the Virgilian-Wolfcampian boundary.
3. A third assemblage is the single, Sphenacodon-dominated local
ity in the upper part of the Arroyo del Agua Formation. By correlation to
Arroyo del Agua, where Seymouria occurs in a similar stratigraphic po
sition, this uppermost assemblage at Cañon del Cobre is of Seymouran
age, approximately late Wolfcampian.
It is extremely difficult to correlate vertebrate fossils in the wholly
REFERENCES
Ash, S. R., 1974, Upper Triassic plants of Cañon del Cobre, New Mexico:
New Mexico Geological Society, Guidebook 25, p. 179-184.
Berman, D. S, 1993, Lower Permian vertebrate localities of New Mexico
and their assemblages: New Mexico Museum of Natural History and
Science, Bulletin 2, p. 11-21.
Berman, D. S and Lucas, S. G., 2003, Aspidosaurus binasser (Amphibia,
Temnospondyli) a new species of Dissorophidae from the Lower Per
mian of Texas: Annals of Carnegie Museum, v. 72, p. 241-262.
Berman, D. S and Sumida, S. S., 1990, A new species of Limnoscelis (Am
phibia,Diadectomorpha) from the Late Pennsylvanian Sangre de Cristo
Formation of central Colorado: Annals of Carnegie Museum, v. 59, p.
303-341.
Berman, D. S and Sumida, S. S., 1995, New cranial material of the rare
diadectid Desmatodon hesperis (Diadectomorpha) from the Late Penn
sylvanian of central Colorado: Annals of Carnegie Museum, v. 64, p.
481-499.
Berman, D. S, Reisz, R. R. and Eberth, D. A., 1987, A new genus and species
of trematopid amphibian from the Late Pennsylvanian of north-central
New Mexico: Journal of Vertebrate Paleontology, v, 7, p. 252-269.
Brinkman, D. and Eberth, D. A., 1986, The anatomy and relationships of
Stereophallodon and Baldwinonus (Reptilia, Pelycosauria): Breviora,
no. 485, p. 1-34.
Carroll, R. L., 1964, Early evolution of the dissorophid amphibians: Mu
seum of Compartive Zoology, Bulletin, v. 131, p. 161-250.
Case, E. C., 1907, Revision of the Pelycosauria of North America: Carnegie
Institution of Washington, Publication no. 55, p. 1-176.
Case, E. C., 1908, Description of vertebrate fossils from the vicinity of
Pittsburgh, Pennsylvania: Annals of the Carnegie Museum, v.4, p. 234-
241.
Case, E. C., 1910, New or little known reptiles and amphibians from the
Permian(?) of Texas: Bulletin of the American Museum of Natural
History, v. 28, p. 168-181.
Case, E. C., 1911, A revision of the Cotylosauria of North America: Carnegie
nonmarine redbeds at Cañon del Cobre to marine chronostratigraphic
concepts such as Virgilian and Wolfcampian. Instead, it is more useful to
place the three successive assemblages of vertebrate fossils at Cañon del
Cobre into the framework of vertebrate biostratigraphy/biochronology
developed by Lucas (2002, 2005, 2006). This framework indicates that
the three vertebrate fossil assemblages at Cañon del Cobre represent
three land vertebrate faunachrons (LVFs). The middle and upper assem
blages belong to the Coyotean and Seymouran LVFs of Lucas (2006),
and the lower assemblage at Cañon del Cobre represents the Cobrean
LVF, an older, entirely Pennsylvanian LVF (Lucas et al., 2005c).
In biochronological terms, the beginning of the Cobrean is the
FAD (first appearance datum) of Eryops, and its end is the beginning of
the Coyotean, which is the FAD of Sphenacodon. Index taxa of Cobrean
time include Diasparactus, Desmatodon and Limnoscelis. The LO of
Eryops is now within the Cobrean characteristic assemblage, but we
believe the actual FAD of the genus (which is known from Pennsylva
nian strataelsewhere: Case, 1908; Vaughn, 1958; Milner, 1996; Harris et
al., 2004) may be older than its currently known LO in Cañon del Cobre.
The base of the Permian is now within the lower Wolfcampian, so the
Permian base is within the Coyotean LVF (Lucas, 2006). The Cañon del
Cobre tetrapod record thus provides a standard of three, stratigraphically
successive vertebrate assemblages (and their corresponding time-succes
sive LVFs) that encompass the Pennsylvanian-Permian boundary (Fig.
4).
ACKNOWLEDGMENTS
The U. S. Forest Service permitted work in Cañon del Cobre.
Numerous colleagues and volunteers assisted in the fieldwork. Adam
Huttenlocker, Joerg Schneider and Ralf Werneburg reviewed the manu
script.
Institution of Washington, Publication no. 145, p. 1-121.
Case, E. C. and Williston, S. W., 1913, Description of a nearly complete
skeleton of Diasparactus zenos Case: Carnegie Institution of Washing
ton, Publication no. 181, p. 17-35.
Cope, E. D., 1878, Descriptions of extinct Batrachia and Reptilia from the
Permian Formation of Texas: Proceedings of the American Philosophi
cal Society, v. 17, p. 505-530.
DiMichele, W. A. and Chaney, D. S., 2005, Pennsylvanian-Permian fossil
floras from two areas of northern New Mexico: New Mexico Museum of
Natural History and Science, Bulletin 31, p. 26-33.
Eberth, D. A. and Berman, D. S, 1983, Sedimentology and paleontology of
Lower Permian fluvial redbeds of north-central New Mexico – prelimi
nary report: New Mexico Geology, v. 5, p. 21-25.
Eberth, D. S. and Berman D. S, 1993, Stratigraphy, sedimentology and
vertebrate paleontology of the Cutler Formation redbeds (Pennsylva
nian-Permian) of north-central New Mexico: New Mexico Museum of
Natural History and Science, Bulletin 2, p. 33-48.
Eberth, D. S. and Brinkman D., 1983, Ruthiromia elcobriensis, a new pely
cosaur from El Cobre Canyon, New Mexico: Breviora, no. 474, p. 1-26.
Eberth, D. S. and Miall, A. D., 1991, Stratigraphy, sedimentology and evo
lution of a vertebrate-bearing, braided to anastomosed fluvial system,
Cutler Formation (Permian-Pennsylvanian), north-central New Mexico:
Sedimentary Geology, v. 72, p. 225-252.
Fracasso, M. A., 1980, Age of the Permo-Carboniferous Cutler Formation
vertebrate fauna from El Cobre Canyon, New Mexico: Journal of Pale
ontology, v. 54, p. 1237-1244.
Fracasso, M.A., 1987, Braincase of Limnoscelis paludis Williston: Postilla,
v. 201, p. 1-22.
Harris, S. K., Lucas, S. G., Berman, D. S and Henrici, A. C., 2004, Vertebrate
fossil assemblage from the Upper Pennsylvanian Red Tanks Member of
the Bursum Formation, Lucero uplift, central New Mexico: New Mexico
Museum of Natural History and Science, Bulletin 25, p. 267-283.
Heaton, M. J., 1980, The Cotylosauria: A reconsideration of a group of

archaic tetrapods; in Panchen, A. L., eds., The terrestrial environment
and the origin of land vertebrates: New York, Academic Press, p. 497
551.
Hook, R. W. 1989. Stratigraphic distribution of tetrapods in the Bowie and
Wichita Groups, Permo-Carboniferous of north-central Texas; in Hook,
R.W., ed., Permo-Carboniferous vertebrate paleontology, litho
stratigraphy and depositional environments of north-central Texas:
Austin, Society of Vertebrate Paleontology, p. 47-53.
Hunt, A. P. and Lucas, S. G., 1992, The paleoflora of the lower Cutler
Formation (Pennsylvanian, Desmoinesian?) in El Cobre Canyon, New
Mexico, and its biochronological significance: New Mexico Geological
Society, Guidebook 43, p. 145-150.
Kemp, T. S., 1980, Origin of mammal-like reptiles: Nature, v. 5745, p. 378
380.
Kempter, K., Zeigler, K., Koning, D. and Lucas, S., 2007, Preliminary
geologic map of the Canjilon SE quadrangle, Rio Arriba County, New
Mexico: New Mexico Bureau of Geology and Mineral Resources, Open
file Geological Map 150, 1:24,000.
Kissel, R. A. and Lehman, T. M., 2002, Upper Pennsylvanian tetrapods
from the Ada Formation of Seminole County, Oklahoma: Journal of
Paleontology, v. 76, p. 529-545.
Langston, W. Jr., 1953, Permian amphibians from New Mexico: University
of California, Publications in Geological Sciences, v. 29, p. 349-416.
Langston, W, Jr. and Reisz, R. R., 1981, Aerosaurus wellesi, new species, a
varanopseid mammal-like reptile (Synapsida: Pelycosauria) from the
Lower Permian of New Mexico: Journal of Vertebrate Paleontology, v.
1, p. 73-96.
Lewis, G. E. and Vaughn, P. P., 1965, Early Permian vertebrates from the
Cutler Formation of the Placerville area, Colorado: U. S. Geological
Survey, Professional Paper,503-C, 50 p.
Lucas, S. G., 1998, Toward a tetrapod biochronology of the Permian: New
Mexico Museum of Natural History and Science, Bulletin 12, p. 71-91.
Lucas, S. G., 2002, Tetrapods and the subdivision of Permian time; in Hills,
L.V., Henderson, C. M., and Bamber, E. W., eds., Carboniferous and
Permian of the world: Canadian Society of Petroleum Geologist Memoir
19, p. 479-491.
Lucas, S.G., 2005, Permian tetrapod faunachrons: New Mexico Museum of
Natural History and Science, Bulletin 30, p. 197-201.
Lucas, S.G., 2006, Global Permian tetrapod biostratigraphy and
biochronology; in Lucas, S. G., Cassinis, G. and Schneider, J. W., eds.,
Non-marine Permian biostratigraphy and biochronology: Geological
Society, London, Special Publication 265, p. 65-93.
Lucas, S.G., and Hunt, A.P., 1992, Triassic stratigraphy and paleontology,
Chama Basin and adjacent areas, north-central New Mexico: New Mexico
Geological Society, Guidebook 43, p. 151-167
Lucas, S.G. and Krainer, K., 2005a, Stratigraphy and correlation of the
Permo-Carboniferous Cutler Group, Chama Basin, New Mexico: New
Mexico Geological Society, Guidebook 56, p. 145-159.
Lucas, S.G. and Krainer, K., 2005b, Cutler Group (Permo-Carboniferous)
stratigraphy, Chama basin, New Mexico: New Mexico Museum of Natu
ral History and Science, Bulletin 31, p. 90-100.
Lucas, S. G., Rinehart, L. F. and Hunt, A. P., 2005a, Eryops (Amphibia,
Temnospondyli) from the Upper Pennsylvanian of El Cobre Canyon,
New Mexico: New Mexico Museum of Natural History and Science,
Bulletin 31, p. 118-120.
Lucas, S. G., Zeigler, K. E., Heckert, A. B., and Hunt, A. P., 2003, Upper
Triassic stratigraphy and biostratigraphy, Chama Basin, north-central
New Mexico: New Mexico Museum of Natural History and Science
Bulletin, v. 24, p. 15-39.
Lucas, S. G., Zeigler, K. E., Heckert, A. B. and Hunt, A. P., 2005b, Review of
Upper Triassic stratigraphy and biostratigraphy in the Chama basin,
northern New Mexico: New Mexico Geological Society, Guidebook 56,
p. 170-181.
Lucas, S.G., Harris, S.K., Spielmann, J.A., Berman, D.S and Henrici, A.C.,
2005c, Vertebrate biostratigraphy and biochronology of the Pennsylva
nian-Permian Cutler Group, El Cobre Canyon, northern New Mexico:
New Mexico Museum of Natural History and Science, Bulletin 31, p.
128-139.
Lucas, S. G., Harris, S. K., Spielmann, J. A., Berman, D. S, Henrici, A. C.,
Heckert, A. B., Zeigler, K. E. and Rinehart, L. F., 2005d, Early Permian
vertebrate assemblage and its biostratigraphic significance, Arroyo del
123
Agua, Rio Arriba County, New Mexico: New Mexico Geological Society,
Guidebook 56, p. 288-296.
Lucas, S. G., Lerner, A. J, Hannibal, J. T., Hunt, A. P. and Schneider, J. W.,
2005e, Trackway of a giant Arthropleura from the Upper Pennsylva
nian of El Cobre Canyon, New Mexico: New Mexico Geological Society,
Guidebook 56, p. 279-282.
Marsh, O. C., 1878, Notice of new fossil reptiles: American Journal of
Science, v. 15, p. 409-411.
Mehl, M. G., 1913, A description of Chenoprosopus milleri, gen. et sp. nov:
Publications of the Carnegie Institute of Washington, v. 181, p. 11-16.
Milner, A. R., 1996, Systematics of the genus Eryops (Amphibia:
Temnospondyli) and its possible biostratigraphical value: Journal of
Vertebrate Paleontology, v. 16, supplement to no. 3, p. 53A.
Reisz, R. R., 1980, The Pelycosauria: a review of phylogenetic relation
ships; in Panchen, A. L., ed., The terrestrial environment and the origin
of land vertebrates: London and New York, Academic Press, p. 553-592.
Reisz, R. R., 1986, Pelycosauria: Encyclopedia of Paleoherpetology, Part
17A, 102 p.
Rinehart, L. F., Lucas, S. G. and Harris, S. K., 2007, Lithology and taphonomy
of an Early Permian Sphenacodon bonebed in Cañon del Cobre, north
central New Mexico: New Mexico Geology, v. 29, p. 63.
Romer, A. S., 1937, New genera and species of pelycosaurian reptiles:
Proceedings of the New England Zoological Club, v. 16, p. 89-96.
Romer, A. S., 1946, The primitive reptile Limnoscelis restudied: American
Journal of Science, v. 244, p. 149-188.
Romer, A. S., 1950, The upper Paleozoic Abo Formation and its vertebrate
fauna; in Colbert, E.H. and Northrop, S.A., eds., Guidebook for the 4th
Field Conferenceof the Society of Vertebrate Paleontology in north
western New Mexico: New York, American Museum of Natural History,
p. 48-55.
Romer, A. S., 1952, Late Pennsylvanian and Early Permian vertebrates of
the Pittsburgh-West Virginia region: Annals of Carnegie Museum, v. 33,
p. 47-112.
Romer, A. S., 1960, The vertebrate fauna of the New Mexico Permian: New
Mexico Geological Society, Guidebook 11, p. 48-54.
Romer, A. S. and Price, L. I., 1940, Review of the Pelycosauria: Geological
Society of America Special Paper, v. 28, p. 1-538.
Schultze, H.-P. and Chorn, J., 1997, The Permo-Carboniferous genus
Sagenodus and the beginning of modern lungfish: Contributions to Zo
ology, v. 67, p. 9-70.
Smith, C. T., Budding, A. J. and Pitrat, C. W., 1961, Geology of the south
eastern part of the Chama Basin: New Mexico Bureau of Mines and
Mineral Resources, Bulletin 75, 57 p.
Sumida, S. S., Berman, D. S and Martens T., 1996, Biostratigraphic correla
tions between the Lower Permian of North America and central Europe
using the first record of an assemblage of terrestrial tetrapods from
Germany: PaleoBios, v. 17, p. 1-12.
Sumida, S. S., Berman, D. S and Martens T., 1998, A new trematopid am
phibian from the Lower Permian of central Germany: Palaeontology, v.
41, p. 605-629.
Vaughn, P. P., 1958, On the geologic range of the labyrinthodont amphibian
Eryops: Journal of Paleontology, v. 32, p. 918-922.
Vaughn, P. P., 1962, Vertebrates from the Halgaito Tongue of the Cutler
Formation, Permian of San Juan County, Utah: Journal of Paleontology,
v. 36, p. 529-539.
Vaughn, P. P., 1963, The age and locality of the Paleozoic vertebrates from
El Cobre Canyon, Rio Arriba County, New Mexico: Journal of Paleon
tology, v. 37, p 283-296.
Williston, S. W., 1911a, A new family of reptiles from the Permian of New
Mexico: American Journal of Science, v. 31, p. 378-398.
Williston, S. W., 1911b, American Permian vertebrates: Chicago, Univer
sity ofChicago Press, 145 p.
Williston, S.W., 1912, Restoration of Limnoscelis, a cotylosaur reptile
from New Mexico: American Journal of Science, v. 34, p. 457-468.
Williston, S. W., 1914, The osteology of some American Permian verte
brates: Journal of Geology, v. 22, p. 364-419.
Williston, S. W., 1915, New genera of Permian reptiles: American Journal
of Science, v. 39, p. 575-579.
Williston, S. W. and Case, E. C., 1913, A description of Edaphosaurus
Cope: Carnegie Institution of Washington, Publication no. 181, p. 71
81.
124
 Lucas et al., eds., 2010, Carb-Permian transition in Cañon del Cobre. New Mexico Museum of Natural History and Science Bulletin 49.

125
DISSOROPHOID RECORD FROM THE UPPER PENNSYLVANIAN
OF CAÑON DEL COBRE, NEWMEXICO

RALFWERNEBURG1, DAVID S BERMAN2 AND SPENCERG. LUCAS3

1 Naturhistorisches Museum Schloss Bertholdsburg, Burgstr. 6, D-98553 Schleusingen, Germany;

2 Carnegie Museum of Natural History, 4400 Forbes Avenue, Pittsburgh, Pennsylvania 15213;

3 New Mexico Museum of Natural History, 1801 Mountain Road N. W., Albuquerque, New Mexico 87104

Abstract—The snout region of a large dissorophoid skull (CM 76873) is described from the Upper Pennsylvanian
El Cobre Canyon Formation of Cañon del Cobre, north-central New Mexico. The following characters are variably
attributable to a wide variety of dissorophoids: 1) large size (skull length about 20 cm); 2) premaxilla massive with
large dorsal exposure; 3) large internarial fontanel; 4) nasals extend well beyond the level of the anterior margin of
the external nares; and 5) lacrimal has a wide marginal entrance into the septomaxilla-naris-complex. Features that
appear characteristic of CM 76873 include: 1) broadly convex snout; 2) a radial pattern of ornamentation on the
dorsal roofing bones; 3) an oval, relatively short naris; 4) paired vomerine depressions; 5) small vomerine fangs; 6)
a narrow, shortanterior embayment of the choana lateral to the paired vomerine fangs; and 7) relatively short, wide
choanae.

INTRODUCTION
Cañon del Cobre has been an important and prolific source of
vertebrates, including the lungfishes Gnathorhiza and Sagenodus, six
taxa of temnospondyl amphibians, two embolomerous anthracosaurs,
three diadectomorphs, a probable captorhinomorph reptile, and seven
pelycosaurian-grade synapsids (e.g., Eberth and Brinkman, 1983; Berman,
1993; Lucas et al., 2005, this volume). In 1983, two specimens of
trematopid, including the greater portion of a skeleton with skull and a
partial skull and portion of the postcranium, were collected from the
Upper Pennsylvanian strata in Cañon del Cobre and were subsequently
described as a new genus and species, Anconastes vesperus (Berman et
al., 1987). Closely associated with these specimens and from the same
bed was a poorly-preserved snout portion of a moderate-sized skull
(CM 76873), which, despite extensive preparation, could not be as
signed to an existing taxon. Numerous poorly preserved, fragmentary
bones were found associated with the snout, but offered no clues to its
identity. Although the snout obviously represents a new taxon, at least
for the Cañon del Cobre locality, there was a reluctance to describe it,
with the hope that future discoveries would clarify its identity. Whereas
this has not happened, the specimen seems important enough to warrant
description.
Institutional Abbreviations: CM = Carnegie Museum of Natu
ral History, Pittsburgh; FMNH = Field Museum of Natural History,
Chicago; MCZ = Museum of Comparative Zoology, Harvard Univer
sity, Cambridge; NMMNH = New Mexico Museum of Natural History,
Albuquerque; SMNS = Staatliches Museum für Naturkunde, Stuttgart;
TUBAF=Technische Universität Bergakademie Freiberg; UBA= Uni
versity of Buenos Aires; UW = University of Washington.
DESCRIPTION (FIGS. 1-2)
The anterior snout region of CM 76873, which is exposed in both
dorsal and palatal views, has a maximum length of 50 mm and a maximum
width of 62 mm. These proportions suggest an original skull length of
about 180-200 mm. Although the skull surfaces exhibit extensive weath
ering, the dorsal roofing bones present an ornamentation pattern of radi
ating ridges or striations. This is unusual for most adult temnospondyl
skulls of comparable size, in which the sculpturing pattern typically
consists of low ridges surrounding small, subcircular depressions (Fig.
3). A large internarial fontanel near the tip of the snout is positioned on
the midline suture of the premaxillae and nasals. The premaxilla is mas
sive and has a relatively large dorsal exposure. The posterior extent of the
alary processes bordering the anterodorsal margin of the external naris
has obviously been shortened by weathering. The only indication of
their truncation is an impression on the nasal. The nasals narrow slightly
as they extend deeply between the alary processes of the premaxillae and
well beyond the level of the anterior margin of the external nares to form
the posterior half of the internarial fontanel. Only the anterior half of the
right lacrimal is preserved, which ends anteriorly in a wide marginal
entrance into the septomaxilla-naris-complex. The naris is broad but
elongated slightly anteroposteriorly, giving it an oval outline.
In the ventral view of the anterior snout region of CM 76873, the
premaxilla forms a narrow, ventral rim of the skull, which is occupied
mainly by the alveolar shelf. The shelf is presumably continued posteri
orly by the maxilla, and the broad area between the lateral shelves is
occupied by the vomers. Paired vomerine depressions, narrowly sepa
rated from one another at the midline, lie anterior and posterior to the
vomer-premaxilla contact. Small, paired fangs are located in circular de
pressions at the anteromedial corner of the choana. Lateral to the paired
fangs there is a narrow, short anterior embayment of the choana.
Between the paired fangs and the embayment of the right vomer is a
small cluster of denticles. A pronounced, rounded ridge without teeth or
denticles borders the medial margin of the choana. The choana appears to
be short and wide, with a width equalling about one fourth the length.
DISCUSSION
The snout fragment, CM 76873, from Cañon del Cobre exhibits a
mosaic of features that are either variably attributable to a wide variety
of dissorophoids and temnospondyls of close affinity (e.g., zatrachydids
and dvinosaurs) or are otherwise unique. Shared features most promi
nent among the former group include: 1) large size (skull length about 20
cm); 2) premaxillae massive and have a relatively large dorsal exposure;
3) large internarial fontanel; 4) nasals extending well beyond the level of
the anterior margin of the external nares; and 5) lacrimal has a wide
marginal entrance into the septomaxilla-naris-complex.
The anterior skull fragment of the Cañon del Cobre specimen CM
76873 is compared with different temnospondyls in dorsal view (Fig.
3A-H). It differs strongly from the amphibamid, micromelerpetontid,
and branchiosaurid dissorophoids in its much greater size, but is well
within the maximum ranges of trematopids and dissorophids. The mas
sive and relatively large dorsal exposure of the premaxilla is a feature seen
only in the trematopids Phonerpeton (Dilkes, 1990) and Acheloma (Olson,
1941; Bolt, 1974; Dilkes and Reisz, 1987). Furthermore, in CM 76873
the nasals extend well beyond the level of the anterior margin of the
external nares, a feature shared in the same two trematopids (Olson,
126

FIGURE 1. Snout of partial dissorophoid skull, CM 76873, from the Upper Pennsylvanian of the El Cobre Canyon Formation in Cañon del Cobre, New
Mexico, dorsal view. Abbreviations: inf, internarial fontanel; l, lachrymal; mx, maxilla; n, nasal; na, naris; pm, premaxilla; sm, septomaxilla.

FIGURE 2. Snout of partial dissorophoid skull, CM 76873, from the Upper Pennsylvanian of the El Cobre Canyon Formation in Cañon del Cobre, New
Mexico, palatal view. Abbreviations: ch, choane; mx, maxilla; pm, premaxilla; v, vomer; vd, vomerine depression.

1941; Bolt, 1974; Dilkes and Reisz, 1987; Dilkes, 1990). The possession
of an internarial fontanel in CM 76873 also occurs in many dissorophoids,
including: amphibamids, such as Doleserpeton (Bolt, 1969), Eoscopus
(Daly, 1994; Werneburg, 1993), Tersomius (Carroll, 1964), Plemmyradytes
(Huttenlocker et al., 2007) and Georgenthalia (Anderson et al., 2008);
branchiosaurids and some micromelerpetontids (Boy, 1972; Werneburg,
1989; Schoch and Milner, 2008); some trematopids, such as Phonerpeton
(Dilkes, 1990) and Acheloma (Olson, 1941; Bolt, 1974; Dilkes and Reisz,
1987); and some dissorophids such as Conjunctio (Carroll, 1964) and
Aspidosaurus (Berman and Lucas, 2003). But, in contrast to these ex
amples, the internarial opening in CM 76873 is much larger. An internarial
fontanel is absent in eryopids and Parioxys. An “internarial pit” was
described in Parioxys by Moustafa (1955), but it is not clearly defined
and appears to be sculptured, possibly due to weathering. In CM 76873
the presence of a wide marginal entrance of the lacrimal into the
septomaxilla-naris-complex is shared by members of all the dissorophoid
families, as well as by intasuchids, eryopids, and Parioxys.
On the basis of the same comparative taxa as above, those features
which appear characteristic to CM 76873 include: 1) broadly convex
snout, as in Aspidosaurus; 2) a radial pattern of ornamentation on the
dorsal roofing bones; 3) an oval, relatively short naris; 4) paired vomerine
depressions; 5) small vomerine fangs; 6) a narrow, short anterior
embayment of the choana lateral to the paired vomerine fangs; and 7)
relatively short, wide choanae. The broadly convex snout in CM 76873
appears to set it apart from nearly all dissorophoids. Moustafa’s (1955)
reconstruction of the skull in Parioxys depicts the proportions of the
snout as appearing to be very similar to those in CM 76873. This is in
sharp contrast, however, to the Parioxys ferricolus specimen MCZ 1162
shown in Figure 3D, which has a much narrower snout. Despite the
strong probability that CM 76873 represents an adult specimen because
of its large size and heavy ossification, its skull roof sculpturing is inter
esting in having a radial pattern of ridges typically seen in juvenile
temnospondyls. In contrast, in similarly sized, presumably adult skulls
of dissorophoids and Parioxys, the sculpturing pattern, although varying
in intensity among taxa, consists of a dense network of pits surrounded
by low, rounded ridges.
Also not seen among the dissorophoids is the relatively short,
slightly oval naris in CM 76873. The nares of Parioxys are triangular and
 127

FIGURE 3. Comparison of CM 76873 in dorsal view with selected other Permo-Carboniferous temnospondyls. A, Dissorophoid, CM 76873, from the
Upper Pennsylvanian El Cobre Canyon Formation in Cañon del Cobre, New Mexico; length of the figured snout region = 5.0 cm; B, Eryops sp., NMMNH
P-46379, from the Upper Pennsylvanian El Cobre Canyon Formation in Cañon del Cobre, New Mexico (see this volume); length of the figured snout
region = 15.9 cm; C, Anconastes vesperus, CM 41711, from the Upper Pennsylvanian El Cobre Canyon Formation in Cañon del Cobre, New Mexico
(Berman et al., 1987); length of the figured snout region = 4.1 cm; D, Parioxys ferricolus, MCZ 1162; length of the figured snout region = 7.4 cm; E,
Acheloma cuminsi, UC 481; length of the figured snout region = 7.7 cm; F, Tersomius texensis, MCZ 1912; length of the figured snout region = 2.2 cm;
G, Conjunctio multidens, UC 673; length of the figured snout region = 4.3 cm; H, Dissorophus multicinctus, MCZ 4168; length of the figured snout region
= 5.9 cm.
128
closer together than in CM 76873. The paired vomerine depressions are
also considered unique to CM 76873. Although vomerine depressions
are sometimes observed in dissorophoids (as well as zatrachydids), they
are usually present in the form of a single, deep depression on the midline
of the vomers, as in Amphibamus (Milner, 1982; Schoch, 2001), Tersomius
(T. texensis: MCZ 1912, personal observation of RW), Micropholis (Boy,
1985), Pasawiops (Fröbisch and Reisz, 2008) and trematopids such as
Acheloma (Olson, 1941; Bolt, 1974) and Phonerpeton (Dilkes, 1990).
The small vomerine fangs in CM 76873 are apparently also absent in
some dissorophoids, Parioxys, and eryopids. Also, in contrast to the
latter group, in CM 76873 the choanae are relatively short and wide. The
paired vomerine fangs produce a narrow, short anterior embayment of
the choana, which is similar to broad-skulled morphs of Micropholis (see
Schoch and Rubidge, 2005, figs. 1D, 2D).
Given the present knowledge of CM 76873, it can be tentatively
considered a member of the Dissorophoidea but without familial or lower
level taxonomic assignment. This specimen may extend the knowledge of
the Pennsylvanian dissorophoids, which had already diversified into the
trematopids, dissorophids, amphibamids, micromelerpetontids and
branchiosaurids.
From the stratigraphic point of view all Pennsylvanian
dissorophoids are interesting for further comparison. Beside the small
REFERENCES
Anderson, J. A., Henrici, A. C., Sumida, S. S., Martens, T. and Berman, D. S.,
2008, Georgenthalia clavinasica, a new genus and species of dissorophoid
temnospondyl from the Early Permian of Germany, and the relation
ships of the Family Amphibamidae: Journal of Vertebrate Paleontology,
v. 28, p. 61-75.
Berman, D. S, 1993, Lower Permian vertebrate localities of New Mexico
and their assemblages: New Mexico Museum of Natural History and
Science, Bulletin 2, p. 11-21.
Berman, D. S, Reisz, R. R. and Fracasso, M, 1981, Skull of the Lower
Permian dissorophid amphibian Platyhystrix rugosus: Annals of Carnegie
Museum, v. 50, p. 391-416.
Berman, D. S, Reisz, R. R. and Eberth, D. A., 1987, A new genus and species
of trematopid amphibian from the Late Pennsylvanian of north-central
New Mexico: Journal of Vertebrate Paleontology, v. 7, p. 252-269.
Berman, D.S, and Lucas, S.G., 2003. Aspidosaurus binasser (Amphibia,
Temnospondyli), a new species of Dissorophidae from the Lower Per
mian of Texas: Annals of Carnegie Museum, v. 72, p. 241-262.
Bolt, J. R., 1969, Lissamphibian origins: Possible protolissamphibian from
the Lower Permian of Oklahoma: Science, v. 166, p. 888-891.
Bolt, J. R., 1974, Evolution and functional interpretation of some suture
patterns in Paleozoic labyrinthodont amphibians and lower tetrapods:
Journal of Paleontology, v. 48, p. 434-458.
Boy, J. A., 1972, Die Branchiosaurier (Amphibia) des saarpfälzischen
Rotliegenden (Perm, SW-Deutschland): Hessisches Landesamt für
Bodenforschung, v. 65, p. 1-137.
Boy, J. A., 1985, Über Micropholis, den letzten Überlebenden der
Dissorophoidea (Amphibia, Temnospondyli; Unter-Trias): Neues
Jahrbuch für Geologie und Paläontologie, Monatshefte, v.1985, p. 29
45.
Carroll, R. L., 1964, Early evolution of the dissorophoid amphibians: Bul
letin of the Museum of Comparative Zoology, Harvard University, v.
131, p. 161-250.
Daly, E., 1994, The Amphibamidae (Amphibia: Temnospondyli), with a
description of a new genus from the Upper Pennsylvanian of Kansas:
The University of Kansas, Miscellaneous Publications, v. 85, p. 1-59.
Dilkes, D.W., 1990. A new trematopsid amphibian (Temnospondyli:
Dissorophoidea) from the Lower Permian of Texas: Journal of Verte
brate Paleontology, v. 10, p. 222-243.
Dilkes, D.W., and R.R. Reisz, 1987. Trematops milleri Williston, 1909
amphibamids Amphibamus, Platyrhinops and Eoscopus, as well as the
trematopids Ecolsonia, Actiobates and Anconastes, a dissorophoid with
probable affinities to Dissorophidae, Platyhystrix, is also known from
the Pennsylvanian and was found together with Eryops in Cañon del
Cobre of New Mexico (compare Werneburg et al., in this volume). In
contrast to CM 76873, Platyhystrix rugosus from the Lower Permian of
Arroyo del Agua in New Mexico (Berman et al., 1981) has a very narrow
skull with a pointed snout region and probably lacks an internarial fon
tanel.
ACKNOWLEDGMENTS
Support for the field work that yielded the dissorophoid snout
CM 76873 was provided by a grant from the New Mexico Bureau of
Mines and Mineral Resources (to DSB). Thanks are also extended to
Amy C. Henrici (CM) and Larry Rinehart (NMMNH) for the prepara
tion of CM 76873. We thank our colleagues J.R. Bolt (FMNH), F.A.
Jenkins Jr. (MCZ), C. Schaff (MCZ), J.W. Schneider (TU BAF) and
R.R. Schoch (SMNS) for the opportunity to study specimens in their
care and for helpful references. This work was supported by a grant of
the Deutsche Forschungsgemeinschaft (to RW: DFG-We 2833/2-1, DFG
We 2833/3-1). Reviews by A. Huttenlocker (UW) and C. A. Marsicano
(UBA) improved the manuscript.
identified as a junior synonym of Acheloma cumminsi Cope, 1882, with
a revision of the genus: American Museum Novitates, no. 2902, p. 1-12.
Eberth, D. S. and Brinkman, D., 1983, Ruthiromia elcobriensis, a new
pelycosaur from El Cobre Canyon, New Mexico: Breviora, no. 474, p.
1-26.
Fröbisch, N. B. and Reisz, R. R., 2008, A new Lower Permian amphibamid
(Dissorophoidea, Temnospondyli) from the fissure fill deposits near
Richards Spur, Oklahoma: Journal of Vertebrate Paleontology, v. 28, p.
1015-1030.
Huttenlocker, A. et
shintoni, gen. K.,sp.Pardo,
nov., an
J. D.
Early
andPermian
Small, amphibamid
B. J., 2007, (Temnospondyli:
Plemmyradytes
Dissorophoidea) from the Eskridge Formation, Nebraska: Journal of
Vertebrate Paleontology, v. 27, p. 316-328.
Lucas, S. G., Rinehart, L. F. and Hunt, A. P., 2005, Eryops (Amphibia,
Temnospondyli) from the Upper Pennsylvanian of El Cobre Canyon,
New Mexico: New Mexico Museum of Natural History and Science,
Bulletin 31, p. 118-120.
Milner, A. R., 1982, Small temnospondyl amphibians from the Middle
Pennsylvanian of Illinois: Palaeontology, v. 25, p. 635-664.
Moustafa, Y. S., 1955, The affinities of Parioxys ferricolus and the phylog
eny of the “eryopsid” amphibians: Bulletin Institute d’Egypte, v. 36, p.
77-104.
Olson, E. C., 1941, The Family Trematopsidae: Journal of Geology, v. 49,
p. 149-176.
Schoch, R. R., 2001, Can metamorphosis be recognized in Palaeozoic am
phibians?: Neues Jahrbuch für Geologie und Paläontologie, Abhandlungen,
v. 220, p. 335-367.
Schoch, R. R. and Milner, A. R., 2008, The intrarelationships and evolu
tionary history of the temnospondyl family Branchiosauridae: Journal
of Systematic Palaeontology, v. 2008, p. 1-23.
Schoch, R. R. and Rubidge, B.S. 2005. The amphibamid Micropholis stowi
from the Lystrosaurus Assemblage Zone of South Africa: Journal of
Vertebrate Paleontology, v. 25, p. 502-522.
Werneburg, R., 1989, Labyrinthodontier (Amphibia) aus dem Oberkarbon
und Unterperm Mitteleuropas - Systematik, Phylogenie und
Biostratigraphie: Freiberger Forschungshefte, v. C 436, p. 7-57.
Werneburg, R., 1993, Ein Schädelrest von Eoscopus lockardi Daly (Am
phibia: Dissorophoidea) aus dem Oberkarbon von Kansas:
Veröffentlichungen Naturhistorisches Museum Schleusingen, v. 7/8, p.
147-149.
 Lucas et al., eds., 2010, Carb-Permian transition in Cañon del Cobre. New Mexico Museum of Natural History and Science Bulletin 49.

129
FIRST PENNSYLVANIAN ERYOPS (TEMNOSPONDYLI)AND
ITS PERMIAN RECORD FROM NEW MEXICO

RALFWERNEBURG1, SPENCERG. LUCAS2, JÖRGW. SCHNEIDER3 AND LARRYF.RINEHART2

1Naturhistorisches Museum Schloss Bertholdsburg, Burgstr. 6, D-98553 Schleusingen, Germany;

2 New Mexico Museum of Natural History and Science, 1801 Mountain Road NW, Albuquerque, New Mexico 87104;
3 TU Bergakademie Freiberg, Institut für Geologie, B.v.Cotta-Str. 2, D-09596 Freiberg, Germany

Abstract—A well-preserved, partial skull of the temnospondyl amphibian Eryops is described from the Pennsyl
vanian strata of the El Cobre Canyon Formation of Cañon del Cobre, Rio Arriba County, New Mexico. The
presence of the autapomorphic interfrontal presents convincing evidence of its assignment to Eryops. The two
other known Pennsylvanian eryopid skulls, Glaukerpeton avinoffi from Pittsburgh, Pennsylvania, and Eryops cf.
E. avinoffi from Sutton, West Virginia, belong to the valid, more aquatically-adapted eryopid genus Glaukerpeton.
The Pennsylvanian Eryops of New Mexico is compared with distinctly defined types of Permian Eryops, includ
ing specimens from the Lower Permian of New Mexico, but at present no species distinction is recognized. The
Lower Permian specimens of Eryops from New Mexico and Utah are tentatively assigned to E. grandis, despite
having a more slender skull with a narrow jugal and relatively fine dermal sculpturing.

INTRODUCTION
The temnospondyl Eryops, one of the best known Permo-Car
boniferous amphibian fossils, occurs at widely distributed sites in the
USA. The general morphology of the eryopids Eryops (Sawin, 1941;
Moulton, 1974) and Onchiodon (Boy, 1990; Werneburg, 2007) is well
known, but the specific taxonomy of the rich Eryops material is unsolved
at present. Therefore, a detailed description of the Eryops skull from the
Pennsylvanian of New Mexico (Lucas et al., 2005a) and comparison
with Permian Eryops would be one step toward a solution of the species
level taxonomic problems of the genus Eryops.
The Eryops skull from the Cañon del Cobre was discovered in
June, 2005, as part of the NMMNH collecting expeditions to Cañon del
Cobre led by SGL. The locality is NMMNH 6121, sec. 18, T24N, R6E.
It is approximately 125 m below the contact of the El Cobre Canyon
Formation and the overlying Arroyo del Agua Formation (Lucas et al.,
2005b). The skull was collected from a muddy sandstone together with
bones of an embolomere and Platyhystrix, as well as pelycosaurian-grade
synapsids. The skull is part of the characteristic assemblage of the Cobrean
land-vertebrate faunachron of Lucas (2006), and all biostratigraphic data
(palynostratigraphy, megafossil plants and vertebrate biostratigraphy)
indicate it is of Late Pennsylvanian age (Lucas et al., 2005b; and this
volume).
Institutional abbreviations: AMNH = American Museum of
Natural History, New York; BMNH = British Museum of Natural His
tory, London; CM = Carnegie Museum of Natural History, Pittsburgh;
CMNH = Cleveland Museum of Natural History; FMNH = Field Mu
seum of Natural History, Chicago; MCZ = Museum of Comparative
Zoology, Harvard University, Cambridge; NMMNH = New Mexico
Museum of Natural History, Albuquerque; SMNS = Staatliches Mu
seum für Naturkunde, Stuttgart; UBA = University of Buenos Aires.
DESCRIPTION AND COMPARISONS
General Morphology (Figs. 1-2)
The skull of NMMNH P-46379 consists of the anterior half and
the left cheek region. The marginal and palatal dentitions are exception
ally well preserved. The length from the anterior end of the premaxilla to
the level of the most posterior end of the quadrate is 41.5 cm. The
midline length is about 33-34 cm, comparable to measurements of other
Eryops skulls (Lucas et al., 2005a, table 1).
The dorsal skull table has a relatively fine and shallow sculpture
of low, narrow small ridges and shallow pits (Fig. 1), which is compa
rable to that of Eryops sp. (MCZ 1914) from the lower part of the
Archer City Formation of Texas (Werneburg, 2007, fig. 7a). The dermal
sculpturing of stratigraphically younger Eryops specimens, for example
the holotype (AMNH 4189) of E. megacephalus (Werneburg, 2007, fig.
6), is much more pronounced, with higher ridges and deeper pits.
The entire New Mexican skull exhibits a dorsal strutting pattern
with large ridges for better mechanical stability of the skull as is already
known in Eryops and Onchiodon (Sawin, 1941; Boy, 1990; Werneburg,
2007). A well-developed, large longitudinal ridge extends from the
postfrontal to the naris (Fig. 1B). Additional ridges run diagonally across
the jugal and lachrymal and as transverse ridges occur between the longi
tudinal ridges in the intraorbital region, the midlength of the nasals and
the premaxillaries. Depressions are present between these ridges.
The skull is dorsoventrally crushed, as is clearly indicated by the
medial displacement of the right jugal below the postorbital, and the
displacement of the left side of the skull roof bones slightly anteriorly.
However, the mid and posterior widths of the skull are affected little by
the distortion, as evidenced by the intact palate (Fig. 2). Therefore, the
preserved shape of the skull is generally narrow with a relative short,
blunt snout region that accurately portrays its true outline. As such, it
has an intermediate shape between the narrow skull morphotype with a
more pointed muzzle of the Eryops megacephalus holotype (AMNH
4189) and the broad snouted “Onchiodon-morphotype” of Eryops sp.
(MCZ 1914, cf. Werneburg, 2007). However, the interorbital region in
NMMNH P-46379 is narrower than in both mentioned morphotypes.
Skull Roof (Figs. 1-2)
The premaxillaries are relatively short and form a rounded, blunt
snout region, as in Eryops sp. (MCZ 1914). The alary process overlies
the nasal. The premaxilla, together with the maxilla, forms a shelf-like
margin of the naris. The unsculptured septomaxilla is oriented ventrolat
erally into the naris. The “angle” in the middle part of the right maxilla is
due to the distortion of the jugal. A small internasal seems to be situated
between the premaxillaries and the nasals. Sometimes such a median
bone is intercalated in the anterior skull table and reflects the ancestral
rostral elements. Werneburg (2007) described a slightly larger bone in the
same position in the eryopid Onchiodon thuringiensis. The posterior
portions of the nasals are greatly widened, so as to greatly restrict the
entrance of the lachrymal into the posteroventral margin of the naris. The
130
FIGURE 1. Eryops sp., NMMNH P-46379, from the Upper Pennsylvanian of Cañon del Cobre in New Mexico; dorsal view. Abbreviations: f = frontal, if
= interfrontal, in = internasal, j = jugal, l = lachrymal, mx = maxilla, n = nasal, pm = premaxilla, po = postorbital, pof = postfrontal, prf = praefrontal,
pt = pterygoid, q = quadrate, qj = quadratojugal, sm = septomaxilla, sq = squamosal.
lachrymal has a wide middle portion. The presence of the interfrontal is
clearly discernible by its radial sculpturing and its anterior and posterior
sutures. It is slightly shorter than the interfrontal in the Eryops
megacephalus holotype. The frontals and therefore the interorbital re
gion are narrow, whereas the orbital rim of the postfrontal is relatively
wide. Prefrontals, frontals and jugals extend to roughly the same anterior
level. The postorbital appears to be short, but its posterior portion is not
preserved. The squamosal is the largest bone of the cheek and is posteri
orly and posteromedially bordered by the quadrate and pterygoid, re
spectively. The quadratojugal is narrow and subequal in length to the
squamosal. The dorsal portion of the quadrate has a narrow exposure,
but is widened, especially in the anteromedial direction.
Palate (Figs. 2-3)
The anterior and marginal palatal bones are well preserved. The
parasphenoid and mid-region of the pterygoids are missing. The entire
surface of the palate is covered with small, tooth-like denticles.
The marginal and palatal dentitions of NMMNH P-46379 are
better preserved than in most other Eryops skulls. All the marginal teeth
have sharp cutting edges on their distal halves (Fig. 3). Stickler (1899, fig.
1) and Williston (1899, pl. 27) had illustrated the tooth morphology of
Eryops with blades. The proximal halves of the larger teeth show the
characteristic ridges of infolded dentine. The same morphology of the
teeth has been described in the eryopid Onchiodon and the archegosauroids
Sclerocephalus and “Cheliderpeton” (Werneburg, 1996). The tooth base
is mediolaterally expanded in the marginal dentition (Fig. 2B), as illus
trated by Stickler (1899, fig. 1A). All marginal teeth of the maxillae and
the palatal tusks are slightly less recurved posteromedially. The teeth of
the premaxillaries are not recurved (with one exception, the tooth in
position 11 of the right premaxilla is more incurved). The premaxilla has
11 to 12 tooth positions; five or six are represented by empty alveoli.
The largest teeth of the premaxilla occupy tooth positions 7 to 11, with
8 and 9 being the largest. The maxilla has 34 tooth positions, 11 being
represented by empty alveoli. The largest teeth in the maxilla are mar
ginal to the choanae and the tusk and socket pair of the palatine; these
occupy tooth positions 4 to 8 of the right maxilla and 3 to 10 of the left
maxilla. These larger tooth positions did not cause a lateral swelling of
the skull margin as is known from other Eryops skulls (e.g., Sawin,
1941). The largest maxillary teeth are smaller than those in the premax
illa.
A depression on the ventral surface of the premaxilla lies close to
the midline. The elongated vomers have a step-like elevation just poste
rior to the level of the vomerine tusk-and-socket pair. The paired fangs
of the vomers have a maximum length of about 14 mm and a basal
diameter of 6-7 mm (Fig. 3B). A raised, denticulated ridge surrounding
the paired fangs continues along or has a posterior extension along the
margin of the short and wide choanae. The palatine is wide anteriorly,
whereas posteriorly it quickly narrows to a splint-like process that
wedges between the ectopterygoid and pterygoid. The palatine fangs are
the largest teeth of the skull, which have a length of about 25 mm and a

FIGURE 2. Eryops sp. NMMNH P-46379, from the Upper Pennsylvanian of Cañon del Cobre in New Mexico; palatal view. Abbreviations: ch = choane,
ec = ectopterygoid, f.pq = paraquadrate foramen, j = jugal, mx = maxilla, pl = palatine, pm = premaxilla, pt = pterygoid, q = quadrate, qj = quadratojugal,
sph = sphenethmoid, sq = squamosal, vd = vomerine depth.
basal diameter of 10 mm. The basal ectopterygoid also has a large tusk
and socket pair, whose size is about 18 mm in length and 9 mm in basal
diameter. The ectopterygoid fangs are larger than those of the vomer and
equal in size to the largest teeth of the premaxilla. This is a clear differ
ence from the eryopid Onchiodon, which has extremely small
ectopterygoid fangs (Werneburg, 2007).
A large portion of the sphenethmoid is preserved, lying posterior
to the vomer and below the interorbital region. The left pterygoid has an
elongated, narrow process lacking denticles, as does the anterior end of
the palatal ramus, and it was probably overlain ventrally by the vomer.
The transverse flange is preserved on the right pterygoid. The right
quadrate ramus, forming the medial wall of the adductor fossa, has a
narrow ventral contact with the quadrate. The quadrate is wide and short
and forms the quadrate condyle. The ventral portion of the quadratojugal
is expanded in an anteromedial direction. Two small paraquadrate fo-
ramina are situated at the posterolateral suture with the quadrate. A larger
third foramen lies anteromedial to these foramina around a ridge in a small
depression on the quadratojugal.
DISCUSSION
The skull, NMMNH P-46379, from the Pennsylvanian of Cañon
del Cobre in New Mexico is doubtless a member of the genus Eryops
(Lucas et al., 2005a). The presence of an interfrontal, which is an
autapomorphic feature of Eryops (Werneburg, 2007), confirms this as
signment.
Two other eryopid skulls are known from the Pennsylvanian of
131
the USA, which have been assigned to Eryops: Glaukerpeton avinoffi
from the Upper Conemough Group at Pittsburgh in Pennsylvania (Romer,
1952), and Eryops cf. E. avinoffi from the Middle Conemough Group of
Sutton in West Virginia (Murphy, 1971). Vaughn (1958) reported an
interfrontal in “Glaukerpeton” avinoffi (contrary to Romer, 1952) and
concluded that Glaukerpeton is a synonym of Eryops. However, we
have restudied both specimens, and they lack an interfrontal and should
be reassigned to Glaukerpeton. The valid genus Glaukerpeton Romer,
1952, may be a more aquatically adapted eryopid of the newly defined
“Actinodon–complex” (Werneburg, 2007). The members of this eryopid
complex were formerly attributed to Onchiodon, and they are distin
guished from Eryops by: (1) a narrow preorbital region of the skull, (2) a
narrow interorbital region, (3) a narrow jugal, (4) a relatively elongate
postorbital portion of the skull with elongated supratemporals, (5) an
elongated interclavicle, and (6) an elongated ventral clavicle.
The type species of Eryops, E. megacephalus Cope, 1877, is well
known from specimen MCZ 1129 (Sawin, 1941) and from the holotype
specimen AMNH 4189 (Werneburg, 2007). Both skulls are from the
Permian Petrolia Formation (Wichita Group) of Texas. However, about
100 skulls of Eryops are known from the Lower Permian (Cisuralian) of
Texas, New Mexico, Utah, Oklahoma, and the Tristate Area (West Vir
ginia, Ohio and Pennsylvania). Milner (1996) provided a brief statement
about a new systematic concept for the genus Eryops, containing four
species based on cranial and dental features, but a detailed analysis has
never been published. Therefore, it is difficult to determine the species
level identity of the New Mexican skull NMMNH P-46379. At this
132
FIGURE 3. Eryops sp., NMMNH P-46379; median incurved labyrinthodont teeth with sharp lateral blades. A, anterior maxilla at margin of choanae; B,
fangs of the right vomer with denticulation of the socket.
stage of knowledge it is appropriate to compare it with distinctly defined
types, such as the holotype of E. megacephalus or Eryops sp., MCZ
1914 (see description above), as well as the known Eryops skulls from
the Lower Permian of New Mexico.
Initially, it is necessary to compare the skull NMMNH P-46379
with the known species of Eryops from the Cutler Group and the Abo
Formation of New Mexico. Langston (1953) discussed the history of
two species. E. reticulates Cope (1881) is a nomen vanum according to
Case (1911) and Langston (1953). Some postcranial material from the
University of California, Berkeley, localities in New Mexico (Spanish
Queen Mine, Quarry butte, Camp quarry and Anderson quarry) is grouped
under the species E. grandis by Langston (1953), which may be smaller
than the other Eryops specimens. According to Douthitt (1917), the
holotype specimen of E. grandis is lost. Therefore, he erected the new
genus Eryopsoides for two skulls from the Baldwin bone bed, the type
locality and stratum typicum of Eryops grandis. But, the holotype of E.
grandis exists in the MCZ collection (a mandible, sacral rib and centra:
MCZ 2731). Romer (1947) did not recognize a basis for the new genus
Eryopsoides and considered it to be a junior synonym of Eryops.
Unfortunately the available holotype material of Eryops grandis
lacks a skull, though the two skulls described by Douthitt (1917) are
from the same locality and were considered as plesiotypes (topotypes)
of E. grandis. Both skulls are illustrated here (Figs. 4-5). MCZ 2733
(Fig. 4) is a crushed skull with a right mandible, but with the same fine
sculpture pattern as in MCZ 2732 (Fig. 5). This skull, MCZ 2732, is
partially reconstructed with a skull length of about 34 cm. The skull
shape in the middle and anterior portions is very slender, which is un
usual for nearly all other Eryops specimens. The jugal seems also to be
narrow. This skull is the same size as, for example, the holotype speci
men of E. megacephalus (Werneburg, 2007). It is possible that the MCZ
2732 skull was incorrectly reconstructed. However, a very similar skull
shape is known from one Eryops skull (CM47817) from Utah (Fig. 6).
This specimen was found in the Halgaito Shale of the Cutler Group
(Wolfcampian) in southeastern Utah. The right jugal is narrow, and an
interfrontal is visible. This suggests that the species Eryops grandis may
be valid for a slender skull, if it is not really an effect of preservation,
with relatively fine dermal sculpture of the same size as E. megacephalus.
On the other hand, the differences in skull morphology between the
broad Pennsylvanian skull from Cañon del Cobre, NMMNH P-46379,
and the newly characterized E. grandis are clearcut. Therefore, the spe
cies E. grandis cannot be applied to all Eryops specimens from New
Mexico.
Surprisingly, another Eryops skull is known from the Cañon del
Cobre in New Mexico. It was collected by P. C. Miller in 1911 and is
housed in the collection of the Field Museum of Natural History, Chi
cago (FMNH-UC 751; Fig. 7). The skull is crushed, but it exhibits a fine
dermal sculpturing and the general dentition pattern. In FMNH-UC 751
at the level of the ectopterygian fangs, there are three large-sized maxil
lary teeth (Fig. 7A). At this level in NMMNH P-46379, the teeth of the
maxillae are clearly smaller (Fig.2). The snout region of FMNH-UC 751
seems to be narrow, as seen in palatal view (similar to E. grandis and
skulls from Texas), but the jugal region is wider (Fig. 7B). Therefore, the
two skulls from Cañon del Cobre are not identical.
Another skull with an approximately 35 cm length is known from
the Cutler Group (Wolfcampian) at CM locality 1070, 1.6 km northwest
of Arroyo del Agua, on the west side of the Rio Puerco (CM 34906;
Lucas et al., 2005c, fig 4E, I; Fig. 8). This skull clearly belongs to the
genus Eryops based on the following features: (1) interfrontal present,
(2) characteristic stabilization system of different sculpture ridges on
dorsal skull roof, (3) coronoids with full denticulation as reported by
Langston (1953) for E. grandis, but also known from Texas Eryops
material (Broom, 1913; Werneburg, 2007), and (4) broad jugal. Other
features in common for both skulls are the fine dermal sculpturing of the
dorsal skull roof, the outline shape of the skull, similar skull lengths, and
 133

FIGURE 4. Eryops grandis; plesiotype MCZ 2733, dorsal skull with right
mandible, from the “Baldwin bonebed” of the Lower Permian El Cobre FIGURE 6. Eryops grandis, CM 47817, dorsal skull, from the Early Permian
Canyon Formation of Arroyo del Agua in New Mexico. Cutler Formation (Halgaito Shale) of the Valley of the Gods in Utah.

FIGURE 7. Eryops sp., FMNH UC 751; skull from the El Cobre Canyon
Formation of the Cañon del Cobre in New Mexico; A, ventral view of
narrow snout region with palatal and maxillary dentition; B, lateral view of
wide left jugal with fine dermal sculpture.

FIGURE 5. Eryops grandis, plesiotype MCZ 2732, reconstructed skull with

original parts in dorsal view, from the “Baldwin bonebed” of the Lower
Permian El Cobre Canyon Formation of Arroyo del Agua in New Mexico.
134
FIGURE 8. Eryops sp., CMNH 34906; skull with left mandible; from the Lower Permian El Cobre Canyon Formation of Arroyo del Agua (1.6 km to NW,
west side of Rio Puerco, CM locality 1070) in New Mexico; A, dorsal skull roof; B, lateral view with mandible.
(probably) not very enlarged maxillary teeth, near the level of the
ectopterygoid fangs. However, some differences in the proportions of
the dorsal skull roof are present. The Permian skull (CMNH 34906, Fig.
8) differs from the Pennsylvanian skull of similar size (NMMNH P
46379) in the following features: (1) a much wider interorbital region, (2)
a more elongated snout region (premaxilla), which is slightly more pointed,
(3) a longer distance from the orbit to the naris, and (4) a more elongated
interfrontal. Therefore, both skulls are different at the species level.
In summary, it is possible to consider that the skull NMMNH P
46379 from the Late Pennsylvanian of Cañon del Cobre clearly belongs
to the genus Eryops, but the species assignment is uncertain. It is not
identical to the Early Permian Eryops specimens from New Mexico. The
Pennsylvanian Eryops of New Mexico thus may represent a new spe
cies. This conclusion will be clarified by a better knowledge of the basal
Eryops specimens from the Permian of Texas.
The discovery of Eryops in the El Cobre Canyon Formation also
sheds new light on Late Pennsylvanian/Early Permian tetrapod/arthro
pod relationships. Lucas et al. (2005d) reported the occurrence of large
Diplichnites tracks made by arthropleurids in the top of fluvial channel
sandstones in the El Cobre Canyon Formation. The co-occurrence of
eryopid amphibians with the giant terrestrial arthropod Arthropleura
was reported and discussed for the first time by Schneider and Werneburg
(1998) based on the example of the Early Permian (Asselian) Manebach
Formation of the Thuringian Forest basin and the Late Pennsylvanian/
Early Permian (Gzhelian/Asselian) Döhlen Formation of the Döhlen
basin, both in Germany. In the Thuringian Forest basin, the eryopid
Onchiodon as well as the pelycosaurid reptile Haptodus have been dis
covered in pebbly channel sandstones of an anastomosing to braided
river system in a floodplain environment (Werneburg, 2007). The vegeta
tion of this landscape consists of gallery forests of calamites along river
courses, as well as swampy areas with coal-forming cordaite forests.
Arthropleura remains occur in shales and sandy siltstones together with
plant remains from different growth sites, which were washed together
during flood events.
In the Döhlen basin, Arthropleura has been discovered in
pyroclastics with mesophilous vegetation between two coal seams

(Schneider and Barthel, 1997). Several meters below, another volcanic
ash horizon contains Limnopus tracks, which probably were produced
by eryopid amphibians (Hausse, 1910). A further occurrence of Limnopus
has been reported together with the arthropleurid track Diplichnites on
the same bedding plane in coastal plain fluvial sandstones of the lower
Conemaugh Formation of Kentucky (Martino and Greb, 2009). The
repeated co-occurrence of eryopid skeletal fossils (Eryops in North
America and its ecological counterpart Onchiodon in Europe) and the
Limnopus tracks, which were likely made by eryopid amphibians, with
Arthropleura body remains or the arthropleurid track Diplichnites, can
be regarded as indications of a characteristic biotope preference of eryopids
and arthropleurids. Obviously, eryopids spent most of their time in
water and only occasionally ventured on land in fluvial environments.
The sparsely vegetated landscape along the river courses, mainly calam
ite-dominated gallery forests, were the preferred biotope of the giant
arthropleurids. Rarely, fully terrestrially-adapted tetrapods in those en
REFERENCES
Berman, D. S, 1993, Lower Permian vertebrate localities of New Mexico
and their assemblages: New Mexico Museum of Natural History and
Science, Bulletin 2, p. 11-21.
Boy, J. A., 1990, Über einige Vertreter Eryopoidea (Amphibia:
Temnospondyli) aus dem europäischen Rotliegend (?höchstes Karbon –
Perm). 3. Onchiodon: Paläontologische Zeitschrift, v. 64, p. 287-312.
Broom, R., 1913, Studies on the Permian temnospondylous stegocephalians
of North America: American Museum of Natural History Bulletin, v. 32,
p. 563-595.
Case, E. C., 1911, Revision of the Amphibia and Pisces of the Permian of
North America: Carnegie Publications, v. 146, p. 1-148.
Cope, E. D., 1877, Descriptions of extinct Vertebrata from the Permian
and Triassic formations of the United States: Proceedings of American
Philosophical Society, v. 17, p. 182-193.
Cope, E. D., 1881, The Permian formation of New Mexico: American
Naturalist, v. 15, p. 1020.
Douthitt, H., 1917, Eryops; Eryopsoides, gen nov. from the New Mexico
Permian: The Kansas University Science Bulletin, v. 10, p. 237-242.
Harris, S. K., Lucas, S. G., Berman, D. S and Henrici, A. C., 2004, Vertebrate
fossil assemblage from the Upper Pennsylvanian Red Tanks Member of
the Bursum Formation, Lucero uplift, central New Mexico: New Mexico
Museum of Natural History and Science, Bulletin 25, p. 267-283.
Hausse, R., 1910, Fossile Tierfährten im Unterrotliegenden des
Steinkohlenbecken im Plauenschen Grunde (des Döhlener Beckens) bei
Dresden: Jahrbuch für das Berg- und Hüttenwesen in Sachsen 1910, p. 3
19.
Langston, W., Jr., 1953, Permian amphibians from New Mexico: Univer
sity of California Publications in Geological Sciences, v. 29, p. 349-416.
Lucas, S. G., 2006, Global Permian tetrapod biostratigraphy and
biochronology; in Lucas, S. G., Cassinis, G. and Schneider, J. W. (eds.),
Non-marine Permian biostratigraphy and biochronology. Geological
Society, London, Special Publication 265, p. 65-93.
Lucas, S. G., Rinehart, L. F. and Hunt, A. P., 2005a, Eryops (Amphibia,
Temnospondyli) from the Upper Pennsylvanian of El Cobre Canyon,
New Mexico: New Mexico Museum of Natural History and Science,
Bulletin 31, p. 118-120.
Lucas, S. G., Harris, S. K., Spielmann, J. A., Berman, D. S. and Henrici, A. C.,
2005b, Vertebrate biostratigraphy and biochronology of the Pennsylva
nian-Permian Cutler Group, El Cobre Canyon, northern New Mexico:
New Mexico Museum of Natural History and Science, Bulletin 31, p.
128-139.
Lucas, S. G., Harris, S. K., Spielmann, J. A., Berman, D. S, Henrici, A. C.,
Heckert, A. B., Zeigler, K. E. and Rinehart, L. F., 2005c, Early Permian
vertebrate biostratigraphy at Arroyo del Agua, Rio Arriba County, New
Mexico: New Mexico Museum of Natural History and Science, Bulletin
31, p. 163-169.
135
vironments include pelycosaurian-grade synapsids, such as Haptodus in
the Thuringian Forest basin and the Döhlen basin, as well as
undeterminated pelycosaurian-grade synapsid remains in the El Cobre
Canyon Formation. Further synecological relationships, such as pos
sible prey-predator relationships, are discussed by Schneider, Lucas,
Werneburg and Rößler (2010, this volume).
ACKNOWLEDGMENTS
We thank our colleagues D. S Berman (CM), J. R. Bolt (FMNH
Chicago), F. A. Jenkins Jr. (MCZ), C. Mehling (AMNH), A. R. Milner
(BMNH), C. Norris (AMNH), M. J. Ryan (CMNH Cleveland), C.
Schaff (MCZ) and R. R. Schoch (SMN Stuttgart) for the opportunity to
study specimens and helpful references. Reviews by D. S Berman (CM),
A. Henrici (CM) and C. A. Marsicano (UBA) improved the manuscript.
This work was supported by a grant of the Deutsche Forschungs
gemeinschaft (DFG-We 2833/2-1, DFG-We 2833/3-1, DFG-Schn 408/
12-1 and 2).
Lucas, S. G., Lerner, A. J., Hannibal, J. T., Hunt, A. P. and Schneider, J. W.,
2005d, Trackway of a giant Arthropleura from the Upper Pennsylva
nian of El Cobre Canyon, New Mexico: New Mexico Geological Society,
Guidebook 56, p. 279-282.
Martino, L.R. and Greb, S. G., 2009, Walking trails of the giant terrestrial
arthropod Arthropleura from the Upper Carboniferous of Kentucky:
Journal of Paleontology, v. 83, p. 140-146.
Milner, A. R., 1996, Systematics of the genus Eryops (Amphibia:
Temnospondyli) and its possible biostratigraphical value: Journal of
Vertebrate Paleontology, v. 16, supplement to no. 3, p. 53A.
Moulton, J. M., 1974, A description of the vertebral column of Eryops based
on the notes and drawings of A.S. Romer: Breviora, v. 428, p. 1-44.
Murphy, J. L., 1971, Eryopsid remains from the Conemaugh Group, Braxton
County, West Virginia: Southeastern Geology, v. 13, p. 265-273.
Romer, A. S., 1947, Review of the Labyrinthodontia: Bulletin Museum
Comparative Zoology Harvard, v. 99, p. 1-368.
Romer, A. S., 1952, Late Pennsylvanian and Early Permian vertebrates of
the Pittsburgh West Virginia region: Annals of the Carnegie Museum, v.
33, p. 47-110.
Sawin, H. J., 1941, The cranial anatomy of Eryops megacephalus: Bulletin
of Museum of Comparative Zoology Harvard College, v. 88, p. 407
463.
Schneider, J. and Barthel, M., 1997, Eine Biocoenose mit Arthropleura im
Rotliegend (?Unterperm) des Döhlen-Becken (Elbezone, Sachsen):
Freiberger Forschungsheft C , v. 466, p. 183-223.
Schneider, J. W. and Werneburg, R., 1998, Arthropleura und Diplopoda
(Arthropoda) aus dem Unter-Rotliegend (Unter-Perm, Assel) des
Thüringer Waldes (Südwest-Saale-Senke): Veröffentlichungen
Naturhistorisches Museum Schleusingen, v. 13, p. 19-36.
Stickler, L., 1899, Über den mikroskopischen Bau der Faltenzähne in Eryops:
Paläontographica, v. 46, p. 85-94.
Vaughn, P. P., 1958, On the geologic range of the labyrinthodont amphibian
Eryops: Journal of Paleontology, v. 32, p. 918-922.
Werneburg, R., 1993, Onchiodon (Eryopidae, Amphibia) aus dem Rotliegend
des Innersudetischen Beckens (Böhmen): Paläontologische Zeitschrift,
v. 67, p. 343-355.
Werneburg, R., 1996, Temnospondyle Amphibien aus dem Karbon
Mitteldeutschlands: Veröffentlichungen Naturhistorisches Museum
Schleusingen, v. 11, p. 23-64.
Werneburg, R., 2007, Der “Manebacher Saurier” – ein neuer großer Eryopide
(Onchiodon) aus dem Rotliegend (Unter-Perm) des Thüringer Waldes:
Veröffentlichungen Naturhistorisches Museum Schleusingen, v. 22, p. 3
40.
Williston, S. W., 1899, Notes on the coraco-scapula of Eryops COPE: The
Kansas University Quarterly, v. 8, p. 185-186.
136

Eryops attacks a xenacanth shark. Artwork by Matt Celeskey.

 Lucas et al., eds., 2010, Carb-Permian transition in Cañon del Cobre. New Mexico Museum of Natural History and Science Bulletin 49.

137
A PARTIAL SKELETON OF OPHIACODON NAVAJOVICUS
(EUPELYCOSAURIA: OPHIACODONTIDAE) FROM THE
UPPER PENNSYLVANIAN OF CAÑON DELCOBRE, NEWMEXICO

SUSANK. HARRIS, SPENCERG. LUCASAND JUSTINA. SPIELMANN

New Mexico Museum of Natural History and Science, 1801 Mountain Road NW, Albuquerque, NM 87104

Abstract—A recently discovered partial skeleton of the pelycosaurian-grade synapsid Ophiacodon navajovicus,
the most complete yet found, is described from the Upper Pennsylvanian interval of the El Cobre Canyon
Formation (Cutler Group) at Cañon del Cobre, Rio Arriba County, New Mexico. The postcranial elements include
vertebrae representative of all regions of the vertebral column, a nearly complete pelvis, a complete right femur and
portions of other limb bones, as well as bones of the right pes. In addition, fragments of a right maxilla and dentary
were also found. Its stratigraphic distribution in New Mexico, Utah and Colorado indicate that O. navajovicus
characterizes the Cobrean land-vertebrate faunachron (lvf).

INTRODUCTION the holotype of a new species, Dimetrodon navajovicus. Ironically, an

Although no longer recognized as the oldest species of the genus
Ophiacodon, O. navajovicus is distinctive in its diminutive size and
remarkable in its abundance in sediments of the El Cobre Canyon Forma
tion in Cañon del Cobre, northern New Mexico. Since its initial descrip
tion as a new species by Case in 1907, only a few key skeletal elements
have been adequately described or illustrated, although four institutions
hold large quantities of specimens identified as O. navajovicus. Here, we
document the most complete specimen of O. navajovicus known (Figs.
1-6) and discuss its utility as an index taxon of the Cobrean land-verte
brate faunachron (lvf).
In this paper, AMNH = American Museum of Natural History,
New York; MCZ = Museum of Comparative Zoology, Cambridge, Mas
sachusetts; NMMNH = New Mexico Museum of Natural History and
Science, Albuquerque; and YPM = Yale Peabody Museum, New Haven.
LOCALITY
The Ophiacodon navajovicus specimen described here, NMMNH
P-43121, is from Cañon del Cobre at NMMNH locality 5660 in section
25, T24N, R6E. This locality is in the middle part of the locally exposed
section of the El Cobre Canyon Formation (Cutler Group). It was pre
served in brownish-red mudstone with blue concretions located approxi
mately 80 m below the contact of the El Cobre Canyon Formation and
the overlying Arroyo del Agua Formation (Lucas and Krainer, 2005). A
majority of the O. navajovicus elements were found in close association
in a small, shallow “pocket,” whereas some elements were weathered
free and collected over an area of approximately 4 by 5 m. Diagnostic,
but isolated, elements of O. navajovicus were abundant at two additional
sites in the same stratigraphic interval.
PREVIOUS STUDIES
In the early 1880s, David Baldwin collected numerous Ophiacodon
fossils from Cañon del Cobre that were later recognized as representative
of a single species. Part of this material reposited in the American Mu
seum of Natural History was assigned several Cope collection lot num
bers (No. 2285, No. 2298 and No. 2299), each of which included mul
tiple individuals (Figs. 7A-E, 8). Although Cope (1888) briefly described
and illustrated a “pes” (later identified as a manus by Romer and Price,
[1940]) from this material without naming a new species, he apparently
wrote on the label of a box containing a characteristic humerus the word
“navajovicus” (Case, 1907, p. 56).
The suggested specific epithet was later read and adopted by Case
(1907) when he selected a humerus, AMNH 4777 from lot No. 2298, as
other humerus, AMNH 4793 from lot No. 2285, was chosen for descrip
tion and illustration instead of the holotypic humerus, which is illus
trated here for the first time (Fig. 7A-B).
The Yale Peabody Museum also received a large number of similar
specimens, sent by Baldwin to O. C. Marsh in the 1880s. After returning
from the Chicago expedition to Cañon del Cobre in 1911, Williston
(1914) compared material he had collected for the University of Chicago
from the surface of the canyon floor to the Baldwin material. The similar
ity of the fossils was so great that it led him to speculate that all of it had
perhaps been recovered from the same bone bed. Consequently, Williston
(1914) erected the new genus Arribasaurus, not only for the two similar
groups of fossils in Chicago and New Haven, but also for the American
Museum material described and named by Case (1907). As justification
for removal of the material from the genus Dimetrodon, Williston (1914,
p. 409) merely stated that “the teeth will exclude it from the
Sphenacodontidae.”
Romer and Price (1940) compared directly the Yale material with
that in the American Museum and concurred with Williston regarding the
conspecific nature of the two groups of fossils. However, they con
cluded that no morphological differences separate “Arribasaurus” from
Ophiacodon that are not arguably the result of ontogenetic changes,
thereby relegating the former taxon to the status of junior subjective
synonym. As a result, Romer and Price (1940) transferred all the material
to the genus Ophiacodon, under the new combination Ophiacodon
navajovicus.
Ophiacodon navajovicus was thought to be endemic to Cañon del
Cobre until a wider geographical range for the species was recognized by
Vaughn (1962) when he identified material as Ophiacodon cf. O.
navajovicus from the Halgaito “tongue” (raised to formation status by
Wengerd and Matheny, 1958) of the Cutler Group in San Juan County,
Utah. The most compelling evidence on which Vaughn (1962, p. 537)
based his identification was the distal portion of a humerus possessing a
“broad entepicondyle” that “looks exactly like that of AMNH 4777, the
holotype of O. navajovicus.” In addition, Vaughn (1962, p. 537) based
the probable presence of O. navajovicus in the Halgaito Formation on
isolated vertebrae characterized by “the typically Ophiacodon pattern,”
combined with “a flattened strip along the ventral midline of the cen
trum” (Vaughn, 1962, p. 537; Sumida et al., 1999).
Another possible occurrence of Ophiacodon navajovicus outside
of Cañon del Cobre was reported by Lewis and Vaughn (1965) from the
Cutler Group of southwestern Colorado. They described and illustrated
an articulated string of five small vertebrae whose centra are character
ized by “a flattened area bounded by a pair of longitudinal ridges” (Lewis
138
and Vaughn, 1965, p. C26). Although the linear measurements of the
vertebrae fall within the ranges listed by Romer and Price (1940, table 3)
for O. navajovicus, Lewis and Vaughn were undecided as to whether
specific assignment should be made to O. navajovicus or Ophiacodon
mirus. However, an illustrated transverse section through one of the
vertebrae (Lewis and Vaughn, 1965, fig. 10A) clearly shows the narrow,
flat, ventral surface of the centrum diagnostic of O. navajovicus.
In the summer of 2004, a NMMNH field party led by one of us
(SGL) collected numerous associated axial and appendicular elements
from the floor of Cañon del Cobre of a single individual of Ophiacodon
navajovicus (NMMNHP-43121, Fig. 1). These elements, some of which
were briefly mentioned and illustrated by Lucas et al. (2005), represent
the single most complete skeleton of O. navajovicus known, and are the
subject of this paper.
DESCRIPTION
As discussed above, all the skeletal material described below was
collected from a small, restricted area of a few square meters. Given the
close association of these fossils, their uniform presentation, the lack of
duplicative elements, the extent to which many elements are readily
articulated to each other and the consistency of their relative sizes, we
interpret these remains as pertaining to a single individual.
Cranial Skeleton
To date, cranial material of Ophiacodon navajovicus is largely
unknown and consists of isolated maxillae, quadrates and braincase ele
ments (Romer and Price, 1940) that have never been described or illus
trated. Recovered cranial elements of NMMNH P-43121 are limited to a
fragmentary maxilla and dentary (Fig. 2A-F).
Maxilla - A fragmentary right maxilla contains two partial, seri
ally adjacent caniniform teeth: anteriorly the basal half of a large canini
form tooth is preserved, followed posteriorly by the anterior half of a
hollow-cored tooth, which may represent a resorption pit. Posterior to
the caniniform teeth are two very small alveoli (Fig. 2A-C). In many
pelycosaurian-grade synapsids, the two largest teeth of the maxilla are
alternately replaced (Reisz, 1986), so that only one is functional at any
given time. The robustly constructed maxillary shelf (preserved length =
26 mm) is enlarged above the caniniform teeth and is supported dorsally
by a narrow buttress attached to the medial surface of the maxilla.
Dentary - A small, right dentary fragment preserves the striated
bases of three marginal teeth, the fractured pulp cavity of a fourth tooth
and an empty alveolus (Fig. 2D-F). As preserved, the length of the
FIGURE 1. Ophiacodon navajovicus, NMMNH P-43121, incomplete skeleton laid out in plan view. Anterior is to the right.
dentary fragment is 25 mm. The teeth are oval in horizontal section, with
the long axis oriented mesiodistally.
Axial Skeleton
Although Romer and Price (1940) state that vertebrae from all
regions of the column of Ophiacodon navajovicus are present in the
material collected by Baldwin, only dorsal and lumbar vertebrae were
selected for description or illustration by previous workers (see Case,
1907, pl. 27, figs. 3-5; Williston, 1914, fig. 11, nos. 8 and 12). However,
Romer and Price (1940, table 3, p. 438) provide a table of measurements
for selected cervical, dorsal, lumbar and caudal vertebrae of O. navajovicus
specimens reposited in the AMNH and YPM collections. Measure
ments of all vertebrae of NMMNH-43121 fall within the ranges given in
this table (measurements for the atlas, axis and sacrals were not given).
NMMNH P-43121 includes a total of 34 vertebrae, some of
which are articulated in short strings, and are representative of all regions
of the vertebral column (Figs. 1, 2-4). Among these are the atlantal
centrum, axis, articulated cervicals and the two sacrals that are described
and illustrated here for the first time.
Atlas-axis complex - The atlantal centrum is a small, cuboidal
element (Fig. 2G-K). It has an anteroposterior length of 10 mm, little
more than half that of other preserved presacral vertebrae. The upper
half of its convex anterior face (width=10 mm) encloses a small foramen
for passage of the notochord, whereas its posterior face (width=15 mm)
bears a deep notochordal funnel. A shallow groove along the length of its
dorsal surface forms the floor of the neural canal. Highly characteristic of
Ophiacodontidae (Romer and Price, 1940) is the ventral surface of the
atlantal centrum, which is firmly fused to the axial intercentrum and
exhibits a round facet on each side of its posterior edge for the reception
of the axial rib capitulum. The ventral surface of the axial intercentrum
bears a prominent, rounded, longitudinal ridge. Total height of the com
bined elements is 17 mm.
The greatly enlarged axial neural spine has a maximum anteropos
terior length of 29 mm at its distal margin (Fig. 2L). The straight poste
rior margin is vertical, whereas the anterior margin is angled 45 degrees
anteriorly from the vertical so that it projects anteriorly 8 mm beyond
the level of the anterior rim of the centrum. The thin anterior edge of the
blade thickens posteriorly to within a very short distance of the poste
rior margin, where it again thins. The prezygapophyses are very small
and project slightly anterolaterally from the surface of the neural arch, in
contrast to the robust, elongate postzygapophyses that project
posterolaterally 4 mm beyond the posterior rim of the centrum. All that
 139

FIGURE 2. Ophiacodon navajovicus, A-R, NMMNH P-43121. A-C, right maxilla fragment in A, lateral, B, medial and C, occlusal views. D-F, right
dentary fragment in D, lateral, E, medial and F, occlusal views. G-K, fused atlas centrum and axial intercentrum in G, posterior, H, anterior, I, right lateral,
J, dorsal and K, ventral views. L-N, axis in L, left lateral, M, ventral and N, anterior views. O-R, three cervical vertebrae in O, anterior, P, posterior, Q,
right lateral and R, ventral views.
140
remains of the transverse processes is a raised rugose, circular outline on
either side of the dorsal margin of the centrum near its anterior rim (Fig.
2L-N). A sharp, midventral keel extends the full length of the 21 mm long
centrum (Fig. 2M).
Postaxial Cervical Vertebrae - Three articulated postaxial cer
vical vertebrae, probably representing positions 3-5 or 4-6, are nearly
complete, with the posteriormost one lacking the posterior portion of
the centrum and a right transverse process (Fig. 2O-R). The neural spines
are inclined anteriorly 50 degrees from the vertical (Fig. 2Q), an indica
tion of the upward curvature of the column in the neck region (Reisz,
1986). In lateral view the cervicals of Ophiacodon navajovicus closely
resemble those of Ophiacodon mirus illustrated by Williston and Case
(1913, fig. 24), but with the oblique orientation of the neural spine being
more pronounced in the four cervicals immediately following the axis.
The neural spines are stout, with a height of 20 mm and a maxi
mum transverse width of 8 mm at the distal terminus, where there is a
lateral boss on both sides of the spine (Fig. 2N-Q). They are oval to
diamond-shaped in sagittal section. The medial margins of the pre and
postzygapophyses are widely separated and their facets are inclined 20
degrees from the horizontal. Unlike the postzygapophyses of the axis,
those of the postaxial cervicals do not extend beyond the level of the
posterior border of the centrum.
The transverse processes originate high on the dorsolateral sur
face of the centrum, almost at the level of the zygapophyses (Fig. 2O-P).
They possess the typical ophiacodontid wing-shaped ventral flange or
web that extends from the diapophysis to the parapophysis at the ante
rior edge of the centrum. Compared to transverse processes of the dorsal
vertebrae of this specimen, they have a slight posteroventral orientation
and are significantly longer (length = 11 mm) and more robust.
In contrast to the sharp, midventral keel of the axis, the centra of
the postaxial cervical vertebrae possess a narrow ventral surface that is
flat and bordered by straight longitudinal edges (Fig. 2R). The centra are
wedge-shaped in cross section. Large, well-ossified intercentra with a
midventral length of 5 mm are articulated between all three centra.
Dorsal Vertebrae – Eleven dorsal vertebrae were recovered. Three
of them are articulated though fragmentary, but the neural spine of the
first is complete and well-preserved (Fig. 3A-C). This spine has a height
of 27 mm, an anteroposterior length of 10 mm, and a transverse width of
4 mm, making it one of the tallest and broadest neural spines of the
preserved column. Williston and Case (1913) divided the 20 dorsal ver
tebrae of Ophiacodon mirus into three regions based on the dimensions
of the neural spines. By analogy with their model, the relative height,
length and width of this neural spine would place these three articulated
vertebrae in the mid-dorsal region among serial positions 14-20, with
other dorsal vertebrae that have relatively tall, broad neural spines of
moderate thickness. An isolated vertebra has a neural spine of similar
proportions and thus most likely represents a mid-dorsal (Fig. 3E-H).
However, it differs from the three articulated dorsal vertebrae in its
posteriorly positioned neural arch that overhangs the posterior border of
the centrum by 5 mm (Fig. 3E).
A single, isolated neural spine is distinctively tall, narrow and thin
(height = 30 mm; anteroposterior length = 8 mm; transverse width = 2
mm), so its probable serial position was more posterior within the range
of vertebral positions 21-23, as these vertebrae possess gracile neural
spines.
Because the orientation of the neural arches of the dorsal vertebrae
of NMMNH P-43121 is not inclined anteriorly like that of the cervicals,
but is instead normal to the long axis of the centrum, the dorsal neural
arches are higher and form a less constricted neural canal. The inclination
and spacing of the articular facets of the zygapophyses remain the same
as in the cervicals, however, the postzygapophyses are not as robust and
elongate. The single preserved transverse process is short and originates
somewhat lower on the dorsal margin of the centrum than those of the
cervicals (Fig. 3C). The maximum lateral length of its horizontal dorsal
surface is 7 mm. Similar to the cervicals, the centra of the dorsal vertebrae
are wedge-shaped in end view, and characterized by a narrow, flat ventral
surface bordered by well-defined edges (Fig. 3D, H).
Four similar, but fragmentary vertebrae are also likely from the
dorsal region of the column, although their exact serial position cannot be
determined due to the absence of their neural spines. Two dorsal proxi
mal rib fragments are anteroposteriorly flattened and have a dorsoventral
height of 17 mm. Anteriorly, the dichocephalus head is triangular in
outline. A small indentation separates the tubercular and capitular sur
faces, which are connected by a thin web of bone.
Lumbar Vertebrae – Two lumbar vertebrae, the better preserved
of which is illustrated in Fig. 3I-J, consist of the centrum and the anterior
portion of the neural arch. The prezygapophyses are very small and the
facets are inclined medially 20 degrees from the horizontal, and posi
tioned close to the midline, which is very different from the condition of
the zygapophyses in the preceding presacral series. In anterior view the
centra are smaller in diameter and rounder than those of the dorsal verte
brae. The narrow midventral surface of the centrum, though flattened, is
very gently concave between two slightly raised parallel edges.
Sacral Vertebrae - The sacrum (Fig. 3K-N) consists of only two
vertebrae, as is the case in all ophiacodontids for which this region of the
vertebral column is known (Romer and Price, 1940; Romer, 1956). The
first sacral vertebra in NMMNH P-43121, though missing most of its
neural arch and spine, has a well-preserved centrum with a length of 14
mm. In anterior view (Fig. 3L) the height of its lateral walls is enclosed by
the fused, bulbous transverse processes, which form a continuous sur
face with the lateral border of the prezygapophyses, giving the centrum
a massive, “swollen” appearance. It is nearly identical to the first sacral
vertebra included in AMNH 4793 (formerly cataloged in specimen lot
AMNH No. 2285) (Fig. 7D). Near the ventral surface of the centrum, at
the terminus of the transverse process, there is a small, ventrolaterally
directed rib facet. The preserved distal end of the first sacral rib (Fig. 3O
P) forms a cupped, fan-like expansion with a thickened anterior border,
broadly similar to that of Ophiacodon illustrated by Romer (1956, fig.
133A). The proximal shaft of the rib is not preserved. As in the lumbar
vertebrae, the preserved anterior zygapophyses are very small and posi
tioned close together at the midline.
The tranverse processes of the second sacral vertebra (Fig. 3K,
M-N) are also bulbous and fused to the lateral walls of the centrum, but
are smaller than those of the first sacral, with an anteroventrally-directed
rib facet that allowed the second sacral rib to support that of the first
sacral, instead of contacting the ilium (Reisz, 1986; see also illustration in
Romer, 1956, fig. 133A). The postzygapophyses are small, positioned
close to the midline and extend beyond the posterior border of the cen
trum by 4 mm. The midventral surface of the centrum of both sacrals, as
in the lumbar vertebrae, is very gently ventrolaterally concave with slightly
raised, paralleledges (Fig. 3N).
Caudal Vertebrae – The first two caudal vertebrae, one of which
is illustrated in Figure 4A-C, have a length of 17 mm with short, moder
ately stout rectangular neural spines. The neural arches are high and
perpendicular to the long axis of the centrum. The one transverse process
is small (transverse length 9 mm), but robust and has a short ventral
extension of the diapophysis. The zygapophyses are very widely sepa
rated, and their articular facets are gently inclined medioventrally. Similar
to the ventral surface of the axis, the anterior caudals have a sharp
midventral keel (Fig. 4C).
Positioned somewhat posteriorly to the first two caudal vertebrae
are caudals with a length of 14 mm, in which the transverse processes
extend halfway down the lateral walls of the centra (Fig. 4D-G). Their
diapophyses are oval in cross section, with the long axis oriented longi
tudinally. Just ventral to the diapophyses are ridges that extend length
wise along the lateral centrum walls. The midventral surface of the cen
trum is v-shaped in cross section, but not as sharp as those of the axis
and anterior caudals (Fig. 4E). Positioned somewhat posteriorly are three
additional caudal vertebrae that are slightly smaller (length = 12 mm)
with faint longitudinal ridges on the lateral walls of the centrum. The
 141

FIGURE 3. Ophiacodon navajovicus, A-P, NMMNH P-43121. A-D, three articulated dorsal vertebrae in A, left lateral, B, anterior, C, posterior and D,
ventral views. E-H, isolated dorsal vertebra in E, left lateral, F, anterior, G, posterior and H, ventral views. I-J, lumbar vertebra in I, right lateral and J,
anterior views. K-N, articulated sacrum in K, left lateral, L, anterior, M, posterior and N, ventral views. O-P, right sacral rib fragment in O, dorsal and P,
ventral views.
142
ventral surface of the centrum is keeled and v-shaped in cross section.
Six of the most posteriorly positioned preserved caudals decrease
in length from 12 mm to 9 mm. Their centra are laterally compressed,
simple cylinders (Fig. 4H-K). Notably, the ventral surface of even the
smallest distal caudal is flat with straight edges.
Pectoral Girdle
Elements of the pectoral girdle of Ophiacodon navajovicus are
rarely preserved. However, Romer and Price (1940, fig. 22) illustrated a
reconstruction of the scapulocoracoid of O. navajovicus based on frag
mentary portions of the dorsal scapular blade and the glenoid area of the
coracoid. NMMNH P-43121 includes a fragment of the ventral portion
of the left scapular blade (Fig. 4L-M). With the exception of a short
length of the thickened anterior border just above the supraglenoid but
tress, all edges are broken. Its maximum preserved anteroposterior length
is 59 mm.
Forelimb
Humerus – Recovered elements of NMMNH P-43121 only in
clude the proximal portion of a right humerus. It has a maximum width of
41 mm (Fig. 4N-P), the same value as that listed for a proximal part of the
humerus, AMNH 4784, by Romer and Price (1940, table 4, p. 156). In
dorsal or ventral view, its outline is broadly triangular with a slightly
rounded, spiral articular surface. The proximal thickened ridge of the well
developed deltopectoral crest broadens distally to a triangular terminus.
Only a short portion of the proximal end of the shaft is preserved, but
indicates that it was slender and round in cross section. These features
compare closely with those of the humerus designated as the holotype of
Ophiacodon navajovicus, AMNH 4777 (Fig. 7A-B), as well as to an
other characteristic humerus illustrated by Case (1907, plate 27, fig 2).
Radius - The radius of NMMNH P-43121 was not recovered.
Ulna – The ulna is represented only by the proximal and distal
ends of the right ulna. The proximal head (Fig. 4 Q-R) possesses a
moderately developed, well-ossified olecranon process. Its transverse
width, measured proximal to the well-defined sigmoid notch, is 24 mm.
The preserved portion of the ulnar shaft is somewhat mediolaterally
compressed and narrows proximally from a width of 18 mm at its convex
distal articular end to 10 mm at its preserved distal end. Measurements
for both proximal and distal ends of the ulna of NMMNH P-43121 fall
within the respective ranges listed for the ulna of Ophiacodon navajovicus
by Romer and Price (1940, table 4).
Manus - The recovered elements of NMMNH P-43121 do not
include bones of the manus.
Pelvic Girdle
To date, known material of the mature pelvic girdle of Ophiacodon
navajovicus is limited to one specimen, MCZ 1595, which consists of an
incomplete right ilium and complete right ischium, though lack of diag
nostic features in these elements makes their assignment to O. navajovicus
questionable (see Appendix). These elements formed the basis of a re
construction of the pelvic girdle illustrated by Romer and Price (1940,
fig. 25). Williston (1914, fig. 11, no. 5) illustrated a small, round pubis
that is notched by the obturator foramen as tentatively diagnostic of the
genus “Arribasaurus.” These characters were later interpreted by Romer
and Price (1940) as ontogenetically, but not taxonomically relevant within
the genus Ophiacodon.
NMMNH P-43121 includes a nearly complete and articulated
pelvis (Fig. 5B). Unless otherwise stated, the following description is
based on the more fully preserved right side (Fig. 5A-C) that notably
includes a nearly complete pubis of a mature individual.
Ilium – In lateral or medial view the preserved anterior portion of
the iliac blade is narrow dorsoventrally and tapers posteriorly to the
level of breakage. An isolated distal end of the posterior process confirms
that the iliac blade ends in a very pointed posterior process. Medial and
lateral walls of the iliac blade form a deep, triangular, dorsally-facing
trough for the axial musculature. (Dorsoventral crushing of the dorsal
trough of the right iliac blade has resulted in widening of its proximal
portion). Ventral to the iliac blade, a short neck is formed with a width of
20 mm. The thickened ventral base of the ilium has a width of 31 mm and
contributes to more than one third of the acetabular surface. The sutural
contact of the ilium with the pubis and ischium form a y-shaped pattern
within the acetabulum.
Pubis – The right and left pubic portions of the median symphy
sis are preserved and meet ventrally along deeply concave, unfinished
surfaces that in life may have been filled with cartilage or connective
tissue. The stout dorsal ridge of the pubis bears a prominent pubic
tubercle more than midway along its length of 70 mm. Its anterior margin
has a truncated appearance terminating in a blunt, unfinished surface that
was likely continued in cartilage. Below the dorsal pubic ridge the bone
thins rapidly, with only fragments of its ventral portion being preserved.
Immediately anteroventral to the anterior edge of the acetabulum the
pubis is pierced by the large, ovoid obturator foramen.
Ischium - The preserved anterodorsal portion of the ischium is
wedge-shaped with a preserved length of 30 mm. Its wide, flat dorsal
margin forms a crescentic ventral acetabular surface. The dorsomedial
edge of the ischium meets the ventral margin of the ilium in an open
suture.
Hind Limb
Femur - A complete, well-preserved right femur has a length of
107 mm (Fig. 5D-G). In ventral view the proximal head is rectangular
with a slightly rounded articular surface with a width of 34 mm. A
smooth, shallow intertrochanteric fossa bears a well-developed internal
trochanter on its anterior border. In these features NMMNH P-43121
conforms closely to the femur of Ophiacodon navajovicus illustrated by
Case (1907, pl. 27, fig. 1).
The shaft is narrow and subcircular in cross section. A weakly
developed adductor ridge arises just below the internal trochanter before
it continues somewhat diagonally across the shaft, with its most promi
nent portion assuming a nearly mid-ventral position. Romer and Price
(1940) illustrated an Ophiacodon navajovicus femur with the typical
Ophiacodon adductor ridge pattern in which the ridge makes an abrupt
posterior turn below the intertrochanteric fossa, after which it extends
distally in a linear path along the posterior edge of the shaft to the
posterior condyle. However, Romer and Price (p. 234) acknowledged
that the O. navajovicus femur available to them was “imperfectly pre
served” and that they reconstructed the ventral ridge system from “only
… a line of rugosities.”
In ventral or dorsal view both distal condyles are convex and fully
ossified, an indication of the maturity of the specimen, with the poste
rior condyle extending much farther distally than that of the anterior
condyle. In direct contrast is a femur illustrated in dorsal view by Williston
(1914, fig. 11, no. 4) in which the distal margin of the bone is almost
straight with no visible development of the articular condyles, an obvi
ously juvenile condition.
Tibia – In anterior or posterior view the proximal end of the tibia
(Fig. 5H-I) is roughly triangular with a maximum width of 31 mm, which
slightly exceeds the range of tibial widths listed by Romer and Price
(1940, table 4) for Ophiacodon navajovicus. The convex articular sur
face (Fig.5J) bears a poorly-preserved cnemial crest. A wide, deep groove
separates the larger anteromedial articular condyle from that of the some
what smaller posterolateral articular condyle. A missing segment sepa
rates the proximal head of the tibia from its shaft, which bears a sharp
ridge along its medial edge and is wedge-shaped in cross section. The
distal margin of the articular surface is convex in dorsal view, but concave
in ventral view, with a width of 21 mm.
Pes
With the exception of the proximal tarsals, the pes of Ophiacodon
navajovicus is largely unknown. Case (1907) illustrates an astragalus in
 143

FIGURE 4. Ophiacodon navajovicus, A-P, NMMNH P-43121. A-C, anterior caudal vertebra in A, left lateral, B, anterior and C, ventral views. D-G, mid
caudal vertebra in D, left lateral, E, ventral, F, anterior and G, posterior views. H-K, posterior caudal vertebra in H, left lateral, I, anterior, J, posterior and
K, ventral views. L-M, left scapular blade fragment in L, lateral and M, medial views. N-P, proximal right humerus in N, ventral, O, dorsal and P, proximal
views. Q-R, right ulna in two fragments in Q, anterior and R, posterior views.
144
FIGURE 5. Ophiacodon navajovicus, A-J, NMMNH P-43121. A, right lateral view of pelvis, C, medial view of right half of pelvis, and B, pelvis in dorsal
view. D-G, right femur in D, dorsal, E, ventral, F, distal and G, proximal views. H-J, right tibia in H, posterior, I, anterior and J, proximal views.
ventral view, whereas Williston (1914, fig. 11, nos. 2 and 6) illustrates
both a juvenile and mature astragalus and calcaneum preserved in articu
lation. The pes of NMMNH P-43121 includes an astragalus, calcaneum,
two articulated tarsals, two complete and several broken metatarsals,
and five partially articulated digits (Fig. 6). These elements were found
closely associated, but not articulated; a reconstruction of the pedal
elements (Fig. 6A) is based on published reconstructions of Ophiacodon
pedes in the literature (e.g. Williston and Case, 1913, fig. 34C-D; Romer
and Price, 1940, fig. 41; Romer, 1956, fig. 195A).
Astragalus - The astragalus is a broad, L-shaped element with an
elongate neck (Fig. 6D-E). The articular facet for the calcaneum is deeply
concave in transverse section and bears a small notch near the distal end
of the contact for passage of the perforating artery. In dorsal view the
distal surface of the horizontal medial process bears a large, rugose facet
for the tibia.
Calcaneum - The calcaneum (Fig. 6B-C) is a wide, ovoid bone
with a broadly concave dorsal surface that nearly reaches the thickened,
roughened peripheral articular margins. Its width is 31 mm. Its lateral
half is thin and flat, whereas its medial portion provides a thickened
wedge for articulation with the astragalus. A small notch for the perforat
ing artery is positioned near the distal end of its contact with the astraga
lus.
Tarsals - Two tarsals are preserved in near articulation with one
another (Fig.6F-G). One, probably the 3rd distal tarsal, is pentangular in
outline with slightly concave dorsal and ventral surfaces. The second is
subequal, but much thicker and quadrangular in outline. The ventral
surface is flat and contrasts with a somewhat rounded dorsal surface that
bears several angular articular surfaces. These elements likely represent
distal tarsals 3 and 4, respectively.
Metatarsals - Metatarsal I, preserved in articulation with its
proximal phalanx, is short with a length of 14 mm and has a greatly
widened proximal articular surface (Fig. 6A, J-K). Another complete
metatarsal with a length of 35 mm most likely represents metatarsal IV
(Fig. 6A, H-I). Portions of three additional metatarsals have broad, flat
 145

FIGURE 6. Ophiacodon navajovicus, A-O, NMMNH P-43121. A, reconstructed right pes in dorsal view. B-C, right calcaneum in B, dorsal and C, ventral
views. D-E, right astragalus in D, dorsal and E, ventral views. F-G, third and fourth tarsals in F, dorsal and G, ventral views. H-I, fourth metatarsal in H,
dorsal and I, ventral views. J-K, articulated pedal digits (from left to right digits 5 to 1) in J, dorsal and K, ventral views. L-O, ungual in L, dorsal, M,
ventral, N, lateral and O, proximal views. Upper left scale bar applies to A, upper right scale bar applies to B-E and middle right scale bar applies to F-O.
146

FIGURE 7. Ophiacodon navajovicus, type and referred specimens. A-E, AMNH 4777, holotype left humerus in A, ventral and B, dorsal views, C, proximal
tibia in anterior view, D, first sacral vertebra in anterior view, E, radius in anterior view. F-G, MCZ 1595(?) incomplete right pelvis in F, lateral(?) and G,
medial(?) views. Left scale bar applies to A-B, middle scale bar applies to C-E and right scale bar applies to F-G.

tened shafts with moderately expanded proximal and distal ends (Fig.
6A).
Phalanges - Articulated phalanges form segments offive digits of
the right pes (Fig. 6A, J-K). Digit 1 is represented by a single phalanx
preserved in articulation with a short metatarsal described above. A prob
able digit 2 includes two slender, articulated phalanges, whereas digit 3
consists of three articulated phalanges that collectively curve slightly to
the left. Digit 4 includes three articulated phalanges and the distal portion
of a third. As articulated the digit is arched slightly dorsally and strongly
curved medially. Digit 5 includes the proximal two articulated phalanges
in which the first is twice the length of the second. Three disarticulated
ungual phalanges, one of which is illustrated in Figure 6L-O, are some
what more laterally compressed than in other species of Ophiacodon.
Their positions shown in the reconstruction (Fig. 6A) are based on size.
DISCUSSION
As currently understood, the postcranial skeleton of Ophiacodon
navajovicus is virtually indistinguishable from that of Ophiacodon mirus,
the only two species of Ophiacodon known from Cutler Group strata of
northern New Mexico (Romer and Price, 1940; Lewis and Vaughn, 1965;
Berman, 1993). Romer and Price (1940) listed only four characters by
which O. navajovicus can be specifically separated from O. mirus: (1)
linear dimensions 4/5 that of O. mirus; (2) ilium more pointed posteri
orly (restated by Reisz [1986, p. 69] as “pointed posterior process of
the iliac blade”); (3) pubic tubercle better developed than in other species
of Ophiacodon in which it is very small or represented only by a rugos
ity; and (4) ventral ridge of femur represented only by a line of rugosities
(but see above description of femur). In addition to these features, how
ever, the postcranial skeleton of O. navajovicus can be readily distin
guished from that of O. mirus on the basis of the ventral surface of the
centrum.
In Ophiacodon mirus variation of the ventral surface of the cen
trum within a single vertebral column typically progresses from a sharp
midventral keel in the cervicals to a thicker, rounded midventral keel in
the dorsals, lumbars and sacrals, followed by anterior caudals with a
sharply keeled, midventral ridge, which becomes smaller in the mid
caudals, where it lies between parallel grooves, and is replaced in the
distal caudals by a rounded ventral surface (Williston, 1914; Kissel and
Lehman, 2002). Thus, the ventral surface of the centrum in O. mirus is
always keeled, but varies from sharp to rounded.
Ophiacodon navajovicus is characterized by a significantly dif
ferent morphological pattern of the ventral surface of the centrum (Figs.
2M and R; 3D, H and N; 4C, E and K). In this species the ventral surface
of the axis typically possesses a sharp midventral keel, which is fol
lowed abruptly in the postaxial cervicals by a distinctively flat midventral
ridge that is defined by straight longitudinal edges. The narrow, flattened,
midventral ridge is retained posteriorly through the presacral column to
the sacrals. The anterior caudals possess a sharply keeled, midventral
 147

FIGURE 8. Ophiacodon navajovicus, referred specimens. A-I, AMNH 4793. A-B, right scapular blade in A, lateral and B, medial views. C-D, left humerus
in C, ventral and D, distal dorsal views. E-F, left acetabulum in E, lateral and F, medial views. G-H, proximal femur in G, ventral and H, dorsal views. I, two
dorsal vertebrae in right lateral view.

ridge, but in the mid-caudals the midventral ridge varies from a somewhat
rounded to a v-shaped keel, which is followed in the most distal caudals
by a flattened ventral surface.
Thus, the ventral surface of the centrum in Ophiacodon navajovicus
is similar to that of Ophiacodon mirus only in the axis and anterior and
middle sections of the caudal region. All other regions of the vertebral
column in O. navajovicus, including most notably the postaxial cervicals
and the distal caudals, are characterized by a narrow, flat, midventral keel
with well-defined lateral margins.
The tendency toward a greater ventral flattening of the centra in
Ophiacodon navajovicus when compared with other species of
Ophiacodon is present also in two genera considered to be primitive
members of a clade that includes Ophiacodon (Brinkman and Eberth,
1986). However, in these two taxa, Baldwinonus, collected from the
Upper Pennsylvanian El Cobre Canyon Formation of Cañon del Cobre
(Romer and Price, 1940), and Stereophallodon, from the Lower Permian
148
Markley Formation (Wichita Group) of Texas (Romer, 1937), the pre
sacral centra can be distinguished from those of O. navajovicus by their
greater size and subrectangular cross-sectional outline.
An additional primitive character-state shared by Ophiacodon
navajovicus and Stereophallodon is the mid-ventral position of the ad
ductor ridge of the femur. Although the adductor ridge is positioned
similarly in two other ophiacodontid genera, Ruthiromia elcobriensis
from Lower Permian (Wolfcampian) strata of Cañon del Cobre, (Eberth
and Brinkman, 1983; Brinkman and Eberth, 1986; Spielmann and Lucas,
this volume) and an ophiacodontid assigned to “Ophiacodon cf. O. mirus
(sic)” from the Upper Pennsylvanian (middle Virgilian) Ada Formation
of Oklahoma (Kissel and Lehman, 2002), it is developed in these taxa to
a much greater degree, so that it forms a well-ossified, prominent crest.
This condition is in sharp contrast to that of O. navajovicus and
Stereophallodon in which a weakly developed ridge is formed from a line
of rugosities (Brinkman and Eberth, 1986).
The morphology of the ventral surface of the centrum in
Ophiacodon navajovicus is potentially advantageous in the identifica
tion of isolated postaxial through precaudal centra when present in com
bination with the typical ophiacodontid wing-shaped transverse process
in which a web of bone extends from the diapophyses to the parapophysis
and the centrum has a wedge-shaped cross-sectional outline. Therefore,
O. navajovicus can be considered of potential use in Late Pennsylvanian
tetrapod biostratigraphy, provided its stratigraphic range can be well
established.
Lucas et al. (2005) constructed a detailed vertebrate biostratigra
phy in the Pennsylvanian-Permian redbeds of the El Cobre Canyon
Formation in Cañon del Cobre that placed into precise lithostratigraphic
position three, temporally successive and distinct vertebrate fossil as
semblages that represent three land-vertebrate faunachrons (lvfs). The
middle and upper assemblages belong to the Coyotean and Seymouran
lvfs that are of Virgilian-Wolfcampian and late Wolfcampian age, respec
tively. Ophiacodon navajovicus is restricted to the lower assemblage at
Cañon del Cobre that represents an older, entirely Pennsylvanian lvf, the
Cobrean.
The Cobrean lvf is defined as beginning with the FAD (first
appearance datum) of the temnospondyl Eryops, and ending with the
beginning of the Coyotean, defined by the FAD of the pelycosaurian
grade synapsid Sphenacodon. Index taxa of Cobrean time currently in
clude the diadectids Diasparactus and Desmatodon and the limnoscelid
Limnoscelis. Other Pennsylvanian tetrapod fossils of this assemblage,
according to Lucas (2005), include Eryops sp., the trematopid
REFERENCES
Brinkman, D. and Eberth, D.A., 1986, The anatomy and relationships of
Stereophallodon and Baldwinonus (Reptilia, Pelycosauria): Breviora,
no. 485, p. 1-34.
Case, E. C., 1907, Revision of the Pelycosauria of North America: Carnegie
Institute of Washington, Publication 55, 176 p.
Cope, E. D., 1888, Systematic catalogue of the species of Vertebrata found
in the beds of the Permian epoch in North America, with notes and
descriptions: American Philosophical Society Transactions, v. 16, p.
285-297.
Eberth, D.A. and Brinkman, D., 1983, Ruthiromia elcobriensis, a new
pelycosaur from El Cobre Canyon, New Mexico: Breviora, no. 474, p.
1-26.
Kissel, R. A. and Lehman, T. M., 2002, Upper Pennsylvanian tetrapods
from the Ada Formation of Seminole County, Oklahoma: Journal of
Paleontology, v. 76, p. 529-545.
Lewis, G. E. and Vaughn, V. V., 1965, Early Permian vertebrates from the
Cutler Formation of the Placerville area, Colorado: U. S. Geological
Survey, Professional Paper, 503-C, 50 p.
Lucas, S. G., 2006, Global Permian tetrapod biostratigraphy and
temnospondyl Anconastes vesperus, the dissorophid Platyhystrix rugusus,
the edaphosaurid Edaphosaurus cf. E. novomexicanus and indetermi
nate eupelycosaurs.
To date, only two potential records of Ophiacodon navajovicus
are known outside of Cañon del Cobre, both of which are in Cutler
Group strata of the Four Corners area. The first of these is from sedi
ments of the Halgaito Formation near Mexican Hat, Utah (Vaughn, 1962).
Arguments regarding the age of the Halgaito Formation are complex,
leaving the position of the Virgilian-Wolfcampian boundary in that unit
unclear. In addition to postcranial material identified as Ophiacodon cf.
O. navajovicus, these strata yield a Coyotean tetrapod assemblage that
includes Diplocaulus, Phlegethonia?, a trimerorhachid, Eryops,
Platyhystrix, Archeria, a limnoscelid, Edaphosaurus, Sphenacodon and
an araeoscelid? (Lucas, 2006).
In the Placerville area of southwestern Colorado the upper part of
the Cutler Formation yielded vertebrae that we believe can be assigned to
Ophiacodon navajovicus based on the description and illustration pro
vided by Lewis and Vaughn (1965, fig. 10A). (See discussion above in
Previous Studies section.) Because the Cutler Formation in this area
remains undivided, the exact age of the strata is unknown. However,
strata of the Placerville area have yielded a tetrapod assemblage that
includes Eryops, Platyhystrix, a seymouriid, Diadectes, a captorhinid?
and a haptodontid (Lewis and Vaughn, 1965; Wideman et al., 2005), taxa
characteristic of the Coyotean lvf (Lucas, 2006), strongly suggesting the
section from which these fossils were recovered encompasses the Virgilian
Wolfcampian boundary.
Clearly, detailed stratigraphic and biochronological resolution,
currently lacking for both the Halgaito Formation, as well as strata of the
Placerville area in the upper part of the Cutler Formation, is needed so
that all localities yielding Ophiacodon navajovicus fossils can be placed
into precise lithostratographic order enabling the lowest occurrence of
Sphenacodon relative to that of O. navajovicus to be determined. Until
such time, O. navajovicus is best considered to be restricted to the
Cobrean lvf, as it is in the El Cobre Canyon Formation in Cañon del
Cobre, the only location in North America where substantial assem
blages of vertebrate fossils are documented as closely bracketing the
Pennsylvanian-Permian boundary.
ACKNOWLEDGMENTS
We thank AMNH and MCZ for collection access. Al Lerner,
Sterling Nesbitt, Randy Pence and Larry Rinehart assisted in fieldwork.
Careful reviews by Dave Berman, Amy Henrici and Sean Modesto greatly
improved the manuscript.
biochronology; in Lucas, S. G., Cassinis, G. and Schneider, J. W., eds.,
Non-marine Permian biostratigraphy and biochronology: London, Geo
logical Society, Special Publications, v. 265, p. 65-93.
Lucas, S. G., Harris, S. K., Spielmann, J. A., Berman, D. S and Henrici, A. C.,
2005, Vertebrate biostratigraphy and biochronology of the Pennsylva
nian-Permian Cutler Group, El Cobre Canyon, northern New Mexico:
New Mexico Museum of Natural History and Science, Bulletin 31, p.
128-139.
Lucas, S. G. and Krainer, K., 2005, Stratigraphy and correlation of the
Permo-Carboniferous Cutler Group, Chama Basin, New Mexico: New
Mexico Geological Society, Guidebook 56, p. 145-159.
Reisz, R. R., 1986, Pelycosauria: Encyclopedia of Paleoherpetology, Part
17A, 102 p.
Romer, A.S., 1937, New genera and species of pelycosaurian reptiles: Pro
ceedings of the New England Zoological Club, 16, p. 89-96.
Romer, A. S., 1956, Osteology of the reptiles: The University of Chicago
Press, 772 p.
Romer, A. S. and Price, L. I., 1940, Review of the Pelycosauria: Geological
Society of America Special Paper, v. 28, p. 1-538.

Sumida, S. S., Lombard, R. E., Berman, D. S and Henrici, A. C., 1999, Late
Paleozoic amniotes and their near relatives from Utah and northeastern
Arizona, with comments on the Permian-Pennsylvanian boundary in
Utah and northern Arizona; in Gillette, D., ed., Vertebrate paleontology
in Utah: Utah Geological Survey, Miscellaneous Publication, 99-1, p.
31-43.
Vaughn, P. P., 1962, Vertebrates from the Halgaito Tongue of the Cutler
Formation, Permian of San Juan County, Utah: Journal of Paleontology,
v. 36, p. 529-539.
Wengerd, S. A. and Matheny, M. L., 1958, Pennsylvanian System of Four
149
Corners region: Bulletin of the American Association of Petroleum
Geologists, v. 42, p. 2048-2106.
Wideman, N. K., Sumida, S. S. and O’Neil, M., 2005, A reassessment of the
taxonomic status of the materials assigned to the Early Permian tetra
pod genera Limnosceloides and Limnoscelops: New Mexico Museum of
Natural History and Science, Bulletin 31, p. 358-362.
Williston, S. W., 1914, The osteology of some American Permian verte
brates: Journal of Geology, v. 22, p. 364-419.
Williston, S. W. and Case, E. C., 1913, Description of a nearly complete
skeleton of Ophiacodon Marsh: Carnegie Institution of Washington,
Publication no. 181, p. 37-59.

APPENDIX
Referred specimens of Ophiacodon navajovicus

AMNH
FR 4776: Forefoot, mostly articulated.
FR 4777: Partial skeleton, vertebrae and forelimbs, humerus, proximal and distal ulna, parts of femur, distal ?tibia and fibula, astragalus
(Fig. 7A-E).
FR 4781: Hindfoot, tibia and parts of other limb bones and several vertebrae.
FR 4783: Fragmentary skeleton (with no. 4782) chiefly vertebrae and foot bones, proximal and distal ulna, proximal tibia and fibula.
FR4784: Fragmentary skeleton including dorsal, sacral and caudal vertebrae, parts of various limb and foot bones, right and left proximal
femora, etc.
FR4789: Skeleton fragments, vertebrae, limb and foot bones.
FR 4791: Skeleton fragments, vertebrae, limb bone fragments, femur, tibia and fibula, etc.
FR 4793: Humerus, half of femur, partial pelvis, vertebrae and foot bones, astragalus and 3 podials and phalanges (Fig. 8A-I).
FR4794: Skull and jaw fragments, articular distal tibia, ulna?, phalanges and scrap.
FR 4799: vertebrae.
FR 4801: Scapulocoracoid fragment, partial ilium, 4 proximal ends and distal end of femora, vertebra and bone fragment.

MCZ
MCZ 1595: An incomplete right pelvic girdle, consisting of a complete ischium and incomplete ilium. Collected by White and Price in 1935
from El Cobre Canyon. They originally identified it as a left pelvic girdle and incorrectly interpreted the ischium as the ilium.
Though originally assigned to Ophiacodon navajovicus by Romer and Price (1940), given the lack of any diagnostic features of
the specimen, its referral to O. navajovicus is questionable (Fig. 7F-G).
MCZ 2007: A large lot of numerous boxes of postcranial elements collected from Cañon del Cobre. Elements include jaw fragments,
fragmentary pectoral girdles, femora, tibiae and fibulae. At least one femur fragment from these lots appears to have the ventral
ridge characteristic of Ophiacodon navajovicus, much of the other material is probably not diagnostic.
MCZ 8079: Nine vertebral centra, some have square ventral keels, while others have triangular ventral keels. The centra with the square
ventral keels can be referred with confidence to Ophiacodon navajovicus. No locality information listed beyond “From El Cobre
Canyon.”
150

Arthropleura against a Late Pennsylvanian moon, 300 year million years ago in Cañon del Cobre, New Mexico. Artwork by Matt Celeskey.
 Lucas et al., eds., 2010, Carb-Permian transition in Cañon del Cobre. New Mexico Museum of Natural History and Science Bulletin 49.

151
RE-EVALUATION OF RUTHIROMIAELCOBRIENSIS (EUPELYCOSAURIA:
OPHIACODONTIDAE?) FROM THE LOWER PERMIAN
(SEYMOURAN?) OF CAÑONDELCOBRE, NORTHERN NEWMEXICO

JUSTINA. SPIELMANN AND SPENCER G. LUCAS

New Mexico Museum of Natural History and Science, 1801 Mountain Rd NW, Albuquerque, NM 87104-1375

Abstract—Ruthiromia elcobriensis is a eupelycosaur from the Lower Permian strata of the Cañon del Cobre, Rio
Arriba County, New Mexico. The taxon is known only from the holotype, a partial skeleton consisting of skull
fragments, isolated dorsal vertebrae, articulated vertebral series of “lumbar,” sacral and anterior caudals, incom
plete pelvic girdle, left humerus and much of the right hindlimb. The type locality of R. elcobriensis is imprecisely
known, but available data suggest it was collected from the lower part of the Arroyo del Agua Formation,
considered Seymouran in age and equivalent to the late Wolfcampian marine stage. Ruthiromia elcobriensis remains
a problematic taxon within the Eupelycosauria. Re-evaluation identifies only two diagnostic characters of the
taxon, the pinched-in presacral centra and distally expansive deltopectoral crest. Comparisons to othereupelycosaurs
indicate that R. elcobriensis displays morphologic features that are present in both varanopids and ophiacodontids,
and given this combination of characteristics R. elcobriensis is assigned tentatively to the Ophiacodontidae, and
likely a close relative of Ophiacodon navajovicus.

INTRODUCTION STRATIGRAPHIC POSITION AND AGE

Eupelycosaurs were the top terrestrial predators of the Early
Permian and significant collections of various eupelycosaur taxa have
been made from Texas and New Mexico. Many eupelycosaurs are known
from numerous specimens (e.g., Sphenacodon from the Arroyo del Agua
bonebeds of north-central New Mexico: Lucas et al., 2005b) and thus
provide an extensive comparative record. Unfortunately, no recent at
tempts have been made to significantly revise eupelycosaurian relation
ships, leaving many species with very general diagnoses that are anec
dotal in nature and genera that are taxonomically oversplit (e.g., 11 recog
nized species of Dimetrodon characterized mostly by roughly defined
size grades: Reisz, 1986). As a result, taxonomic uncertainty makes
incomplete eupelycosaurian specimens difficult to diagnose, especially
when lacking a complete skull or dorsal vertebrae (e.g., the non-axial
postcrania of Dimetrodon and Sphenacodon are identical).
Unfortunately, this is the case with Ruthiromia elcobriensis, a
taxon named by Eberth and Brinkman (1983) based on an incomplete
skeleton, consisting of the occipital and quadrate fragments, posterior
left mandible, atlantal centrum and axial intercentrum, cervical rib, four
disarticulated dorsal vertebrae, a disarticulated neural arch and partial
neural spine, an incomplete, articulated vertebral column and pelvis, and
an incomplete distal portion of the right hindlimb. Eberth and Brinkman
(1983, p. 22-23) assigned Ruthiromia elcobriensis to the Varanopidae
based on it possessing “excavated neural arches and the presence of a
shallow fossa on the proximal surface of the deltopectoral crest,” fea
tures that in combination are unique to Varanopidae (Langston and Reisz,
1981).
One of us (JAS) recently re-examined the holotype and only
known material of Ruthiromia elcobriensis and found discrepancies in
the original description of Eberth and Brinkman (1983). Here, we discuss
the stratigraphic position and probable age of Ruthiromia elcobriensis,
modify its diagnosis, reevaluate its taxonomic position within
Eupelycosauria and photographically document portions of the holo
type for the first time.
Institutional abbreviations: MCZ, Museum of Comparative
Zoology, Harvard University, Cambridge, MA; and NMMNH, New
Mexico Museum of Natural History and Science, Albuquerque NM.
Eberth and Brinkman (1983, p. 1-2) provided the following sum
mary of the stratigraphic position of Ruthiromia elcobriensis: “In 1965
A. Lewis and S. Olsen collected the larger part of the postcranial skeleton
and disarticulated cranial fragments of a pelycosaur from the west wall of
the canyon.” The canyon they refer to is Cañon del Cobre (= El Cobre or
Copper Canyon), Rio Arriba County, New Mexico. Most of the strata
exposed in Cañon del Cobre are Late Pennsylvanian to Early Permian in
age (Lucas and Krainer, 2005; Lucas et al., 2005a, this volume), including
the west wall where the holotype of R. elcobriensis was collected. We
attempted to relocate the excavation site of the holotype of R. elcobriensis,
but failed because of the lack of specific information about the site’s
position.
Based on the detailed stratigraphy of Lucas and Krainer (2005)
and mapping of Kempter et al. (2007: though see modifications by Lucas
et al., this volume) we can now tentatively indicate the stratigraphic
position of the site where the Ruthiromia elcobriensis holotype was
collected. Along the west wall of the Cañon del Cobre the Arroyo del
Agua Formation is exposed, with the underlying El Cobre Canyon For
mation forming the canyon floor (Kempter et al., 2007; Lucas et al., this
volume). The lithology of these formations is distinct, with the El Cobre
Canyon Formation being brown in color with numerous rhizolith hori
zons and the Arroyo del Agua Formation being orange with abundant
calcrete nodules. The “orange” (moderately reddish brown) siltstone
matrix of the partially prepared vertebral column of R. elcobriensis,
(Lucas and Krainer, 2005, table 1) indicates the holotype was collected
from the Arroyo del Agua Formation. This then means that it is from the
uppermost tetrapod assemblage zone in Cañon del Cobre (Lucas et al.,
2005a), which includes the diadectomorph Diadectes, the sphenacodontid
eupelycosaur Sphenacodon ferox and the dissorophid amphibian
Platyhystrix. Lucas et al. (2005a) and Lucas (2005, 2006) assigned this
upper assemblage zone to the Early Permian Seymouran land-vertebrate
faunachron, based on correlations with the type section of the Arroyo
del Agua Formation at Mesa Montosa, where Seymouria, the principal
index taxon of the Seymouran, is present. Lucas (2006) indicates that the
Seymouran lvf is equivalent to the late Wolfcampian marine stage.
152
SYSTEMATIC PALEONTOLOGY
Suborder EUPELYCOSAURIA
Family ?OPHIACODONTIDAE
Genus Ruthiromia Eberth and Brinkman, 1983
Ruthiromia elcobriensis Eberth and Brinkman, 1983
Figs. 1-5
1983 Ruthiromia elcobriensis: Eberth and Brinkman, figs. 2-8, 10
15.
Holotype: MCZ 3150, partial skeleton consisting of occiput and
quadrate fragments (Fig. 1B-C), a posterior portion of the left mandible,
atlantal centrum and axial intercentrum, a cervical rib, four disarticulated
dorsal vertebrae (e.g., Fig. 1A); an isolated disarticulated neural arch and
partial neural spine, an incomplete, articulated vertebral column and
pelvis (Fig. 2), left humerus (Fig. 3) and an incomplete, distal portion of
a right hindlimb (Figs. 4-5).
Referred specimens: Only known from the holotype.
Type locality and horizon: Arroyo del Agua Formation, Cutler
Group in Cañon del Cobre, Rio Arriba County, New Mexico. See above
for interpretation of the stratigraphic position of the type locality, exact
collection site unknown.
Revised diagnosis (modified from Eberth and Brinkman,
1983): A large eupelycosaur distinguished from all others by the follow
ing features: dorsal and lumbar centra pinched-in laterally, just below the
level of the notochordal pit, giving the centra an hourglass shape in cross
section; the deltopectoral crest of the humerus is massively developed
and continues distally as a sharp ridge below the area of the pectoralis
muscle insertion.
Description: Eberth and Brinkman (1983) described the holo
type Ruthiromia elcobriensis in detail, obviating the need for further
of
description here (other than the emendations noted below).
Discussion: Eberth and Brinkman (1983, p. 2) diagnosed
Ruthiromia elcobriensis as a varanopid on the basis of the five following
characters “[1] linear measurements approximately 30 percent greater
than those of Varanodon…[2] the first eight centra directly anterior to
the pelvis deeply pinched-in laterally, just below the level of the noto
chordal pit, giving the centra an hour-glass appearance; [3] each of the
eight centra displays a broad rounded ventral surface; [4] the deltopectoral
crest of the humerus is massively developed and continues distally as a
sharp ridge below the area of pectoralis muscle insertion; [5] the neck of
FIGURE 1. Ruthiromia elcobriensis, MCZ 3150, holotype. A, isolated dorsal
vertebra in left lateral view, note pinched-in centrum and broad rounded
ventral centrum surface (illustrated in Eberth and Brinkman, 1983, fig.
7[left]). B-C, opisthotic/supraoccipital fragment in B, posterior and C,
anterior views. Abbreviations: op, opisthotic; st, supratemporal.
the astragalus represents 35 percent of the total height of that element.”
Our reexamination of the holotype and relevant literature allows an evalu
ation of these characteristics:
[1] Linear measurements approximately 30 percent greater than
those of Varanodon: Eberth and Brinkman (1983) interpreted Ruthiromia
elcobriensis as a varanopid and used its large size to distinguish it from
all other varanopids. However, this was prior to the recognition of
Watongia meireri as a varanopid and collection of additional larger speci
mens of Varanodon agilis, which have larger humeri than R. elcobriensis
(humeral length: R. elcobriensis = 105.6 mm; Watongia meieri = 130 mm;
Varanodon agilis = 124 mm; data from Eberth and Brinkman, 1983, table
1 and Reisz and Laurin, 2004, table 1). Thus, Ruthiromia is no longer the
largest varanopid. We interpret Ruthiromia as demonstrating characteris
tics of both varanopids and ophiacodontids, and cannot assign it to
either family with confidence (see below for further discussion). The
size of R. elcobriensis is not considered a useful diagnostic feature, as its
size, based on femur length, is considerably smaller than the largest
species of Ophiacodon (femoral length of R. elcobriensis: 116.5 mm
from Eberth and Brinkman, 1983, p. 9, table 1 and Ophiacodon
retroversus: 140-192 mm from Romer and Price, 1940, p. 465, table 4).
[2] Dorsal and lumbar centra deeply pinched-in laterally, just
below the level of the notochordal pit, giving the centra an hour-glass
appearance: this is the most distinctive feature of Ruthiromia elcobriensis
and contrasts with the morphology of the centra of all other non-varanopid
eupelycosaurs (Fig. 6). We have modified the original character as listed
by Eberth and Brinkman (1983) because the three isolated dorsal centra
described also possess this character, indicating this feature may be
present in all presacral centra, not just the centra immediately anterior to
the sacrum.
[3] Each of the eight [posteriormost presacral] centra displays a
broad rounded ventral surface: we have removed this feature from the
diagnosis because it is also present in various ophiacodontids (e.g.,
Stereophallodon and Clepsydrops: Romer and Price, 1940, fig. 17) and
thus cannot be used to distinguish Ruthiromia elcobriensis from all other
pelycosaurian-grade synapsids. However, it can be noted that these
ophiacodontids lack the pinched-in presacral centra of R. elcobriensis.
[4] The deltopectoral crest of the humerus is massively developed
and continues distally as a sharp ridge below the area of pectoralis muscle
insertion: this prominent crest is retained in the diagnosis of R. elcobriensis,
as it distinguishes this taxon from all other pelycosaurian-grade synapsids.
This feature is also independent of size, as Dimetrodon limbatus, which
is larger than R. elcobriensis (humerus length of R. elcobriensis: 105.6
mm from Eberth and Brinkman, p. 9, table 1 and D. limbatus: 180-208
mm from Romer and Price, 1940, p. 456, table 4) lacks such a distal
extension of the deltopectoral crest (Romer and Price, 1940, pl. 29C and
Brinkman, 1988, fig. 3).
[5] The neck of the astragalus represents 35 percent of the total
height of that element: we have removed this feature from the diagnosis
because a long neck of the astragalus is also present in Ophiacodon
navajovicus (Harris et al., this volume, fig. 9a, d-e). In O. navajovicus,
NMMNH P-43121, the height of the neck of the astragalus (13.3 mm) is
38.8 percent of the total height of the astragalus (34.3 mm). Thus, a
relatively tall astragular neck cannot be used to separate Ruthiromia
elcobriensis from all other pelycosaurian-grade synapsids.
Overall, only two of the five characteristics originally diagnosed
by Eberth and Brinkman (1983) are still useful in identifying Ruthiromia
elcobriensis. However, the pinched-in dorsal and lumbar centra and dis
tally expansive deltopectoral crest are unique among eupelycosaurs and
allow R. elcobriensis to be recognized from isolated postcranial fossils.
TAXONOMIC PLACEMENT
Removal from the Varanopidae
As noted above, Eberth and Brinkman (1983) identified two fea
tures – the excavated neural arches and the presence of a shallow fossa on
 153

FIGURE 2. Ruthiromia elcobriensis, MCZ 3150, holotype. A-C, vertebral column and pelvis. A-B, vertebral column in A, left lateral and B, dorsal views.
C, Close-up of dorsal neural spine and arch in dorsolateral view. Vertebrae are numbered in series, with 1 being anterior. Abbreviations: c, centrum; ex,
excavation in neural spines; is, ischium; n, neural spines: na, neural arch; p, pubis.
154
FIGURE 3. Ruthiromia elcobriensis, MCZ 3150, A-F, holotype left humerus in A, dorsal, B, ventral, C, posterior, D, anterior, E, proximal and F, distal
views. Abbreviations: dpc, deltopectoral crest; ent, entepicondyle.
the proximal deltopectoral crest–that identified Ruthiromia elcobriensis
as a varanopid, following Langston and Reisz (1981), who noted that
these two features in combination are unique to the Varanopidae. Our
reexamination of the holotype of R. elcobriensis has called these two
features into question, and in so doing modified the taxonomic placement
of R. elcobriensis.
Most of the vertebrae of Ruthiromia elcobriensis are represented
by incomplete centra. Only four of the eight “lumbar” vertebrae (verte
brae 5-8 of the presacral series) possess enough neural that are complete
enough for identification of excavations. Of these four, only two are
completely unobstructed, whereas the other two have matrix filling the
neural arch excavations. Thus, only four “lumbar” vertebrae demonstra
bly possess the excavated neural arches (Fig. 2), and this is the only
character shared between varanopids and R. elcobriensis, compared to
the centra and femoral morphology common to both ophiacodontids and
R. elcobriensis.
 155

FIGURE 4. Ruthiromia elcobriensis, MCZ 3150, A-F, holotype right femur in A, dorsal, B, posterior, C, anterior, D, ventral, E, proximal and F, distal
views. G-J, holotype left femur fragment in G, ventral, H, anterior, I, dorsal and J, proximal views. K-M, holotype fibula in K, lateral, L, medial and M,
proximal views. Abbreviations: ar, adductor ridge.
156
FIGURE 5. Ruthiromia elcobriensis, MCZ 3150, A-B, incomplete right tibia and pes in A, dorsal and B, dorsolateral views.
The shallow fossa on the deltopectoral crest of the humerus could
not be confirmed in our examination of the holotype. It appears to be the
result of damage to the specimen and is far less prominent than the
deltopectoral crest fossa in some varanopids (compare Fig. 3B-E and
Eberth and Brinkman, 1983, fig. 11B-C, with the humerus of Aerosaurus
greenleeorum illustrated by Langston and Reisz, 1981, fig. 15B-D and
the Richards Spur varanodontine illustrated by Maddin et al., 2006, fig.
4A-C).
Placement within the Ophiacodontidae
Whereas Reisz (1986) only tentatively placed Ruthiromia
elcobriensis in the Varanopidae, our re-evaluation of the taxon has led us
to place it outside of this family and place it tentatively within the
Ophiacodontidae. R. elcobriensis possesses excavated neural arches, but
lacks a deltopectoral crest fossa, precluding it from inclusion in the
Varanopidae. Berman et al. (1995) provided a diagnosis for the
Ophiacodontidae that included ilium possesses dorsal groove or trough
and posterior process of ilium is slender and extends close to the level of
the posterior limit of the ischium, two features present in R. elcobriensis
(compare R. elcobriensis in Eberth and Brinkman, 1983, fig. 12a-b to
Ophiacodon in Reisz, 1986, fig. 22b-c). In addition, both R. elcobriensis
and Ophiacodon navajovicus share an adductor ridge that extends across

FIGURE 6. Dorsal vertebrae of Ruthiromia elcobriensis compared to various
other eupelycosaurs in cross section. Images from Romer and Price (1940,
fig. 17) and Eberth and Brinkman (1983, figs. 7-8). Higher level taxonomy
after Reisz (1986). Note pinched-in centra of R. elcobriensis.
the shaft with its most prominent portion assuming a nearly mid-ventral
position (Harris et al., this volume), though the adductor ridge of R.
elcobriensis is more well developed, extending away from the femoral
shaft a much greater distance than the adductor ridge of O. navajovicus
(Fig. 4). The presence of adductor ridges on the femora of both these taxa
REFERENCES
Berman, D.S, Reisz, R.R., Bolt, J.R. and Scott, D., 1995, The cranial anatomy
and relationships of the synapsid Varanosaurus (Eupelycosauria:
Ophiacodontidae) from the Early Permian of Texas and Oklahoma:
Annals of Carnegie Museum, v. 8, p. 172-180.
Brinkman, D., 1988, Size-independent criteria for estimating relative age in
Ophiacodon and Dimetrodon (Reptilia, Pelycosauria) from the Admiral
and lower Belle Plains formations of west-central Texas: Journal of
Vertebrate Paleontology, v. 8, p. 172-180.
Eberth, D.A. and Brinkman, D., 1983, Ruthiromia elcobriensis, a new
pelycosaur from El Cobre Canyon, New Mexico: Breviora, no. 474, 26
p.
Evans, D.C. and Reisz, R.R., 2006, An Early Permian varanodontine
varanopid (Synapsida: Eupelycosauria) from the Richards Spur locality,
Oklahoma: Journal of Vertebrate Paleontology, v. 26, p. 957-966.
Harris, S.K., Lucas, S.G. and Spielmann, J.A., 2010, The postcranial skel
eton of Ophiacodon navajovicus (Eupelycosauria: Ophiacodontidae)
from the Upper Pennsylvanian of Cañon del Cobre, New Mexico: New
Mexico Museum of Natural History and Science, Bulletin 49.
Kempter, K., Zeigler, K., Koning, D. and Lucas, S., 2007, Preliminary
geologic map of the Canjilon SE quadrangle, Rio Arriba County, New
Mexico: New Mexico Bureau of Geology and Mineral Resources, Open
file Geological Map 150, 1:24,000.
157
eliminates one of the diagnostic features of O. navajovicus listed by
Romer and Price (1940) and Reisz (1986), leaving the pointed posterior
process of the ilium and the well-developed pubic tubercle as the diag
nostic characteristics of the taxon. Also, R. elcobriensis possesses centra
morphology similar to other ophiacodontids and astragalus proportions
comparable to O. navajovicus, to the exclusion of all othereupelycosaurs.
Nevertheless, O. navajovicus cannot be synonymous with R. elcobriensis,
as it lacks excavated neural arches, pinched-in dorsal and lumber centra
and a prominent distal expansion of the deltopecotral crest.
Eberth and Brinkman (1983, fig. 16C) provided four possible
taxonomic positions for Ruthiromia elcobriensis based on the derived
feature of excavated neural arches: as the sister taxon to the Varanopidae;
the sister taxon to (Edaphosauridae + Sphenacodontidae); within
Edaphosauridae as the sister taxon to an undescribed taxon from the
Pennsylvanian of Kansas; and the sister taxon to the Sphenacodontidae.
It should be noted that Eberth and Brinkman (1983, fig. 16C) placed the
Varanopidae as the sister taxon to the Caseidae, a phylogenetic position
not followed by later studies (e.g., Reisz, 1986, fig. 41). Reisz (1986, fig.
41) placed Varanopidae as the sister taxon to (Ophiacodontidae +
(Edaphosauridae + (Sphenacodontidae + Therapsida))). This study fol
lows most other classifications that place Varanopidae and
Ophiacodontidae as closely related, with either Varanopidae (Reisz, 1986,
fig. 41) or Ophiacodontidae (Reisz, 1986, fig. 40[lower] modified from
Reisz, 1980, fig. 17) as the more inclusive taxon. Unfortunately, Reisz
(1986, table 3) did not provide any postcranial characteristics to identify
the various eupelycosaurian families in his analysis, though R. elcobriensis
does possess a disproportionately large entepicondyle (Fig. 3), a feature
used by Reisz (1986, p. 66) to identify the Ophiacodontidae. Given the
similarities in the femoral morphology and proportions of the astragalus
between R. elcobriensis and Ophiacodon navajovicus we tentatively
place R. elcobriensis in the Ophiacodontidae, noting that it is likely most
closely related to O. navajovicus.
ACKNOWLEDGMENTS
Chuck Schaff provided collection access for JAS at the MCZ.
Susan Harris, David Berman and Stuart Sumida provided helpful reviews
that improved the manuscript.
Langston, W., Jr. and Reisz, R.R., 1981, Aerosaurus wellesi, new species, a
varanopseid mammal-like reptile (Synapsida: Pelycosauria) from the
Lower Permian of New Mexico: Journal of Vertebrate Paleontology, v.
1, p. 73-96.
Lucas, S.G., 2005, Permian tetrapod faunachrons: New Mexico Museum of
Natural History and Science, Bulletin 30, p. 197-201.
Lucas, S.G., 2006, Global Permian tetrapod biostratigraphy and
biochronology; in Lucas, S.G., Cassinis, G. and Schneider, J.W., eds.,
Non-marine Permian biostratigraphy and biochronology: Geological
Society, London, Special Publication 265, p. 65-93.
Lucas, S.G. and Krainer, K., 2005, Stratigraphy and correlation of the
Permo-Carboniferous Cutler Group, Chama Basin, New Mexico: New
Mexico Geological Society, Guidebook 56, p. 145-159.
Lucas S.G., Harris, S.K., Spielmann, J.A., Berman, D.S and Henrici, A.C.,
2005a, Vertebrate biostratigraphy and biochronology of the Pennsylva
nian-Permian Cutler Group, El Cobre Canyon, northern New Mexico:
New Mexico Museum of Natural History and Science, Bulletin 31, p.
128-139.
Lucas, S.G., Harris, S.K., Spielmann, J.A., Berman, D.S, Henrici, A.C.,
Heckert, A.B., Zeigler, K.E. and Rinehart, L.F., 2005b, Early Permian
vertebrate biostratigraphy at Arroyo del Agua, Rio Arriba County, New
Mexico: New Mexico Museum of Natural History and Science, Bulletin
158
31, p. 163-169.
Lucas, S.G., Spielmann, J.A. and Krainer, K., 2010, Summary of geology of
Cañon del Cobre, Rio Arriba County, New Mexico: New Mexico Museum
of Natural History and Science, Bulletin 49, this volume.
Reisz, R.R., 1980, The Pelycosauria: A review of phlogenetic relationships;
in Panchen, A.L., ed., The terrestrial environment and the origin of land
vertebrates: Academic Press, London, p. 553-592.
Reisz, R.R., 1986, Pelycosauria: Encyclopedia of Paleoherpetology, Part
17A, 102 p.
Reisz, R.R., 2004, A reevaluation of the enigmatic Permian synapsid Watongia
and its stratigraphic significance: Canadian Journal of Earth Science, v.
41, p. 377-386.
Romer, A.S. and Price, L.I., 1940, Review of the Pelycosauria: Geological
Society of America, Special Paper 28, p. 1-538.
 Lucas et al., eds., 2010, Carb-Permian transition in Cañon del Cobre. New Mexico Museum of Natural History and Science Bulletin 49.

159
REDESCRIPTION OF THE CRANIALANATOMY OF SPHENACODON FEROX
MARSH (EUPELYCOSAURIA: SPHENACODONTIDAE) FROM THE LATE
PENNSYLVANIAN-EARLY PERMIAN OF NEW MEXICO

JUSTINA. SPIELMANN1, LARRYF. RINEHART1, SPENCERG. LUCAS1,

DAVID S BERMAN2,AMYC. HENRICI2 AND SUSANK. HARRIS1
1 New Mexico Museum of Natural History and Science, 1801 Mountain Rd. NW, Albuquerque, NM 87104-1375;

2 Section of Vertebrate Paleontology, Carnegie Museum of Natural History, 4400 Forbes Ave., Pittsburgh, PA 15213

Abstract—Sphenacodon is a pelycosaurian-grade synapsid, best known from the Lower Permian of Rio Arriba
County, northern New Mexico. Of the two species (S. ferox and S. ferocior), S. ferox is known from comparatively
little skull material, and the skulls historically assigned to the taxon are heavily reconstructed and composed offew
actual cranial elements. Here, we report on a new, nearly complete skull of S. ferox that demonstrates numerous
cranial differences between the two species of Sphenacodon. S. ferox has a temporal range from the Late Pennsyl
vanian (late Virigilian: Coyotean lvf) through the Early Permian (late Wolfcampian: Seymouran lvf), whereas S.
ferocior is restricted to the Coyotean lvf, but does span the Pennsylvanian-Permian boundary.

INTRODUCTION
Sphenacodon ferox is a pelycosaurian-grade synapsid
(Eupelycosauria: Sphenacodontidae: Sphenacodontinae) best known from
the Lower Permian of Rio Arriba County, northern New Mexico (Fig. 1).
S. ferox is known from numerous, incomplete skulls, most of which have
been extensively reconstructed. Here, we review all known S. ferox skulls
(Figs. 2-9, 13) and contrast their cranial morphology with that of S.
ferocior, whose skull morphology is well documented (Eberth, 1985),
and describe the most complete skull of S. ferox known, NMMNH P
55367 (Figs. 13-14), recently collected from the eastern wall of Cañon
del Cobre (El Cobre Canyon), a classic Permo-Carboniferous collecting
area (Lucas et al., this volume). This results in a revision of the diagnoses
of the two species of Sphenacodon.
Institutional abbreviations: AMNH, American Museum of
Natural History, New York; CM, Carnegie Museum of Natural History,
Pittsburgh; FMNH, Field Museum of Natural History, Chicago;
NMMNH, New Mexico Museum of Natural History and Science, Albu FIGURE 1. Pennsylvanian-Permian fossil localities in northern New Mexico,
querque; UCMP, University of California Museum of Paleontology, Rio Arriba County, that yield Sphenacodon fossils (Arroyo del Agua area
Berkeley; UMMP, University of Michigan Museum of Paleontology, and Cañon del Cobre).
Ann Arbor; YPM, Yale Peabody Museum, New Haven.
goid. Unfortunately, the reconstruction of this skull followed the com
mon practice in the early 20th century of trying to blend the recon
SKULLS OF SPHENACODON FEROX
structed elements with the actual skeletal materialso that they are indis
Several incomplete skulls of Sphenacodon ferox have been de tinguishable from each other. We have summarized what is clearly visible
scribed since the taxon was erected by Marsh (1878), and were the basis on the specimen, though the only way to determine the full extent of the
of several reconstructions. Here, we first review the most significant of
reconstruction would be to reprepare the specimen. UCMP 34226 can
these skulls listing the elements used in the reconstruction and discuss be referred to Sphenacodon ferox on the basis of its nearly linear ventral
the basis of their referral to S. ferox. This is followed by description of
maxillary margin, overall small size and reduced maxillary step (see be
the much more complete skulls NMMNH P-58231 and P-55367. low).
YPM 806 (Holotype) AMNH 4779
The holotype of Sphenacodon ferox is a left mandible and associ AMNH 4779 is an extensively reconstructed skull in the region
ated skull fragments (Fig.2). These additional cranial fragments (a right
anterior to the temporal fenestra (Fig. 4). The elements used in the
premaxilla, left and right maxilla and a right mandible fragment) clearly reconstruction include fragments of the left maxilla and jugal, and the
demonstrate features that identify the specimen as S. ferox, most nota right nasal, frontal, maxilla, lacrimal and jugal. Some of the inferred su
bly the linear ventral maxillary margin (see below). tures have been incorrectly traced on the reconstructed portions of the
UCMP 34226 skull (e.g., the frontal-nasal suture), and there is no indication of the
anterior portion of the maxilla-lacrimal suture. Based on examination of
UCMP 34226 is an extensively reconstructed skull (Fig. 3) with the curvature of the ventral maxillary margin, AMNH 4779 may be a
an associated mandible (not illustrated). The skull elements used in the composite specimen composed of bones of both Sphenacodon ferocior
reconstruction are the right premaxilla, maxilla, jugal and possible ptery and S. ferox. The left maxilla is linear (Fig. 4C) and thus conforms to S.
160

FIGURE 2. Sphenacodon ferox, YPM 806, holotype of S. ferox. A, right maxilla in lateral view. B, right mandible fragment in lateral view. C, left
premaxilla in lateral view. D, left maxilla in lateral view. E, left mandible in lateral view.

FIGURE 3. Sphenacodon ferox, UCMP 34226, reconstructed skull in lateral view.
ferox, whereas the right maxilla is slightly convex ventrally (Fig. 4A) and
thus closer to the definition of S. ferocior, though given its relatively
small size it likely pertains to S. ferox.
FMNH 35
Originally described by Williston (1918), this skull is the basis for
the most influential skull reconstruction of Sphenacodon ferox, as it was
considered the best known skull of the species at the time (Figs. 5B-C,
6). Unfortunately, much of the skull has been reconstructed on the basis
of fragments of the maxillae, nasals, prefrontals, frontals, postfrontals,
right lacrimal, jugals and right squamosal (Fig. 7). The nearly linear ven
tral maxillary margin clearly indicates assignment to Sphenacodon ferox.
FMNH 1218
FMNH 1218 consists of associated skull fragments, including
portions of the right nasal, lacrimal and maxilla, and a nearly complete
right mandible (Fig. 8). The maxilla preserves portions of the two canini
form teeth and eight postcaniniforms. Based on its small size and the
nearly linear ventral maxillary margin, FMNH 1218 is referable to
Sphenacodon ferox (see below).
NMMNH P-58231
NMMNH P-58231 consists of three right maxillae, one left man
dible and two skull fragments (Fig. 9), thus representing at least three
different skulls. The maxillae exhibit nearly linear ventral margins with a
greatly reduced maxillary step, so they are referable to Sphenacodon
ferox.
Summary
As demonstrated above, all of the skulls of Sphenacodon ferox,
other than NMMNH P-55367 (described below), are very incomplete,
generally represented by the maxillae and other fragmentary elements.
Hence, the longstanding diagnosis based solely on maxillae characters
and overall size. While the maxillae characters can still be used to diag
nose S. ferox, newly recognized skull characters based on NMMNH P
55367 and described below provide additional features that allow for a
161
more rigorous differentiation between the two species of Sphenacodon.
LOCALITY 5379:AGE, LITHOLOGYAND TAPHONOMY
NMMNH P-55367 was collected from NMMNH locality 5379,
a site originally found by David Eberth and described by Berman (1993),
who reviewed the vertebrate assemblages of Cañon del Cobre and deter
mined that the assemblage containing locality 5379 is Wolfcampian in
age. The NMMNH locality 5379 site is stratigraphically high in the
Lower Permian Arroyo del Agua Formation (Lucas and Krainer, 2005),
~50 m below the base of the overlying Upper Triassic Chinle Group.
Topographically, the quarry is located on the eastern wall of Cañon del
Cobre approximately 120 m (400 feet) above the canyon floor (Fig. 10).
The bonebed has yielded both cranial and postcranial elements of
Sphenacodon ferox, in addition to fossils of the diadectomorph Diadectes
and the temnospondyl Platyhystrix (Lucas et al., 2005). NMMNH P
55367 was located approximately 0.6 m above the main bonebed level (in
Figure 10F the main bonebed is at the feet of the worker, and the skull is
in the highwall at the level of the chisel).
The stratigraphic position of NMMNH locality 5379 places it
within the upper assemblage zone of tetrapod fossils found in Cañon del
Cobre (Lucas et al., 2005). Lucas et al. (2005) assigned this upper assem
blage zone to the Early Permian Seymouran land-vertebrate faunachron
of Lucas (2005, 2006), based on correlations with the Arroyo del Agua
Formation near Arroyo del Agua in Rio Arriba County, New Mexico
(Fig. 1), where Seymouria, the principal index taxon of the Seymouran, is
present. Lucas (2006) indicates that the Seymouran lvf is equivalent to
the late Wolfcampian marine stage.
Lithology and taphonomy: The lithology and taphonomy of the
NMMNH locality 5379 quarry were first reported by Rinehart et al.
(2007). The quarry facies comprise poorly sorted (vfL to cL), finely
laminated, dark reddish brown (10R 3/4), immature arkosic sandstone
with pale yellowish-green (10GY 7/2) banding at ~10 cm intervals (Fig.
10B-D, F). The yellowish green bands are similar in composition to the
reddish brown rock and apparently represent chemically reduced layers.
The quarry site lithology is typical of the Cutler Group throughout
northern New Mexico. In situ rhizoliths are common in the quarry, and
occasionally disrupt the fossil bones, including NMMNH P-55367 (see
162

FIGURE 4. Sphenacodon ferox, AMNH 4779, heavily reconstructed skull anterior to the temporal fenesta in A, right lateral, B, dorsal and C, left lateral
views. Note incorrect and incomplete tracing of sutures on the reconstructed portions of the skull.
 163

FIGURE 5. Reconstructions of the skulls of A, Sphenacodon. ferocior in lateral view, from Romer and Price (1940), and B-C, Sphenacodon ferox in B,
lateral and C, dorsal views, from Williston (1918).
164
FIGURE 6. Reconstructions of the skeleton of A-B, Sphenacodon ferox in lateral view, A, from Williston (1918) and B, from Romer and Price (1940),
and C, Sphenacodon ferocior, in lateral view from Romer and Price (1940).
below for details).
Field jackets were marked with a north-south arrow during collec
tion, and this index was used during preparation to ascertain the direc
tional orientation of the bones. A small sample of elongate (length greater
than or equal to four times width) bones was found to trend N40°E with
reasonably close alignment (± 35°) (Fig. 11). Such alignment probably
indicates gentle to moderate flow. Additionally, the fossil bones lie flat on
the bedding planes, which indicates a low energy depositional environ
ment (Voorhies, 1969).
Hydraulic sorting of the bones is evident. Using the Voorhies
group definitions of Behrensmeyer (1975) (Appendix 1), which are based
on mammal bones, we categorized and plotted all the prepared bones
from the quarry (Fig. 12). The three Voorhies groups are based on diffi
culty of transport by flowing water. Group I comprises bones that are
easily transported by floating or saltation, and Group II bones have
traction on the bottom and are more difficult to transport. They require
higher flow rates than Group I, and Group III bones lag and are very
difficult to transport. The proportions of the Voorhies groups relative to
the proportions of each category of bone in a complete skeleton thus
indicates the extent of hydraulic sorting (Fig. 12). In unsorted deposits,
the bones are found in the same proportions as in a complete skeleton. If
the deposit is impoverished in Group I elements, winnowing is indi
cated, and if both Group I and Group II elements are impoverished
relative to Group III, a lag deposit is indicated (Shipman, 1981; Lyman,
1994). Our ternary plot (Fig. 12) shows Group I to be severely depleted
(representing 20%) and Group II to be considerably impoverished (rep
resenting 33%), whereas Group III is greatly enriched (47%) over the
expected proportion. Thus, the ternary plot indicates a heavily win
nowed or lag deposit.
In summary, the microstratigraphy of the quarry, the alignment of
 165

FIGURE 7. Sphenacodon ferox, NMMNH C-4771, cast of FMNH 35, reconstructed skull in A, left lateral and B, right lateral views. Labeled elements
indicate actual skeletal material. Abbreviations: d, dentary; f, frontal; j, jugal; max, maxilla; na, nasal; pf, postfrontal; po, postorbital; prf, prefrontal;
sq, squamosal.
166

FIGURE 8. Sphenacodon ferox, FMNH 1218, portions of the right nasal, lacrimal, maxilla and nearly complete right mandible in A, medial and B, lateral
views.
 167

FIGURE 9. Sphenacodon ferox, NMMNH P-58231, A, associated maxillae and mandible fragments in medial and lateral views. B-C, isolated right maxilla
in B, lateral and C, medial views.
168

FIGURE 10. NMMNH locality 5379 where NMMNH P-55367 was collected. A-C, Photographs of the quarry at various points during the 2005
excavation. D, Small jackets collected from the quarry; NMMNH P-55367 is in the jacket at the bottom center of the photograph. E, NMMNH P-55367
as it was exposed in the field. F, Collecting of NMMNH P-55367; note that the primary quarry level is at the feet of the person in the photograph and
NMMNH P-55367 was collected from approximately 0.6 m above this horizon at the position of the chisel.

FIGURE 11. Rose plot of elongate bone orientations from L-5379 showing
NE-SW alignment.
FIGURE 12. A ternary plot of the relative abundances (%) of the Voorhies
groups at L-5379. The uppermost data point represents the expected
proportions if the bones were unsorted. The lower point represents the
measured proportions in the quarry. A heavily winnowed or lag deposit is
indicated.
the bones and the preponderance of Voorhies Group III elements indicate
that the bonebed was probably accumulated as a lag deposit on a flood
plain that was subject to periodic sheetfloods of low to medium velocity
and lacking turbulent flow.
NMMNH P-55367: IDENTIFICATION
Identifying sphenacodontines based solely on skull characteris
169
tics is difficult given how similar the two North American species of
Sphenacodon, S. ferox and S. ferocior, are to one another and to the
various species of Dimetrodon and Ctenospondylus. Sphenacodon fos
sils are found predominantly in New Mexico, whereas Dimetrodon fos
sils are found predominantly in Texas and rarely in New Mexico (e.g.,
Berman, 1977; Lucas et al., 2009).
We identify NMMNH P-55367 as Sphenacodon ferox based on
the following features: the specimen was found in a nearly monotaxic
assemblage that included postcrania identifiable as Sphenacodon (Lucas
et al., 2005); overall size of the skull, i.e., the skull metrics are virtually
identical to those of S. ferox and dissimilar from S. ferocior (discussed
below); lesser development of maxillary step; lesser curvature of the
ventral margin of the maxilla; and additional cranial features that are
slightly varied and more primitive than those found in S. ferocior. The
mediolaterally compressed, blade-like neural spines of the dorsal verte
brae differentiate Sphenacodon from Dimetrodon, Ctenorhachis,
Ctenospondylus and all other sphenacodontids, and only this morphol
ogy of dorsal vertebra is present in the locality 5379 bonebed (Lucas et
al., 2005).
The reduced maxillary step and less convex ventral maxillary mar
gin of NMMNH P-55367 identify it as Sphenacodon ferox (Figs. 13
14). These features contrast with a large maxillary notch and convex
ventral maxillary margin in S. ferocior and have been used by Romer and
Price (1940) and Reisz (1986, p. 77-78) to distinguish the two species
from one another.
Based on the closed sutures of the skull we interpret NMMNH
P-55367 as representing an adult individual. Its overall size indicates that
the skull conforms more closely to the size range of Sphenacodon ferox
than to that of S. ferocior, which can be up to 40% larger than S. ferox
(Figs. 13-14). The skull metrics of NMMNH P-55367 compared to
average skull sizes of Sphenacodon ferox and S. ferocior, as reported by
Romer and Price (1940), are summarized in Table 1.
SYSTEMATIC PALEONTOLOGY
Suborder EUPELYCOSAURIA Kemp, 1982
Family SPHENACODONTIDAE Marsh, 1878
Subfamily SPHENACODONTINAE Romer and Price, 1940
Genus Sphenacodon Marsh, 1878
Type species: Sphenacodon ferox Marsh, 1878.
Included species: The type species and Sphenacodon ferocior
Romer and Price, 1940.
Diagnosis: Differs from other sphenacodontines by possessing
tall, blade-like neural spines that are shorter than the neural spines of
Dimetrodon, Ctenospondylus and Ctenorhachis.
Distribution: Upper Pennsylvanian-Lower Permian (Coyotean
and Seymouran lvfs) of New Mexico and Utah.
Discussion: “Oxydon” brittanicus von Huene (1908) is a
sphenacodontine known only from the holotype, which consists of a
part and counterpart of a maxilla, from the Kenilworth Sandstone of
England, originally named by von Huene (1908). This taxon was as
signed by Paton (1974) to Sphenacodon based on its similar size, pro
portions and relative size of the caniniform tooth, who retained the
specific name on the basis of its geographic isolation from all other
occurrences of Sphenacodon. Reisz (1986, p. 78) followed this assign
ment, but noted that it may prove to be invalid. Eberth (1985) pointed
out that based on the criteria provided by Paton (1974), “Oxydon”
brittanicus could also be assigned to Dimetrodon booneorum and is best
designated as Sphenacodontine incertae sedis until additional material
suggests otherwise. We follow Eberth (1985) and retain only two species
of Sphenacodon, S. ferox and S. ferocior.
An argument could be made for synonymizing the two species of
Sphenacodon, given their morphologically identical postcrania and the
few, subtle cranial differences, which some workers (e.g., Eberth, 1985)
have interpreted as merely reflecting size and/or ontogeny. However,
170

FIGURE 13. Sphenacodon ferox, NMMNH P-55367, nearly complete skull in A, left lateral and B, right ventrolateral views.
 171

FIGURE 14. Sphenacodon ferox, NMMNH P-55367, line drawing of nearly complete skull in A, left lateral and B, right ventrolateral views. Skeletal
elements labeled. Stippled shading is exfoliated bone, black shading is openings in the skull, light gray shading is reconstruction and dark gray shading is
adhered matrix. Abbreviations: bs, basisphenoid; ec, ectopterygoid; in, internal narial opening; f, frontal; j, jugal; la, lacrimal; ltf, lateral temporal
fenestra; max, maxilla; na, nasal; p, parietal; pal, palatine; pf, postfrontal; pmx, premaxilla; po, postorbital; pp, postparietal; prf, prefrontal; ps,
parasphenoid; pt, pterygoid; q, quadrate; qj, quadratojugal; s, stapes; sm, septomaxilla; sq, squamosal; srf, sclerotic ring fragments; so, supraoccipital; st?,
supratemporal; t?, tabular?; v, vomer.
172
TABLE 1. Comparative skull metrics of S. ferox, S. ferocior (Romer and
Price, 1940) and P-55367. All measurements in mm.
NMMNH P-55367, being the most complete S. ferox skull known,
contains new information and its description below identifies numerous
characters that distinguish S. ferox from S. ferocior. The NMMNH speci
men is especially important because it is an adult skull, and thus elimi
nates the problem of comparing different ontogenetic stages of growth
by allowing for comparisons between fully adult individuals of the two
species.
Sphenacodon ferox Marsh, 1878
1878 Sphenacodon ferox Marsh, p. 410
1907 Elcabrosaurus baldwini Case, p. 28
1913 Scoliomus puercensis Williston and Case, p. 60, fig. 37
1940 Sphenacodon ferox: Romer and Price, p. 326-327
1986 Sphenacodon ferox: Reisz, p. 77, figs. 4, 18c, 25g, 26e, 27g
Holotype: YPM 806, left mandible (Fig. 2E). Clearly associated
elements of YPM 806 include the left maxilla, a left mandibular fragment,
right premaxilla and maxilla (Fig. 2A-D). Given their association these
elements should be considered elements of the holotype, due to all the
skull elements being similarly proportioned and likely derived from a
single skull. See above under Skulls of Sphenacodon ferox for addi
tional comments on this specimen.
Type locality and horizon: Baldwin bonebed (Fig. 15), near
Arroyo del Agua, Rio Arriba County, New Mexico, from the El Cobre
Canyon Formation, Cutler Group (Berman, 1993).
Principal referred material: Skulls of Sphenacodon ferox in
clude:UCMP 34226, reconstructed skull based on partial right premax
illa, and left maxilla and left jugal (Fig. 3); NMMNH P-55367, nearly
complete skull, crushed mediolaterally; NMMNH P-58231, maxillae,
mandibles and skull fragments (Fig. 9); AMNH 4779, mostly recon
structed skull based on right nasal, frontal, maxilla, lacrimal, jugal, pos
torbital, postfrontal and left maxilla and jugal, all elements in various
states of completeness (Fig. 4); FMNH 35, mostly reconstructed skull
using various skull fragments incorporated in the reconstruction (Fig. 7);
FMNH 1218, incomplete skull and mandible (Fig. 8).
Revised diagnosis: Aspecies of Sphenacodon distinguished from
S. ferocior by the following characters: overall smaller size, as much as
40% smaller than the largest individuals of S. ferocior; four premaxillary
teeth; three precaniniform maxillary teeth in adult specimens; nearly
linear ventral maxillary margin; greatly reduced maxillary step; dorsal
lamina of maxilla less developed, resulting in a wedge-shaped lacrimal
and a shorter sutural contact between the maxilla and nasal; orbital pro
cess of the frontal not as greatly thickened; anteroventral margin of the
temporal fenestra forms a right angle; squamosal-jugal suture does not
extend posteriorly beyond the posterior extent of the temporal fenestra;
parietals with circular, concave embayments on anterior margin; and
denticle field on pterygoid flange does not extend to its lateral margin.
Discussion: A summary of the diagnostic features of S. ferox is
provided below under the heading Diagnostic cranial features of
Sphenacodon ferox.
Description: Description of NMMNH P-55367 is provided be
low; all other type and referred material is discussed above under the
heading Skulls of Sphenacodon ferox.
Sphenacodon ferocior Romer and Price, 1940
1940 Sphenacodon ferocior: Romer and Price, p. 327-329.
1985 Sphenacodon ferocior: Eberth, p. 1-40, figs. 1-39.
1986 Sphenacodon ferocior: Reisz, p. 78, figs. 21i, 23j, 30j, 38.
Holotype: MCZ 1489, skull and anterior vertebrae.
Type locality and horizon: San Diego Canyon, Sandoval County,
New Mexico, precise locality unknown. Abo Formation, Lower Per
mian.
Principal referred specimens: YPM 818, partial skull and nearly
complete presacral vertebral column, pectoral and pelvic girdles, hu
merus and femur from “Rito Puerco,” probably a misreading of Rio
Puerco, by David Baldwin, late 19th century, in the Arroyo del Agua area,
Rio Arriba County, NM. Various UCMP and MCZ specimens dis
cussed and illustrated by Eberth (1985) including: UCMP 34184, left
postfrontal; UCMP 34192, partial lacrimal; UCMP 34194, incomplete
parietal; UCMP34196, partial skull; UCMP34201, left splenial; UCMP
34202, left articular; UCMP 34218, prearticular, occiput and braincase;
UCMP 34219, left quadrate and quadratojugal; UCMP 83047, right
premaxilla; UCMP 83185, right prefrontal; UCMP 81393, left frontal;
UCMP 83060, right maxilla; UCMP 83203, right nasal; UCMP 83208,
partial skull; UCMP 83251, supraoccipital; UCMP 83285, left palatine;
UCMP 83378, right surangular; UCMP 83397, right angular; UCMP
83459, right coronoid; UCMP 83478, right dentary; UCMP 123288,
partial stapes; UCMP 125583, partial stapes.
Revised diagnosis: Aspecies of Sphenacodon distinguished from
S. ferox by the following features: overall large size, up to 40% larger
than the largest individuals of S. ferox; possessing three premaxillary
teeth; one or two precaniniform maxillary teeth in adult specimens; con
vex ventral maxillary margin; prominent maxillary step; dorsal maxillary
process that results in a more rounded anterior end of the lacrimal and a
longer maxilla-nasal suture; orbital process of the frontal thickened; tem
poral fenestra with a rounded anteroventral margin; a squamosal-jugal
suture that extends posteriorly beyond the posterior temporal fenestra;
anterior parietal has shallow, elliptical sutural surface, no embayments;
and tooth field of the pterygoid extends to its lateral margin. See Figure 4
for graphic illustration of these characteristics.
NMMNH P-55367: DESCRIPTION AND COMPARISON
NMMNH P-55367 is a nearly complete skull of Sphenacodon
ferox that has been obliquely crushed mediolaterally to the extent that
the left lateral side and the skull table are visible in the same plane (Figs.
13-14). The following elements are not preserved in NMMNHP-55367:
tabulars, epipterygoid, basioccipital, exoccipital, supraoccipital,
opisthotic, sphenethmoid. Portions of left preorbital region of the skull
and cheek region are patched with epoxy putty (for strength and stabili
zation). In addition, left caniniform 2 has been reconstructed based on an
impression in the surrounding matrix. Some of the cranial elements ante
rior to the antorbital reconstruction are partially exfoliated, further com
plicating suture identification.
Oblique crushing has resulted in most of the right side of the skull,
the palate and braincase being visible in a single plane. The right lacrimal
is missing part of its central portion. The posteroventral margin of the
maxilla is missing. The posterior part of the skull and braincase are
missing, including most of the squamosal and basioccipital.
Here, we describe the skull of Sphenacodon ferox based on
NMMNH P-55367, the most complete skull known for this taxon, and
compare it to descriptions of S. ferocior (Eberth, 1985) and of other
sphenacodontines.
Premaxilla
Both premaxillae are present, complete and articulated, with the
right slightly offset dorsally from the left, due to oblique crushing (Figs.
13-14). The dorsal process is posterodorsally curved and ends in a point

FIGURE 15. Photograph of the Baldwin bonebed taken by E.C. Case. White arrow highlights the level of the quarry.
medial to the anterior end of the nasal. Due to the disarticulation of the
nasals (see below) the exact suture between these elements and the pre
maxillae is not clear. Four tooth sockets are present (Fig. 16A-B). In
ventral view the sinusoidal suture with the anterior margin of the maxilla
is angled posteromedially, so that in lateral view the fourth (posteriormost)
premaxillary tooth is positioned directly below the premaxilla-maxilla
suture. Thus, the premaxillary process of the maxilla overlaps laterally
the maxillary process of the premaxilla. The anterior two premaxillary
teeth are conical and not recurved, whereas the posterior two premaxil
lary teeth are much smaller, approximately half the size of the anterior
teeth, and are recurved. Due to crushing, the suture between the premax
illae and vomers is not completely clear. Crushing also prevents a clear
view of the septomaxilla-maxilla suture and the posterior premaxillary
and prepalatal foramina, which are well described in Sphenacodon ferocior
by Eberth (1985, p. 9, fig. 4).
Septomaxilla
The left septomaxilla is complete, whereas the right is incomplete
and disarticulated (Figs. 13A, 14A). The description is based on the left
173
septomaxilla, which is only visible in lateral view. The septomaxilla
appears as a crescent-shaped element in the posterior half of the external
naris. It extends for the full height of the external naris and contacts
anteriorly the premaxilla. The medial shelf appears to be high on the
septomaxilla and more similar to the left septomaxilla illustrated by
Eberth (1985, fig. 5C) in MCZ 1489, than the right illustrated by Eberth
(1985, fig. 5A) in UCMP34196. The anterior foramen is located directly
under the medial shelf and is directed ventromedially. A small, posteri
orly-directed, triangular notch is present at the posteroventral margin of
the septomaxilla that leads to the septomaxillary foramen. The
septomaxilla of NMMNHP-55367 differs from that of S. ferocior (Eberth,
1985, fig. 5) in the following ways: has an overall crescent-shaped out
line that is not taller than wide (Eberth, 1985, fig. 5A) or polygonal
(Eberth, 1985, fig. 5C) and is more recessed into the naris with a promi
nent posteroventral rim.
Nasal
Both nasals are present (Figs. 13A, 14A). The nasal contacts the
lacrimal ventrolaterally with a finely digitate suture anteriorly, whereas
174

FIGURE 16. Closeup of Sphenacodon ferox skull, NMMNH P-55367, highlighting diagnostic features of S. ferox. A-B, premaxillae and anterior maxillae
in ventrolateral views; four premaxillary teeth indicated by arrows, premaxilla-maxilla suture and fourth premaxillary tooth socket outlined. C-D,
maxillae in ventrolateral view; three maxillary precaniniform teeth indicated by arrows, premaxilla-maxilla suture and first caniniform tooth socket
outlined. Abbreviations: mx pc1-pc3, maxillary precaniniform teeth one through three; mx 1-2, maxillary caniniform teeth one and two; pmx t1-t4,
premaxillary teeth one through four.

posteriorly the suture is obscured by exfoliated bone. The turbinal ridge
is clearly visible on the ventral surface of the nasal, also noted in
Sphenacodon ferocior by Eberth (1985, p. 12-13, fig. 6A).
Frontal
Both frontals are present (Figs. 13A, 14A). The anterior process
of the frontal tapers anteriorly to a triangular point that fits in the poste
rior margin of the nasal and is only clearly visible on the right side, due to
reconstruction. The frontal, like the nasal, possesses a turbinal ridge. The
orbital process of the frontal does not appear to be as thickened in
Sphenacodon ferox as in S. ferocior (Eberth, 1985, fig. 7A), yet the
orbital processes of both the prefrontal and postfrontal are thickened in
S. ferox.
Parietal
Both parietals are present and essentially complete, although the
right parietal is missing the supratemporal sutural surface (Figs. 13A,
14A). The midline suture is indistinct and the pineal foramen was dam
aged by a fossilized root (rhizolith). On the anterior margin of the pari
etal the embayment for the postfrontal is deep and semicircular (Fig.
17C-D), whereas in Sphenacodon ferocior (Eberth, 1985, fig. 8A) it is
represented by a shallow, concave sutural scar on the dorsal surface of
the anterior margin (Fig. 17E-F). Also, no features mark the tabular
sutural surface, which is distinguished in S. ferocior by a shallow ellipti
cal embayment, similar to the embayment for the postfrontal.
Postparietal
An unpaired postparietal is present, an element not known in
Sphenacodon ferocior (Eberth, 1985, p. 15) (Figs. 13A, 14A). As hy
pothesized by Eberth (1985), the postparietal has a low, squat, triangu
lar outline, with the apex directed ventrally. The postparietal of NMMNH
P-55367 is bowed anteriorly and is bordered dorsally and laterally by the
parietal, which overlap it marginally. A faint, radial, surface texture is
present on the margins of the postparietal. Although obscured by crush
ing, the postparietal was likely bordered ventrally by the supraoccipital
and ventrolaterally by the tabulars, as in other sphenacodontines (e.g.,
Dimetrodon limbatus: Reisz, 1986, fig. 15).
Prefrontal
The left prefrontal is nearly complete (except for some minor
areas of reconstruction) and in articulation with neighboring elements,
whereas the right is incomplete, missing the margin bordering the lacrimal
(Figs. 13-14). The thickening of the orbital margin is prominent, espe
cially compared to the unthickened orbital process of the frontal (see
above). The depression immediately anterior to the orbital deepens to
ward the well-developed orbital ridge, more so than illustrated in
Sphenacodon ferocior (Eberth, 1985, fig. 9A). Also, the thickening of the
prefrontal continues anteriorly and dorsal to the orbital process, a fea
ture not seen in S. ferocior. The area of the suture with the lacrimal has
been lost, but the prefrontal-frontal suture is preserved.
Postfrontal
Both postfrontals are present, complete and in place (Figs. 13
14). The postfrontal forms the posterodorsal margin of the orbit, where,
like the prefrontal, it is greatly thickened. Overall, the element is qua
drangular in dorsal view. There are no differences in this element between
the two species of Sphenacodon.
Postorbital
Both postorbitals are present, though the left is more complete
and well articulated (Figs. 13-14). The postorbital forms the posterior
margin of the orbit, but it is not thickened as are the prefrontal and
175
postfrontal. Also, it forms the anterodorsal margin of the lateral temporal
fenestra. The ventral process of the postorbital overlaps the postorbital
process of the jugal and contacts the parietal with a distinct suture. The
thin posterior process borders the entire lateral length of the wing-like
posterior extension of the parietals. There are no differences in this
element between the two species of Sphenacodon.
Maxilla
Both maxillae are present, though the left is reconstructed par
tially and the right is incomplete. There are three precaniniform, two
caniniform and 12 postcaniniform tooth positions, for a total of 17 (Figs.
13-14, 16C-D). The left maxilla preserves teeth in precaniniforms 2 and
3, caniniform 2 and postcaniniforms 1, 2, 9 and 10. The right maxilla
preserves teeth in precaniniform 1-3, caniniform 2 and postcaniniform 1,
2 and 10-12. Three precaniniform teeth in an adult Sphenacodon ferox
distinguishes it from S. ferocior, which generally has one, and rarely two,
precaniniform teeth in adults, though juveniles can have up to four.
Overall, the ventral margin of the maxilla is only slightly convex, com
pared to being strongly convex in S. ferocior; this feature has long been
used to distinguish between S. ferox and S. ferocior (Romer and Price,
1940; Reisz, 1986). The dorsal maxillary process is present on the ante
rior maxilla, resulting in maxillae and lacrimals that are shaped differently
than in S. ferocior.
Lacrimal
As noted above, both lacrimals are present but incomplete (Figs.
13-14), with the right partially disarticulated and missing its central
portion, whereas the left is partially exfoliated, further confirming Eberth’s
(1985, p. 18) observation that they are thin elements. The ventral margin
of the anterior lacrimal is bounded by a strongly digitate suture with the
maxilla and dorsally by a finer digitate suture with the nasal. The anterior
portion of the lacrimal is wedge-shaped, rather than rounded, as in S.
ferocior. This difference may be the result of a less well-developed dorsal
maxillary lamina that borders the lacrimal anteriorly in S. ferocior.
Jugal
Both jugals are present, with the left being complete, whereas the
right is missing its anterior margin (Figs. 13-14). Eberth (1985, p. 18, fig.
15) noted two distinct morphologies of the posterior process of the jugal
in Sphenacodon ferocior: one with only a slight downward angulation of
the posterior process and the other with a strong downward angulation.
The posterior process in NMMNHP-55367 is significantly downturned
and is much shorter than in S. ferocior. However, in NMMNH P-55367
the angle between the body of the jugal and the posterior process ap
pears to be more acute.
Squamosal
Both squamosals are very incomplete (Figs. 13-14). The left squa
mosal is represented by a small triangular fragment sutured to the poste
rior process of the jugal, which together form the anterior portion of the
zygomatic arch. A small portion of the right squamosal is preserved at
the posterior margin of the temporal fenestra. Based on these incomplete
elements, there is no obvious differences between the two species of
Sphenacodon.
Supratemporal
An incomplete element is identified tentatively as the left su
pratemporal (Figs. 13A, 14A). It is articulated with the lateral process of
the parietal and is curved posteroventrally. In sphenacodontids the su
pratemporal is a splint-like bone whose anterior portion is sutured to a
shallow groove in the posterolateral wing of the postparietal, whereas
the medial margin of the posterior portion contacts the lateral margin of
the tabular.
176

FIGURE 17. Closeup of Sphenacodon ferox skull, NMMNH P-55367, highlighting diagnostic features of S. ferox. A-B, dorsal orbital margin in lateral view;
note the unthickened orbital process of the frontal, sutures between the prefrontal-frontal, frontal-postfrontal and postfrontal-postorbital are outlined.
C-D, parietals in dorsal view; parietals are outlined and the deep, circular embayments are indicated by arrows. E-F, comparison of the parietals of E,
Sphenacodon ferox (based on NMMNH P-55367) and F, S. ferocior (based on UCMP 34194 and MCZ 1489, from Eberth, 1985, fig. 8A). Abbreviations:
emb, parietal embayments; op f, orbital process of the frontal; pf, postfrontal; po, postorbital; prf, prefrontal.

Vomer
The paired vomers are present and visible in ventral view. They
form a prominent ridge at the midline, which borders the medial margin of
the internal narial opening (Figs. 13B, 14B). Posteriorly, the vomers may
be unique in wedging deeply between the anterior ends of the palatines.
The posterior suture with the palatine is V-shaped with the paired vomers
wedging between the anterior ends of the palatine, whereas the anterior
suture with the premaxilla is obscured. Posteriorly, the vomers expand
slightly before contacting the anteromedial corners of the palatine, then
narrow abruptly to short processes that overlap the distal ends of the
pterygoids.
Palatine
The palatine is represented by only the incompletely preserved
left. It is subtriangular, and the tooth field abuts the vomer-palatine
suture (Figs. 13B, 14B). The palatine contribution to the posterior mar
gin of the internal naris is not visible. A narrow rectangular denticle field
covers the medial half of the palatine and is continuous with that of the
pterygoid.
Ectopterygoid
Both ectopterygoids are present and have a narrowly
subrectangular outline with a slightly convex ventral surface. The
retroangular posterior portion of the ectopterygoid is incised into the
anterolateral area of the transverse flange of the pterygoid. It appears to
be greatly reduced compared to the ectopterygoid in other sphenacodontids
(Figs. 13B, 14B). Unfortunately, examples of this element are unknown
for S. ferocior (Eberth, 1985, p. 22).
Pterygoid
Both pterygoids are presumably complete, but only their poste
rior portions are visible in ventral view (Figs. 13B, 14B, 18C-D). The left
and right transverse flanges possess six (two are missing) and seven
teeth, respectively, which are aligned in a single row along their posterior
margin. The anterior margin of the transverse flange has a triangular
concavity, similar to the condition in S. ferocior (Eberth, 1985, fig. 21B).
Anterior to the transverse row of teeth, the triangular dental field nar
rows anteriorly to an apex and closely borders the medial margin of the
pterygoid and does not reach its lateral margin (Fig. 18A-B). The den
ticles increase in size toward the center of the field. Eberth (1985, fig.
21B) illustrated the tooth field of S. ferocior as more polygonal, extend
ing to the lateral margins of the pterygoid, with no concentration of larger
denticles.
Quadrate
Both quadrates are complete, but visible only in ventral view,
although the right has been damaged by a rhizolith that extended through
the pineal foramen and formed a circular hole in its posterolateral margin
(Figs. 13B, 14B, 18C-D). As noted by Eberth (1985, p. 24), the dorsal
process of the quadrate is essentially diamond-shaped and exceedingly
thin.
Parabasisphenoid
The parabasisphenoid complex of the braincase is preserved in
ventral view (Figs. 13B, 14B, 18C-D). In overall morphology this ele
ment appears similar to the parabasisphenoid of Sphenacodon ferocior
(Eberth, 1985, p. 26-27, fig. 26) except that the fossa between the basal
tubers of NMMNH P-55367 is more pronounced and lacks the small, u
shaped depression just posterior to the cartoid foramina found in S.
ferocior.
Stapes
The right stapes is represented in NMMNH P-55367 only by the
177
footplate, which is articulated with the posterior braincase (Figs. 13B,
14B, 18C-D).
Sclerotic Ring
A sclerotic ring is present in the left orbit of P-55367 (Fig. 19A).
The ring is apparently folded and flattened (forming a short, flattened,
hollow cylinder exposed in lateral view). Fourteen ossicles are visible,
represented as either complete or partial elements, or as impressions in
the matrix. Space exists for four or five additional ossicles, which may be
present, but obscured by other elements of the skull. Walls (1942) desig
nated sclerotic ossicles as plus (+) or minus (-) for those that overlap, or
are overlapped by, both of their neighboring ossicles, respectively and (i)
for those that show an imbricated pattern in which they both overlap and
are overlapped by each of their respective neighboring ossicles. The
sclerotic ossicles in NMMNH P-55367 are imbricated (i) with the ex
ception of a single anteriormost (+) ossicle in the nasal position of the
ring (Fig. 19B). The outer, orbital edge of the ring is paper-thin and
translucent, whereas the inner, corneal boundary is much thicker and
opaque.
The visible portion of the ring is flattened into a nearly straight
structure that is 47 mm long. If more ossicles are present in the covered
portion of the ring, it could have a circumference of up to 70 mm. The
radial distance from the corneal aperture to the orbital edge of the ring is
11 mm. Individual imbricated ossicles are pear-shaped. The wider, cor
neal portion forms the corneal aperture and overlaps half of each of the
adjacent ossicles, whereas the narrower, orbital portion abuts its adja
cent ossicles (Fig. 19B). In the best-exposed ossicle, the orbital margin is
2.9 mm wide and the corneal margin is 6.5 mm wide.
DIAGNOSTIC CRANIAL
FEATURES OF SPHENACODON FEROX
Figure 20 provides graphic illustration of the diagnostic character
istics of the two Sphenacodon species. Here, we briefly discuss the
diagnostic cranial features that distinguish S. ferox from S. ferocior and
other sphenacodontines.
1. The overall smaller size of Sphenacodon ferox: The skull of S.
ferox is as much as 40% smaller than that of S. ferocior (Figs. 20-21).
This is a well-established characteristic that has been used to distinguish
the two species since Romer and Price (1940) originally described and
named S. ferocior.
One possibility to explain the size difference is ontogeny, with
the smaller S. ferox being the juvenile of the larger S. ferocior. In order to
test this idea we used a statistical approach to determine if there are two
significantly different size classes representing the two species of
Sphenacodon, S. ferox and S. ferocior. Scapulocoracoid data from Romer
and Price (1940) were used because they provide the largest sample size
available in the literature (Table 2). Our first approach was to apply
ANOVA testing using PAST software (Hammer et al., 2001) to deter
mine if two size classes are present, but the tests failed to run because of
inadequate sample size in the S. ferocior data. Likewise, the sample size
was inadequate for assessment of shape through PC analysis. Our next
approach was to use probability plotting. In a probability plot of scapu
lae lengths (Fig. 22) the five data points representing S. ferox appear to
be distributed normally as indicated by the straight line fit on the normal
probability scale and show a mean value and standard deviation (sigma)
of 134 mm and 9.3 mm, respectively. The calculated coefficient of varia
tion (CV = (sigma/mean value) *100) is a very reasonable 6.9%. CVs in
this range (between 4% and 10%, and most often 5% or 6%) usually
indicate that the data are relatively “pure,” that is, they include only
animals of similar age class and taxonomic assignment (Simpson et al.,
1960).
Admittedly, the number of data points, which is well below an
optimal sample size, does not inspire confidence in the result. However,
low scatter about the regression line (R2 = 0.99) indicates that the stan
178

FIGURE 18. Closeup of Sphenacodon ferox skull, NMMNH P-55367, highlighting diagnostic features of S. ferox. A-B, pterygoid in ventral view;
pterygoid-maxilla suture outlined and maximum extent of pterygoid tooth field outlined by dotted line. C-D, posterior skull/braincase in ventral view.
Abbreviations: bs, basisphenoid; ic, internal carotid foramen; mx, maxilla; ps, parasphenoid; pt, pterygoid; q, quadrate; Q. pr, quadrate process of the
pterygoid; qj, quadratojugal; s, stapes.

dard error of the estimate is low, thus increasing confidence (Glantz,
2005). The value of the CV, which implies that the data derive from a
single taxon and age class, also raises confidence in the result.
On the normal probability scale (Fig. 22) the +1, +2, and +3 sigma
probability levels occur at 84%, 97.5%, and 99.85%, respectively. If the
two Sphenacodon ferocior data points are translated horizontally to the
S. ferox distribution (the curve fit line), they intersect the distribution at
approximately the +4 sigma level (off the plot to the right). Thus, there
is a less than one in 10,000 probability (P<0.0001) that these data points
actually belong to the same distribution. Forcing the regression line to the
extreme limits of the error bars does not change this result significantly.
Haldane (1948) asserted that a four sigma difference in the metric charac
ters of two populations of fossils alone was sufficient to establish them
as “almost certainly” different species. The likelihood that both S. ferocior
data points represent extreme outliers of the S. ferox distribution thus is
vanishingly small. We conclude that within the limits of a very small
sample size, we have shown with reasonable confidence that S. ferox and
S. ferocior embody two distinct size classes that represent two species
and are not part of a single ontogenetic trajectory.
2. Possession of four premaxillary teeth: Eberth (1985) noted that
only three premaxillary teeth occur in Sphenacodon ferocior, based on
14 specimens, thus clearly distinguishing it from S. ferox, which has four
premaxillary teeth (Fig. 16A-B). Although Eberth (1985, p. 34), follow
ing Langston (1952), noted that there are “three premaxillary teeth in all
known specimens of S. ferox,” the only skull with a complete premaxil
lary dentition is NMMNH P-55367; all other skulls of S. ferox are
extremely incomplete and/or heavily reconstructed (see above for our
summary of known S. ferox skulls). In general, the various species of
 179

FIGURE 19. Sphenacodon ferox, NMMNH P-55367. A, Closeup of sclerotic ring in lateral view. The leftmost ossicle in the photo is the single visible (+)
ossicle. B, Schematic sketch of the imbricated ossicles that comprise the bulk of the sclerotic ring.
180

FIGURE 20. Comparison of Sphenacodon ferox and S. ferocior skulls. Skulls reproduced at the same size for comparison, not to scale, except in inset (1),
reproduced at the same size for comparison. Diagnostic features are numbered: (1) overall size; (2) number of premaxillary teeth; (3) number of
precaniniform maxillary teeth in adult specimens; (4) ventral maxillary margin shape; (5) development of maxillary step; (6) possession/absence of dorsal
maxillary process and shape of anterior lacrimal; (7) anteroventral temporal fenestra shape; (8) position of the squamosal-jugal suture; and (9) extent of
the tooth field on the pterygoid. S. ferocior redrawn from Eberth (1985, fig. 3a) and S. ferox based on NMMNH P-55367.

Dimetrodon possess three premaxillary teeth, with D. grandis having
two and D. natalis having five (Romer and Price, 1940, p. 332). Thus,
neither S. ferocior nor any species of Dimetrodon possess four premax
illary teeth.
3. Three precaniniform maxillary teeth in adults: Injuvenile speci
mens of Sphenacodon ferocior, the precaniniform maxillary tooth count
can range from 1 to 4 teeth. However, adult specimens of S. ferocior
possess only one or two (Eberth, 1985, table 1). This is clearly distinct
from the three precaniniform maxillary teeth of adult S. ferox (Fig. 16C
D).
4. Linear ventral maxillary margin: Eberth (1985, p. 33) noted that
the degree of curvature of the ventral margin of the maxilla in both
Sphenacodon ferox and S. ferocior increases during ontogeny. He based
this on seven specimens of S. ferox and 13 specimens of S. ferocior,
though no specimen numbers for the illustrated material are provided,
making identification of the specimens used and thus reproducibility of
this observation exceedingly difficult. It is also unclear why Eberth (1985)
identified maxillae with curved ventral margins as S. ferox, with no addi
tional characters to support his identification. In addition, the maxillae
illustrated by Eberth (1985, p. 33) are isolated, so their taxonomic iden
tification is questionable. NMMNH P-55367, unquestionably an adult,
possesses a linear ventral maxillary margin that can be distinguished
easily from the convex maxillary margin of S. ferocior. Thus, a linear
ventral maxillary margin remains a useful feature in distinguishing adult
skulls of S. ferox from those of S. ferocior.
5. Greatly reduced maxillary step: Eberth (1985, p. 33) noted that
the maxillary step is well developed in all specimens of Sphenacodon
ferocior except the smallest (based on 7 specimens of S. ferox and 13
specimens of S. ferocior). When comparing adult skulls of Sphenacodon,
S. ferocior always has a larger, more developed maxillary notch than S.
ferox. Thus, this feature, originally used by Romer and Price (1940),
remains useful to distinguish the two species of Sphenacodon.
6. Dorsal lamina of maxilla less developed, resulting in a wedge
shaped lacrimal and a shorter sutural contact between the maxilla and
nasal: Comparing the skulls of the two species of Sphenacodon (Fig. 20),
the lacrimal of S. ferocior is bordered anteriorly by the maxillary dorsal
process, a small triangular extension of the maxilla, just behind the exter
nal narial opening. This process is not present in S. ferox, resulting in an
 181

FIGURE 21. Comparing the size of Sphenacodon ferox and S. ferocior skulls. All skulls are to the same scale. Skulls aligned by the anterior margin of their
orbits (gray bar). From top to bottom: UCMP 34226 (S. ferox); NMMNH P-55367 (S. ferox) (photo reversed horizontally for comparison); AMNH 4779
(S. ferox); and UCMP 34196 (S. ferocior). Black bars approximate length of the S. ferocior skull for comparison. Note that NMMNH P-55367 is similar
in overall size to S. ferox and considerably smaller than S. ferocior.
182
TABLE 2. Sphenacodon scapulocoracoid data from Romer and Price (1940).
Scapulocoracoid measurements are used because they produce the largest
statistical sample. Length refers to distance from “top of anterior end of
glenoid to top of scapula.” WM, Walker Museum, now part of the Field
Museum of Natural History.
FIGURE 22. Probability plot of scapulae lengths in Sphenacodon ferox and
S. ferocior. Error bars represent the standard error. Both S. ferocior data
points occur at the ~+4 sigma probability level on the S. ferox line indicating
extremely low probability that they belong to the same distribution
(P<0.0001).
anterior lacrimal that is more wedge-shaped and extends farther anteri
orly than in S. ferocior. This can be seen most clearly if one compares the
respective lengths of the nasal-maxilla suture in the two species.
Characters 4 through 6 all relate to the maxilla, and thus are likely
part of a single larger functional complex. In a phylogenetic analysis all
three of these characters would have identical distributions and our dis
cussion of these characteristics is merely a means of detailing the features
of the maxilla.
FIGURE 23. Temporal distribution of Sphenacodon.
7. Orbital process of the frontal not thickened: Sphenacodon ferox
is the only sphenacodontid that does not possess a thickened orbital
process of the frontal (Fig. 17A-B), including S. ferocior (Eberth, 1985,
p. 13-14) and Dimetrodon. This is especially interesting given that both
the prefrontal and postfrontals that bracket the orbital process are thick
ened in S. ferox.
8. Temporal fenestra with a right angle at the anteroventral corner:
In Sphenacodon ferocior, the anteroventral margin of the temporal fenes
tra is rounded, giving the fenestra a subelliptical shape. In S. ferox, the
anteroventral margin of the temporal fenestra forms a right angle (Fig.
20), resulting in a reversed D-shaped outline.
9. The squamosal-jugal suture does not extend posteriorly be
yond the temporal fenestra: Eberth (1985, fig. 3) described and illus
trated two distinct morphs of Sphenacodon ferocior, based primarily on
the shape of the jugal. In both, the squamosal-jugal suture extends be
yond the posterior margin of the temporal fenestra. In contrast, this
suture in S. ferox is not as long and does not extend beyond the posterior
margin of the fenestra (Fig. 20).
10. Parietals with circular embayments: The parietals of
Sphenacodon ferox have a shape distinct from those of S. ferocior (Eberth,
1985, fig. 8). In S. ferox the parietals have hemispherical embayments on
their anterior margins (Fig. 17C-E), whereas in S. ferocior only a broadly
concave sutural scar for the postfrontals is present (Fig. 17F).
11. Dental field on the palatal ramus of the pterygoid does not

TABLE 3. Stratigraphic distribution of key Sphenacodon cranial specimens
discussed in the text.
183
extend to its lateral margin: Eberth (1985, p. 22, fig. 21) described and
illustrated the dental field on the palatal ramus of the pterygoid of
Sphenacodon ferocior as extending to the element’s lateral margin, whereas
in NMMNH P-55367 it does not (Fig. 18A-B).
STRATIGRAPHIC DISTRIBUTION
Fossils of Sphenacodon are known principally from New Mexico
(Fig. 1). The only non-New Mexican records appear to be from the
Halgaito Formation, Organ Rock Shale and Cedar Mesa Sandstone of the
Arizona-Utah borderland (Monument Valley) (Vaughn, 1962, 1964;
Sumida et al., 1999). In New Mexico, Sphenacodon fossils are known
from the: (1) Red Tanks Member of the Bursum Formation in the Lucero
uplift (Harris et al., 2004); (2) upper part of the El Cobre Canyon
Formation in the Arroyo del Agua area of Rio Arriba County (Berman,
1993; Lucas et al., 2005); (3) El Cobre Canyon and Arroyo del Agua
formations in Cañon del Cobre, Rio Arriba County (Lucas et al., 2005);
(4) Abo Formation in the vicinity of Jemez Springs (San Diego Canyon);
(5) Sangre de Cristo Formation near Villaneuva (Vaughn, 1964; Berman,
1993); (6) Abo Formation near Tularosa, Otero County (Vaughn, 1969);
and (7) the Abo Formation at Gallina Well in Socorro County (Berman,
1993; Spielmann et al., 2009). The stratigraphic distribution of key
Sphenacodon fossils is summarized in Table 3.
Lucas (2006) used the first appearance datum (FAD) of
Sphenacodon to define the beginning of the Coyotean land-vertebrate
faunachron, an interval of time that encompasses the Pennsylvanian
Permian boundary (Fig. 23). Thus, the lowest occurrence (LO) of
Sphenacodon in New Mexico (and in the Halgaito Formation) is in strata
of demonstrable Virgilian (Late Pennsylvanian) age (Red Tanks Member
of the Bursum Formation). Its youngest New Mexican record is in strata
of Seymouran (late Wolfcampian) age – Arroyo del Agua Formation.
Both species – S. ferocior and S. ferox – co-occur in Coyotean age strata,
but only S. ferox is known from the Seymouran, so it is the youngest
known Sphenacodon species.
ACKNOWLEDGMENTS
Pat Holroyd (UCMP), Bill Simpson (FMNH), Walter Joyce and
Dan Brinkman (YPM), Carl Mehling (AMNH) and Chuck Schaff (MCZ)
provided collection access. Sean Modesto and Stuart Sumida provided
helpful reviews that improved the manuscript.
184
REFERENCES
Behrensmeyer, A.K., 1975, The taphonomy and paleoecology of Plio
Pleistocene vertebrate assemblages east of Lake Rudolf, Kenya: Bulletin
of the Museum of Comparative Zoology, v. 146, p. 474-578.
Berman, D.S, 1977, A new species of Dimetrodon (Reptilia: Pelycosauria)
from the Lower Permian of north-central New Mexico: Journal of
Paleontology, v. 51, p. 108-115.
Berman, D.S, 1993, Lower Permian vertebrate localities of New Mexico
and their assemblages: New Mexico Museum of Natural History and
Science, Bulletin 2, p. 11-21.
Case, E.C., 1907, Revision of the Pelycosauria of North America: Carnegie
Institution of Washington, Publication 55, 176 p.
Eberth, D.A., 1985, The skull of Sphenacodon ferocior, and comparisons
with other sphenacodontines (Reptilia: Pelycosauria): New Mexico Bu
reau of Mines and Mineral Resources, Circular 190, 40 p.
Glantz, S.A., 2005, Primer of Biostatistics: New York, McGraw-Hill Inc.,
520 p.
Haldane, J.B.S., 1948, Suggestions as to quantitative measurement of rates
of evolution: Evolution, v. 3, p. 51-56.
Hammer, O., Harper, D.A.T. and Ryan, P.D., 2001, PAST: Paleontological
statistics software package for education and data analysis, v.4, p. 9.
Harris, S.K., Lucas, S.G., Berman, D.S and Henrici, A.C., 2004, Vertebrate
fossil assemblage from the Upper Pennsylvanian Red Tanks Member of
the Bursum Formation, Lucero uplift, central New Mexico: New Mexico
Museum of Natural History and Science, Bulletin 25, p. 267-283.
Huene, F.v., 1908, Neue und verkannte Pelycosaurier: Reste aus Europe:
Centralblatt fur Mineralogie, Geologie und Palaontologie, v. 14, p. 431-
434.
Langston, W., Jr., 1952, Permian vertebrates of New Mexico [Ph.D. the
sis]: University of California, Berkeley, p. 141-153.
Lucas, S.G., 2005, Permian tetrapod faunachrons: New Mexico Museum of
Natural History and Science, Bulletin 30, p. 197-201.
Lucas, S.G., 2006, Global Permian tetrapod biostratigraphy and
biochronology; in Lucas, S.G., Cassinis, G. and Schneider, J.W., eds.,
Non-marine Permian biostratigraphy and biochronology: Geological
Society, London, Special Publication 265, p. 65-93.
Lucas, S.G. and Krainer, K., 2005, Stratigraphy and correlation of the
Permo-Carboniferous Cutler Group, Chama Basin, New Mexico: New
Mexico Geological Society, Guidebook 56, p. 145-159.
Lucas S.G., Harris, S.K., Spielmann, J.A., Berman, D.S and Henrici, A.C.,
2005, Vertebrate biostratigraphy and biochronlogy of the Pennsylva
nian-Permian Cutler Group, El Cobre Canyon, northern New Mexico:
New Mexico Museum of Natural History and Science, Bulletin 31, p.
128-139.
Lucas, S. G., Spielmann, J. A., Rinehart, L. F., and Martens, T., 2009,
Dimetrodon (Amniota: Synapsida: Sphenacodontidae) from the Lower
APPENDIX
Voorhies Group definitions used in this study; after Behrensmeyer (1975).
Permian Abo Formation, Socorro County, New Mexico: New Mexico
Geological Society, Guidebook 60, p. 281-284.
Lyman, R.L., 1994, Vertebrate taphonomy, Cambridge Manuals in Archae
ology: Cambridge, Cambridge University Press, 524 p.
Marsh, O.C., 1878, Notice of new fossil reptiles: American Journal of
Science, v. 15, p. 409-411.
Paton, R.L., 1974, Lower Permian pelycosaurs from the English midlands:
Palaeontology, v. 17, p. 541-552.
Reisz, R.R., 1986, Pelycosauria: Encyclopedia of Paleoherpetology, Part
17A, 102 p.
Rinehart, L.F., Lucas, S.G. and Harris, S.K., 2007, Lithology and taphonomy
of an early Permian Sphenacodon bonebed in Canon del Cobre, north
central New Mexico: New Mexico Geology, v. 29, no. 2, p. 63.
Romer, A.S. and Price, L.I., 1940, Review of the Pelycosauria: Geological
Society of America, Special Paper 28, p. 1-538.
Shipman, P., 1981, Life history of a fossil: An introduction to taphonomy
and paleoecology: Cambridge, Harvard University Press, 222 p.
Simpson, G.G., Roe, A. and Lewontin, R.C., 1960, Quantitative Zoology:
New York, Harcourt, Brace and World Inc., 440 p.
Spielmann, J.A., Lucas, S.G., Berman, D.S and Henrici, A.C., 2009, An Early
Permian (Wolfcampian-Seymouran) vertebrate fauna from the Abo
Formation (Scholle Member), Gallina Well, Socorro County, NM: New
Mexico Geological Society, Guidebook 60, p. 69-70.
Sumida, S. S., Lombard, R. E., Berman, D. S and Henrici, A. C., 1999, Late
Paleozoic amniotes and their near relatives from Utah and northeastern
Arizona, with comments on the Permian-Pennsylvanian boundary in
Utah and northern Arizona: Utah Geological Survey, Miscellaneous Pub
lication, 99-1, p. 31-43.
Vaughn, P.P., 1963, The age and locality of the Late Paleozoic vertebrates
from El Cobre Canyon, Rio Arriba County, New Mexico: Journal of
Paleontology, v. 37, p. 283-286.
Vaughn, P.P., 1964, Vertebrates from the Organ Rock Shale of the Cutler
Group, Permian of Monument Valley and vicinity, Utah and Arizona:
Journal of Paleontology, v. 38, p. 567-583.
Vaughn, P.P., 1969, Early Permian vertebrates from southern New Mexico
and their paleozoogeographic significance: Los Angeles County Mu
seum Contributions in Science, v. 166, p. 1-22.
Voorhies, M.R., 1969, Taphonomy and population dynamics of the early
Pliocene vertebrate fauna, Knox County, Nebraska: Contributions to
Geology, Special Paper No. 1.
Walls, G.L., 1942, The vertebrate eye and its adaptive radiation: Cranbrook
Institute of Science Bulletin 19, p. 1-785.
Williston, S.W., 1917, The phylogeny and classification of reptiles: Contri
butions of the Walker Museum, v. 2, p. 61-71.
Williston, S.W. and Case, E.C., 1913, A description of Edaphosaurus Cope:
Carnegie Institution of Washington, Publication 181, p. 71-81.
 Lucas et al., eds., 2010, Carb-Permian transition in Cañon del Cobre. New Mexico Museum of Natural History and Science Bulletin 49.

185
REDESCRIPTION OFTHE SKULL OF LIMMOSCELIS PALUDIS
WILLISTON(DIADECTOMORPHA: LIMNOSCELIDAE) FROM THE
PENNSYLVANIAN OF CAÑON DELCOBRE, NORTHERN NEWMEXICO

DAVID S BERMAN1, ROBERTR. REISZ2 AND DIANESCOTT2

1Carnegie Museum of Natural History, Section of Vertebrate Paleontology, Pittsburgh, Pennsylvania, USA 15213,
bermand@carnegiemnh.org; 2 Department of Biology, University of Toronto at Mississauga, 3359 Mississauga Rd.,
Mississauga, Ontario, Canada L5L 1C6, rreisz@utm.utoronto.ca, dscott@utm.utoronto.ca

Abstract—The skull of the Late Pennsylvanian diadectomorph Limnoscelis paludis Williston, 1911, is redescribed
in detail on the basis of the holotype YPM 811, the only one of three known specimens to include the skull.
Although originally collected by David Baldwin in 1879 in Cañon del Cobre, Rio Arriba County, north-central New
Mexico, the anatomy of the skull has been the subject of only two, relatively recent, comprehensive, detailed
descriptions. In addition to inaccuracies, these two studies contain numerous, contradictory interpretations. A
primary goal of the description presented here is an attempt to resolve these deficiencies. In the more recent of the
previous descriptions the skull had been separated into 17 pieces along pre-existent fractures and more extensively
prepared to facilitate study. Most were prepared further and illustrated for the present study, permitting descrip
tion of internal structures typically not available in intact specimens. On the basis of new anatomical information
a revised reconstruction of the cranium and mandible is presented. In addition, a reconstruction of the entire
skeleton of L. paludis in lateral and dorsal views and a whole-body restoration in lateral view is presented, but
without discussion. A more accurate, detailed description of the skull of L. paludis allows comparisons with other
diatectomorphs in light of a greatly expanded knowledge of their anatomy.

INTRODUCTION
The Late Pennsylvanian Limnoscelis paludis Williston, 1911, has
always been considered as one of the most important late Paleozoic
tetrapods because of the widely accepted view of its close phylogenetic
association with the origin of amniotes. Since its original description by
S. W. Williston in a series of publications (Williston, 1911a, b, 1912),
there have been only two detailed studies of the anatomy of the cranium
and mandible. The first was by A. S. Romer (1946), who reinterpreted
the anatomy of the cranium (defined here as the skull without the lower
jaw) on the basis of the growing knowledge of early reptiles. His recon
struction of the cranium contained only a few major errors, mainly in the
temporal and occipital regions (Berman, 2000). Subsequently, M. A.
Fracasso (1987) described the skull (defined here as including the cra
nium and mandible) in great detail in his Ph.D. thesis, but published only
a redescription of the braincase (Fracasso, 1983), although accompanied
also by reconstructions of the cranium in ventral and occipital views
without comments. More recent restudies (Berman and Sumida, 1990;
Berman et al., 1992; Berman, 2000) of selective areas of the cranium have
noted numerous, important errors in the descriptions by Romer (1946),
but most notably by Fracasso (1983, 1987). A reinterpretation of the
occipital region of the cranium by Berman (2000) corrected some of the
errors in their descriptions and reconstructions.
The continued interest in the skull anatomy of Limnoscelis paludis
since Williston’s studies (1911a, b, 1912) has resulted in several at
tempts at further preparation. Fracasso (1987) facilitated his study by
separating the skull into 17 pieces along existing, natural breaks, which
allowed an unprecedented access to internal structures. The skull was
still in a disassembled state at the initiation of the present study, but for
purposes of illustration was partially reassembled into seven pieces.
Unfortunately, some important features of the skull were lost or com
promised during preparation by previous investigators. As an example,
clearly the dermal sculpturing must have been much more extensive and
pronounced than the current state of the skull indicates, but most impor
tantly, some details of the occiput and dentition were lost.
Although named and initially described by Williston in 1911, the
materials pertaining to Limnoscelis paludis were originally collected by
David Baldwin in 1879 in Cañon del Cobre, Rio Arriba County, north
central New Mexico, while employed by O. C. Marsh of the Yale Peabody
Museum. These included an essentially complete, articulated skeleton
(YPM 811) and two partial postcranial skeletons (MCZ 1947 and 1948,
formerly YPM819 and 809, respectively). According to Fracasso (1980,
1983), the discovery of the specimen now believed to be the holotype,
YPM 811, which included the only known skull of L. paludis, was first
reported to Marsh by Baldwin in a letter dated September 19, 1879.
However, as is so common in vertebrate paleontology, its significance
was overlooked, probably due to the distraction of the discovery at that
time of the much more spectacular Como Bluff dinosaurs by another of
Marsh’s collecting teams and his rival E. D. Cope. As a consequence, the
L. paludis specimens were not studied until 30 years later by Williston
(1911a, b, 1912).
As the type species and best known member of Limnoscelidae, a
detailed study of the skull morphology of Limnoscelis paludis will per
mit a more rigorous evaluation of its relationships with Diadectidae and
Tseajaiidae, which together with Limnoscelidae constitute
Diadectormorpha. Recent phylogenetic analyses (Berman et al., 1992;
Lee and Spencer, 1997; Berman, 2000; Kissel and Reisz, 2004) strongly
support the hypothesis that the three families form a natural group, the
Diadectomorpha. Most importantly, however, the present study will
provide a greater anatomical basis to help resolve the controversy of the
relationship of diadectomorphs to the amniotes. Amniota is divisible
into two universally accepted clades, Synapsida (extinct mammal-like
forms and their mammalian descendants) and Reptilia (all amniotes ex
clusive of Synapsida), and two hypotheses have been proposed relating
Diadectomopha to these two clades: firstly, the diadectomorphs com
prise a nonamniote sister clade to all amniotes, with Diadectomorpha
and Amniota forming a monophyletic clade which constitutes the most
recent definition of Cotylosauria (Laurin and Reisz, 1995, 1997; Reisz,
2007); and secondly, the Diadectomorpha and Synapsida comprise an
unnamed, primitive sister clade to Reptilia (Berman et al., 1992; Lee and
Spencer, 1997; Berman, 2000). The latter hypothesis would mandate the
assignment of Diadectomorpha to Amniota. However, in order to avoid
any confusion in the description and comparisons of L. paludis pre
sented below, the more widely accepted former hypothesis that
186
diadectomorphs comprise a nonamniote sister clade to all amniotes is
followed.
Regarding the intrarelationships of the Diadectomorpha, there is
wide consensus that Diadectes Cope, 1878, and presumably also the
diadectids Desmatodon Case, 1908, and Orobates Berman, Henrici, Kissel,
Sumida, and Martens, 2004, on the one hand, and Tseajaia Vaughn, 1969,
on the other, share a more recent common ancestor than either does with
Limnoscelis (Berman et al., 1992; Kissel and Reisz, 2004; Reisz, 2007).
A thorough reconsideration of the phylogenetic relationships of the
Diadectomorpha to Amniota, however, must await further study and
will be presented at a future date.
Despite a long history of extensive handling and repeated at
tempts at preparation and description, the skull of Limnoscelis paludis
remains inadequately described. We have therefore undertaken that task
here and present a revised, detailed description accompanied by a series
of illustrations of disassociated portions of the cranium and mandible, as
well as reconstructions of the cranium and mandible in several views. To
this is added a revised reconstruction of the entire skeleton in lateral and
dorsal views, as well as a whole-body restoration (Fig. 1), but without
discussion.
In the description that follows, comparisons of Limnoscelis paludis
with other diadectomorphs is limited to a few taxa whose cranial and
mandible anatomies are adequately known: the Late Pennsylvanian L.
dynatis Berman and Sumida, 1990, the only other valid member of the
genus (Berman and Sumida, 1990); among the Diadectidae the Early
Permian type genus Diadectes, which has recently been redescribed in
detail (Berman et al., 1992; Berman et al., 1998; Berman, 2000), and the
closely related, primitive diadectids Desmatodon hesperis Vaughn, 1969
(Vaughn, 1969, 1972; Berman and Sumida, 1995) and Orobatis pabsti
Berman, Henrici, Kissel, Sumida, and Martens, 2004 (Berman et al.,
2004) from the Late Pennsylvanian and Early Permian, respectively; and
the Early Permian Tseajaia campi Vaughn, 1964, sole member of
Tseajaiidae, the holotype of which was described by Vaughn (1964) and
Moss (1972), but whose cranial anatomy description was more recently
updated on the basis of a subsequently discovered specimen (Berman et
al., 1992). In the description that follows the repetitive citation of these
references has been eliminated as an unnecessary redundancy. However,
whenever a partial list of the comparative diatectomorphs is given, it
should be understood that the relevant morphological data are either
lacking or inconclusive for the excluded taxa.
ABBREVIATIONS
The following acronyms refer to institutional repositories of speci
mens: CM, Carnegie Museum of Natural History, Pittsburgh, Pennsyl
vania; MCZ, Museum of Comparative Zoology, Harvard University,
Cambridge, Massachusetts; MNG, Museum der Natur, Gotha, Ger
many; YPM, Yale Peabody Museum, New Haven, Connecticut.
Key to abbreviations used in figures for anatomical structures: an,
angular; ar, articular; bo, basioccipital; c1, c2, anterior, posterior
coronoids; cp, cultriform process; d, dentary; e, epipterygoid; ec,
ectopterygoid; ex, exoccipital; f, frontal; j, jugal; l, lacrimal; m, maxilla;
n, nasal; op, opisthotic; p, parietal; pal, palatine; pf, postfrontal; pm,
premaxilla; po, postorbital; pp, postparietal; pra, prearticular; prf, pre
frontal; ps, parasphenoid; pt, pterygoid; q, quadrate; qj, quadratojugal;
sm, septomaxillae; sp, splenials; sph, sphenethmoid; sq, squamosal; so,
supraoccipital; st, supratemporal; su, surangular; t, tabular; tt, tooth; v,
vomer.
SYSTEMATIC PALEONTOLOGY
COTYLOSAURIA Cope, 1880
DIADECTOMORPHAWatson, 1917
LIMNOSCELIDAE Williston, 1911
Limnoscelis paludis Williston, 1911
Revised Diagnosis - Diadectomorph cotylosaur that can be dis
tinguished from non-limnoscelid diadectomorphs by the following char
acters: elongated snout in which the nasal is twice the length of the
parietal and has a greatest length equal to 50% of the midline skull length;
the premaxillae united externally by a strongly interdigitating suture to
form a massive, bluntly rounded snout tip in which the ventral margin is
inclined anteroventrally; long vomerine process of premaxilla extends to
a level at least equal to that of the maxillary process of the premaxilla;
anterior two teeth of premaxilla and second and third teeth of dentary
dominate the upper and lower marginal series as long, massive canines;
long anterior and posterior processes of nasals taper laterally from mid
line to extend far beyond their midline union; anterior portions of frontals
converge strongly on the midline to wedge deeply between posterior
processes of the nasals; broad, dorsally arching suture separates the
postorbital and supratemporal from the squamosal in which the lateral
margin of postorbital and supratemporal project ventrally and slightly
laterally about 3 to 5 mm beyond lateral surface of squamosal in a struc
ture likened to the eave of a house roof; lacrimal long and narrow, with a
height only slightly exceeding that of the maxilla; posterior process of
jugal along ventral margin of skull roof extends much farther as it narrows
in a wedge-like terminus below the anterior end of quadratojugal; absence
of a continuous single row of well-developed teeth that extends along the
medial margin of premaxillary process of vomer and palatal process of
pterygoid; narrow, arcuate flange projects medially and slightly dorsally
from a level just above the ventral edge of the quadrate ramus; lateral
surface of dentary at the symphysis of mandibles swollen into a mas
sive, hemispherical structure; prearticular and articular form a large, mas
sive, anteriorly inflected pterygoideus process; a low, broadly convex
coronoid eminence of mandible formed by the surangular.
Comment - Although several limnoscelids have been named and
described on the basis of fragmentary materials, a recent taxonomic re
view of the family (Wideman, 2002, Wideman et al., 2005) persuasively
demonstrates that only two taxa are valid, Limnoscelis paludis and L.
dynatis.
DESCRIPTION AND COMPARISONS
General -The skull of YPM 811 is essentially complete, missing
only the hyoid elements and stapes (Fig. 2). Although there has been
some loss of portions of the roofing bones and postmortem compression
of the cranium, this has been easily compensated for in producing reli
ably accurate reconstructions in several views (Figs. 3-4). Modest over
all dorsoventral crushing with a slight oblique component directed vent
rolaterally to the right side has caused the most obvious distortions of
the cranium. This has distorted the entire cranium in several ways (Figs.
2, 5-7, 12): 1) overall ventral depression of the skull table by approxi
mately 2 cm; 2) minor ventral vaulting of the snout along the midline
union of the nasals and frontals; 3) inward buckling of the postorbital
cheeks, resulting in a strong ventrolateral splaying of the cheeks from the
skull table; 4) breakage and displacement of the strut-like dorsal pro
cesses of the epipterygoids; and 5) some telescoping of the postorbital
bones of the right temporal region. Little distortion of the mandibles has
occurred, and they are tightly attached to the skull without any displace
ment at the jaw articulation. However, as a consequence of their tight
attachment to the skull, particularly at the jaw articulation, the man
dibles have been displaced laterally due to the outward splaying of the
postorbital cheek regions of the cranium. Concomitant with this dis
placement the jaw has shifted slightly to the left side of the cranium, the
angle between the mandibles has increased, and the left mandible has
been displaced dorsally to lie against the left side of the snout and slightly
lateral to the upper tooth row.
The cranium and mandibles are massively built but low in lateral
profile, with a robust snout region that appears slender only because of
the elongation of the antorbital region (Figs. 2-4). The massiveness and
structural strength of the snout region is most prominently exemplified
by the large size of the premaxilla, but the entire cranium is also robustly
constructed, as indicated by the thickness of the skull roof elements,
187
188

FIGURE 2. Limnoscelis paludis, holotype, YPM 811. A, dorsal, B, ventral (palatal), and C, left lateral views of skull.
 189

FIGURE 3. Reconstruction of Limnoscelis paludis, holotype, YPM 811. Cranium in A, dorsal, and B, lateral views; C, left mandible in lateral and medial
views.
190
particularly the circumorbital and ventral marginal bones (Fig. 9). The
interorbital and intertemporal regions of the cranium contribute to an
unusually broad, essentially flat skull table, whereas the dorsal elements
of the snout are slightly domed in transverse section. In lateral profile the
nearly vertical postorbital cheek region is low and only slightly taller
than the snout. The external nares and orbits face laterally, being visible
only slightly in dorsal view. The pineal foramen is transversely oval in
outline. The sculpturing of the skull table ends posteriorly in a very low,
occipital ridge that extends transversely just anterior to the posterior
margin of the parietal, then curves posterolaterally for a short distance
along the anterolateral margin of the tabular. The ridge presumably delin
eates the boundary of the posterior extent of the direct skin attachment
to the skull roof. Posterior to the occipital ridge the skull table slopes
posteroventrally to occupy a plane somewhat intermediate between the
horizontal plane of the skull table and the nearly vertical occiput. This
narrow, smoothly finished, posterior portion of the skull table is bowed
slightly anteriorly in dorsal view and includes a very thin strip of the
posterior margin of the parietal, the entire, single, median postparietal,
and a posterolaterally directed, horn-like portion of the tabular.
In palatal view of the cranium (Fig. 4A) the long anteroposterior
length of the internal nares, which occupies nearly 30 % of the skull
length, reflects the great elongation of the snout. A moderately deep,
anteroposteriorly elongate, narrow, oval, basin-like depression straddles
the contact of the medial margin of the ectopterygoid with the pterygoid
just anterior to the transverse flange. An identical depression is seen in
Diadectes (Olson, 1947), whereas in Orobates there is a less well-de
fined, shallow depression of the ectopterygoid adjacent to its sutures
with the pterygoid and palatine. The narrow interpterygoid vacuity
occupies half the length of the palate anterior to the basicranial articula
tion. The subtemporal adductor fossae, the largest paired openings on
the ventral surface of the cranium, are located between the quadrate
ramus of the pterygoid and the contributions of the quadratojugal and
jugal to the ventral rim of the cranium. Their large size undoubtedly
reflects a massive development of the adductor musculature.
The occipital region of the skull (Figs. 2, 3A, 4B) is broad and low
and dominated by a large occipital plate that is inclined slightly
anterodorsally, which is barely sufficient to expose almost its entire
surface, as well as the foramen magnum, in dorsal view.
The mandibles are massively built despite having a relatively low
lateral profile (Figs. 2C, 13-14). They are markedly shorter than the
skull, suggesting that when tightly closed the large anterior mandibular
teeth must have fitted into the anterior ends of the internal nares. The
symphyseal area is massive, the coronoid prominence is low and ex
ceeded in height by the largest marginal teeth at the symphysis, and the
medial fenestra is narrow but long, with a length equaling 28% of that of
the mandible. The sculpturing pattern of the lateral surface mandible at
its massive anterior end is dominated by coarse pitting, whereas most of
the posterior half of the lateral surface exhibits a very fine pattern of
alternating grooves and ridges radiating from a central area of minute pits
on the surangular.
Skull Roof
Premaxilla -The premaxillae (Figs. 3A, B, 4A, 6, 8, 11) are united
along the dorsal midline in a strongly interdigitating suture and form a
massive, bluntly rounded snout tip in which the ventral margin in lateral
view is inclined anteroventrally. The surface sculpturing of the premax
illa is more pronounced than on other parts of the skull roof, despite
some weathering. This is due in part to the premaxillary sculpturing
being associated with a series of foramina that probably allowed passage
of blood vessels and nerves to the skin of the snout. Three processes
unite the premaxilla with the maxilla, vomer, and nasal. Joined, narrow
dorsal processes of the premaxillae extend posterodorsally along the
midline between anterior processes of the nasals. Fracasso (1983) incor
rectly described the dorsal processes as much shorter and wider and as
enclosing a midline internasal bone. The maxillary process of the premax
illa, as in Tseajaia, Desmatodon, Orobates, and Diadectes, forms nearly
the anterior two-thirds of the ventral border of the external naris. In
Limnoscelis paludis it extends posteriorly as a vertical, wing-like blade
that contacts the medial surface of the premaxillary process of the max
illa to about the level of its second tooth. Although Fracasso (1983) did
not recognize a palatal, vomerine process of the premaxilla in L. paludis,
additional preparation has revealed a long, narrow pair of vomerine pro
cesses with rounded ventral margins exposed in a narrow, midline gap
between the anterior ends of the vomers that clasped them together. The
exposed ventral margins of the vomerine processes reveal a posterior
extent equal to at least that of the maxillary processes of the premaxillae,
but their true posterior extent may be greater. Although the vomerine
process of an isolated premaxilla of L. dynatis was described as short,
reexamination indicates that it is incomplete distally, and its length is
therefore indeterminate. The vomerine processes are extremely short in
Tseajaia, Orobates, and Diadectes.
Each premaxilla has four tooth positions, with the first two being
represented by a tooth and socket pair but with reversed serial replace
ment patterns. The first two teeth greatly dominate the series in mas
siveness and length, although exhibiting a marked decrease in size poste
riorly. The massiveness of the premaxilla is undoubtedly related to the
large size and deep implantation of this pair of teeth. The much smaller
posterior two premaxillary teeth exhibit a dramatic decrease in size pos
teriorly. The premaxillary teeth are similar, however, in being essentially
subcircular in cross-section with the basal halves maintaining a constant
diameter and the distal halves or crowns tapering to a slightly recurved
point. In addition, they exhibit very fine, narrowly spaced, groove-like
striations that are most noticeable at their bases. Cross-sections of the
teeth reveal the striations to be indicative of a labyrinthine infolding of
the surface enamel that appears to penetrate the full thickness of the
dentine the entire length of the root and crown. Loss of the surface bone
of the premaxillary teeth precludes a description of their superficial
features. The marginal teeth in Limnoscelis paludis are similar to those of
other diadectomorphs in their deep implantation and apparent infolding
of the surface enamel deeply into the dentine.
Berman and Sumida (1990) noted one possible noteworthy differ
ence between the premaxillae in Limnoscelis paludis and L. dynatis. In
Vaughn’s (1969) description of the right premaxilla in L. dynatis he recog
nized only two teeth. However, as noted by Berman and Sumida (1990),
what Vaughn (1969) described as the base of the larger and more anterior
of the two, actually represents two teeth. But, as they cautioned, the
two teeth may represent a stage of one tooth replacing the other in a
single tooth socket and therefore should be counted as a single functional
tooth.
Septomaxilla - The septomaxillae (Figs. 3B, 6, 8) are nearly
complete, but have been distorted slightly by the dorsoventral crushing
of the cranium, but together they permit a reliable reconstruction of its
anatomy and relationship with neighboring elements. It is a squat ele
ment that has in general an hourglass outline with large, flared ventral and
dorsal ends, the former being substantially larger, which are separated by
a short, narrow, pillar-like dorsal process. Ventrally it contacts mainly
the premaxilla and very narrowly the maxilla and lacrimal along the ven
tral margin of the naris, whereas dorsally it appears to have ended short
of a contact with the nasal border of the naris. The dorsal process is
pierced anteroposteriorly by a septomaxillary canal (not visible in illus
trations). In all these features it is closely comparable to the septomaxillae
in Orobates and Diadectes. The septomaxilla in diadectomorphs is very
similar to that in basal, pelycosauria-grade synapsids, differing mainly in
the latter having a dorsal contact with the nasal (Romer and Price, 1940;
Wible et al., 1990).
Nasal - The nasals (Figs. 3A, B, 6, 8) are extremely thick, and in
overall length are the longest of the paired midline skull table elements,
reflecting the great elongation of the snout. They not only form the
dorsal surface of the snout, but also contribute significantly to the vent
rolateral curvature of the cheek. Their external surface retains at least

FIGURE 4. Reconstruction of cranium of Limnoscelis paludis, holotype, YPM 811, in A, ventral (palatal), and B, occipital views.
remnants of the original ornamentation. Long anterior and posterior pro
cesses of the nasals taper laterally as they extend far beyond their mid
line union. The anterior process contacts the entire lateral margin of the
dorsal process of the premaxilla to form the dorsal margin of the naris and
end in a short, narrow, anteroventrally directed splint of bone. The pos
terior processes of the nasals, on the other hand, are deeply wedged
between the anterior portions of the frontal and prefrontal.
Frontal - As in all diadectomorphs, the paired frontals (Figs. 3A,
5-6, 9) are relatively narrow and excluded widely from the orbital margin
by a wide contact between the postfrontal and prefrontal. In contrast to
the condition in other diadectomorphs, however, where the frontals are
narrowly subrectangular in outline, in Limnoscelis paludis the anterior
portions converge strongly on the midline, wedging deeply between
posterior processes of the nasals, whereas posteriorly they expand gradu
ally to their contact with the parietals. The anterior half of their midline
contact is essentially straight, whereas posteriorly it is strongly inter
digitating.
Parietal - The large parietals (Figs. 3A, 5, 10) dominate the pos
torbital skull table. Their great width is due in part to a well-defined,
subrectangular lateral lappet that extends a short distance anterolaterally
and represents either a replacement or an incorporation of the
intertemporal bone (Berman et al., 1992, 1998). The anterior, lateral, and
posterior margins of the lappet are bordered by the postfrontal, postor
191
bital, and supratemporal, respectively. An unusual feature is an extremely
narrow, spike-like projection of the posterolateral corner of the parietal
that partially separates the tabular and supratemporal. The parietal not
only has a strongly interdigitating suture with the frontal, which on the
ventral surface of the skull roof underlies extensively the frontal. A break
through the left side of the skull table reveals that the parietal also
narrowly underlies the supratemporal. The sculpturing of the parietals
ends posteriorly in a very low occipital ridge narrowly anterior to and
parallel with their posterior margin. The midline parietal suture is inter
rupted at its midlength by a large pineal foramen, and, although it is
relatively smaller than that in Tseajaia, Orobates, and Diadectes, it occu
pies 30% of the midline length of the parietal.
Tseajaia, Orobates, and Diadectes also possess a lateral lappet of
the parietal, but it differs from that in Limnoscelis paludis in having a
relatively greater transverse width and projecting directly laterally. The
contacts of the lateral lappet to adjacent elements are identical in L.
paludis, Orobates, and Diadectes, but in Tseajaia the lateral margin also
contacts the squamosal. The parietal of Tseajaia is also distinguishable
by its lack of the spike-like projection of its posterolateral corner that
partially separates the tabular and supratemporal.
Postparietal - The postparietal (Figs. 3A, 4B, 5, 12) is a single,
median element with a trapezoidal outline in which the transverse width
is about five times the midline length, and its smoothly finished surface
192
lacks the typical, midline ridge of cotylosaurs. The postparietal is also a
single element in Orobates and Diadectes, but paired in Tseajaia. As in
other diadectomorphs, the dorsal surface of the postparietal, although
exhibiting a slight anterior bowing, occupies a plane intermediate be
tween the horizontal skull table and the slightly anterodorsally inclined
plane of the occiput. In non-cotylosaurs the postparietal occupies a
horizontal plane that is continuous with skull table, whereas in amniotes
it occupies a near vertical plane that is continuous with the occiput. The
occipital margin of the postparietal overlaps the dorsal edge of the su
praoccipital in a rounded, slightly overhanging projection in what ap
pears to be a loosely sutured contact.
Prefrontal - The large prefrontal (Figs. 3A, B, 5-8, 10), is almost
entirely restricted to the dorsal skull table as essentially an
anteroposteriorly elongated rectangular, plate-like structure. It exhibits
an otherwise strong ventral curvature into the vertical cheek surface of
the snout. A transverse break through the skull roof at a level just anterior
to the orbits (Fig. 9) reveals the prefrontal as forming a massive, out
wardly directed, wedge-shaped, supraorbital shelf. The greater lower
portion of the ventrolateral surface of the shelf has an extensive contact
with the dorsal half of the medial surface of the lacrimal. Posterior to its
contact with the lacrimal the free, ventromedial, angular extension of the
shelf is quickly reduced to a low, rounded ridge that is continued a short
distance along the internal surface of the prefrontal-frontal contact and
then onto the postfrontal (Fig. 10A). Remnants of the surface sculptur
ing of the prefrontal, as in the nasal, are present as shallow pits and faint
ridges and grooves.
Postfrontal - The postfrontals (Figs. 3A, B, 5, 7, 10) are well
preserved and almost entirely restricted to the dorsal skull table, where
they appear subrectangular in outline. They are plate-like except for a
slight ventral curvature along the orbital margin, especially at the
posterodorsal margin of the orbit. The medial margin parallels the dorsal
orbital margin and contacts mainly the frontal, but also occupies the
right-angle notch between the anterior margin of the parietal and its
lateral lappet. The anterior and posterior contacts with the prefrontal
and postorbital, respectively, are oriented normal to the dorsal orbital
rim. Near its posterior, orbital margin the ventral surface thickens greatly
and from it a transversely expanded, flange-like process that projects a
short distance ventrally along the posterior orbital wall.
Postorbital - Exposure of the postorbital is divided nearly equally
between the skull table and cheek (Figs. 3A, B, 5, 7). On the skull table
it contacts the entire lateral margin of the lateral lappet of the parietal
between contacts with the anterior postfrontal and posterior supratem
poral. In lateral view the cheek portion of the postorbital forms the
greater portion of the posterior rim of the orbit from which it projects
posteriorly in a wing-like outline. Its orbital margin is greatly thickened
and ends ventrally in a laterally overlapping contact with a short, narrow,
dorsal orbital process of the jugal. Superficially this suture is directed
posterodorsally and continues in a broad, dorsally arching contact be
tween the ventrolateral margins of the postorbital and supratemporal
and the dorsal margin of the squamosal. This suture is unusual in that the
postorbital and supratemporal project about 3 to 5 mm beyond the
lateral surface of the squamosal in what can be likened to the eave of a
house roof. Whereas the external surface of the eave is a continuation of
the sculptured skull table, the internal surface is smoothly beveled to an
acute-angled edge along its free margin.
Maxilla - The long, low maxilla, which extends between the
midlength levels of the external naris and orbit, reflects the skull’s profile
(Figs. 3A, B, 4A, 8, 11, 13-14). In lateral view the ventral edge of the
maxilla is slightly convex for almost its entire length except at its anterior
end, where it is redirected slightly anteroventrally beneath the external
naris. Here it not only overlaps laterally the maxillary process of the
premaxilla, but also contributes to the ventral shelf of the external naris
to support partially the base of the septomaxilla. The lateral surface of
the maxilla extends a short distance above a thick alveolar shelf as a low,
vertical, dorsal lamina. A transverse break passing through the snout a
short distance anterior to the orbits (Fig. 9) reveals the dorsal lamina as
being slightly thinner and subequal in height to the alveolar shelf. At this
level the dorsal lamina has a strongly interdigitating suture with the
ventral margin of the anterior end of the suborbital process of the jugal, as
well as sheathing its lateral and medial surfaces. The entire dorsal margin
of the maxilla appears to be narrowly overlapped laterally by the lacri
mal in a nearly straight suture.
The maxillary teeth are set in very deep sockets. The implanta
tion, however, cannot be considered strictly as thecodont, because the
ventral surface of the alveolar shelf slopes steeply ventrolingually, creat
ing a superficially pleurodont-like implantation (Figs. 9, 11). In addition,
reabsorption dental pits are clearly visible at the base of the lingual
margin of several teeth (Fig. 14). The tooth row, which includes 23 tooth
positions, extends nearly the entire length of the maxilla with only the
short, subnarial, premaxillary process being edentulous. The maxillary
teeth exhibit a gradual but steady increase in size serially from the first to
about the fifth tooth, then gradually decrease to the end of the series.
Although all the previously exposed teeth have lost their superficial
features, the fourth, left maxillary tooth (Fig. 14), which was not previ
ously exposed, exhibits a sharp, very narrow, low, cutting ridge that
extends the entire fore and aft lengths of the crown and passes through
the crown tip. The distal third of the crown, as well as the three preced
ing teeth, is slightly recurved. Presumably these features were present in
all the maxillary except possibly those of the posterior end of the series.
In contrast to the condition in Limnoscelis paludis, the maxillary
dentitions in L. dynatis and Tseajaia include a pronounced caniniform
tooth at the anterior end and the midlength of the series, respectively.
The upper and lower marginal dentitions in Desmatodon, Orobates, and
Diadectes are unique in that the teeth of the premaxilla and the anterior
ends of the maxilla and dentary are incisiform and much larger than the
succeeding cheek teeth, which are molariform.
Lacrimal - The entire length of the right lacrimal (Fig. 8) has been
severely crushed and fractured, whereas the left (not illustrated) is nearly
intact, but both have been buckled medially (Fig. 9). Restored (Fig. 3A,
B) the lacrimal appears as an elongate, rectangular, strap-like structure
that extends between substantial contributions to the external naris and
orbit. For most of the lacrimal’s anterior length, however, it exhibits a
rather slight ventrolateral curvature in transverse section, so that in the
reconstructed skull nearly its upper half is visible in dorsal view. Anteri
orly the lacrimal narrows slightly to match the height of the maxilla, as it
forms the greater ventral portion of the posterior rim of the external
naris. Here it joins with the maxilla in forming a small, medially directed,
narial shelf at the posteroventral corner of the naris. Above the narial
shelf the lacrimal duct (not visible in the illustrations) exits the narial
margin of the lacrimal. The ventral margin of the lacrimal contacts the
entire dorsal margin of the maxilla except for a narrow contact with the
suborbital process of the jugal at the anteroventral corner of the orbit. A
shallow, broadly concave depression of the posterior third of the lacrimal
deepens slightly posteriorly, where it is bordered by the thickened,
rounded orbital rim. A small V-shaped notch on the lateral surface of the
orbital rim marks the exit of a single, large foramen for the lacrimal duct.
In overall proportions the lacrimals in Tseajaia, Orobates, and
Diadectes differs from that in Limnoscelis paludis in being relatively
much shorter and taller, with the height greatly exceeding that of the
maxilla.
Jugal - Although the jugals have suffered severe breakage and
some distortion (Figs. 13-14), their structure can be reconstructed with
confidence (Fig. 3B). It is a relatively long, triradiate element, extending
posteriorly from the level of the anterior orbital margin to shortly be
yond the midlength of the postorbital cheek. A broad, anterior suborbital
process forms almost the entire ventral orbital rim, whereas a short,
narrow, dorsal process contributes to the posterior orbital margin. The
distal end of the dorsal process of the jugal, which is exposed in cross
section in the right (Fig. 7, 10) and medial view in the left (Fig. 10), is
laterally overlapped by the ventral, orbital process of the postorbital.
 193

FIGURE 5. Limnoscelis paludis, holotype, YPM 811. Posterior portion of cranium in dorsal view. A, Illustrated as preserved, and B, emphasizing preserved
limits of elements.
194
FIGURE 6. Limnoscelis paludis, holotype, YPM 811. Anterior portion of cranium in dorsal view. A, Illustrated as preserved, and B, emphasizing preserved
limits of elements.
However, as seen on the left side of the posterior portion of the cranium
exposed in the ventral view (Fig. 10), dorsoventral crushing has tele
scoped the contact between the two processes, producing the false ap
pearance depicted in the reconstructions by Romer (1946) and Fracasso
(1980, 1983) that the jugal lacks a dorsal orbital process. Although the
dorsal process in the reconstruction presented here is short, a longer
process, as in other diadectomorphs, is likely. This is most strongly
supported by the dorsal process of the isolated jugal of Limnoscelis
dynasties, which is much longer and formed nearly the ventral half of the
posterior orbital margin. In L. paludis, as well as probably in L. dynatis,
the posterior process is extremely long, extending along the ventral mar
gin of the skull roof to a short distance beyond the midlength level of the
postorbital cheek (Fig. 13). The dorsal margin of the process quickly
descends to form an acute-angled terminus below the anterior end of the
quadratojugal. As a result of its great length the posterior process of jugal
accounts for about 28% of the ventral margin of the skull roof. In marked
contrast, the posterior process in Tseajaia, Orobates, and Diadectes
ends well short of the midlength level of the postorbital cheek and is
directed away from the ventral margin of the skull roof as it ends wedged
between the quadratojugal and squamosal.
Quadratojugal - From about midlength of the postorbital cheek
the quadratojugal forms the posteroventral corner of the skull roof, with
both lateral and occipital exposures. Its lateral exposure (Figs. 3B, 13)
increases in height posterodorsally in a broadly concave contact with the
squamosal from a narrow contact with the distal end of the dorsal margin
of the posterior process of the jugal to the occipital margin of the skull
roof. The quadratojugal then thickens as it curves abruptly medially onto
the occipital surface of the skull roof (Figs. 3B, 4, 10, 12, 14). Here its
exposure is abruptly narrowed dorsally by a ventral, occipital expansion
of the squamosal before contacting the lateral half of the posterior sur
face of the articular condyle of the quadrate. The quadratojugal ends
medially in a short, sheet-like expansion that with the quadrate form the
lateral and medial margins of the quadrate foramen, respectively. The full
dorsal extent of the quadratojugal-quadrate contact above the quadrate
foramen is covered by the lateral margin of the occipital exposure of the
squamosal (Figs. 4, 12). However, within the subtemporal adductor
chamber, as in amniotes, the squamosal-quadrate suture is supported by
a dorsal process of the quadratojugal that overlaps their contact (Fig. 10).
Whereas the pattern of sculpturing on the lateral exposure of the
quadratojugal is similar to that seen on the jugal and squamosal, it is
greatly subdued on its occipital surface.
Squamosal - In lateral view the squamosal occupies most of the
postorbital cheek region, forming a rather flat, disk-like structure whose
dorsal margin forms a broad, dorsally arching contact with the postor
bital and supratemporal (Figs. 3B, 5, 7, 13). Two unusual features are
exhibited by the squamosal along this contact: 1) sculpturing covers all of
its surface except for a narrow band of smoothly finished bone along its
dorsal margin; and 2) its dorsal sutural margin is expanded dorsomedially

FIGURE 7. Limnoscelis paludis, holotype, YPM 811. Posterior portion of cranium in right lateral view. A, Illustrated as preserved, and B, emphasizing
preserved limits of elements.
in the form of a well-developed, approximately 5 mm wide flange that
thins to an acute-angled edge. The flange contacts the ventral surfaces of
the postorbital and supratemporal 3 to 5 mm from their ventrolateral,
eave-like margin. An identical, medially directed sutural flange has also
been described in an isolated right squamosal of Limnoscelis dynatis.
The eave-like structure of the contact of the squamosal with the
postorbital and supratemporal in Limnoscelis paludis is unique among
Paleozoic tetrapods and has been variously explained as either a “line of
weakness” (Romer, 1946), a “kinetic junction” (Panchen, 1972), the
“temporal line” site where the upper temporal fenestra of synapsids
195
evolved (Kemp, 1980), or simply a “loose junction” (Fracasso, 1983).
Fundamental to these interpretations is the flawed observation that the
contact between the cheek and skull table is a loose junction that was
inherited from osteolepiform antecedents, where it was integral to an
intracranial kinetic complex. On the basis of the redescription given here,
however, the so-called line of weakness in L. paludis must be reinter
preted as representing a complex, sutural contact that was surely immo
bile. Ventral exposure of the skull roof (Fig. 10) further strengthens this
conclusion in revealing that the dorsomedially directed sutural flange of
the squamosal extends beyond the postorbital-supratemporal to contact
196
also the ventral surface of the posterolateral corner of the lateral lappet
of the parietal. In addition, the anterior corner of the squamosal medially
overlaps the sutural contact between the postorbital and jugal (Fig. 7),
effectively uniting these elements in an immobile union. Although the
unique structural relationship of the union between the postorbital cheek
and skull table can no longer be characterized as a line of weakness, a
plausible, alternative functional explanation also cannot be offered.
The posterior margin of the cheek exposure of the squamosal
(Figs. 3B, 4B, 5, 7, 12) curves abruptly medially onto the occipital
surface of the skull roof in what can be referred to as the occipital flange
of the squamosal. Ventrally the flange has a short, bluntly pointed pro
cess that overlaps the quadratojugal. At the posterior margin of the
lateral exposure of the squamosal the occipital flange is demarcated by an
approximately 1 mm, step-like, depression of its surface that produces a
ridge between the two exposures. The two exposures of the squamosal
are also distinguished by a dramatic change in surface texture; a very fine
sculpturing pattern of alternating grooves and ridges radiating from a
central area of minute pits covers all of its lateral cheek exposure except
the narrow, smoothly finished band along the dorsal margin, whereas the
occipital flange is entirely smoothly finished. The occipital flange of the
squamosal not only extends medially across the posterior surface of the
dorsal extension of the quadratojugal, but also a narrow margin of the
dorsal extension of the quadrate above the articular condyle. Dorsally the
occipital flange extends about 2 mm beneath a freely projecting margin of
an occipital exposure of the supratemporal.
Supratemporal - The supratemporal is best described as con
sisting of two components (Berman, 2000). One, referred to as the proxi
mal, horn-like component, is triangular in outline, finely sculptured, and
occupies the greater portion of the posterolateral, horn-like extension of
the skull table (Fig.3A, B, 4B, 5, 7, 12). Posteriorly it curves slightly
ventrally, where the greater distal portion of medial margin of the horn
like component is defined by an approximately 1 mm, step-like depres-
sion of surface, producing a ridge that lines up with the identical ridge
along the lateral margin of the occipital flange of the squamosal. The ridge
also demarks the second component of the supratemporal, a thin, flat,
smoothly finished structure referred to as the distal occipital flange. It
extends a short distance ventrally before curving abruptly medially around
the ventrolateral margin of a much smaller similar component of the
tabular (Figs. 3A, B, 4B, 5, 12). An approximately 2 mm space separates
the flange from the underlying occipital flange of the squamosal. As
restored (Fig. 4B), the distal occipital flange of the supratemporal forms
the ventrolateral and most of the ventral margin of what has been identi
fied (Romer 1946; Berman, 2000) as the posttemporal fenestra and ends
in a bluntly rounded process just a few mm from the distal end of the
paroccipital process of the opisthotic. However, in Romer’s (1946) de
scription (p. 158) and reconstruction (fig. 8, p. 154) of the occiput of
Limnoscelis paludis the distal, occipital flange of the supratemporal
(misidentified as part of the tabular) ends in a broad contact with the
distal margin of the paroccipital process. The same relationship between
these two structures was proposed by Berman (2000), which is sup
ported by three lines of evidence: 1) the distal end of the more complete
left distal occipital flange of the supratemporal is at least marginally
incomplete; 2) the left paroccipital process of the opisthotic, as seen in
occipital view (Fig. 12), is definitely incomplete distally, as it is notice
ably shorter than the right. Weathering and/or aggressive preparation
could easily account for the approximately 3 mm gap separating the two
structures on the left side; and 3) the very narrow separation could also
be accounted for by the dorsoventral crushing of the skull, which has
resulted in the cheeks being splayed ventrolaterally from their original,
more vertical orientation (Fig. 12). All these altered features were either
overlooked or ignored by Kissel and Reisz (2004) and Reisz (2007), who
stated that there is no direct evidence of a contact between the supratem
poral and distal end of the paroccipital.
The supratemporal in Orobates and Diadectes can also be de
scribed (Berman, 2000; Berman et al., 2004) as divisible into the same
two components, but somewhat modified from that of Limnoscelis
paludis. Similarly, a proximal, horn-like component forms almost the
entire posterolateral corner of the skull table. The distal occipital flange
component in both diadectids, however, differs in forming a greatly
elongated, narrow, process-like extension that curves abruptly ventrally
into the plane of the occiput. Whereas the ventral extension is separated
by a narrow gap from the occipital surface of the quadrate, rather than
the squamosal, the distal end narrowly contacts the distal end of the
paroccipital process. In Tseajaia the supratemporal appears to consist
only of the proximal, horn-like component, which also occupies nearly
the entire posterolateral extension of the skull table. Interestingly, it
appears to contact the lateral margin of the postparietal, rather than that
of the tabular.
Tabular - The tabular (Fig. 3A, 4B, 5, 7, 12) is a complex bone
with large exposures on both the skull-roof table and occiput that can be
described as divisible into three components (Berman, 2000). A proxi
mal, horn-like component joins with the much larger, complementary
component of the supratemporal in forming the posterolateral corner of
the skull table. This component, considered part of the skull table, is
exposed almost entirely as a flat, smoothly finished surface that slopes
slightly posteroventral at the same plane as the postparietal, intermedi
ate between the skull table and the nearly vertical occiput. The distal end
of the horn-like component does exhibit, however, a distinct, low, dome
like swelling of unknown function. A second much smaller component,
referred to as the distal occipital flange of the tabular, is smoothly con
tinuous with ventral margin of the distal end of the proximal, horn-like
component. It is poorly preserved on both sides of the skull roof, but is
most complete on the left. It faces posteriorly and slightly medially and
has the form of a smoothly surfaced, thin, flange-like structure with a
broadly rounded, ventrolateral margin that overlaps the medial margin of
the distal occipital flange of the supratemporal. As restored, the tabular
supratemporal suture ends distally at the lateral margin of what has been
identified (Romer, 1946; Berman, 2000) as the posttemporal fenestra.
Romer (1946) and Berman (2000) suggested that the distal occipital
flange of the tabular may have continued farther medially, forming the
ventral margin of the posttemporal fenestra and ending in a narrow con
tact with the dorsolateral corner of the paroccipital process of the
opisthotic. This contact, as in the case of the complementary structure
of the supratemporal, was likely lost due to weathering and/or excessive
preparation. This interpretation is supported by a small, triangular frag
ment of what appears to be part of the distal occipital flange of the right
tabular that retains its contact with the posterior surface of the dorsolat
eral corner of the paroccipital process. In addition, immediately above
the contact is a small, smoothly rounded notch in the lateral margin of the
occiput (lost on the right side of the occiput) that undoubtedly repre
sents the medial border of the posttemporal fenestra. Fracasso (1983,
1987) misinterpreted the occipital-flanges of the supratemporal and tabu
lar and the squamosal exposure on the occipital surface of the cheek as
parts of a long, ventral extension of the tabular that extended along the
occipital surface of the skull cheek to nearly the ventral margin of the
skull.
The remaining component of the tabular, the ventromedial occipi
tal flange (referred to as the tabular cone by Fracasso, 1983, 1987), is
restricted to the occiput and occupies the medial angle formed between
the other two components of the tabular. It has the form of a thick, flat
plate that broadens slightly as it extends medially to contact the lateral
margin of the supraoccipital region of the occiput and the posterior
surface of the dorsolateral corner of the paroccipital process of the
opisthotic. For most of its extent it exhibits the same roughened, surface
texture as the occiput. The smoothly finished, lip-like structure of the
occipital margin of the postparietal is continued laterally onto the tabular
in a broad, dorsal arch that defines the boundary between the proximal
horn-like component and the ventromedial occipital flange of the tabular.
The lateral portion of the ventral margin of the ventromedial occipital
flange of the tabular forms the dorsal margin of the posttemporal fenes
 197

FIGURE 8. Limnoscelis paludis, holotype, YPM 811. Anterior portion of cranium in right lateral view. A, Illustrated as preserved, and B, emphasizing
preserved limits of elements.
198
tra.
The same three components of the tabular in Limnoscelis paludis
can be recognized, though greatly modified, in Orobates and Diadectes
(Berman, 2000; Berman et al., 2004). In both diadectids the posterolat
eral corner of the sculptured, proximal, horn-like component of the tabu
lar curves abruptly ventrally from the skull table into the plane of the
occiput in the form of a greatly elongated, narrow process. Distally it
ends in a narrow contact with the dorsal margin of a greatly shortened
paroccipital process. A narrow, smoothly finished, flange-like extension
of the medial margin of the proximal, horn-like component, considered
equivalent to the combined distal medial and occipital flange components
of the tabular, contacts the lateral margin of the occiput. It encloses the
posttemporal fenestra laterally and extends a short distance farther ven
trally to contact the dorsal margin of the paroccipital process. In Tseajaia
all that is visible of the tabular is a narrow, dorsally sculptured, thumb
like process, presumably the proximal, horn-like component, which
projects a short distance beyond the distal end of its contact with the
ventral surface of the same component of the supratemporal. Together
they project posteroventrally over the dorsal margin of the occipital
surface of the quadrate.
Palate
Vomer - The vomers are long in Limnoscelis paludis, equaling
almost 30% of the total cranium length, and form the entire medial margin
of the internal nares (Figs. 4A, 11). The approximately anterior half of
the vomer, the premaxillary process, has the form of a vertical,
subrectangular sheet that closely approaches, possibly even contacts,
the ventral surface of the nasal, and is narrowly separated from its mate
along the mid-sagittal plane. Anteriorly they nearly overlap the ventral
margin of the vomerine process of the premaxilla and end anteriorly in a
short, forked margin. Posteriorly the premaxillary processes of the vomers
thicken along their ventral margin as they curve strongly medially to
contact one another. Farther posteriorly the vertical, sheet-like vomer
twists about its long axis as it diverges from the midline to form a poste
rolateral, wing-like extension with the lateral surface now facing ventro
laterally and slightly anteriorly. The posterolateral wing of the vomer
narrowly overlaps the ventral, lateral margin of the wedge-shaped, united
anterior ends of the pterygoids. Posteriorly the vomer ends in a V
shaped, marginal contact with the anterior margin of the palatine. A
short, narrow field of minute denticles, which closely parallels the pos
terior end of the midline union of the premaxillary processes, is contin
ued posteriorly onto the pterygoid. In Tseajaia, Orobates, and Diadectes
a single row of well-developed teeth closely parallels the medial margin
of the premaxillary process.
Fracasso’s (1983, 1987) description of the vomer had several
important errors. Because he was unable to detect the midline palatal
suture between the premaxilla and vomer, he misinterpreted the maxil
lary process of the premaxilla as a lateral, wing-like extension of the
vomer. Furthermore, he incorrectly indicated the vomer-palatine suture
as nearly straight and directed slightly anterolaterally, and positioned
slightly anterior to the posterior margin of the internal naris, which
falsely portrays the palatine as having a short entrance into the medial
margin of the internal naris.
Palatine - In ventral view (Figs. 4A, 9, 11) the palatine has an
anteroposterior elongate, subrectangular outline. A channel-like depres-
sion extends along the lateral half of its ventral surface that deepens as it
extends to the posterior margin of the internal naris. Laterally the channel
is bounded by a thin, ventrally reflected, lip-like flange that contacts the
medial surface of the alveolar shelf of the maxilla. Anteriorly the lateral
flange of the palatine continues for a short distance as it tapers ventrally
to a sharply pointed process. A transverse section of the cranium at the
midlength level of the palatines (Fig. 9) reveals a pronounced, dorsal
thickening of their lateral margin to form a thick, buttress-like contact
with the medial surface of the alveolar shelf. The transverse section also
reveals a vertical thickening of the palatine medial to the channel on its
ventral surface. At its medial margin the palatine bifurcates into thinning,
shelf-like processes between which the lateral margin of the pterygoid is
clasped. The much thinner, dorsal shelf of the bifurcation extends the
farthest medially and, although not visible in the cross-section, is in turn
overlapped dorsally at its anterior and posterior ends by the medial
margins of the vomer and ectopterygoid, respectively. In the anterior,
suborbital region of the skull roof the palatine also contacts the internal
area of juncture of the maxilla, jugal, and lacrimal. A narrow field of
minute denticles extends along the posterior half of the medial margin of
the ventral surface of the palatine.
Ectopterygoid - Although smallest of the palatal elements (Fig.
4A), the greatly thickened structure of the ectopterygoid unites firmly
the palate with the suborbital region of the skull roof. As in the palatine,
it thickens greatly laterally, forming not only a buttress-like contact with
the medial surface of the alveolar shelf of the maxilla, but also with the
jugal. The posterolateral corner of the ectopterygoid has a short, verti
cally rounded edge that narrowly contributes to the anterior margin of
the narrow, notch-like anterior extension of the subtemporal adductor
fossa. The posterior border of the ectopterygoid is thickened ventrally
toward its transverse contact with the transverse flange of the pterygoid,
so as to be slightly convex ventrally in transverse section. The medial
margin of the convexity borders an anteroposteriorly elongated, narrow,
oval basin-like depression that deepens posteriorly as it extends along
the greater posterior length of the ectopterygoid-pterygoid contact. The
depression does not correspond to the suborbital foramen seen in some
primitive amniotes, which is located typically at the juncture between
the palatine, ectopterygoid, and pterygoid. However, at or near the junc
ture of these elements a very similar depression is also present in Orobates
and Diadectes. The ectopterygoid is the only edentulous element of the
palate.
Pterygoid - The pterygoid is divisible into the three standard
components: anterior palatal process, transverse flange, and posterior
quadrate process. Anteriorly the long palatal processes (Figs. 4A, 9, 11)
converge on the midline, wedging a short distance between the postero
lateral wings of the vomers. A field of fine denticles covers the palatal
process except for a narrow margin bordering the interpterygoid vacuity.
In Tseajaia, Orobates, Desmatodon, and Diadectes a single row of large
teeth, supported by a well-developed ridge, extends along the medial
margin of the palatal process. With the exception of the vomer not being
known in Desmatodon, the pterygoid tooth row is smoothly continuous
with that of the premaxillary process of the vomer. A short, rectangular
lateral expansion at the posterior end of the palatal process contacts the
medial margin of the ectopterygoid. The oval depression described above
as extending along this contact is bordered medially by a well-developed,
buttress-like ridge of the pterygoid.
Between the levels of the posterior margin of the transverse flange
and basicranial articulation the pterygoid is constricted to a short, neck
like strut with a dorsoventrally thickened, rounded medial margin (Figs.
4A, 10). On its lateral margin a web-like flange curves anterolaterally a
short distance onto the posterior margin of the transverse flange. The
posterior margin of the interpterygoid vacuity is closed on either side by
a short, broad, posteromedially directed basal process of the pterygoid.
The process has a triangular notch for the reception of the basipterygoid
process of the braincase, which is roofed dorsally by the base of the
epipterygoid to form a freely mobile basicranial articulation. A similar
basicranial articulation has been described in Diadectes and Orobates.
The transverse flange of the pterygoid (Figs. 4A, 10, 14) is dors
oventrally thickened into a massive structure, particularly toward its
posterior, tooth-bearing margin. The bases of a single, transverse row of
eight or nine large teeth dominate the dentition along the posterior margin
of the transverse flange. Their basal diameters vary considerably, but in
general exhibit a pronounced serial increase in size to a mid-serial posi
tion of three teeth of subequal size. The teeth are implanted in deep
sockets and are bordered laterally and anteriorly by a narrow field of

FIGURE 9. Limnoscelis paludis, holotype, YPM 811. Posterior view of anterior portion of cranium at cross-sectional break through cranium at level of
anterior margins of orbits.
small denticles. As in amniotes, the transverse flange projects ventrolat
erally far below the ventral margin of the cranium (Fig. 3B).
The quadrate ramus of the pterygoid (Figs. 4A, 10, 14) extends
posterolaterally from the level of the basicranial articulation to the pos
teromedial edge of the occipital exposure of the quadrate and can be
divided into two distinct components. One forms a thin, tall,
dorsomedially inclined lamina-like structure that ends posteriorly con
tacting the entire medial surface of the dorsal process of the quadrate just
above its medial condyle. In ventral view all that is visible of this compo
nent is a thickly rounded, keeled ventral edge, which medially borders the
subtemporal adductor fossa. The second component of the quadrate
ramus forms a prominent, narrowly arcuate flange that projects medially
and slightly dorsal to the level of the keeled ventral edge of the quadrate
ramus. The arcuate flange, present also in Limnoscelis dynatis, is appar
ently a unique feature of the limnoscelids among the diadectomorphs.
Palatoquadrate Complex
Quadrate - The morphology of the quadrate probably represents
the primitive condition for diadectomorphs and possibly all cotylosaurs
in having a large ventral condyle and a dorsal process that consists of two
vertical sheets or laminae that arise from the condyle. One lamina, re
ferred to here as the occipital lamina (Figs. 4B, 10, 12), extends directly
above the posterior margin of the condyle as a transversely narrow,
subrectangular plate that nearly reaches the entire dorsal height of the
subtemporal adductor chamber. Within the chamber it forms the medial
half of the posterior wall. The greater portion of the occipital lamina is
also exposed on the occipital surface of the chamber, where it occupies
199
essentially a transverse vertical plane that is inclined slightly
dorsomedially. Its surface is slightly convex, so as to be smoothly con
tinuous with the curvature of the posterior margin of the cheek region of
the skull roof, where it is overlapped narrowly by the squamosal. The
occipital lamina in Tseajaia, Orobates, and Diadectes differs only in that
the occipital surface is flat and inclined slightly anterodorsally. The me
dial margin of the occipital lamina curves anteromedially, where it is
continuous with the second, vertical lamina of the dorsal process (Figs.
4A, 7, 10, 12), referred to here as the pterygoid lamina. It extends ante
riorly and slightly medially and has the form of an expansive, subtriangular,
sheet that forms approximately the posterior half of the medial wall of
the subtemporal adductor chamber. Proximally the pterygoid lamina prob
ably reached the height of the occipital lamina, but anteriorly its dorsal
margin tapers ventrally to the ventral margin of the subtemporal adduc
tor chamber and is overlapped medially by approximately the posterior
half of the quadrate ramus of the pterygoid. In amniotes the dorsal
process of the quadrate lacks an occipital lamina, consisting only of the
relatively flat, vertical pterygoid lamina and, therefore, does not contrib
ute to the posterior wall of the adductor chamber.
Although the articular surfaces of both condyles are only partially
exposed due to the jaw being tightly attached to the cranium, their ap
proximate shape can be discerned (Figs. 4, 10, 12, 14). As in amniotes,
the condyle is greatly expanded mediolaterally and divided by a well
developed, anteroposteriorly oriented, saddle-like groove into two dis
tinct condylar facets. The axis of the slightly longer medial facet is ori
ented slightly anteromedially, whereas that of the lateral facet is oriented
strongly anteromedially.
200

FIGURE 10. Limnoscelis paludis, holotype, YPM 811. Posterior portion of cranium in ventral (palatal) view. A, Illustrated as preserved, and B,
emphasizing preserved limits of elements.

Just above the medial margin of the medial condyle the quadrate
and adjacent to the distal end of the quadrate ramus of the pterygoid
there is a deep pocket of unfinished bone (Figs. 4B, 12) that presumably
received the distal end of the stapes, although this element is unknown in
Limnoscelis paludis.
Epipterygoid - Although rarely exposed in tetrapods, both
epipterygoids are present in Limnoscelis paludis. As in amniotes, where
this bone is known, the epipterygoid consists of a large basal plate (Figs.
4A, 7, 10) and a dorsal, narrow, vertical, rod–like process. Only the base
of the right dorsal process is visible in lateral view in any of the figures
(Fig.7). The epipterygoids are preserved in place and essentially com
plete except for the distal halves of both dorsal processes being sepa
rated narrowly by a break and reoriented due to the dorsoventral crusting
of the skull so that the right and left distal portions are redirected poste
riorly and anteriorly, respectively. Despite the jagged margins of the
basal plates, due to small loss of bone, they have an expansive contact
with the dorsolaterally facing surface of the quadrate ramus of the ptery
goid from the level of the basicranial articulation to approximately the
midlength of the ramus. The basal plate is thickest along its ventrolateral
margin in the area of the basicranial articulation, where it roofs the V
shaped notch in the basal process of the pterygoid to provide a ventrally
facing articular facet for the basipterygoid process of the braincase (Figs.
4A, 10). The dorsal process of the epipterygoid arises from the medial
margin of the dorsal surface of the basal plate as a conical structure that
gradually narrows to a slender, rod-like structure. It reaches the ventral
surface of the parietal to provide a structural brace between skull roof
and palate. A small recess on the ventral surface of the parietal at the level
of the posterior orbital margin and lateral to the sphenethmoid marks the
location of its contact with the skull roof (not visible in figures). In
previous interpretations of L. paludis (Romer, 1946; Fracasso, 1983,
1987) the role of the epipterygoid in the basicranial articulation was not
recognized.
Braincase
General - The braincase was described and illustrated as pre
served and reconstructed by Fracasso (1983, 1987). In contrast to the
present study, he removed portions of the dermal skull that obscured the
braincase in dorsal and much of its lateral views. This was not attempted
here in order to avoid further damage to the cranium, and a detailed
description of the braincase in these views is deferred to Fracasso (1983,
1987). The description of the braincase presented here, therefore, is
limited to palatal and occipital views.
Palatal view - The fused basisphenoid and parasphenoid, the
basiparasphenoid complex, is extensively exposed in ventral view of the
braincase (Figs. 4A, 10). As exposed, the narrow cultriform process
extends the entire length of the interpterygoid vacuity, reaching the level
of the posterior border of the internal naris. Its lance-like outline widens
quickly but smoothly to a greatest width in the first third of its posterior
length and then gradually narrows to an acutely angled point. In cross
section at the level of the anterior margins of the orbits it has the form of
a narrow V-shaped trough with a well-developed ventral keel (Fig. 9).
The trough of the posterior portion of the process supports the ventral
margin of the sphenethmoid (Fig. 7). The stout basipterygoid process
(Figs. 4A, 10) is triangular in outline with a transverse anterior margin
and is directed ventrolaterally into the notch of the same outline in the
basal process of the pterygoid. Only the anterior surface of the
basipterygoid process exhibits a well-defined, roughened articular sur
face, which is transversely oval in outline.
Posterior to the basipterygoid processes the smoothly finished
basiparasphenoid expands greatly laterally, with the lateral margins be
ing dominated by narrow, well-developed, rounded ridges of the cristae
ventrolaterales (Figs. 4, 10). The cristae increase in height as they merge
anteromedially to form a deep, concave posteromedial border between
them. The lateral margins of the cristae extend dorsally for a short dis
tance to a narrow zone of unfinished bone that separates them from being
201
continuous with the smoothly finished lateral walls of the prootic region
of the braincase. Posteriorly the cristae end in a thin, ventrolaterally
convex, sleeve-like structure that partially encloses the fenestra ovalis
anteriorly. The thin sheet of bone spanning the depressed area between
the cristae extends onto the anterior margin of the ventral surface of the
basioccipital in a strongly interdigitating suture. A low, midventral ridge
between the cristae is continued posteriorly onto the basioccipital, where
it merges with the raised, rounded, ventral lip of the occipital condyle.
Posterior and immediately dorsal to the level of the distal free margins of
the cristae ventrolaterales the opisthotic portion of the braincase is ex
panded greatly laterally into two distinct structures. Suturally adjoining
the lateral margin of the basioccipital and projecting posteroventrally
and slightly laterally from the base of the paroccipital process of the
opisthotic is a short, gutter-shaped structure originally termed the otic
trough by Fracasso (1983, 1987). Because Fracasso (1987, p. 9) was
unable to detect the suture separating the opisthotic from the basioccipi
tal and exoccipital, he mistakenly described the otic trough as appearing
“to be continuous with the basioccipital, as there is no evidence of
sutures or other separation between the trough bases and the basioccipi
tal.” The otic trough, with its anterolaterally facing, gutter-like channel
forms essentially an extension of the posteromedial border of the fenes
tra ovalis. An otic trough has also been described in Diadectes and
Limnoscelis dynatis, as well as in some primitive basal synapsids (Berman,
2000). Although an otic trough was not described by Berman et al.
(2004) in Orobates, reexamination of one specimen (MNG 8760) has
revealed its presence. An otic trough was also tentatively identified in
Tseajaia (Berman et al., 1992). In both species of Limnoscelis a distinct
cleft separates the otic trough from the base of the paroccipital process.
Occiput - The supraoccipital and opisthotic are indistinguishably
fused into a single complex (Figs. 4B, 12). Romer (1946) and Fracasso
(1983, 1987) misinterpreted symmetrical fractures, most likely due to
the dorsoventral crushing of the skull, as the supraoccipital-opisthotic
sutures extending directly dorsolaterally from the midheight of the lateral
margin of the exoccipital to the lateral margin of the occiput. The
supraoccipital-opisthotic suture was also misidentified by Berman (2000)
and Berman et al. (1992) as extending dorsolaterally a short distance
from the middorsal margin of the foramen magnum that limited the su
praoccipital exposure to rather narrow, dorsolaterally arching bands. The
apparent absence of a supraoccipital-opisthotic suture in the holotype
of Limnoscelis paludis is best attributed to late adult fusion. The suture
is, however, easily traced in the isolated otico-occipital portion of the
braincase of L. dynatis, which lacks the exoccipitals and basioccipital.
Here the suture extends dorsolaterally from about midheight of the lat
eral margin of the foramen magnum in a broadly sigmoid course that ends
at the lateral margin of the occiput, where the occiput would have con
tacted the lower portion of the medial margin of the ventromedial occipi
tal flange of the tabular. Using the occiput of L. dynatis as a guide, the
opisthotic in L. paludis would approximate a shallow depression that
encompasses the ventrolateral region of the occiput and includes the
entire paroccipital process. Its dorsally convex margin ends with the
paroccipital process contacting about the ventral third of the medial
margin of the ventromedial occipital flange of the tabular. On the basis of
the right side of the occiput, the supraoccipital-opisthotic suture would
have ended distally near or just above the medial margin of the small,
rounded notch in the lateral margin of the occiput that is identified here as
the posttemporal fenestra. There is a further deepening of the depression
on the distal portion of the paroccipital process that begins in a step-like,
medially convex arc and appears to extend dorsally beneath the lip-like,
ventral margin of the ventromedial occipital flange of the tabular. The
surface of the supraoccipital portion of the occiput is in contrast rela
tively flat.
In a reexamination of the otico-occipital portion of the braincase
in Limnoscelis dynasties Berman (2000) described a suture that divides
the supraoccipital into lateral halves. It extends from the middorsal mar
gin of the foramen magnum to the dorsal margin of the occiput and is
202

FIGURE 11. Limnoscelis paludis, holotype, YPM 811. Anterior portion of cranium in ventral (palatal) view. A, Illustrated as preserved, and B,
emphasizing preserved limits of elements.

FIGURE 12. Limnoscelis paludis, holotype, YPM 811. Cranium in occipital view. A, Illustrated as preserved, and B, emphasizing preserved limits of
elements.
traceable also on the interior, posterior wall of the cranial cavity. This
was compared to a similar division of the supraoccipital in two
anthracosaurian amphibians. There is no evidence in either species of
Limnoscelis of a synotic bone or sutural evidence of its former presence,
which Fracasso (1987, p. 11) described in L. paludis by as “a small
median wedge-shaped element situated posterodorsally between the otic
capsules…tightly apposed against the overlying anteroventral surface of
the supraoccipital.”
The basioccipital and exoccipitals (Figs. 4, 12) are indistinguish
ably fused into a single complex in Limnoscelis paludis. The exoccipitals
undoubtedly did not consist, as Fracasso (1983, 1987) described, solely
of small, triangular, wing-like elements restricted to the lateral margins of
the foramen magnum. Rather, they almost certainly included, as is stan
dard in tetrapods and illustrated here, a short, ventral, column-like struc
ture, which, along with the dorsal, wing-like process, contacted the pos
terior surface of the supraoccipital-opisthotic in forming the lateral walls
of the foramen magnum. Furthermore, expanded bases of the exoccipitals
contributed undoubtedly to the dorsolateral portions of the occipital
condyle and met or almost met on the floor of the foramen. Fracasso
(1983, 1987) was unable to identify the jugular foramen and mistook an
extremely small pocket of matrix at the ventrolateral margin of the dorsal,
wing-like process of the left exoccipital as the foramen for the hypoglo
ssal nerve XII. Additional preparation has located both these foramina in
their normal positions, with the hypoglossal foramen on the lateral sur
face of the neck of the occipital condyle (not visible in the figures) and
203
the large jugular foramen, presumably for cranial nerves IX-XI and the
jugular vein, on the exoccipital-opisthotic suture at the ventral, lateral
margin of the dorsal, wing-like process of the exoccipital (Figs. 4, 10, 12).
Mandible
Dentary - The dentary occupies a large proportion of the lateral
surface of the jaw, extending over more than two-thirds of its anterior
length (Figs. 3C, 13). Most noticeable, at the symphysis the lateral
surface of the dentary is swollen into a massive, hemispherical structure,
undoubtedly to accommodate the deep, large roots of the anterior three
teeth. On the dorsal surface of the symphysis and medial to the three
anteriormost teeth the dentaries form a large, flat, rugose-surfaced plate
(Figs. 15A, B), which may have provided additional reinforcement to the
symphysis. Its midline suture is irregular, so that the left dentary makes
a significantly greater contribution to the plate. In lateral view the dentary
ends posteriorly in typical cotylosaurian pattern, in an oblique contact
with the angular and surangular as it narrows posterodorsally in a sharp,
wedge-shaped process that rises very slightly as it extends onto the base
of the low coronoid eminence of the surangular. In medial view (Fig. 14)
the dentary is nearly entirely covered by other elements except for a
short anterior portion of the alveolar shelf.
Transverse breaks of the right and left mandibles at the approxi
mate levels of the sixth and tenth tooth positions (Fig. 15C, D), respec
tively, reveal a marked reduction in the transverse thickness of the dentary
posteriorly that is accompanied by a corresponding expansion the
204
205
206
Meckelian canal. Concomitantly, there is a ventral expansion of the
dentary from a level just below the Meckelian canal to the ventral midline
of the mandible that gradually thickens to equal that of its more anterior
cross-section. The ventral expansion, however, is excluded from the
lateral wall of the expanded Meckelian canal by a thin, sheet-like flange
of the angular. As in the maxilla, the lateral margin of the alveolar shelf of
the dentary is noticeably higher than the medial margin, creating a super
ficially pleurodont-like implantation of the teeth, particularly anteriorly
in the series, and the teeth are set in very deep sockets (Fig. 15C, D).
There are places for 23 dentary teeth. The second and third teeth greatly
dominate the series, with the second being slightly reduced. The first
tooth is reduced considerably compared to the second and third, but is
still slightly larger, especially in its diameter, than the largest remaining
teeth of the series, which decrease gradually in size posteriorly. The
teeth are identical in structure to those of the upper marginal series,
which is best exemplified by the five teeth posterior to the canine-like
third and fourth of the left mandible (Fig. 13). Crenulations at the base of
the roots are indicative of a labyrinthine enfolding (Figs. 13, 14-15).
Splenial – In medial view of the mandible (Figs. 3C, 14) the
splenial extends posteriorly from its contact with its mate at the
posteroventral portion of the symphysis to about midlength of the man
dible. In this course the splenial is restricted to a very narrow, short
exposure along the ventrolateral margin of the dentary posterior to its
symphyseal swelling (Figs. 3C, 13). A deep pocket in the posterior
surface of the symphysis is floored by the union of transversely narrow,
shelf-like structures of the splenials (Fig. 15A, B). For a short distance
posterior to the symphyseal region on the medial surface of the mandible
the splenial expands quickly into a tall, flat, vertical plate that contacts
dorsally the ventral half of the medial surface of the alveolar shelf of the
dentary (Fig. 15C). Farther posteriorly the splenial bifurcates around the
anterior end of the medial fenestra (identification of this opening follows
Welles, 1941, rather than Fracasso’s (1983, 1987) identification of an
inframandibular fenestra) into two, narrow, strap-like posterior pro
cesses of very unequal lengths (Figs. 3C, 14). The much longer ventral
process extends nearly the entire length of the medial fenestra, medially
overlapping the angular beneath the fenestra, as it narrows slightly to end
in a bluntly pointed margin. The much shorter dorsal, posterior process
of the splenial extends between the ventral and dorsal margins of the
anterior ends of the anterior coronoid and medial fenestra, respectively,
as it overlaps medially the anterior end of the prearticular.
The cross-sectional breaks of the right and left mandibles at the
levels of the sixth and tenth tooth positions, respectively, reveal a dra
matic restructuring of the splenials (Fig. 15C, D). Anteriorly the cross
sectional area of that portion of the splenial exposed on the ventrolater
ally surface of the mandible has a massive, rectangular structure, which is
excluded from the ventral margin of the Meckelian canal by the angular.
In contrast, that portion of the splenial exposed on the medial surface of
the mandible appears as a vertical flange that bounds the Meckelian canal
as it extends dorsally to a narrow contact at the midheight level of the
medial surface of the alveolar shelf. Farther posteriorly, at the level of the
tenth tooth of the cross-sectional break of the left mandible the thick,
ventrolateral portion of the splenial is replaced by a ventral expansion of
the dentary, resulting in a greatly narrowed contact of the splenial with
the dentary at the ventral midline of the mandible. The vertical, medial
flange of the splenial, on the other hand, has become thickened, reori
ented dorsomedially, greatly shortened dorsally to the ventral margin of
the medial fenestra, and excluded from the Meckelian canal by the angu
lar.
Coronoids - Two coronoids are present, referred to here as the
anterior (c1) and posterior (c2) coronoids, and exposed only in medial
view of the mandible (Figs. 3C, 14). The anterior coronoid and the ante
rior half of the posterior coronoid form a narrow, rectangular strip that
widens slightly as it extends posteriorly in contact with the dorsal half of
the medial surface of the alveolar shelf of the dentary (Fig. 15D). The
ventral margin of the coronoids narrowly contact the medial surface of
the dorsal margin of the prearticular except anteriorly, where the contact
is continued a short distance along the dorsal margin of the splenial. A
shagreen of denticles increases in density and width posteriorly as it
extends across the posterior half of the anterior coronoid and the poste
rior coronoid anterior to the adductor fossa. At the anterior margin of the
adductor fossa a thin, wing-like extension of the posterior coronoid ex
tends along the dorsal margin of the medial surface of the surangular,
nearly reaching the summit of a low coronoid eminence of the surangular.
Contrary to Fracasso’s (1983, p. 194) description, there is no indication
of an expansion of the anterior end of the anterior coronoid to form a
“lance-like configuration” that nearly reaches the mandibular symphy
sis. This was likely confused with the large, flat, rugose-surfaced plate of
bone formed by the dentaries on the dorsal surface of the symphysis.
As in Limnoscelis paludis, Tseajaia possesses two coronoids,
with the posteriormost one contacting the dorsal margin of the medial
surface of the surangular in a low cornonoid eminence at the anterolateral
border of the adductor fossa. In contrast to L. paludis the diadectids
Desmatodon, Orobates, and Diadectes have a single coronoid, which
forms a high, well-developed, prominent coronoid eminence at the ante
rolateral border of the adductor fossa that is visible in lateral and medial
view.
In contrast to Fracasso’s (1983) description of the medial fenestra
as an ovate opening that narrows anteriorly to an acute angle, here it is
interpreted as having a narrow, near-elliptical outline that narrows only
slightly anteriorly. Tseajaia, Orobates, Desmatodon, and Diadectes also
possess a medial fenestra of the same size, general proportions, and
relationship to bordering elements as in Limnoscelis paludis. Otherwise,
in cotylosaurs a medial fenestra is absent with the possible exception of
the ophiacodonts, basal synapsids Ophiacodon and Varanosaurus, where
a very small, possibly equivalent structure is present at the angular
prearticular suture near its anterior juncture with the splenial.
Prearticular - The prearticular is visible only in medial view of
the mandible (Figs. 3C, 14). For most of its anterior length it has basi
cally the outline of a vertically narrow, anteroposteriorly elongated plate
that extends from the posterior margin of the mandible to nearly the level
of the anterior margin of the medial fenestra. In this course the prearticular
can be divided into three regions of anteriorly decreasing vertical widths.
The prearticular is most greatly expanded posterior to the adductor
fossa, where it contacts the medial surface of the articular a short dis
tance below the glenoid facet and ventrally overlaps the medial surface of
the dorsal margin of the angular. At its posterior margin the prearticular
joins with the articular in forming the ventral, triangular portion of a very
large, vertical, anteriorly inflected, shelf-like structure of the pterygoideus
or medial process for presumably the attachment of the pterygoideus
musculature. The posterior margin of the process is angled anteroventrally,
so as to intersect the vertical, anterior margin ventrally in a blunt point.
At its posterior margin the process is massive, but is quickly reduced to
a thin edge along its anterior margin, where it is separated from the medial
face of the mandible by about 3 cm, forming a deep, anteriorly facing,
smoothly finished, convexity. The free anterior edge of the process is
roughly finished. A low ridge near the ventral margin of the medial surface
of the expanded, posterior, plate-like portion of the prearticular extends
across its contribution to the inside curvature of the pterygoideus pro
cess to its ventral tip. In Tseajaia there is a low but prominent, medially
projecting ridge at the posteromedial corner of the prearticular that ex
tends anteroventrally on the posterior portion of the prearticular’s con
tact with the medial surface of the articular and adjacent to the posterior
rim of the adductor fossa. It is apparently a greatly reduced equivalent of
the pterygoideus process in Limnoscelis paludis, presumably serving
the same function, whereas Desmatodon, Orobates, and Diadectes lack
any indication of the process.
The mid-region of the prearticular is marked by a posterior, step
like narrowing of its dorsal and ventral margins that extends the length of
the adductor fossa, as it forms the upper half of the lateral wall of the
adductor chamber and ventrally overlaps the medial surface of the dorsal
 207

FIGURE 15. Limnoscelis paludis, holotype, YPM 811. Symphyseal portion of lower jaw in dorsal view. A, Illustrated as preserved, and B, emphasizing
preserved limits of elements. Cross-sectional, posterior views of C, right mandible at level of sixth tooth, and D, left mandible at level of tenth tooth.
208
margin of the angular. A further narrowing of the prearticular anterior to
the adductor fossa defines its anterior region. It borders the dorsal margin
of the medial fenestra except for the short distance at its anterior end,
where it is overlapped medially by the splenial. At the cross-sectional
break of the left mandible at the level of the tenth dentary tooth position
(Fig. 15D) the medially exposed prearticular contacts the ventral half of
the medial surface of the alveolar shelf of the dentary. However, at the
cross-sectional break of the right mandible at the level of the sixth dentary
tooth position (Fig. 15C) the depth of the prearticular is greatly reduced.
Here it narrowly contacts on the ventral half of the medial margin of the
alveolar shelf and is overlapped medially by the splenial.
Angular - The angular is extensively exposed on the lateral and
medial surfaces of the posterior half of the mandible (Figs. 3C, 13-14).
Here it forms the lateral and medial walls of the ventral half of the
adductor chamber, which converge ventrally to the broadly convex,
posteroventral angle of the mandible in a sharply angled, V-shaped, keel
like structure. The keeled edge extends to within a short distance from
the posterior end of the mandible. Here only the laterally exposed por
tion of the angular continues posteriorly, narrowing to an acute angle at
the posterodorsal corner of the mandible as its posterior margin curves
medially to contact the lateral margin of the posterior surface of the
pterygoideus process. Almost the entire dorsal margin of the laterally
exposed portion of the angular contacts the surangular in a broadly con
vex suture. Anteriorly the angular has a sinuous contact with the dentary
that quickly angles anteroventrally to the ventral margin of the mandible,
where it continues a short distance as a narrow, splint-like exposure
between the ventral margins of the dentary and splenial. The lateral
surface of the angular bears a fairly rugose sculpturing that extends to the
ventral edge of the mandible.
In medial view of the mandible the angular extends anteriorly to
the posteroventral margin of the medial fenestra. Here its medial surface
is overlapped by the distal end of the ventral, posterior process of the
splenial, limiting the exposure of the angular to two, short, splint-like
processes: one extending along the ventral margin of the fenestra and the
other, as described above, along the ventral margin of the mandible.
Farther anteriorly the angular looses its superficial exposure as it contin
ues within the Meckelian canal to at least a short distance beyond the
level of the anterior margin of medial fenestra. The cross-sectional break
of the left mandible at the level of the tenth marginal tooth (Fig. 15D)
exposes an extensive, well-developed angular that is divisible into two
portions. A narrow, rectangular portion is slightly imbedded into the
internal surface of the splenial. It has a broad, midsagittal contact with
the dentary that is separated from the ventral margin of the mandible by
a narrow dentary-splenial contact. The second portion of the angular
extends dorsally from the ventromedial surface of the internal surface of
the first portion as a thin, flange-like extension that sheaths the ventral
half of the internal surface of the dentary. Thus, at this level of the
mandible the angular bounds the ventrolateral and the ventral, medial half
of the Meckelian canal. At the cross-sectional break of the right mandible
at the level of the sixth marginal tooth (Fig. 15C) all that remains of the
angular is a greatly reduced, rectangular portion, which floors the greatly
constricted Meckelian canal and retains much reduced contacts with the
splenial and dentary.
Surangular - Exposures of the surangulars on the medial and
lateral surfaces (Figs. 3C, 13-14) of the mandibles are partially hidden by
the close apposition of the mandibles to the cranium. The surangular
forms the lateral wall of the adductor fossa, thickening dorsally to a
smoothly rounded, free dorsal margin that is elevated slightly into a
broadly convex coronoid eminence. In lateral view the surangular has
roughly a diamond outline, with the anterior portion wedging between
the dentary and angular, whereas posteriorly it narrows into a rather
attenuated process that curves slightly medially as it contacts the articu
lar.
Articular - The articular (Figs. 3C, 14) forms the posterior corner
of the mandible, and its concave anterior surface forms the posterior wall
of the adductor fossa. It is most extensively exposed in medial view of
the mandible (Fig. 14), where its posterior margin forms the upper por
tion of the large, anteriorly inflected, flange-like pterygoideus process.
Projecting posteriorly from the upper margin of the posterior surface of
the pterygoideous process is a short, rounded, rugose knob that is prob
ably equivalent to a retroarticular process. An anterior process extends
from the anteromedial edge of the articular body. It has the form of a
vertical, transversely thin flange that is partially exposed above the ex
panded posterior portion of the prearticular that contacts its medial
surface. Farther anteriorly it is visible only in dorsal view of the man
dible, as it sheaths the lateral surface of the prearticular to approximately
the midlength of the medial wall of the adductor fossa. In Desmatodon
the anterior process of the articular extends along the lateral surface of
the prearticular to at least level of the posterior third of the medial
foramen, as indicated by its very narrow exposure along the posterior
dorsal border of the foramen, whereas in Diadectes the process extends
to the level of the anterior end of the foramen.
Unfortunately, the articular glenoid surface of both articulars are
mostly hidden from view by their tight articulation with the quadrate
condyles (Fig. 14), but it is possible to discern that the glenoid is divided
into two, anteroposteriorly elongated concave surfaces that are sepa
rated by a ridge of approximately the same orientation.
SUMMARY
The above description, in addition to expanding on the most re
cent detailed descriptions of the holotypic skull of Limnoscelis paludis
presented by Fracasso (1983, 1987) since that of Romer’s (1946), also
recognizes the following major errors:
1) the dorsal processes of the premaxillae are incorrectly described
as short and wide and as enclosing an internasal bone, rather than as long,
narrow, united processes that extend posterodorsally along the midline
in the absence of an internasal;
2) failed to recognize presence of palatal vomerine processes of
the premaxillae, which extend posteriorly as long, narrow processes on
either sided of the midline of the anterior portion of the palate;
3) misinterpreted the jugal as lacking a the dorsal orbital process,
which is present as a long, narrow process that probably formed nearly
the ventral half of the posterior orbital margin;
4) eave-like structural contact of the squamosal with the postor
bital and supratemporal was misinterpreted as a “loose junction” be
tween the cheek and skull table, rather than as a complex, sutural contact
that was surely immobile;
5) misinterpreted the occipital-flanges of the supratemporal and
tabular and the exposure of the squamosal on the occipital surface of the
cheek as parts of a long, ventral extension of the tabular occipital surface
of the skull cheek that nearly reached its ventral margin;
6) failed to recognize the midline palatal suture between the vomer
and premaxilla and misinterpreted maxillary process of the premaxilla as
a lateral, wing-like extension of the vomer;
7) misinterpreted the vomer-palatine suture as nearly straight,
directed slightly anterolaterally, and positioned slightly anterior to the
posterior margin of the internal naris and, therefore, falsely portrayed
the palatine as having a short entrance into the medial margin of the
internal naris. Rather, the posterior margin of the vomer has a V-shaped,
marginal contact with the anterior margin of the palatine at a level slightly
posterior to the posterior margin of the internal naris and, therefore, the
palatine does not enter the internal naris;
8) did not recognize the moderately deep, anteroposteriorly elon
gate, narrow, basin-like depression on the ventral surface of the palate
that straddles the contact of the medial margin of the ectopterygoid with
the pterygoid;
9) failed to recognize the participation of the epipterygoid in the
basicranial articulation, which dorsally roofs a V-shaped notch in the
basal process of the pterygoid to provide a ventrally facing articular facet
for the basipterygoid process of the braincase;

10) failed to identify the presence of the suture marking the con
tact between the opisthotic and the basioccipital-exoccipital complex
and mistakenly described the otic trough as a structure of the basioccipi
tal rather than the opisthotic;
11) misinterpreted paired symmetrical fractures on the posterior
surface of the occiput (most likely due to the dorsoventral crushing of
the skull) as the supraoccipital-opisthotic sutures, which are absent due
to probable late adult fusion;
12) mistakenly identified the presence a synotic bone of the brain
case for which there is also no sutural evidence of its former presence;
13) misidentifed the exoccipitals as consisting solely of small,
triangular, wing-like element restricted to the lateral margin of the fora
men magnum. They almost certainly included also the presence of a
short, ventral, column-like structure that extended between the dorsal,
wing-like portion and an expanded base that contributed to the dorsolat
eral portions of the occipital condyle, with the paired elements forming
the lateral walls and at least the greater portion of the floor of the foramen
magnum;
14) did not recognize the presence of the jugular foramen and
misidentified the location of the foramen for the hypoglossal nerve XII.
Both foramina are present in their standard locations, with the former on
the exoccipital-opisthotic contact at the ventral, lateral margin of the
REFERENCES
Berman, D.S, 2000, Origin and early evolution of the amniote occiput:
Journal of Paleontology, v. 74, p. 938-956.
Berman, D. S, and Sumida, S.S., 1990, A new species of Limnoscelis (Am
phibia, Diadectomorpha) from the Late Pennsylvanian Sangre de Cristo
Formation of central Colorado: Annals of Carnegie Museum, v. 59, p.
303-341.
Berman, D.S, Sumida, S.S., and Lombard, R.E., 1992, Reinterpretation of
the temporal and occipital regions in Diadectes and the relationships of
diadectomorphs: Journal of Paleontology, v. 66, p. 481-499.
Berman, D.S, and Sumida, S.S., 1995, New cranial material of the rare
diadectid Desmatodon hesperis (Diadectomorpha) from the Late Penn
sylvanian of central Colorado: Annals of Carnegie Museum, v. 64, p.
315-336.
Berman, D.S, Reisz, R.R., Bolt, J.R., and Scott, D., 1995, The cranial
anatomy and relationships of the synapsid Varanosaurus (Eupelycosauria:
Ophiacodontidae) from the Early Permian of Texas and Oklahoma:
Annals of Carnegie Museum, v. 64, p. 99-133.
(Diadectomorpha:
Berman, Diadectidae)
D.S, Sumida, from Martens,
S.S., and the Early Permian of central
T., 1998, Ger
Diadectes
many, with description of a new species: Annals of Carnegie Museum, v.
67, p. 53-93.
Berman, D.S, Henrici, A.C., Kissel, R.A., Sumida, S.S., and Martens, T.,
2004, A new diadectid (Diadectomorpha), Orobates pabsti, from the
Early Permian of central Germany: Bulletin of Carnegie Museum of
Natural History, v. 35, p. 1-36.
Fracasso, M.A., 1980, Age of the Permo-Carboniferous Cutler Formation
vertebrate fauna from El Cobre Canyon, New Mexico: Journal of Pale
ontology, v. 41, p. 1256-1261.
Fracasso, M.A., 1983, Cranial osteology, functional morphology, system
atics of Limnoscelis paludis Williston [Ph.D. dissertation]: Yale Uni
versity, New Haven, Connecticut, 624 p.
Fracasso, M.A., 1987, Braincase of Limnoscelis paludis Williston: Postilla,
no. 201, p. 1-22.
Kemp, T.S., 1982, Mammal-like reptiles and the origin of mammals: Aca
demic Press, London, 363 p.
Kissel, R.A., and Reisz, R.R., 2004, Ambedus pusillus, new genus, new
species, a small diadectid (Tetrapoda: Diadectomorpha) from the Lower
Permian of Ohio, with a consideration of diadectomorph phylogeny:
Annals of Carnegie Museum, v. 73, p. 197-212.
209
dorsal, wing-like portion of the exoccipital and the latter on the lateral
surface of the neck of the occipital condyle;
15) misidentified the presence of a “lance-like configuration” of
the anterior end of the anterior coronoid as nearly reaching the mandibu
lar symphysis, which likely was confused with a single, large, flat, rug
ose-surfaced plate of bone formed by dentaries on the dorsal surface of
the symphysis;
16) misinterpreted the medial fenestra of the lower jaw as having
an ovate outline that narrows anteriorly to an acute angle, rather than a
narrow, near-elliptical outline that narrows only slightly anteriorly.
ACKNOWLEDGMENTS
The authors wish to express their thanks to Yale Peabody Mu
seum for the extended loan of the holotype of Limnoscelis paludis. We
are also greatly indebted to Mark A. Klingler of the Carnegie Museum of
Natural History for his exacting, full skeletal reconstructions and
whole-body restoration of L. paludis and expertise in composing and
formatting the figures. Special gratitude is extended to the reviewers, S. S.
Sumida (California State University, San Bernardino) and Robert Sullivan
(The State Museum of Pennsylvania, Harrisburg), whose careful, con
structive comments and criticisms had considerable impact on the im
provement of this paper.
Laurin, M., and Reisz, R.R., 1995, A reevaluation of the amniote phylog
eny: Zoological Journal of the Linnean Society, v.113, p. 165-223.
Laurin, M., and Reisz, R.R., 1997, A new perspective on tetrapod phylog
eny, in Sumida, S.S., and Martin, K.L.M., eds., Amniote Origins, Com
pleting the Transition to Land: Academic Press, San Diego, p. 9-59.
Lee, M.Y.S., and Spencer, P.S., 1997, Crown-clades, key characters and
taxonomic stability: when is an Amniote not an Amniote? in Sumida,
S.S., and Martin, K.L.M., eds., Amniote Origins, Completing the Tran
sition to Land: Academic Press, San Diego, p. 61-84,
Moss, J.L., 1972, The morphology and phylogenetic relationships of the
Lower Permian tetrapod Tseajaia campi Vaughn (Amphibia:
Seymouriamorpha). University of California Publications in Geological
Sciences, v. 98, p. 1-72.
Olson, E.C., 1947, The family Diadectidae and its bearing on the classifica
tion of the reptiles: Fieldiana, Geology, v. 7, p. 2-53.
Panchen, A.L., 1972, The interrelationships of the earliest tetrapods, in
Joysey, K.A., and Kemp, T.S., eds., Studies in Vertebrate Evolution:
Winchester Press, New York, p. 65-87,
Reisz, R.R., 2007, The cranial anatomy of basal diadectomorphs and the
origin of amniotes, in Anderson, J.S., and Sues, H-D., eds., Major Tran
sitions in Vertebrate Evolution: Indiana University Press, Bloomington
and Indianapolis, p. 228-252.
Romer, A.S., 1946, The primitive reptile Limnoscelis restudied: American
Journal of Science, v. 244, p. 149-188.
Romer, A.S., and Price, L. I., 1940, Review of the Pelycosauria: Geological
Society of America Special Paper 28, 538 p.
Vaughn, P.P., 1964, Vertebrates from the Organ Rock Shale of the Cutler
Group, Permian of Monument Valley and vicinity, Utah and Arizona:
Journal of Paleontology, v, 38, p. 567-583.
Vaughn, P.P., 1969, Upper Pennsylvanian vertebrates from the Sangre de
Cristo Formation of central Colorado: Contributions in Science, Los
Angeles County Museum of Natural History, v. 164, 1-28 pp.
Vaughn, P.P., 1972, More vertebrates, including a microsaur, from the
Upper Pennsylvanian of central Colorado: Contributions in Science,
Los Angeles County Museum of Natural History, v. 223, p. 1-30.
Welles, S., 1941, The mandible of a diadectid cotylosaur: University of
California Publications in Geological Sciences, v. 25, p. 423-432.
Wideman, N.K., 2002, The postcranial skeleton of the late Paleozoic
Family Limnoscelidae and its significance for diadectomorph taxonomy:
210
Journal of Paleontology, v. 22, p. 119A.
Wideman, N.K., Sumida, S.S., and O’Neil, M., 2005, A reassessment of the
taxonomic status of the materials assigned to the Early Permian tetra
pod genera Limnosceloides and Limnoscelops, in, Lucas, S.G. and Zeigler,
K.E., eds., The Nonmarine Permian: New Mexico Museum of Natural
History and Science Bulletin, 30, p. 358-362.
Wible, J.R., Miao, D., and Hopson, J.A., 1990, The septomaxillae of fossil
and Recent synapsids and the problem of the septomaxillae of
monotremes and armadillos: Zoological Journal of the Linnean Society,
v. 98, p. 203-228.
Williston, S.W., 1911a, A new family of reptiles from the Permian of New
Mexico: American Journal of Science, v. 31, p. 378-398.
Williston, S.W., 1911b, American Permian Vertebrates: Chicago University
Press, 145 p.
Williston, S.W., 1912, Restoration of Limnoscelis, a cotylosaur reptile
from New Mexico: American Journal of Science, v. 34, p. 457-468.
 Lucas et al., eds., 2010, Carb-Permian transition in Cañon del Cobre. New Mexico Museum of Natural History and Science Bulletin 49.

211
REDESCRIPTION OF THE POSTCRANIAL SKELETON OF LIMNOSCELIS
PALUDIS WILLISTON (DIADECTOMORPHA: LIMNOSCELIDAE) FROM
THE UPPER PENNSYLVANIAN OFELCOBRE CANYON, NORTHERN NEWMEXICO

NATALIAK. KENNEDY
Department of Ecology and Evolutionary Biology, University of California Los Angeles, 621 East
Charles Young Drive, Los Angeles, California 90024, U.S.A.; natalia.k.kennedy@ucla.edu

Abstract—The Late Pennsylvanian diadectomorph tetrapod Limnoscelis is known from two species, the most
complete of which is L. paludis from El Cobre Canyon, Rio Arriba County, New Mexico. The most complete
specimen of L. paludis is Yale Peabody Museum specimen 811, which is redescribed here. Distinctive features of
the axial skeleton include an atlas-axis complex demonstrating the characteristic anteriorly-directed processes of
the atlantal and axial intercentra, and alternation of the neural spine height in certain regions of presacral vertebrae.
There is a single coracoid plate, and the scapula is robust; however, the earlier suggestion of the presence of a
suprascapular cartilage is questionable. The humerus is extremely robust with heavily constructed condyles and
muscular attachment processes. The radius and ulna are relatively short, measuring approximately 60 and 85% the
length of the humerus, respectively. The ilium demonstrates the well-developed lateral iliac shelf characteristic of
diadectomorphs, and the anteroventral margin of the acetabulum extends anteriorly in a narrow channel that
reaches toward the anterior border of the pubis. The femur is notable in its prominently developed rugosities for
the puboischiofemoralis internus and ischiotrochantericus muscles. The tibia and fibula range from 77 to 80% of
the length of the femur. Both manus and pes are incomplete in this specimen of Limnoscelis.

INTRODUCTION enemies (Williston, 1912). Romer (1946) redescribed Limnoscelis, con

The Late Pennsylvanian Limoscelidae is a pivotal taxon in under
standing the origin and early evolution of amniotes (e.g., Williston, 1911a,b,
1912; Romer, 1966; Berman et al., 1992; Sumida et al., 1992; Sumida,
1997). It is part of the more inclusive tetrapod grouping, the Diadecto
morpha, which is generally considered to be the sister taxon to Amniota,
as traditionally defined (Heaton, 1980), but has also been suggested to be
a basal member of Amniota (Berman et al., 1992; Lee and Spencer, 1997).
The Diadectomorpha is composed of three families: Diadectidae,
Tseajaiidae and Limnoscelidae. The latter two families have a strictly
North American record, whereas the family Diadectidae is found both in
North America and central Europe (Berman et al., 1997). Significantly,
the limnoscelids have been suggested as basal members of the grouping,
making them a reasonable model for the skeletal structure characteristic
of extremely primitive amniotes.
Although Limnoscelis is vital to the studies of early amniote ori
gins and evolution, the greater portion of the limnoscelid postcranial
skeleton has never been fully described or illustrated. Previous studies
have focused on the skull (Fracasso, 1983, 1987; Berman et al., 1992;
Berman, 2000), the atlas-axis complex (Sumida et al., 1992), and the
vertebral column (Sumida, 1990). Because the Middle Pennsylvanian
“limnoscelid” Limnostygis has been proposed as a combination of
ophiacodontid and captorhinid specimens (Wideman and Sumida, 2004),
and Limnoscelis is the only well-defined member of the Limnoscelidae
(Wideman et al., 2005), L. paludis is included in this study, and com
pared to L. dynatis when differences are present.
Limnoscelis paludis was first described by Williston (1911a, b)
based mainly on two specimens: the nearly complete, articulated skel
eton of L. paludis (the holotype, YPM 811) and a nearly complete
skeleton with some articulated portions (MCZ 1948, formerly YPM
809). Unfortunately, he illustrated only representative parts of the skel
eton instead of the skeleton as a whole, and included only a brief descrip
tion. At the time, Williston classified Limnoscelidae in the Class Reptilia.
The following year, Williston (1912) expanded his study of L. paludis,
reconstructing it as a slow, crawling reptile with a long body and long tail
that hunted slow-moving prey. He suggested that L. paludis probably
lived a semi-terrestrial lifestyle in marshes and hid in the water from its
cluding that it was indeed a reptile, but that its anatomy was so general
ized that it could be “regarded as representing the common stem of all
lines of reptilian descent.”
Limnoscelis dynatis was described from the Upper Pennsylva
nian of Fremont County in central Colorado (Berman and Sumida, 1990).
It was distinguished from L. paludis by a few differences in cranial
elements, and Berman and Sumida (1990) suggested that L. paludis is
more derived than L. dynatis. Although Berman and Sumida assigned L.
dynatis to Diadectomorpha, they placed it in the Class Amphibia.
Here the holotypic postcranial skeleton of Limnoscelis paludis
(YPM 811) is fully described for the first time, with dorsal and ventral
illustrations of the skeleton.
Institutional Abbreviations: CM, Carnegie Museum of Natural
History, Pittsburgh, Pennsylvania; FMNH, Field Museum of Natural
History, Chicago, Illinois; MCZ, Museum of Comparative Zoology,
Harvard University, Cambridge, Massachusetts; YPM , Yale Peabody
Museum, New Haven, Connecticut.
Anatomical Abbreviations: cor, coracoid; delt, deltoid process;
ect, ectepicondyle; ent, entepicondyle; ent for, entepicondylar fora
men; f, femur; fi, fibula; gl for, glenoid foramen; h, humerus; icl, inter
clavicle; int, intermedium; mc, medial centrale; ole, olecranon process;
r,radius; rade, radiale; sup, supinator process; t, tibia; u, ulna; ule,
ulnare; I-V, digits 1-5.
MATERIALS AND METHODS
The description of the postcranial skeleton of Limnoscelis is pri
marily based on the almost complete articulated holotype of L. paludis
(YPM 811). Two other specimens of Limnoscelis exist (MCZ 1947 and
1948, formerly YPM 819 and 809, respectively). The pes of MCZ 1948
was used to supplement this description of Limnoscelis, because it is not
complete in the holotype. I could not study the remainder of MCZ 1948
or MCZ 1947, and would like to do so in the future. These materials
were compared to the disarticulated postcranial materials of L. dynatis
(CM 47653), and differences were noted.
Length, height and width measurements were made on all available
skeletal elements. The processes and depressions were measured, as far
212
as possible, at a right angle to the sagittal and dorsoventral axes. Speci
mens were measured in as many ways as possible as determined by the
element and its preservation. In addition, photographs were also taken.
The holotypic postcranial skeleton of Limnoscelis paludis is well
preserved and nearly complete; however, it is partially encased in plaster
in the dorsoventral midsection, making access to some of the dorsal and
ventral aspects of the specimen difficult, and sometimes impossible
(Figs. 1,4,5). Much of the skeleton is visible in dorsal view, with the
exception of most of the ribs, portions of the pectoral and pelvic girdles
and the vertebrae. The ventral anterior portion has been prepared, expos
ing the interclavicle and the ventral surfaces of some of the vertebrae and
ribs and a portion of the ventral surface of the pelvis are visible. Unfor
tunately, at some point in time, the plaster supporting the holotype of L.
paludis was cracked, and this crack was repaired with plaster (Fig. 4),
obscuring some of the ventral skeleton from view. Furthermore, none of
the ventral surfaces of the limb bones are visible in YPM 811. For the
sake of completeness, the only other valid limnoscelid, L. dynatis, was
used for comparison.
Specimen and locality data for the specimens examined include:
YPM 811, holotype of Limnoscelis paludis, complete skeleton; Upper
Pennsylvanian El Cobre Canyon Formation, Cutler Group, El Cobre
Canyon, Rio Arriba County, northern New Mexico. Collected by David
Baldwin. CM 47653, holotype of L. dynatis, partial associated skeleton
of a single individual including premaxilla, maxilla, partial occiput, atlantal
neural arch, dorsal vertebrae, ribs, partial pectoral and pelvic limbs. Up
per Pennsylvanian Sangre de Cristo Formation, Fremont County, central
Colorado. Collected by Peter P.Vaughn.
AXIAL SKELETON
Atlas-axis Complex: The atlas-axis complex of Limnoscelis
paludis was first described by Williston (1911a, b). However, due to
incomplete preparation of the specimen (YPM 811), only the dorsal
aspect of the complex was described. Sumida (1990) redescribed the
atlas-axis complex in more detail; nonetheless, only the dorsal aspect
was described. This complex was finally fully prepared and described in
complete detail by Sumida et al. (1992). Sumida et al. (1992) also de
scribed the atlas-axis complex of L. dynatis, which, due to its fragmen
tary nature, provided additional insight only into the medial side of the
left neural arch.
Re-examination of materials ascribed to Limnoscelis suggests no
significant emendations to the work of Sumida et al. (1992) are neces
sary, so only a brief summary is presented here. Both the atlas and axis
are multipartite elements in L. paludis (Sumida, 1992). The axial
intercentrum is fused to the ventral aspect of the axial pleurocentrum,
FIGURE 1. Limnoscelis paludis, holotype, YPM 811, in dorsal view.
precluding ventral exposure of that element in the axial column. Most
notably, both the atlantal and axial intercentra bear ventromedial pro
cesses that are directed anteriorly into furrows of the exoccipital and
atlantal intercentrum, respectively. There is little or no development of
an atlantal neural spine from the paired atlantal neural arches; however,
the axial neural arch is robust and very well developed. A proatlas was
present.
Dorsal Vertebrae and Ribs: Williston (1911a, b, 1912) first
described the vertebrae of Limnoscelis paludis (YPM 811) as being rather
uniform, with swollen neural arches and neural spines all of about the
same length. Additional descriptions of L. paludis and the disarticulated
vertebrae of L. dynatis (CM47653; Berman and Sumida, 1990; Sumida,
1990; Sumida et al., 1992) showed that the size and shape of the verte
brae of limnoscelids vary significantly throughout the column.
Limnoscelis paludis (YPM 811) has 26 presacral vertebrae. The
third to seventh dorsal vertebrae occupy the level between the scapular
blades of the pectoral girdle (Fig. 2). The centra are amphicoelous and
notochordal. The centra of the most anterior vertebrae are small in diam
eter compared to more posterior vertebrae and are approximately 25%
longer than they are wide (Sumida, 1990). The planes of the anterior and
posterior zygapophyses slope anteromedially at about 30° and 25°,
respectively. The neural spines vary in height between tall and short
spines in a random pattern. The neural spines in presacral vertebrae 3-7
are the least expanded of the presacral column, but have a greater ante
rior-posterior measure. The third and fifth neural spines are well-devel
oped with bases about 9 mm in transverse diameter. Usually, a fore and
aft midline ridge is present on the ends of these tall spines. On the other
hand, the fourth neural spine is narrow and low with a longitudinal ridge.
Furrows on either side of the neural spine may have facilitated passage of
interspinus musculature (Sumida, 1990, 1997). Isolated anterior verte
brae in L. dynatis confirm this pattern (Fig. 3). The neural spine on the
sixth vertebra of the holotype is tall, measuring approximately 16.7 mm
in height and 11 mm in transverse basal diameter. The seventh and tenth
neural spines are also tall. Even though the height alteration pattern does
not have a precise arrangement and is not identical in all specimens, it is
clear that some type of alteration of height and structure does occur in
every specimen in the anterior dorsal vertebrae.
Stout, spatulate ribs are visibly associated with dorsal vertebrae
three through seven (Fig. 2). The ribs on the right side of YPM 811 are
directed dorsolaterally, with some vertebrae abutting the scapula, whereas
the ribs on the left side are directed ventrally. The anteriormost ribs are
approximately 17 cm long and increase serially in length posteriorly at
the level of the pectoral girdle. The ribs have a single head, but with
subequally developed areas for the capitulum and tubercle. Tubercular

FIGURE 2. Limnoscelis paludis, holotype, YPM 811, vertebrae 3-7 and
associated ribs. Cranial end of column is toward top of page.
and capitular facets are visible in anterodorsal and posteroventral views
of the rib, respectively. As preserved, all ribs in limnoscelids are single
headed. Romer (1946) hypothesized that the proximal ends of the ribs
had cartilaginous caps to allow the movement or exit of the vertebral
artery between the capitulum and tubercle. The shaft of the ribs is nar
row and roughly oval in cross-section. A low, triangular protrusion is
visible on the posteroventral margin of the shaft just distal to its narrow
est point. The distal ends are very thin dorsoventrally and spatulate in
shape.
The more posterior vertebrae of the presacral series differ from
those at the anterior end of the series. The centra are significantly longer
than they are wide and they have beveled anterior and posterior edges
ventrally, presumably for reception of the dorsal intercentra. A midventral
depression is formed by the edges of these anterior and posterior ridges
by the middle of the column. The zygapophyseal planes of the middor
sal vertebrae slope medially about 15° without a posterior component as
in the anterior vertebrae. The neural spines of vertebrae 12, 14, 16, 17,
and 19 are tall, in a manner similar to those described in the more anterior
vertebrae, whereas vertebrae 13 and 15 have virtually no neural spine,
but only a low longitudinal ridge. Although the spine of vertebra 18 is
tall, it is only about 5 mm in transverse width and much narrower than
those of the anterior vertebrae. Clearly, the pattern of neural spine height
alteration continues in the middorsal vertebrae. Isolated dorsal vertebrae
in Limnoscelis dynatis also exhibit the high- and low-spined morphology
(Fig. 3).
Most of the ribs in the middle and posterior of the column are not
visible in dorsal view in Limnoscelis paludis (YPM 811); only the left rib
of vertebra 10, and partial right ribs of vertebrae 10, 11, 12 and 13, are
visible. They are much narrower and longer than the pectoral ribs and are
sharply recurved. In cross-section, the size and shape of the rib shaft
remain constant throughout the rib’s length. In ventral view, the ribs of
vertebrae 9-16 are visible (Fig. 4). These ribs extend in natural position
from the vertebrae to about their midpoint. At the midpoint, the ribs
collapse on themselves ventrally and their distal ends point
posteromedially, most likely indicating postmortem crushing of these
ribs. The ribs are slender in ventral view, but with their distal ends
slightly dilated. A much shorter rib (5.5 cm) lies close to the ventral
surface of vertebra 17. This rib ventrally overlies another short rib that is
mostly hidden by matrix. This rib is visible in lateral view and has a
213
broadened proximal end with a long, narrow shaft.
The most posterior dorsal vertebrae exhibit serial changes in the
structure of the centra, pedicels, zygapophyses, transverse processes,
and neural spines. The centra continue to be longer than they are wide.
The posterior zygapophyses slope ventromedially 15 to 20 degrees,
whereas those of the anterior zygapophyses are horizontal. The trans
verse processes and neural arch pedicels are proportionately shorter
than those in the anterior portion of the column. Vertebrae 17 to 26 have
tall neural spines except the 22nd, which appears to have had a low,
ridge-like spine, which, though not preserved is evidenced by a long,
slender break scar. In some of the vertebrae coarse projections extend
from the lateral surface of the ventral portion of the neural spine. The tip
of the last neural spine is strongly bifid. All the presacral vertebrae up to
23 have been described as having accompanying ribs (Williston, 1912).
This description appears to be based on the holotype, but may have also
included supplementary information from MCZ 1947 and 1948. Cur
rently, only the left rib of vertebra 18 is present in the holotype, and the
ribs of CM 47653 are disarticulated, so they cannot be assigned to a
particular vertebra. The 18th left rib of L. paludis is shaped similarly to
the preceding ribs except for being about 50% shorter and less recurved.
Sacral Vertebrae and Ribs: Differing interpretations of the sac
ral region of Limnoscelis paludis have been presented. Williston (1911a,b,
1912) described a single sacral vertebra in the holotype with a following
large caudal vertebra, whereas Romer (1946) suggested that Limnoscelis
was a transitional model between “…a one-ribbed and two-ribbed condi
tion.” Sumida (1990) described two sacral vertebrae and ribs. Examina
tion of YPM 811, the only specimen with an intact pelvic girdle, sug
gests that L. paludis likely had two sacral vertebrae. Although the neural
spines have been broken off and the vertebrae are not visible in ventral
view, the first sacral vertebra is obvious and well-developed. It is a large,
robust vertebra, approximately 5.3 cm wide across the anterior zygapo
physes, with swollen neural arches and widely-spaced anterior zygapo
physes. The first sacral ribs are robust and extend laterally for a short
distance and then are abruptly angled posteriorly. The distal ends of the
ribs are wide and point ventrally, where they contact the ilium almost
vertically. The second vertebra is much smaller than the first and has
relatively much narrower neural arches and more closely-spaced anterior
zygapophyses. The neural spines are also missing on this vertebra. Con
fusion over the presence of one or two sacral vertebrae is understandable
given the morphology of the associated ribs. The second sacral rib is
represented only by 1.8 cm, but retains its contact with the ilium. It is far
less robust than the first. Even though this rib does not directly contact
the second sacral vertebra, it is clear from its size and placement that it
distally articulated with the posterior margin of the second sacral verte
bra.
Caudal Vertebrae and Ribs: In his restoration of Limnoscelis
paludis, Williston (1912) based his description of it as having approxi
mately 60 caudal vertebrae on the almost complete articulated tail of
MCZ 1947. He described the holotype as having 23 caudal vertebrae.
Although approximately 60 caudal vertebrae are currently found
in the holotype of L. paludis, this does not necessarily suggest that this
is the accurate tail count for this specimen. For accuracy, the following
description is based only on the first 23 caudal vertebrae.
The first few caudal vertebrae resemble the second sacral vertebra.
Caudal centra are significantly longer than they are wide, but decrease
serially in length posteriorly. Intercentra are present. The neural arches
are not swollen, and the transverse processes project directly laterally.
The neural spines are tall and blade-like and present only at the anterior
end of the series and decrease serially in height quickly posteriorly. No
alteration of spine height is present. Chevrons in the holotype are de
scribed by Williston (1911a,b, 1912) as first appearing with the third
caudal, and ribs are fused to the first 10 or 11 caudals and distally curve
abruptly posteriorly (Sumida, 1990). However, in the holotype the chev
rons are not visible due to its encasement in plaster and only the first two
left caudal ribs are visible.
214
FIGURE 3. Limnoscelis dynatis, holotype, CM 47653. Dorsal vertebrae demonstrating variable neural spine height. A, Tall-spined and B, low-spined
vertebra.
APPENDICULAR SKELETON
Pectoral Girdle and Forelimb: The pectoral girdle is completely
preserved only in YPM 811, but is not completely exposed. However,
Limnoscelis dynatis (CM 47653) provides important information that is
not visible in the holotype. CM 47653 has a partial pectoral girdle
preserved, including the right clavicle and both scapulocoracoids.
The limnoscelid pectoral girdle consists of an interclavicle and
paired cleithra, clavicles, and scapulocoracoids (Fig. 9A). The interclavicle
is known only in Limnoscelis paludis, but only recently has been ex
posed in ventral view. However, the articulation of the interclavicle ob
scures the ventral surface of the anterior portion of the expanded ante
rior, diamond-shaped head of the interclavicle. What is visible is com
posed of a large, robust head with a long posterior process. The visible
posterior portion of the head is slightly convex ventrally, and in poste
rior view has the shape of a small arc (Fig. 5). The shaft is dilated
anteriorly where it contacts the head and slightly waisted posteriorly,
continuing to a rounded end.
The cleithra of limnoscelids are small, vestigial splints of bone
that contact the upper portion of the anterior margin of the scapular
blade (Williston, 1911a,b). The clavicles are larger bones located ventral
to the cleithra along the anterior margin of the scapular blade. They are
partially visible in YPM 811. The clavicles consist of a dorsal stem and
ventral plate that meet at an angle of approximately 110°, forming a half
sling-like shape. A posteriorly directed lamina on the lateral margin of the
dorsal stem contacts the anterior margin of the scapular blade. A groove
on the upper portion of the lateral surface of the clavicular stem likely
received the ventral end of the cleithrum. The ventral plate turns sharply
dorsally to form a high ridge, which is reduced as it continues as the
dorsal stem. This plate is subdivided into subequal anterior and posterior
portions by a deep groove on the medial margin. The ventral plate is
bowed ventrally and is slightly sculptured with dense, transverse striae
(Berman and Sumida, 1990). In the fully articulated YPM 811, the ante
rior suture between the two clavicles is not apparent.
Characterization of the positioning of the scapulocoracoid in
limnoscelids is problematic because the scapulocoracoid is not fully
visible in the only articulated specimen of YPM 811. Also, the pectoral
girdle is partially crushed in this specimen, making the relative positions
of the scapular and coracoid regions even more difficult to characterize.
The scapulocoracoid is composed of a dorsal scapula and a posterior
ventral coracoid. The scapular blade is relatively short dorsoventrally,
but expanded in its dorsal region anteroposteriorly. The dorsal margin of
the scapular blade in Limnoscelis paludis is thick and convex. The ante
rior border is almost vertically straight, whereas distally the posterior
border curves posteriorly.
Williston (1911a) and Romer (1946) both suggested that the dor
sal portion of the scapular blade had a cartilaginous suprascapula at
tached in Limnoscelis paludis. If there was a cartilaginous suprascapular
element, the base of attachment for this element would have been thick in
L. paludis, but would probably have been too thin for an extensive
suprascapula in L. dynatis (Berman and Sumida, 1990). The
scapulocoracoid is preserved in a single plane in L. dynatis, and Berman
and Sumida (1990) suggested that the coracoid plate had a significant
ventromedial curvature. Unfortunately, the coracoid plate is not com
pletely visible in L. paludis (Figs. 4-5), so a ventromedial curvature of
the coracoid plate cannot be confirmed. A triangular, supraglenoid but
tress faces posterolaterally and is well developed at the base of the
scapula. A vertically expanded supraglenoid foramen is located near its
dorsal apex. The existence of a suture between the anterior and posterior
coracoids in L. paludis was described by Williston (1911a), but this
suture’s existence cannot be confirmed in YPM 811, as this portion of
the girdle is not visible, and the suture is not present in L. dynatis.
However, an angular notch at approximately the same level as the cora
coid suture in L. dynatis may indicate their separate ossifications. The
coracoid plate in L. dynatis is smooth and thin, where it would have been
covered ventrally by the clavicle and interclavicle. The glenoid fossa is
screw-shaped and faces slightly ventrally and posterolaterally at its
anterior end and dorsally at its posterior end. The glenoid fossa is sup
ported by the supraglenoid buttress dorsally and a flange of thickened
bone anteroventrally. A coracoid foramen probably lies directly anterior
to this flange in a deep fossa that undercuts the anterior portion of the
glenoid fossa in L. dynatis (Berman and Sumida, 1990).
Part of the left dorsal scapular blade is visible in YPM 811, as are
ventral parts of both coracoid plates. On the left side, the coracoid is

FIGURE 4. Limnoscelis paludis, holotype, YPM 811, in partial ventral view showing detail of anterior portion of postcranial skeleton.
crushed, and the glenoid fossa is present as a flat basin. A small foramen,
which may be the glenoid foramen, also known as the coracoid foramen,
extends dorsally from a small depression in the coracoid. The coracoid
extends posteriorly beyond the level of the distal end of the interclavicle
stem. Repair of the posterior portion of the coracoid has resulted in its
slight displacement posteriorly and medially as evidenced by the offset
of a fracture on the medial side of the element. The right coracoid is
slightly better preserved. The glenoid foramen is crushed closed, but the
glenoid fossa is well-preserved. The coracoid foramen and fossa are
located on the posterior end of the coracoid, directly lateral to the poste
rior end of the interclaviclar stem. Despite these preservational defects,
there appear to be no significant differences between the pectoral girdle
in YPM 811 and the better preserved examples in Limnoscelis dynatis.
Elements of both forelimbs are present in Limnoscelis paludis,
but their exposure in YPM 811 is obscured partially by plaster, espe
cially on the ventral surface (Figs. 4-5). However, due to postmortem
anteroposterior crushing, the ventral surface of the left humerus of YPM
811 is partially visible (Fig. 6).
The humerus of Limnoscelis paludis (Figs. 6,9) is an extremely
stout bone, approximately 13.2 cm long. It resembles two tetrahedra set
one on top of the other and offset approximately 90°. It has a large,
broad, quadrangular entepicondyle that is convex in outline. The
ectepicondyle is robust, extends significantly anteriorly, and has a con
cave outline. A large entepicondylar foramen appears to have been crushed
closed in YPM 811. Stout supinator and deltoid processes project ante
riorly, and a pectoral process projects posteriorly. A large, hook-shaped
notch separates the supinator and deltoid processes. A narrow ridge
215
extends from the dorsal surface of the ectepicondyle to the proximal
articular surface. Based on the descriptions of the humeri in L. paludis
and L. dynatis, Sumida (1997) reconstructed the distal, ventral aspect of
the humerus in Limnoscelis (Fig. 9C).
The radius is a short, stout element with flared dorsal and ventral
ends and a very short, narrow shaft connecting them. It is about 58% of
the length of the humerus in Limnoscelis paludis. The dorsal surface is
slightly convex with a small, blunt, midline ridge that runs the length of
the shaft. The proximal end is only slightly expanded, and the proximal
facet is slightly cupped for the articulation with the capitulum of the
humerus. The distal end of the radius is more expanded than the proximal
end, with proximodistal parallel striae on its dorsal surface.
The ulna is approximately 84% of humeral length in Limnoscelis
paludis and more heavily built than the radius. A narrow shaft is deeply
concave along its medial margin, but has a nearly straight lateral margin.
The proximal end is flared medially to the semilunar or sigmoid notch,
which is matched on the medial margin on the distal end. The dorsal
surface is slightly convex. The distal articular surface is clearly divided
into two facets for articulation with the carpals (Berman and Sumida,
1990). A slightly laterally facing facet articulates with the ulnare and
pisiform, whereas the slightly medially facing surface articulates with
the intermedium. Interestingly, Williston (1911a) described the ventral
surface of the radius and ulna in L. paludis as more flattened than the
convex dorsal surface. This may indicate slight postmortem compres-
sion on the ventral surface. The ulna and radius have been reconstructed
in Figure 9E, based on the elements in L. paludis and L. dynatis.
The manus is only preserved in the holotype of Limnoscelis
216
FIGURE 5. Limnoscelis paludis, holotype, YPM 811. Oblique ventral view
of pectoral girdle and associated structures.
paludis. The right manus has three proximal carpal bones, whereas the
left manus has four. Williston (1911a,b) described the right manus as also
having four proximal carpals, but only three are currently visible. The
four carpal bones are from lateral to medial: the pisiform, ulnare,
intermedium and radiale. The pisiform is semi-oval, tapered slightly on
its lateral margin, articulating with the ulna on its proximal margin and the
ulnare on its medial margin. The ulnare is the largest of the carpal bones,
almost circular in shape, with a ridge running from the lateral margin
approximately three-fourths of the width of the bone, where it splits
into a small proximo-distal ridge. A small fossa is visible on either side of
the ridge. The ulnare articulates with the ulna proximally and intermedium
medially, but does not seem to have clearly defined facets for articula
tions with any other carpal elements (Williston, 1911a,b). The
intermedium is a small, dorsoventrally thick bone with two distinct
ridges running proximo-distally on its lateral and medial sides. A fossa
separates the two ridges in the left intermedium. In the right manus, the
intermedium articulates with the ulna, ulnare, and radius, whereas in the
left manus, the intermedium has shifted slightly medially, giving the
misleading impression of articulating only with the ulnare and radius.
The radiale is the smallest of the four proximal carpals, almost oval in
shape, flat on its dorsal surface and has a straight, flat radial border. It
contacts the intermedium, but does not have a large articular surface for
it. Williston (1911a,b) described the ventral surface of these proximal
carpal bones as flattened.
Three distal carpal elements are preserved in YPM 811. Origi
nally, Williston (1911a,b) described three distal carpals in both manus;
however, only one is visible in the right and three in the left. The distal
carpals are all smaller than the proximal carpals and are all roughly circu
lar in outline. Williston’s (1911a,b) illustration of the right manus indi
cates one directly proximal to metacarpal III, one lying between distal
carpal III and the intermedium, and one proximal to metacarpal IV. How
ever, the one distal carpal present in the right manus is located directly
proximal to metacarpal III, whereas the carpals present in the left manus
are located proximal to metacarpal V, proximal to and directly in between
metacarpals III and IV, and proximal to metacarpal II. If distal carpal II is
preserved in approximately correct position, then it may tentatively be
identified as the medial centrale. The positional information for the other
FIGURE 6. Limnoscelis paludis, holotype, YPM 811. Dorsal view of left
forelimb and manus.
two carpals of the left manus is not sufficient for a confident identifica
tion. This indicates that the one remaining carpal in the right manus is
most likely correctly positioned, whereas the three distal carpals in the
left manus may have shifted from their natural positions. These distal
carpals most likely represent the centrale and the third and fourth distal
carpals (Williston, 1911a,b).
Williston (1911a) described digits II, III, and IV as being pre
served complete, except for the distal, or ungual, phalanges of digits II
and IV. The ungual phalanx of digit I is also described as missing, as well
as the phalanges of digit V not being correctly articulated to the metacar
pal. However, those digits were preserved in the right manus, allowing
for a reconstruction of both hands (Williston, 1911a). The phalangeal
formula for the manus in limnoscelids appears to be 2-3-4-5-3. The
phalanges serially decrease in length distally. It is important to note that
the carpals and tarsals are possibly not fully ossified peripherally, and
some of the more distal elements were probably not ossified at all, and
this may affect the completeness of the specimen. The overall shape of
the manus is rather broad, with the ungual phalanges shaped like small
hooves, with a thin rounded distal edge, which may have had a keratinous
covering in life (Williston, 1911a).
Pelvic Girdle and Hindlimb: The pelvic girdle is partially pre
served in Limnoscelis paludis. In YPM 811 the complete pelvis is pre
served in articulation; however, due to its having been embedded in
plaster only a portion of the dorsal and ventral surfaces are visible. The
pelvic girdle is reconstructed in Figure 9B based on specimens of L.
paludis and L. dynatis.
The iliac blade is subrectangular in lateral outline except for the
posterior process, which extends posteriorly. Romer (1946) suggested
 217
that this posterior extension of the ilium was tipped with cartilage and
served as an attachment point for caudal tendons and ligaments. There is
no anterior expansion of the iliac blade. The ilium is proportionally short
and wide. A low sinuous or wavy ridge, the iliac shelf, extends
anteroposteriorly across the dorsal portion of the blade (Romer, 1946).
This shelf is a characteristic unique to diadectomorphs (Berman and
Sumida, 1990).A lateral iliac ridge and depression are present in Seymouria
and the pelycosaurian-grade synapsid Ophiacodon (Romer and Price,
1940; Berman and Sumida, 1990). However, the lateral iliac shelf is
developed to a much greater degree in Limnoscelis and in all other
diadectomorphs for which data on the ilium are available, thus making
this well-developed iliac shelf exclusively a diadectomorph characteris-
tic.
The pubis extends anteriorly beyond the level of the anterior
margin of the ilium and has a large anterior process, which has a rounded
convex anterior outline in Limnoscelis paludis. This is unlike the ilium,
which has a more quadrangular outline. An anteroposteriorly-directed
obturator foramen is partially crushed, yet visible in the right pubis of
YPM 811. Williston (1911a,b) described the pelvis of YPM 811 as
having a large, ventral midline keel formed by a ventral deepening of the
pubic and ischiadic symphyses. As preserved, the medial surface of the
ilium is convex and is touched by the two sacral ribs on the right side, and
by extrapolation, on the left side as well.
Currently, the ischium is not visible in Limnoscelis paludis, but
has been described in detail by Williston (1911a,b) and Romer (1946).
The ischium has a large posterior extension and is slightly concave in
posterior outline.
The acetabulum in the holotype of Limnoscelis paludis is only
partially visible in the right pelvis. What is visible suggests it is large and
oval with its long axis directed anteroposteriorly. The acetabulum is
supported dorsally by a small, ventrolaterally-expanded buttress in the
ilium. Notably, the anteroventral margin of the acetabulum has been
illustrated as reaching anteriorly, but not to the pubic margin, in YPM
811 (Romer, 1946).
Most elements of the hindlimb are present in the holotype of
Limnoscelis paludis (Fig. 7). However, the pes is only preserved com
pletely in articulation in MCZ 1948. The femur of L. paludis is very
robust with large, expanded ends connected by a relatively short, narrow FIGURE 7. Limnoscelis paludis, holotype, YPM 811. Dorsal aspect of
shaft, and has an average length of approximately 12.4 cm. It is visible in femur, tibia and fibula.
dorsal view in YPM 811.
The proximal head is angled posteriorly with an almost straight proximal articular surface is slightly sigmoid in outline with the anterior
anterior border and a concave posterior border. The dorsal surface of the end curving ventrally. The cnemial crest is well-developed originating on
proximal head is convex and smooth, except for a significantly rugose the dorsal surface of the proximal head and terminating proximally in a
area on its posterior margin where the puboischiofemoralis internus and small posterior knob. Posterior to the crest is a broad, shallow concavity.
ischiotrochantericus muscles probably inserted. The distal end of the The element narrows substantially to produce a well-defined shaft, then
femur is divided into two distinct condyles that articulate with the tibia expands again, although not as much as the proximal head. The dorsal
and fibula. The distal condyles of the femur are bulbous in shape and surface of the distal head is smooth and has a rounded convex distal
expanded dorsoventrally more than the proximal head. The anterior mar margin.
gin of the femur has a straight border, although it angles slightly anteri The fibula is longer than the tibia, being on average 80% of the
orly towards its distal end. The posterior margin of the femur expands length of the femur in the holotype of Limnoscelis paludis. As in the
posteriorly at both proximal and distal ends to accommodate the articu tibia, it is visible in only dorsal view in YPM 811. Its proximal head is
lar surfaces. The distal articular surface of the posterior condyle is shaped slightly flared with a proportionally longer shaft and a flared distal end.
like a parallelogram, with the dorsal edge slightly longer than the ventral The anterior margin is more convex, whereas the posterior margin is
margin. A small intercondylar fossa is present between the two condyles. almost straight. The articular surface of the proximal head is rugose, with
The surface of the distal end of the femur is smooth and convex, and the a slightly convex dorsal margin that is greatly expanded posteriorly. The
posterior condyle is slightly longer than the anterior condyle. The dorsal surface of the proximal head is flat, and appears much thicker
anteroproximal portion of the internal trochanter is visible in dorsal view dorsoventrally than the distal head. The dorsal surface of the distal head
in YPM 811. is slightly concave and has very fine parallel striae. The distal articular
The tibia is a robust bone with flared proximal and distal ends. surface appears very thin dorsoventrally and longer anteroposteriorly
Both tibiae are present in YPM 811, although only their dorsal surfaces than that of the proximal head.
are visible. It is on average 77% of the length of the femur in Limnoscelis Williston (1911a,b) described the hind foot of YPM 811 as only
paludis. The tibia of YPM 811 is thick dorsoventrally, especially at its being partially preserved in the left, and those bones being weathered and
proximal end. The tibia is anteroposteriorly broad, with an almost straight mostly disarticulated. However, this cannot be confirmed as the hind
anterior margin and a distinctly concave posterior margin, due primarily foot bones of YPM 811 have been reconstructed with plaster, and it is
to a more extensive posterior flaring of the proximal and distal ends. The impossible to determine which bones are the original elements Williston
218
FIGURE 8. Limnoscelis paludis, MCZ 1948 (formerly YPM 809). Dorsal
aspect of the right pes.
described without preparation of the specimen. Examination of the cur
rent mount of the left pes of MCZ 1948 (Fig. 8) indicates that the
relationship of elements is not biologically reliable, and thus a complete
current description of the pes of MCZ 1948 is not possible. Because of
these limitations and to facilitate comparison with other taxa, Williston’s
(1911 a,b) description of the pedes of MCZ 1948 is included here as the
general model in Limnoscelis paludis.
In MCZ 1948 (formerly YPM 809) the pedes are much better
preserved, with two tarsals and partial digits present in the left pes and
two tarsals and complete digits present in the right pes (Williston, 1911a,b,
1912). In the pedes of MCZ 1948, Williston (1911a,b) identified the
proximal tarsals as the fibulare and the fused tibiale and intermedium. He
illustrated the fibulare as an almost circular element and the joined tibiale
intermedium as cuboidal, with articular surfaces for the fibula and tibia,
separated by a slight notch. However, he admitted the possibility that
the tibiale was cartilaginous and the two proximal tarsals are the fibulare
and intermedium. Romer (1946) agreed with this view, and I concur. If
the tibiale and intermedium were fused, the element would be much larger
than what is present in the current mount of the right pes of MCZ 1948.
It is most likely that the two proximal tarsals in this specimen are the
fibulare and the intermedium, with the tibiale missing. Williston (1911
a,b) also recovered some disarticulated phalanges.
The phalanges of the right pes in MCZ 1948 were found in two
separated blocks whose surfaces had been damaged during excavation,
thus distorting the articular surfaces of the phalanges and making their
serial positions difficult to determine. However, Williston (1911a,b)
articulated them on the basis of their similar morphologies and similar
matrix in a seamless, anatomical association. The phalanges of the pes are
similar to those of the manus except for being slightly broader. The
phalanges serially shorten distally. The ungual phalanges are hoof-like. If
the above interpretation is accepted, then the phalangeal count for the
right pes of MCZ 1948 is 2-3-4-5-4 (Fig. 8). Romer (1946) suggested
that the incomplete preservation of the carpus and tarsus is due to
imperfect ossification of the elements. Berman and Henrici (2003) have
have described in greater detail the pedes in the diadectid diadectomorphs
Diadectes and Orobates, and suggest it served as a mechanism to allow
anterior orientation of the manus and pes during locomotion. Of note,
Williston (1911 a, b; 1912) based further description of the hind foot of
L. paludis on FMNHUC 650. This specimen has since been identified as
“Diadectoides,” a taxon that Olson (1947) synonymized with Diadectes.
Olson’s synonymy reinforces the similarity of the pes in Limnoscelis
and diadectids, and lends some support to the potential application of
Berman and Henrici’s (2003) functional interpretation to Limnoscelis.
Their functional interpolation should be considered tentatively accepted
until the discovery of additional or more complete specimens of the pes
in Limnoscelis.
SUMMARY
The high degree of ossification of the postcranium and robust
development of axial and appendicular processes confirm the highly
terrestrial nature of Limnoscelis. Although this study is not intended as
the basis of a phylogenetic analysis of Limnoscelis or the
Diadectomorpha, its results do provide data relevant to future studies of
the group. Limnoscelis demonstrates variability in neural spine height, a
condition common in many basal amniotes and their close relatives. As in
basal amniotes, Limnoscelis has a phalangeal formula of 2-3-4-5-3 for the
manus and 2-3-4-5-4 for the pes.
This study reinforces the contention that the anterior processes
of the atlantal and axial intercentra, and the external iliac shelf are well
developed in Limnoscelis and likely defining features of the
Diadectomorpha, as recently suggested by Sumida et al. (1997) and
Sumida (1997).
A variety of morphological differences are present between the
holotype of Limnoscelis paludis and L. dynatis. In the pectoral girdle and
hindlimb, differences can be seen in the scapulocoracoid, humerus, radius
and ulna. First, the dorsal margin of the scapulocoracoid is thinner and
less convex in L. dynatis than that in L. paludis. Also, the humerus in L.
paludis is 2.2 cm longer than that in L. dynatis. The radius is slightly
longer in L. dynatis than L. paludis, with the former being about 62% of
humeral length, while the latter is only 58% of humeral length. As in the
humerus, the ulna (84% and 80% of humeral length, respectively) of L.
paludis is longer than that in L. dynatis.
Similar differences are also present in the pelvic girdle and hindlimb.
The femora are approximately 1.3 cm in length on average shorter in
Limnoscelis dynatis than those in L. paludis. The femora in L. paludis
also appear more robust and stout in proportion than those in L. dynatis.
Both the tibia and fibula are longer in L. dynatis than they are in L.
paludis. In L. dynatis, the tibia and fibula are 80% and 88% of the femoral
length, whereas the tibia and fibula in L. paludis are 77% and 80% of
femoral length.
A synapomorphy of Limnoscelis is the presence of an anteriorly
directed channel extending from the acetabulum to the anterior edge of
the pubis. Whereas this channel is illustrated by Romer (1946), it is not
clearly visible in YPM 811. However, its presence is confirmed by its
occurrence in L. dynatis. The anteroventral margin of the acetabulum in
this specimen extends anteriorly as a narrow channel that reaches the
anterior border of the pubis (Berman and Sumida, 1990). Pelletier et al.
(2007) have described a similar, though shorter, channel in the Permian
diadectomorph Tseajaia. The presence of a strongly developed adductor
ridge of the femur is not unique to Limnoscelis, however, Pelletier et al.
(2007) have shown that the degree of angulation of this feature may be
useful in distinguishing the genera of Diadectomorpha. Although this
feature is difficult to assess in YPM 811, it is visible in CM 47653. This
study confirms the analysis of Pelletier et al. (2007) that the angulation
of the adductor crest relative to the long axis of the femur in Limnoscelis
dynatis is greater than that in either Tseajaia or Diadectes. The construc
tion of the pes in Limnoscelis presents intriguing potential for compari
son with the more thoroughly described condition in diadectid
diadectomorphs (Berman and Henrici, 2003); however, a functional in
terpretation is outside the scope of this review.
ACKNOWLEDGMENTS
I thank Dr. Stuart Sumida for sharing his knowledge of limnoscelids
 219

FIGURE 9. Reconstructions of various elements of Limnoscelis. A, Left lateral aspect of pectoral girdle, B, Left lateral view of pelvic girdle, C, Distal
ventral aspect of left humerus, D, Ventral view of left femur, E, Dorsal aspect of right radius and ulna, F, Dorsal aspect of left tibia and fibula. The
reconstruction is of slightly disarticulated elements that were laid flat for ease of viewing. Also, the preaxial curvature evident in the tibia may be slightly
exaggerated in the illustration due to the fact that elements available for examination were somewhat flattened. All scale bars equal 1 cm. Adapted from
Sumida (1997).
220
and for all his support throughout my time at California State University
San Bernardino (CSUSB) and beyond, both with this project and others.
This study fulfilled in part the requirements for the Masters of Science at
California State University, San Bernardino. Sincere thanks are also due
my thesis committee: Drs. Joan Fryxell, Anthony Metcalf, and Jim
Ferrari. I am grateful to the staff of the Raymond M. Alf Museum of
Paleontology for all of their support and patience. My sincere apprecia
tion is also extended to Dr. David S Berman and Ms. Amy Henrici of the
Carnegie Museum of Natural History for all of their assistance in study
ing the holotypic skeleton of Limnoscelis, as well as their generous
hospitality. Thank you also to Chuck Schaff of the Museum of Com
parative Zoology for his above-and-beyond-the-call help with studying
REFERENCES
Berman, D.S, and Henrici, A. C., 2003, Homology of the astragalus and
structure and function of the tarsus of Diadectidae: Journal of Paleontol
ogy, v. 77, p. 172-188.
Berman, D.S, and Sumida, S.S., 1990, A new species of Limnoscelis (Am
phibia, Diadectomorpha) from the Late Pennsylvanian Sangre de Cristo
Formation of central Colorado: Annals of Carnegie Museum, v. 59, p.
303-341.
Berman, D.S, Sumida, S.S, and Lombard, R.E., 1992, Reinterpretation of
the temporal and occipital regions in Diadectes and the relationships of
the diadectomorphs: Journal of Paleontology, v. 66, p. 481-499.
Berman, D.S, Sumida, S.S., and Lombard, R.E., 1997, Biogeography of
primitive amniotes; in Sumida, S.S. and Martin, K.L.M., eds., Amniote
origins: Completing the transition to land. San Diego, Academic Press,
p. 85-139.
Berman, D.S. 2000. Origin and early evolution of the amniote occiput:
Journal of Paleontology, v. 74, p. 938-956.
Fracasso, M.A., 1980, Age of the Permo-Carboniferous Cutler Formation
vertebrate fauna from El Cobre Canyon, New Mexico: Journal of Pale
ontology, v. 54, p. 1237-1244.
Fracasso, M.A., 1983, Cranial osteology, functional morphology, system
atics and paleoenvironment of Limnoscelis paludis Williston [Ph.D.
dissertation]: Yale University, New Haven, 624 p.
Fracasso, M.A., 1987, Braincase of Limnoscelis paludis Williston: Postilla,
no. 201, p. 1-22.
Lee, M.Y.S., and Spencer, P., 1997, Crown clades, key characters and taxo
nomic stability: when is an amniote not an amniote?; in Sumida, S.S. and
Martin, K.L.M., eds., Amniote origins: Completing the transition to
land. San Diego, Academic Press, p. 61-84.
Olson, E. C., 1947, The family Diadectidae and its bearing on the classifica
tion of reptiles: Fieldiana: Geology, v. 11, p. 3-53.
Pelletier, V., Sumida, S.S., Walliser, J., and Berman, D.S., 2007, Hindlimb
and borrowing specimens. Sincere thanks also to Jessica Cundiff of the
Museum of Comparative Zoology for her help with the history and
study of the MCZ Limnoscelis specimens. In the CSUSB Vertebrate
Paleontology laboratory, thanks are acknowledged to Gavan Albright
and Kim Scott for assistance with the illustrations, and to Ken Noriega
who provided support and assistance during this study. Thanks are also
due to Brett Austin Kennedy for his support and encouragement through
out the challenges of my graduate studies. Thanks to Dr. Dave S Berman,
Ms. Amy Henrici, Dr. Spencer G. Lucas, and Dr. Stuart S. Sumida for
thorough reviews of drafts of this manuscript; they helped make this a
significantly clearer and more concise paper. All mistakes or misinterpre
tations however, remain my own. This project was funded in part by the
California State University, San Bernardino ASI and IRP grants.
structure of the rare Lower Permian diadectomorph Tseajaia campi
Vaughn: Journal of Vertebrate Paleontology, v. 27, p. 128A.
Romer, A.S., 1946, The primitive reptile Limnoscelis restudied: American
Journal of Science, v. 244, p. 149-188.
Romer, A.S., 1966, Vertebrate paleontology, Third Edition. Chicago, The
University of Chicago Press, 468 p.
Sumida, S. S., 1990, Vertebral morphology, alternation of neural spine height,
and structure in Permo-Carboniferous tetrapods, and a reappraisal of
primitive modes of terrestrial locomotion: University of California
Publications in Zoology, v. 122, p. 1-133.
Sumida, S.S., 1997, Locomotor features of taxa spanning the origin of
amniotes; in Sumida, S.S. and Martin, K.L.M., eds., Amniote origins:
Completing the transition to land. San Diego, Academic Press, p. 353-
398.
Sumida, S. S., Lombard, R. E. and Berman D. S, 1992, Morphology of the
atlas-axis complex of the late Palaeozoic tetrapod suborders
Diadectomorpha and Seymouriamorpha: Philosophical Transactions of
the Royal Society of London, Series B, v. 336, p. 259-273.
Wideman, N. K., and Sumida, S. S., 2004, Taxonomic status of the tetrapod
Limnostygis relictus and its bearing on the temporal distribution of basal
amniotes: Journal of Vertebrate Paleontology , v. 24, p. 129A.
Wideman, N. K., Sumida, S. S., and O’Neil, M., 2005, A reassessment of the
taxonomic status of the materials assigned to the Early Permian tetra
pod genera Limnosceloides and Limnoscelops: New Mexico Museum of
Natural History and Science, Bulletin, 30. p. 358-362.
Williston, S.W., 1911a, A new family of reptiles from the Permian of New
Mexico: American Journal of Science, v. 31, p. 378-398.
Williston, S.W., 1911b, American Permian Vertebrates. Chicago, The Uni
versity of Chicago Press, 145 p.
Williston, S.W., 1912, Restoration of Limnoscelis, a cotylosaur reptile
from New Mexico: American Journal of Science, v. 34, p. 457-468.
 Lucas et al., eds., 2010, Carb-Permian transition in Cañon del Cobre. New Mexico Museum of Natural History and Science Bulletin 49.
TYPOTHORAX COCCINARUM (ARCHOSAURIA: STAGONOLEPIDIDAE)
FROM THE UPPERTRIASSIC (REVUELTIAN)PETRIFIED FOREST FORMATION,
ELPUERTOCITO, CAÑON DELCOBRE, RIOARRIBA COUNTY, NEWMEXICO
JUSTINA. SPIELMANN AND SPENCER G. LUCAS
New Mexico Museum of Natural History and Science, 1801 Mountain Rd. NW, Albuquerque, NM 87104-1375
Abstract—Specimens of the Late Triassic aetosaur Typothorax coccinarum from localities near El Puertocito in
the Cañon del Cobre, Rio Arriba County, New Mexico, consists of dorsal osteoderms, vertebrae and other
postcrania. These fossils are from the Painted Desert Member of the Petrified Forest Formation, as are other
records of T. coccinarum from the Chama basin, and confirm the Revueltian age of these strata.
INTRODUCTION
Aetosaurs (Archosauria: Stagonolepididae) are a suborder of
archosaurs whose fossils are known from the Upper Triassic strata of
North and South America, Greenland, Europe, India, North Africa and
Madgascar (Heckert and Lucas, 2000). Aetosaur fossils are primarily
osteoderms (bony armor) that formed a carapace around their neck,
trunk and tail. Associated and articulated specimens are rare (Huntet al.,
1993; Heckert et al., 2010), but individual osteoderms have taxonomic
utility, allowing identification to genus or species level (e.g., Long and
Ballew, 1985; Long and Murry, 1995; Heckert and Lucas, 2000).
Numerous Late Triassic vertebrate localities in Rio Arriba County
yield aetosaur fossils (Hunt and Lucas, 1989, 1993; Lucas and Hunt,
1992; Long and Murry, 1995). These sites include the Snyder (Typothorax
coccinarum and Rioarribasuchus chamaensis) (Heckert et al., 2003),
Canjilon (Typothorax coccinarum) (Zeigler et al., 2002a; Nesbitt and
Stocker, 2008) and Hayden (Typothorax coccinarum) (Irmis et al., 2007;
A. Downs, personal commun.) quarries, all within the Petrified Forest
Formation (Revueltian) and one locality from the Rock Point Formation
(Apachean) that yields a nearly complete skeleton of Aetosaurus that
awaits description. While vertebrate fossil collecting in Cañon del Cobre
(= El Cobre Canyon) has been ongoing for over a century, this effort has
been primarily focused on Pennsylvanian and Permian strata. Few Trias
sic specimens have been reported from the canyon. Here, we describe
various postcranial fragments of three specimens of Typothorax
coccinarum (NMMNH P-56493 and AMNH 7634, 7635) and summa
rize their significance for regional biostratigraphy.
Institutional abbreviations: AMNH, American Museum of
Natural History, New York and NMMNH, New Mexico Museum of
Natural History and Science, Albuquerque.
PROVENANCE
NMMNH P-56493 was collected from NMMH locality 7141, a
pale green mudstone slope in an unsurveyed section of the southwest
corner of the Canjilon SE7.5’ quadrangle, immediately west-southwest
of El Puertocito (Fig. 1). Numerous outcrops of the Late Triassic Poleo,
Petrified Forest and Rock Point formations bracket both east and west
sides of El Puertocito, which forms the natural southern boundary of the
Cañon del Cobre (Fig. 1) (Kempter et al., 2007; Lucas et al., this volume).
Kempter et al. (2007) mapped these outcrops as Petrified Forest Forma
tion/Rock Point Formation undivided, however, based on the green, slope
forming mudstones and stratigraphic position, NMMNH locality 7141
is in the Painted Desert Member of the Petrified Forest Formation.
The location that the AMNH material was collected from cannot
be so easily identified. All the material was collected by S.J. Olsen and A.
Lewis in 1954 during an expedition for the Museum of Comparative
Zoology, Harvard University, and exchanged with the AMNH in 1956.
The elements were originally identified as “Thecodont” and later “Thec
221
odont, Aetosaurid?” Both AMNH 7634 and 7635 have similar preserva
tion, and it can be surmised that they were collected from the same
locality and cataloged separately based on elements, with the vertebral
fragments given one number (AMNH 7634) and the osteoderms, limb
bones and unidentified postcrania given another (AMNH 7635). The
locality information listed for AMNH 7634 and 7635 is “Chinle, Trias
sic, El Cobre Canyon, New Mexico,” El Cobre Canyon being one of the
other names used for the Cañon del Cobre. Given that Upper Triassic
strata are limited to the southern end of the canyon, around El Puertocito,
the AMNH material was likely collected not far from the NMMNH
material.
NMMNH P-56493
NMMNH P-56493 consists primarily of poorly preserved
osteoderm fragments with a handful of well preserved vertebral centra,
diagnostic osteoderm fragments and a left proximal humerus.
Vertebral centra
Five vertebral centra are present: one anterior dorsal (Fig. 2A-B),
one mid dorsal (Fig. 2C-E), one posterior dorsal (Fig. 2F-H), one sacral
(Fig. 2I-J) and one posterior caudal (Fig. 2K-M).
The anterior dorsal consists of the body of the centrum, missing
both ends, one side of the neural arch and the base of a transverse process
(Fig. 2A-B). The centrum has a poorly developed ventral keel, and the
base of the transverse process is subtriangular in cross section. We iden
tify it as an anterior dorsal because of the position of the transverse
process high on the neural arch and its relatively large size; cervical
vertebrae are extremely reduced in size compared to dorsal vertebrae in
Typothorax coccinarum (Long and Murry, 1995, p. 104; Heckert et al.,
2010).
The mid dorsal is more complete, missing all pre- and
postzygapophyses, the neural spine and lateral ends of the transverse
process (Fig. 2C-E). The anterior articular surface of the centrum is
circular (Fig. 2C), whereas the posterior surface is elliptical, and the
anterior surface is offset slightly higher than the posterior surface (Fig.
2D). The centrum is arched in lateral view and is double keeled (Fig. 2E).
The neural canal is filled with matrix and cannot be discerned. The trans
verse processes are rectangular in dorsal view and project posterolaterally.
The lack of a distinct capitular facet identifies the vertebra as a dorsal,
and the position of the transverse process on the neural arch at the level
of the dorsal margin of the centrum places it in the middle of the series.
The posterior dorsal is very incomplete, preserving only the body
of the centrum and its posterior articular surface and portions of the
neural arch (Fig. 2F-H). The centrum has a single keel and an elliptical
posterior articular surface. The large size of the vertebra and total lack of
a distinct capitular facet suggest it is a posterior dorsal, very close to the
sacrum.
222

FIGURE 1. Index map of the Cañon del Cobre. NMMNH locality 7141, where NMMNH P-56493 was collected, marked with a star.
 223

FIGURE 2. Typothorax coccinarum, NMMNH P-56493. A-B, Anterior dorsal vertebra in A, dorsal and B, end views. C-E, Mid dorsal vertebra in C,
anterior, D, right lateral and E, ventral views. F-H, Posterior dorsal vertebra in F, anterodorsal, G, right lateral and H, posterior views. I-J, Sacral vertebra
in I, posterior and J, ventral views. K-M, Posterior caudal vertebra in K, left lateral, L, ventral and M, anterior views.
224
The sacral vetebra is only represented by an incomplete centrum
and the bases of the transverse processes (Fig. 2I-J). The centrum is not
keeled, and both anterior and posterior articular surfaces are incomplete.
The bases of the transverse processes are very large, running
anteroposteriorly nearly the entire length of the centrum, identifying the
element as a sacral centrum.
The posterior caudal vertebra is nearly complete, missing only
one of its transverse processes and the dorsal tip of the neural spine (Fig.
2K-M). The double keeled centrum is elongate and cylindrical, with
circular articular ends. The transverse process is reduced in size and
splint-like. The neural spine, as preserved, is low, and no indication of
pre- or postzygapophyses is present. The elongate centrum, splint-like
transverse process and lack of zygapophyses confirm the identification
as a posterior caudal vertebra.
Osteoderm fragments
No complete osteoderms were collected, but all fragments pos
sess a consistent patterning offine, randomly arranged pits (Fig. 3C-H).
All osteoderm fragments are flat or only slightly keeled; none are flexed,
as in lateral elements, and thus are paramedian osteoderms. The random
pattern of fine pitting is a diagnostic feature of Typothorax coccinarum
(Heckert and Lucas, 2000; Heckert et al., 2010), contrasting with
FIGURE 3. Typothorax coccinarum, NMMNH P-56493. A-B, Proximal left humerus in A, anterior and B, posterior views. C-D, Osteoderm fragment in
C, dorsal, D, ventral and E, end view. F-H, Osteoderm fragments in dorsal view.
Typothorax antiquum, which has coarser, less dense pitting on its para
median osteoderms (Lucas et al., 2002), and Rioarribasuchus chamaensis,
which has a prominent boss with fine ridges radiating out from it (Zeigler
et al., 2002b; Lucas et al., 2006).
Left proximal humerus
The left proximal humerus is broken just distal to the end of the
deltopectoral crest (Fig. 3A-B). The articular surface of the humerus is
tetralobate and offset from the shaft. The deltopectoral crest, when in
articulation, is directed ventrally. All these features are consistent with
the humeri of aetosaurs (Long and Murry, 1995).
AMNH 7634 AND 7635
None of the vertebral material in AMNH 7634 was complete
enough to describe, though given its preservation it likely came from the
same locality as AMNH 7635, as noted above, and may represent parts
of the same individual.
The identifiable material from AMNH 7635 comprises a right iliac
blade fragment (Fig. 4A-B), a proximal humerus (Fig. 4C-D), both proxi
mal ulnae (Fig. 4E-I), proximal and distal fragments of a femur (Fig. 5A
F), a proximal tibia (Fig. 5G-I), a proximal ?fibula (Fig.5J-L), one com
plete (Fig. 6A) and two fragmentary paramedian osteoderms (Fig. 6B-C)
 225

FIGURE 4. Typothorax coccinarum, AMNH 7635. A-B, Iliac blade fragment in A, medial and B, lateral views. C-D, Proximal right humerus in C, posterior
and D, anterior views. E-H, Proximal ulnae in E, G, posterior, F, I, anterior and H, medial views.
226
and two complete caudal ventral osteoderms (Fig. 6D-E).
Iliac blade fragment
The posterior portion of a heavily concreted, right iliac blade is
preserved (Fig. 4A-B). It is subrectangular in lateral outline and has a
prominent ridge running posteriorly along its medial side. This medial
ridge is more pronounced in Typothorax coccinarum (compare Fig. 4A to
Long and Murry, 1995, fig. 106) than in other aetosaurs such as
Desmatosuchus haplocerus (Long and Murry, 1995, fig. 91) and
Stagonolepis wellesi (Long and Murry, 1995, fig. 80).
Proximal humerus
A right proximal humerus is preserved, though heavily concreted,
with almost no natural bone surface showing (Fig. 4C-D). It is identical
in morphology to the proximal humerus described in NMMNHP-56493
(e.g., tetraradiate proximal articulation, proximal humerus offset from
shaft, etc.).
Proximal ulnae
The left and right proximal ulnae are identical, with an undevel
oped, blunt olecranon process, shallow articular surface for contacting
the distal humerus and anteroposterior expansion of the proximal ulna
compared to the compressed shaft (Fig. 4E-I). Both ulnae are approxi
mately the same size, suggesting they come from the same individual.
The L-shaped olecranon process is identical to the ulnae of Typothorax
coccinarum described by Heckert et al. (2010).
Proximal and distal femoral fragments
An isolated femoral head is complete with both the proximal ends
of the greater and fourth trochanters preserved (Fig. 5A-C). The distal
left femur fragment preserves both condyles, the lateral condyle pos
sessing a large, subtriangular, posteriorly-directed process and the lateral
condyle possessing a lower, posterolaterally-directed process (Fig. 5D
F). The larger lateral condyle has been noted in other specimens of
Typothorax coccinarum (Martz, 2002; Heckert et al., 2010). A fossa is
developed between these processes on the posterior side of the element.
Proximal tibia
The left proximal tibia is expanded at its articular end, and the
articular surface has an elliptical to kidney-shaped outline (Fig. 5G-I).
Proximal ?fibula
The proximal ?fibula has two distinct articular surfaces on its
proximal end, one subcircular and slightly larger than the other, rectangu
lar surface (Fig. 5J-L). In mediolateral view, an elongate fossa extends
distally down the element. We tentatively identify it as a proximal fibula
based on its overall size; it is approximately the same size as the tibia and
much larger than the ulnae, suggesting it cannot be a radius. However,
descriptions of Typothorax coccinarum fibulae do not describe two dis
tinct, proximal articulations (Long and Murry, 1995; Martz, 2002; Heckert
et al., 2010), thus making identification uncertain.
Osteoderms
Five osteoderms, three paramedian and two caudal ventral, are
complete enough for description (Fig. 6). The three paramedians are
complete (Fig. 6A), nearly complete (Fig. 6B) and fragmentary (Fig. 6C),
respectively. Both the complete and nearly complete paramedians are
right sacral or caudal osteoderms based on the presence of a small,
subtriangular dorsal boss that projects beyond the posterior margin of
the rest of the osteoderm. This boss only occurs on sacral and caudal
osteoderms, increasing in size down the tail (Heckert et al., 2010). Also,
both share prominent anterior bars, ventral keels and the random, rela
tively coarse pitting characteristic of Typothorax coccinarum. The frag
mentary osteoderm preserves only the right half of a right paramedian
osteoderm, but shares with the other osteoderms a pattern of fine pit
ting.
The caudal ventral osteoderms are distinct from the paramedians;
they are square to subrectangular in outline, and lack anterior bars or
prominent bosses (Fig. 6D-E). One is larger than the other and is inter
preted as anterior. Their overall morphology is consistent with caudal
ventral osteoderms of Typothorax coccinarum described by Heckert et
al. (2010).
DISCUSSION
The morphology of all elements from NMMNH P-56493 and
AMNH 7634 and 7635 is consistent with Typothorax coccinarum, most
notably the diagnostic patterning of the paramedian osteoderms (Long
and Murry, 1995; Heckert and Lucas, 2000; Heckert et al., 2010).
The preservation of all three specimens is consistent, with creamy
white to black bone, some elements partially concreted with yellow-gray
matrix and all elements possessing dark staining or spots, possibly the
result of manganese. In addition, no elements are duplicated between the
specimens. While it cannot be definitively demonstrated that all three
specimens were collected from the same locality, especially given the
incomplete data for the AMNH specimens, it suggests that they all were
recovered from the same stratigraphic interval, i.e., the Painted Desert
Member of the Petrified Forest Formation.
Typothorax coccinarum is an index fossil of the Revueltian land
vertebrate faunachron of Lucas (1998). Its type material and other speci
mens are known from localities in the Painted Desert Member of the
Petrified Forest Formation east of Cañon del Cobre in the Chama basin
(e.g., Lucas and Hunt, 1992; Heckert et al., 2005). The occurrence of T.
coccinarum in the Painted Desert Member of the Petrified Forest For
mation near El Puertocito thus is consistent with its stratigraphic distri
bution elsewhere in the Chama basin.
While Late Triassic tetrapod fossils are rare locally in the Cañon
del Cobre, the recognition of Typothorax coccinarum within the Painted
Desert Member of the Petrified Forest Formation further confirms its
presence within stata of Revueltian and its utility as a biostratigraphic
index taxon regionally (Lucas et al., 2007).
ACKNOWLEDGMENTS
Randy Pence and Susan Harris discovered and collected NMMNH
P-56493. Carl Mehling provided collection access at the AMNH. An
drew Heckert and Adrian Hunt provided helpful reviews that improved
the manuscript.
 227

FIGURE 5. Typothorax coccinarum, AMNH 7635. A-C, Femoral head in A, posterior, B, anterior and C, articular views. D-F, Left distal femur in D,
anterior, E, posterior and F, distal views. G-I, Left proximal tibia in G, medial, H, lateral and I, articular views. J-L, Proximal fibula in J-K, mediolateral
views and L, proximal views.
228

FIGURE 6. Typothorax coccinarum, AMNH 7635. A, Complete right paramedian osteoderm in dorsal view. B, Nearly complete paramedian osteoderm in
dorsal view. C, Fragmentary paramedian osteoderm in dorsal view. D-E, Caudal ventral osteoderms in ventral view.

REFERENCES
Heckert, A.B. and Lucas, S.G., 2000, Taxonomy, phylogeny, biostratigra
phy, biochronology, paleobiogeography, and evolution of the Late Tri
assic Aetosauria (Archosauria: Crurotarsi): Zentralblatt für Geologie und
Paläontologie, Teil I 1998, Heft 11-12, p. 1539-1587.
Heckert, A.B., Zeigler, K.E. and Lucas, S.G., 2003, Aetosaurs (Archosauria:
Stagonolepididae) from the Upper Triassic (Revueltian) Snyder quarry,
New Mexico: New Mexico Museum of Natural History and Science,
Bulletin 24, p. 115-126.
Heckert, A.B., Lucas, S.G., Sullivan, R.M., Hunt, A.P. and Spielmann, J.A.,
2005, The vertebrate fauna of the Upper Triassic (Revueltian: lower
mid Norian) Painted Desert Member (Petrified Forest Formation: Chinle
Group) in the Chama basin, northern New Mexico: New Mexico Geo
logical Society, Guidebook 56, p. 302-318.
Heckert, A.B., Lucas, S.G., Rinehart, L.F., Spielmann, J.A., Celeseky, M.D.
and Hunt, A.P., 2010, Articulated skeletons of the aetosaur Typothorax
coccinarum Cope (Archosauria: Stagonolepididae) from the Upper Tri
assic Bull Canyon Formation (Revueltian: early-mid Norian), eastern
New Mexico, USA: Journal of Vertebrate Paleontology, in press.
Hunt, A.P. and Lucas, S.G., 1989, Late Triassic vertebrate localities in New
Mexico; in Lucas, S.G. and Hunt, A.P., eds., Dawn of the age of dinosaurs
in the American Southwest: Albuquerque, New Mexico Museum of Natu
ral History, p. 72-101.
Hunt, A.P. and Lucas, S.G., 1993, Stratigraphy and vertebrate paleontology
of the Chinle Group (Upper Triassic), Chama basin, north-central New
Mexico: New Mexico Museum of Natural History and Science, Bulletin
2, p. 61-69.
Hunt, A.P., Lucas, S.G. and Reser, P.K., 1993, A complete skeleton of the
stagonolepipid Typothorax coccinarum from the Upper Triassic Bull
Canyon Formation of east-central New Mexico, USA: New Mexico
Museum of Natural History and Science, Bulletin 3, p. 209-212.
Irmis, R.B., Nesbitt, S.J., Padian, K., Smith, N.D., Turner, A.H., Woody, D.
and Downs, A., 2007, A Late Triassic dinosauromorph assemblage from
New Mexico and the rise of dinosaurs: Science, v. 317, p. 358-361.
Kempter, K., Zeigler, K., Koning, D. and Lucas, S., 2007, Preliminary
geologic map of the Canjilon SE quadrangle, Rio Arriba County, New
Mexico: New Mexico Bureau of Geology and Mineral Resources, Open
file Geological Map 150, 1:24,000.
Long, R.A. and Ballew, K.L., 1985, Aetosaur dermal armor from the Late
Triassic of southwestern North America, with special reference to mate
229
rial from the Chinle Formation of Petrified Forest National Park: Mu
seum of Northern Arizona, Bulletin 47, p. 45-68.
Long, R.A. and Murry, P.A., 1995, Late Triassic (Carnian and Norian)
tetrapods from the southwestern United States: New Mexico Museum of
Natural History and Science, Bulletin 4, 254 p.
Lucas, S.G., 1998, Global Triassic tetrapod biostratigraphy and
biochronology: Palaeogeography, Palaeoclimatology, Palaeoecology, v.
143, p. 347-384.
Lucas, S.G., Heckert, A.B. and Hunt, A.P., 2002, A new species of the
aetosaur Typothorax (Archosauria: Stagonolepididae) from the Upper
Triassic of east-central New Mexico: New Mexico Museum of Natural
History and Science, Bulletin 21, p. 221-233.
Lucas, S.G. and Hunt, A.P., 1992, Triassic stratigraphy and paleontology,
Chama basin and adjacent areas, north-central New Mexico: New Mexico
Geological Society, Guidebook 43, p. 151-172.
Lucas, S.G., Hunt, A.P. and Spielmann, J.A., 2006, Rioarribasuchus, a new
name for an aetosaur from the Upper Triassic of north-central New
Mexico: New Mexico Museum of Natural History and Science, Bulletin
37, p. 581-582.
Lucas, S.G., Hunt, A.P., Heckert, A.B. and Spielmann, J.A., 2007, Global
Triassic tetrapod biostratigraphy and biochronology: 2007 status: New
Mexico Museum of Natural History and Science, Bulletin 41, p. 229
240.
Lucas, S.G., et al., this volume [Geology]
Martz, J.W., 2002, The morphology and ontogeny of Typothorax coccinarum
(Archosauria: Stagonolepididae) from the Upper Triassic of the Ameri
can Southwest [M.S. thesis]: Texas Tech University, Lubbock, 279 p.
Nesbitt, S.J. and Stocker, M.R., 2008, The vertebrate assemblage of the
Late Triassic Canjilon quarry (northern New Mexico, USA) and the
importance of apomorphy-based assemblage comparisons: Journal of
Vertebrate Paleontology, v. 28, p. 1063-1072.
Zeigler, K.E., Heckert, A.B. and Lucas, S.G., 2002a, A tale of two sites: A
taphonomic comparison of two Late Triassic (Chinle Group) vertebrate
fossil localities from New Mexico: New Mexico Museum of Natural
History and Science, Bulletin 21, p. 285-290.
Zeigler, K.E., Heckert, A.B. and Lucas, S.G., 2002b, A new species of
Desmatosuchus (Archosauria: Aetosauria) from the Upper Triassic of
the Chama basin, north-central New Mexico: New Mexico Museum of
Natural History and Science, Bulletin 21, p. 215-219.