ANKYLOSAURS FROM THE PRICE RIVER QUARRIES, CEDAR MOUNTAIN FORMATION
(LOWER CRETACEOUS), EAST-CENTRAL UTAH
and REESE BARRICK
Department of Earth Sciences, Denver Museum of Nature & Science, 2001 Colorado Blvd., Denver, CO 80205;
Prehistoric Museum, College of Eastern Utah, Price, UT 84501
ABSTRACT—A new large nodosaurid ankylosaur, Peloroplites cedrimontanus, is described from a partial skull and
postcranial skeleton found at the PR-2 Quarry located at the base of the Mussentuchit Member of the Cedar Mountain
Formation in central Utah. The specimen is about the same size as the contemporary nodosaurid Sauropelta edwardsorum
from the Cloverly Formation of Montana, and is of an individual approximately 5–5.5 m long. The skull of Peloroplites
differs from that of Sauropelta in the vertical orientation of the suspensorium, non-domed cranium and broad, square
premaxillary beak. The quarry is near and roughly at the same level as the CEM Quarry that produced the holotype of
the ankylosaurid Cedarpelta bilbeyhallorum. The postcrania of Cedarpelta is described and illustrated based on the
paratype and new material. These elements clearly establish that Cedarpelta is closer to Ankylosaurus than to Sauropelta.
As a primitive ankylosaurid, there is no a priori reason to assume that the tail club was present. Based on recent finds in
China, a hypothesis is presented that the tail club is a derived feature in non-shamosaurine (i.e., ankylosaurine) anky-
The past decade has seen the discovery of numerous new an-
kylosaur taxa globally. These discoveries have shed much new
light on ankylosaur evolution and relationships, while at the
same time creating many new questions. The once clear di-
chotomy of two families, the Nodosauridae and Ankylosauridae
(Coombs, 1978), is now blurred with the recognition of a third
family, the Polacanthidae (Carpenter, 2001) based on ankylo-
saurs from the Jurassic (Carpenter et al., 1998) and Lower Cre-
taceous (Kirkland, 1998). Furthermore, the Ankylosauridae is
now subdivided into the subfamilies Shamosaurinae and Anky-
losaurinae (Tumanova, 1983; Vickaryous et al., 2001). The Po-
lacanthidae is not yet universally accepted (e.g., Vickaryous et
al., 2004), but it has been independently substantiated by its
distinctive armor histology (Scheyer and Sander 2004).
Ankylosaurs from the Lower Cretaceous Cedar Mountain
Formation of eastern Utah (Fig. 1) have proven to be uncom-
monly rich and diverse compared to similar aged strata. Whereas
most Lower and even middle Cretaceous strata from North
America only contain a single ankylosaur taxon each (Carpenter
and Kirkland, 1998), the Cedar Mountain Formation has at least
six, probably more (Carpenter et al., 1999, 2002). These ankylo-
saurs are not evenly distributed throughout the formation, how-
ever. Polacanthid ankylosaurs (Gastonia burgei, Gastonia n.sp.
and large unnamed taxon, Kirkland, 1998; Carpenter et al., 2007;
Kirkland and Madsen, 2007) occur in the Yellow Cat through to
basal Ruby Ranch Members, thus apparently become extinct
during the earliest Aptian. The oldest ankylosaurid, Cedarpelta
bilbeyhallorum appears at the base of the Mussentuchit Member
and was briefly described by Carpenter et al. (2001). The oldest
nodosaurid is an unnamed large taxon that appears about in the
middle of the Ruby Ranch (Warren and Carpenter, 2004) and
the oldest named taxon is Animantarx ramaljonesi from low in
the Mussentuchit Member. A previous report of cf. Sauropelta
sp. from the older Poison Strip Member by Carpenter et al.
(1999), following Kirkland (1992, 1996), is probably in error. The
specimen may in fact belong to a new, large polacanthid as dis-
cussed further below.
A large nodosaurid from the Cedar Mountain Formation was
previously mentioned by Burge and Bird (2001). More of that
nodosaurid is now known and is named below. The new taxon
comes from the Price River II quarry (a.k.a., PR-2), which is
about 24.5 km southeast of Price, Emery County, Utah (Fig. 1).
PR-2 has been reported previously as occurring in the Ruby
Ranch Member (Burge and Bird, 2001; Burge et al., 2000), but
the dark, carbonaceous nature of mudstones identifies the strata
as the Mussentuchit Member. In addition to the new nodosaurid,
the quarry has produced four individuals of a new brachiosaurid
sauropod, an iguanodontid, material referable to the ankylosau-
rid Cedarpelta bilbeyhallorum (described below), a turtle (Burge
and Bird, 2001; Burge et al., 2000; Coulson et al., 2004), and a
pterosaur. Radiometric dates (∼109–116 Ma) place the quarry at
the Aptian-Albian boundary (Burton et al., 2006) based on the
2004 revised geologic time scale (Gradstein and Ogg, 2004).
Thus, this part of the Cedar Mountain Formation is equivalent
with at least part of the Cloverly Formation of Wyoming and
Montana (∼108.5 +/− 0.2 Ma, Burton et al., 2006).
Institutional Abbreviations—AMNH, American Museum of
Natural History, New York, New York; CEUM, College of East-
ern Utah, Prehistoric Museum, Price, UT; GISPS, Geological
Institute, Section of Paleontology and Stratigraphy, Ulan Bataar,
Mongolia; HMNS, Hayashibara Museum of Natural Sciences,
Okayama, Japan; KUVP, Kansas University, Vertebrate Pale-
ontology, Lawrence, Kansas; SMU, Southern Methodist Univer-
sity, Dallas, Texas; USNM, United States National Museum
(now the National Museum of Natural History), Washington,
Journal of Vertebrate Paleontology 28(4):1089–1101, December 2008
© 2008 by the Society of Vertebrate Paleontology
Comparative Materials—Comparisons are made with the po-
lacanthid Gastonia (CEUM 1307, and thousands of bones at the
College of Eastern Utah Prehistoric Museum and Denver Mu-
seum of Nature & Science (ms in preparation); nodosaurids Ani-
mantarx (CEUM 6228), Edmontonia (AMNH 5581; HMNS-001;
USNM 11868), Pawpawsaurus (SMU 73203), Sauropelta
(AMNH 3032), and Silvisaurus (KUVP 10296); ankylosaurids
Ankylosaurus (AMNH 5214), Euoplocephalus (AMNH 5404),
and Saichania (GISPS100/1305).
Order ANKYLOSAURIA Osborn 1923
Family NODOSAURIDAE Marsh 1890
PELOROPLITES, gen. nov.
Etymology—“monstrous heavy one”derived from the Greek
peloros “monstrous, gigantic; and hoplites, meaning heavily
armed; and as a subjunctive, a heavily armed soldier.
Type Species—Peloroplites cedrimontanus.
Diagnosis—as for the type and only species.
PELOROPLITES CEDRIMONTANUS, sp. nov.
Etymology—Cedrus Latin for ‘cedar’and mont- Latin com-
bining form of mountain, meaning “from Cedar Mountain”in
reference to the formation.
Type Locality and Stratum—Price River II Quarry (Locality
number EM 372), base of the Mussentuchit Member, Cedar
Mountain Formation, Emery County, Utah. Exact locality infor-
mation is available from the Prehistoric Museum, College of
Diagnosis—Skull lacking premaxillary teeth, occiput slopes
anterodorsally, no prominent lateral temporal notch dorsally as
in Sauropelta; cranial “horns”small, and blunt, paroccipital pro-
cess project laterally, not posterolaterally as in Sauropelta; quad-
rate vertical, not anteriorly bowed or sloped anteroventrally;
odontoid very short, unlike Sauropelta, axis centrum short, as
long as tall. Coracoid:scapula proportions about 2:3, similar to
Animantarx,Edmontonia, unlike 1:3 in Sauropelta.
Holotype—Partial skull CEUM 26331.
Paratypes—Cervicals CEUM11342, CEUM 11354, CEUM
11682, CEUM 26283, CEUM 36714; dorsals: CEUM 8080,
CEUM 11355, CEUM 11654, CEUM 11702, CEUM 11720,
CEU12829, CEUM 36701; synsacrum CEUM 50418, CEUM
11889; caudals: CEUM 19481, CEUM 22684, CEUM 31961,
CEUM 31971, CEUM 32003, CEUM 32022, CEUM 36696,
CEUM 11330, CEUM 36703, CEUM 36709, CEUM 39582,
CEUM 50410, CEUM 11624, CEUM 11681, CEUM 11632,
CEUM 12755, CEUM 12846, CEUM 14382, CEUM 36186;
Chevron CEUM 11735; fragments; CEUM 19507, CEUM 11670,
CEUM 14381. Scapula-coracoid: CEU11706, CEUM 11719; hu-
merus: right CEUM 10614, CEUM 11334, right CEUM 11639,
CEUM 12831; radii: CEUM 11631, CEUM 11655, CEUM 11712,
CEUM 11718; ulnae: CEUM 11708, CEUM 11347, CEUM
11888, CEUM 11669; ilia: CEUM 26342, right CEUM 50427;
pubis CEUM 11626; ischium: distal end CEUM 11313; Femora:
head CEUM 22682, CEUM 11319, CEUM 11321, CEUM 11676;
Tibiae: fragment CEUM 12837, CEUM 11640, CEUM 11710;
fibula: CEUM 12836, CEUM 12158; metacarpals: CEUM 12187,
CEUM 12188, CEUM 12189, CEUM 12191, CEUM 12193;
metatarsal: CEUM 11605; metapodials: CEUM 10638, CEUM
11606, fragment CEUM 11617, CEUM 11639, CEUM 11641,
CEUM 11649, fragment CEUM 11665, CEUM 11666, CEUM
11672, CEUM 11673, CEUM 11700, CEUM 14390; fragment
CEUM 26286 CEUM 19527, CEUM 31239, CEUM 35866; pha-
langes: CEUM 11357, CEUM 11618, CEUM 11623, CEUM
11882, CEUM 11646, CEUM 11656, CEUM 12190, CEUM
12192, CEUM 12219, CEUM 12220, CEUM 12222, CEUM
12834, CEUM 20562, CEUM 32011, CEUM 50383; unguals:
CEU11645, CEUM 11647, CEUM 12164, CEUM 11671, CEUM
12218, CEUM 12221, CEUM 12223, CEUM 12797, CEUM
13286, CEUM 26324, CEUM 19553, CEUM 32002; armor:
CEUM 1344, CEUM 11971, CEUM 11346, CEUM 11350,
CEUM 11973, CEUM 11356, CEUM 11607, CEUM 11650,
CEUM 11697, CEUM 12825, CEUM 36404, CEUM 14050,
CEUM 12388, CEUM 12867, CEUM 26291, CEUM 20561,
CEUM 26333, CEUM 34578, CEUM 38748, CEUM 14331.
A reconstructed skull is shown in Figure 2. It is estimated to
have been 56 cm long and had a maximum width between the
dorsal orbital rims of 35.5 cm, which is about the same width as
Sauropelta AMNH 3035 (Carpenter and Kirkland, 1998). The
FIGURE 1. Stratigraphic section (A) and location map (B)ofPR2
relative to the Cedar Mountain outcrop belt (black) in eastern Utah.
Modified from Kirkland and Madsen (2007). Naturita Formation is used
in place of Dakota Formation for the reasons given by Witzke and Lud-
vigson (1994: 69-73). Naturita Formation, introduced by Young (1960,
1965), is the next available name for these strata.
JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 28, NO. 4, 20081090
snout of Peloroplites tapers anteriorly and terminates abruptly at
a relatively broad premaxillary beak, as compared to Silvisaurus.
The premaxillae are fused along their mid-line (Fig. 3A–C).
Although the left premaxilla is damaged along its lateral margin,
the width of the premaxillary beak is estimated to be 18 cm. The
premaxillae are dorsoventrally thick, whereas they are thin in
Gastonia (unknown in Animantarx). The dorsal surface of the
premaxillae is rugose for the keratinous beak (smooth in Gasto-
nia), as well as arched in anterior view; in addition, the beak has
a broad, inverted U-shaped notch that is less developed than in
Gastonia. A groove is present near the lower margins of the beak
in anterior view (Fig. 3C). This groove continues on to the pala-
tal side (Fig. 3B) defining the lateral edge of the tomial ridge.
This ridge is inset from the lateral margins and the beak is con-
cave between them.
Both the left and right prefrontals and lachrymals are fused
(Fig. 4). The presence of the lachrymal is inferred from the lach-
rymal foramen seen within the anterior orbital wall (Fig. 4F, L).
Both sets of prefrontal-lachrymals are triangular in lateral and
dorsal views. Their external surfaces have rugose sculpturing
composed of irregular pits, anastomosing “valleys”and ridges.
This sculpturing is especially prominent over the orbits. A faint,
shallow groove is present just anterior to the orbit, which extends
onto the dorsal surface of the prefrontal. Similar grooves are
seen in other nodosaurids, such as Edmontonia, and probably
delineate the margins between adjacent keratinous scales in the
living animal. Medially, the prefrontal-lachrymals are divided by
the anterior orbital wall, which separates the orbit from the nasal
cavity. The wall is pierced by the lachrymal fenestra. The dorsal
surface of the orbit is concave.
The postorbital, squamosal, jugal, quadratojugal, and quadrate
are coossified on both side of the skull and all traces of sutures
are missing (Figs. 5, 6). The identity of the elements is inferred
based on topography relative to the landmarks of the orbit and
lateral temporal fenestra. In addition, the right complex is at-
tached to the braincase and dorsal part of the skull. The external
surfaces are rough and pitted, although not to the extent seen
over the orbits. The postorbital horncores (dorsal coronux of
Blows, 2001) are very low, conical structures that project dorso-
laterally. They are much less prominent than seen in many no-
dosaurids, such as Pawpawsaurus or Sauropelta, or even in the
polacanthid Gastonia The jugal-quadratojugal horncores (ven-
tral coronux of Blows, 2001) are not prominent, appearing as
low, localized thickening of bone; they are very prominent and
tapering in the polacanthid Gastonia, as well as in other nodo-
saurids, such as Pawpawsaurus and Animantarx, but moderately
developed in Sauropelta,Edmontonia and Silvisaurus.
The ventral rim of the orbit is probably composed of the jugal
(sutures have been obliterated). It forms a medial-lateral narrow
floor to the orbit. The posterior rim of the orbit is assumed to be
FIGURE 2. Reconstructed skull of Peloroplites cedrimontanus CEUM
26331. A, lateral; B, dorsal; C, posterior. Scale in cm.
FIGURE 3. Premaxillary beak of Peloroplites cedrimontanus CEUM
26331.1: A, dorsal; B, palatal; C, anterior views. Maxillary fragment with
tooth CEUM 26331.6: D, medial side showing tooth; E, lateral side; F,
close-up of tooth. Abbreviations:g, grove; n, notch; tr, triturating ridge.
Scale in cm.
CARPENTER ET AL.—CEDAR MOUNTAIN ANKYLOSAURS 1091
composed of the postorbital and jugal as it typically does in other
dinosaurs. The posterior orbital wall is more complete on the
right side and shows that the wall did not completely separate the
orbit from the temporal fossa. Laterally on the skull, the orbit is
widely separated from the lateral temporal fenestra (Fig. 5A) as
in Edmontonia and unlike Pawpawsaurus. This great width sug-
gests a broad postorbital and jugal, but this needs to be con-
firmed from a younger specimen where sutures are visible. The
squamosal is fused to the head of the quadrate, although crush-
ing of the left side has displaced the head medially some. The
quadrates are slightly bowed anteriorly (seen best in Fig. 6A),
whereas they are more strongly bowed in Edmontonia,Animan-
tarx and Pawpawsaurus, and moderately bowed in Silvisaurus.
They are not inset from the lateral sides of the skull, as they are
in most nodosaurids, including Sauropelta (compare Fig. 6F with
Carpenter and Kirkland, 1998, fig. 9A). In these other nodosau-
rids, the inset of the quadrate recesses the lateral temporal fe-
nestra from the lateral margin of the skull and is manifest in
dorsal view as the lateral temporal notch (Carpenter and
Kirkland 1998). The lateral temporal fenestra, best seen on the
right side, is a vertical rectangle with rounded corners and is
almost keyhole shape.
The frontoparietal region is slightly domed in contrast to the
more extreme condition seen in Panoplosaurus (Carpenter,
1990) and possibly Sauropelta. In posterior or occiput view (Fig.
6C), the dome is moderately arched laterally in a manner similar
to Animantarx,Edmontona and Silvisaurus. The nuchal margin
is scalloped and overhangs the sutural region to the paroccipital
process and supraoccipital. The paroccipital process is twisted
along its length so that the posterior surface faces obliquely
downward. In Edmontonia and Animantarx the entire paroccipi-
tal faces obliquely downward and in Pawpawsaurus it faces
mostly posteriorly. The supraoccipital crest is weakly developed
as is typical for nodosaurids. The facet for the proatlas is seen on
the right side, although it is partially damaged. The exoccipital is
damaged on the left side and the exoccipital-basioccipital suture
of the right side is fused. The amount the exoccipital contributes
to the occipital condyle cannot thus be determined. The foramen
magnum is crushed somewhat so that its exact shape cannot be
determined. The occipital condyle has the typical hemispherical
shape of nodosaurids. In addition, the dorsal surface of the con-
dyle neck is slightly concave as in Pawpawsaurus.
Ventrally, the basioccipital-basisphenoid suture is fused and
assuming that the basitubera marks the approximate location of
the suture, the basioccipital is twice as long as the basisphenoid.
The lateral wall of the braincase is damaged making identifica-
tion of the fenestra somewhat problematic, and therefore we
refrain from doing so at this time. The parasphenoid extends
anterodorsally from the basisphenoid (Fig. 6B) and is fused to
the underside of the cranium. A fragment of the displaced pos-
terior pterygoid plate is present anterior to the parasphenoid
(Fig. 6D, E). The fragment is concave as in Edmontonia and
probably represents the left pterygoid. The adductor fossa is
deep and subtriangular, and is partially separated from the or-
bital by a postorbital wall.
A maxillary fragment retains a single tooth (Fig. 3D–F). The
tooth is large and reminiscent of some teeth referred to Prioco-
nodon from the Aptian-Albian Arundel Formation of Maryland
FIGURE 5. Left posterior section of skull (fused postorbital + jugal +
quadratojugal + quadrate + squamosal) of Peloroplites cedrimontanus
CEUM 26331.4: A, lateral; B, posterior; C, medial views. Abbreviations:
j/jq, jugal + quadratojugal; jh, jugal “horn”;ltf, lateral temporal fenestra;
o, orbit; po, postorbital; poh, postorbital “horn”; q, quadrate; sq, squa-
mosal. Scale in cm.
FIGURE 4. Left prefrontal + lachrymal of Peloroplites cedrimontanus CEUM 26331.2: A, lateral; B, medial into the posterior portion of the nasal
cavity; C, dorsal; D, ventral into the orbit; Eanterior; F, posterior. Right prefrontal + lachrymal of Peloroplites cedrimontanus CEUM 26331.3: G,
lateral; H, medial into the posterior portion of the nasal cavity; I, dorsal; J, ventral into the orbit; K, anterior; L, posterior views. Abbreviations:dow,
dorsal orbital wall; g, groove; lf, lachrymal fenestra; nc, nasal cavity; o, orbit; ow, orbital wall. Scale in cm.
JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 28, NO. 4, 20081092
(Carpenter and Kirkland, 1998), more so than of teeth of Ed-
montonia (see Coombs, 1990). The tooth of Peloroplites cedri-
montanus has an extensive wear facet that extends the entire face
of the crown (Fig. 3F) as is seen in ankylosaurid teeth (e.g.,
Carpenter 2004:fig. 10). The tooth has 7 denticles on the anterior
margin and 6 on the posterior margin.
Both left (Fig. 7A–C) and right (Fig. 7D–F) rear portions of
the mandibles are preserved, although the upper half of the right
portion is folded over. These rear portions include the articular,
angular, surangular and prearticular. Laterally, the posterior sur-
face is remodeled and gives the appearance of a coossified os-
teoderm. However, a crack on the right portion, which reveals
the internal surface, does not support such an interpretation.
Medially, the adductor fossa is large and lateral-medially wider
than in Edmontonia. In addition, the fossa is separated from the
articular by a wall (Fig. 7F); such a wall is absent in Animantarx,
Edmontonia and Sauropelta. The articular cotyle is deeper rela-
tive to the size of the bone than it is in Edmontonia and Ani-
mantarx, and may correspond to the deeper and thicker retro-
articular as well. The enlarged adductor fossa for a large adduc-
tor, thick, deep retroarticular, deep articular cotyle, and
apparently large teeth suggest a strong correlation, possibly re-
lating to a diet of tougher plants than is typically thought of for
ankylosaurs (Coombs and Maryanska, 1990).
The vertebrae of Peloroplites were identified by comparison
with those of Edmontonia and Sauropelta, as well as overall large
size compared with those associated with the paratype of Cedar-
pelta from the CEM site. The total vertebral count is unknown.
The atlas has not yet been found, but it was certainly not coos-
sified to the axis as in Edmontonia (Gilmore, 1930). The axis is
nearly complete, lacking parts of the neural spine and postzyg-
apophyses (Fig. 8A, B). The fragment of neural spine shows that
it was more recumbent than in Sauropelta (Fig. 8C). The centrum
is almost as tall as it is long, in marked contrast to Sauropelta and
Edmontonia where the centrum is significantly longer than tall.
FIGURE 7. Left rear portion of mandible of Peloroplites cedrimonta-
nus CEUM 12850: A, lateral; B, medial; C, dorsal views. Right rear
portion of mandible CEUM 26336: D, lateral; E, dorsal; F, medial. Ab-
breviations:af, adductor fossa; an, angular; ar, articular; cop, coronoid
process; pra, prearticular; rp, retroarticular process; sa, surangular; w,
wall of bone. Scale in cm.
FIGURE 6. Rear portion of skull of Peloroplites cedrimontanus CEUM 26331.5: A, right lateral; B, medial; C, posterior; D, anterior; E, ventral;
F, dorsal views. Abbreviations:af, adductor fossa; bo, basioccipital; bs, basisphenoid; bt, basitubera; fo, fenestra ovalis; fm, foramen magnum; fp,
frontal + parietal; jh, jugal “horn”;ltf, lateral temporal fenestra; nm, nuchal margin; oc, occipital condyle; paf, proatlas facet; pop, paroccipital
process; ps, parasphenoid; pt, pterygoid; q, quadrate; soc, supraoccipital crest. Scale in cm.
CARPENTER ET AL.—CEDAR MOUNTAIN ANKYLOSAURS 1093
Furthermore, the ventral side of the centrum is concave, rather
than straight as in Sauropelta. The odontoid is short, but is long
in Sauropelta. The odontoid overhangs a crescent-shaped, shal-
low articular facet for the atlas. A small, flat surface, which prob-
ably represents the prezygapophysis, is seen on the right side
near the neural canal. The diapophysis is short and located just
posterior of the midline on the centrum (Fig. 8A). This is in
contrast to Sauropelta where it occurs on the neural arch. A large
parapophysis is located anteriorly on the centrum, rather than
medially as in Sauropelta.
The post-atlas cervical vertebrae consist mostly of centra (Fig.
8D–U), although one nearly complete anterior cervical is present
(Fig. 8D–F). The centrum of this vertebra is short, being as long
as tall. It is also sloped, so that the posterior articular face is
lower than the anterior face, a feature described by Gilmore
(1930) in Edmontonia; none of the cervicals of Sauropelta show
this degree of sloping, nor does the cervical of Cedarpelta. The
articular face is circular, rather than heart-shaped as in Edmon-
tonia nor hexagonal or horizontally ellipsoid as in Sauropelta and
Cedarpelta (Fig. 9). The neural spine is incomplete, but enough
remains to show that it was anteroposteriorly narrow and erect,
rather than short and inclined as Cedarpelta (Fig. 9). Further-
more, the anterior margin of the spine extends down between the
prezygapophyses as a ridge. The neural arch is anteroposteriorly
short, tall and erect. As a result, the postzygapophyses do not
reach the level of the posterior face of the centrum, in marked
contrast to most nodosaurids, including Edmontonia and Sauro-
pelta; the postzygapophysis extend well past the centrum in Ce-
darpelta (Fig. 9). The prezygapophysis is short and angled up-
FIGURE 8. Vertebrae of Peloroplites cedrimontanus. Axis of P.ced-
rimontanus CEUM 12812 (A,B) compared with Sauropelta AMNH 3035
(C): A,C, lateral; B, anterior views. Anterior cervical (3?) CEUM 36714
in D, anterior; E, lateral; F, posterior. Anterior dorsal CEUM 26283 in
G, anterior; H, lateral; I, posterior views. Posterior dorsal CEUM 36701
in J, anterior; K, lateral; L, posterior. Anterior caudal CEUM 50410 in
M, anterior; N, lateral; O, posterior. Anterior mid-caudal CEUM 36696
in P, anterior; Q, lateral; R, posterior. Distal caudal CEUM 12846 in S,
anterior; T, lateral; U, posterior. Abbreviations:af, atlas facet; cr, coos-
sified caudal rib; di, diapophysis; in, intervertebral notch for nerve; nc,
neural canal; ns, neural spine; o, odontoid; pa, parapophysis; poz,
postzygapophysis; pz, prezygapophysis; r, coossified rib. Scale in cm.
FIGURE 9. Vertebrae of Cedarpelta bilbeyhallorum: Anterior cervical
(3rd?) CEUM 10396: A, anterior; B, lateral; C, posterior. Synsacrum
CEUM 12163: D, dorsal; E, ventral. Anterior caudal CEUM 10258: F,
anterior; G, lateral. Anterior mid-caudal CEUM 10412: H, anterior; I,
lateral. Mid-caudal CEUM 10404: J, anterior; K, lateral. Abbreviations:
cr, coossified caudal rib; di, diapophysis; ns, neural spine; pa, parapophy-
sis; poz, postzygapophysis; pz, prezygapophysis. Scale in cm.
JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 28, NO. 4, 20081094
wards, whereas it is extremely large in Cedarpelta (Fig. 9). The
diapophysis is moderately long and steeply angled ventroposte-
riorly, rather than horizontally as in Sauropelta. As on the axis,
the parapophysis is large and located adjacent to the anterior
margin of the centrum.
Dorsal vertebrae are represented by several vertebrae, includ-
ing both anterior (Fig. 8G–I) and posterior (Fig. 8J–L). The
centra are amphiplatyan and lack the nodochordal projections
seen in Edmontonia. The centrum of the anterior dorsal vertebra
is shorter than tall as in Edmontonia, whereas the centrum is as
long as or longer than tall in Sauropelta. The ventral margin of
the centrum in Peloroplites is strongly concave, and less so in
Sauropelta. The prezygapophyses are steeply angled as in Sau-
ropelta, whereas they are considerably less so in Edmontonia.
The neural spine is incomplete, but the anterior margin does
extend between the prezygapophyses. The transverse processes
are angled upwards and terminate in subtriangular diapophyses.
Ventrally, a ridge extends along the transverse process to the
parapophysis located on the neural arch. The arch is short and
the neural canal subcircular. The conjoined postzygapophyses do
not extend beyond the posterior face of the centrum.
The posterior dorsal centra are as long as they are tall, thus
resembling more closely those of Sauropelta than Edmontonia,
which are taller than long. Unfortunately, the neural spines are
incomplete, thus their height is unknown. The ribs are coossified
with the vertebrae as is typical for ankylosaurs.
The synsacrum consists of six coossified vertebrae (Fig. 10A,
B), as in Silvisaurus (Carpenter and Kirkland, 1998) of which
three are probably true sacrals. The other vertebrae include a
dorsal and two caudal vertebrae. In light of other ankylosaur
synsacrals, it seems doubtful that only a single dorsal was fused
to the sacrals, although the number of additional dorsals is un-
known. The neural spines for all the synsacral vertebrae are
missing, as are most of the neural arches. These appear to have
been lost prior to burial because of encrustation over the surface.
Ventrally a pair of ridges delineates the synsacral groove as in
Edmontonia. The sacral ribs are damaged, except for the right
second. This rib retains a portion of the acetabular facet.
Caudals are represented by vertebrae from different parts of
the tail (Fig. 8M–U). The anterior caudal (Fig. 8M–O) has a
centrum that is wider than tall, unlike both Sauropelta and Ed-
montonia where height is about equal to width; it is taller than
wide in Cedarpelta (Fig. 9). The centrum is proportionally longer
relative to height as in Sauropelta, rather than short as in Ed-
montonia and Cedarpelta (Fig. 9). The caudal ribs are fused to
the centrum and project laterally, rather than ventrally as in
Edmontonia or anteroventrally as in Sauropelta. Proximally
there is no dorsal process or “shoulder”developed on the caudal
rib as there is in Edmontonia, and the distal ends of the ribs are
expanded as in both Edmontonia and Sauropelta, but not in
Cedarpelta (Fig. 9). The mid-caudal vertebra (Fig. 8P, R) is the
most complete vertebra from the tail region. This specimen
FIGURE 10. Girdle parts of Peloroplites cedrimontanus. Synsacrum CEUM 50427 and iliumCEUM 50418: A, dorsal; B, ventral. Scapula+coracoids
CEUM 11706: C, dorsal; D, lateral; E, ventral; F, medial. Pubis CEUM 11626: G, dorsal; H, lateral; I, medial. The acromion is damaged. Abbre-
viations:a, acetabulum; ac, acromion; cf, coracoid fenestra; co, coracoid; gl, glenoid; i, facet for ilium; ip, ischial peduncle; is, facet for ischium; pa,
postacetabular portion of right ilium; pop, postpubic process; pp, pubic peduncle; pr, preacetabular portion of right ilium; prp, prepubic process; sc,
scapula; sr, sacral ribs; vg, ventral groove. Scale in cm.
CARPENTER ET AL.—CEDAR MOUNTAIN ANKYLOSAURS 1095
shows that the caudal neural spines were laterally expanded as in
Sauropelta and Edmontonia; it is narrow and blade-like in Ce-
darpelta (Fig. 9). The postzygapophyses completely overhang the
posterior face of the centrum, in marked contrast to both Sau-
ropelta and Edmontonia; the prezygapophyses project dorsally
more as well. The neural arch is thick and encloses a subcircular
neural canal. The centrum is circular rather than heart-shaped as
in Edmontonia or hexagonal as in Sauropelta. The caudal rib
projects laterally and slightly ventrally, although the rib on the
right side is distorted ventrally. The distal caudal is damaged, but
shows that the centrum is elongated relative to its height (Fig.
8S–U), whereas the height and length are about the same in
Cedarpelta (Fig. 9). The neural arch is low and encloses a circular
The right scapula and coracoid are coossified (Fig. 10C–F).
They are almost complete, except for a small portion along the
dorsal margin of the coracoid and a small portion of the scapula.
The acromion process of the scapula is damaged and may be
pathological due to avulsion of the deltoideus clavicularis (scapu-
lohumeralis anterior of Coombs, 1978). The remnant of the acro-
mion process indicates that it was not as low towards the glenoid
as in Sauropelta, but rather more similar to the condition seen in
Edmontonia and to a lesser extent Animantarx. The scapular
blade is intermediate in shape between the arcuate Edmontonia
and Animantarx and the straighter Sauropelta. The posterior
margin of the scapula is rounded (Fig. 10D). The glenoid is large
and deep as it is in most ankylosaurs. A trace of the suture
between coracoid and scapula is visible within the glenoid. The
coracoid is elongated and almost as long as the scapula. It is
pierced by the coracoid foramen which is circular on the lateral
surface and ellipsoid on the medial surface where it extends onto
the scapula as a groove.
Parts of two humeri are known (Fig. 11A, B), the left one
being more complete. The proximal end is incomplete, including
the deltopectoral crest. Nevertheless, enough of the crest re-
mains to show that the humeral shaft is elongate as in Sauropelta,
Edmontonia,Animantarx, and Gastonia. Distally, the radial and
ulnar condyles are widely separated by a deep intercondylar
notch. The radial condyle is significantly lower relative to the
ulnar condyle than seen in Sauropelta or Edmontonia.
The left radius and left and right ulna are complete (Fig. 11C–
E). The radius lacks the extreme expanded ends seen in Sauro-
pelta and more closely resembles that of Edmontonia, although
more slender. The ulna is long and straight, rather than bowed as
in Sauropelta; it is short and massive in Cedarpelta and Gastonia.
The olecranon process is tall and massive, and partially over-
hangs the humeral notch; such an overhang does not occur in
Sauropelta. The radial notch on the lateral surface is practically
nonexistent due to lateral crushing.
Carpals are unknown and are probably not present in anky-
losaurs. A partial manus was found in loose association and in
proximity to some of the forelimb material. The manus has a
complete set of metacarpals (Fig. 12) making it one of the few
known specimens. Lambe (1919) and Sternberg (1921) illustrate
and described a partial manus for Panoplosaurus (I-III), Ostrom
(1970) for Sauropelta (I-V), Carpenter (1990) for Edmontonia
(I-IV) and Carpenter and Kirkland (1998) for Nodosaurus (II-
IV?). The complete metacarpal set of Peloroplites allows reas-
sessment of the identities of the metacarpals in Carpenter
(1990:fig 21.17): left metacarpal I fig. 17E; right metacarpal I fig.
17F, left metacarpal II fig. 17D, left(?) metacarpal III fig. 17C,
right metacarpal IV fig 17A, and left metacarpal IV fig 17B.
Metacarpals I-IV of Edmontonia and Sauropelta and I-III of
Panoplosaurus are more similar in length than is the case of
The proximal ends of the metacarpals are faceted and fit snug-
gly against one another (Fig. 12B). Metacarpal I is the broadest
in the manus. It is blocky and sub-rectangular, with a subtrian-
gular proximal end, and no distal condyles. Interestingly, cera-
topsians also have a similar shaped metacarpal I (see Brown,
1917). Metacarpal I of Peloroplites differs from that of Edmon-
tonia and Panoplosaurus in being proportionally shorter com-
pared to its width. The remaining metacarpals of Peloroplites are
hourglass shaped, with expanded proximal and distal ends. Meta-
carpal II is about 25% taller than metacarpal I. Its proximal end
is subtriangular and the distal condyles are weakly separated.
Metacarpal II is proportionally more slender in Panoplosaurus;
FIGURE 11. Forelimb elements of Peloroplites cedrimontanus:A,
right humerus CEUM 11704; B, left humerus CEUM 10614. Left radius
CEUM 11655: A, medial; left ulna CEUM 11347 B, in lateral; right ulna
CEUM 11708 C, in lateral. Abbreviations:dp, deltopectoral crest; hn,
humeral notch; in, intercondylar notch; op, olecranon process; rc, radial
condyle; rf, radial facet; rh, radius head; uc, ulnar condyle. Scale in cm.
FIGURE 12. Manus of Peloroplites cedrimontanus:A, extensor side; B,
proximal ends of metacarpals showing how they nest against one an-
other. I-CEUM 12193, CEUM 12192, CEUM 12218; II-CEUM 12188,
CEUM 12190; III-CEUM 12189, CEUM 12222, CEUM 12220, CEUM
12221; IV- CEUM 12187, 12219; V- CEUM 12191, 12223. B] CEUM
12193, CEUM 12188, CEUM 12189, CEUM 12187, CEUM 12191.
JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 28, NO. 4, 20081096
it is stockier in Peloroplites and Edmontonia (assuming it has
been correctly identified). Metacarpal III is the most robust and
usually the longest in the manus in all families of ankylosaurs.
The distal condyles are separate in Peloroplites and Edmontonia,
but apparently not in Sauropelta and Nodosaurus. Metacarpal IV
is about the same length as metacarpal I, although considerably
more slender. Proximally it is subpentagonal, with a slight con-
cave margin for metacarpal. Distally, the condyles are not sepa-
rate, but form a single surface. Metacarpal V is the smallest of
the series, although slightly more robust than metacarpal IV. The
distal end is rounded, with no development of separate condyles.
Few of the phalanges have been identified and their placement
is mostly estimated. They are all anteroposteriorly short (Fig.
12A), have very shallow proximal articular surfaces and poorly
developed or no distal condyles. Being so short, lateral collateral
ligament pits are absent. The longest and largest phalanx is I-1.
The distal unguals are somewhat disc-shape, being wider than
long, and rounded.
The pelvis is represented by a right ilium (Fig. 10A, B) and left
pubis (Fig. 10G–I). The ilium is incomplete distally and a little
medially. The preacetabular process or blade apparently di-
verges considerably, possibly 55°in contrast to 28°in Edmonto-
nia and 39°in Sauropelta. It is possible, however, that this high
degree of divergence is an artifact of not having a complete
medial surface. The postacetabular process is broad and short as
it is in Sauropelta and Edmontonia. The lateral surface of the
blade is almost straight rather than concave as in Sauropelta and
to a lesser extent in Edmontonia. The pubis is robust compared
to that of Edmontonia and Sauropelta. The lateral face of the
pubis forms the anterior wall of the acetabulum (DiCroce and
Carpenter in prepararation). The preacetabular process is short,
straight and thick; it is proportionally longer and thinner in both
Edmontonia and Sauropelta. The postpubic process is short and
is probably not complete; it is angled posteroventrally.
The left and right femora are complete (Fig. 13A–D). They are
relatively straight shafted and only slightly waisted below the
fourth trochanter. The head is set at a slight upward angle. The
crista trochanteris is low and not as prominent as in Sauropelta,
and the greater trochanter is not as rugose either. An oblique
transverse ridge is present below the greater trochanter (Fig.
13E). It is damaged on the left femur and is better seen on the
right. The fourth trochanter is moderately well developed and is
located about mid-shaft on the posteromedial side. Unlike Sau-
ropelta, the shaft is not strongly waisted or constricted below the
The right tibia is slightly crushed towards the lateral distal end
(Fig. 13F–I), but is otherwise complete. The cnemial crest is short
and rounded in profile. The tibial shaft is thick throughout its
length, not slender-waisted as in Sauropelta (the medial view of
the tibia shown by Ostrom 1970, pl. 26C, D is upside down). The
astragalus is not fused to the distal end of the tibia, which is
unusual, but not unknown, in ankylosaurs. The fibular process
extends well below the astragalar facet.
Of the hind foot, only a right metatarsal (II?) and ungual (I?)
are known (Fig. 13J–M). As with most quadrupedal dinosaurs,
the metatarsal is proportionally shorter and more robust than the
metacarpals. The proximal end of the metatarsal is sloped later-
ally in anterior view, as in Sauropelta, and the distal condyles
well developed, especially posteriorly. The ungual is broad
throughout its length and is distally truncated.
Family ANKYLOSAURIDAE Brown, 1908
CEDARPELTA BILBEYHALLORUM Carpenter, Kirkland,
Burge, and Bird, 2001
Holotype—CEUM 12360 partial skull.
Paratypes—Disarticulated skull: including left premaxilla
CEUM 10405; left nasal fragment CEUM 10410; right prefrontal
CEUM 10421; right lachrymal CEUM 10560; right postorbital
CEUM 10352; jugal fragment CEUM 10598; left frontal CEUM
10325; parietal CEUM 10332; right squamosal CEUM 10345;
left quadrate with attached quadrojugal CEUM 10417; right
quadratojugal CEUM 10561; braincase CEUM 10267; left sur-
angular CEUM 10270; left angular CEUM 10529. Vertebrae:
cervical centrum CEUM 11288; dorsal centra CEUM 10258,
CEUM 10409, CEUM 10442, CEUM 10360; synsacrum of 2 dor-
sals, 4 sacrals, and 1 caudal CEUM 12163; first? caudal CEUM
10258; anterior caudals CEUM 10258, CEUM 10387, CEUM
10366; mid-caudals CEUM 10255, CEUM 10257, CEUM 10260,
CEUM 10261, CEUM 10262, CEUM 10349, CEUM 10400,
CEUM 10412; posterior caudals CEUM 10404, CEUM 10407.
Appendicular: right partial humerus CEUM 10258; left ulna:
CEUM 10425; left ischium, CEUM 10266; partial right ischium
CEUM 10537; fragment of right ilium CEUM10375; cervical ribs
CEUM 10248, CEUM 10445; metacarpals CEUM, 10254,
CEUM 10356, CEUM 10430, CEUM10449, CEUM 10984; pha-
langes CEUM 10247, CEUM 9970; unguals CEUM 9922, CEUM
10253. Armor: keeled plates CEUM 10526, CEUM 10359,
CEUM 10394, CEUM 10431, CEUM 10459, CEUM 10248; com-
pressed conical plates CEUM 10359, CEUM 9960, CEUM 9962,
CEUM 10548, CEUM 10414,CEUM 10441; flat osteoderm
FIGURE 13. Hindlimb material of Peloroplites cedrimontanus: Right
femur CEUM 11319: A, anterior; B, posterior. Left femur CEUM 11321:
C, anterior; D, posterior. E, Close-up of trochanter ridge. Right tibia
CEUM 11640: F, lateral; G, anterior; H, medial; I, proximal. Metatarsal
CEUM 35866: J, anterior; K, lateral? UngualCEUM 26324: L, dorsal; M,
lateral? Abbreviations:4th, fourth trochanter; af, astragalar facet; c,
cnemial crest; cz, crushed zone; ff, fibular facet; fh, femur head; gt,
greater trochanter; in, intercondylar notch; lc, lateral condyle; le, lateral
epicondyle; mc, medial condyle; pf, popliteal fossa; tr, trochanter ridge.
Scale in cm.
CARPENTER ET AL.—CEDAR MOUNTAIN ANKYLOSAURS 1097
Holotype and Paratype Locality—CEM (Locality number
EM 419), base of the Mussentuchit Member (not top of Ruby
Ranch as given by Carpenter et al, 2001).
Referred Specimens—cervical CEUM 10396; caudals CEUM
10412, CEUM 10404; coracoid CEUM 10371; humeri CEUM
10256, CEUM 11629; ischium CEUM 10266; femur CEUM
11334; tibia CEUM 11640.
Additional Locality—Price River, 2 Quarry (Locality number
EM 372), base of the Mussentuchit Member, Cedar Mountain
Formation, Emery County, Utah. Exact locality information is
available from the Prehistoric Museum, College of Eastern Utah.
The skull parts of Cedarpelta bilbeyhallorum are described by
Carpenter and coauthors (2001) and a preliminary reconstruc-
tion of the skull is presented in Figure 14. Originally, the indi-
vidual skull parts were cataloged separately, but now have been
subsumed under a single catalog number. The description of the
skull is supplemented below with that of the postcrania. The
bones of Cedarpelta from PR-2 were separated from those of
Peloroplites by comparison with the paratype material from the
CEM site. Other bones were assigned to Cedarpelta based on
their similarity to other ankylosaurids, such as Ankylosaurus,
Euoplocephalus or Saichania, which differed from those of no-
dosaurids, such as Sauropelta and Edmontonia.
The cervical vertebrae include a complete anterior cervical
(Fig. 9A–C), possibly the third based on the large size of the
prezygapophyses and position of the parapophysis low and an-
teriorly located on the centrum. The neural spine is low and
dorsoposteriorly tapered. Unfortunately the neural spines on the
cervical vertebrae of Peloroplites are too incomplete for com-
parison, although it is clear that the spine does continue as a
sharp ridge between in prezygapophyses, but not in Cedarpelta.
The cervical centrum of Cedarpelta is as long anteroposteriorly
as it is tall, and is wider than it is long. In contrast, the cervical
centrum of Peloroplites is longer than tall, and is as wide as it is
The dorsal vertebrae of Cedarpelta are too damaged for de-
scription. The synsacrum of Cedarpelta is complete, save for
some of the neural spines and some of the sacral ribs (Fig. 9D,
E). Assuming that there are the usual three sacrals, then there
are four sacrodorsals and two sacrocaudals. The sacral groove
developed on the ventral side is restricted to the three sacrals.
The groove is not bounded laterally by a ridge as it is in Pelo-
roplites. The neural spines are coossified in typical ankylosaur
fashion. The anterior caudal is illustrated for the first time in
Figure 9F, G. Unlike Ankylosaurus, the chevron is not coossified
to the centrum. The centrum differs from Peloroplites in that it is
almost as tall as it is wide, whereas it is considerably wider than
tall in the latter. Furthermore, the centrum is longer relative to
its height in Peloroplites than in Cedarpelta. Although the neural
spine is missing in Peloroplites, its base shows that it was wide,
whereas it is narrow in Cedarpelta. The prezygapophysis over-
hangs the centrum in Cedarpelta, but not in Peloroplites. The
mid-caudals of both ankylosaurs are more similar than are the
anterior caudals, except for the thin neural spine of Cedarpelta
versus the wider spine in Peloroplites. Distal caudals are not yet
known, thus the degree of fusion of the distal caudals anterior to
the club (if present) is unknown.
Of the pectoral girdle, only the left coracoid has been found
(Fig. 15A). The dorsal (⳱anterior edge when in anatomical po-
sition) is incomplete. It is plate-like and is pierced by a moder-
ately large foramen. This coracoid foramen is located far from
the scapular suture.
The humerus from PR2 is more complete (Fig. 15B–E) than
the one from PR1 that was associated with the holotype skull
(Fig. 15F, G). Both show the distinctive short shaft between the
deltopectoral crest and radial condyle of ankylosaurids. The
complete humerus is proportionally short and broad, with the
deltopectoral crest situated lower than in Ankylosaurus or Euo-
plocephalus, but this may be the result of anteroposterior crush-
ing because a similar situation occurs in the crushed left humerus
of Ankylosaurus (see Carpenter 2004:fig. 16). This crushing has
also affected the deltopectoral crest, which flares laterally. The
internal tuberosity is prominently distinct, unlike the condition
in Euoplocephalus or Ankylosaurus. Distally, the radial condyle
is significantly lower than the ulnar condyle, whereas in Anky-
losaurus the condition is reversed (Carpenter 2004, fig. 16); in
Euoplocephalus the ulnar condyle is lower than the radial on the
left humerus, but are at the same relative position in the right.
FIGURE 14. Reconstruction (cast) of the skull of Cedarpelta bilbeyhallorum in A, left lateral; B, dorsal; C, anterior; D, ventral. Scale in cm.
JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 28, NO. 4, 20081098
The left ulna is complete, although laterally crushed (Fig.
15H–J). It closely resembles that of the Mongolian Talarurus as
figured by Maryanska (1977:fig. 11A
). The olecranon is trian-
gular in profile and is lower than in Peloroplites (compare with
Fig. 11D, E). The shaft is significantly less elongate and more
robust than Peloroplites as well. The humeral notch is less poorly
developed, and the radial notch only moderately. The two radii
are similar (Fig. 15K–M), although one is distorted. The least
damaged radius, a left, is short and robust, whereas it is much
longer and proportionally narrower in the nodosaurid Peloro-
plites (compare with Fig. 11C).
The ischia are relatively straight shafted (Fig. 16), a character
frequently seen in ankylosaurids (e.g., Ankylosaurus,Saichania
and Pinacosaurus). The proximal end is expanded to form the
medial wall of the acetabulum. On the medial side is a unique
node (Fig. 16C), of unknown function, as previously noted by
Carpenter and coauthors (2001).
The femur of Cedarpelta (Fig. 17A–D) has the typical short
and stocky form of ankylosaurids (except Pinacosaurus), com-
pared with the long, slender nodosaurid femur of Peloroplites.
The head is well developed compared to other ankylosaurids,
although Saichania approaches this condition. The fourth tro-
chanter is non-existent, in typical ankylosaurid condition. Dis-
tally, the medial condyle is considerably larger than the lateral
condyle (they are almost equal in Ankylosaurus) and the lateral
condyle projects lower; in Ankylosaurus,Euoplocephalus,
Saichania, and Pinacosaurus the two condyles are at the same
level. The intercondylar notch is more prominent than in Anky-
losaurus, but is present in most other ankylosaurids.
The tibia is damaged proximally and distally due, in part, to
being crushed against other bones (Fig. 17E–J). Overall, it is very
robust for its length, being more like Saichania than Euoplo-
cephalus. In lateral view the proximal end is anteroposteriorly
expanded mostly due to the cnemial crest. This expansion is
greater relative to tibial length than that seen in Euoplocephalus,
but less than seen in Saichania. The fibula is straight and nearly
parallel-sided, except distally where it flares slightly (Fig. 17K,
L). Overall, it more closely resembles that of Saichania than the
tapering fibula of Ankylosaurus and Euoplocephalus.
The Cedar Mountain Formation has proven to be unusually
rich compared with most Cretaceous formations in ankylosaur
diversity and in the numbers of specimens recovered. The nodo-
saurid Peloroplites is large, matched only in size by Sauropelta
from the Cloverly Formation. It is possible that some of the large
nodosaurid material from the Cedar Mountain Formation that
has been questionably identified as Sauropelta may actually be-
long to Peloroplites. If true, that would extend the stratigraphic
range down to the middle of the Ruby Ranch Member based on
FIGURE 15. Forelimb material of Cedarpelta bilbeyhallorum.A, cora-
coid CEUM 10371. Left humerus CEUM 11629: B, anterior; C, posterior;
D, lateral; E, medial. Right humerus CEUM 10256 in F, anterior; G,
posterior. Left ulna CEUM 11669: H, lateral; I, medial; J, anterior. Left
radius CEUM 11718 in K, lateral; L, anterior. Right radius CEUM 11631
in M, medial? Abbreviations:dp, deltopectoral crest; h, head; hn, hu-
meral notch; in, intercondylar notch; it, internal tuberosity; lsr, lateral
supracondylar ridge; op, olecranon process;rc, radial condyle; rf, radial
facet; rh, radius head; uc, ulnar condyle. Scale in cm.
FIGURE 16. Left ischium of Cedarpelta bilbeyhallorum CEUM 10266:
A, lateral; B, medial showing tuberosity (arrow); C, proximal showing
medial tuberosity (arrow). Abbreviations:ac, acetabulum; if, facet for
ilium; pf, facet for pubis. Scale in cm.
CARPENTER ET AL.—CEDAR MOUNTAIN ANKYLOSAURS 1099
a specimen described by Warren and Carpenter (2004). One
specimen questionably identified as Sauropelta (Carpenter et
al.,1999), cannot be referred to either that genus or to Peloro-
plites. The specimen, from the Poison Strip Sandstone Member
of the Cedar Mountain Formation, was originally described by
Bodily (1969) as Hoplitosaurus, from the Lakota Formation of
South Dakota, based on similar compressed, triangular spines.
Such spines are characteristic of polacanthid ankylosaurs (Car-
penter, 2001), which also includes Hoplitosaurus. Therefore, the
Bodily specimen probably represents an unnamed large polacan-
thid ankylosaur rather than a nodosaurid.
As for Cedarpelta, the postcrania clearly establishes it as an
ankylosaurid (contra Vickaryous et al. 2004) based on the crite-
ria established first by Coombs (1978). Therefore, the presence
of premaxillary teeth in Cedarpelta is a plesiomorphic character
in ankylosaurids as Carpenter and coauthors (2001) originally
concluded. The shamosaurine ankylosaurs (Cedarpelta,Shamo-
saurus and Gobisaurus) now include Zhongyuansaurus luoyan-
gensis recently described by Xu and coauthors (2007) as a no-
dosaurid. The specimen includes a crushed skull, partial man-
dible, partial and complete cervical and dorsal vertebrae,
anterior caudal centra, distal end of caudal vertebrae, left hu-
merus, pubis, both ischia, and ribs. The specimen was considered
to be a nodosaurid based on the narrowness of the skull (ratio of
skull length to skull width ⳱1.4:1) and lack of an ossified tail
club. However, the skulls of shamosaurines retain the plesiomor-
phic narrow skull as measured across the dorsal surface of the
cranium (Shamosaurus 1.3:1; Gobisaurus 1.3:1; Cedarpelta ∼1.8:
1); thus, the ratio of Zhongyuansaurus cannot be used to identify
it as nodosaurid. In addition, the skull of Zhongyuansaurus
shows ankylosaurid, and specifically shamosaurine features
(Vickaryous et al., 2004), including low, blocky profile, covered
lateral temporal fenestra, narrow premaxillary beak, posterolat-
erally directed paroccipital process, and crescent-shaped occipi-
tal condyle. In addition, it has the characteristic straight ischium
of ankylosaurids (Coombs, 1978), and resembles that of Cedar-
pelta (Fig. 16). The end of the tail, however, clearly lacks the
distinctive osteoderms coossified with the vertebrae into a tail
club (Xu et al., 2007:fig. 2D).
The tail club has long been considered as diagnostic of the
Ankylosauridae (e.g., Coombs, 1978; Carpenter, 2001; Vickary-
ous et al., 2004), despite the club only being known among the
ankylosaurine ankylosaurids (e.g., Euoplocephalus Coombs,
1995). The fusion of the distal caudal osteoderms into the tail
club apparently occurred late in life, at least in Pinacosaurus,
because no tail club is known for 1.5 m long juveniles (Dong et
al., 1989). If Zhongyuansaurus is accepted as a shamosaurine
ankylosaurid rather than a nodosaurid, then the tail club is an
autapomorphy of adult Ankylosaurinae. Based on this hypoth-
esis, Cedarpelta, as well as the other shamosaurines, lacked a tail
In conclusion, the Cedar Mountain Formation of eastern Utah
has produced the greatest diversity of Early Cretaceous ankylo-
saurs in the world. This diversity will increase over the next few
years as specimens currently under study are described. The no-
dosaurid Peloroplites cedrimontanus and shamosaurine ankylo-
saurid Cedarpelta bilbeyhallorum are comparable in size to the
5-5.5 m long Sauropelta edwardsorum from the contemporary
Cloverly Formation of Montana. What factors drove ankylosaurs
to attain such large size at the Aptian-Albian boundary remain
We thank Ben Creissler for his suggestions of the genus and
species names. Thanks to Mike Brett-Surman (USNM), Char-
lotte Holton and Mark Norell (AMNH), John Ostrom and Lyn
Murrey (YPM) for access to specimens under their care. Thanks
for review comments by Jim Kirkland and an anonymous re-
Blows, W. T. 2001. Dermal armor of the polacanthine dinosaurs; pp.
363–385 in K. Carpenter (ed.), The Armored Dinosaurs. Indiana
University Press, Bloomington, Indiana.
Bodily, N. M. 1969. An armored dinosaur from the Lower Cretaceous of
Utah. Brigham Young University Geology Studies 16:35–60.
Brown, B. 1908. The Ankylosauridae, a new family of armored dinosaurs
from the Upper Cretaceous. American Museum of Natural History
Brown, B. 1917. A complete skeleton of the horned dinosaur Monoclo-
nius, and description of a second skeleton showing skin impressions.
Bulletin of the American Museum of Natural History 37:281–306.
Burge, D. L., and J. H. Bird 2001. Fauna of the Price River II Quarry of
eastern Utah. Journal of Vertebrate Paleontology 21 (3, Supple-
Burton, D., B. W. Greenhalgh, B. B. Britt, B. J. Kowallis, and, W. Elliot.
2006. New radiometric ages from the Cloverly Formation, Wyo-
ming: implications for contained dinosaur faunas. Geological Soci-
ety of America Abstracts with Programs 38(7) 52.
Carpenter, K. 1990. Ankylosaur systematics: example using Panoplosau-
rus and Edmontonia (Ankylosauria, Nodosauridae). IN Carpenter,
K., and Currie, P. (eds.), Dinosaur Systematics: Perspectives and
Approaches. Cambridge University Press, New York: 282–298.
Carpenter, K. 2001. Phylogenetic analysis of the Ankylosauria. IN Car-
FIGURE 17. Hindlimb material of Cedarpelta bilbeyhallorum: Right
femur CEUM 11676: A, anterior; B, posterior. Left femur CEUM 11334
in C, anterior; D, posterior. Left tibia CEUM 11640: E, anterior; F,
lateral; G, posterior; H, medial; I, proximal; J, distal. Fibula CEUM
12836: K, lateral; L, medial. Abbreviations:a, astragalus; c, cnemial
crest; cz, crushed zone; ff, fibular facet; fh, femur head; gt, greater tro-
chanter; in, intercondylar notch; lc, lateral condyle; le, lateral epicondyle;
mc, medial condyle; pf, popliteal fossa; tf, tibial facet. Scale in cm.
JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 28, NO. 4, 20081100
penter, K. (ed.), The Armored Dinosaurs. Indiana University Press,
Carpenter, K. 2004. Redesciption of Ankylosaurus magniventris Brown
1909 (Ankylosauridae) from the Upper Cretaceous of the Western
Interior of North America. Canadian Journal of Earth Science 41:
Carpenter, K., and J. I. Kirkland. 1998. Review of Lower and Middle
Cretaceous ankylosaurs from North America; pp. 249–270 in, S.G.
Lucas, J. I. Kirkland and, J. W. Estep (eds.), Lower and Middle
Cretaceous Terrestrial Ecosystems. New Mexico Museum of Natu-
ral History and Science Bulletin 14.
Carpenter, K., C. Miles, and K. Cloward. 1998. Skull of a Jurassic anky-
losaur (Dinosauria). Nature 393: 782–783.
Carpenter, K., J. I. Kirkland, D. Burge, and J. Bird. 1999. Ankylosaurs
(Dinosauria: Ornithischia) of the Cedar Mountain Formation, Utah,
and their stratigraphic distribution; pp. 244–251 in D. Gillette (ed),
Vertebrate Paleontology in Utah. Utah Geological Survey Miscel-
laneous Publication 99–1.
Carpenter, K., J. I. Kirkland, D. Burge, and J. Bird. 2001. Disarticulated
skull of a new primitive ankylosaur from the Lower Cretaceous of
eastern Utah; pp. 211–238 in K. Carpenter (ed.), The Armored Di-
nosaurs. Indiana University Press, Bloomington, Indiana.
Carpenter, K., T. DiCroce, D. Gilpin, B. Kinneer, F. Sanders, V. Tidwell,
and A. Shaw. 2002. Origins of the Early and “Middle”Cretaceous
dinosaurs of North America: implications for plate tectonics. Pro-
ceedings of the International Symposium on New Concepts in Glob-
al Tectonics, Otero College, La Junta, Colorado: 289–308.
Coombs, W. 1978. The families of the ornithischian dinosaur order An-
kylosauria. Journal of Paleontology 21:143–170.
Coulson, A., R. Barrick, W. Straight, S. Decherd, and J. Bird. 2004.
Description of the new brachiosaurid (Dinosauria: Sauropoda) from
the Ruby Ranch Member (Cretaceous: Albian) of the Cedar Moun-
tain Formation, Utah. Journal of Vertebrate Paleontology 24 (3,
Coombs, W. 1978. The families of the ornithischian dinosaur order An-
kylosauria. Journal of Paleontology 21:143–170.
Coombs, W. P. 1990. Teeth and taxonomy in ankylosaurs; pp. 269–279 in,
K. Carpenter, and, P. J. Currie (eds.), Dinosaur Systematics: Per-
spectives and Approaches. Cambridge University Press, New York,
Coombs, W., and T. Maryanska. 1990. Ankylosauria; pp. 456–483 in D.
Weishampel, P. Dodson, and H. Osmolska (eds.), The Dinosauria.
University of California Press, Berkeley, California.
Dong, Z., Currie, P. J., and Russell, D. A. 1989. The 1988 field program
of the Dinosaur Project. Vertebrata PalAsiatica 27:233–236.
Gilmore, C. 1930. On dinosaurian reptiles from the Two Medicine For-
mation of Montana. U.S. National Museum Proceedings 77:1–39.
Gradstein, F. M., and Ogg, J. GF. 2004. Geologic Time Scale 2004—why,
how, and where next? Lethaia 37:175–181.
Kirkland, J. I. 1992. Dinosaurs define a two-fold Lower Cretaceous zo-
nation of the Cedar Mountain Formation, central Utah. Geological
Society of America Abstracts and Programs 24: 22.
Kirkland, J. I. 1996. Biogeography of North America’s Mid-Cretaceous
dinosaur faunas, losing European ties and the first great Asian-
North American interchange. Journal of Vertebrate Paleontology,
16 (3, Supplement): 45A.
Kirkland, J. I. 1998. A polacanthine ankylosaur (Ornithischia: Dinosau-
ria) from the Early Cretaceous (Barremian) of Eastern Utah. IN
Lucas, S. G., J. I. Kirkland, and J. W. Estep (eds.), Lower and
Middle Cretaceous Terrestrial Ecosystems. New Mexico Museum of
Natural History and Science Buletin 14:271–281.
Kirkland, J. I. and Madsen, S. K. 2007. The Lower Cretaceous Cedar
Mountain Formation, eastern Utah: the view up an always interest-
ing learning curve. Geological Society of America Rocky Mountain
Section Guidebook, 1–108.
Kirkland, J. I., B. Britt, D. Burge, K. Carpenter, R. Cifelli, F. Decourten,
J. Eaton, S. Hasiotis, and T. Lawton. 1997. Lower to Middle Creta-
ceous dinosaur faunas of the Central Colorado Plateau: a key to
understanding 35 million years of tectonics, sedimentology, evolu-
tion, and biogeography. Brigham Young University Geology Studies
Lambe, L. 1919. Description of a new genus and species (Panoplosaurus
mirus) of an armoured dinosaur from the Belly River Beds of Al-
berta. Transactions of the Royal Society of Canada ser. 4, 13:39–50.
Marsh, O. C. 1890. Additional characters of the Ceratopsidae, with no-
tice of new Cretaceous dinosaurs. American Journal of Science, ser.
Maryanska, T. 1977. Ankylosauridae (Dinosauria) from Mongolia. Pal-
aeontologia Polonica 37: 85–151.
Osborn, H. H. 1923. Two Lower Cretaceous dinosaurs of Mongolia.
American Museum Novitates 95:1–24.
Ostrom, J. H. 1970. Stratigraphy and paleontology of the Cloverly For-
mation (Lower Cretaceous) of the Bighorn Basin area, Wyoming
and Montana. Peabody Museum of Natural History Bulletin 35:
Scheyer, T. M. and Sander, P. M. 2004. Histology of ankylosauir osteo-
derms: implications for systematic and function. Journal of Verte-
brate Paleontology 24:874–893.
Sternberg, C. M. 1921. A supplemental study of Panoplosaurus mirus.
Transactions of the Royal Society of Canada 15: 93–102.
Tumanova TA. 1983. The ﬁrst ankylosaur from the Lower Cretaceous of
Mongolia. Sovmestnaya Sovetsko-Mongol’skaya Paleontogichekaya
Ekpeditsiya, Trudy 24:110–120 (In Russian with English summary).
Warren, D. and K. Carpenter. 2004. A large nodosaurid ankylosaur from
the Cedar Mountain Formation of Utah. Journal of Vertebrate Pa-
leontology 24 (3, Supplement): 126A.
Witzke, B. J., and Ludvigson, G. A. 1994. The Dakota Formation in Iowa
and its type area: 43 78 G. W. Shurr, G. A. Ludvigson and, R. H.
Hammond (eds), Perspectives on the eastern margin of the Creta-
ceous Western Interior Basin. Geological Society of America, Spe-
cial Paper 287.
Vickaryous, M. K., T. Maryan´ska and D. B. Weishampel. 2004. Ankylo-
sauria: pp 363–392 D. Weishampel, P. Dodson, and H. Osmolska
(eds.), The Dinosauria. University of California Press, Berkeley,
Vickaryous, M. K., A. P. Russell, P. J. Currie, and Zhao X.-J. 2001. A
new ankylosaurid (Dinosauria: Ankylosauria) from the Lower Cre-
taceous of China, with comments on ankylosaurian relationships.
Canadian Journal of Earth Sciences 38:1767–1780.
Young, R. G. 1960. Dakota Group of Colorado Plateau. Bulletin of the
American Association of Petroleum Geologists 44:156–194.
Young, R. G. 1965. Type section of the Naturita Formation. Bulletin of
the American Association of Petroleum Geologists 49:1512–1516.
Xu L., Lu J., Zhang X., Jia S., Hu W., Zhang J., Wu Y., and Ji Q. 2007.
New nodsaurid ankylosaur from the Cretaceous of Ruyang, Henan
Province. Acta Geologica Sinica 81:433–438.
Submitted May 28, 2007; accepted March 12, 2008.
CARPENTER ET AL.—CEDAR MOUNTAIN ANKYLOSAURS 1101