Content uploaded by Viktor Tereshchenko
Author content
All content in this area was uploaded by Viktor Tereshchenko on Nov 01, 2018
Content may be subject to copyright.
175
ISSN 0031-0301, Paleontological Journal, 2007, Vol. 41, No. 2, pp. 175–188. © Pleiades Publishing, Ltd., 2007.
Original Russian Text © V.S. Tereschenko, 2007, published in Paleontologicheskii Zhurnal, 2007, No. 2, pp. 59–72.
INTRODUCTION
The Neoceratopsia are only represented in Mongo-
lia by the primitive groups Protoceratopidae Granger et
Gregory, 1923 and Bagaceratopidae Alifanov, 2003.
They are rather abundant and usually well preserved,
have great biogeographic and stratigraphic significance
(Marya ska and Osmólska, 1975; Dodson and Currie,
1990; Russell, 1993; Nessov, 1995, 1997; Brinkman
et al., 1998); in some cases, they alone allow the iden-
tification of horizons or formations (Alifanov, 2000,
2003). However, they are often represented by isolated
fragments of the postcranial skeleton. Since it is diffi-
cult to identify these specimens in field conditions,
researchers pay little attention to this material and pre-
dominantly collect cranial specimens, disregarding the
axial skeleton, although vertebrae, particularly verte-
bral columns, are in fact as informative as skulls (Tere-
schenko, 1991, 2001, 2004a, 2004b, 2004c; Tere-
schenko and Alifanov, 2003; Sukhanov and Tere-
schenko, 2004). However, the use of this material
requires the restoration of the vertebra’s sequence in the
vertebral column and recognition of morphological
characters of each vertebra. On this basis, specimens
need to be identified and the number of individuals in
collections needs to be established. For this reason, a
key to protoceratopoid vertebrae that takes into account
the position in the vertebral column is developed; this
allows the recognition of distinctions in vertebrae
between the two protoceratopoid families.
The key to vertebrae should be preceded by general
information concerning the morphology of the verte-
bral column and of particular vertebrae of protocerato-
poids.
The vertebral column of protoceratopoids usually
consists of 9 cervical, 12 dorsal, 8 sacral, and 38–
50 caudal vertebrae (Brown and Schlaikjer, 1940,
n
′
1942; Sternberg, 1951; Marya ska and Osmólska,
1975). Exceptions are
Leptoceratops gracilis
Brown,
1914 (Sternberg, 1951),
Udanoceratops tschizhovi
Kurzanov, 1992,
Bainoceratops efremovi
Tereschenko
et Alifanov, 2003,
“Udanoceratops”
sp. (Tereschenko,
2004c), and
Archaeoceratops oshimai
Dong et Azuma,
1997 (Dong and Azuma, 1997; You and Dodson, 2003),
which have 13 dorsal vertebrae. The sacral region of
L. gracilis
consists of eight vertebrae (Tereschenko and
Alifanov, 2003; Tereschenko, 2004c) rather than seven
(as was previously proposed by Sternberg (1951)),
because the sacral region comprises vertebrae of which
the ribs are connected to the ilium through special con-
tact areas rather than vertebrae with fused centra. These
contact areas differentiated early in ontogeny, before
the fusion between the elements of vertebrae into the
definitive sacrum (Sukhanov and Tereschenko, 2004).
The fusion between vertebral centra, which is usually
taken in protoceratopoids and other dinosaurs as a cri-
terion of the sacral region, depends on age and individ-
ual features of animals; therefore, it is rejected here
(Tereschenko, 2001).
In protoceratopoids, the initial presacral part of the
vertebral column includes 23 vertebrae, and is marked
by the presence of the first true sacral rib between ver-
tebrae 23 and 24 (Sukhanov and Tereschenko, 2004;
Tereschenko, 2004c), which is distinguished by size,
massiveness, and the connection with the pubic buttress
of the ilium. This rib retains the initial position relative
to its vertebra, coming into contact with the centra of
two adjacent vertebrae (the preceding vertebra has a
posterior semifacet and the succeeding vertebra has an
anterior semifacet of the facet of the parapophysis) and
the neurapophysis of its vertebra; this distinguishes it
from the presacral vertebrae, which are only connected
to the neurapophysis of their own vertebra (for conve-
n
′
Key to Protoceratopoid Vertebrae (Ceratopsia, Dinosauria)
from Mongolia
V. S. Tereschenko
Paleontological Institute, Russian Academy of Sciences, Profsoyuznaya ul. 123, Moscow, 117997 Russia
e-mail: valifan@paleo.ru
Received December 26, 2005
Abstract
—A detailed nomenclature and a measurement system for vertebrae of Mongolian Protoceratopoidea
are proposed. A key to vertebrae that allow the determination of the region of the vertebral column and the serial
number of each vertebra within each region is developed. Distinctions in vertebral column between two proto-
ceratopoid families (Protoceratopidae and Bagaceratopidae) are summarized.
DOI:
10.1134/S0031030107020086
Key words
: Protoceratopoids, Ceratopsia, postcranial skeleton, vertebrae, Mongolia.
176
PALEONTOLOGICAL JOURNAL
Vol. 41
No. 2
2007
TERESCHENKO
nience, they are designated as the
upper ribs,
in contrast
to the
lower ribs
of the sacrum and tail). Thus, the
24th vertebra corresponds to the only primary sacral
vertebra of the amphibian ancestor of reptiles (principal
sacral vertebra after Romer, 1956). In protoceratopoids,
this vertebra occupies a constant position in the verte-
bral column, while the boundaries of generally
accepted regions vary depending on the taxon. In
Pro-
toceratops
and
Bagaceratops
, the sacrum includes two
dorsosacral and five sacrocaudal vertebrae, while
Udanoceratops
and its close relative
Leptoceratops
(Kurzanov, 1992; Tereschenko, 2004c) have only one
dorsosacral and six sacrocaudal vertebrae, i.e., the
sacrum is displaced caudally for one vertebra.
Previously, it was generally accepted that the lum-
bar vertebrae have transverse processes formed by the
fusion between rudimentary ribs and centra. At present,
all posterior dorsal vertebrae lacking ribs (Gurtovoi
et al., 1978) or having ribs that are not connected to the
sternum (Dzerzhinsky, 1998) are also referred to the
lumbar region. In the available material, specimen PIN,
no. 3143/5 of
Protoceratops andrewsi
retains seven
posterior dorsal vertebrae and the pelvic girdle in natu-
ral articulation. The ribs of three posterior dorsal verte-
brae of this specimen are short, coming into contact dis-
tally, directed to the ventrocranial surface of the preac-
etabular process of the ilium. The ribs of two posterior
dorsal vertebrae adjoin this surface; they were probably
connected by connective tissue. In
Udanoceratops
tschizhovi,
the rib of the 11th dorsal vertebra is short,
curved posteriorly, and directed to the preacetabular
process, like the rib of the 10th dorsal vertebra of
P. andrewsi.
Thus, three posterior dorsal vertebrae of
protoceratopoids are modified and have short ribs,
which are not connected to the sternum. Therefore, it
seems plausible to assign them to an unusual lumbar
region, which is distinguished by both morphological
and biomechanical features from the preceding dorsal
vertebra. Based on the above, the vertebral column of
protoceratopoids is divided into 9 cervical, 9 (10) dor-
sal or thoracic, 3 lumbar, 8 sacral, and 38–50 caudal
vertebrae. In the text, figures, and tables, the serial
numbers of vertebrae are accompanied by letters desig-
nating the regions as follows: (c) cervical, (th) thoracic,
(lm) lumbar, (sc) sacral, and (cd) caudal.
The keys given below are proposed based on the
study of the protoceratopoid skeletons and their frag-
ments housed in the Paleontological Institute of the
Table 1.
Protoceratopoids, the axial skeleton of which was examined to produce the key
Species Specimen
PIN, no.* Material Axial skeleton
Protoceratops andrewsi
Granger et Gregory, 1923 614/31 vertebrae 2sc-2cd
614/32 vertebrae 5-7sc, 1-3cd, ?5cd, ?14-18cd**
614/35 vertebrae 6c-9th
614/61 vertebrae and pectoral girdle 1-3th
3143/5 skeleton 1c-8sc
3143/7 skeleton 4
Ò
-17cd
3143/9 skeleton 1c-25cd
3143/12 skeleton without skulls 1th-3 lm, 1-8sc for articulation with pelvic
region
3143/16 vertebrae 5c, 9c-3th, 3 l m-4cd, 6-13cd
Bagaceratops
sp. 614/29 vertebrae 5c-2sc, 7sc-1cd, ?10-15cd
614/34 vertebrae 1th, 3-7th, 5-7sc, 1-3cd, ?10cd
614/53 skeleton fragment (pelvic region) 1sc-8cd, ?15-18cd
3142/7 skeleton 6
Ò
-6sc
3143/11 vertebrae ?9-23cd
4550/3 vertebrae 6c-9th
Bainoceratops efremovi
Tereschenko et Alifanov, 2003 614/33 vertebrae 6-7
Ò
, 9c, ?1th, 5-7th, 9th-3 lm, 8sc-3cd,
?10-11cd, ?18-19cd
Udanoceratops tschizhovi
Kurzanov, 1992 3907/11 skeleton 5c-2cd, 4-10cd, ?13-31cd, ?35cd, ?42cd
Udanoceratops
aff.
tschizhovi
(“
Udanoceratops
” sp.) 4046/11 skeleton 2c-13cd
*
Collection numbers correspond to the following localities in Mongolia: (614) Bain-Dzak, (3907) Udan-Sair, (3142) Khermin-Tsav,
(3143) Tugrikiin-Shire, (4046) Baga-Tariach, and (4550) Gilbentu.
**
(?) incomplete preservation of vertebrae prevents precise identification of the serial number.
PALEONTOLOGICAL JOURNAL
Vol. 41
No. 2
2007
KEY TO PROTOCERATOPOID VERTEBRAE (CERATOPSIA, DINOSAURIA) FROM MONGOLIA 177
Russian Academy of Sciences (PIN), and collected by
the Mongolian Paleontological Expedition of the Acad-
emy of Sciences of the USSR (1946–1949), Joint
Soviet–Mongolian Paleontological Expedition (1969–
1991), and Mongolian–Russian Paleontological Expe-
dition (1992 to the present time) in Late Cretaceous
localities, in the Dzhadokhskaya (Tugrikiin-Shire,
Bain-Dzak, Udan-Sair, Baga-Tariach) and Barun Goyot
(Khermin-Tsav, Gilbentu) formations (Table 1). Addi-
tional material includes a cast of the skeleton of
Lepto-
ceratops gracilis,
the original of which is stored in the
National Museum of Canada (NMC, no. 8887).
The vertebrae were identified in two stages; first, the
region of the vertebral column was determined; then,
the serial number of each vertebra within the region
was established. This allows subsequent comparisons
between taxa and identification of taxa based on iso-
lated vertebrae. This was formerly a stumbling block
for the study of vertebral columns and isolated verte-
brae of all dinosaurs. We frequently deal with postcra-
nial elements without skulls that clearly differ from
those of known taxa. We face a dilemma whether to
establish new taxa, although a number of taxa have
already been described based only on skulls, or to use
open nomenclature before the study of all available
material in which skulls are associated with the postc-
ranial skeleton. Although some vertebrae allow the
identification of species, in this paper, preference is
given to the determination of families.
Note that the thoracic region of
Protoceratops
and
Bagaceratops
contains one less vertebra than that of
Udanoceratops
, and the boundary of the sacral region
varies. This results in different serial numbers of verte-
brae in the thoracic and sacral regions of the vertebral
column; therefore, the serial numbers of respective ver-
tebrae of
Udanoceratops
are sometimes given in paren-
theses. For example, the record 3(2)sc means 3sc of
Protoceratops
or
Bagaceratops
and 2sc of
Udanocer-
atops
.
The standard terminology used in anatomical
descriptions of vertebrae is insufficient to describe the
surface sculpture, which not only varies in different
taxa, but also follows certain patterns of change within
the same vertebral column. Therefore, it is proposed to
recognize certain elements on the vertebral surface,
including true crests, sharp margins of vertebral ele-
Table 2.
Comparison of nomenclatures proposed for crestlike structures of dinosaurian vertebrae
Osborn and Mook, 1921 Janensch, 1929 Wilson, 1999 This paper
laminae Leisten laminae cristae
infraprediapophysial vorderer Centrodiapophysialleiste anterior centrodiapophyseal craniodiapophysalis
infrapostdiapophysial hinterer Centrodiapophysialleiste posterior centrodiapophyseal caudodiapophysalis
infradiapophysial Centrodiapophysialleiste infradiapophysalis
horizontal Präzygodiapophysialleiste prezygodiapophyseal praediapophysalis
supradiapophysial Supradiapophysialleiste,
Lateraspuinalleiste spinodiapophyseal spinodiapophysalis
horizontal Postsygopophysialleiste postzygodiapophyseal postzygodiapophysalis
horizontal Paradiapophysialleiste paradiapophyseal paradiapophysalis
infraprezygapophysial Infrapräzygapophysialleiste
(mediale, laterale) centroprezygapophyseal infraprezygapophysalis
supraprezygapophysial Suprapräzygapophysialleiste spinoprezygapophyseal spinoprezygapophysalis
intraprezygapophysial Intrapräzygapophysialleiste intraprezygapophyseal interpraezygapophysalis
infrapostzygapophysial Infrapostzygapophysialleiste
(mediale, laterale) centropostzygapophyseal infrapostzygapophysalis
Infrahyposphenalleiste
suprapostzygapophysial,
suprahyposphenal Suprapostzygapophysialleiste
(mediale, laterale) spinopostzygapophyseal
(medial and lateral)* spinopostzygapophysalis
intrapostzygapophysial Intrapostzygapophysialleiste intrapostzygapophyseal interpostzygapophysalis
infraprezygapophysial vorderer Centroparapophysialleiste anterior centroparapophyseal infrapraeparapophysalis
oblique hinterer Centroparapophysialleiste posterior centroparapophyseal infrapostparapophysalis
oblique Präzygoparapophysialleiste prezygoparapophyseal prezygoparapophysalis
prespinal Präspinalleiste prespinal praespinalis
postspinal Postspinalleiste postspina postspinalis
* Cases of bifurcations of crests into two are absent from protoceratopoids and, hence, they are not considered here.
178
PALEONTOLOGICAL JOURNAL
Vol. 41
No. 2
2007
TERESCHENKO
ments, edges, facets, etc. This has already been
achieved for sauropods, which have unusual vertebrae
(Osborn and Mook, 1921; Wilson, 1999); however, in
that case they were termed
laminae.
When naming sur-
face elements, we use a mixed principle, taking into
account both connection with certain vertebral ele-
ments, such as processes, centrum, etc. (as Wilson did),
and their relative position, i.e., anterior, posterior, dor-
sal, ventral, etc. (as Osborn and Mook did) (Table 2).
The following designations are used for vertebral struc-
tures: (
c
) crest, (
p
) process, (
v
) ridge, (
i
) incisure,
(
f
) facet, and (
s
) suture.
The major vertebral elements are the centrum and
neurapophysis, which are always connected by the cen-
troneurapophysial suture (Fig. 1,
scnph
). The neurapo-
physis, and occasionally the centrum, have special pro-
cesses: the unpaired neural spine (spinous process),
paired prezygapophyses and postzygapophyses (
prz
and
ptz
), and the so-called transverse processes, includ-
ing the diapophyses (
d
), or, less often, synapophyses (
s
),
and parapophyses (
p
); in some vertebrae, they are
located on the neurapophysis, in others, on the centrum.
The parapophyses are in the shape of a process or a
facet. The sculptured facets of the zygapophyses of
some thoracic vertebrae have one or two longitudinal
ridges (
vfaz
). In the caudal region, the single-headed
ribs are fused with the vertebral centrum and neurapo-
physis to form the costal processes (
pc
). The centrum
has six basic surfaces; the dorsal and ventral surfaces fre-
quently have dorsal and ventral longitudinal crests (
cv
),
which project out of the surface of the centrum (in the
caudal region, the ventral crest is paired); in addition,
there are the cranial and caudal articular facets and two
lateral surfaces. The ventral crests of the caudal verte-
ciz
(a) (b) (c) (d)
(e) (f) (g) (h)
prz
csprd
d
ivc
p
cpp
ivca
cid
cv
cptd
p
ciprp
csptz
ccd
d
cprzd cpd
vfaz
iiprz
ciptp
scnph
pd
prz
cprd
cprzd
ptz
iiptz
ptz
ccs
s
csptz
cpts
cps
cptzs
cprs pc
scsprz cspts
pzr
cv
fah
Fig. 1.
Protoceratopoid vertebrae, elements of vertebrae and surface sculpturing: (a, c, d, h)
Protoceratops andrewsi
Granger et Gre-
gory, 1923: (a) specimen PIN, no. 614/35, 6c; (c) specimen PIN, no. 3143/5, 5th; (d, h) specimen PIN, no. 3143/7: (d) 9th and
(h) 3cd; (b)
Bagaceratops
sp., specimen PIN, no. 614/29, 8c; (e)
Udanoceratops tschizhovi
Kurzanov, 1992, specimen PIN,
no. 3907/11, 3lm; and (f)
Bainoceratops efremovi
Tereschenko et Alifanov, 2003, specimen PIN, no. 614/33, 3lm; (a, c, e) left lateral
view; (b, d, g) dorsal view; (f) posterior view, and (h) ventral view. Designations: (
ccd
) crista caudodiapophysalis, (
ccs
) cr. caudosy-
napophysalis, (
cid
) cr. infradiapophysalis, (
ciprp
) cr. infrapraeparapophysalis, (
ciptp
) cr. infrapostparapophysalis, (
ciz
) cr. interzyg-
apophysalis, (
cpd
) cr. paradiapophysalis, (
cpp
) cr. postparapophysalis, (
cprd
) cr. praediapophysalis, (
cprs
) cr. praespinalis,
(
cprzd
) cr. praezygodiapophysalis, (
cps
) cr. postspinalis, (
cptd
) cr. postdiapophysalis, (
cpts
) cr. postsynapophysalis, (
cptzd
) cr.
postzygodiapophysalis, (
cptzs
) cr. postzygosynapophysalis, (
csprd
) cr. suprapraediapophysalis, (
csprz
) cr. spinoprezygapophysalis,
(
csptz
) cr. spinopostzygapophysalis, (
cv
) cr. ventralis, (
d
) diapophysis, (
fah
) facies articularis haemalis, (
iiprz
) incisura interpraezyg-
apophysalis, (
iiptz
) inc. interpostzygapophysalis, (
ivc
) inc. vertebralis cranialis, (
ivca
) inc. vertebralis caudalis, (
p
) parapophysis,
(
pc
) processus costarum, (
prz
) praezygapophysis, (
ptz
) postzygapophysis, (
s
) synapophysis, (
scnph
) sutura centroneurapophysalis,
and (
vfaz
) vallum facii articularis zygapophysi. Scale bar, 10 mm.
PALEONTOLOGICAL JOURNAL
Vol. 41 No. 2 2007
KEY TO PROTOCERATOPOID VERTEBRAE (CERATOPSIA, DINOSAURIA) FROM MONGOLIA 179
brae form two ventrocaudal facets for the buttresses of
the haemal arch (fah). At the midlength, the centrum is
divided by a constriction (minimum breadth of the cen-
trum), while its cranial and caudal ends are widened.
When connected, the vertebrae show three interverte-
bral foramina: two lateral foramina (intervertebralia
laterale), which are formed by the caudal and cranial
vertebral incisures (ivca and ivc, respectively), and one
dorsal foramen (intervertebrale dorsale), which is
formed by the interpostzygapophysial and interprezyg-
apophysial incisures (iiptz and iiprz). The caudal verte-
bral incisure is the posterior surface of the base of
postzygapophyses and buttresses of the neural arch,
while the cranial incisure is the anterior surface of the
base of the prezygapophyses and buttresses of the neu-
ral arch of the succeeding vertebra. The interpostzyga-
pophysial incisure is formed by the caudal surface,
while the interprezygapophysial incisure is formed by
the cranial margin of the neural arch between the
postzygapophysis and prezygapophysis, respectively. If
the surface of the incisures of the dorsal intervertebral
foramen is ridged, the names of these ridges coincide
with the names of the incisures. However, the ridged
surface of the lateral intervertebral foramen should
preferably be named the infraprezygapophysial and
infrapostzygapophysial crests.
A crest that extends on the dorsal or ventral surface
of any transverse process, i.e., (p) parapophysis,
(d) diapophysis, (s) synapophysis, etc., is designated by
the prefix supra- or infra-, e.g., supradiapophysial,
infradiapophysial (cid); while crests on the cranial or
caudal surfaces, by the prefix cranio- or caudo-, e.g.,
caudodiapophysial (ccd), caudosynapophysial (ccs). If
the crest passes on the dorsocranial or dorsocaudal sur-
face, the prefixes suprapre- or suprapost- are used, e.g.,
supraprediapophysial (csprd), suprapostdiapophysial;
if it passes on the ventrocranial or ventrocaudal surface,
the prefixes are infrapre- and infrapost-, e.g., infra-
preparapophysial (ciprp), infrapostparapophysial (ciptp).
However, if such a crest is horizontal, we use the pre-
fixes pre- or post-, e.g., prediapophysial (cprd), postdi-
apophysial (cptd), postsynapophysial (cpts), and post-
parapophysial (cpp) = margo ventralis after Hoffstetter
and Gasc (1969). The cranial and caudal edges of the
neural spine are termed the prespinal and postspinal
crests (cprs, cps), respectively. In some cases, it is nec-
essary to mark the extent to which they stretch on the
neural arch between the prezygapophysis and postzygapo-
physis. A crest that connects the lateral edges of the
prezygapophysis or postzygapophysis is named the
interzygapophysial crest (ciz) = margo lateralis after
Hoffstetter and Gasc (1969). A crest on the lateral surface
of the prezygapophysis or neural spine is named the lat-
eroprezygapophysial or laterospinal. If crests are located
between two processes (or their facets), their names are
composed of the names of both processes, e.g., prezyg-
apophysial (cprzd), postzygapophysial (cptzd), spino-
prezygapophysial (csprz), paradiapophysial (cpd), spino-
postzygapophysial (csptz), postzygapophysial (cptzs), etc.
The identification of vertebrae requires certain mor-
phometric parameters that follow change in size and
position of particular elements along the vertebral col-
umn. Some important parameters are shown in Fig. 2.
The length (L), height (H), and breadth (B) of the ver-
tebral centrum are essential measurements, characteriz-
ing their shape and proportions. The dorsal (ventral)
length of the centrum is the greatest distance between
the cranial and caudal edges of the dorsal (ventral) sur-
face of the centrum. The height and breadth of the cra-
nial and caudal surfaces of the centrum coincide with
the vertical and transverse diameters of the centrum.
The medial length and height of the centrum are calcu-
lated as half the sum of the dorsal and ventral lengths
and the cranial and posterior heights of the centrum,
respectively. The length of the neurapophysis (Fig. 2, 1)
corresponds to the vertebral length at the pre- and
postzygapophyses. The height of the buttresses of the
neural arch (Fig. 2, 5) is measured on the posterior side
of vertebra from the suture between the centrum and the
neurapophysis to the base of the neural spine. The pro-
jection of the pre- and postzygapophyses (Fig. 2, 2, 3)
are connected with the depth of the cranial and caudal
incisures, respectively. The inclination of the vertebral–
costal articulation (10) is determined based on the incli-
nation of the line connecting the centers of the facets of
the parapophysis and diapophysis relative to the longi-
tudinal axis of the centrum (conventional horizontal,
O–O', the posterior angle is measured). The angle of
articulation of vertebral structures with the pelvis, the
sacropelvic articulation, is measured similarly, i.e.,
based on the centers of contact areas of the diapophysis
and sacral rib or on the facets on the medial surface of
the ilium. The inclination of the prezygapophyses to the
horizontal in the sagittal plane (13) corresponds to the
angle between the line extending along the long axis of
the prezygapophysis and the longitudinal axis of the
centrum. The inclination of the diapophyses to the hor-
izontal in the transverse plane (14) coincides with the
angle between the horizontal and the line extending
along the long axis of the diapophysis. The angle of
divergence of postzygapophyses in the frontal plane (12)
is formed by their long axes. The height of the neural
spine is measured from its base to the apex on the pos-
terior side of the vertebra (4).
KEY TO VERTEBRAE IN REGARD TO THEIR
POSITION IN THE VERTEBRAL COLUMN
1 (4) The parapophysis and diapophysis are isolated
from each other and articulated with a two-headed rib…
………presacral vertebrae (except for the atlas and the
posterior lumbar vertebra);
2 (3) The parapophysis is on the centrum, the diapo-
physis is on the neurapophysis; the centroneurapophy-
sial suture is usually below the upper margin of the cen-
trum; the anterior breadth of the centrum is approxi-
mately equal to the posterior breadth (it is occasionally
greater than the posterior breadth because the parapo-
180
PALEONTOLOGICAL JOURNAL Vol. 41 No. 2 2007
TERESCHENKO
physis projects as an anteriorly positioned process); the
prezygapophyses project about twice as far as the
postzygapophyses; the lateral intervertebral foramen is
extended somewhat dorsoventrally……cervical region;
3 (2) The diapophysis and parapophysis are on the
neurapophysis; the centroneurapophysial suture is in
line with the upper margin of the centrum; the anterior
breadth of the centrum is greater than the posterior
breadth:
a (b) The diapophyses are relatively long; the pro-
cesses of the dia- and parapophyses are usually fused in
a united paradiapophysial process, but their facets are
always separate (the distance between them is greater
than the length of the parapophysial facet), the postzyg-
apophyses project approximately 2.5 times as far as the
prezygapophyses; the lateral intervertebral foramen is
circular…………………………………thoracic region;
b (a) The diapophyses are relatively short; the facets
of the diapophysis and parapophysis are positioned
extremely close to each other (the distance between
them is at most the same as the length of the parapophy-
sial facet) or even fused in a united synapophysial facet;
the postzygapophyses project approximately 1.5–2 times
as far as the prezygapophyses……………lumbar region;
4 (1) The facet for the single-headed rib is located
between the neurapophysis and the centrum (or, less
often, between the centra of two adjacent vertebrae; in
this case, the facet is divided into the caudal semifacet
of the preceding vertebra and the cranial semifacet of
the succeeding vertebra)………sacral–caudal vertebrae;
a (b) The facets for the rib, which is connected to the
ilium, vary widely in size, are located on the centrum
and neurapophyses; the diapophyses vary in length; the
parapophyses are not in the shape of processes; some
vertebrae have two semifacets for the parapophysial
part of rib; the cranial and caudal surfaces of the cen-
trum are usually coarse……………………sacral region;
b (a) the facet for the rib is relatively small,
restricted to the suture between the centrum and
neurapophysis, the diapophyses are absent, but ribs
(a) (b) (c)
HaOO'
1
2
10
Lv
36
13
4
5
7
Ba
Bp
Ha
15
14
(d) (e) (f)
11
9128
Fig. 2. Protoceratopid vertebrae, measurements of vertebral elements: (a–d) Protoceratops andrewsi Granger et Gregory, 1923:
(a) specimen PIN, no. 614/35, 8c; (b–d) specimen PIN, no. 3143/7, 3lm; (e, f) Udanoceratops tschizhovi Kurzanov, 1992, specimen
PIN, no. 3907/11, 9th; (a, b, e) left lateral view, (c) posterior view, (d) dorsal view, and (f) frontal view. Designations: (1) length of
the neurapophysis, (2) length of the projection of the postzygapophysis, (3) length of the projection of the prezygapophysis,
(4) height of the neural spine, (5) height of the buttresses of the neural arch, (6) length of the base of the neural spine, (7) length of
the projection of the synapophysis, (8) vertebral breadth at the prezygapophyses, (9) vertebral breadth at the postzygapophyses,
(10) angle of inclination of the vertebral–costal articulation to the conventional horizontal, (11) length of the facet of the parapo-
physis, (12) angle of divergence of the postzygapophyses, (13) angle of inclination of the prezygapophysis to the horizontal in sag-
ittal plane, (14) angle of inclination of the diapophysis to the horizontal in transverse plane, (15) angle of inclination of the facet of
the prezygapophysis to the horizontal in transverse plane, (O–O') longitudinal axis of the centrum, (Lv) ventral length of the centrum,
(ça) anterior height of the centrum, (Ça) anterior breadth of the centrum, and (Çp) posterior breadth of the centrum. Scale bar, 20 mm.
PALEONTOLOGICAL JOURNAL Vol. 41 No. 2 2007
KEY TO PROTOCERATOPOID VERTEBRAE (CERATOPSIA, DINOSAURIA) FROM MONGOLIA 181
connected to the centrum and neurapophysis are
present (in the case of fusion, they are in the shape of
costal processes); in the posterior half of the region,
ribs are sometimes reduced; the ventral surface of the
centra usually has two crests, which expand caudally to
form a pair of facets for the haemal arch…caudal region.
Cervical Region (Figs. 1a, 1b, 2a, 3a–3d)
Three anterior cervical vertebrae of adults are fused
(Fig. 3a).
1 (2) The vertebra is very short (at most half as long
as other cervical vertebrae) and low, lacks a neural
spine, has a large, deep fossa for articulation with the
occipital condyle; the centrum is rounded ventrally; the
neurapophysis is represented by two small elements,
each with an arched narrow process directed dorsopos-
teriorly to the anterior margin of the buttress of the epis-
tropheal arch…………………………………1c (atlas).
2 (1) The vertebrae have well-developed neurapo-
physes, with postzygapophyses and neural spine, elon-
gated centrum, the length of which is equal to, or
greater than, the height; the centrum of each vertebra
has more or less developed ventral crests; the parapo-
physis and diapophysis are separate and project as pro-
cesses.
3 (4) The neurapophysis is very high, approximately
1.5 times higher than the centrum; the anteroposterior
breadth of the neural spine is greater than the length of
the centrum; prezygapophyses are absent, only the
anterior margin of the upper part of the buttress of the
neurapophysis has a small facet for articulation with the
process of the neurapophysis of the atlas; the interzyg-
apophysial crest is very weak; the process of the
parapophysis is better developed than the diapophysis…
………………………………………2c (epistropheus).
4 (3) The pre- and postzygapophyses are well-devel-
oped; the neural spine is narrow, its breadth is less than
the length of the centrum; the interzygapophysial crest
is either present or absent.
5 (6) The diapophysis and parapophysis project
approximately equally from the vertebral surface; the
parapophysis is located at the anterior edge of the cen-
trum, the diapophysis is at the level of the middle of the
centrum near the centroneurapophysial suture, so that
the vertebral–costal articulation is inclined to the con-
ventional horizontal at an angle of about 55°; the diapo-
physis is flattened dorsoventrally; the center of the base
of the diapophysis is at the level of the bottom of the
spinal canal; the interzygapophysial crest is well devel-
oped…………………………………………………3c.
6 (5) The diapophysis projects to a greater extent
than the parapophysis and is positioned on the buttress
(a) (b) (c) (d) (e)
(f) (g) (h) (i) (j)
Fig. 3. Cervical and thoracic vertebrae of protoceratopoids: (a, c) Bagaceratops sp., specimen PIN, no. 4550/3: (a) 1–3c, and (c) 6c;
(b, e, f, i, j) Protoceratops andrewsi Granger et Gregory, 1923, specimen PIN, no. 614/35: (b) 6c and (e, f) 3th; (i, j) specimen PIN,
no. 3143/7, 9th; (d) Udanoceratops tschizhovi Kurzanov, 1992, specimen PIN, no. 3907/11, 7c; and (g, h) Bagaceratops sp., spec-
imen, PIN, no. 614/29, 3th; (a, e, g, i) left lateral view, (b–d) dorsal view, and (f, h) frontal view. Scale bar, 10 mm.
182
PALEONTOLOGICAL JOURNAL Vol. 41 No. 2 2007
TERESCHENKO
of the neurapophysis above the centroneurapophysial
suture (somewhat above the bottom of the spinal canal);
the inclination of the vertebral–costal articulation to the
horizontal is about 90°.
7 (16) The ventral length of the centrum is greater
than the dorsal length; the facet of the parapophysis is
at the level of the midheight of the centrum and does not
expand onto the neurapophysis; the facets of the pre-
and postzygapophyses are at approximately the same
level relative to the horizontal; in dorsal view, the
diapophysis is positioned close to the line of the
prezygapophysis; in transverse plane, it is inclined at
more than 30° to the horizontal; in cross section, the
diapophysis is in the shape of a dorsoventrally extended
oval………………………………………………4–8c.
8 (13) The neural spine is relatively narrow (in sag-
ittal plane); in cross section, its base is wedge-shaped,
with its center located close to the midlength of the
neurapophysis; the parapophysis projects only slightly;
the diapophysis is relatively long and, in the transverse
plane, is directed ventrolaterally or laterally; the long
axis of the facet of the diapophysis is inclined posteri-
orly and ventrally at less than 80° to the horizontal; the
supraprediapophysial crest is present; the buttresses of
the neural arch cover the upper third of the centrum; the
vertebral breadth at the prezygapophyses is equal to, or
less than, the breadth at the postzygapophyses; the
neurapophysis has a well-developed interzygapophys-
ial crest……………………………………………4–6c.
9 (12) The diapophysis is directed ventrolaterally;
the center of the base of the diapophysis is at the level
of the midheight of the buttress of the neurapophysis;
the pre- and postzygapophyses project approximately
equally; the vertebral breadth at the prezygapophyses is
less than the breadth at the postzygapophyses; the neu-
ral spine is directed anteriorly, it is at most as high as
the vertebral centrum.
10 (11) The facet of the parapophysis is at the level
of the midheight of the centrum; the vertebral–costal
articulation is inclined at about 70° to the horizontal;
the diapophysis deviates ventrally relative to the hori-
zontal at an angle of 8°–10°; the center of the base of
the diapophysis is below the midheight of the buttress
of the neurapophysis; the long axis of the facet of the
diapophysis is inclined at about 10° to the horizontal;
the neural spine is approximately as high as the verte-
bral centrum; the neural spine narrows sharply dorsally
………………………………………………………4c.
11 (10) The facet of the parapophysis is above the
midheight of the centrum; the vertebral–costal articula-
tion is inclined at about 80° to the horizontal; the diapo-
physis is inclined ventrally at 5° to the horizontal; the
center of the base of the diapophysis is at the midheight
of the buttress of the neurapophysis; the axis of the
facet of the diapophysis is inclined at approximately
30° to the horizontal; the neural spine is approximately
half as high as the vertebral centrum…………………5c.
12 (9) The diapophysis is directed laterally; the cen-
ter of the base of the diapophysis is above the midheight
of the buttress of the neurapophysis; the axis of the
facet of the diapophysis is inclined at 50°–60° to the
horizontal; the postzygapophysis projects approxi-
mately 1.5 times greater than the prezygapophysis; the
vertebral breadth at the prezygapophyses is equal to the
breadth at the postzygapophyses; the vertebral–costal
articulation is directed vertically; the neural spine is
approximately half as high as the centrum; the neural
spine is directed dorsally……………………………6c.
13 (8) The neural spine is relatively wide, the cross
section of its base is an extended and anteriorly nar-
rowed oval, the center of the base is located caudal to
the midlength of the neurapophysis; the neural spine is
approximately half as high as the vertebral centrum; the
facet of the parapophysis is located directly on the cen-
trum; the diapophysis is directed dorsolaterally; the
prediapophysial crest is present instead of the suprapre-
diapophysial crest; the axis of the facet of the diapophy-
sis is oriented almost vertically; the buttresses of the
neurapophysis cover the centrum dorsally over its
upper third; the postzygapophysis projects approxi-
mately two times greater than the prezygapophysis; the
vertebral breadth at the prezygapophyses is greater than
at the postzygapophyses; the interzygapophysial crest
is weak or absent…………………………………7–8c.
14 (15) The ventral length of the centrum is approx-
imately 1.3 times greater than the dorsal length; the
diapophysis projects dorsolaterally at less than 30° to
the horizontal; in dorsal view, the base of the diapophy-
sis is approximately in line with the facet of the prezyg-
apophysis; the interzygapophysial crest is poorly pro-
nounced; the axis of the vertebral–costal articulation is
oriented vertically; the neural spine is directed dorsally
………………………………………………………7c.
15 (14) The ventral length of the centrum is less than
1.3 times greater than the dorsal length; the diapophysis
projects at about 30° to the horizontal; the base of the
diapophysis is in line with the caudal two-thirds of the
facet of the prezygapophysis, including the postfacet
space; the postdiapophysial crest is present; the verte-
bral–costal articulation is inclined at an angle of 80° to
the horizontal; the neural spine is inclined posteriorly…
………………………………………………………8c.
16 (7) The ventral length of the centrum is equal to
the dorsal length; the facet of the parapophysis partially
expands onto the neurapophysis; the facet of the
postzygapophyses is located substantially higher than
that of the prezygapophyses; the diapophysis is posi-
tioned somewhat posterior to the prezygapophysis and
is inclined dorsolaterally at more than 30° to the hori-
zontal; infra-, prezyga-, and postdiapophysial crests are
well-developed, so that the cross section of the diapo-
physis is subtriangular; the neural spine is approxi-
mately as high as the vertebral center; the neural spine
is inclined posteriorly…………………………………9c.
PALEONTOLOGICAL JOURNAL Vol. 41 No. 2 2007
KEY TO PROTOCERATOPOID VERTEBRAE (CERATOPSIA, DINOSAURIA) FROM MONGOLIA 183
Thoracic Region (Figs. 1c, 1d, 2e, 2f, 3e–3j)
The distance between the facets of the parapophysis
and diapophysis is equal to, or greater than, the length
of the facet of the parapophysis; the postzygadiapophy-
sial crest is present; the neural spine is approximately
1.5 times as high as the vertebral centrum.
1 (8) The distance between the facets of the parapo-
physis and diapophysis varies widely, but is always
greater than the length of the facet of the parapophysis;
the facet of the parapophysis is located at the base of the
paradiapophysial process, or is displaced to the middle
of this process; the ventral length of the centrum is
greater than, or equal to, the dorsal length; the postzyg-
adiapophysial crest passes into the lateral margin of the
postzygapophysis at the midlength of the base of the
neural spine or somewhat caudally; the paradiapophys-
ial crest is present; the postzygapophyses project more
than 2.5 times as far as the prezygapophyses; the diapo-
physis is triangular in cross section, with a wide dorsal
surface……………………………………………1–5th.
2 (3) The center of the facet of the parapophysis is
positioned lower than the interprezygapophysial inci-
sure; the ventral surface of the centrum has a crest; the
ventral length of the centrum is greater than the dorsal
length; the facets of the postzygapophyses are located
markedly higher than the prezygapophyses; the verte-
bral–costal articulation is inclined at more than 50° to
the horizontal………………………………………1th.
3 (11) The center of the facet of the parapophysis is
above the level of the interprezygapophysial incisure,
while the facets of the postzygapophyses are located
somewhat higher than the prezygapophyses………2–7th.
4 (5) The ventral length of the centrum is greater
than the dorsal length; the base of the neural spine is
less than half as long as the neurapophysis; the facet of
the prezygapophysis is smooth, while the facet of the
postzygapophysis has surface sculpturing (one or two
longitudinal ridgelike crests are present)……………2th.
5 (4) The ventral length of the centrum is approxi-
mately equal to the dorsal length; the base of the neural
spine is more than half as long as the neurapophysis;
the facets of the prezygapophyses and postzygapophy-
ses have surface sculpturing………………………3–5th.
6 (7) The postzygapophyses project approximately
twice as far as the prezygapophyses;
a (b) The paradiapophysial and ventral crests on the
centrum are well pronounced………………………3th;
b (a) The paradiapophysial crest is sometimes weak;
the ventral crest on the centrum is hardly discernible;
the length of the facet of the prezygapophysis is less
than its breadth………………………………………4th.
7 (6) The postzygapophyses project approximately
2.5 times as far as the prezygapophyses; the paradiapo-
physial crest is sometimes hardly discernible; the ven-
tral crest on the centrum is absent……………………5th.
8 (1) The distance between the facets of the parapo-
physis and diapophysis is reduced, at most as long as
the facet of the parapophysis because the facet of the
parapophysis is located on the paradiapophysial pro-
cess close to its end; the ventral length of the centrum
is less than the dorsal length; the postzygadiapophysial
crest passes into the lateral margin of the postzygapo-
physis anterior to the midlength of the base of the neu-
ral spine; the paradiapophysial crest is absent; the pre-
diapophysial crest is present; the postzygapophyses
project at most 2.5 times as far as the prezygapophyses;
the diapophysis is circular in cross section……6–9(10)th.
9 (12) The facets of the pre- and postzygapophyses
have surface sculpturing composed of one or two longi-
tudinal ridgelike crests.
10. The postzygapophyses project less than
2.5 times as far as the prezygapophyses…………6–7th.
11 (3) The center of the facet of the parapophysis is
positioned in line with the interprezygapophysial inci-
sure, while the facets of the postzygapophyses are
approximately in line with the prezygapophyses; the
postzygapophyses project approximately 2.5 times as
far as the prezygapophyses……………………8(8–9)th.
12 (9) The facet of the prezygapophysis has sculp-
turing, while the facet of the postzygapophysis is
smooth; the center of the facet of the parapophysis is
below the interprezygapophysial incisure………9(10)th.
Lumbar Region (Figs. 1e, 1f; 2b–2d)
1 (2) The distance between the facets of the parapo-
physis and diapophyses is less than the length of the
facet of the parapophysis, i.e., the head and the tubercle
of the shortened rib are pulled close together; the ven-
tral surface of the centrum is smooth………………1lm.
2 (1) The facets of the parapophysis and diapophysis
are fused or only slightly isolated; the ventral surface of
the centrum has a crest.
3 (4) The facets of the parapophysis and diapophysis
come into contact, the first is located slightly ventral to
the second…………………………………………2lm.
4 (3) The neurapophysis has a synapophysis with a
single circular facet for a single-headed rib; the facet of
the prezygapophyses exceeds in area that of the postzyg-
apophyses; the postzygapophyses project approximately
1.5 times as far as the prezygapophyses; the spinopo-
stzygapophysial crest passes on the caudal surface of
the neural spine and postzygapophysis, terminating
short of the margin of the facet to form a small projec-
tion or a short process………………………………3lm.
Sacral Region (Figs. 4a–4f)
Three anterior sacral vertebrae of adults are usually
fused by their centra and zygapophyses, the other ver-
tebrae are only fused by either centra or zygapophyses,
depending on the individual age and taxon. The verte-
bral elements and vertebrae of very young animals
remain separate.
184
PALEONTOLOGICAL JOURNAL Vol. 41 No. 2 2007
TERESCHENKO
1 (8) The vertebral centra are relatively short and
wide (the anterior breadth of the centrum is equal to, or
greater than, the length); the ventral crest is weak or
absent; the vertebral breadth at the prezygapophyses is
approximately 1.5 times greater than at the postzygapo-
physes; the diapophysis is positioned at the same height
as the prezygapophysis; the neural spine is directed
strictly vertically………………………………1–4(3)sc.
2 (3) The vertebrae have only strongly reduced (or
shortened) “upper” ribs, which can come into contact
with the ilium, although it lacks articular facets; verte-
brae are only connected to the ilium through the diapo-
physes; the anterior breadth of the centrum is less than
the posterior breadth; the middle constriction of the
vertebral centrum is weak……………………1–2(1)sc;
a (b) The caudal surface of the centrum may be
coarse, the centrum lacks a facet for the sacral rib……
………………1sc (Protoceratops and Bagaceratops);
b (a) The cranial surface of the centrum may be
smooth, but the caudal surface is coarse; on the poste-
rior side, the centrum has a caudal semifacet for the sac-
ral rib of the succeeding vertebra…………………2(1)sc.
3 (2) The vertebrae have sacral ribs and a double
articulation with the ilium through the elongated diapo-
physes, the facet of which is displaced ventrally and
combined with the increased facet of the parapophysis;
each centrum has cranial and caudal semifacets; the
proximal end of the rib is connected ventrally to the
centra of two adjacent vertebrae; dorsally, it is con-
nected (without a space) to the buttress of the neural
arch up to the diapophysis of the second vertebra of a
pair; the centrum is strongly constricted at the middle
(the cranial or posterior breadth of the centrum is 1.5–
2 times greater than the minimum breadth); the facet for
the rib on the centrum is restricted to its cranial half; the
anterior breadth of the centrum is greater than, or equal
to, the posterior breadth; the cranial and caudal surfaces
of the centrum are rough…………3–6(2–5)sc (see 10).
4 (9) The sacropelvic articulation is inclined at more
than 90° to the horizontal (the diapophysis adjoins the
ilium more cranially than the rib); the anterior breadth
of the centrum is approximately equal to the posterior
breadth……………………………………3–4(2–3)sc.
5 (6) In lateral view, the lower border of the suture
between the rib and the centrum almost coincides with
the ventral surface of the centrum; the sacropelvic artic-
ulation is inclined at 130°–135° to the horizontal; the
ribs are directed slightly posteriorly………………3(2)sc.
6 (5) A different position of the border of the suture
between the rib and the centrum; the ribs are directed
laterally; the anterior breadth of the centrum is approx-
imately equal to its length…………4(3)–7sc (see 10b).
7. The lower border of the suture between the rib
and the centrum is lower than the midheight of the cen-
(a) (b) (c) (d) (e)
(f) (g) (h) (i) (j) (k)
prz
fpcd
Fig. 4. Sacral and caudal vertebrae of protoceratopoids: (a–c, j) Protoceratops andrewsi Granger et Gregory, 1923, specimen PIN,
no. 3143/7: (a–c) 1sc and (j) 3cd; (d, e) Bagaceratops sp., specimen PIN, no. 614/29, 2sc; (f–i, k) Udanoceratops tschizhovi
Kurzanov, 1992, specimen PIN, no. 3907/11: (f) 8sc, (g–i) 1cd, and (k) ?13cd; (a, d, j) left lateral view, (f, g, k) right lateral view,
(b, h) posterior view, (c) dorsal view, (e) frontal view, (i) ventral view. Designations: (fpcd) facies parapophysi caudalis; for other
designations, see Fig. 1. Scale bar, 20 mm.
PALEONTOLOGICAL JOURNAL Vol. 41 No. 2 2007
KEY TO PROTOCERATOPOID VERTEBRAE (CERATOPSIA, DINOSAURIA) FROM MONGOLIA 185
trum; the sacropelvic articulation is inclined at 100°–
105° to the horizontal……………………………4(3)sc.
8 (1) The vertebral centra are relatively long (the
length of the centrum is greater than the anterior
breadth); the ventral crest is frequently well pro-
nounced; the vertebral breadth at the prezygapophyses
is less than 1.5 times greater than at the postzygapophy-
ses; the diapophysis is short (or even absent) and posi-
tioned lower than the prezygapophysis; the neural
spines are inclined caudally; the lower border of the
suture between the rib and the centrum is at the level of
the midheight of the centrum…………………5(4)–8sc.
9 (4) The sacropelvic articulation is positioned at
less than 90° to the horizontal (the diapophysis adjoins
the ilium more caudally than the rib)……………(see 10);
a (b) The sacropelvic articulation is inclined at about
80° to the horizontal……………………………5(4)sc;
b (a) The sacropelvic articulation is inclined at about
45° to the horizontal………………………………6(5)sc.
10 (3) The vertebrae are only connected to the ilium
through single-headed ribs; the diapophyses are
reduced; the centrum has one costal facet, which is
shifted somewhat caudally along the centrum; the cen-
tra are weakly constricted at the middle; the anterior
breadth of the centra is less than the posterior breadth…
…………………………………………………(see 9);
a (b) The costal facet is in the middle part of the ver-
tebral centrum; the ribs are directed laterally; the cra-
nial surface of the centrum is rough, while the caudal
surface is smooth……………………………7(6–7)sc;
b (a) The costal facet is in the caudal part of the cen-
trum; the ribs are directed somewhat cranially; the ante-
rior breadth of the centrum is greater than its length; the
cranial and caudal surfaces of the centrum are smooth…
………………………………………………8sc (see 6).
Caudal Region (Figs. 2g, 2h, 4g–4k)
1 (11) The costal process is well developed, the cen-
ter of its base coincides with the midlength of the cen-
trum; the height of the centrum is less than, or equal to,
the breadth; the pre- and postzygapophyses project
approximately equally; in the transverse plane, the
facet of the prezygapophyses is inclined at less than 75°
to the horizontal; the neural spine (measured from a
point somewhat above its base) is craniocaudally at
least half as long as the neurapophysis; the center of the
base of the neural spine is approximately at the level of
the caudal third of the length of the centrum; the neural
spines are approximately twice as high as the vertebral
centrum…………………………………………1–11cd.
2 (7) The vertebral centra are inversely heterocelous
(see Tereschenko, 2004a); the ventral length of the cen-
trum is less than the dorsal length; the height of the cen-
trum is usually less than its breadth; the vertebral
breadth at the prezygapophyses is equal to, or less than,
the breadth at the postzygapophyses; the lower border
of the suture between the rib and the centrum is at the
level of the midheight of the centrum, or slightly above;
in transverse plane, the facet of the prezygapophyses is
inclined at less than 65°–70° to the horizontal; the cos-
tal processes project for at least half the anterior breadth
of the centrum; the neurapophysis is at most 1.5 times
as long as the centrum; the neural spines are flattened
laterally; the neural spine is craniocaudally at least half
as long as the neurapophysis……………………1–5cd.
3 (4) The vertebral breadth at the prezygapophyses
is less than the breadth at the postzygapophyses; the
prezygapophysis projects for less than half the length of
its facet; the lower border of the suture between the rib
and the centrum is at the level of the midheight of the
centrum; the neurapophysis is less than 1.5 times as
long as the centrum; the ventral surface of the centrum
is smooth; in transverse plane, the facet of the prezyga-
pophyses is inclined at 50°–55° to the horizontal…1–2cd.
4 (3) The vertebral breadth at the prezygapophyses
is greater than, or equal to, the breadth at the postzyga-
pophyses; the prezygapophysis projects for half the
length of its facet; the lower border of the suture
between the rib and the centrum is slightly above the
midheight of the centrum; the neurapophysis is approx-
imately 1.5 times as long as the centrum; the centrum
occasionally has two ventral crests; in transverse plane,
the facet of the prezygapophyses is inclined at about
60° to the horizontal………………………………3–5cd.
5 (6) The vertebral breadth at the prezygapophyses
is equal to the breadth at the postzygapophyses……3cd.
6 (5) The vertebral breadth at the prezygapophyses
is greater than at the postzygapophyses;
a (b) If the centrum has two ventral crests, the facets
for the haemal arch are poorly pronounced…4(4–5) cd;
b (a) The centrum has two well-developed ventral
crests and facets for the haemal arch………………5cd.
7 (2) The articular surfaces of the centra are slightly
concave; the ventral length of the centrum is equal to
the dorsal length; the height of the centrum is equal to
its breadth; the vertebral breadth at the prezygapophy-
ses is greater than at the postzygapophyses; the lower
border of the suture between the rib and the centrum is
above the midheight of the centrum; the neural spines
are oval in cross section.
8. The costal processes project for a distance of at
most half of the anterior breadth of the centrum; the
neurapophysis is 1.5–2 times as long as the centrum;
the neurapophysis is 2–2.5 times as long as the neural
spine.
9 (10) The lower border of the suture between the rib
and the centrum is at the level of the upper third of the
height of the centrum; the prezygapophysis projects for
a distance of slightly more than half the length of its
facet; in transverse plane, the facets of the prezygapo-
physes are inclined at 65°–70° to the horizontal; the cos-
tal processes project for a distance of half of the ante-
rior breadth of the vertebral centrum………………6cd.
186
PALEONTOLOGICAL JOURNAL Vol. 41 No. 2 2007
TERESCHENKO
10 (9) The suture between the rib and the centrum
passes onto the dorsolateral surface of the centrum; the
prezygapophysis projects for almost the entire length of
the facet; the costal processes project for less than half
of the anterior breadth of the centrum………7–11(12)cd.
11 (1) The costal processes are weakly developed or
absent, the centers of their bases are located somewhat
posterior to the midlength of the centrum; the height of
the centrum is greater than its breadth; the postzygapo-
physes occasionally project to a greater extent than the
prezygapophyses of the same vertebra; in transverse
plane, the facets of the prezygapophyses are inclined at
more than 75° to the horizontal; the neurapophysis is
3−4 times as long as the neural spine; the center of the
base of the neural spine is close to the line of the caudal
articular surface of the centrum……………12–35(40)cd.
12 (15) The costal processes are poorly developed;
the postzygapophyses project to a greater extent than
the prezygapophyses; the neurapophysis is longer than
the centrum; the facet of prezygapophyses is inclined in
the transverse plane at an angle of 75°–80° to the hori-
zontal; the neurapophysis is approximately three times
as long as the neural spine; the neural spines are 2.5–
4 times as high as the centrum……………12–22(20)cd.
13 (14) The postzygapophyses project approxi-
mately 1.5 times as far as the prezygapophyses; the
neurapophysis is 1.5–1.4 times as long as the centrum…
………………………………………………12–18cd.
14 (13) The postzygapophyses project less than
1.5 times as far as the prezygapophyses; the neurapo-
physis is 1.3–1.2 times as long as the centrum…………
……………………………………………19–22(20)cd.
15 (12) The costal processes are absent; the zygapo-
physes are rudimentary and lack facets; the prezygapo-
physes and postzygapophyses project approximately
equally; the prezygapophyses are inclined in the trans-
verse plane at 85°–90° to the horizontal; the neurapo-
physis is shorter than, or equal to, the centrum; the
neurapophysis is almost four times as long as the neural
spine; the neural spines are at most two times as high as
the centrum……………………………23–40(21–50)cd;
a (b) The centrum is approximately as long as the
neurapophysis…………………………23–35(21–40)cd;
(a) The centrum is longer than the neurapophysis…
………………………………………36–40 (41–50)cd.
DISTINCTIONS IN VERTEBRAE BETWEEN
THE TWO PROTOCERATOPOID FAMILIES
Protoceratopidae
(Figs. 1a, 1c–1h, 2, 3b, 3d–3f, 3i, 3j, 4a–4c, 4f–4k)
Presacral vertebrae. The centra are short and low
(the ratios of the mean length and height of the centrum
to the anterior breadth range from 0.5 to 1.1 and from
0.8 to 1.2, respectively). The neural arch is short (the
ratio of the neurapophysial length to the vertebral
breadth at the postzygapophyses ranges from 0.9 to 2.2)
and high, particularly, in the anterior part of the thoracic
region, where the height of its buttresses is on average
about 50% of the total height of the centrum and spinal
canal. The postzygapophyses are usually positioned at
an angle of more than 90°. The cranial surface of the
centrum of 1–5(6)c is slightly extended transversely,
while that of 1–6(7)th is circular.
Cervical vertebrae. The centrum of the axis (epist-
ropheus) is 1.3 times as long as that of the atlas. The
centrum of 2c is slightly rhomboidal in lateral view; the
postparapophysial crest, with the laterocaudal projec-
tion, is well developed; the diapophysis is positioned in
line with, or anterior to the midlength of the neurapo-
physial buttresses; the postdiapophysial crest is fused
with the lateral surface of the neurapophysis at the level
of the midheight of the buttress of the neural arch, or
below this level; a groove extends along the entire cau-
dal surface of the neural spine. 3c has well-pronounced
paradiapophysial and supradiapophysial crests. The
neural spines of 2c and 3c are equal in height. The
diapophyses of 6c are directed laterally or caudolater-
ally. The spinopostzygapophysial crest of 7–9c is well
developed.
Dorsal vertebrae. The center of the parapophysial
facet of 1th is at the level of, or above the midheight of
the buttress of the neurapophysis. In 3th, the vertebral
breadth at the prezygapophyses is the same as at the
postzygapophyses; the vertebral–costal articulation is
inclined at an angle of 40°–60° to the horizontal. The
postzygadiapophysial crest of 7(8)th passes into the
surface of the postzygapophysis at the midlength of the
base of the neural spine. The centrum of 3lm is circular
or subtriangular; the strongly concave facet of the syn-
apophysis of this vertebra is smooth or coarse.
Sacral and caudal vertebrae. The lower margins of
the right and left facets of the postzygapophyses of 8sc
are either widely spaced or fused to form a crest. The
heterocelous pattern is occasionally poorly pronounced
in the centra of 1–5cd. The centra decrease in size from
7cd to 11cd, while the neurapophysis retains almost the
same length. The neural spines of 12–20cd are high
(2.5–3 times as high as the centrum); the haemal pro-
cesses are relatively low (at most 1.2 times as high as the
centrum and the buttress of the neural arch taken
together); the cross section at the midheight of the neural
spine of these vertebrae is circular or laterally flattened.
Bagaceratopidae (Figs. 2b, 3a, 3c, 3g, 3h, 4d, 4e)
Presacral vertebrae. The centra are long and high
(the ratios of the mean length and mean height of the
centrum to its anterior breadth range from 0.8 to 1.3 and
from 1.0 to 1.6, respectively). The neural arch is long
(the ratio of the neurapophysial length to the vertebral
breadth at the postzygapophyses ranges from 1.1 to 2.6)
and low, particularly, in the anterior part of the thoracic
region, where the buttresses of the neural arch are about
45% of the total height of the centrum and spinal canal.
PALEONTOLOGICAL JOURNAL Vol. 41 No. 2 2007
KEY TO PROTOCERATOPOID VERTEBRAE (CERATOPSIA, DINOSAURIA) FROM MONGOLIA 187
The postzygapophyses of the presacral vertebrae are
usually diverge at an angle of at most 90°. The cranial
surface of the centra of 1–5(6)c is circular, while that of
1–6(7)th is dorsoventrally extended.
Cervical vertebrae. The centrum of the axis is
1.5 times as long as that of the atlas. The centrum of 2c
is square in lateral view; the postparapophysial crest,
with the laterocaudal projection, is hardly discernible;
the diapophyses are located posterior to the midlength
of the neurapophysial buttress; the postdiapophysial
crest goes up and is fused with the laterocaudal surface
above the midheight of the buttress of the neural arch;
a groove on the caudal surface of the neural spine is
only seen within the lower half of the spine. 3c lacks
paradiapophysial and supradiapophysial crests. The
neural spine of 3c is lower than that of 2c. The diapo-
physes of 6c project craniolaterally. 7–9c lack a spino-
postzygapophysial crest.
Dorsal vertebrae. The center of the parapophysial
facet of 1th is at the level of, or below the midheight the
buttress of the neurapophysis. In 3th, the vertebral
breadth at the prezygapophyses is slightly less than at
the postzygapophyses; the vertebral–costal articulation
is inclined at about 30° to the horizontal. The postdi-
apophysial crest of 7(8)th passes into the surface of the
postzygapophyses cranial to the midlength of the base
of the neural spine. The centrum of 3lm is heart-shaped;
the slightly concave facet of the synapophysis of this
vertebra is always smooth.
Sacral and caudal vertebrae. The lower margins of
the right and left facets of the postzygapophyses of 8sc
are closely spaced, but do not fuse into a crest. The het-
erocelous pattern of 1–5cd is well pronounced. The
centra decrease in size from 7cd to 11cd, while the
neurapophyses increase, becoming approximately
1.7 times as long as the centrum. The neural spines on
12–20cd are very high (at least three times as high as
the centrum); the haemal processes are moderately high
(1.5 times as high as the centrum and buttress of the
neural arch taken together); the cross section at the mid-
height of the neural spines of these vertebrae usually is
craniocaudally flattened.
ACKNOWLEDGMENTS
I am grateful to V.B. Sukhanov (PIN) for supervis-
ing this research and to A.N. Kuznetsov (Zoological
Museum of Moscow State University) for valuable dis-
cussions and perusal of the manuscript.
This study was supported by the Russian Founda-
tion for Basic Research (project no. 04-04-48829) and
the Russian State Program for Support of Leading Sci-
entific Schools (project nos. NSh-1840.2003.4 and
NSh-6228.2006.4).
REFERENCES
1. V. R. Alifanov, “Macrocephalosaurs and Early Evolu-
tionary Stages of Lizards of Central Asia,” Tr. Paleontol.
Inst. Ross. Akad. Nauk 272, 1–126 (2000).
2. V. R. Alifanov, “Two New Dinosaurs of the Infraorder
Neoceratopsia (Ornithishia) from the Upper Cretaceous
of the Nemegt Depression, Mongolia,” Paleontol. Zh.,
No. 5, 77–88 (2003) [Paleontol. J. 37 (5), 524–534
(2003)].
3. D. B. Brinkman, M. J. Ryan, and D. A. Eberth, “The
Paleogeographic and Stratigraphic Distribution of Cer-
atopsids (Ornithischia) in the Upper Judith River Group
of Western Canada,” Palaios 13 (2), 160–169 (1998).
4. B. Brown and E. M. Schlaikjer, “The Structure and Rela-
tionship of Protoceratops,” Ann. New York Acad. Sci. 40
(3), 133–265 (1940).
5. B. Brown and E. M. Schlaikjer, “The Skeleton of Lepto-
ceratops with the Description of a New Species,” Am.
Mus. Novit., No. 1169, 1–15 (1942).
6. P. Dodson and P. Currie, “Neoceratopsia,” in The Dino-
sauria, Ed. by D. B. Weishampel, P. Dodson, and
H. Osmólska (Univ. California Press, Berkeley, 1990),
pp. 593–618.
7. Zh.-M. Dong and Y. Azuma, “On a Primitive Neocer-
atopsian from the Early Cretaceous of China,” in Sino–
Japanese Silk Road Dinosaur Expedition, Ed. by
Zh.-M. Dong (China Ocean Press, Beijing, 1997),
pp. 68–89.
8. F. Ya. Dzerzhinsky, Comparative Anatomy of Vertebrates
(Mosk. Gos. Univ., Moscow, 1998) [in Russian].
9. N. N. Gurtovoi, B. S. Matveev, and F. Ya. Dzerzhinsky,
Practical Zootomy of Vertebrates: Amphibians and Rep-
tiles (Vyssh. Shkol., Moscow, 1978) [in Russian].
10. R. Hoffstetter and J.-P. Gasc, “Vertebrae and Ribs of
Modern Reptiles,” in Biology of the Reptilia, Ed. by
C. Gans, A. d’A. Bellairs, and T. S. Parsons, Vol. 1: Mor-
phology A (London–New York, 1969), pp. 201–335.
11. W. Janensch, “Die Wirbelsäule der Gattung Dicraeosau-
rus,” Palaeontographica, Suppl. No. 2, 35–133 (1929).
12. S. M. Kurzanov, “Gaint Protoceratopids from the Upper
Cretaceous of Mongolia,” Paleontol. Zh., No. 3, 81–93
(1992).
13. T. Marya ska and H. Osmólska, “Protoceratopsidae
(Dinosauria) of Asia,” Palaeontol. Polon., No. 33, 133–
182 (1975).
14. L. A. Nessov, Dinosaurs of Northern Eurasia: New Data
on the Composition of Assemblages, Ecology, and
Paleogeography (St. Petersburg, 1995) [in Russian].
15. L. A. Nessov, Nonmarine Vertebrates of the Cretaceous
Period of Northern Asia (St. Petersburg, 1997) [in Rus-
sian].
16. H. E. Osborn and C. C. Mook, “Camarasaurus, Amphi-
coelias, and Other Sauropods of Cope,” Mem. Am. Mus.
Natur. Hist. 3, 247–387 (1921).
17. A. S. Romer, Osteology of the Reptiles (Univ. Chicago
Press, Chicago, 1956).
18. D. A. Russell, “The Role of Central Asia in Dinosaurian
Biogeography,” Can. J. Earth Sci. 30 (10–11), 2002–
2012 (1993).
19. C. M. Sternberg, “Complete Skeleton of Leptoceratops
gracilis Brown from the Upper Edmonton Member on
n
′
188
PALEONTOLOGICAL JOURNAL Vol. 41 No. 2 2007
TERESCHENKO
Red Deer River, Alberta,” Bull. Nat. Mus. Can., No. 123,
225–255 (1951).
20. V. B. Sukhanov and V. S. Tereschenko, “The Structure
and Formation of the Sacrum in Dinosaurs,” in Interna-
tional Conference on the Problems of Paleontology of
Central Asia (35 Years of Joint Russian–Mongolian
Paleontological Expedition), May 27–28, 2004 (Paleon-
tol. Inst. Ross. Akad. Nauk, Moscow, 2004), pp. 56–58
[in Russian].
21. V. S. Tereschenko, “On the Reconstruction of the Verte-
bral Column of Protoceratops,” Paleontol. Zh., No. 2,
86–96 (1991).
22. V. S. Tereschenko, “Sexual Dimorphism in the Postcra-
nial Skeleton of Protoceratopids (Neoceratopsia, Proto-
ceratopsidae) from Mongolia,” Paleontol. Zh., No. 4,
79–89 (2001) [Paleontol. J. 35 (4), 415–425 (2001)].
23. V. S. Tereschenko, “About Heterocelous Vertebrae in
Horned Dinosaurs (Protoceratopidae, Neoceratopsia),”
Paleontol. Zh., No. 2, 81–86 (2004a) [Paleontol. J. 38
(2), 200–205 (2004a)] [in Russian].
24. V. S. Tereschenko, “Some Adaptive Features of the Skel-
eton Structure of Horned Dinosaurs (Protoceratopoidea,
Neoceratopsia) Mongolia,” in International Conference
on the Problems of Paleontology of Central Asia
(35 Years of Joint Russian–Mongolian Paleontological
Expedition), May 27–28, 2004 (Paleontol. Inst. Ross.
Akad. Nauk, Moscow, 2004b), pp. 60–63 [in Russian].
25. V. S. Tereschenko, “Morphology of the Axial Skeleton
of Horned Dinosaurs (Neoceratopsia: Protocerato-
poidea) from the Upper Cretaceous of Mongolia,” Can-
didate’s Dissertation in Biology (Paleontol. Inst. Ross.
Akad. Nauk, Moscow, 2004c).
26. V. S. Tereschenko and V. R. Alifanov, “Bainoceratops
efremovi, a New Protoceratopid (Neoceratopsia) from
the Bain-Dzak Locality (Southern Mongolia),” Paleon-
tol. Zh., No. 3, 71–80 (2003) [Paleontol. J. 37 (3), 293–
302 (2003)].
27. J. A. Wilson, “A Nomenclature for Vertebral Laminae in
Sauropods and Other Saurischian Dinosaurs,” J. Vertebr.
Paleontol. 19 (4), 639–653 (1999).
28. H.-L. You and P. Dodson, “Redescription of Neoceratop-
sian Dinosaur Archaeoceratops and Early Evolution of
Neoceratopsia,” Acta Palaeontol. Polon. 48 (2), 261–272
(2003).
