Content uploaded by David Martill
Author content
All content in this area was uploaded by David Martill on Jun 02, 2014
Content may be subject to copyright.
Irritator challengeri, a Spinosaurid (Dinosauria: Theropoda) from the Lower Cretaceous of
Brazil
Author(s): Hans-Dieter Sues, Eberhard Frey, David M. Martill and Diane M. Scott
Source:
Journal of Vertebrate Paleontology,
Vol. 22, No. 3 (Sep. 19, 2002), pp. 535-547
Published by: Taylor & Francis, Ltd. on behalf of The Society of Vertebrate Paleontology
Stable URL: http://www.jstor.org/stable/4524248 .
Accessed: 02/06/2014 06:43
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
.
The Society of Vertebrate Paleontology and Taylor & Francis, Ltd. are collaborating with JSTOR to digitize,
preserve and extend access to Journal of Vertebrate Paleontology.
http://www.jstor.org
This content downloaded from 148.197.61.116 on Mon, 2 Jun 2014 06:43:12 AM
All use subject to JSTOR Terms and Conditions
Journal of Vertebrate
Paleontology
22(3):535-547,
September
2002
C 2002
by
the
Society
of Vertebrate
Paleontology
IRRITATOR
CHALLENGERI,
A
SPINOSAURID
(DINOSAURIA: THEROPODA)
FROM
THE
LOWER
CRETACEOUS OF
BRAZIL
HANS-DIETER
SUES1,
EBERHARD
FREY2,
DAVID M.
MARTILL3,
and
DIANE
M. SCOTT4
'Department
of
Palaeobiology, Royal
Ontario
Museum,
100
Queen's
Park, Toronto,
Ontario M5S 2C6 and
Department
of
Zoology, University
of
Toronto, Toronto,
Ontario,
M5S
3G5, Canada, hdsues@rom.on.ca;
2Abteilung
ftir
Geowissenschaften,
Staatliches
Museum
fiir
Naturkunde
Karlsruhe,
Erbprinzenstrasse
13,
D-76133
Karlsruhe,
Germany;
3School of Earth and
Environmental
Sciences,
University
of
Portsmouth,
Burnaby
Road,
Portsmouth
PO1
3QL,
U.K.;
4Department
of
Zoology, University
of
Toronto at
Mississauga, Mississauga,
Ontario,
L5L
1C6,
Canada
ABSTRACT-The
holotype
of Irritator
challengeri
Martill et
al.,
1996
from
the Romualdo Member of the Santana
Formation
(Lower Cretaceous)
in
northeastern Brazil
represents
the most
complete
skull of a
spinosaurid
known to
date. The now
fully prepared specimen provides
much new information on the cranial structure in
these
enigmatic
predatory
dinosaurs. The skull
is
remarkably
narrow,
especially
in the
region
of
the
elongated
snout. The maxillae are
in broad
contact
along
the
midline,
forming
an extensive
secondary bony
palate.
The
maxillary
teeth have
straight
or
slightly
recurved,
conical
crowns,
with
thin,
fluted enamel
and distinct but smooth carinae. As
in
Baryonyx
walkeri,
the
anterior and ventral
processes
of the lacrimal
meet at a more acute
angle
than
in
most non-avian
theropod
dinosaurs.
The braincase is short
anteroposteriorly
but
deep dorsoventrally, extending ventrally
far below the
occipital condyle.
Irritator
challengeri
most
closely
resembles
Spinosaurus
aegyptiacus
in
the structure of its
teeth,
but more extensive
comparisons
between the two
taxa are
currently
impossible
due to the limited amount of cranial
material known for
the latter.
INTRODUCTION
The Romualdo Member of
the Santana Formation
in
north-
eastern Brazil is well-known for the
abundance,
taxonomic di-
versity,
and often
exceptional preservation
of its vertebrate fos-
sils,
especially
fishes
(Maisey,
1991; Martill,
1993).
In
recent
years,
skeletal remains of dinosaurs have
been
repeatedly
re-
ported
from this unit
(Frey
and
Martill, 1995;
Kellner and Cam-
pos,
1996; Kellner, 1996, 1999;
Martill et
al., 1996;
Martill et
al.,
2000).
Although
the
material recovered to date is
fragmen-
tary,
it is
typically
well
preserved.
The
most remarkable
among
these
fossils is the
nearly
complete
skull of an
unusual
theropod
dinosaur,
which
is
housed
in
the collections of the
Staatliches
Museum
fuir
Naturkunde
Stuttgart
(SMNS 58022).
In its
unpre-
pared
state,
it was
originally
identified as the skull of a
large
pterosaur.
Its dinosaurian affinities
were first
recognized by
Martill et al.
(1996),
who
designated
this
specimen
as the ho-
lotype
of a new
taxon,
Irritator
challengeri,
which
they
placed
in
the
Maniraptora.
Kellner
(1996)
found no evidence
in
sup-
port
of
maniraptoran
affinities and referred Irritator to the
Spi-
nosauridae,
a
group
of
rather
poorly
known but distinctive te-
tanuran
theropods.
This
assignment
has been
widely accepted
(Charig
and
Milner,
1997;
Sereno et
al., 1998;
Taquet
and Rus-
sell,
1998),
and is
fully supported by
the
present study.
The Santana Formation is a
poorly
defined
stratigraphic
unit
of
the
Araripe
Group
in
the
intracontinental
Araripe
Basin of
northeastern
Brazil. Strata of the
Araripe Group crop
out
at the
foot of the
Chapada
do
Araripe,
a vast
plateau
on the borders
of the
states of
Ceard,
Pernambuco,
and
Piauf,
with
possible
correlative
deposits
in
basins to
the south and west
(Martill,
1993).
The Santana Formation
constitutes a diverse suite of
sedimentary
rocks. In
its lower
portion,
a series of
variegated
clays
and
silty
clays
with channel
sandstones,
conglomerates,
and
thin bituminous shales is
overlain
by silty, greyish-green
clays
with abundant
carbonate concretions. The matrix enclos-
ing
the
holotype
of Irritator
challengeri
closely corresponds
to
that of the
concretions from the Romualdo
Member
of
this unit.
The Romualdo Member is a well-known source of vertebrate
fossils,
and there is little doubt that the
specimen
described
here
was derived from it. As
partial
confirmation of this
provenance,
preparation
of the skull
yielded
fossils of the ostracod Patter-
soncypris
as well as isolated scales of
the
ichthyodectid
fish
Cladocyclus,
both of which are
commonly
found
in
concretions
from the Romualdo Member. D. M. M.
recently
showed
a
pho-
tograph
of the
unprepared
skull to several
fossil dealers
in
San-
tana do Cariri.
One
dealer who had
also worked as a collector
recalled the
specimen
and recollected that it was found near
Buxexd,
a site near Santana do Cariri. The nature of the con-
cretion as well as its color and texture are consistent with con-
cretions from this
region.
Buxex6
is a small
farming community
located at about 650
m
above sea-level on the flanks of the
Chapada
do
Araripe
in
the south of
Ceari,
about 5
km
south
of the famous
fossil-producing region
of Santana do Cariri.
It
is
only
accessible
by rough
track. There are
several
localities
in
this
region
where fossils are collected from the Romualdo
Member,
and teams of local collectors move from
place
to
place.
It is
possible
that the
holotype
of
Irritator
challengeri
came from one of several localities in the
valley
of the
Caririacu
River,
and we are unable to determine the
provenance
of this
specimen
more
precisely
than to confirm its derivation from the
Romauldo Member of the Santana Formation.
Despite
the abundance and taxonomic
diversity
of its
fossils,
the
precise
stratigraphic
age
of the Romualdo Member remains
uncertain
(Martill, 1993).
Gardner
(1846)
considered the San-
tana Formation Cretaceous in
age
based on
lithological
simi-
larities to the Cretaceous strata
in
the south of
England.
Braun
(1966)
proposed
an
Early
Cretaceous
age
for this formation on
the basis of similarities to the Wealden strata of northwestern
Europe.
Lima
(1978)
regarded
the Santana Formation as
Aptian
in
age
based on
palynological
data. More recent work
by
Pons
et al.
(1990)
suggested
a late
Aptian
to
early
Albian date for
the
underlying
Crato
Formation and a mid-
to late
Albian
age
for bituminous shales with fossils of the
aspidorhynchid
fish
Vinctifer
from the Romualdo Member. Between the Crato For-
535
This content downloaded from 148.197.61.116 on Mon, 2 Jun 2014 06:43:12 AM
All use subject to JSTOR Terms and Conditions
538 JOURNAL OF VERTEBRATE
PALEONTOLOGY,
VOL.
22,
NO.
3,
2002
m-,
pnn o
is
ao.
so
FIGURE 2. Skull of Irritator
challengeri,
SMNS 58022
(holotype),
in
dorsal view. Scale bar
equals
5 cm.
DESCRIPTION
The retouched
photograph
of SMNS 58022
published by
Martill
et al.
(1996:fig.
2)
presented
a left lateral view of
the
then
largely unprepared
fossil. The
principal openings
of
the
skull
(shown
in
black for
contrast)
were still filled with matrix
but the latter has since been removed
to reveal additional
detail,
including previously
unrecorded
cranial bones. The
alleged
frontoparietal
crest does not
belong
to the
specimen
as
no actual
contact can be established
between this
fragment
of indeter-
minate bone and the skull roof. The dorsal
surfaces of the
right
frontal and both
parietals
as
well as the
posterodorsal
portion
of the
supraoccipital
have been
largely destroyed, possibly
due
to
exposure along
the
edge
of the
concretion
containing
the
skull.
SMNS 58022 was
preserved lying
on its side
in
an
early
diagenetic
calcareous concretion. The crude initial
preparation
undertaken
by
the local
collector inflicted considerable
damage
to the fossil because the hard limestone of the
concretion does
not
always separate cleanly
from the enclosed bone. Thus
pres-
ervation of the
bony
surfaces on the
left side of the skull
(in-
cluding
the
braincase)
is
generally poor,
with
extensive fractur-
ing
as well as loss of bone
in
many places.
However,
the
bones
of the
right
side of the
skull
are,
for the most
part, excellently
preserved.
The skull has been
subjected
to some
crushing,
mainly
on
the
left side
of the snout and the
right
side of
the
postorbital region.
The anterior end of
the
snout,
comprising
most of both
premaxillae
and the anterior
ends of both
maxillae,
is broken off. The broken
surface
is clean and
unweathered,
suggesting
that the rostral
portion
of the snout was lost
only
during
or after the
recovery
of
the fossil. Numerous
large
and
small
fractures,
some due to
septarian cracking
of the surround-
ing
concretion,
traverse the cranial bones.
A
major
break ex-
tends almost
vertically through
the
skull at the level of the
anterior ends of the antorbital
fenestrae and was
crudely
re-
paired
with
epoxy car-body
filler
by
the collector.
Some
bony
surfaces sustained
damage during
a
poorly
executed
attempt by
an unknown
party
to
employ
acid
preparation.
The
right postorbital,
squamosal,
and
quadrate,
the left
quad-
ratojugal,
and the
right angular
as well as both
coronoids,
den-
taries
(except possibly
the
posterior part
of the left
element)
and
splenials
were
probably already
lost
prior
to
fossilization.
The
parietals, premaxillae,
and
supraoccipital
are
only
incom-
pletely preserved.
Most of the
tooth-bearing portion
of the
right
maxilla was
separated
from the skull
during
collecting
and is
now
preserved
as a detached
fragment.
The left
jugal
was
slightly
rotated clockwise
anteroventrally.
The left
postorbital
became detached
along
its sutural
contacts and
dropped
into the
left orbit
prior
to burial.
Similarly,
the
left
squamosal
and
quad-
rate were disarticulated. The former is
now
preserved
medial
to the left
jugal,
whereas
the latter was
displaced
forward and
is now
exposed through
the
right
antorbital fenestra. The left
pterygoid
became
separated
and
somewhat rotated clockwise
ventrally;
its lateral surface was
exposed by preparation.
The
posterior portion
of the left mandibular
ramus is no
longer
in
articular contact with the
skull,
but lies close to
it and its lateral
surface is visible. The
posterior portion
of the
right
mandibular
ramus was
displaced
so that the ventral
edge
of the
surangular
(which
is
exposed
in
both lateral and medial
views)
is
now
adjacent
to the alveolar
margin
of the
right
maxilla.
Skull
In
dorsal
view,
discounting
some
crushing
on
both
sides,
the
skull is
remarkably
narrow
throughout
its entire
length.
The
long
and low snout is
subtriangular
in
transverse section at its
broken anterior end where it is much
deeper
(75 mm)
than wide
(48
mm).
Although
the
anterior end of the snout is not
pre-
served,
the
length
of the
preorbital region
of
the skull was more
This content downloaded from 148.197.61.116 on Mon, 2 Jun 2014 06:43:12 AM
All use subject to JSTOR Terms and Conditions
SUES
ET
AL.-SKULL
OF IRRITATOR 539
than
twice
that of the
postorbital
region
and
significantly
ex-
ceeded the
greatest
height
of the skull. The sides of the snout
are
relatively
flat and
steeply
inclined
dorsomedially.
The
long
axes
of
the antorbital fenestra and orbit
extend
posterodorsally.
Similarly,
the
postorbital region
of
the
skull
roof faces
poster-
odorsally,
and the vertical axis of the braincase extends ante-
roventrally.
The
skull has an estimated
length
of
about 600
mm,
assuming
rostral
proportions
similar to
those for
Baryonyx
walkeri
(Charig
and
Milner,
1997)
and
measured
from the
tip
of
the
snout
to
the
occipital condyle.
(The
much
higher
estimate
published
by
Martill
et
al.
[1996]
was based on the then
largely
unprepared specimen including
the
alleged frontoparietal
crest.)
It is about 165 mm
high
at the level of the
preorbital
bar.
As
the sutures between the
basioccipital, exoccipital-opisthotic,
and
supraoccipital
are still
discernable,
SMNS
58022
may
rep-
resent an individual that was not
fully
mature at the time of
death
(see
Currie and
Zhao,
1994).
The external surfaces of the
cranial bones are smooth where this condition can
be
confi-
dently
assessed on the
specimen.
However,
Carr
(pers.
comm.)
noted the
presence
of "immature" bone
grain
on the
surangular.
The
oval external naris
is
about three times
longer
antero-
posteriorly
than tall
dorsoventrally
(76
mm vs.
approximately
25
mm
for
the
left
naris).
It is
situated well
behind the anterior
end of the
elongate
snout.
The
large
antorbital fenestra has
a
longest
(posterodorsal)
diameter of about 145
mm
(left
open-
ing).
Its
anterior
margin
is
rounded,
but
its dorsal
and
ventral
margins converge posterodorsally, giving
the
opening
an
obliquely ellipsoidal
outline.
The
"preantorbital
fenestra" and
"maxillary process"
identified
by
Martill et al.
(1996:fig.
3a,
"mxp"
and
"p.an.f")
are,
in
fact,
the choana and
part
of the
palatine
and
vomer,
respectively;
these structures are
exposed
in
lateral view
through
the antorbital fenestra
on
both sides
of
the snout.
The
maxilla lacks
accessory
antorbital
(maxillary)
and
promaxillary
fenestrae.
Only
the
posterior edge
of the me-
dial wall of the antorbital fossa was
exposed
along
the antero-
dorsal
margin
of
the antorbital fenestra.
Posteriorly,
the antor-
bital fossa is
weakly
incised onto the lateral surface
of
the
lac-
rimal
and does not extend onto the lateral surface of
the
jugal.
The
tall,
somewhat
keyhole-shaped
orbit is round and widest
in
its dorsal
portion
(which
housed the
eyeball),
but,
more
ven-
trally,
its anterior
and
posterior margins converge
anteroven-
trally.
The
longest
(posterodorsal)
diameter of the
right
orbit is
130 mm. The small
"supratemporal
fenestra" identified
by
Martill et al.
(1996:fig.
3a,
"stf")
represents
an artificial hole
in the left sidewall of the braincase
just
behind the dorsal con-
dyle
of
the
laterosphenoid.
The
medial
margins
of
the
actual
supratemporal
fenestrae are formed
by
the
more
or
less
vertical
lateral
surfaces of
the
parietals
and
laterosphenoids.
The
great-
est
anteroposterior
diameter of the
(right)
infratemporal
fenestra
is 77 mm.
Premaxilla-Only
the
posterodorsal extremity
of the
pre-
maxilla is
preserved.
Posteriorly,
the
premaxilla
bifurcates
to
form
the
anterior and anteroventral
margins
of the external nar-
is.
The
short ventral
process
of
this bifurcation contributes
to
the
shelf-like
ventral
margin
of the
narial
fenestra,
whereas
the
(now
largely destroyed)
dorsal
process
contacted
the
anterior
end
of the
nasal.
Maxilla--Anteriorly,
the maxilla forms a
greatly elongated
and
dorsoventrally deep
subnarial
ramus,
which contributes
most
of
the concave ventral
margin
of
the external naris and
separates
the nasal and
premaxilla
below this
opening.
The lat-
eral surfaces of the maxillae are
only
slightly
inclined dorso-
medially
towards each other. Most of the lateral
portion
of
the
right
maxilla has been broken
off,
exposing
a
large,
conical
maxillary
antrum
(sensu
Witmer,
1997a),
which extends
in the
body
of the
bone
forward to
the
posteroventral
margin
of the
external naris. The
maxilla
is
deeply embayed posteriorly
by
the anterior
and ventral
margins
of
the
antorbital
fenestra and
forms the medial
bony
wall
to
the antorbital
fossa and
maxillary
antrum. The medial wall
of the
maxillary
antrum is
perforated
by
a
large
oval
opening,
as
in
Allosaurus
(Witmer, 1997a).
Me-
dially,
the maxillae
broadly
contact each
other to form a sec-
ondary
bony
palate;
a distinct
ridge,
which
is
clearly
visible
in
transverse section
at
the
broken
anterior
end of the
snout,
ex-
tends
along
their median suture. The slender
ascending
process
of
the maxilla curves
posterodosally
and
tapers
towards
the
posterior
end,
which
overlaps
the anterior ramus of the lacrimal
above the antorbital fenestra.
The
left
maxilla
preserves
nine tooth crowns or
stumps
of
crowns;
the first and third tooth show details
of
the
crown. Most
of
the
dentigerous portion
of the
right
maxilla
was
separated
from
the
skull
by
the collector but can be
readily
fitted
back
onto
it,
using
a
major
vein
of
dark,
sparry
calcite
traversing
the
bone and a broken anterior tooth
as landmarks
(Fig.
IB).
The
right
maxilla holds nine
mostly damaged
teeth,
and there is the
broken
base
and a
partial
impression
in the matrix
of the tenth
crown at the
incomplete posterior
end
of
the bone.
The
broken
dorsal surface of the
posterior part
of this
jaw
fragment
shows
oblique
cross-sections of the functional teeth
and of sometimes
two
replacement
teeth in different
stages
of
eruption.
The
deep-
ly
implanted, vertically
oriented teeth are
well
separated
from
each other more
anteriorly.
Posteriorly,
the
maxillary
tooth
row
terminates below
the anterior end of the antorbital
fenestra,
well
forward
of the orbit. The broken anterior end of
the snout re-
veals
that
the
long,
apparently
slender root
of the first
preserved
tooth in the left maxilla extends
dorsally
close to the level of
the ventral
margin
of
the external naris.
The roots of the teeth
from
opposing
sides
converge medially,
almost
reaching
the
midline.
The
first
and second
preserved
tooth
in
the
left maxilla
have
the
tallest
crowns,
with
heights
of 32 and
over 40
mm,
respectively,
and a mesiodistal diameter of about
13
mm;
their
relative
position
in the
maxilla is uncertain
due to the loss
of
the anterior end of
that
bone.
The crowns of the more
posterior
maxillary
teeth are shorter.
The left maxilla
preserves
nine teeth
partially
or
completely,
and the
complete
tooth row
probably
comprised
at least
11
teeth. The crown of
the last tooth in the
left
maxilla was in the
process
of
eruption
at the
time of
death,
and
only
its
tip
is
exposed.
All tooth crowns
are
conical,
ta-
pering
rather
evenly
towards the
apex,
and
straight
or at most
slightly
recurved.
They
are round
in transverse section rather
than
labiolingually
flattened as is
typical
for
theropod
teeth.
The
lumen of the
pulp
cavity
is
small. The
anterior
(mesial)
and
posterior
(distal)
carinae
extend the full
height
of
the crown to
its
apex. They
are distinct but devoid
of
serrations;
their relative
position
on the tooth crowns does
not
vary along
the
maxillary
tooth row.
The
thin enamel
is
preserved
in extensive
patches
on several
maxillary
tooth crowns.
At
higher magnification,
it
has a
granular
texture,
as
in
Baryonyx
(Charig
and
Milner,
1997:
fig.
19;
Sereno et
al.,
1998:fig.
2E).
The labial surface of
the
left first and the
right eighth
tooth crown bears
fine,
short
enam-
el "wrinkles" that
extend
obliquely
toward
the
posterior
carina,
and
the
carina
appears
"beaded" at
higher magnification
on the
right eighth
tooth
(Fig.
5).
The enamel is
distinctly
fluted
on
all
maxillary
teeth,
with vertical
ridges extending
the
height
of
the crown on both
the
labial and
lingual
surfaces. The labial
surfaces of
the
well-preserved
seventh and
eighth
right
tooth
crowns each bear
seven
vertical
ridges.
In
Baryonyx
walkeri,
only
the
lingual
surface of the tooth crowns is
fluted
(Charig
and
Milner,
1997),
but
fluting
is
present
on both the labial
and
lingual
surfaces of
tooth crowns referred to
Spinosaurus
from
the
Kem Kem
beds
of southern Morocco
(Kellner
and
Mader,
1997).
Nasal--The
long,
narrow nasals
form much of the anterior
portion
of the skull
roof,
extending
from the external nares
back
beyond
the anterior
margin
of
the orbit.
Dorsally,
a
longitudinal
crest is
developed along
the
straight
internasal
suture
(which
is
This content downloaded from 148.197.61.116 on Mon, 2 Jun 2014 06:43:12 AM
All use subject to JSTOR Terms and Conditions
540
JOURNAL
OF
VERTEBRATE
PALEONTOLOGY,
VOL.
22,
NO.
3,
2002
still
discernable
more
anteriorly).
Its
height
cannot
be deter-
mined with
certainty
due to
extensive
breakage
along
much of
its
dorsal
edge.
The crest
appears
to
reach its
greatest
depth
in
the
region
above
the
antorbital fenestra.
Posteriorly,
it
decreases
in
height
and
terminates in
a median
knob-like,
somewhat dor-
soventrally
flattened
projection
(part
of which
has been
abrad-
ed)
that
projects
over the
anterior end
of the
frontals and
ap-
pears
to be
flanked on
either
side
by
a
narrow
opening
in
the
skull
roof.
The latter is
situated on
either
side of the
posterior
end of
the nasals
at their
sutural
contacts with
the
prefrontal
laterally
and the
frontal
posteriorly.
The
presence
of a
"post-
nasal fenestra" in
Baryonyx
walkeri was
inferred
by Charig
and
Milner
(1997)
on the basis
of the
shape
of
the
posterior
end of
the
fused nasals.
However,
it is
likely
that
the
openings
in
SMNS
58022 are
the result of
dorsal
displacement
of the
nasals,
which has
slightly separated
these bones
at their
sutural contacts
with the
frontal,
lacrimal,
and
prefrontal.
This
interpretation
is
supported
by
the fact
that the left
opening
is
less
distinct than
the
right.
Additional cranial
material
is needed
to confirm
the
presence
of
the
postnasal
fenestra in
Baryonyx
and
Irritator.
The
nasal
contacts
the maxilla
laterally along
a
nearly
straight,
forward-sloping
suture.
Anteriorly,
it is
emarginated
by
the
posterior
margin
of
the
external naris.
The nasal
does
not
enter
into the
dorsal
margin
of the
antorbital
fenestra and
fossa. The
nasals are
not fused
medially
throughout
their entire
length,
and
the
internasal suture
is
clearly
visible
anteriorly
where
the dorsal
surface of
the snout
was
damaged.
Lacrimal-The
lacrimal forms
the
slightly
thickened
poster-
odorsal
margin
of the
antorbital
fenestra
and the
anteroposte-
riorly
broad
bony
bar
separating
the
orbit from
the
antorbital
fenestra.
The
long
axes of its
anterior
and ventral
processes
enclose an
angle
of
only
about
400,
similar
to the
condition in
Baryonyx
(about
350;
Charig
and
Milner,
1997).
The ventral
process
of the
lacrimal is
distinctly
inclined
posterodorsally
and,
in
lateral
view,
flares
ventrally
to
form a
broad
flange,
which
reaches
its
greatest
anteroposterior
width at
the sutural
contacts with the
jugal
and
maxilla.
The
extensive anterior
por-
tion of
this
flange
contacts the
maxilla in an
interdigitating
su-
ture;
its
posterior
portion
is
thicker and forms
the
anterior mar-
gin
of
the orbit.
The
slender
anterior
process
of the
lacrimal
contacts
the nasal
dorsally
and
medially,
and is
overlapped by
the
tapering
dorsal
process
of
the
maxilla
anteriorly
close to the
posterodorsal
apex
of
the
antorbital
fenestra.
Posterodorsally,
the lateral
surface of
the better
preserved
right
lacrimal bears a
prominent,
anteriorly facing
recess
at the
junction
of
the ante-
rior and
ventral
rami. This
recess is
partially
concealed in
lateral
view
by
a thick
bony ridge
along
the
posterodorsal
margin
of
the
lacrimal,
but it is
more
exposed
than
in
Baryonyx
walkeri.
It
contains
no
obvious
openings
into the
body
of
the
lacrimal,
but it is
possible
that several
pits
within
the recess
may repre-
sent
foramina that are
still filled
with
matrix.
Dorsally,
the lac-
rimal
is
narrowly
exposed
on
the skull roof.
Its
posterodorsal
cornual
process
is
lower
and
smoother than
the
tuberosity
in
B.
walkeri
(Charig
and
Milner,
1997:fig.
5A).
Prefrontal--The
rather
large
prefrontal
is
triangular
in
lat-
eral view
and
robust. Its
lateral
edge
forms
the
rugose,
thick
anterior
portion
of
the
dorsal rim of
the
orbit,
which
continues
anterolaterally
onto
the
lacrimal. Its
ventral
process
extends
far
anteroventrally
along
the
medial
surface of the
lacrimal.
Pos-
teromedially,
the
prefrontal
inserts into
a broad
anterolateral
notch on
the
frontal.
Dorsally,
it is well
exposed
on the
skull
roof
and
forms
the
lateral
margin
of the
postnasal
fenestra.
Frontal-The
frontals
broadly
contact
the
parietals
along
a
transverse suture
posteriorly.
Dorsally,
they
form a
distinct
ridge
along
their
median
sutural
contact.
Their lateral
margins
converge
only
slightly
anteriorly.
The
supratemporal
fossa ex-
tends
forward
onto
the
dorsal surface
of the
posterolateral
por-
tion of
the
frontal.
Most of
the dorsal
surface of
the
frontal
is
gently
concave
transversely
and smooth.
Ventrally,
low
cristae
cranii border
a
narrow
median
sulcus
for the
olfactory
tracts
of
the
forebrain;
the
interfrontal
suture
is discernable
in the
roof
of
this
groove.
The frontal
forms
the thin
posterior portion
of
the dorsal rim of the orbit.
Parietal-In dorsal
view,
the
parietals
are
much
expanded
anteriorly
and
posteriorly
so that their lateral
edges
bordering
the
supratemporal
fossae
are
distinctly
concave.
They
form
a
nearly
straight
transverse
suture
with
the frontals
anteriorly.
The
interparietal
suture
can still
be traced
more
anteriorly,
but
most
of the
posterodorsally
facing
dorsal surface
of the
parietals
has
been
destroyed.
The low
occipital
wings
of
the
parietals
extend
posterolaterally
as well
as
ventrally. Distally,
each
wing
tapers
and becomes
slightly
twisted
to assume
an
almost
vertical
ori-
entation,
resting
on the dorsal
edge
of the
paroccipital process.
The
lateral surface
of the
occipital
wing
bears
a
depressed
area
for contact
with
the
squamosal.
Postorbital-The
postorbital
is
represented
by
the
disartic-
ulated
left element.
It formed
the
posterodorsal
margin
of
the
orbit
and
the anterior
portion
of the
supratemporal
bar.
The
postorbital
is more
or less
T-shaped
in lateral
view,
with
a
dor-
sal
portion
formed
by
the anterior
and
posterior
processes
and
a broad ventral
process,
the
distal end
of which
is still
buried
in the matrix.
The
anterolateral
surface
of
the
postorbital
does
not
form
a
"horn"
or
rugosity.
The anterior
process
curves
anteromedially
to contact
the frontal
and
parietal.
It is
more
robust
than
the
short,
tapering
posterior
process.
The
ventral
process
formed
the
more dorsal
portion
of the
postorbital
bar.
It
appears
to lack
any
trace
of a
suborbital
process.
Jugal-The
jugal
is
dorsoventrally deep
and
mediolaterally
compressed.
Its much
expanded
anterior
portion
meets
the
max-
illa
in
a
deeply
interdigitating
suture
anteriorly
and the
broad
ventral
process
of the lacrimal
dorsally.
The
sutural contact
be-
tween
the lacrimal
and maxilla
excludes
the
jugal
from
partic-
ipating
in the
posterior
margin
of the antorbital
fenestra
(contra
Martill
et
al.,
1996).
The dorsal
edge
of the
jugal
is notched
by
the narrow
ventral
margin
of the orbit.
On
the better
preserved
right
element,
a low
ridge
extends
longitudinally
above the
ven-
tral
margin.
The
tall,
posterodorsally
directed
dorsal
process
of
the
jugal
forms
a
long, steeply
forward-sloping
surface
for
ar-
ticular
contact
with the ventral
process
of the
postorbital.
The
slender
posterior
process
of
the
jugal
bifurcates
and
receives
the anterior
process
of the
quadratojugal posteriorly,
forming
the
infratemporal
bar.
The dorsal
prong
of this bifurcation
is
more slender
and shorter
than the
ventral one.
Squamosal-The
displaced
left
squamosal
is
preserved
me-
dial to
the left
jugal.
The lateral
surface
of
its anterior
process
bears
a
posteriorly
tapering groove
for
the
reception
of
the
pos-
terior
process
of the
postorbital.
The
dorsal surface
is
convex
anteriorly
and concave
posteriorly,
indicating
a shallow
supra-
temporal
fossa that
is
bordered
laterally by
a low
ridge.
A
short
posterior
process
arises
just
behind
a ventral
concavity
on
the
squamosal
for
the
reception
of
the
proximal
head
of the
quad-
rate.
Its distal
end is
dorsoventrally expanded
and
curves
ven-
trally.
The
ventral
process
of the
squamosal
is
broad
antero-
posteriorly.
Quadratojugal--The
L-shaped
quadratojugal
is
represented
by
the
complete
right
element.
It forms
most of
the ventral
and
posterior
margin
of
the
large
infratemporal
fenestra. The
quad-
ratojugal
has a
tall,
anteroposteriorly
broad,
and
nearly
vertical
dorsal
process,
which
is divided
by
a lateral
ridge
into a
broad
posterolaterally
facing
and a narrower
anterolaterally
oriented
surface,
which
probably
contacted the ventral
process
of
the
squamosal.
It
would have
contacted the
ventral
process
of
the
squamosal
dorsally.
The
more
slender,
tapering
anterior
process
of
the
quadratojugal
inserts into a
slot on
the
infratemporal
process
of the
jugal.
Quadrate--A
displaced
bone
partially
visible
through
the
This content downloaded from 148.197.61.116 on Mon, 2 Jun 2014 06:43:12 AM
All use subject to JSTOR Terms and Conditions
SUES ET AL. SKULL
OF IRRITATOR
541
SO
sPB~l
po. p
PO.C
bo
b.p
FIGURE 3. Braincase of
Irritator
challengeri,
SMNS 58022
(holo-
type),
in
occipital
view. Scale bar
equals
5 cm.
right
antorbital fenestra
right
behind the
palatine represents
the
?left
quadrate.
It is tall and has
a vertical shaft
terminating
in
the
proximal
articular head as well as an
extensive,
flange-like
portion
for
contact with the
pterygoid.
The
proximal
head has
a rounded
triangular
outline in
dorsal view and forms a
cap
that
is
offset
by
a
lip
from the remainder of the bone.
Palatine-The tall and thin
vomeropterygoid process
of the
palatine
(sensu
Witmer
[1997a])
borders the choana and is ex-
posed
in
lateral view
through
each antorbital fenestra.
Anteri-
orly,
it forms a
deep
and stout
flange
for contact with
the vomer.
Pterygoid-The
pterygoid
is a
long
and thin bone. A distinct
concavity
marks the
articular contact for the
basipterygoid
pro-
cess
of the
basisphenoid.
A
short but
dorsoventrally deep, wing-
like
process
for contact with the
quadrate
rises
steeply posterior
and lateral to the
basipterygoid joint.
The
anteriorly extending
palatine
ramus of the
pterygoid
is
long,
thin,
and
almost
straight.
Braincase-As
in
Baryonyx
walkeri
(Charig
and
Milner,
1997),
the
braincase
(Figs.
3,
4)
is short
anteroposteriorly
but
deep
dorsoventrally.
The sutural outlines of
its constituent
bones
are,
for
the most
part,
still
discernable,
especially
on the
well-preserved right
lateral
surface. The
greatest
width of the
braincase,
as measured
posteriorly
across the
(slightly
abraded)
distal ends of the
paroccipital processes,
is about 100 mm. The
foramen
magnum
has a
concave dorsal and an almost
straight
ventral
margin.
It is
25
mm
wide and 20
mm
high.
The dorsal
margin
of the
opening
is formed
by
the
supraoccipital,
and its
lateral
and ventral
margins
are made
up by
the
exoccipitals
except
for a narrow
median contribution to the ventral
margin
from
the
basioccipital.
The
occipital condyle
is almost
hemi-
spherical
but
is flattened
dorsally.
It is
largely
formed
by
the
basioccipital except
for the
dorsolateral
corners,
which are con-
tributed
by
the
exoccipitals.
Its articular
surface faces
slightly
posteroventrally.
The
condyle
is set off from
the braincase
by
a short neck.
The
olfactory
tracts extend in about the same
horizontal
plane
as the
forebrain,
and the hindbrain
parallels
the
forebrain at a more ventral level.
Is.c
-IV
Is
os
f
.fm
f.o
S
FIGURE 4. Braincase of Irritator
challengeri,
SMNS 58022
(holo-
type),
in
right
lateral view.
Scale bar
equals
5 cm.
FIGURE 5. Crown of the
right eighth maxillary
tooth of Irritator
challengeri
(SMNS
58022,
holotype)
in
labial view. Anterior to the
right
of the
figure.
Scale bar
equals
1
cm.
This content downloaded from 148.197.61.116 on Mon, 2 Jun 2014 06:43:12 AM
All use subject to JSTOR Terms and Conditions
542
JOURNAL
OF VERTEBRATE
PALEONTOLOGY,
VOL.
22,
NO.
3,
2002
Supraoccipital-The
supraoccipital
is
not fused
to the
ex-
occipital-opisthotic
(otoccipital).
In
occipital
view,
its dorsal
portion
bears a
conspicuous,
slightly
posterodorsally
projecting
nuchal crest
along
the
midline,
which
is
subtriangular
in
coronal
section
and
served
as
the site of
attachment for
the
ligamentum
nuchae. The
posterior
surface of
the
supraoccipital
is
deeply
recessed on
either side of this
crest,
presumably
for
the insertion
of M.
rectus
capitis.
Breakage
reveals that
the
supraoccipital
is
V-shaped
in
coronal
section.
A
foramen
in
the
notch at
the
dorsal
contact between
the
forward
sloping,
wing-like
lateral
process
(which
presumably represents
the fused
epiotic,
as in
extant
birds and
crocodylians;
Walker,
1990)
and
the dorso-
median
process
of the
supraoccipital
on either
side
represents
the
posterior
opening
for
a
vascular
canal. This
canal
opens
anteriorly
in
a small
foramen
situated on or
close to the
suture
between
the
prootic
and
laterosphenoid,
and
probably
served
for the
passage
of V.
cerebralis
media
(V.
occipitalis
externa of
authors).
Ventrally,
a short
median
process
of the
supraoccipital
enters into
the
dorsal
margin
of
the foramen
magnum.
Orbitosphenoid-The
delicate
orbitosphenoid
is
clearly
dis-
cernable on
the
right
side of
the
braincase.
Dorsally,
a
large
foramen,
presumably
for the
passage
of N.
trochlearis
(IV),
is
situated
ventral and
medial to
the
dorsal
condyle
of the
later-
osphenoid
on the
suture
between
the latter
bone and
the orbi-
tosphenoid.
Just
ventral to this
foramen on
the
right
side
of
the
braincase,
there are two
smaller
openings,
at
least one of
which
represents
the exit
for N.
oculomotoricus
(III).
Ventromedially,
the
orbitosphenoids
meet and
enclose
between them
a
large
opening,
which
presumably
represents
the
passage
for N.
op-
ticus
(II).
Laterosphenoid-The
laterosphenoid
extends
anterolaterally
from
the front of
the
braincase.
It forms
a
transversely
expanded
dorsal
condyle
for
articular
contact
with the
frontal,
parietal,
and
presumably postorbital.
This
condyle
has a
smooth,
round-
ed articular
surface.
The lateral
surface of
the
laterosphenoid
is
concave
anteroposteriorly
and
convex
dorsoventrally.
Posteri-
orly,
a
single
large
foramen
for
the
passage
of N.
trigeminus
(V)
and of
V.
cerebralis
media is
enclosed
between the
later-
osphenoid
and
prootic.
A
deep
sulcus,
which
probably
carried
ramus
ophthalmicus
of
V,,
and a
less well
developed groove
extend
anterodorsally
from
the
foramen
along
the almost
flat
anterolateral
surface of the
laterosphenoid
toward the dorsal
condyle.
This
indicates that
the
three
branches of N.
trigeminus
still left
the
braincase
together
through
a
single
common
open-
ing
and
diverged
only
after
exiting
the
braincase.
Prootic-The
prootic
is an
irregularly shaped,
robust
ele-
ment.
Posteroventral to
the
trigeminal
foramen,
a slit-like
open-
ing
for
the
passage
of N.
facialis
(VII)
is situated
on a thick
crest,
which
becomes a
laterally
projecting
flange
and
extends
onto
the
basisphenoid
more
ventrally.
The latter
represents
the
otosphenoidal
crest
(sensu
Witmer,
1997b;
crista
prootica
of
authors).
Just
posterior
to the
crest,
there is an
extensive,
deep
recess that
continues
laterally
as
the broad
stapedial
groove
along
the
anteroventral
surface of
the
paroccipital
process.
It
contains
a
pair
of
large
foramina,
which are
separated
from
each
other
by
a
slender crista
interfenestralis. The
anterior and
slight-
ly
more
dorsally
situated
opening represents
the
fenestra
ovalis
(fenestra
vestibuli of
authors)
into
which fitted
the
head of
the
stapes,
and the
posterior
foramen is the
metotic
foramen
(sensu
Walker,
1990)
for the
passage
of
N.
glossopharyngeus,
N.
va-
gus,
and
N.
accessorius
(IX-XI).
Although
the sutural
contact
between
the
prootic
and
the
exoccipital-opisthotic
is
not
clear,
at
least
the
posterior
and
dorsal
margins
of the
metotic
foramen
are
formed
by
the
exoccipital-opisthotic.
There
is
no trace of
a
fenestra
pseudorotunda
(cochleae).
Basioccipital
and
Basisphenoid--Together
with
the basi-
occipital,
the
basisphenoid
(which
appears
to
be
indistinguish-
ably
fused
with the
parasphenoid)
forms a
transversely
narrow,
somewhat
apron-like
structure
that
extends
ventrally
far
below
the
occipital
condyle.
The
basioccipital
bears
a
median
depres-
sion
rather
than
a
ridge
between
the
occipital
condyle
and
basal
tubera.
The
basal tubera
are indistinct.
Just below
each
tuber,
an area
for
muscle
insertion
("oval
scar"
sensu
Bakker
et
al.,
1988)
is
developed
along
the
posterolateral
margin
of the
ba-
sicranium.
The
transversely
concave
posterior
surface
of
the
basisphenoid
is
marked
by
a
deep,
dorsoventrally
oval
median
recess
and,
more
dorsally,
a median
opening
on the sutural
con-
tact between
the
basioccipital
and
basisphenoid.
Both
features
are
part
of what
is
commonly
referred
to as the
basisphenoid
sinus,
which
is
part
of the median
pharyngeal system
(Witmer,
1997b).
This
differs
from
the
condition
in
Baryonyx
walkeri
(Charig
and
Milner,
1997:fig.
9A)
where
a
deep
median
"fur-
row"
on
the
posterior
surface
of the
basisphenoid tapers
dor-
sally
to terminate
on the suture
between
the
basioccipital
and
basisphenoid.
The
basipterygoid
processes
project
anteroven-
trally
and
diverge only
slightly
ventrolaterally. Medially,
they
are linked
by
the
transversely
concave
ventral
edge
of the
body
of the
basisphenoid.
The articular
surface
of each
process
faces
ventrally
and
slightly
medially.
Just anterior
to the sella
turcica,
the
anterodorsally
projecting
cultriform
process
arises
from
the
bases
of the
basipterygoid
processes.
It bears
a
blunt,
slightly
recurved
dorsomedial
projection;
Currie
(1985)
interpreted
a
similar
structure
in Troodon
as
marking
the
posteroventral
ex-
tent
of the
(otherwise
unossified)
interorbital
septum.
Anter-
oventral
to the foramen
for
N. facialis
(VII)
and
the
otosphen-
oidal
crest,
the
anteroposteriorly
gently
concave
lateral
surface
of the
basisphenoid
is marked
by
a
prominent,
dorsoventrally
oval recess.
This
depression
contains
two
large
foramina
(which
are
fully
exposed
on the
right
side
of the
braincase)
and
rep-
resents
the anterior
(lateral)
tympanic
recess
(Witmer,
1997b).
The
larger
opening
is situated
on the
anterodorsal
margin
of
the
depression
and
probably
represents
the
posterior
entrance
of
the canal
for A. carotis
interna
(cerebralis),
which
would
have
passed
anterodorsally
to the sella turcica.
The other
fora-
men is located
posteroventral
to the
former;
it
opens
into
the
body
of
the
basisphenoid
and
may
have
been
pneumatic
in
or-
igin.
Exoccipital-Opisthotic-As
in most
dinosaurs,
the
exoccip-
ital
and
opisthotic
appear
to
be
indistinguishably
fused into
a
single
element
(otoccipital).
The
compound
bones do not
meet
along
the midline
but are
separated
by
a
narrow
ventral
process
of
the
supraoccipital
dorsal
to
the foramen
magnum
and
by
the
basioccipital
below
that
opening.
The lateral
surface
of the
con-
dylar portion
of the
exoccipital
is
pierced by
a
single
foramen
for the
passage
of
N.
hypoglossus
(XII)
just
anterior to
the
occipital
condyle.
(This
opening
is
currently exposed
only
on
the
left side
of the
braincase.)
The
robust,
short
paroccipital
process
projects
posterolaterally
and
ventrally
so that
its
distal
tip
is
situated at
about
mid-height
of
the
occipital
condyle.
Its
proximal
end
is
thick,
but
the
process
gradually tapers postero-
laterally.
The distal
end
is
neither
expanded
nor
extended
down-
ward
as,
for
example,
in
Allosaurus
(Gilmore, 1920;
Madsen,
1976).
In
lateral
view,
the
lower half of
the
distal end
of
the
process
bears
a
concavity
for
possible
contact
with the
quadrate.
The
concave
dorsolateral
surface of the
bone is
flattened
by
the
articular
surface
for the
squamosal.
The
ventral
margin
of
the
latter
extends
anterodorsally
to the
proximal
end of the
contact.
Thus
the
medial
process
of
the
squamosal
fitted
into a
broad
groove
formed
by
the
otoccipital
and
parietal.
Stapes--On
the
right
side
of the
skull,
a
slender,
rod-like
bone is
preserved
lying
across the
infratemporal
fenestra
and
the lateral
surface
of
postorbital
process
of
the
jugal.
It
is
55
mm
long
and has
slightly
expanded
and
flattened
proximal
and
distal
ends. The
bone is
identified as a
stapes
because it
closely
resembles
the
stapes
of
Allosaurus
(Madsen,
1976)
and
Dro-
maeosaurus
(Colbert
and
Ostrom,
1958).
The
presence
of
this
This content downloaded from 148.197.61.116 on Mon, 2 Jun 2014 06:43:12 AM
All use subject to JSTOR Terms and Conditions
SUES ET
AL.-SKULL OF
IRRITATOR
543
element
in
SMNS 58022 is
noteworthy
because
only
a few
examples
of a
complete stapes
in
non-avian
dinosaurs have
been
reported
to date.
Mandible
The
left mandibular ramus is
represented
by
the
articulated
surangular, angular,
articular,
and
prearticular,
and the
right by
the articulated
surangular
and articular. The
coronoid,
dentary,
and
splenial
are
not
preserved
for either lower
jaw.
A
possible
exception may
be the
posterior
end
of
the left
dentary,
but
there
is no clear
sutural
separation
from the
surangular
posteriorly
and
the
fragment
in
question
shows
no trace of
alveoli.
Surangular-The
large
surangular
is
a more or less
vertical
and
dorsoventrally
deep
bone,
which
comprises
most of
the
dorsolateral
portion
of the
mandibular
ramus
behind the tooth
row. A
prominent
shelf
extends
along
the lateral
surface of the
surangular
from the level
of the
posterior
end of
the external
mandibular
fenestra back to the
region
anterior to the
glenoid
facet of
the
jaw joint.
The dorsal
surface of
this shelf faces
obliquely
dorsolaterally
and is
concave
anteroposteriorly
as
well as
transversely. Anteriorly,
the
shelf fades into
the dorso-
ventrally
gently
convex lateral surface
of
the
surangular.
The
posterior
surangular
foramen is
small;
it is
clearly
visible on
the medial
surface of the
right
element.
The medial
surface of
the
surangular
is
broadly
concave
dorsoventrally except
for a
prominent
ridge along
the dorsal
margin
of the extensive
ad-
ductor fossa.
Anteromedially,
this
ridge
bears a distinct
articular
facet,
possibly
for
contact with the
coronoid;
the latter
passes
onto the
dorsal surface of
the
surangular.
Angular-The
angular,
which is
incompletely preserved
on
the left
mandibular
ramus,
forms the
distinctly
concave
ventral
margin
of the
external mandibular fenestra.
Behind
this
open-
ing,
it forms a
thin
sheet of bone. There is no
clearly
discern-
able sutural
separation
from the
surangular,
and it
is not clear
whether
the
angular
contacted the
surangular
anterior to the
external
mandibular fenestra.
Prearticular-The
prearticular
is
exposed
in
medial view
on
the left
mandibular ramus. It is
a
curved,
long,
and thin
bone,
which is
dorsoventrally
expanded
at either end and
borders the
adductor fossa
ventromedially.
The
prearticular
is most
robust
at
its
posterior
end,
which covers the ventromedial
aspect
of
the
articular. Its dorsal
edge
is concave.
The
anterior
portion
of
the
prearticular
forms a
vertical,
dorsoventrally expanded plate
of
bone,
which
presumably
contacted
the
splenial
and the me-
dial surface of
the
dentary.
A
narrow
surface
along
the dorsal
margin
of the
anterior end
may
have
been for contact
with the
coronoid.
Articular-The articular
contacts
the
prearticular
antero-
medially
and the
surangular
laterally.
Ventrally,
it is
plate-like
and
wedged
between
the
surangular
laterally
and the
prearti-
cular
medially.
The
articular facet for
the
mandibular
condyle
of the
quadrate
is
concave
anteroposteriorly
and
delimited
by
transverse
bony ridges
anteriorly
and
posteriorly. Immediately
posterior
to the
posterior
transverse
ridge,
the
dorsomedial
edge
of the articular is
distinctly
concave
anteroposteriorly.
A
small
foramen,
which
may
have served
for
passage
of the
chorda
tympani,
is situated
just posteromedial
to the
posterior
trans-
verse
ridge.
The articular
forms an
anteroposteriorly
short re-
troarticular
process,
which turns
posteromedially
behind the
facet for
the mandibular
condyle
of the
quadrate.
The
process
is
dorsolaterally-ventromedially
flattened
so
that its
anteropos-
teriorly gently
convex
lateral surface faces
dorsolaterally.
RELATIONSHIPS OF
IRRITATOR CHALLENGERI
Irritator shares
with
Baryonyx
and the
probably congeneric
Suchomimus
(see below)
the
following
derived character-states
in
the skull
and dentition
that are absent
in most or
all other
known non-avian
theropod
dinosaurs:
(1)
The
skull is remark-
ably
narrow
throughout
its entire
length, especially
in
its
rostral
region.
(2)
The
long
but low external
nares are situated far
back
on the sides
of the
elongated
snout rather
than near the
rostral
end of the
snout.
(3)
The
maxillae
broadly
contact
each other
medially,
forming
an extensive
secondary bony palate.
(4)
The
maxilla forms
a
greatly elongated
subnarial
ramus,
which
sep-
arates
the
premaxilla
and
nasal below the external
naris and
is
longer
anteroposteriorly
than
deep
dorsoventrally.
(5)
The
max-
illary
teeth
have
straight
or
slightly
recurved crowns that
are
round
to subcircular
in
transverse
section,
rather than
labiolin-
gually
flattened.
(6)
If the
holotype
of
"Angaturama
limai"
indeed
represents
the anterior end of the snout of
Irritator chal-
lengeri,
the latter also
shares
with
Baryonyx
the
apomorphic
presence
of seven
premaxillary
teeth.
(7)
The
fused
posterior
portions
of the nasals terminate
in
a knob-like
median
projec-
tion
posteriorly.
(8)
A
narrow fenestra
(postnasal
fenestra;
Charig
and
Milner,
1997)
appears
to be
present
between the
frontal,
posterior
end of the
conjoined
nasals,
and
prefrontal
on
either side of the skull
roof;
this feature is
convergently
present
in
the
coelophysoid theropod
Syntarsus
(Rowe, 1989).
(9)
The
anterior
and ventral
processes
of
the lacrimal enclose a
much
more acute
angle
(about
350-40')
than in
other
theropod
taxa
(about
75'-90o).
(10)
The braincase
is short
anteroposteriorly
but
deep
dorsoventrally,
extending ventrally
far below
the oc-
cipital condyle.
(11)
The
basipterygoid processes
of the
basi-
sphenoid
are
elongate
and
diverge only slightly
ventrolaterally.
These features
support recognition
of the
family-level
taxon
Spinosauridae
Stromer,
1915,
as defined and
diagnosed by
Ser-
eno
et al.
(1998).
Combining
information
from
the
cranial
bones of
Baryonyx
walkeri and
"Suchomimus"
tenerensis,
Sereno et al.
(1998:fig.
2)
reconstructed the skull
of
spinosaurid theropod
dinosaurs
as
long,
low,
and
narrow
throughout
its entire
length. They
criti-
cized the reconstruction
of
the skull of B. walkeri
by
Charig
and
Milner
(1997:fig.
1)
as
exaggerating
the vertical
height
of
the cranium in
the
occipital
region by
"unnatural ventral dis-
placement"
of
the
quadrate
to the
paroccipital
process
(Sereno
et
al.,
1998:1301).
However,
as the more
distal
portions
of
the
paroccipital processes
in the
holotype
of B.
walkeri
(BMNH
R9951)
are not
preserved (Charig
and
Milner,
1997:fig.
9),
it is
not obvious which
reference
point
was used
by
Sereno et al.
(1998)
as the basis for their assertion. The skull of L
challengeri
demonstrates that the more
posterior
region
of the
skull was
indeed
deeper dorsoventrally
than
the snout
(Figs.
1,
6).
The
Spinosauridae
had
a
wide,
apparently mainly
Gondwan-
an distribution
during
the
Early
and
early
Late Cretaceous
(Charig
and
Milner, 1997;
Sereno et
al.,
1998).
Stromer
(1915)
named
Spinosaurus aegyptiacus
on
the basis of associated
skel-
etal remains from
the
Upper
Cretaceous
(Cenomanian)
Baha-
riya
Formation of the
Bahariya
Oasis
in
the Western
Desert of
Egypt.
The
holotype
consists of an edentulous
fragment
of a
maxilla,
a
partial
mandible
with four
teeth,
isolated
teeth,
ver-
tebrae
from all
parts
of the
column,
incomplete
thoracic
ribs,
and
gastralia
elements
(Stromer,
1915,
1936).
With the
possible
exception
of a caudal
vertebra,
Stromer
interpreted
this material
as
representing
the skeletal
remains of a
single
individual. Un-
fortunately,
the
holotype
and
only
known
specimen
was
lost
during
the destruction of
the
Palaiontologische Staatssammlung
in
Munich
by
a British air raid
in
1944. Stromer
(1934a)
iden-
tified additional
postcranial
bones from
Bahariya
as
"Spinosau-
rus
B"
based on their
proportionately
smaller size and
some
morphological
differences.
However,
Sereno et al.
(1998)
re-
assigned
this material
(which
was
destroyed along
with the ho-
lotype
of
S.
aegyptiacus)
to the
allosauroid
Carcharodontosau-
rus saharicus.
Buffetaut
(1989)
identified two
jaw fragments
and an isolated
tooth from
the
?Cenomanian-age
Kem
Kem
beds of southern
Morocco as
Spinosaurus.
Russell
(1996)
re-
This content downloaded from 148.197.61.116 on Mon, 2 Jun 2014 06:43:12 AM
All use subject to JSTOR Terms and Conditions
544 JOURNAL OF VERTEBRATE
PALEONTOLOGY,
VOL.
22,
NO.
3,
2002
FIGURE 6. Partial reconstruction of the skull and mandible of Irritator
challengeri
in left lateral
view,
based
primarily
on the better
preserved
right
side of SMNS 58022. Scale bar
equals
5 cm. Broken lines indicated restored
portions.
The
height
of the
nasal crest and the
depth
of the
postdentary portion
of the mandibular ramus are uncertain. The
supralabial
foramina on the lateral surface of the maxilla are based on the condition
in
Baryonyx
tenerensis
(Sereno
et
al.,
1998).
This content downloaded from 148.197.61.116 on Mon, 2 Jun 2014 06:43:12 AM
All use subject to JSTOR Terms and Conditions
SUES
ET
AL.--SKULL
OF IRRITATOR
545
ferred several vertebrae from
the same
region
to
this
genus,
and
Taquet
and Russell
(1998)
figured
and
briefly reported
on
the
anterior
portion
of
a
snout.
Additional cranial material from
the
Kem
Kem
beds
will be
described
by
Milner
(in
prep.).
Russell
used
proportional
differences in
the cervical vertebrae to distin-
guish
a new
species,
Spinosaurus
maroccanus,
for the
Moroc-
can
material,
but Sereno et al.
(1998)
considered
these
differ-
ences
taxonomically
insignificant
and
synonymized
S.
maroc-
canus with
S.
aegyptiacus.
To
date,
the
most
completely
known
spinosaurid theropod
is
Baryonyx
walkeri
from
the Lower
Cretaceous
(Barremian)
Up-
per
Weald
Clay
of
Ockley,
Surrey (England; Charig
and
Milner,
1986,
1997).
Buffetaut
(1989,
1992)
first
recognized
its
close
relationship
to
Spinosaurus.
The
holotype
of B. walkeri is an
incomplete
but
well-preserved
skeleton
(BMNH R9951),
which
includes
largely
disarticulated and
fragmentary
bones of the
skull and
mandible and was
described
in
detail
by Charig
and
Milner
(1997).
Sereno
et al.
(1998:1302)
diagnosed
B.
walkeri
by
"fused nasals with a median
crest
terminating
posteriorly
in
a
cruciate
process,
a solid
subrectangular
lacrimal
horn,
a
marked
transverse constriction of the
sacral or anterior caudal
centra,
a
well-formed
peg-and-notch
articulation between the
scapula
and
coracoid,
an
everted
distal
margin
of
the
pubic
blade,
and a
very
shallow
fibular fossa."
Charig
and
Milner
(1997)
and Martill and Naish
(2001)
listed
additional
occur-
rences
of isolated teeth as well
as isolated
postcranial
bones
referable to B.
walkeri
(or
related
taxa)
from the Lower Cre-
taceous
(mainly
Barremian)
of
England
and the
Isle
of
Wight.
A
maxilla
fragment
from
the
Barremian-age
Enciso
Group
of
La
Rioja, Spain
has
also
been referred to
Baryonyx
(Viera
and
Torres,
1995).
Taquet
(1984),
Taquet
and
Russell
(1998),
and
Sereno
et al.
(1998)
reported
on
a series of
specimens
of a
Baryonyx-like
spinosaurid
from continental
strata of the Lower Cretaceous
(Aptian-Albian)
Tegama Group
at
Gadoufaoua
(locality
GAD
5),
Niger. Taquet
and Russell
(1998)
named
Cristatusaurus
lap-
parenti
on
the basis
of
conjoined
premaxillae
and associated
fragments
of a
maxilla
and
dentary. They distinguished
this
taxon from
Baryonyx
walkeri
solely
on
the basis
of the
"bre-
virostrine
condition"
of the
premaxilla.
Sereno
et al.
(1998)
considered this
difference
uninformative,
and
Charig
and Mil-
ner
(1997)
identified
the
material discovered
by
Taquet
as
Bary-
onyx
sp.
indet.
Shortly
after the
publication
of
the
paper by
Taquet
and
Russell,
Sereno et
al.
(1998)
briefly
announced
the
discovery
of additional skeletal
remains,
including
a
snout and
a
partial postcranial
skeleton,
from the
same
locality. They
as-
signed
these
specimens
to
a new
genus
and
species,
Suchomi-
mus
tenerensis,
which
they distinguished
from
Baryonyx
walk-
eri
by
the broader and taller neural
spines
of the
dorsal,
sacral,
and
anterior
caudal
vertebrae,
robust humeral
tuberosities,
much
enlarged
olecranon
that
is offset from
the humeral
articulation,
and
hook-shaped
radial
ectepicondyle.
There exists at
present
no
evidence to indicate the
presence
of
more than one taxon of
spinosaurid
in
the faunal
assemblage
from
GAD 5.
We
concur
with Milner
(in
prep.)
that
the anatomical
differences between
the material
reported by
Sereno et al.
(1998)
and
Baryonyx
walkeri
only
warrant
recognition
of the
former as
a distinct
species
of
Baryonyx,
B.
tenerensis. The
generic
nomina
Cris-
tatusaurus and
Suchomimus should be considered
subjective
ju-
nior
synonyms
of
Baryonyx.
Buffetaut
and
Ingavat
(1986)
de-
scribed some
unusual
teeth from
the
Upper
Jurassic Sao Khua
Formation of
northeastern Thailand as
Siamosaurus
suteethorni
and
tentatively
referred this taxon
to the
Spinosauridae.
How-
ever,
the
currently
available material
is insufficient
for
estab-
lishing
even dinosaurian affinities for
Siamosaurus.
Sereno
et al.
(1994, 1998)
and Holtz
(2000)
have established
the
phylogenetic position
of
Spinosauridae
among
basal Tetan-
urae,
obviating
the need
for
a cladistic
analysis
in this
paper.
Sereno
et al.
(1994, 1998)
grouped Spinosauridae
with
Torvo-
sauridae
(comprising Eustreptospondylus
from the
Middle
Ju-
rassic
[Callovian]
Oxford
Clay
Formation
of
England
and
Tor-
vosaurus
from the
Upper
Jurassic
[Kimmeridgian-Tithonian]
Morrison Formation of the
western
United
States)
and
Afrov-
enator from
the Lower Cretaceous
Tiourar6n
beds of
Niger
in
a
clade
Spinosauroidea
("Torvosauroidea"
of
Sereno et
al.,
1994).
However,
the
comprehensive phylogenetic analysis
of
Theropoda
by
Holtz
(2000)
placed
Spinosauridae
as the
most
basal
clade
of
Tetanurae and considered
Eustreptospondylus,
Torvosaurus,
and
Afrovenator
as
progressively
more derived
within
Tetanurae. Of the
apomorphies
listed
by
Sereno et
al.
(1998)
in
support
of
Spinosauroidea,
Irritator
shares
the
pres-
ence of a subnarial
process
of the
maxilla
that is
longer
than
deep
and the
dorsoventrally
narrow anterior ramus of the lac-
rimal.
However,
the derived
character-state "lacrimal
anterior
ramus,
length:
... less
(1)
than 65%
of
the ventral ramus"
cited
by
Sereno et al.
(1998:1302)
is absent
in
Irritator
(SMNS
58022,
right
side).
Sereno
et
al.
(1998)
distinguished
two clades
among Spino-
sauridae:
Baryonychinae,
comprising
Baryonyx
and Suchomi-
mus,
and
Spinosaurinae,
for
Spinosaurus
and
Irritator.
At
pre-
sent,
the latter
grouping
is united
only
by
two
derived
features
of the
teeth
(tooth
crowns with
distinct but
non-serrated
carinae
and fluted
enamel
on both
the labial
and
lingual
surfaces)
and
possibly
the wide
spacing
of
the
maxillary
teeth.
In
Baryonyx,
the carinae are
very finely
serrated
(Martill
and
Hutt,
1996;
Charig
and
Milner,
1997;
Sereno et
al.,
1998).
The
presence
of a
spinosaurid
theropod
in
the Santana For-
mation is not
surprising
in
view of
the connection between east-
ern
Brazil and
West
Africa
during
at least
part
of the
Early
Cretaceous,
which is
reflected
by
numerous
sister-group
pair-
ings among
continental fish and
tetrapod
taxa from these
re-
gions
(e.g.,
Forey
and
Grande,
1998).
FUNCTIONAL INFERENCES
As noted
above,
Baryonyx
and
Irritator
share
a
suite of cran-
iodental features
that
distinguish
them
from
most
known non-
avian
theropod
dinosaurs.
The
straight
or
at most
slightly
re-
curved,
conical
tooth
crowns in
these two
taxa are
round or
oval in
transverse
section,
rather than
labiolingually
flattened,
and their mesial and distal
carinae have either
very
fine serra-
tions
(Baryonyx)
or none
at all
(Irritator).
This
type
of teeth
was less suited
for
"grip-and-rip"
cutting
than the teeth with
more
coarsely
serrated
carinae in
other
non-avian
theropods
(Abler,
1992),
and
instead
may
have
been
used
primarily
for
impaling
and
holding prey
items
(Charig
and
Milner,
1997).
The
long,
narrow snout of
Baryonyx
forms an
expanded
anterior
end
("terminal
rosette"
sensu
Charig
and
Milner,
1997)
with
six or seven
teeth
in
each
premaxilla, resembling
the
spatulate
tip
of the snout
in
predominantly piscivorous, longirostrine
cro-
codyliforms
such as the extant
gharial
(Gavialis
gangeticus)
and
various extinct taxa. The terminal
expansion
of the
snout
of
"Angaturama"
is
less
prominent
than
that of
Baryonyx
(Kellner
and
Campos,
1996).
The
external
nares are
situated
well back
behind
the anterior
end of the
snout
in
Baryonyx
and
Irritator.
Taquet
(1984),
Buffetaut
(1989),
Charig
and Milner
(1986,
1997),
Milner
(1996),
Martill et al.
(1996)
and Sereno
et
al.
(1998)
interpreted
this
combination
of
features as indi-
cating
piscivorous
habits for
spinosaurid
dinosaurs. The
pre-
sumed
gastric
contents of
the
holotype
of
Baryonyx
walkeri
(BMNH
R9951)
contain etched
scales of the holostean fish
Lep-
idotes and thus
are
consistent with
this
dietary
inference
(Char-
ig
and
Milner,
1997).
However,
they
also
included
abraded
and(or)
etched bones of a small
individual
of the
ornithopod
dinosaur
Iguanodon,
suggesting
that B.
walkeri
fed on terres-
trial
vertebrates
as well as
fish.
Most extant
faunivorous
tetra-
This content downloaded from 148.197.61.116 on Mon, 2 Jun 2014 06:43:12 AM
All use subject to JSTOR Terms and Conditions
546
JOURNAL OF
VERTEBRATE
PALEONTOLOGY,
VOL.
22,
NO.
3,
2002
pods
are
opportunistic
feeders,
and there exists no
reason
not
to
assume
similarly
broad
dietary
habits for
non-avian
theropod
dinosaurs.
The
deeply
implanted,
vertically
oriented teeth of Irritator
with their
conical,
straight
or at
most
slightly
recurved crowns
are
suitable for
apical
loading parallel
to
their
long
axes,
which
is consistent with
their
use
for
impaling
and
holding prey.
The
narrow snout with its
nearly
vertical sides and more
or
less
convex dorsal surface would have
facilitated forceful
biting
down on
prey
because this
type
of
rostral
shape
(oreinirostral
condition)
is
less
susceptible
to
dorsoventral
bending
stresses
(Busbey,
1995).
It
clearly
differs from the
long,
tubular snout
(platyrostral
condition)
in
Gavialis and various extinct crocod-
yliforms
that have been
interpreted
as
piscivorous (Busbey,
1995).
The
extensive
secondary bony palate
in
spinosaurid
the-
ropods
would also
have served
to
reduce
bending
stresses
act-
ing
on the snout
(Thomason
and
Russell,
1986).
We
hypothe-
size
transmission of the
vertically
directed forces
applied along
the tooth row
along
the roots
of
the teeth to
the
occipital
seg-
ment of
the
skull
through
the
dorsal
strut
formed
by
the
nasals,
which was
probably
further
strengthened by
the
median crest.
The
posterodorsal
inclination
of
the
postorbital
region
of
the
skull
suggests
a
reorientation
of
M. adductor mandibulae ex-
ternus.
Barghusen
(1973)
interpreted
a
comparable
feature
in
basal
therapsids
as
reflecting
a
posterodorsal
line
of
action
for
the adductor
jaw
muscle to resist anteroventral
displacement
of
the mandible
by prey
seized with the
enlarged
anterior teeth.
This
functional
interpretation
is
equally plausible
for
spino-
saurid
theropods,
which
have
a
premaxillary
dentition
suitable
for
seizing
prey
(Charig
and
Milner,
1997).
The skull
of Irritator does not
appear
to be
well-suited
for
catching
and
processing large,
resistant
prey.
Its
structure
dif-
fers
from that in
other
large
theropod
dinosaurs such
as
Allo-
saurus
(Rayfield
et
al.,
2001)
and
Tyrannosaurus
(Erickson
et
al.,
1996),
presumably reflecting
different modes of
feeding.
Most
likely
spinosaurid theropods rapidly
and
forcefully
seized
smaller
prey,
which
was then
processed
by
dorsoventral motion
of
the
head
facilitated
by
the
powerful
neck musculature.
(Ex-
tensive side-to-side
striking
movements
of the
head,
as em-
ployed by
extant
crocodylians,
appear
unlikely
in view
of the
narrow
occiput
as
well
as
the weak
development
of
the
basal
tubera.)
Whereas fish
formed
part
of the diet
in
at least B. walk-
eri,
there is
nothing
to
suggest
that
spinosaurids
were exclu-
sively
or even
predominantly piscivorous.
Previous anatomical
comparisons
between the
feeding
apparatus
of
crocodylians
and
spinosaurid
theropods
were
based
only
on
superficial
resem-
blances.
The
postcranial
skeleton of
Baryonyx
lacks
any
obvious
spe-
cializations
suggestive
of an
aquatic
or
semiaquatic
mode of
life
(Charig
and
Milner,
1997).
Charig
and Milner
(1986,
1997)
interpreted
the
greatly
enlarged
and
strongly
curved
ungual
of
manual
digit
I as a
"gaffing"
device for
catching
fish,
but this
intriguing hypothesis
remains untestable in
the absence
of
a
close
analogue among
extant
tetrapods.
ACKNOWLEDGMENTS
We are
greatly
indebted to R. Wild
(Staatliches
Museum
ftir
Naturkunde
Stuttgart)
for the extended
loan
of
the material and
his
continuing
interest
in
this
study.
D.
Pulert
(Toronto)
pre-
pared
the
drawings
of
the
reconstructed skull
reproduced
in
Fig.
6. The
remaining
illustrations are the work of D.
M. S. We
thank
T.
D. Carr
(Royal
Ontario
Museum)
who
assisted
our
work
through
a
thorough
critique
of the
manuscript.
A.
C.
Mil-
ner
(The
Natural
History
Museum,
London)
and
T.
R.
Holtz,
Jr.
(University
of
Maryland,
College
Park)
provided
construc-
tive
reviews.
H.-D.
S.
gratefully
acknowledges support
from the
Royal
Ontario Museum Foundation.
LITERATURE CITED
Abler,
W. L.
1992. The serrated
teeth
of
tyrannosaurid
dinosaurs,
and
biting
structures in other animals.
Paleobiology
18:161-183.
Bakker,
R.
T.,
M.
Williams,
and
P.
Currie.
1988.
Nanotyrannus,
a new
genus
of
pygmy tyrannosaur
from the latest Cretaceous of
Mon-
tana. Hunteria
1(5):1-30.
Barghusen,
H.
R.
1973.
The
adductor
jaw
musculature
of
Dimetrodon
(Reptilia, Pelycosauria).
Journal of
Paleontology
47:823-834.
Benton,
M.
J.,
S.
Bouaziz,
E.
Buffetaut,
D.
Martill,
M.
Ouaja,
M. Sous-
si,
and
C. Trueman.
2000. Dinosaurs and other fossil
vertebrates
from fluvial
deposits
in
the
Lower Cretaceous
of
southern Tunisia.
Palaeogeography,
Palaeoclimatology, Palaeoecology
157:227-246.
Berthou,
P.-Y.
1990.
Le bassin
d'Araripe
et
les
petits
bassins intracon-
tinentaux voisins
(NE
de
Bresil):
formation et evolution dans
le
cadre
de
l'ouverte
de
l'Atlantique equatorial.
Comparison
avec
les
bassins ouest-africains
situ6s dans le m8me
contexte;
pp.
113-134
in
D.
de
A.
Campos,
M.
S. S.
Viana,
P
M.
Brito,
and G.
Beurlen
(eds.),
Atas
do I
Simp6sio
sobre a Bacia do
Araripe
e Bacias
In-
teriores do
Nordeste,
Crato,
14-16 de Junho de
1990,
Anais.
Crato,
Ceara.
Bouaziz,
S.,
E.
Buffetaut,
M.
Ghanmi,
J.-J.
Jaeger,
M.
Martin,
J.-M.
Mazin,
and
H.
Tong.
1988. Nouvelles
d6couvertes
de
vert6br6s
fos-
siles
dans l'Albien du
Sud tunisien. Bulletin de la
Soci6t6
g6olo-
gique
de France
(8)
4:335-339.
Braun,
O.
P.
G.
1966.
Estratigrifia
dos sedimentos da
parte
interior da
regido
Nordeste do Brasil
(bacias
de
Tucano-Jatobi,
Mircundiba e
Araripe).
Boletim da
Divisno
de
Geologia
e
Mineralogia
do
De-
partamento
Nacional
de
Produgqo
Mineral,
Rio
de Janeiro
263:1-
75.
Buffetaut,
E.
1989.
New
remains
of
the
enigmatic
dinosaur
Spinosaurus
from the Cretaceous of
Morocco and the affinities between
Spi-
nosaurus and
Baryonyx.
Neues
Jahrbuch
fOr
Geologie
und
Paldion-
tologie,
Monatshefte
1989:79-87.
.
1992. Remarks on the
Cretaceous
theropod
dinosaurs
Spino-
saurus and
Baryonyx.
Neues
Jahrbuch
ftir
Geologie
und
Paldion-
tologie,
Monatshefte 1992:88-96.
-
,
and R.
Ingavat.
1986. Unusual
theropod
dinosaur
teeth
from
the
Upper
Jurassic of Phu
Wiang,
northeastern Thailand. Revue
de
Pal6obiologie
5:217-220.
Busbey,
A.
B.
1995.
The
structural
consequences
of
skull
flattening
in
crocodilians;
pp.
173-192 in J. J. Thomason
(ed.),
Functional Mor-
phology
in
Vertebrate
Paleontology. Cambridge University
Press,
Cambridge
and New York.
Charig,
A.
J.,
and A. C. Milner.
1986.
Baryonyx,
a remarkable new
theropod
dinosaur.
Nature 324:359-361.
,,
and -. 1997.
Baryonyx
walkeri,
a
fish-eating
dinosaur
from the Wealden of
Surrey.
Bulletin
of
the Natural
History
Mu-
seum,
Geology
Series 53:11-70.
Colbert,
E.
H.,
and
J.
H.
Ostrom.
1958.
Dinosaur
stapes.
American
Museum
Novitates 1900:1-20.
Currie,
P. J.
1985. Cranial
anatomy
of
Stenonychosaurus
inequalis
(Saurischia,
Theropoda)
and its
bearing
on
the
origin
of
birds. Ca-
nadian Journal of Earth
Sciences 22:1643-1658.
, and
X.-J.
Zhao.
1994.
A
new carnosaur
(Dinosauria,
Thero-
poda)
from
the
Jurassic of
Xinjiang, People's Republic
of China.
Canadian Journal
of
Earth Sciences
30:2037-2081.
[Listed
as "Oc-
tober-November
1993,"
but
actually published
on
February
10,
1994.]
Erickson,
G.
E,
S. D. Van
Kirk,
J.
Su,
M.
E.
Levenston,
W.
E.
Caler,
and
D.
R.
Carter. 1996. Bite-force estimation for
Tyrannosaurus
rex from
tooth-marked bones. Nature
382:706-708.
Forey, P.
L.,
and L.
Grande. 1998. An African twin to the Brazilian
Calamopleurus
(Actinopterygii:
Amiidae).
Zoological
Journal
of
the Linnean
Society
123:179-195.
Frey,
E.,
and D. M. Martill.
1995.
A
possible
oviraptorosaurian theropod
from the
Santana Formation
(Lower Cretaceous,?Albian)
of Brazil.
Neues
Jahrbuch
fuir Geologie
und
Palhontologie,
Monatshefte 1995:
397-412.
Gardner,
G.
1846. Travels
in
the Interior of
Brazil,
Principally
through
the
Northern
Provinces,
and the
Gold and Diamond
Districts,
dur-
ing
the Years
1836-1841.
Reeve,
Benham and
Reeve, London,
XVI
+
562
pp.
Gauthier,
J. A.
1986. Saurischian
monophyly
and the
origin
of
birds;
This content downloaded from 148.197.61.116 on Mon, 2 Jun 2014 06:43:12 AM
All use subject to JSTOR Terms and Conditions
