New information on the skull of the enigmatic theropod Spinosaurus with remarks on its size and affinities

Abstract

New specimens of the unusual theropod Spinosaurus cf. S. aegyptiacus from the Late Cretaceous (early Cenomanian) of Morocco reveal new information about the structure of the snout and the very large adult body size attained by the species. The external naris is retracted farther caudally on the snout than in other spinosaurids and is bordered exclusively by the maxilla and nasal. The fused nasals preserve a longitudinal, fluted crest. The size of the snout suggests that Spinosaurus may well have exceeded the maximum adult body size of other large Cretaceous theropods such as Tyrannosaurus and Giganotosaurus. The new material also supports the monophyly of the Spinosaurinae and the separation of Spinosaurus and Irritator.

Full text

NEW INFORMATION ON THE SKULL OF THE ENIGMATIC THEROPOD SPINOSAURUS,
WITH REMARKS ON ITS SIZE AND AFFINITIES
CRISTIANO DAL SASSO
1
, SIMONE MAGANUCO
2
, ERIC BUFFETAUT
3
, and MARCO A. MENDEZ
4
1
Museo Civico di Storia Naturale di Milano, Corso Venezia 55, 20121 Milano, Italy, cdalsasso@yahoo.com;
2
Museo Civico di Storia Naturale di Milano, Corso Venezia 55, 20121 Milano, Italy, simonemaganuco@iol.it;
3
CNRS, 16 cour du Liégat, 75013 Paris, France, eric.buffetaut@wanadoo.fr;
4
University of Chicago, Department of Organismal Biology and Anatomy, 3424 W 54th Street, Chicago, IL 60632, USA,
mamendez@uchicago.edu
ABSTRACT—New specimens of the unusual theropod Spinosaurus cf. S. aegyptiacus from the Late Cretaceous (early
Cenomanian) of Morocco reveal new information about the structure of the snout and the very large adult body size
attained by the species. The external naris is retracted farther caudally on the snout than in other spinosaurids and is
bordered exclusively by the maxilla and nasal. The fused nasals preserve a longitudinal, fluted crest. The size of the snout
suggests that Spinosaurus may well have exceeded the maximum adult body size of other large Cretaceous theropods such
as Tyrannosaurus and Giganotosaurus. The new material also supports the monophyly of the Spinosaurinae and the
separation of Spinosaurus and Irritator.
INTRODUCTION
In 1912, Ernst Stromer discovered several bones of a large,
long-snouted, sail-backed predator from the Cenomanian of Ba-
harija (Egypt) that he named Spinosaurus aegyptiacus (Stromer,
1915). Unfortunately, this material was destroyed during an air
raid in the Munich bombing of April 1944 (Taquet, 1984; Taquet
and Russell, 1998; Sereno et al., 1998). Although remains have
been described subsequently from Morocco (Buffetaut, 1989;
Russell, 1996), Tunisia (Bouaziz et al., 1988; Buffetaut and
Ouaja, 2002) and Algeria (Taquet and Russell, 1998), none of
them have significantly furthered knowledge about this unusual
theropod. In the present paper, we describe specimens referable
to Spinosaurus cf. S. aegyptiacus that provide new information
on its anatomy, size, and relationships.
Fossil Location and History
MSNM V4047—Specimen MSNM V4047, now housed in the
collections of the Museo di Storia Naturale di Milano (MSNM
V4047), was found in southern Morocco in 1975 and remained in
a private collection until 2002. The specimen was reported to
have been found east of the town of Taouz, within the red beds
underlying the Hammada du Guir plateau, and more precisely in
the area called Kem Kem. More specific field data were not
recorded but sediment adhering to the bone is closely consistent
with the Kem Kem red sandstone both in colour, composition,
and texture. An isolated fish vertebra associated with the speci-
men is embedded between the right second premaxillary alveo-
lus and its erupting tooth and can be tentatively referred to
?Onchopristis sp. (Stromer, 1926:taf I, fig. 7), a sawfish that is
very abundant in the Kem Kem beds.
UCPC-2—Specimen UCPC-2 (University of Chicago Paleon-
tological Collection) was discovered in the Kem Kem beds in
Northern Morocco (Location: N30° 02⬘W5°12⬘, near the out-
post in Keneg ed Dal) during the 1996 expedition led by Dr.
Sereno (University of Chicago). The expedition produced a
myriad of fossils from river washes, one being a partially eroded
pair of fused nasals preserving a fluted crest. Initially thought to
be an unidentifiable fragment, the specimen remained within the
collection cases at the University of Chicago until 2002. It was
then re-evaluated thanks to its multiple key characteristics that
suggested spinosaur affinities.
Horizon—On the basis of stratigraphical and paleontological
evidence (Wellnhofer and Buffetaut, 1999), the Kem Kem “Con-
tinental Red Beds” (Russell, 1996) can be referred to the early
Cenomanian.
SYSTEMATIC PALEONTOLOGY
THEROPODA Marsh, 1881
TETANURAE Gauthier, 1986
SPINOSAUROIDEA Stromer, 1915, sensu Sereno et al., 1998
SPINOSAURIDAE Stromer, 1915, sensu Sereno et al., 1998
SPINOSAURUS cf. S. AEGYPTIACUS Stromer, 1915
Comments—The genus Spinosaurus was erected in 1915 by
Stromer (1915), with Spinosaurus aegyptiacus as type species, on
the basis of an incomplete skeleton, including a dentary and
long-spined dorsal vertebrae. Rauhut (2003) questioned this as-
sociation, claiming that the dorsal vertebrae lack the strong
pneumatization and laminae of the Baryonychinae and instead
are comparable to the dorsal vertebrae of the Allosauroidea.
The absence of these derived vertebral features in Spinosaurus,
however, may represent a plesiomorphic feature shared by Spi-
nosaurus aegyptiacus and the Allosauroidea. Consequently, we
regard the cranial material and the tall-spined dorsal vertebrae
of the type material (Stromer, 1915) as belonging to the same
individual. A second species of Spinosaurus,S. maroccanus, was
erected by Russell (1996). The holotype of S. maroccanus is
based on inadequate material: a single cervical vertebra thought
to be distinguishable from S. aegyptiacus by its “relatively
greater central and neural arch length” (Russell, 1996:fig. 9).
Rauhut (2003) attributed this difference to the more rostral po-
sition of the vertebra in the cervical series. Similarly, in our
opinion, the attribution of a snout from Algeria to S. maroccanus
(Taquet and Russell, 1998) cannot be supported. For all these
reasons we regard S. maroccanus as a nomen dubium, following
Sereno et al. (1998). The only other specimens of Spinosaurus for
which skull material is known are the type specimen of S. aegyp-
tiacus, which includes a piece of maxilla with four alveoli (not
figured but described by Stromer [1915]), and a referred maxil-
lary fragment described by Buffetaut (1989). S. aegyptiacus is the
only species of Spinosaurus that we regard as valid and to which
we refer the new specimens MSNM V4047 and UCPC-2. First-
Journal of Vertebrate Paleontology 25(4):888–896, December 2005
© 2005 by the Society of Vertebrate Paleontology
888

hand comparison of the craniodental material previously re-
ferred to Spinosaurus (isolated teeth [Bouaziz et al., 1988]; max-
illa fragment [Buffetaut, 1989]; snout [Taquet and Russell, 1998];
dentary fragment [Buffetaut and Ouaja, 2002]) with that of
MSNM V4047 and UCPC-2 do not reveal significant differences
within this genus. At present there is no evidence for the occur-
ring of more than one species of Spinosaurus in the Albian-
Cenomanian of North Africa.
DESCRIPTION
Specimen MSNM V4047
MSNM V4047 consists of a large snout, 988 mm long, pre-
served from the tip of the rostrum to the rostral portion of the
antorbital fenestra. It includes both of the premaxillae and max-
illae and the rostral part of the nasals, all well preserved in three
dimensions (Fig. 1).
Premaxillae—The conjoined premaxillae form a long and
slender rostrum that belongs to a mature animal, as shown by the
fact that the sagittal suture is fused dorsally, and is clearly visible
only at the rostral end. The tip of the rostrum bears numerous
enlarged pits (neurovascular foramina) on its outer wall (Fig.
2A) and forms a spatulate terminal rosetta, emphasized caudally
by deep emarginations. In dorsal and in ventral view, this rosetta
is almost circular in outline, whereas in Baryonyx (Charig and
Milner, 1997) and Suchomimus (Sereno et al., 1998) it is more
oval in shape, gradually tapering caudally. Caudal to the rosetta,
the articulated premaxillae taper, never exceeding a width of 80
mm. At the level of the external nares they measure only 29 mm
in width. In ventral view, the medial portions of the premaxillae
form a pair of elongate but massive elements (“stout ridges”in
Charig and Milner, 1997:16), which are clearly divided from the
lateral dentigerous portions by a premaxillary groove. Rostrally,
the two elements meet medially with a strongly interdigitate me-
dian suture, about 60 mm long, whereas caudally they are sepa-
rated by a narrow, deep, median gap. Contrary to the premax-
illae referred to Spinosaurus maroccanus (Taquet and Russell,
1998), which bear 7 alveoli, MSNM V4047 bears 6 alveoli on
each side. It is questionable whether varying premaxillary tooth
count can be regarded as a diagnostic feature. In Baryonyx
(Charig and Milner, 1997) there are 6 teeth on the left and 7 on
the right premaxilla. In MSNM V4047, alveolus 1 is much smaller
than its counterpart in the Baryonychinae, whereas it is compa-
rable to that of S. maroccanus and to that of Angaturama (Kell-
ner and Campos, 1996); alveoli 2 and 3 are the largest; alveoli 4
and 5 are coupled and separated from the other premaxillary
teeth by two short diastemata. A third larger, asymmetrical dia-
stema (up to 76 mm on the right side) is present between alveo-
lus 6 and the maxillary teeth. All the alveoli having a diameter
less than 35 mm (measurement that approaches the maximum
width of the premaxillary dentigerous portion) are circular. This
width represents a constraint that forced the largest alveoli (left
2 and right 2 and 3) to grow farther along their mesiodistal axis
and to become slightly compressed labiolingually. The preserved
portions of the tooth crowns (left 3 and right 3) closely resemble
the teeth previously assigned to Spinosaurus (Stromer, 1915;
Bouaziz et al., 1988), both being nearly straight, elongate, coni-
cal, and sub-circular in transverse section. In lateral view, the
dentigerous margin of the premaxilla resembles S. maroccanus
(Taquet and Russell, 1998) in being strongly downturned to-
wards the tip, such that the front of the rostrum is not elevated
above the line of the maxillary tooth row, as in Baryonyx (Charig
and Milner, 1997), Suchomimus (Sereno et al., 1998), and An-
gaturama (Kellner and Campos, 1996). In the diastema region,
the rostrum has a sub-oval cross section with minimum circum-
ference of 303 mm; the dentigerous margin is smoothly concave
and, at scale, seems to fit on the convex dorsal margin of the
rostral portion of the dentary figured by Stromer (1915:taf I, fig.
12a). As in S. “maroccanus”(pers. obs.), the lateral wall of the
diastema region is marked by three depressions. Judging by
Stromer’s figures (1915:taf I, fig. 12b) we can infer that the ros-
tral portion of the dentary of Spinosaurus was mediolaterally
wider than its premaxillary counterpart (the diastema region), so
that the largest dentary teeth (2–4) were visible when the jaws
closed, occupying the above-mentioned three depressions.
Maxillae—Due to the intimate rostral intrusion of the laminar
rostromedial processes of the maxillae (see below), the premax-
illary-maxillary connection is very complex. In lateral view, the
rostral margin of the maxillae rises caudally at some 40 degrees
from its tip to the level of the second maxillary alveolus. Above
the third maxillary alveolus, the two bones interlock via two
peg-like processes, immediately below a foramen that is homolo-
gous with the subnarial foramen of other Saurischia. Caudal to
the foramen, the maxillary margin curves and flattens to project
horizontally along the ventral margin of the premaxilla. The lat-
eral surface of each maxilla bears a complete row of large vas-
cular foramina that runs parallel to the alveolar margin. In ven-
tral view, a deep septum formed by two unfused vertical laminae
(290 mm long) emerges from the thin median gap that divides
the massive premaxillary median elements. Although some au-
thors interpreted these laminae as the rostral portion of the
paired vomers (Charig and Milner, 1997), these bones are not
fused and must be regarded as rostromedial processes of the
maxillae (“anteromedial processes of the maxillae”in Sereno et
al., 1998). This is demonstrated by the fact that in MSNM V4047
there is a clear bone continuity, via a very thin caudal passage,
between the laminae and the maxillary rami (Fig. 2B). The ros-
tralmost portions of the rostromedial maxillary processes do not
contact medially; thus, it is possible to see underneath the gap
between the medial rami of the articulated premaxillae. The
labial edges of the maxillae appear wavy and nearly parallel each
other, with the caudal half only slightly wider than the rostral
one. In ventral view, the lingual half of each maxilla appears to
be divided from the labial half by a wavy groove. Because of the
incompleteness of other spinosaurid taxa, this groove was inter-
preted as a suture and the lingual half of the maxilla was misin-
terpreted as a vomer (Taquet and Russell, 1998). Nonetheless, at
a closer view of the broken caudal edges of each maxilla of
MNSM V4047 in cross section (Fig. 2C), we see that the groove
does not penetrate the bone. The maxillary grooves and their
premaxillary continuations contain the resorption pits of the
teeth that in turn expose several replacement teeth (Fig. 2E). On
the medial wall of the labial half of the maxilla, some irregular
rugose areas can be seen through the grooves. Those areas cor-
respond to the not well-defined interdental plates of Spinosaurus
cf. S. aegyptiacus (Buffetaut, 1989) and S. maroccanus (Taquet
and Russell, 1998; pers. obs.). The maxillae meet broadly along
the midline, forming a deep, acutely arched (35°to 40°) second-
ary palate that matches the subtriangular pattern of their outer
wall. The vomers and the palatines are lacking, but the attach-
ment areas of the latter are preserved at the level of the tenth
alveolus as symmetrical scars on the caudomedial wall of the
maxillae. Caudally, the maxillae are broken off at the contact
with the jugals; on the right lateral side, the notch that received
the forward-pointing maxillary process of the jugal can be seen
(Fig. 2D). As a result of the fracture, only the rostral parts of the
antorbital fossae and fenestrae are visible. The preserved part of
the rostromedial wall of the antorbital fossa resembles that of
Suchomimus in being confined to the rostral end of the antor-
bital fenestra (Sereno et al., 1998). The maxilla of MSNM V4047
additionally resembles that of Suchomimus (Sereno et al., 1998)
in having a simple conical pneumatocoel that extends rostrally
into the body of the maxilla. A complete series of 12 subcircular
alveoli is preserved on both maxillae: they are nearly identical in
shape and spacing to the maxillary alveoli in the other specimens
DAL SASSO ET AL.—SKULL OF SPINOSAURUS 889

FIGURE 1. Photos and line drawings of MSNM V4047 in dorsal, lateral and ventral views. Abbreviations:antfe, antorbital fenestra; antfo,
antorbital fossa; aj, articular surface for jugal; ap, articular surface for palatine; en, external naris; m, maxilla; m1–12,1
st
–12
th
maxillary alveoli; mg,
maxillary groove; n, nasal; npm, nasal process of maxilla; p1–6,1
st
–6
th
premaxillary alveoli; pg, premaxillary groove; pm, premaxilla; rpm, rostro-
medial processes of maxillae; sf, subnarial foramen. Scale bar equals 20 cm.
JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 25, NO. 4, 2005890

FIGURE 2. Close-ups of MSNM V4047. Abbreviations as in Figure 1. A, rostral view of the rostrum, showing the particularly developed pits
emerging on the outer wall of the premaxillae. Scale bar equals 5 cm. B, ventral view of the specimen, showing the bone continuity (arrow) between
the lingual halves of the maxillae and their thin rostro-medial processes. Scale bar equals 2 cm. C, cross section of the maxillae, viewed from upside
down. Their caudal fracture highlights that the maxillary groove (arrow) does not penetrate the bone, thus it cannot be misinterpreted as a suture.
D, right lateral view of the caudal portion of the snout, showing the preserved part of the antorbital fossa and fenestra, and the notch for the jugal
attachment. E, right maxillary groove (arrow) with the erupting maxillary tooth 6. F, left external naris.
DAL SASSO ET AL.—SKULL OF SPINOSAURUS 891

of Spinosaurus (Stromer, 1915; Buffetaut, 1989; Taquet and Rus-
sell, 1998). As in S. maroccanus (Taquet and Russell, 1998), their
size increases abruptly from 1 to 4 (the basal circumference of
the teeth increases from 42 mm to 146 mm), and decreases
gradually from 5 to 12 (in S. maroccanus only alveoli 1 to 9 are
preserved), resulting in an interdental gap that approaches in
length the alveolar diameter toward the rear of the maxillae.
Except for the replacement teeth and one fully grown tooth (left
4), the tooth crowns are missing or broken at their bases. The
fourth left tooth crown is preserved from its base to one third of
its reconstructed height, it also resembles the premaxillary teeth
yet it is a bit more recurved. The roots are deeply implanted and
converge medially, occupying almost the complete depth of the
maxillae. As in the premaxillae, the largest teeth and alveoli (3 to
5 right and 3 to 4 left) were forced to grow compressed labiolin-
gually by the width of the maxillae. The emerging tips of the
replacement teeth are nearly straight, slightly flattened labiolin-
gually and possess carinae lacking serrations. The thin enamel
layer, where preserved, bears fine vertical ridges, denser lin-
gually than labially: this fluting is variably present in some of the
isolated teeth referred to Spinosaurus (Bouaziz et al., 1988).
External Nares—The external nares, which perforate the
snout bilaterally, occur as a pair of openings that are remarkably
small relative to the snout size and compared to those of other
theropods (Rauhut, 2003). An autapomorphy of Spinosaurus re-
vealed in this specimen is the position of the external nares,
which are dramatically retracted to the level of maxillary alveoli
9–10. Unlike the condition in other dinosaurs (Sereno, 1999;
Holtz, 2000), including baryonychines (Charig and Milner, 1997;
Sereno et al., 1998) and perhaps Irritator (Sues et al., 2002), in
MSNM V4047 the premaxilla does not bifurcate caudally to bor-
der the nasal cavity, but rather is completely excluded from its
boundary (Fig. 2F). Uniquely, the entire concave ventral margin
of each naris is formed by the main body of the maxilla and its
thin, upturned nasal process, whereas the straight, dorsal margin
is formed by the upper one of two finger-like, rostral rami of the
nasal bone. This nasal ramus terminates rostral to the external
naris, at the level of alveoli 7–8. The shape of the external naris
is oval but terminates rostrally in an acute angle. According to
Witmer (pers. comm.), Spinosaurus may have been similar to
other tetrapods in having rostrally placed fleshy nostrils, situated
in the rostral-most portion of the very subtle narial fossa extend-
ing from the external naris all the way rostrally up to the sub-
narial foramen. The complexity of this structure and other pa-
leobiology-related features will be the basis of another investi-
gation.
Nasals—The premaxillae taper caudally and meet a median
spike of the nasals above the external nares. The preserved por-
tion of the conjoined nasals is 280 mm long. In lateral view, the
dorsal margin of the nasals projects horizontally as the morpho-
logical continuation of the narrow premaxillae, without any trace
of a crest. However, the passage from the premaxillae to the
nasals is marked, on the sagittal line, by an abrupt shift from an
inverted U-shaped to an inverted V-shaped cross-section, sug-
gesting that the inter-nasal suture might be the point of origin of
a crested structure. The nasals of Irritator rise dorsally between
the external nares and the antorbital fossa, whereas in the same
position the nasals of MSNM V4047 have a straight dorsal mar-
gin.
Specimen UCPC-2
UCPC-2 (180 mm in length, 52 mm in width at its widest point,
62 mm in height at its apex; Fig. 3) consists of the caudal portion
of a pair of conjoined narrow nasals articulated with a small
fragment of a left maxilla. In lateral profile, the specimen pre-
serves two peculiar characteristics: a ridge-like fluted crest and
an inverted-V shape in at the rostral and caudal views. In right
lateral profile (Fig. 3A, C) the specimen shows a smooth, non-
eroded, surface. The rostral-most portion also preserves a strong
sutural surface composed of horizontally oriented pores where
the nasal would join the portion of the maxilla that borders
dorsally the antorbital fenestra. Meanwhile, the caudal-most
lower portion preserves another sutural surface that is slightly
hidden by a curved protrusion and represents the attachment
area for the rostral-most portion of the lachrymal. The shape of
the rim formed by the nasal here is very much like the one in
Baryonyx (Charig and Milner, 1997). The convex lateral surface
of the nasal just above the sutural surfaces preserves fine longi-
tudinal wrinkles. These wrinkles are similar to the surface tex-
ture on the bones of MSNM V4047, but are less dense. The
convex lateral surface gradually becomes at first slightly curved
and then vertical until it forms the real crest. This transitional
point is characterized also by a change in surface texture from
wrinkled to a smooth texture marked by shallow, elongate, and
parallel depressions inclined caudally. Above this transitional
surface there is a series of oscillating convex protrusions that
form the right side of the fluted crest. The texture on the surface
of the crest is very much like the lateral transitional surface
below. This texture suggests the presence of highly vascularized
soft tissue tightly bound to it (pers. obs. on extant comparative
material). Nonetheless, erosion atop the crest does not allow for
a distinct height measurement. As expected to be consistent with
MSNM V4047, UCPC-2, in left lateral profile, shows a portion of
the maxillary border of the antorbital fenestra without traces of
the wall of the antorbital fossa. Also, the dorsoventral height of
both the nasal and maxilla is consistent, at scale, with the ideal
continuity of MSNM V4047.
In dorsal view, erosion has given way to insight into pneumat-
ics of the nasal. At least 7 foramina that housed blood vessels are
visible dorsally, each other apart, within the deep spongy bone.
Each foramen appears to run deep into the bone but not through
it and possibly forms a network within that may have regulated
blood flow. The smooth medial surfaces of the two enclosing
nasals indicate the presence of hollows within the crest. In ven-
tral view (Fig. 3B, D), UCPC-2 resembles both MSNM V4047
and Baryonyx (Charig and Milner, 1997) in having a narrow,
smooth and rather flat ventral surface. The surface also preserves
two foramina that are longitudinally stretched over the rostral-
most portion of the specimen. Additionally, there is a gradual
expansion toward the caudal end of the fossil, where it would
join with the frontal, a feature also evident in Irritator (Sues et
al., 2002) and Baryonyx (Charig and Milner, 1997). Notably, a
portion of co-ossified frontals figured by Russell (1996:fig. 18b)
and referred to Theropoda indet. shows a ventral surface that
would match perfectly the shape of UCPC-2. Considerably deep
grooves form along the edge where the sutural surfaces occur
with the maxilla and the lachrymal, as in Baryonyx (Charig and
Milner, 1997). The rostral-most portion of the sutural surface of
the lachrymal runs medial to the caudal-most portion of the one
with the maxilla; this is particularly evident alongside the pre-
served left maxilla. In rostral and caudal (Fig. 3E, F) views, a key
spinosaurid characteristic is visible, that of the inverted-V shape,
which is also present in both MSNM V4047 and in the nasal
fragments of Baryonyx (Charig and Milner, 1997). Unlike those
of Baryonyx, however, the conjoined nasals of UCPC-2 apically
form an inflated, hollow, and fluted crest. As mentioned above,
since the bone has been eroded, its height and further morphol-
ogy is not known. However, we believe the crest to be a single
crest that expands but does not diverge. Additionally, the rostral-
most portion of UCPC-2 resembles the shape of the caudal-most
portion of the nasals of MSNM V4047, which preserves a slight
lateral constriction at the apex that corresponds to a concave
surface of UCPC-2. This makes for the transition between the
nasal and the crest visible in Figure 5. Unlike Baryonyx, UCPC-2
does not have a thin parasagittal crest (Charig and Milner, 1997);
JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 25, NO. 4, 2005892

FIGURE 3. Stereophotographs (A,B, and F) and line drawings (C,D, and E) of the fused nasals UCPC-2. Aand C, right lateral profile. Band
D, ventral view. Eand F, caudal view showing inverted V-shape. Abbreviations:al, articular surface for lachrymal; am, articular surface for maxilla;
cr, crest; fl, fluted portion; fo, foramen; m, maxilla; pnc, pneumatic cavity. Scale bars equal 10 cm (A,B,C, and D)and5cm(Eand F).
DAL SASSO ET AL.—SKULL OF SPINOSAURUS 893

instead it is bulky and fluted with a presence of foramina. The
only crested theropod that comes close to this specimen is Irri-
tator (Sues et al., 2002). Sues et al. (2002) describe the crest of
Irritator to be a single entity that stems from the inter-nasal
suture. However, the skull of Irritator does not keep a level plane
like that of Spinosaurus, but rather tapers rostrally and does not
seem to have a fluted portion to the longitudinal crest. This
level-plane character is key in assigning UCPC-2 to S. cf. S.
aegyptiacus, which keeps a level plane as can be seen in both
MSNM V4047 (Fig. 4) and S. maroccanus (Taquet and Russell,
1998).
DISCUSSION
Specimen Affinities and Taxonomy
Baryonychinae and Spinosaurinae—The craniodental fea-
tures of MSNM V4047 and UCPC-2 support recognition of the
family-level taxon Spinosauridae Stromer, 1915, as defined and
diagnosed by Sereno et al. (1998) and discussed by Sues et al.
(2002). According to the phylogenetic analysis of the Spinosau-
ridae by Sereno et al. (1998), within this derived clade of basal
Tetanurae, two taxa have been recognized, the Baryonychinae
and the Spinosaurinae. Herein, a comparison of previously de-
scribed spinosaurid specimens (Table 1) with MSNM V4047 and
UCPC-2 supports and strengthens the monophyly of the Spino-
saurinae. We agree with Sereno et al. (1998) in recognizing some
features of the snout that differentiate the Baryonychinae (pre-
maxillary alveolus 1 slightly smaller in diameter than alveoli 2
and 3; curved tooth crowns; teeth with fine serrations) from the
Spinosaurinae (premaxillary alveolus 1 less than one half of the
diameter of alveoli 2 and 3; unserrated teeth; tooth crowns
hardly curved or straight). By comparing the snouts (Fig. 4), we
have found an additional difference between the Baryonychinae
(external naris retracted to the first half of the maxillary tooth
row) and the Spinosaurinae (external naris retracted farther cau-
dally). The maxillary tooth count could be an additional distinc-
tive feature, Suchomimus having 22 maxillary teeth and the Spi-
nosaurinae 12 well-spaced maxillary teeth; however, the caudal
portion of the maxilla in Baryonyx is not known, so the presence
of a high number of maxillary teeth in the Baryonychinae can be
inferred only on the basis of the strong resemblance of the lower
jaws of both Baryonyx and Suchomimus, which bear more than
30 teeth (a clear autapomorphy of that taxon). Moreover, the
external nares of the Baryonychinae seem to be larger than in
the Spinosaurinae, but their exact size and shape cannot be es-
tablished because in Baryonyx and Suchomimus the rostral por-
tion of the nasals is missing. As mentioned in the description, the
nasal crest of Spinosaurus clearly differs from that of Baryonyx.
The features of the nasal crest may eventually characterize the
Spinosaurinae, but the poor preservation in Irritator renders a
diagnosis on the basis of this element impossible at the moment.
Irritator and Spinosaurus—The Spinosaurinae include Spi-
nosaurus,Irritator, and Angaturama. According to several au-
thors (Charig and Milner, 1997; Sereno et al., 1998; Buffetaut and
Ouaja, 2002; Sues et al., 2002), Angaturama limai is a junior
synonym of Irritator challengeri, as the two holotypes in all like-
lihood pertain to the same taxon (Fig. 4c, d). For this reason,
information about the rostrum in Irritator is here taken from
Angaturama (Kellner and Campos, 1996).
Due to the lack of adequate cranial material of Spinosaurus
(Fig. 4a), Sues et al. (2002) suggest that Irritator could be con-
generic with it. On the basis of both MSNM V4047 and UCPC-2,
we can confirm that the two taxa are closely related but show
also some differences. In particular, in Irritator the external nares
are retracted to the mid-part of the maxillary tooth row; the
premaxillae bifurcate caudally and their ventral rami participate
to the rostral margin of the external nares (although their con-
tribution in bordering the external nares seems to be markedly
lower than in the Baryonychinae); the dentition is less massive
than in Spinosaurus, as pointed out by Taquet and Russell
(1998); in ventral view the labial margins of both premaxillae and
FIGURE 4. Comparison among the known spinosaurid snouts: a,Spi-
nosaurus maroccanus (Taquet and Russell, 1998); b,Spinosaurus cf. S.
aegyptiacus MSNM V4047; c,Angaturama (Kellner and Campos, 1996);
d,Irritator (Sues et al., 2002); e,Suchomimus (Sereno et al., 1998); f,
Baryonyx (Charig and Milner, 1997). Shaded areas represent unknown
parts of the snout; missing elements of Spinosaurus maroccanus and
Baryonyx are based respectively on MSNM V4047 and Suchomimus;
Irritator and Angaturama are shown as two parts of the same taxon (see
text). Scale bar equals 20 cm.
TABLE 1. List of the spinosaurid specimens used in the comparison of
craniodental features
Spinosaurinae MSNM V4047 Spinosaurus cf. aegyptiacus
UCPC-2 Spinosaurus cf. aegyptiacus
IMGP 969-1 Spinosaurus cf. aegyptiacus (Buffetaut,
1989)
MNHN SAM 124 Spinosaurus maroccanus (Taquet
and Russell, 1998)
USP GP/2T-5 Angaturama limai (Kellner and
Campos, 1996)
SMNS 58022 Irritator challengeri (Sues et al., 2002)
Baryonychinae BMNH R9951 Baryonxy walkeri (Charig and Milner,
1986, 1997)
MNHN GDF 366 Cristatusaurus lapparenti (Taquet
and Russell, 1998)
MNN GDF501 Suchomimus tenerensis (Sereno et al.,
1998)
Institutional Abbreviations—MSNM, Museo di Storia Naturale di Mi-
lano; UCPC, University of Chicago Paleontological Collection; IMGP,
Institut und Museum für Geologie und Paläontologie of the Georg-
August-Universität(Göttingen); MNHN, Muséum National d’Histoire
Naturelle (Paris); USP, Universidade de São Paulo; SMNS, Staatliches
Museum für Naturkunde Stuttgart; BMNH, The Natural History Mu-
seum (London); MNN, Musée National du Niger (Niamey).
JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 25, NO. 4, 2005894

maxillae are subparallel, without marked constriction caudal to
the rosette; in lateral view the rostrum is less elongated than in
Spinosaurus. In Spinosaurus the external nares are retracted to
the level of the caudal half of the maxillary tooth row; the max-
illae entirely border the rostro-ventral and caudal margins of
external nares, excluding both premaxillae and nasals; the ros-
tral-most part of the rostrum is strongly downturned and reaches
the level of the dentigerous margin of the maxillary tooth row,
whereas in Irritator the ventral margin of the premaxilla is el-
evated above the maxillary tooth row as in the Baryonychinae.
Finally, the skull of Irritator tapers rostrally whereas that of Spi-
nosaurus maintains a level plane. Therefore, on the basis of the
above-mentioned differences, the separation of the two genera is
clearly warranted.
Skull Size and Hypothetical Body Size
The exceptional size of MSNM V4047 indicates that it repre-
sents the largest known spinosaurid skull (Fig. 4). The snout of
Suchomimus, measured from the tip of the rostrum to the notch
for the jugal attachment, is only 60% that of MSNM V4047,
whereas the reconstructed snout of Irritator is less than half that
size. MSNM V4047, measured both from the tip of the rostrum
to the caudal margin of maxillary alveolus 7 and from one lateral
margin to the other (at the level of the maxillary alveolus 7), is
about 21.5–24.5% larger than the snout of S. maroccanus. Our
tentative reconstruction of the skull (Fig. 5B), based on MSNM
V4047, UCPC-2 and other spinosaurid specimens (Fig. 5A),
gives a total skull length of about 175 cm. The toothed half of the
lower jaw is from the holotype (Stromer, 1915), and the sister-
taxon of Spinosaurus,Irritator (Sues et al., 2002), was used for
the unknown part of the skull and the rear portion of the lower
jaw. Some parts of the parietals, not preserved in Irritator, are
from the Baryonychinae (Charig and Milner, 1997; Sereno et al.,
1998). Due to the uncertain maturity and small size of Irritator
(Sues et al., 2002), in drawing some bones we have also taken
into consideration the degree of variation of shape and robust-
ness relative to age and size in the skull of other theropods, as
well demonstrated in tyrannosaurids (Carr, 1999 ; Currie, 2003).
With regard to body size, a comparison between the known el-
ements (lower jaw, ribs and dorsal centra) of the holotype of S.
aegyptiacus (Stromer, 1915) with the Baryonychinae (Sereno et
al., 1998; Charig and Milner, 1997), suggests that it was about
20–30% larger than Suchomimus and Baryonyx, rivalling in size
other giant theropods such as Tyrannosaurus (Brochu, 2003) and
the Carcharodontosauridae (Sereno et al. 1996; Calvo and Coria,
2000). Spinosaurus specimen MSNM V4047 is roughly 20% big-
ger than the holotype (Stromer, 1915); therefore, it represents
potentially the largest known theropod dinosaur. As some post-
cranial elements (i.e., the limb bones and the caudal vertebrae)
are hitherto unknown in Spinosaurus, it is difficult to reconstruct
accurately its body proportions, so the real size of MSNM V4047
can be only tentatively hypothesised. With an appropriate de-
gree of caution, the size of the whole animal (Fig. 5C) can be
calculated by reconstructing the skeleton on the basis of both
the remains of the holotype of Spinosaurus (Stromer, 1915) and
Suchomimus (Sereno et al., 1998). The estimated length for
MSNM V4047 is about 16–18 m, and presuming that the body
proportions were the same as for Suchomimus (Sereno et al.,
1998), using Seebacher’s (2001) method we obtain a weight
around 7–9t.
ACKNOWLEDGMENTS
We thank T. Holtz Jr. for the revision of a first version of the
manuscript. Many thanks to four anonymous referees, as well as
to M. Carrano, D. Naish, O. Rauhut, S. Sampson, and L. Witmer
for their critical and helpful comments. D. Naish, S. Sampson,
and P. Sereno provided useful suggestions on the presentation of
the manuscript. We thank also P. Taquet for access to specimens
MNHM SAM 124 and MNHN GDF 366, P. Sereno for access to
specimen MNN GDF501, F. Bacchia for information about the
Kem Kem area, A. Cau for unpublished data about theropod
phylogeny, L. Magnoni and D. Affer for specimen preparation,
L. Spezia for photos, and E. Bianchi for helpful discussion on
extant comparative material. M. Mendez thanks P. Sereno, who
offered him the opportunity for a great start on a career in pa-
leontology. Drawings are by M. Auditore (Fig. 1), S. Maganuco
(Figs 4, 5) and M. Mendez and C. Abraczinskas (Fig. 3).
LITERATURE CITED
Bouaziz, S., E. Buffetaut, M. Ghanmi, J.-J. Jaeger, M. Martin, J.-M.
Mazin, and H. Tong. 1988. Nouvelles découvertes de vertébrés fos-
siles dans l’Albien du Sud tunisien. Bulletin de la Société.
Géologique de France (8) 4:335–339.
Brochu C.A. 2003. Osteology of Tyrannosaurus rex: Insight from a
FIGURE 5. A, explanatory sketch showing on which material the skull reconstruction here proposed is based. The snout and the crested nasals
(photos) pertain respectively to MSNM V4047 and UCPC-2, the dentary (black) is based on the holotype of Spinosaurus aegyptiacus, and the
remaining parts of the skull (grey) are modified from Irritator. For more details see the text. B, reconstruction of the skull of Spinosaurus cf. S.
aegyptiacus. With an estimated skull length of 175 cm, it represents the largest known spinosaurid skull and one of the largest theropod skulls. Scale
bar equals 20 cm. C, estimated size of MSNM V4047 (body length about 17 m) compared, from left to right, with Homo sapiens and the largest known
individuals of Giganotosaurus (Calvo and Coria, 2000), Tyrannosaurus (Brochu, 2003) and Suchomimus (Sereno et al., 1998). Scale bar equals 2 m.
DAL SASSO ET AL.—SKULL OF SPINOSAURUS 895

Nearly Complete Skeleton and High-Resolution Computed Tomo-
graphic Analysis of the Skull. Society of Vertebrate Paleontology
Memoir 7:1–138.
Buffetaut, E. 1989. New remains of the enigmatic dinosaur Spinosaurus
from the Cretaceous of Morocco and the affinities between Spino-
saurus and Baryonyx. Neues Jahrbuch für Geologie und Paläontolo-
gie, Monatshefte 1989:79–87.
Buffetaut, E., and M. Ouaja. 2002. A new specimen of Spinosaurus (Di-
nosauria, Theropoda) from the Lower Cretaceous of Tunisia, with
remarks on the evolutionary history of the Spinosauridae. Bulletin
de la SociétéGéologique de France 173:415–421.
Calvo, J. O., and R. Coria. 2000. New specimen of Giganotosaurus caro-
linii (Coria & Salgado, 1995), supports it as the largest theropod ever
found. Gaia 15:117–122.
Carr, T. D. 1999. Craniofacial ontogeny in Tyrannosauridae (Dinosauria,
Coelurosauria). Journal of Vertebrate Paleontology 19:497–520.
Charig, A. J., and A. C. Milner. 1997. Baryonyx walkeri, a fish-eating
dinosaur from the Wealden of Surrey. Bulletin of the Natural His-
tory Museum, London, Geology Series 53:11–70.
Currie, P. J. 2003. Allometric growth in tyrannosaurids (Dinosauria:
Theropoda) from the Upper Cretaceous of North America and
Asia. Canadian Journal of Earth Sciences 40:651–665.
Gauthier, J. A. 1986. Saurischian monophyly and the origin of birds; pp.
1–55 in K. Padian (ed.), The Origin of Birds and the Evolution of
Flight. Memoirs of the California Academy of Sciences 8.
Holtz, T. R., Jr. 2000. A new phylogeny of the carnivorous dinosaurs.
Gaia 15:5–61.
Kellner, A. W. A., and D. de A. Campos. 1996. First Early Cretaceous
theropod dinosaur from Brazil with comments on Spinosauridae.
Neues Jahrbuch für Geologie und Paläontologie, Abhandlungen,
Stuttgart 199:151–166.
Marsh, O. C. 1881. Principal characters of American Jurassic dinosaurs.
Part V. American Journal of Science 21:417–423.
Rauhut, O. W. M. 2003. The interrelationships and evolution of basal
theropod dinosaurs. Special Papers in Palaeontology 69:1–213.
Russell, D. A. 1996. Isolated dinosaur bones from the middle Cretaceous
of the Tafilalt, Morocco. Bulletin du Muséum National d’Histoire
Naturelle, Paris, Série 4 18:349–402.
Seebacher, F. 2001. A new method to calculate allometric length-mass
relationships for dinosaurs. Journal of Vertebrate Paleontology 21:
51–60.
Sereno, P. C. 1999. The evolution of dinosaurs. Science 284:2137–2147.
Sereno, P.C., D. B. Dutheil, M. Iarochene, H. C. E. Larsson, G. H. Lyon,
P. M. Magwene, C. A. Sidor, D. J. Varricchio, and J. A. Wilson.
1996. Predatory dinosaurs from the Sahara and Late Cretaceous
faunal differentiation. Science 272:986–991.
Sereno, P. C., A. L. Beck, D. B. Dutheuil, B. Gado, H. C. Larsson, G. H.
Lyon, J. D. Marcot, O. W. M. Rauhut, R. W. Sadleir, C. A. Sidor, D.
Varricchio, G. P. Wilson, and J. A. Wilson. 1998. A long-snouted
predatory dinosaur from Africa and the evolution of spinosaurids.
Science 282:1298–1302.
Stromer, E. 1915. Ergebnisse der Forschungsreisen Prof. E. Stromers in
den Wüsten Ägyptens. II. Wirbeltier-Reste der Bahariye-Stufe (un-
terstes Cenoman). 3. Das Original des Theropodes Spinosaurus ae-
gyptiacus nov. Gen., nov. Spec. Abhandlungen der Königlich Bay-
erischen Akademie der Wissenschaften, Mathematisch-Physika-
lische Klasse, München 28:1–28.
Stromer, E. 1926. Ergebnisse der Forschungsreisen Prof. E. Stromers in
den Wüsten Ägyptens. II. Wirbeltier-Reste der Baharîje-Stufe (un-
terstes Cenoman). 7. Stomatosuchus inermis Stromer, ein schwach
bezahnter Krokodilier. 8. Ein Skelettrest des Pristiden Onchoprisitis
numidus Haug sp. Abhandlungen der Bayerischen Akademie der
Wissenschaften, Mathematisch-naturwissenschaftliche Abteilung,
30:1–22.
Sues, H.-D., E. Frey, D. M. Martill, and D. M. Scott. 2002. Irritator
challengeri, a spinosaurid (Dinosauria: Theropoda) from the Lower
Cretaceous of Brazil. Journal of Vertebrate Paleontology 22:
535–547.
Taquet, P. 1984. Une curieuse spécialisation du crâne de certains Dino-
saures carnivores du Crétacé: Le museau long et étroit des Spino-
sauridés. Comptes Rendus de l’Académie des Sciences, Paris, série
II, 299:217–222.
Taquet, P., and D. A. Russell. 1998. New data on spinosaurid dinosaurs
from the Early Cretaceous of the Sahara. Comptes Rendus de
l’Académie des Sciences, Paris, Sciences de la Terre et des Planètes
327:347–353.
Wellnhofer, P., and E. Buffetaut. 1999. Pterosaur remains from the Cre-
taceous of Morocco. Paläontologische Zeitschrift 73:133–142.
Submitted 10 March 2005; accepted 25 May 2005.
JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 25, NO. 4, 2005896