Although nonavian theropod have received considerable interest in the last years, their ontogeny still remains poorly understood, especially the ontogenetical changes affecting their skull (Rauhut
and Fechner, 2005). The quadrate, for instance, is preserved in several embryos and juvenile specimens belonging to many clades of theropods such as the Tyrannosauridae (Carr, 1999), Compsognathidae (Dal Sasso and Maganuco, 2011), Therizinosauroidea (Kúndrat et al., 2007), Oviraptoridae (Norell et al., 1994; Norell et al., 2001; Weishampel et al., 2008) and Troodontidae (Varrichio et al., 2002) but very little is usually said about the anatomy of this bone and no one has ever investigated ontogenetical variation in the nonavian theropod quadrate. The discovery of two quadrates belonging to embryos of the sinraptorid Lourinhanosaurus antunesi from Portugal and five isolated quadrates pertaining to juvenile, subadult and adult specimens of Spinosauridae from Morocco fills this gap and allows some ontogenetic information to be drawn for this bone in these two specific clades of Theropoda.
I
Quadrate ontogeny in Sinraptoridae
Two isolated quadrates (ML565-10; ML565-150; fig. 1 and 2)
were discovered among the skeletal remains of several embryos
ascribed to the sinraptorid (Allosauroidea) Lourinhanosaurus
antunesi (Mateus et al., 1998; Ricqlès et al., 2001; Mateus, 2005)
from the Lourinhã Formation (Kimmeridgian – Tithonian, Upper
Jurassic) of Portugal. Formerly regarded as a basal Allosauroidea
(Mateus, 1998) and a eustreptospondylid (Mateus, 2005; Mateus et
al., 2006), Lourinhanosaurus antunesi is currently classified among
Sinraptoridae (Benson, 2009). Absence of cranial material in the
holotype of Lourinhanosaurus autunesi, a mature specimen, does not
allow direct comparison between the embryos and adult, thereby the
two quadrates have been compared to the well-preserved quadrates
of the most closely related taxon Sinraptor dongi (Currie, 2006).
ML565-150 (Fig. 1) and ML565-10 (Fig. 2) are two incomplete
left quadrates missing the dorsal part of the quadrate body, the cotylus,
and part of the pterygoid flange. ML565-10 is the best preserved
one but the bone was lost in the 1990s and only one photograph
and two drawings of it remain (Fig. 2). The bone surface of the
second remaining quadrate ML565-150 has been damaged during
preparation but the general morphology is still preserved.
Although incomplete, the two quadrates share numerous features
with the quadrate of allosauroids, and especially Sinraptor dongi. Such
as basal allosauroids, the pterygoid flange is slightly medially curved
and gets attached to the quadrate body well-above the mandibular
articulation, the ventral quadratojugal contact shows a well-developed
quadratojugal process projecting anteriorly and the posterior fossa
is deep, strongly ventro-dorsally elongated and positioned on the
quadrate body in between the quadrate shaft and the quadratojugal
contact. The quadrate shaft is rod-shaped and well-delimited at one half
of the quadrate body and the shaft gets flared dorsal to the entocondyle
without reaching the later, like in Allosaurus sp. (SMA 127) but different
from Sinraptor dongi (IVPP 10600). As seen in the neovenatorid
Aerosteon riocoloradensis (MCNA-PV 3137) but not in allosaurids
and sinraptorids, the ventral quadratojugal contact is straight instead
of convex and the ventral margin of the pterygoid flange is not folded
medially or medio-dorsally. Finally, ML565-10 and ML565-150 share
with Sinraptor dongi a piriform ventral quadratojugal contact facing
postero-laterally, a dorsal quadratojugal contact forming an elongated
line and laterally positioned, and a very shallow medial fossa.
When directly compared to other allosauroid taxa, the two
quadrates show a few differences that can most likely be linked
to the embryonic stage of the individuals. Unlike all allosauroids,
the quadrate foramen is absent so that both ventral and dorsal
quadratojugal contact are connected. In addition, rather than
displaying the typical allosauroid mandibular articulation displaying
two elliptical, globular and prominent condyles separated by a
deep intercondylar sulcus, the mandibular articulation of the two
embryonic quadrates is made of two shallow and not well-delimited
ento- and ectocondyle barely separated by a broad and very shallow
intercondylar sulcus. The latter seems to run perpendicular to the long
axis passing through the mandibular articulation on the contrary of
the diagonally oriented intercondylar sulcus of allosauroids. Whether
the absence of a medially-folded ventral margin of the pterygoid
flange and a short parabolic flange projecting laterally just above
the ventral quadratojugal contact, two features present in Sinraptor
dongi, is an intertaxic variation among sinraptorids or an ontogenetic
feature cannot be determined.
Quadrate ontogeny in Spinosauridae
Five isolated quadrates from the Kem Kem beds (lower
Cenomanian, Upper Cretaceous) of South-eastern Morocco (MSNM
V6896; Eldonia Coll.) are determined to be from a same taxon of
Spinosauridae based on the presence of a thick and prominent quadrate
shaft, a quadrate foramen ventro-dorsally elongated, an elongated
tear-drop shaped dorsal quadratojugal contact, a subquadrangular
pterygoid ala reaching the entocondyle on its ventral part, an elongated
and helicoidal ectocondyle, and a concavity on the anterior side of the
entocondyle, a combination of features only visible in the quadrates of
the spinosaurids Baryonyx walkeri (BMNH R.9951) and Suchomimus
tenerensis (MNN GAD 502). Spinosaurid materials are common in
he Kem Kem beds of Morocco and this clade is only represented by
the genus Spinosaurus so far (e.g. Buffetaut, 1989; Russel, 1996; Dal
Sasso et al., 2005). Therefore, and although no quadrate have been
preserved in this taxon hitherto, the five quadrates most likely belong
to Spinosaurus sp.
Based on size and the presence and absence of features linked
to ontogenetical variations, the smallest quadrate is interpreted
to belong to a juvenile (Fig. 3) whereas three medium-sized
quadrates most likely belong to subadult specimens and the largest
one pertains to a fully grown adult. Sequence lists of ontogenetic
haracter transformations (maturity dependant characters) are the
following:
• Juvenile state (Fig. 3): Ventral and dorsal quadratojugal
contact smooth (contact with the quadratojugal weak and
perhaps mobile) and ectocondyle poorly delimited.
• Subadult state 1: Intercondylar sulcus shallow and poorly
delimited, entocondyle poorly delimited, cotyle poorly
developed, and absence of a ventral projection of the dorsal
quadratojugal suture.
• Subadult state 2 (Fig. 4): Dorsal quadratojugal suture with two
longitudinal grooves.
• Adult state: Dorsal quadratojugal contact deeply excavated
(strong contact in between the quadrate and quadratojugal)
and ectocondyle, entocondyle and cotyle well-delimited.