ArticlePDF Available

A Theropod Dinosaur from the Late Jurassic Cañadón Calcáreo Formation of Central Patagonia, and the Evolution of the Theropod Tarsus

Authors:

Abstract and Figures

A fragmentary postcranial skeleton from the Late Jurassic (Oxfordian-Tithonian) Cañadón Calcáreo Formation of Chubut, Argentina, represents a new taxon of theropod dinosaur, which is here described as Pandoravenator fernandezorum gen. et sp. nov. This material represents the first Late Jurassic theropod known from Argentina. Pandoravenator fernandezorum is characterized by strongly elongated postzygapophyses in the caudal vertebrae and an unusual tarsal joint, with the astragalus showing two distal tubercles and a very low and laterally inclined ascending process. Phylogenetic analyses indicate that the new taxon is a basal tetanuran, although its exact phylogenetic position within basal tetanurans remains uncertain, due to the fragmentary nature of the remains and the lack of consensus among the different phylogenetic analyses. The tarsus of P. fernandezorum shows an intermediate morphology between that of basal theropods and more derived tetanurans. It is especially noteworthy for the presence of a suture between the distal astragalar condyles and the anteroproximal extension of the astragalus, including the ascending process. This indicates that a separate ossification of the ascending process of the astragalus was present in this taxon, and, in a phylogenetic context, thus provides evidence that the origin of the ascending process and the astragalar body from separate ossifications was already present at the base of Averostra. © 2017 Asociacion Paleontologica Argentina. All Rights Reserved.
Content may be subject to copyright.
A THEROPOD DINOSAUR FROM
THE LATE JURASSIC CAÑADÓN
CALCÁREO FORMATION OF CENTRAL
PATAGONIA, AND THE EVOLUTION
OF THE THEROPOD TARSUS
1SNSB, Bayeri sche St aatssamml ung für Paläon tologie und Ge ologie, Departme nt of Earth and Env ironmenta l Scien ces, Ge oBioCente r, Ludw ig-Maximi lians-
University, Richard-Wag ner-Str. 10, 80333 München, Germany.
2CONICET, Museo Paleontológico Egidio Feruglio, Av. Fontana 140, U9100GYO Trele w, Argentina.
OLIVER W. M. RAUHUT1
DIEGO POL2
Also appearing in this issue:
Two new taxa unveil the
previously unrecognized diversity
of Coelophysidae in the Late Triassic
of South America.
A new ornithomimosaur taxon
from the Early Cretaceous of Niger
and new anatomical data on
Nqwebasaurus from South Africa.
Murusraptor had a brain morphology
similar to tyrannosaurids but
neurosensorial capabilities
resembling that of allosauroids.
Submitted: April 4th, 2017 - Accepted: October 12th, 2017 - Published online: November 1st, 2017
To cite this article:Oliver W.M. Rauhut, and Diego Pol (2017). A theropod dinosaur from the Late Jurassic Cañadón
Calcáreo Formation of central Patagonia, and the evolution of the theropod tarsus. Ameghiniana 54: 506–538.
To link to this article: http://dx.doi.org/10.5710/AMGH.12.10.2017.3105
PLEASE SCROLL DOWN FOR ARTICLE
539
AMGHB2-0002-7014/12$00.00+.50
A THEROPOD DINOSAUR FROM THE LATE JURASSIC
CAÑADÓN CALCÁREO FORMATION OF CENTRAL PATAGONIA,
AND THE EVOLUTION OF THE THEROPOD TARSUS
OLIVER W. M. RAUHUT1, AND DIEGO POL2
1SNSB, Bayerische Staatssammlung für Paläontologie und Geologie, Department of Earth and Environmental Sciences, GeoBioCenter, Ludwig-Maximilians-University,
Richard-Wagn er-Str. 10, 80333 München, Germany. o.rauhut@lrz.uni-muenchen.de
2CONICET, Museo Paleontológico Egidio Feruglio, Av. Fontana 140, U9100GYO Trele w, Argentina. dpol@mef.org.ar
Abstract. A fragmentary postcranial skeleton from the Late Jurassic (Oxfordian–Tithonian) Cañadón Calcáreo Formation of Chubut, Argentina,
represents a new taxon of theropod dinosaur, which is here described as Pandoravenator fernandezorum gen. et sp. nov. This material repre-
sents the first Late Jurassic theropod known from Argentina. Pandoravenator fernandezorum is characterized by strongly elongated postzy-
gapophyses in the caudal vertebrae and an unusual tarsal joint, with the astragalus showing two distal tubercles and a very low and laterally
inclined ascending process. Phylogenetic analyses indicate that the new taxon is a basal tetanuran, although its exact phylogenetic position
within basal tetanurans remains uncertain, due to the fragmentary nature of the remains and the lack of consensus among the different
phylogenetic analyses. The tarsus of P. fernandezorum shows an intermediate morphology between that of basal theropods and more derived
tetanurans. It is especially noteworthy for the presence of a suture between the distal astragalar condyles and the anteroproximal extension of
the astragalus, including the ascending process. This indicates that a separate ossification of the ascending process of the astragalus was
present in this taxon, and, in a phylogenetic context, thus provides evidence that the origin of the ascending process and the astragalar body
from separate ossifications was already present at the base of Averostra.
Key words. Theropoda. Tetanurae. Jurassic. Tarsals.
Resumen. UN DINOSAURIO TERÓPODO DE LA FORMACIÓN CAÑADÓN CALCÁREO DEL JURÁSICO TARDÍO DE PATAGONIA CENTRAL, Y LA
EVOLUCIÓN DEL TARSO DE LOS TERÓPODOS. Un esqueleto postcraneano fragmentario de la Formación Cañadón Calcáreo del Jurásico Tardío
(Oxfordiano–Titoniano) de Chubut, Argentina, representa un nuevo taxón de dinosaurio terópodo, el cual es descripto aquí como Pandorave-
nator fernandezorum gen. et sp. nov. Este material representa el primer terópodo conocido del Jurásico Tardío de Argentina. Pandoravenator
fernandezorum se caracteriza por sus postzigapófisis caudales fuertemente alargadas y una articulación tarsal inusual, con el astrágalo mostrando
dos tubérculos distales y un proceso ascendente muy bajo e inclinado lateralmente. Diversos análisis filogenéticos indican que el nuevo taxón
es un tetanuro basal, aunque su posición precisa dentro de los tetanuros basales permanece incierta debido a la condición fragmentaria de los
restos y la falta de consenso en los diferentes análisis filogenéticos del grupo. El tarso de P. fernandezorum muestra una morfología intermedia
entre la de los terópodos basales y tetanuros más derivados. Es especialmente destacable la presencia de una sutura entre los cóndilos dis-
tales y la extensión anteroproximal del astrágalo, incluyendo el proceso ascendente. Esto indica que este taxón poseía un centro de osificación
separado en el proceso ascendente del astrágalo, y en un contexto filogenético, podría proveer evidencia de que el origen del proceso ascendente
a partir de un centro de osificación separado ocurrió en la base de Averostra.
Palabras clave. Theropoda. Tetanurae. Jurásico. Tarso.
ALTHOUGH recent phylogenetic analyses indicate that most
major lineages of theropod dinosaurs reach back to the
Jurassic (e.g., Choiniere et al., 2010; Rauhut et al., 2010; Xu et
al., 2010; Pol and Rauhut, 2012; Carrano et al., 2012; Rauhut
et al., 2016), our understanding of Jurassic theropod evolu-
tion is largely based on the fossil record from the Northern
Hemisphere (Rauhut and López-Arbarello, 2008). For the
Late Jurassic, the only Gondwanan unit that has yielded
identifiable remains is the Tendaguru Formation of southern
Tanzania, from where a number of ceratosaurian and basal
tetanuran taxa are known (Janensch, 1920, 1925, 1929;
Rauhut, 2005a, 2011; Rauhut and Carrano, 2016). In South
America, Late Jurassic theropods have so far only been re-
ported from the Tithonian Toqui Formation of southern
Chile, from where Salgado et al. (2008) described fragmen-
tary remains that were recently referred by Novas et al.
(2015) to the highly unusual basal tetanuran Chilesaurus.
The latter represents the only formally named possible
GONDWANAN PERSPECTIVES
ISSN 0002-7014
AMEGHINIANA - 2017 - Volume 54 (5): 539 – 566
theropod taxon from the Late Jurassic of South America;
however, it should be noted that its theropod affinities are
disputed, as Chilesaurus has recently been re-interpreted
as an ornithischian (Baron and Barrett, 2017).
The Oxfordian–?Tithonian Cañadón Calcáreo Forma-
tion of Chubut Province, Argentina, has yielded numerous
dinosaur remains in recent years, mainly representing
sauropods. Theropods seem to be rare in this formation, but
several fragmentary specimens have been found in the
past 15 years. Here we describe a new taxon of theropod
from a new locality, Caja de Pandora, located on the west-
ern banks of the Río Chubut, some 12 km north of the village
of Cerro ndor. The new taxon is interpreted as a basal
tetanuran and bears peculiar features in the tarsus that
have implications for understanding the evolution of the
astragalus and its ascending process in Theropoda.
GEOLOGICAL AND PALAEONTOLOGICAL SETTING
The Caja de Pandora locality (Fig. 1) is situated in the
outcropping area of the Cañadón Calcáreo Formation, some
1000 m west of the fish locality of Puesto Almada (López-
Arbarello et al., 2013), in the middle course of the Río
Chubut, Chubut Province, Argentina. The Cañadón Calcáreo
Formation is a unit of terrestrial sediments that were de-
posited in the Cañadón Asfalto Basin, a large, north-north-
west–south-south-east trending depocenter in northern
central Chubut Province (Figari and Courtade, 1993; Figari,
2005). It represents the post-rift phase of the basin (Figari
and Courtade, 1993; Figari, 2005; Figari et al., 2015) and
comprises a mainly lacustrine basal part, followed by flu-
vial and overbank deposits. A tuff at the top of the lacus-
trine section at Puesto Almada has recently been dated at
157 Ma (latest Oxfordian;neo et al., 2013). The section
at this locality has been referred to as Puesto Almada
Member of the Cañadón Asfalto Formation by other work-
ers (e.g., Cabaleri et al., 2010; Gallego et al., 2011; Volkheimer
et al., 2015), but, despite the differences in nomenclature,
there is agreement in the Late Jurassic age of these rocks
(Cabaleri et al., 2010).
Due to faulting, the section in which the Caja de Pandora
locality is placed cannot directly be correlated with the
Puesto Almada locality, since there is only a restricted
outcrop of the Cañadón Calcáreo Formation in this area, and
much of the outcrop is covered by debris from Miocene
basalts that cap the formation on top. However, parts of
the outcropping sediments at the locality consist of fine-
grained, laminated sandy siltstones, which are similar to in-
tercalated lacustrine levels in the transition between the
lacustrine and fluvial part of the section at Puesto Almada,
and thus we consider it very likely that these sediments
represent this part of the section. Thus, the age of this lo-
cality probably very closely corresponds to the radiometric
date obtained at Puesto Almada (for further discussion of
the geology of the Cañadón Calcáreo Formation see Figari,
2005; Cúneo et al., 2013; López-Arbarello et al., 2013; Figari
et al., 2015).
In 2002, a fieldwork team of the Museo Paleontológico
Egidio Feruglio, led by one of us (OR), discovered two distal
caudal vertebrae, articulated with their chevron at Caja de
Pandora. The remains were collected from the surface, and
an intensive search of the area by the whole team in the
next day recovered additional caudal vertebral fragments,
but failed to identify the exact level and location where
these remains came from. In January 2009, the senior au-
thor again visited the site, and a number of additional frag-
ments, mainly of hindlimb elements, were found. Again, an
AMEGHINIANA - 2017 - Volume 54 (5): 539 – 566
540
Figure 1. Satellite image showing the Caja de Pandora locality in the
Puesto Almada area, west of the Chubut river in central Chubut
Province, Argentina. Abbreviation: Ea., Estancia.
intensive search led to the recovery of a few more bones,
but no remains could be found in situ. Finally, the distal
part of a left tibia, an astragalus and a few other fragments
(including remains of two articulated pedal phalanges)
were found by DP in 2017. Despite the disparate times of
discovery, all elements were found in the same erosional
gully in an area of six to eight square meters. Furthermore,
the recovered elements are of matching size, preservational
quality, and, in the case of bones known from both left
and right side, of matching morphology (allowing for some
deformation), so that all of this material seems to represent
a single individual. The hindlimb elements found in 2009
were recovered from an even more restricted area of some
two square meters, further strengthening the interpreta-
tion that at least these remains belong to the same ani-
mal.
Vertebrate remains have been found in situ at the Caja
de Pandora locality in a hard, silicified layer of laminated fine
sandstone of some 50 cm thickness, which is brown to dark
red in colour. Fossils found in this layer were mainly fish
(Luisiella sp.) and isolated unidentified bones. Although
the preservation and attached sediment of the theropod
bones agree with the characteristics of this layer, no thero-
pod remains could be discovered in situ, and all remains
were found some 5 m below this layer.
Apart from these remains, the Cañadón Calcáreo For-
mation has yielded an abundant fish fauna from the basal
lacustrine part (the Almada fauna; López-Arbarello et al.,
2008), including the coccolepid Condorlepis groeberi (Bor-
das, 1943; López-Arbarello et al., 2013) and the basal teleost
Luisiella feruglioi (Bordas, 1943; Bocchino, 1967; Sferco,
2012; Sferco et al., 2015), as well as the basal crocodyliform
Almadasuchus, found towards the top of the lacustrine
section at Puesto Almada (Pol et al., 2013), the sauropod
dinosaurs Tehuelchesaurus (Rich et al., 1999; Carballido et al.,
2011) and Brachytrachelopan (Rauhut et al., 2005), an
unidentified brachiosaurid (Rauhut, 2006), and a diplodocid
(Rauhut et al., 2015).
Institutional Abbreviations. FMNH, Field Museum of Natural
History, Chicago, USA; MB, Museum für Naturkunde Berlin,
Berlin, Germany; MCNA, Museo de Ciencias Naturales y
Antropológicas (J. C. Moyano), Mendoza, Argentina; MOR,
Museum of the Rockies, Bozeman, USA; MPEF, Museo Pa-
leontológico Egidio Feruglio, Trelew, Argentina.
SYSTEMATIC PALEONTOLOGY
DINOSAURIA Owen, 1842
THEROPODA Marsh, 1881
TETANURAE Gauthier, 1986
Pandoravenator fernandezorum gen. et sp. nov.
Figures 2–8
Etymology.From Pandora, referring to the type locality ‘Caja
de Pandora’ and venator, Greek for hunter. The species name
honours the Fernandez family, including Daniel Fernández
and the late Victoriano Fernández and his daughters and
sons (especially Abel). The family has helped in many ways
the exploration of the Museo Paleontológico Egidio Feruglio
on their land in the Upper Jurassic rocks of central Chubut
for more than twenty years.
Holotype.MPEF PV 1773-3, distal end of right femur and
proximal end of right tibia; 1773-6, proximal end of right
fibula; 1773-5a, b, fragment of the proximal shaft of the left
tibia and proximal sections of the shafts of the right tibia
and fibula; 1773-4, distal end of articulated right tibia,
fibula, astragalus and calcaneum; 1773-7, partial right dis-
tal tarsal 4; 1773-8, proximal end of right metatarsal IV;
1773-9, articulated left distal tarsals 3 and 4 and proximal
ends of metatarsals II to IV; 1773-10, articulated distal ends
of right metatarsals II to IV, proximal ends of phalanges II-
1 and IV-1, and ungual of digit I; 1773-11, distal end of left
metatarsal II; 1773-12, distal end of left metatarsal III;
1773-13-17, fragments of five pedal phalanges; 1773-18,
pedal ungual; 1773-27–28, mid-shaft of a slender long
bone (?fibula).
Referred material.MPEF PV 1773-21, fragments of two ar-
ticulated distal mid-caudal vertebrae; MPEF PV 1773-19,
one-and-a-half distal caudal vertebrae with articulated
chevron; MPEF PV 1773-20, parts of two articulated distal
caudal vertebrae; MPEF PV 1773-2, glenoid region of left
scapulocoracoid; MPEF PV 1773-1, distal end of left
humerus; MPEF PV 1773-36, fragment of the distal end of
the left tibia (MPEF PV 1773-36a) and articulated partial
left astragalus (MPEF PV 1773-36b); MPEF PV 1773-37,
fragments of two articulated pedal phalanges.
Comments.Although we consider all of the material to repre-
sent a single individual (see above), we restrict the holotype
RAUHUT AND POL: NEW JURASSIC THEROPOD FROM PATAGONIA
541
to the hindlimb elements in order to avoid confusion in case
future discoveries demonstrate that remains of more than
a single individual or taxon are present at this locality. As
noted above, the discovery of the hindlimb elements at the
same time in a very restricted spot and the fact that all of
these elements are of matching size and repeated elements
(mainly metatarsals and distal tarsals) are of matching
morphology leave little doubt that at least these remains
represent the same animal. The very fragmentary scapulo-
coracoid (MPEF PV 1773-2) and distal humerus (MPEF PV
1773-1) were found at the same time as the hindlimb ele-
ments just a few meters below the latter within an erosional
gully; their preservation and the slightly further displace-
ment indicate that these elements might have eroded out
some time earlier. Their size is consistent with that of the
hindlimb elements. The caudal vertebrae found in 2002
(MPEF PV 1773-19–21) were recovered from the same
general area (as noted above, within a 6–8 m2range) and,
as far as this can be established, their size is within the
range expected for an animal of the size represented by the
hindlimb elements; the referral of these remains should
thus be seen as somewhat more tentative. The fragmen-
tary tibia (MPEF PV 1773-36a), astragalus (MPEF PV 1773-
36b) and phalanges (MPEF PV 1773-37) recovered in 2017
are eroded and seem to be deformed. Nevertheless, the
elements were again found in the same small area and the
discernible morphology matches that of the holotype of
Pandoravenator.
Type locality and horizon.Caja de Pandora locality, ca. 1 km
west of the fish locality of Puesto Almada, Chubut Province,
Argentina (Fig. 1). The material was found in a section of
lacustrine and fluvial silt to sandstones that can be referred
to the basal part of the Cañadón Calcáreo Formation, latest
Oxfordian to Kimmeridgian.
Diagnosis.Small to medium-sized theropod dinosaur that
can be diagnosed by an astragalus with anterior and distal
lateral tubercles adjacent to the contact with the calcaneum,
an anteriorly facing part of the astragalar body that is offset
from the distal part by a transverse groove anteriorly, an as-
cending process of the astragalus that is laminar, but very
low (about one third of the height of the astragalar body),
triangular in outline in anterior view, and inclined laterally,
and the shaft, but not proximal end, of metatarsal III being
strongly constricted between the shafts of metatarsals II
and IV. A further potential apomorphy, based on the referred
material, is the postzygapophyses of posterior caudal ver-
tebrae being elongate to overhang approximately half of
the length of the following vertebral centrum.
DESCRIPTION
Caudal vertebrae
Most of the recovered posterior caudal vertebrae are
poorly preserved, but MPEF PV 1733-19 (Fig. 2) and MPEF
PV 1733-20, both of which include the remains of two ar-
ticulated posterior caudal vertebrae and are reasonably
well-preserved, provide information on the morphology of
these elements. The fact that both include strongly elon-
gated posterior caudal vertebrae, but are derived from dif-
ferent parts of the caudal column, as evidenced from their
quite different sizes, indicates that the tail of Pandoravena-
tor was long and slender.
MPEF PV 1733-19 represents the more anterior section
of the posterior tail. The more anterior of the two caudal
vertebrae preserved in articulation is almost complete, with
only minor parts of the anterior end missing, whereas the
more posterior element is only represented by its anterior
half. The vertebral centra are elongate, with the anterior
vertebra being 56 mm long (with an additional 1–2 mm
missing at the anterior end) and 18 mm high posteriorly.
The vertebral body is slightly wider than high (21 vs. 18 mm
at the posterior end), notably constricted ventrally (to a
minimal height of approximately 12 mm in the anterior
vertebra), but only slightly transversely (minimal width is 16
mm). Due to the articulated state of all vertebral fragments
found, the exact shape and nature of the articular ends
cannot be established, although they seem to be amphi- to
platycoelous. A weakly-developed, but notable lateral lon-
gitudinal ridge is present at about midheight of the verte-
bral body, and the ventral side of the vertebra is offset from
the lateral sides by notable edges, giving the centrum a
hexagonal outline in cross-section, as in many basal teta-
nurans. The ventral surface between these edges is flat to
very slightly concave transversely. The posterior end of the
centrum extends slightly further ventrally than the ante-
rior end to form the chevron facets. These seem to be de-
veloped as small, separate, posteroventrally directed facets
in the more anterior vertebra, where they are partially
hidden by the articulated chevron.
AMEGHINIANA - 2017 - Volume 54 (5): 539 – 566
542
On the lateral side, the ventral margin of the basis of
the massive prezygapophysis extends posteroventrally to
the midlength of the centrum, where it reaches half of its
height (Fig. 2.1–2), as in the posterior caudal vertebrae of
the noasaurid Elaphrosaurus (Rauhut and Carrano, 2016),
the basal tetanuran Allosaurus (Madsen, 1976) and basal
coelurosaurs, such as ornithomimosaurs (e.g., Osmólska et
al., 1972). Anteriorly, it is developed as a stout, rounded
ridge, which becomes more indistinct posteroventrally and,
at its posteroventral end, is separated from the longitudi-
nal lateral ridge of the centrum by a shallow, narrow, slightly
curved groove. The broken posterior end of the more pos-
terior vertebra shows the hollow interior of the centrum,
surrounded by bone walls of ca. 4 mm laterally and 2 mm
ventrally, and apparently devoid of internal septa or tra-
beculae.
The neural arch extends over almost the entire length of
the centrum, its base being separated from the anterior
margin of the latter by some 5 mm and from its posterior
end by 9 mm. It is low, being approximately 12 mm high
above its posterior base, and encloses a small, round neu-
ral canal. The prezygapophyses are massive and strongly
elongated, the elements of the more posterior vertebra of
MPEF PV 1773-19 extending over ca. 37 mm or approxi-
mately 65% of the length of the preceding vertebra (Fig. 2),
as in Allosaurus (Gilmore, 1920; Madsen, 1976) and many
coelurosaurs (e.g., Osmólska et al., 1972), but in contrast to
the generally shorter posterior caudal prezygapophyses
in Ceratosaurus (Gilmore, 1920; Madsen and Welles, 2000)
and non-averostran theropods. The prezygapophyses are
more anteriorly than dorsally directed, in contrast to the more
strongly dorsally inclined prezygapophyses in Majunga-
saurus (O’Connor, 2007). The base of the prezygapoph-
ysis of the more posterior vertebra is 11 mm high at its
highest point, just above the anterior end of the centrum,
where its ventral margin also shows a notable lateral
bulge in dorsal view. Anteriorly, the tongue-shaped prezy-
gapophyses narrow to a point. The ventral margins of the
prezygapophyses are generally more massive than the
dorsal rims, which are developed as thin laminae and are
narrowly spaced, being 6 mm apart posteriorly and 3 mm
towards their anterior ends in the more posterior vertebra
of MPEF PV 1773-19 (Fig. 2.3).
The bases of the prezygapophyses extend posteriorly
to about the midlength of the centrum, where their dorsal
margins merge into the roof of the neural arch. In this area,
the neural arch extends slightly dorsally towards the base
of the postzygapophyses. Although the exact morphology
of the latter is difficult to establish, due to the fully articu-
lated preservation, the postzygapophyses seem to be placed
on a long, single, rod-like posterior process that extends
posteriorly to about the half-length of the following verte-
RAUHUT AND POL: NEW JURASSIC THEROPOD FROM PATAGONIA
543
Figure 2. Caudal vertebrae of Pandoravenator fernandezorum, MPEF
PV 1733-19; 1–2, outline drawing and photograph in left lateral
view; 3, photograph in dorsal view. Abbreviations: ap, anterior
process; ch, chevron; dg, dorsal groove; prz, prezygapophysis; poz,
postzygapophysis of the preceding vertebra; ri, lateral longitudinal
ridge on the vertebral centrum. Scale bar= 10 mm.
bral centrum. There is no neural spine, but the dorsal side
of this rod-like process is excavated by a narrow, but deep
longitudinal groove (Fig. 2.3), as in Condorraptor (Rauhut,
2005b) and Allosaurus (Madsen, 1976).
Two articulated fragments of probably posterior middle
caudal vertebral centra (MPEF PV 1773-21) show a convex
ventral surface and a hollow interior of the centrum, but
otherwise do not provide any useful morphological infor-
mation. The more posterior caudal vertebrae MPEF PV
1773-20 are closely comparable to, but less well-preserved
than the vertebrae described above. The lateral longitudinal
ridge and the chevron facets are less distinctive than in the
more anterior vertebrae, and, with an anterior width of 14
mm and a height of ca. 16 mm, the vertebral body is slightly
higher than wide.
Haemapophysis
A single haemapophysis is preserved in articulation
with the caudal vertebrae of MPEF PV 1733-19. It is almost
complete, missing only the distalmost tip. The haemapophy-
sis is dorsoventrally short and anteroposteriorly expanded
distally, as in Allosaurus (Madsen, 1976) and other basal
tetanurans. Whereas the anterior expansion and most of
the ventral border are rounded, the larger posterior expan-
sion seems to have been pointed, but its tip is broken off.
Proximally, the expansion is separated from the slender ar-
ticular processes for the vertebrae by a marked, angular
“shoulder” posteriorly, and the anterior process anteriorly,
the presence of which represents a theropod synapomor-
phy (Rauhut, 2003). The dorsal margin of the anterior
process is aligned with the dorsal margin of the posterior
“shoulder”. The anterior process is more elongated than
in Allosaurus (Madsen, 1976) and tapers to a sharp point
anteriorly.
Pectoral girdle
The glenoid region of the left scapulocoracoid is pre-
served (MPEF PV 1773-2). Although scapula and coracoid
are preserved in articulation, they are not fused, the suture
between the two elements being well visible on both me-
dial and lateral sides (Fig. 3). Unfortunately, the fragment is
too incomplete to say anything about the general shape of
the scapulocoracoid, but some details of the glenoid and its
surrounding areas are discernible.
The broken base of the scapular blade is oval in cross-
section, narrowing posteriorly. It is rather massive, being
13 mm wide transversely at the widest part. The medial
side of the scapula bulges medially towards the glenoid, al-
though the exact extent of this bulge cannot be established,
since the surface of the bone is abraded in this area. On the
lateral side, a well-developed lateral lip is present at the
posterior end of the glenoid facet (Fig. 3). Between the me-
dial bulge and lip, the ventral surface just behind the glenoid
forms a wide, triangular surface. Unlike the situation in most
theropods, in which this area is transversely convex, this
surface is flat, although this might be due to abrasion in this
area. The scapular glenoid facet accounts for 14 mm of the
total 24 mm of the glenoid. Unlike the situation in basal
theropods, in which the glenoid facet is usually mainly an-
teroventrally directed, it is more laterally directed (Fig. 3.3),
at an angle of approximately 40° from the vertical, similar to
the condition in maniraptorans (Norell and Makovicky, 1999;
Rauhut, 2003). The scapular glenoid facet is transversely
widest at the level of the lateral lip posteriorly, where its
width is approximately 15 mm, but narrows to 12 mm an-
teriorly, before it slightly expands again towards the con-
tact with the coracoid. Thus, its lateral margin is markedly
concave in ventral view (Fig. 3.3). At the suture with the
coracoid, the glenoid is 14 mm wide. The scapular glenoid
facet is both anteroposteriorly and transversely slightly
concave. A shallow, oval depression is present on the
lateral side of the scapula above the glenoid, between the
lateral lip and the coracoid suture (Fig. 3.1). The coracoid
suture extends dorsally and slightly anteriorly from the
glenoid to approximately 11 mm above the lateral rim of the
latter. Here, it curves posteriorly, more so on the medial
side than on the lateral side, so that the suture is trans-
versely oblique in the broken dorsal cross-section, with
the coracoid extending further posteriorly medially than
laterally. Some 20 mm above the glenoid, the suture flexes
again dorsally and extends almost straight towards the
dorsal break on the medial side.
Although only the glenoid region of the coracoid is pre-
served, the structure of the preserved parts indicates that
it was relatively short anteroposteriorly, unlike the elon-
gate coracoids of dromaeosaurids (Norell and Makovicky,
1999) and other maniraptorans. The coracoid is transversely
widest at the glenoid, and narrows dorsally and anteriorly.
AMEGHINIANA - 2017 - Volume 54 (5): 539 – 566
544
RAUHUT AND POL: NEW JURASSIC THEROPOD FROM PATAGONIA
545
Towards the anteroventral end of the preserved part, the
bone widens again, slightly more so medially than laterally,
to form the basis of the broken ventral process of the cora-
coid. A weak oblique ridge on the lateral side of this expan-
sion probably corresponds to the coracoid tubercle (‘biceps
tubercle’). The biceps tubercle is developed as an oblique
ridge in other taxa, including Allosaurus (Madsen, 1976), but
it is more prominent and triangular in basal theropods, such
as Coelophysis rhodesiensis (Raath, 1977) and Zupaysaurus
(Ezcurra and Novas, 2007). Between the expansion and the
glenoid, the ventral margin of the coracoid forms a sharp
edge, from which the lateral side of the bone gradually ex-
tends dorsally (Fig. 3); thus, a longitudinal groove, as it is
present in some theropods, is absent in this region in Pan-
doravenator.
The glenoid facet of the coracoid is oval in outline and
has a raised and considerably thickened lateral rim. Al-
though the facet accounts for 10 mm of the complete gle-
noid along the longitudinal axis of the scapulocoracoid, its
length along its oblique long axis is approximately 14 mm. It
is oval in outline and up to 16 mm wide transversely. As in
the scapular glenoid facet, the facet of the coracoid is some-
what posteroventrolaterally directed and is slightly convex
transversely, so that its lateral parts are more strongly
laterally oriented than the medial parts.
Figure 3. Left scapulocoracoid of Pandoravenator fernandezorum, MPEF PV 1733-2; 1–2, outline drawing and photograph in lateral view; 3,
stereophotograph in ventral view. Abbreviations: co, coracoid; ct, coracoid tubercle; d, depression; gl, glenoid; ll, lateral lip; sc, scapula. Scale
bar= 10 mm.
Humerus
The distal end of the left humerus (MPEF PV 1773-1) is
broken right above the distal condyles (Fig. 4), so nothing
can be said about the shape of the humeral shaft. The
distal end is, as preserved, 38 mm wide transversely. The
medial, ulnar condyle is 15 mm deep anteroposteriorly,
whereas the lateral, radial condyle measures 19 mm. The
ulnar condyle is obliquely anteromedially-posterolaterally
oriented, 22 mm wide transversely, and slightly widens an-
teroposteriorly from its lateral to its medial end (Fig. 4.2).
The anteromedial expansion is mainly due to a stout, tuber-
cle-like medial entepicondyle, which is largely eroded. The
posterior side of the distal humerus is slightly concave be-
tween the entepicondyle and the lateral end of the ulnar
condyle, which projects posteriorly. The distal articular sur-
face of the ulnar condyle is slightly convex anteroposteri-
orly and more notably so transversely, so that the medial
part extends slightly more proximally than the lateral part.
The ulnar condyle is separated from the radial condyle
by a deep, broad, V-shaped incision posteriorly and a broad
concavity distally (Fig. 4). The radial condyle has its long axis
oriented anteroposteriorly, and is 19 mm wide in this direc-
tion and 14 mm transversely. It has a triangular outline in
distal view, with a rounded anterior point (Fig. 4.2). Its dis-
tal articular surface is placed slightly more proximally than
the lateral part of the ulnar condyle. It is almost flat ante-
riorly, only very slightly convex transversely, but extends in
a convex arch onto the posterior side of the distal humerus
posteriorly. The medial part of the articular surface on the
posterior side of the humerus extends slightly more proxi-
mally medially than laterally, and thus forms the lateral bor-
der of the trough-like incision between the two condyles.
The anterolateral side of the condyle is flat. On the anterior
side of the distal humerus, there is a large depression
laterally above the radial condyle and the lateralmost part
of the ulnar condyle (Fig. 4.1). This depression gradually
deepens from a broad ridge extending from the anterior
point of the radial condyle proximally towards the medial
side, and is bordered by a marked, slightly concave rim me-
dially above the lateral part of the ulnar condyle. The exact
extent of this depression cannot be established due to the
proximal incompleteness of the bone. The lateral side of the
distal end is notably flat, in contrast to the usually antero-
posteriorly convex surface in most other theropods.
Femur
Only the distal end of the right femur is preserved (MPEF
PV 1773-3; Fig. 5). A crystal infilled cavity is visible at the
proximal break, indicating that the bone was hollow, as in
most theropods. The shaft of the femur was oval to sub-
triangular in outline, but seems to be slightly deformed. As
preserved, the distal end is 56 mm wide transversely, but a
few mm might be missing at the anterolateral side of the
eroded lateral condyle. A shallow, broad depression is pres-
ent on the medial part of the anterior side of the distal part
of the femur, as in many other theropods (Rauhut, 2003),
but only its distalmost part is preserved. It is bound medially
by a rounded, longitudinal medial edge on the anterior part
of the medial side of the femur. This edge widens distally
towards the distal condyle into a flat, triangular medial
surface. However, in contrast to other theropods, such as
Allosaurus (Madsen, 1976), Megalosaurus (Benson, 2010) and
AMEGHINIANA - 2017 - Volume 54 (5): 539 – 566
546
Figure 4. Distal end of left humerus of Pandoravenator fernandezo-
rum, MPEF PV 1733-1; 1, photograph in anterior view; 2, photograph
in distal view. Abbreviations: ent, entepicondyle; rc, radial condyle;
uc, ulnar condyle. Scale bar= 10 mm.
Condorraptor (Rauhut, 2005b), this expansion is restricted
to the distalmost part of the femur, and thus the resulting
triangular area is wider anteroposteriorly than high proxi-
modistally. The lateral part of the anterior side is convex
transversely. The extensor groove on the anterior side of
the femur was obviously very shallow or even absent,
similar to the situation in e.g. Dilophosaurus (Welles, 1984),
Cryolophosaurus (Smith et al., 2007), and Eustreptospondylus
(Sadleir et al., 2008), but in contrast to many basal tetanu-
rans, which usually have a well-developed extensor groove
(e.g., Madsen, 1976; Currie and Zhao, 1993; Rauhut, 2005b).
The distal condyles are strongly convexly rounded an-
teroposteriorly (Fig. 5). Only a slight concavity marks the
distinction between the tibial and fibular condyles, but a
small, moderately deep depression, which does not seem to
be an artifact of preservation, is present in the central part
of the articular end. The medial side of the distal end ex-
tends slightly further distally than the lateral side and both
the tibial condyle and the crista tibiofibularis are inclined
laterally (Fig. 5), though this might be exaggerated by de-
formation. The posterior expansion of the tibial condyle is
subequal in size to the crista tibiofibularis, from which it is
separated by a wide posterior intercondylar groove (Fig.
5.5). No infrapopliteal ridge is present, in contrast to coelo-
physoids and other basal theropods (Rowe, 1989; Tykoski,
2005). As in most theropods, the crista tibiofibularis is
slightly offset from the distal end and from the lateral
margin of the femur. A broad longitudinal depression is
present between the fibular condyle and the crista laterally.
There are two longitudinal ridges on the lateral side that
extend from the distal condyle proximally. They are sepa-
rated by ca. 7mm, and whereas the more posterior ridge
ends at about the level of the proximal end of the crista
tibiofibularis, the more anteriorly placed ridge extends
some 10 mm further proximally.
RAUHUT AND POL: NEW JURASSIC THEROPOD FROM PATAGONIA
547
Figure 5. Distal end of right femur and proximal end of tibia of Pandoravenator fernandezorum, MPEF PV 1733-3; 1–2, outline drawing and
stereophotographs in posterior view; 3–4, outline drawing and photograph in medial view; 5, photograph in distal view. Abbreviations: cc, cnemial
crest; d, depression; fc, fibular condyle; fe, femur; tc, tibial condyle; ti, tibia. Scale bar= 10 mm.
Tibia
The tibia is represented by a fragment of the proximal
part of the left shaft, a section of 96 mm of the proximal
part of the right mid-shaft (MPEF PV 1773-5; Fig. 6), and
the distal end of the left (MPEF PV 1773-36a) and right
(MPEF PV 1773-4; Fig. 7) elements. Furthermore, a frag-
ment of bone that is attached to the posteromedial side of
the distal part of the femur probably represents the proxi-
mal end of the right tibia (Fig. 5.1–2). A part of the mid-shaft
of the right fibula is preserved in articulation with the tibial
shaft, and the distal end of the right tibia is articulated with
the distal part of the fibula, astragalus and calcaneum.
As in the femur, the broken ends show that the tibia was
hollow. If the interpretation of the fragment preserved with
the femur as the proximal end of the right tibia is correct,
the cnemial crest was rather small, anteriorly and very
slightly proximally directed, and has a rounded anterior out-
line (Fig. 5.1, 4), similar to the condition in Condorraptor
(Rauhut, 2005b), but unlike the usually much larger and
more rectangular crest in ceratosaurs (e.g., Gilmore, 1920;
Madsen and Welles, 2000; Carrano et al., 2002; Carrano,
2007; Rauhut, 2011; Pol and Rauhut, 2012; Rauhut and
Carrano, 2016). In proximal view, there is no deeply-notched
incisura tibialis separating the cnemial crest from the fibular
condyle, but the former seems to extend rather gradually
towards the latter structure, as in most non-tetanuran
theropods (Rauhut, 2003). The posterior part of the proxi-
mal end is damaged, but a shallow, broad depression in the
preserved part posteriorly, between the medial side of the
proximal part of the tibia and the fibular condyle, indicates
that a posterior incision between these two structures was
present, as in most non-avian theropods (Rauhut, 2003).
Only a fragment of the proximal part of the shaft,
showing the basis of the cnemial crest, is present for the
left tibia. This fragment confirms that the cnemial crest
was stout and its lateral side is only slightly concave to-
wards the fibular condyle. There is no indication of any
lateral ridge or crest connecting the proximal end with the
fibular crest, as it is present in non-tetanuran theropods
(Rauhut, 2003). The fibular crest is preserved in the section
of the mid-shaft of the right tibia (Fig. 6). It is a stout lateral
crest that distally gradually lowers into the shaft. Below the
crest, the tibial shaft is broader transversely (33 mm) than
anteroposteriorly deep (ca. 28 mm) and has a triangular
AMEGHINIANA - 2017 - Volume 54 (5): 539 – 566
548
Figure 6. Proximal shaft of right tibia and fibula of Pandoravenator fernandezorum, MPEF PV 1733-5; 1–2, outline drawing and stereopho-
tographs in anterior view. Abbreviations: fc, fibular crest; fi, fibula. Scale bar= 10 mm.
outline, with a flattened anterior side and a bulging pos-
terior side. This posterior bulge becomes more pronounced
proximally towards the proximal break. On the anterior side,
a longitudinal depression seems to be present and becomes
more conspicuous towards the distal break of the shaft,
although it cannot be ruled out that this is due to deforma-
tion.
The distal end of the right tibia is generally well pre-
served (Fig. 7), but several details cannot be established
because they are hidden by the articulated distal fibula
and proximal tarsals. The distal end of the tibia is expanded
transversely to a width of 67 mm and measures approxi-
mately 30 mm in anteroposterior direction, although the
width might be slightly exaggerated by anteroposterior
compression. Although covered by the proximal tarsals, the
distal outline of the tibia is obviously broad and triangular,
unlike the more rectangular outline in non-averostran
theropods, but as in most non-maniraptoran averostrans
(Rauhut, 2003). Whereas the medial malleolus is not notably
expanded from the distal shaft, the lateral malleolus ex-
pands laterally into a rounded, lobe-like expansion that is
directed laterally and slightly posteriorly and backs the
distal end of the fibula posteriorly (Fig. 7.3–5). The mor-
phology of the lateral malleolous is unlike the indistinct or
more rectangular distal expansion seen in non-averostran
theropods (e.g., Ezcurra and Novas, 2007; Nesbitt and
Ezcurra, 2015), or the gradual expansion in Cryolophosaurus
(FMNH PR 1821; Smith et al., 2007), Elaphrosaurus (Rauhut
and Carrano, 2016), and Ceratosaurus (Madsen and Welles,
2000), but similar to the situation in Piatnitzkysaurus (Bo-
naparte, 1986), Allosaurus (Madsen, 1976), and Tugulusaurus
(Rauhut and Xu, 2005), although the latter two taxa have a
more pronounced medial expansion of the distal part of the
tibia. In posterior view, the line of suture between the tibia
and the astragalus is slightly sigmoidal, extending further
proximally just distal to the anteroposteriorly broadest part
of the distal part of the tibia, which is situated at approxi-
mately one third of the width of the bone from the medial
side (Fig. 7.5). The lateral malleolus extends only very
slightly further distally than the medial malleolus, unlike the
situation in several basal tetanurans, including Torvosaurus
(Britt, 1991) and carcharodontosaurids (Brusatte and Se-
reno, 2008), in which the latter extends notably further
distally. On the anterior side of the distal part of the tibia, a
pronounced step extends from the medial side laterally
and slightly proximally and forms the proximal border of the
attachment area for the astragalus (Fig. 7.1–2), as in most
non-coelurosaurian basal averostrans (e.g., Madsen, 1976;
Britt, 1991; Currie and Zhao, 1993; Carrano, 2007), with the
exception of noasaurids (Rauhut and Carrano, 2016). At
about half the width of the bone, the step flexes proximally
and extends almost vertically, only very slightly laterodor-
sally inclined, towards the proximal break. Here, the step
has an anteroposterior depth of approximately 8 mm. The
fragment of the distal end of the left tibia is strongly de-
formed and highly incomplete, providing no additional in-
formation.
Fibula
The proximal end of the right fibula (MPEF-PV 1773-6)
lacks the posterior border of the bone and the shaft is
broken some 50 mm below the proximal end. As in most
theropods, the proximal end is kidney-shaped, but the
transverse width remains subequal throughout the pre-
served part and does not get notably narrower posteriorly,
as it is the case in most taxa. In lateral view, the proximal
articular surface is slightly concave anteroposteriorly, with
this concavity being mainly due to the presence of a small,
rounded proximal expansion at the anterior end. With an es-
timated anteroposterior length of approximately 40 mm
and a transverse width of 16 mm, the proximal end is rather
robust. The lateral side of the proximal fibula is evenly con-
vex anteroposteriorly, and the medial side gently concave.
However, a deep medial groove is absent. Such a groove is
present as a rather narrow, more posteriorly placed de-
pression in most non-tetanuran theropods (e.g., Rowe,
1989; Madsen and Welles, 2000; Carrano, 2007), and as a
wide, deeply excavated groove with sharp margins in
Elaphrosaurus (Rauhut and Carrano, 2016) and many teta-
nurans (Rauhut, 2003), with the notable exception of mega-
losauroids (Benson, 2010). Fine longitudinal striations are
present on the surface of the medial side. Towards the dis-
tal break, the bone is deformed, so that the broken attach-
ment of the shaft faces anterodistolaterally instead of
distally.
A part of the mid-shaft of the right fibula is preserved in
articulation with the tibia (MPEF PV 1773-5; Fig. 6). This in-
cludes the section that articulates with the fibular crest of
RAUHUT AND POL: NEW JURASSIC THEROPOD FROM PATAGONIA
549
the tibia and slightly more distal parts of a total length of
56 mm. The proximal part of the preserved section is flat-
tened transversely and slightly expanded anteroposteriorly
to a width of 16 mm. There is no indication of the iliofibu-
laris tubercle, which is found in this area in other theropods
(e.g., Currie and Zhao, 1993; Madsen and Welles, 2000;
Brochu, 2003), and the posterior margin of the bone is more
markedly thickened (ca. 9 mm) than the anterior edge (ca. 6
AMEGHINIANA - 2017 - Volume 54 (5): 539 – 566
550
Figure 7. Distal end of right tibia, fibula, astragalus, and calcaneum of Pandoravenator fernandezorum, MPEF PV 1733-4; 1–2, outline drawing
and stereophotographs in anterior view; 3–4, outline drawing and stereophotographs in lateral view; 5, photograph in posterior view; 6, pho-
tograph in medial view; 7, photograph in distal view. Abbreviations: asc, ascending process of astragalus; ast, astragalus; at, anterior tubercle;
cal, calcaneum; dt, distal tubercle; fi, fibula; gr, groove; no, notch; ti, tibia. Scale bars= 10 mm.
mm). Below the insertion area of the m. iliofibularis, the
fibula narrows slightly towards the distal shaft, and has
an oval outline at the distal break, which is 10 mm wide
anteroposteriorly and 8 mm transversely at the posterior
margin. As in the more proximal section, the anterior mar-
gin is slightly thinner. The medial side of the fibula is convex
anteroposteriorly and thus lacks a longitudinal depression,
as it is found in some theropods (e.g., Madsen, 1976).
The distalmost 42 mm of the right fibula is preserved in
articulation with the tibia, astragalus and calcaneum (MPEF
PV 1773-4; Fig. 7). At the proximal break, the bone is obliquely
semioval with a very slightly transversely concave pos-
teromedial side. It is approximately 11 mm wide anteropos-
teriorly and measures 7 mm in anterolateral-posteromedial
direction. Distally, the bone rapidly expands, and the distal
end has a maximal anteroposterior width of 26 mm.
Whereas the posterior part of this expansion is triangular in
lateral view (Fig. 7.3–4) and begins some 17 mm above the
distal end, the anterior part is developed as an extensive
bony lamina that gradually extends from approximately 35
mm above the distal end anterodistally (Fig. 7.1–2). This
lamina is flexed medially in its anteriormost part and con-
tacts the lateral side of the ascending process of the as-
tragalus, as is the case in Coelophysis rhodesiensis (Rauhut,
2003: fig. 55A) and Camposaurus (Ezcurra and Brusatte,
2011). This morphology is notably different from that seen
in most theropods, including non-averostrans, such as
Liliensternus (Rauhut, 2003: fig. 55B), Zupaysaurus (Ezcurra
and Novas, 2007), Dilophosaurus (Welles, 1984), ceratosaurs
(e.g., Madsen and Welles, 2000), and tetanurans (e.g., Al-
losaurus: Madsen, 1976; Sinraptor: Currie and Zhao, 1993;
Eustreptospondylus: Sadleir et al., 2008; Tyrannosaurus:
Brochu, 2003), in which the expansion is less pronounced
and the distal end is more robust. Although the suture be-
tween the fibula and the ascending process of the astra-
galus is tight, they are not fused, unlike the condition in
adult abelisauroids (Carrano and Sampson, 2008) and at
least some non-averostran theropods (e.g., Nesbitt and
Ezcurra, 2015). The distal articular surface of the fibula is
strongly convex anteroposteriorly, facing distally at is pos-
terior end and anterodistally at its anterior end (Fig. 7.3–4).
The posterior margin of the fibula forms a small, posteriorly
directed, triangular tubercle just above the distal articular
surface.
Astragalus
The right astragalus and calcaneum are preserved in ar-
ticulation with the tibia and fibula (MPEF PV 1773-4), and
the medial half of the left astragalus (MPEF PV 1773-36b)
has been preserved in articulation with a fragment of the
distal left tibia. The astragalus is one of the most distinc-
tive elements of Pandoravenator. It is a stout bone with the
lateral side being considerably narrower anteroposteriorly
than the medial side (Fig. 7.7), unlike the situation in Li-
liensternus (Rauhut, 2003: fig. 49B) and Elaphrosaurus
(Rauhut and Carrano, 2016), but as in Zupaysaurus (Ezcurra
and Novas, 2007) and tetanurans (e.g., Eustreptospondylus:
Sadleir et al., 2008; Wiehenvenator: Rauhut et al., 2016;
Allosaurus: Gilmore, 1920; Tyrannosaurus: Brochu, 2003).
The complete right element is parallelogram-shaped in dis-
tal outline, but the medial side of the left element shows
that this might be partially due to anteroposterior com-
pression, and the original shape might have been rather
trapezoidal, narrowing laterally. The right astragalus
measures approximately 47 mm transversely, is 33 mm
wide medially, but only 28 mm wide laterally; whereas the
transverse width might be slightly exaggerated, the antero-
posterior measurements should be seen as minimal esti-
mates, due to compression. The maximal proximodistal
height of the right astragalus (excluding the ascending
process) is 27 mm. In medial view, the distal astragalar body
is kidney-shaped, narrowing posteriorly (Fig. 7.6). The me-
dial side is almost flat proximodistally, but gently convex an-
teroposteriorly. A high, triangular (in medial view) part of
bone is present proximal to the anteroproximal corner of the
main astragalar body and is separated from the latter by a
deeply incised notch in the right element (Fig. 7.6). This
separation is only indicated by a shallow groove in the left
astragalus. In the right astragalus, the anteroproximal ex-
pansion extends proximally for ca. 11 mm (13 mm in the left
element) and articulates proximally with the step on the
anterior side of the tibia, as in other non-coelurosaurian
theropods. In anterior view of the right astragalus, the notch
between the main astragalar body and the anteroproximal
expansion extends laterally in a distally slightly convex arch
as a well-developed, narrow, but deep and suture-like
groove (Fig. 7.1–2). This groove ends some 10 mm medial to
the contact with the calcaneum, where it turns sharply
proximally and becomes a thin, slightly laterally inclined line
RAUHUT AND POL: NEW JURASSIC THEROPOD FROM PATAGONIA
551
that extends towards the lateral margin of the ascending
process of the astragalus. Both the groove and this line are
reminiscent of a suture, but the clearly defined groove is ab-
sent in the left element, although its course can be followed
as a fine line on the anterior side of the bone towards its
lateral break. In the right astragalus, the proximal margin of
this anteroproximally part of the bone follows the course
of the distal border of the step on the anterior side of the
tibia until the latter flexes abruptly proximally (as described
above). From there, the margin expands into the ascending
process of the astragalus. It first flexes steeply proximo-
laterally for approximately 6 mm and then extends much
less steeply proximolaterally over 13 mm until it contacts
the anteromedial margin of the fibula (Fig. 7.1–2). The
lateral margin of the ascending process of the astragalus
is inclined proximolaterally, so that the whole process has a
slight lateral inclination, as in Liliensternus (Rauhut, 2003:
fig. 49A), Zupaysaurus (Ezcurra and Novas, 2007), and Cera-
tosaurus (Madsen and Welles, 2000), but unlike the more
vertical process in most tetanurans (e.g., Allosaurus: Madsen,
1976). In total, the ascending process is very low (maximally
10 mm or 37% of the maximal height of the astragalar body,
although 1–2 mm might be missing proximally due to ero-
sion), laminar, but rather stout anteroposteriorly, and its
base extends over only 13 mm, or approximately 28%, of the
breadth of the astragalar body, being placed entirely on
the lateral half of the latter. Apart from being laminar, the
ascending process of Pandoravenator is comparable in
height, width and placement, with that of many non-ave-
rostran theropods, such as Liliensternus (Rauhut, 2003) and
Zupaysaurus (Ezcurra and Novas, 2007). Among cerato-
saurs, which have a laminar ascending process, the posi-
tion and placement of the process are comparable at least
with those of Ceratosaurus (Madsen and Welles, 2000),
Elaphrosaurus (Rauhut and Carrano, 2016) and abelisaurids
(Carrano, 2007), although in the latter it seems to be higher.
The distal articular surface of the astragalus is strongly
convex anteroposteriorly and is concavely arched in ante-
rior view (Fig. 7.1–2). In posterior view, the tibia-astragalus
contact extends further distally on the lateral than on the me-
dial side. The astragalus is closely appressed, but not fused
to the calcaneum. There are two large, unusual tubercles at
the suture between the two elements. The larger of these is
placed laterally at the anteroproximal end of the astragalar
body, lateral to the base of the ascending process (Fig. 7.1–2).
It is ca. 14 mm high proximodistally and semioval in outline.
The second tubercle is placed laterally on the distal surface
of the astragalus (Figs. 7.1–2, 8.7). It is triangular in outline,
ca. 11 mm in length anteroposteriorly, and delimited ante-
riorly by a shallow, oval to rectangular depression on the
distal surface, which is 8 mm long anteroposteriorly and 6 mm
wide transversely. Unlike the condition in Allosaurus (Welles
and Long, 1974; Madsen, 1976) and other tetanurans (e.g.,
Britt, 1991; Currie and Zhao, 1993), there is no lateral notch
in the anterolateral side of the distal surface of the astra-
galus for a medial process of the calcaneum (Fig. 7.1–2).
Calcaneum
The calcaneum (MPEF PV 1773-4) is a high, but thin disc
of bone. Its maximal transverse width is approximately 13
mm, or 28% of the width of the astragalus, very similar to
the ratio seen in Elaphrosaurus (Rauhut and Carrano, 2016)
and Sinraptor (Currie and Zhao, 1993), whereas the width of
the calcaneum is less than 20% of the width of the astra-
galus in coelurosaurs (e.g., Kobayashi and Barsbold, 2005;
Rauhut and Xu, 2005; Brusatte et al., 2012). The anterodis-
tal and posteromedial parts of the bone are damaged, so
that the exact proximodistal height cannot be established,
but it was about 31–33 mm. In lateral view, the proximal
articular facet for the fibula is horizontal in its posterior
half, but flexes steeply anteroproximally in its anterior half
(Fig. 7.3–4). Thus, the anterior end of the calcaneum ex-
tends approximately 11 mm further proximally than the
posterior end. On the lateral side, the fibular articular sur-
face extends all the way to the posterior end of the astra-
galus, but a small articular facet for the tibia is present
posteromedially of this surface on the calcaneum (Fig. 7.3–
4). The lateral side of the calcaneal body is very slightly con-
cave both anteroposteriorly (Fig. 7.7) and proximodistally,
with a large, inverted teardrop-shaped depression present
close to the anterior margin (Fig. 7.3–4), as in Allosaurus
(Madsen, 1976), Elaphrosaurus (Rauhut and Carrano, 2016),
and an astragalocalcaneum referred to Aerosteon (MCNA-
PV-3139). This depression is 8 mm high proximodistally
and maximally 5 mm wide anteroposteriorly. The distal ar-
ticular surface of the calcaneum is partly damaged, but
based on the preserved portion it seems to be strongly
convex anteroposteriorly and slightly convex transversely.
AMEGHINIANA - 2017 - Volume 54 (5): 539 – 566
552
Distal tarsals
The distal tarsals III and IV are preserved in articulation
with the metatarsus of the left side (MPEF PV 1773-9), but
the lateral half of the right distal tarsal IV (MPEF PV 1773-
7) was found in isolation, together with the proximal end of
the right metatarsal IV. Distal tarsal III is the smaller of the
two distal tarsals, as in Allosaurus (Madsen, 1976), but un-
like the situation in Coelophysis bauri (Padian, 1986), Coelo-
physis rhodesiensis (Raath, 1977), Dilophosaurus (Welles,
1984), tyrannosaurs (Lambe, 1917; Brusatte et al., 2012),
and Deinonychus (Ostrom, 1969), in which these bones are
of similar size or distal tarsal III is larger than distal tarsal
IV (e.g., Padian, 1986); Sinraptor is somewhat intermediate
in that the distal tarsal III is slightly smaller than distal tarsal
IV (Currie and Zhao, 1993). It is rectangular to trapezoidal in
outline, with rounded edges and caps the central part of
metatarsal III and a small part of the lateral margin of
metatarsal II (Fig. 8.5–6). There is no posterolateral expan-
sion, as it is present in Coelophysis (Padian, 1986) and Sin-
raptor (Currie and Zhao, 1993). The bone is 21 mm long
anteroposteriorly and maximally 25 mm wide transversely.
It is proximodistally thin anteriorly, but thickens posteriorly,
with the thickest part (ca. 5 mm) being the posteromedial
corner (Fig. 8.3–4). The proximal surface is slightly trans-
versely convex anteriorly, and notably convex anteropos-
teriorly posteriorly.
Distal tarsal IV is larger than distal tarsal III, also ap-
proximately trapezoidal in outline, and caps the proximal
end of metatarsal IV (Fig. 8.5–6). In anterior view of the ar-
ticulated left tarsus and metatarsus, the proximal end of
metatarsal IV is placed slightly distal to the proximal ends of
metatarsals II and III, so that the proximal surface of distal
tarsal IV is level with the proximal surface of metatarsal III
(Fig. 8.1–2). The long axis of distal tarsal IV is oriented an-
terolaterally-posteromedially, and it is 23 mm wide ante-
riorly, 29 mm long along its long axis, and approximately 14
mm wide posteriorly. The anteromedial corner of the bone
is pointed, whereas the anterolateral edge is rounded and
slightly raised, similar to the condition in Allosaurus (Mad-
sen, 1976) and Sinraptor (Currie and Zhao, 1993). The ante-
rior side between these two edges forms an almost straight
line. The medial side of the bone marks a concave arch that
contacts the lateral side of the anterior half of metatarsal
III. The lateral margin of the distal tarsal is very slightly
concave posterior to the rounded anterolateral edge, much
less so than in many other basal theropods (Tykoski, 2005),
but comparable to the condition of Sinraptor (Currie and
Zhao, 1993). Posteriorly, the lateral corner of the bone is
more broadly rounded than the medial edge, and the pos-
terior margin between them is straight.
The proximodistal thickness of the bone varies con-
siderably. The anteromedial part of the bone bulges distally
to articulate with a corresponding depression in the proxi-
mal articular surface of metatarsal IV, and forms a high,
triangular surface adjacent to metatarsal III in anterior view
(Fig. 8.1–2). This anteromedial expansion reaches a thick-
ness of 10 mm along the contact with metatarsal III in the
left tarsus. The bone rapidly thins laterally to a minimal
thickness of approximately 3 mm. The lateral edge is raised
proximally to a thickness of 5.5 mm, so that the proximal
articular surface is notably concave in its lateral part. The
posterolateral part of the distal tarsal IV bulges distally to
form a semioval expansion posteriorly at the contact be-
tween metatarsal IV and III (Fig. 8.3–4). This bulge reaches
a maximal thickness of 9 mm. The disarticulated partial right
distal tarsal IV demonstrates that the distal articular sur-
face is saddle-shaped between the anterior and posterior
distal expansions.
Metatarsus
The metatarsus is represented by the proximal ends of
the left metatarsals II to V (MPEF PV 1773-9), which are
preserved in articulation, the isolated proximal end of the
right metatarsal IV (MPEF PV 1773-8), the articulated
distal ends of right metatarsals II to IV (MPEF PV 1773-10),
and the isolated distal ends of left metatarsals II (MPEF PV
1773-11) and III (MPEF PV 1773-12). Unfortunately, no
complete metatarsal shaft is preserved, so nothing can be
said about the length of the metatarsus. However, many of
the broken shaft fragments clearly show that, as is the case
with the other long bones, the metatarsals were hollow in-
teriorly (Fig. 8.8). In the following description, the distal ends
of metatarsals II and III are mainly described based on the
isolated ends of the left elements; the right metatarsals
confirm the observations based on these elements, al-
though the shaft of especially the right metatarsal III is
strongly deformed.
Metatarsal II has a large, semioval to subtriangular
RAUHUT AND POL: NEW JURASSIC THEROPOD FROM PATAGONIA
553
proximal articular surface (Fig. 8.5). The long axis of the ar-
ticular end is oriented anterolaterally-posteromedially.
Thus, the proximal end is 35 mm long along its long axis,
and measures maximally 22 mm perpendicular to it. The
lateral margin seems to be very slightly convex anteropos-
teriorly, but is largely covered by the medial margin of dis-
tal tarsal III. The anterior and anteromedial margin of the
bone form a gradual curve, whereas the posteromedial
AMEGHINIANA - 2017 - Volume 54 (5): 539 – 566
554
Figure 8. Articulated proximal left metatarsus and distal tarsals of Pandoravenator fernandezorum, MPEF PV 1733-9; 1–2, outline drawing and
stereophotographs in anterior view; 3–4, outline drawing and stereophotographs in posterior view; 5–6, outline drawing and stereopho-
tographs in proximal view; 7–8, outline drawing and photograph in distal view showing the cross-section of metatarsals II-III-IV. Abbreviations:
dt, distal tarsal; mt, metatarsal. Scale bar= 10 mm.
margin has a concave indentation (Fig. 8.5–6), as in Cerato-
saurus (Gilmore, 1920), Majungasaurus (Carrano, 2007),
megalosauroids (e.g., Britt, 1991; Sadleir et al., 2008), allo-
sauroids (e.g., Madsen, 1976; Currie and Zhao, 1993), and
tyrannosaurs (Brochu, 2003), but unlike the convex margin in
Liliensternus (Huene, 1934), Coelophysis rhodesiensis (Raath,
1977) and Elaphrosaurus (Rauhut and Carrano, 2016). This
indentation is due to a small, triangular posterior flange that
extends from the proximal shaft of metatarsal II posteriorly
in lateral view. Apart from this posterior flange, the proximal
end is more strongly expanded anteriorly than posteriorly.
The shaft rapidly narrows, and is 16 mm wide anteroposte-
riorly and 13 mm mediolaterally at the distal break (45 mm
below the proximal end). Here, the bone is anteromedially-
posterolaterally flattened and has a semioval cross-section.
The distal end of metatarsal II is stout and has a well-de-
veloped distal articular surface, which forms a round con-
vex arch that extends over approximately 180° from the
dorsal surface to the ventral surface of the bone (Fig. 9.8).
There is no extensor groove on the dorsal surface of the
bone, only the area adjacent to the distal articular surface
shows a very weakly concave depression (Fig. 9.1–2, 5).
A well-developed and deep collateral ligament fossa is
present on the lateral side of the bone, separated by some
11 mm from the extremity of the distal end (Fig. 9.8). It is
oval in outline and 10 mm long proximodistally and 6 mm
high anteroposteriorly in the left metatarsal. A small,
sharply defined, triangular lateral tubercle is present on the
ventral margin of the lateral side, just below the anterior
end of the collateral ligament fossa (Fig. 9.7). On the medial
side, the collateral ligament fossa is only indicated by a
large, but shallow depression adjacent to the distal articu-
lar surface (Fig. 9.6). The lateral side of the distal articular
surface extends further distally than the medial side, so
that the surface is inclined mediodistally at an angle of
approximately 60° from the long axis of the shaft (Fig. 9.1–
2, 5). The distal articular surface is transversely convex in
its dorsal part, but subdivided into two distinct condyles in
the ventral part (Fig. 9.7). These condyles diverge at an
angle of approximately 65°, and the medial condyle is only
about half the width of the lateral and tapers ventrally,
whereas the lateral condyle is rounded. The lateral condyle
is strongly transversely convex. At the proximal break, some
35–40 mm above the distal end, the shaft is oval in outline,
with the long axis of the oval oriented slightly obliquely
from anterolateral to posteromedial, and is approximately
16 mm wide transversely and 15 mm wide anteroposte-
riorly.
The proximal end of metatarsal III is hourglass-shaped
in proximal view, as it is usual in basal tetanurans (e.g., Mad-
sen, 1976; Britt, 1991; Currie and Zhao, 1993), but unlike
the elongate condition in Dilophosaurus (Welles, 1984), the
T-shaped outline in Ceratosaurus (Gilmore, 1920) and
Elaphrosaurus (Rauhut and Carrano, 2016), or the more
blocky appearance in Majungasaurus (Carrano, 2007). The
posterior end of the articular surface is displaced medially
in comparison to the anterior end, and there seems to be a
constriction between these ends, although most of this
area is covered by distal tarsal III (Fig. 8.5–6). Furthermore,
whereas the anterior margin faces mainly anteriorly, the
posterior margin faces slightly posterolaterally. The articu-
lar surface is 17 mm wide anteriorly, 16 mm wide posteri-
orly, and 41 mm long anteroposteriorly. The anterior end
bulges slightly proximally and the shaft rapidly narrows dis-
tally (Fig. 8.1–2). Whereas the anterior end fades into the
shaft gradually, there is a large, rectangular tubercle poste-
riorly. This tubercle is 16 mm wide transversely, 14 mm high
proximodistally, has a flat posterior surface, and is offset
from the shaft by a distinct, concave step, similar to the
situation found in Allosaurus (Madsen, 1976). At the distal
break, some 42 mm below the proximal end, the posterior
part of the shaft of metatarsal III is strongly pinched be-
tween the shafts of metatarsals II and IV, which almost
meet posterior to metatarsal III (Fig. 8.7–8). It has the same
inclination of the long axis as the proximal end and seems to
be slightly constricted between the anterior and posterior
sides. It is approximately 16 mm long anteroposteriorly, ca.
15 mm wide anteriorly and 13 mm wide posteriorly. The
distal end of metatarsal III expands rapidly from a slender
shaft. In the isolated left metatarsal (MPEF PV 1733-12),
the shaft is compressed with a laterodorsally-ventromedially
oriented long axis of the outline at the proximal break, some
60 mm proximal from the distal end. It is ca. 18 mm wide
transversely and measures ca. 17 mm anteroposteriorly.
The shaft is somewhat compressed posterolaterally-an-
teromedially, giving it an 8-shaped cross-section at the
break, although this might be exaggerated by deformation.
Just distal to the break, there is a small medial flange for the
RAUHUT AND POL: NEW JURASSIC THEROPOD FROM PATAGONIA
555
contact with the distal metatarsal II. Thus, in articulation,
the distal articular end of metatarsal II is placed entirely
proximal to the expanded articular end of metatarsal III
(Fig. 9.1–2, 5). The distal end of metatarsal III expands to a
transverse width of 27 mm and an anteroposterior height
of 25 mm medially and 21 mm laterally (Fig. 9.10). Collateral
ligament pits are found on both the medial and lateral side
of the distal end, but whereas the ligament pit on the medial
side is developed as a moderately deep oval depression
with a medially raised dorsal margin (Fig. 9.9), that of the
lateral side is deeper and better defined, but also oval in
outline (Fig. 9.11). The lateral collateral pit is approximately
AMEGHINIANA - 2017 - Volume 54 (5): 539 – 566
556
Figure 9. Distal end of metatarsals and proximal ends of pedal phalanges of Pandoravenator fernandezorum. 1–4, Outline drawing and
stereophotographs of distal ends of right metatarsals II-III-IV, pedal ungual I, and phalanges II-1 and IV-1 (MPEF PV 1733-10) in 1–2, ante-
rior and 3–4, posterior views; 5, photograph in anterior view of left metatarsals II (MPEF PV 1733-11) and III (MPEF PV 1733-12); 6–8, pho-
tographs of left metatarsal II (MPEF PV 1733-11) in 6, medial, 7, distal, and 8, lateral views; 9–11, photographs of left metatarsal III (MPEF
PV 1733-12) in 9, medial, 10, distal, and 11, lateral views. Abbreviations: cg, collateral ligament groove; eg, extensor groove; mt, metatarsal;
ri, ridge; I-2, II-1, IV-1, pedal phalanges. Scale bars= 10 mm.
8 mm long proximodistally and maximally 7 mm high an-
teroposteriorly. On the anterior side of the distal end there
is a small, rounded extensor groove, with a better-defined
lateral than medial wall (Fig. 9.2). The groove is approxi-
mately 8 mm long proximodistally and 6 mm wide trans-
versely, and is placed more on the lateral than on the
medial side of the metatarsus. Just proximal to the thick-
ened lateral edge of this groove, the bone narrows on the
lateral side, forming a flat lateral surface for the contact
with metatarsal IV. Thus, all three central metatarsals
were obviously tightly appressed over their entire length.
The distal articular surface of metatarsal III extends over
an arc of approximately 180° from the anterior surface of
the distal end to its posterior surface (Fig. 9.9, 11). It is
broad, anteroposteriorly strongly convex, and subdivided
into a slightly wider lateral and a narrower medial part by a
very shallow transverse concavity (Fig. 9.10). Posteriorly,
there are two small, widely diverging, flat condyles, sepa-
rated by a very wide, V-shaped division.
The proximal end of metatarsal IV consists of a sub-
round to subtriangular central articular surface and a long,
posteromedially curved medial process at the posterior
end, as it is usually found in tetanuran theropods. However,
this process is broken off in the isolated proximal end of
the right metatarsal IV and largely hidden by the articulated
distal tarsal IV and metatarsal V in the left element, and
not much can be said about its morphology. Only the distal-
most tip can be seen in articulation with the posterolateral
corner of metatarsal III. This tip shows that the distal ex-
treme of the process tapers to a point. The proximal end of
the left metatarsal IV is 21 mm long anteroposteriorly, 17
mm wide transversely anteriorly, and 33 mm wide poste-
riorly, including the posteromedial process. The lateral
part of the proximal articular end is almost round in the left
metatarsal (Fig. 8.1–4), but more triangular in the right
element. The medial and posterior sides of this part form
almost straight edges in both elements. The disarticulated
right metatarsal IV shows that the proximal articular sur-
face has a large depression anteriorly for the articulation
with the distal tarsal IV, with slightly raised anterior and
lateral edges. The shaft of metatarsal IV is only slightly
constricted distally, with the main proximal expansions
being anteriorly and, to a lesser extent, laterally. The poste-
rior and medial sides of the shaft remain flat over the en-
tire preserved length, and the posterior side is offset from
the lateral side by a pronounced rounded edge. The anterior
and lateral sides are also mainly flat, but curve into one
another anterolaterally in a broad curve. Thus, at the distal
break, the shaft of metatarsal IV is subrectangular in out-
line (Fig. 8.7–8), 20 mm wide transversely and 13 mm long
anteroposteriorly. The distal end of the right metatarsal
IV is preserved in articulation with metatarsals II and III and
with a fragment of the first digit of the fourth toe (Fig. 9).
Hence, some details of its morphology are difficult to es-
tablish. At the proximal break, the shaft of metatarsal IV is
triangular in outline, with a flat medial and posterior sur-
face, and a curved anterolateral side. The distal end is
transversely narrow, expands considerably dorsoventrally
from the shaft and is triangular in outline, tapering antero-
medially. There are no collateral ligament grooves, but only
a large, shallow depression on the medial side (Fig. 9.3–4),
which is straight to slightly convex anteroposteriorly. The
distal articular surface seems to be much more restricted
than in metatarsals II and III, only slightly convex antero-
posteriorly, and does not extend much onto the anterior and
posterior sides of the bone (Fig. 9.4). It is subdivided in its
posterior part into a broader medial condyle, and a narrow,
widely diverging, posteriorly tapering lateral condyle. The
distal end of the metatarsal is approximately 23 mm high
anteroposteriorly and maximally 21 mm wide transversely.
As in the case of metatarsal II, the distal articular end of
metatarsal IV is placed entirely proximal to the distal ar-
ticular end of metatarsal III (Fig. 9.1–2).
Metatarsal V is preserved in articulation on the poste-
rior side of the posteromedial flange of metatarsal IV (Fig.
8.3–4), its proximomedial corner contacting the lateral side
of the proximal posterior bulge of metatarsal III. Its proxi-
mal end is slightly damaged, and the distal end is broken
off. The proximal end of the metatarsal is triangular in out-
line. It is 12 mm wide transversely, ca. 9 mm deep laterally,
and tapers to a rounded point medially. The proximal sur-
face steeply ascends from distolaterally to proximome-
dially, although this might partially be due to erosion.
Towards the distal break, some 36 mm below the proximal
end, the bone becomes slightly flattened, oval in outline
and flexes very slightly laterally, as it is the case in many
tetanuran theropods (Rauhut, 2003). The distal break is 10
mm wide transversely and 6 mm deep anteroposteriorly.
RAUHUT AND POL: NEW JURASSIC THEROPOD FROM PATAGONIA
557
Pedal phalanges
Several fragments of pedal phalanges are preserved.
The proximal ends of pedal phalanges II-1 and IV-1 are pre-
served in articulation with their respective metatarsals
(Fig. 9.1–4). Phalanx II-1 is a stout element, 26 mm high and
approximately 22 mm wide proximally. The proximal ar-
ticular surface is concave, semioval in outline, with a flat-
tened posterior side. On the medial side of the posterior
(ventral) surface, a well-developed longitudinal ridge is
present (Fig. 9.3–4). The ridge disappears some 22 mm
distal to the proximal end and is separated from a pro-
nounced posteromedial edge of the bone by a shallow de-
pression proximally. At the distal break, the phalanx is oval
in cross-section, with a slightly oblique long axis. A crystal-
filled interior cavity shows that the phalanx was hollow, as
is the case with all the other long bones of Pandoravenator.
The pedal phalanx IV-1 has an asymmetric proximal end,
fitting the distal articular surface of metatarsal IV. It is 21
mm wide and ca. 22 mm high. The proximal articular surface
seems to have a semitriangular outline with a flat posterior
side. Posteriorly, the proximal end of the phalanx is slightly
concave transversely, mainly due to a pronounced, slightly
posteriorly expanded lateral margin. Other fragments of
proximal ends of pedal phalanges correspond closely to this
morphology. They are semioval to subtriangular in outline
and have slightly transversely concave posterior surfaces.
In the proximal end of a more distal pedal phalanx, the
proximal articular surface is subdivided into slightly
asymmetrical concavities by a weakly developed medial
ridge. Fragments of distal ends of pedal phalanges indicate
strongly gynglimoidal articulations, with large and deep
collateral ligament pits. Two partial pedal unguals are pre-
served. One is the ungual of the right pedal digit I (Fig. 9.3–
4), which is preserved in approximately its life position with
the distal ends of metatarsals II to IV. Its proximal end is
damaged. The ungual is moderately curved, rather stout
and has a single, ventrally placed claw groove on either
side. As preserved, it is 35 mm long and 16 mm high and
approximately 7 mm wide posteriorly. The dorsal surface is
rounded. The other pedal ungual is only represented by its
proximal end, which is 18 mm high and 10 mm wide. The
proximal articular end is weakly subdivided by a vertical
ridge, and a small flexor tubercle is present close to the
proximal end. The dorsal margin of this ungual is sharp, and
the cross-section asymmetrical, so that the claw groove is
placed further dorsally on one side than on the other. This
ungual might represent that of digit IV.
DISCUSSION
Systematic position of Pandoravenator
To test the systematic position of Pandoravenator, we
included this taxon in three different phylogenetic matrices
that includes a broad array of basal theropods. The first one
is based on the matrix published by Carrano et al. (2012)
with the subsequent modifications introduced by Rauhut
et al. (2016). The second dataset is based on Smith et al.
(2008) with the modifications introduced by Novas et al.
(2015). The third dataset is that published by Rauhut
(2003). We chose these matrices to establish the general
affinities of the new taxon within the major clades of
Theropoda, since they include a broad range of taxa and
representatives of the major theropodan subclades rather
than being focused on a specific subclade (e.g., Benson et al.,
2010; Brusatte et al., 2010; Pol and Rauhut, 2012; Loewen
et al., 2013). The character scorings of Pandoravenator for
these three data matrices are given in the Appendix. The
data matrices were analyzed in TNT v. 1.1 (Goloboff et al.,
2008) using equally weighted parsimony and the character
ordering settings used in the original analyses.
The results of the three different phylogenetic analyses
place Pandoravenator basally within Tentanurae. The spe-
cific positions retrieved for the new taxon within Tetanurae
vary (Fig. 10), which is expected, based on its fragmentary
nature and the different taxon and character sampling
regimes of the three datasets. Nonetheless, our goal was to
test the affinity of Pandoravenator with tetanurans, which is
supported by multiple characters in the three phylogenetic
analyses. The following list gathers the derived characters
that support the inclusion of Pandoravenator in Tetanurae in
the conducted phylogenetic analyses: mid caudal chevrons
L-shaped (Rauhut, 2003: char. 130.1); large anterior process
of chevron base (Novas et al., 2015: char. 224.1); biceps tu-
bercle of coracoid developed as a posteroventrally oriented
ridge (Carrano et al., 2012: char. 227.2; present in allo-
sauroids); distal end of femur with an anteroposteriorly
oriented trough separating lateral and medial convexities
(Carrano et al., 2012: char. 216.1; present in allosauroids and
coelurosaurs); ridge on lateral side of tibia for connection
AMEGHINIANA - 2017 - Volume 54 (5): 539 – 566
558
with fibula clearly separated from the proximal articular
surface (Novas et al., 2015: char. 358.2); ascending process
of astragalus laminar (Carrano et al., 2012: char. 331.1); as-
tragalar condyles significantly expanded proximally on an-
terior side of tibia and face anterodistally (Rauhut, 2003:
char. 217.1); astragalus proximal margin of the medial sur-
face deeply concave anteroposteriorly in medial view
(Novas et al., 2015: char. 382.1; present also in abelisau-
roids); outline of proximal articular surface of metatarsal III
hourglass-shaped (Novas et al., 2015: char. 398.1); well-de-
veloped posteromedial flange on proximal end of metatarsal
IV for articulation with metatarsal III (Novas et al., 2015:
char. 402.1).
As noted above, the position of Pandoravenator is res-
tricted to basal nodes within Tetanurae (Fig. 10), given the
presence of plesiomorphic features that are modified in
more derived tetanurans, such as a rounded medial epi-
condyle of the femur (Carrano et al., 2012: char. 310.0), dis-
tal extensor groove of femur absent (Novas et al., 2015:
char. 351.0; Carrano et al., 2012: char. 311.0), and fibular
condyle on proximal end of tibia confluent with cnemial
crest anteriorly in proximal view (Rauhut, 2003: char. 204.0).
Pandoravenator and its implications for the evolution
of the tetanuran tarsus
One of the unusual features of Pandoravenator is the
morphology of the proximal tarsals. The taxon differs from
all other theropods in the development of two notable tu-
bercles on the astragalus at the junction with the calcaneum
(Fig. 7). One of these is semicircular and placed directly
below the contact of the astragalus with the anteromedial
flange of the fibula, whereas the other is found on the dis-
RAUHUT AND POL: NEW JURASSIC THEROPOD FROM PATAGONIA
559
Figure 10. Phylogenetic positions of Pandoravenator fernandezorum
in the three phylogenetic data matrices used in this study. The trees
represent summarized consensus trees collapsing major clades of
averostran theropods. 1, phylogenetic results using the dataset of
Rauhut et al. (2016), based on Carrano et al. (2012); dashed lines
represent the alternative positions retrieved for Pandoravenator in the
most parsimonious trees, allied to different basal nodes of Tetanu-
rae (results based on reduced strict consensus); 2, phylogenetic re-
sults using the dataset of Novas et al. (2015), based on Smith et al.
(2008); 3, phylogenetic results using the dataset of Rauhut (2003).
tal surface of the bone and seems to be triangular, although
it is not completely preserved. These tubercles probably ar-
ticulated with distal tarsal IV and, maybe, the lateral edge
of distal tarsal II. Their functional significance is unclear, but
these structures have not been described in the literature
or observed by the authors in any other theropod, and are
thus regarded as autapomorphies of Pandoravenator.
Origins of tetanuran apomorphies in the tibia and as-
tragalus
In basal theropods the ascending process of the astra-
galus is usually placed on the lateral half of the bone, with
a well-developed shelf for the reception of the distal end of
the fibula lateral to it. The process itself is triangular in out-
line, usually with a vertical lateral border and gradually rising
proximolaterally (see Welles and Long, 1974; Ezcurra and
Novas, 2007). In proximal view, the process is blocky, its
anteroposterior width reaching approximately half of the
width of the astragalus laterally, and its posterolateral
corner is connected to the posterior margin of the bone by
an oblique ridge that marks the border between the tibial
and fibular facets (e.g., Huene, 1934: pl. 15, fig. 17). The dis-
tal part of the tibia is quadrangular to rectangular in outline
and has a deep slot anterolaterally into which the ascending
process of the astragalus fits (see Ezcurra and Novas,
2007).
In basal tetanurans, in contrast, the distal end of the
tibia is relatively broader and triangular in outline, and the
anterolateral notch is reduced to a narrow, oblique shelf on
the anterior side of the distal end of the bone (see Rauhut,
2003: figs. 44–45). The ascending process of the astragalus
is anteroposteriorly reduced, so that it forms a thin lamina
of bone that overlaps the anterior side of the distal part of
the tibia and contacts the distal part of the fibula with its
usually slightly thickened lateral edge. In basal tetanurans,
the ascending process is triangular to lobular in outline in
anterior view and extends further laterally than in basal
theropods, so that the facet for the fibula on the astragalar
body is reduced and faces mainly laterally (e.g., Torvosaurus,
Britt, 1991; Sinraptor, Currie and Zhao, 1993, Eustreptos-
pondylus, Sadleir et al., 2008). Furthermore, the ascending
process is relatively higher: i.e. as high as or higher than the
astragalar body. The posterior ridge connecting the as-
cending process with the posterior rim of the astragalus is
absent, as the expanded distal end of the tibia extends
laterally onto the calcaneum. The astragalar body is also
modified in tetanurans in comparison to basal theropods.
The distal articular surface is extended anteriorly so that
the articular surface for the distal tarsals and metatarsals
faces anterodistally. A notable transverse groove trans-
verses the articular condyles of the astragalus in most basal
tetanurans (Welles and Long, 1974; Rauhut, 2003).
The situation in basal averostrans and ceratosaurs is
variable, and thus the ancestral tetanuran condition is diffi-
cult to establish. The probably immediate outgroup taxon to
averostrans, Tachiraptor, has only the distal end of the tibia
preserved, and the morphology of its distal end was dis-
cussed by Langer et al. (2014). In the tibia of Tachiraptor, the
notch in the anterolateral end of the distal part of the tibia
is reduced in width and approaches the condition of the
shelf on the anterior surface of the distal part of the tibia in
tetanurans. Furthermore, the tibia is transversely expanded,
with an especially expanded lateral malleolus, resulting in a
roughly parallelogram-shaped outline that is considerably
wider transversely than deep anteroposteriorly. In basal
ceratosaurs, the situation is notably different in Elaphro-
saurus and Ceratosaurus, two of the taxa for which the tibia
and astragalus are known. In Ceratosaurus, the situation is
furthermore complicated by the fact that the astragalus
and calcaneum are fused to one another and preserved in
articulation, if not fused with the tibia (Gilmore, 1920; Mad-
sen and Welles, 2000). Nevertheless, the morphology of the
distal part of the tibia in this taxon seems to be similar to
the tetanuran condition, in that it is considerably expanded
transversely and triangular in distal outline, with a stout,
oblique shelf on the anterior side that extends over ap-
proximately half the width of the bone. However, the astra-
galus has a more transitional morphology. Although the
expanded tibia extends laterally onto the calcaneum and the
ascending process is laminar rather than blocky, the latter is
low, triangular and has a well-developed facet laterally for
the reception of the fibula. Furthermore, the distal condyles
are mainly distally directed, with a very limited overlap onto
the anterior side of the distal end of the tibia (Madsen and
Welles, 2000). In Elaphrosaurus, the distal part of the tibia
is strongly expanded transversely and broadly triangular in
distal outline (Janensch, 1925; Rauhut and Carrano, 2016).
The anterior side of the distal end is flat and lacks an oblique
AMEGHINIANA - 2017 - Volume 54 (5): 539 – 566
560
shelf, thus resembling the condition in coelurosaurs and
some derived abelisauroids (Rauhut, 2003, 2012). As in Ce-
ratosaurus, the astragalar condyles are oriented distally and
the ascending process is very low and transversely narrow
(MB R 4960; Rauhut and Carrano, 2016).
The morphology of the distal part of the tibia and tarsus
in Pandoravenator fits well in this general panorama, espe-
cially if this taxon represents one of the most basal teta-
nurans (e.g., Fig. 10.2–3). The distal part of the tibia of
Pandoravenator is broad and triangular in distal outline, but
has a robust oblique shelf that extends to approximately
half of the width of the distal end, similar to the condition in
Ceratosaurus (Madsen and Welles, 2000). Likewise, the
ascending process of the astragalus is laminar, but re-
markably low and triangular in outline, pointing proximo-
laterally. However, in contrast to ceratosaurs (including
Ceratosaurus), the distal condyles overlap the anterior sur-
face of the distal part of the tibia so that the condyles are
anterodistally directed, as in all tetanurans. Thus, Pando-
ravenator shows a novel combination of characters that may
represent a transitional morphology in the evolution from
the basal theropod tarsus to the modified tetanuran condi-
tion. The tarsus of Pandoravenator suggests that the ex-
tension of the astragalar condyles onto the anterior surface
of the tibia (and the consequent functional changes in the
mesometatarsal articulation) preceded the development of
the high ascending astragalar process in the evolution of
Tetanurae.
Astragalus ascending process as a separate ossification
A feature of special interest in the astragalus of Pando-
ravenator is the presence of a deep and continuous groove
along the ventral end of the anteroproximal expansion of
the distal condyles in the right element (Figs. 7.2, 11). Ex-
tant paleognath birds have a broad ascending process
connected to the astragalus that resemble the condition of
non-avian theropods (McGowan, 1985), whereas neognath
birds have a “pretibial bone” that ossifies as a proximal pro-
jection of the calcaneum (Martin et al., 1980). The homology
of these structures and the astragalar ascending process of
non-avian theropods have been controversially discussed
in the 1980’s and embryological information led to alterna-
tive interpretations. Some studies argued that these two
processes originate from cartilaginous projections of the
astragalus in all birds (McGowan, 1984, 1985), whereas
others interpreted the “pretibial bone” as a development
from a neomorphic cartilaginous precursor separated from
the astragalus and calcaneum in neognaths (Martin et al.,
1980; Martin and Stewart, 1985). Ossa-Fuentes et al. (2015)
have recently clarified the early cartilaginous development
in birds and indicated that the ascending process of both
paleognath and neognath birds originates from the primor-
dial cartilagenous intermedium. Ossa-Fuentes et al. (2015)
showed that birds have a developmental pathway unseen
in other amniotes, in that the cartilaginous intermedium
first expands proximally along the distal part of the tibia
as the ascending process and later fuses with the tibiale
(condylar region of astragalus) and calcaneum to form a
single large cartilaginous cap for the tibia.
Although Ossa-Fuentes et al. (2015) have clarified the
developmental decoupling of the intermedium from the as-
tragalus and even noted that a separate ossification may
have been present already in enanthiornithines, a major
pending question is when this unique process appeared in
the evolution of the lineage to birds. As noted by these au-
thors, the decoupling event may have occurred as early as in
dinosauriforms, but data from unfused tarsal elements is
scarce among non-avian dinosaurs. Within theropods, pre-
vious reports (Welles, 1983) on a suture between the as-
cending process and the astragalar body in the basal
theropod Dilophosaurus have been reconsidered and the
alleged suture line cannot be distinguished from other mul-
tiple breaks in the astragalus of this taxon (Ossa-Fuentes et
al., 2015). On the other hand, the highly unusual tarsus of
the heterodontosaurid Fruitadens shows a small separate
ossification at the proximal end of the ascending process of
the astragalus (Butler et al., 2012), but the unusual mor-
phology of the tarsus in general and the position of the
separate ossification make its homology with the avian con-
dition questionable (Ossa-Fuentes et al., 2015).
The tarsus of Pandoravenator provides additional evi-
dence on the homology of the tarsal structures in birds and
non-avian theropods and has thus implications for our un-
derstanding of the evolution of the theropodan tarsus. We
interpret the very conspicuous groove on the anterior side
of the right astragalus of Pandoravenator as the remnants
of a suture between two different centers of ossification.
Assuming the homology of the separate ascending process
RAUHUT AND POL: NEW JURASSIC THEROPOD FROM PATAGONIA
561
of Pandoravenator with the intermedium of birds that ossi-
fies into the ascending process in these animals (Ossa-
Fuentes et al., 2015), this discovery thus presents the first
direct evidence that supports the appearance of the unique
bird developmental pathway at least in basal tetanurans,
much earlier than the origin of birds.
Interestingly, an at least partial and usually shallow
groove across the anterior side of the astragalar condyles
is found in several ceratosaurs and many basal tetanurans
(Welles and Long, 1974; Rauhut, 2003). This groove often
coincides roughly with the boundary between the main as-
tragalar body and the anteroproximal expansion of the as-
tragalar condyles, even in cases where the latter is very
restricted in its expansion, as in the ceratosaur Ceratosaurus
(Madsen and Welles, 2000). In basal tetanurans the groove
also extends across the anteroproximal expansion of the as-
tragalar condyles (e.g., Allosaurus, Magnosaurus, Torvosaurus,
sinraptorids; Rauhut, 2003: p. 122) and is sometimes noted
as a notch between the distal astragalar body and the an-
teroproximal expansion of the bone, closely comparable in
position to the probable suture seen in Pandoravenator (Fig.
11). Pandoravenator helps interpreting the groove commonly
found in ceratosarus and basal tetanurans as a remnant of
the suture between two different centers of ossification in
the theropod astragalus, likely representing the separate
ossifications of the tibial cartilage as the astragalar body
and the intermedium as the ascending process described
for birds (Ossa-Fuentes et al., 2015). The horizontal groove
has been found as a probable averostran synapomorphy
(e.g., Rauhut, 2003; Carrano and Sampson, 2008) and is
correlated with other changes of the tarsal joint in these
animals, such as the lateral expansion of the distal part of
the tibia to back the distal fibula and articulate with the
calcaneum and the change from a rather massive, pyrami-
dal ascending process to a sheet-like ascending process of
the astragalus. These changes probably imply significant
modifications in theropod locomotion from non-averostran
to averostrans, but functional and biomechanical details
about this apparently major change still need to be studied.
It is also interesting to note that coelurosaurs lack the
groove across the anteroproximal expansion of the astra-
galar condyles described above, but bear a more proximally
located discontinuous groove between the convex condyles
and the laminar ascending process, which has been noted
for many coelurosaur groups, including Archaeopteryx and
basal birds (Mayr et al., 2007; Chiappe et al., 2007). Based on
the variation of the proximodistal and lateromedial exten-
sion of the intermedium cartilage among extant birds
(Ossa-Fuentes et al., 2015: suppl. fig. 2), the condition of
coelurosaurs may thus represent a proximal shift of the
boundary between the tibial and the intermedium cartilages.
CONCLUSIONS
The remains described here represent the first thero-
pod for the Cañadón Calcáreo Formation fauna, a Late
Jurassic assemblage previously represented by diplodocoid
and macronarian neosauropods, basal crocodylomorphs,
and coccolepid and basal teleosts fishes. As such, it is also
the first representative of this clade from the Late Jurassic
of Argentina in general. Pandoravenator contributes to the so
far limited knowledge on the Late Jurassic theropod assem-
blages from Gondwana, until now restricted to the cerato-
saurs and basal tetanurans of the Tendaguru Formation
(Tanzania) and the bizarre Chilesaurus from the Toqui For-
mation (Chile). The Tendaguru Formation records are so far
the most diverse, and the tetanurans and ceratosaurs
from this unit have been interpreted as indicating a thero-
pod fauna more similar to those from the Cretaceous of
Gondwanan than to other Late Jurassic theropod faunas
from the northern hemisphere (Rauhut, 2011). The thero-
AMEGHINIANA - 2017 - Volume 54 (5): 539 – 566
562
Figure 11. Left astragali in medial view. 1, Pandoravenator fernande-
zorum, MPEF PV 1733-36b. 2, Allosaurus fragilis, MOR 693. Abbrevia-
tions: asc, ascending process; no, notch. Scale bars= 10 mm.
pod records from the Late Jurassic of South America do not
provide evidence either in favor nor contradicting this bio-
geographical scenario. In the case of Chilesaurus, interpreted
as a basal tetanuran by Novas et al. (2015), its bizarre mor-
phology is so autapomorphic that is hard to find particular
affinities with other theropod species. In the case of Pando-
ravenator, it probably represents a basal branch of Tetanu-
rae, but the fragmentary nature of the known remains
precludes a more precise assessment of its phylogenetic
affinities.
The significance of Pandoravenator, instead, relies more
on its information regarding the evolution of a set of particu-
lar features that characterize the averostran or tetanuran
tarsus. The morphology of the distal part of its tibia and
proximal tarsals has a combination of plesiomorphies and
apomorphies that sheds light on the origin of the tetanuran
tarsus. Plesiomorphies include the presence of a relatively
deep shelf along half of the anterior surface of the distal
part of the tibia and a low and triangular ascending process
of the astragalus, resembling the condition of some cerato-
saurs. Tetanuran apomorphies include the extension of the
articular surface of the astragalus onto the anterior surface
of the distal part of the tibia and a triangular outline of the
distal margin of the bone.
Finally, the ascending process of the astragalus of Pan-
doravenator has a deep and continuous groove across the
anterior surface of the astragalar condyles that is inter-
preted as a sutural contact between two centers of ossifi-
cation that would form the astragalus. This resembles the
two centers of ossification recognized in the astragalus of
extant birds and leads us to interpret a similarly located
groove in ceratosaurs and non-coelurosaur tetanurans as a
possible remnant of a suture. This interpretation implies
that the unique developmental pattern seen in the astra-
galus of extant birds arose at least at the base of Averostra,
which provides a minimum age for the event of develop-
mental decoupling of the intermedium from the astragalus
in the late Early Jurassic when the ceratosaur and tetanuran
lineages diverged from one another.
ACKNOWLEDGEMENTS
Fieldwork and research were funded by DFG RA1012/9-1 and RA
1012/18-1 (to O.W.M.R.), ANPCyT PICT 0808 and 1288 (to D.P.), and
by the Fundación Egidio Feruglio. Thanks are due to the participants
of the 2002, 2009 and 2017 Jurassic expeditions. We thank L. Austin
who conducted all illustrations and took most photographs used in
this work. Critical comments by the editors, as well as M. Ezcurra, J.
Choiniere, and A. Otero greatly helped to improve the manuscript.
We also thank the Secretaría de Cultura de la Provincia del Chubut
for granting permits to conduct fieldwork.
REFERENCES
Baron, M.G., and Barrett, P.M. 2017. A dinosaur missing-link? Chi-
lesaurus and the early evolution of ornithischian dinosaurs. Bio-
logy Letters 13: 20170220.
Benson, R.B.J. 2010. A description of Megalosaurus bucklandii (Di-
nosauria: Theropoda) from the Bathonian of the UK and the re-
lationships of Middle Jurassic theropods. Zoological Journal of
the Linnean Society 158: 882–935.
Benson, R.B.J., Carrano, M.T., and Brusatte, S.L. 2010. A new clade
of archaic large-bodied predatory dinosaurs (Theropoda: Al-
losauroidea) that survived to the latest Mesozoic. Naturwis-
senschaften 97: 71–78.
Bocchino, A. 1967. Luisiella inexcutata gen. et. sp. nov. (Pisces, Clu-
peiformes, Dussumieridae) del Jurásico superior de la Provincia
de Chubut, Argentina. Ameghiniana 4: 91–100.
Bonaparte, J.F. 1986. Les Dinosaures (Carnosaures, Allosauridés,
Sauropodes, Cétiosauridés) du Jurassique moyen de Cerro Cón-
dor (Chubut, Argentine). Annales de Paléontologie 72: 247–289,
326–386.
Bordas, A.F. 1943. Peces del Cretácico del Río Chubut (Patagonia).
Physis 19: 313–318.
Britt, B.B. 1991. Theropods of Dry Mesa Quarry (Morrison Forma-
tion, Late Jurassic), Colorado, with emphasis on the osteology
of Torvosaurus tanneri. BYU Geology Studies 37: 1–72.
Brochu, C.A. 2003. Osteology of Tyra<