A new specimen of the elasmosaurid plesiosaur Zarafasaura oceanis from the Upper Cretaceous (Maastrichtian) of Morocco

Abstract

A new specimen of the Moroccan elasmosaurid plesiosaur Zarafasaura oceanis from the Maastrichtian (Upper
Cretaceous) phosphate deposits of Morocco is described. A partial skeleton with an associated skull is mounted ondisplay at the Wyoming Dinosaur Center, USA. The short preorbital region of the skull and contact of the squamosal with the parietals, jugal and postorbital, allow for identification to the genus. However, large amounts of the skull and mounted skeleton are reconstructed so description is difficult in all but a cursory way. In recent years, the popularity of Moroccan material among collectors and the general public has increased the number of display specimens in museums across the world. As plesiosaurs remain very scarce in these rich marine vertebrate strata, any information expanding our understanding of the osteology of a known species warrants study. The new specimen of Zarafasaura provides information on postcranial elements which were previously unknown, along with additional information regarding a fairly complete skull that preserves elements that are missing in the holotype specimen.

Full text

Paludicola 9(2):97-109 April 2013
© by the Rochester Institute of Vertebrate Paleontology
97
A NEW SPECIMEN OF THE ELASMOSAURID PLESIOSAUR ZARAFASAURA OCEANIS
FROM THE UPPER CRETACEOUS (MAASTRICHTIAN) OF MOROCCO
Dean R. Lomax1, 2 and William R. Wahl³
¹Doncaster Museum & Art Gallery, Chequer Rd, Doncaster, DN1 2AE, UK
2School of Earth, Atmospheric and Environmental Sciences, The University of Manchester, Oxford Road,
Manchester, M13 9PL, UK, dean.lomax@doncaster.gov.uk
³BigHorn Basin Foundation, 110 Carter Ranch RD, Thermopolis, Wyoming 82443, wwahl2@aol.com
ABSTRACT
A new specimen of the Moroccan elasmosaurid plesiosaur Zarafasaura oceanis from the Maastrichtian (Upper
Cretaceous) phosphate deposits of Morocco is described. A partial skeleton with an associated skull is mounted on
display at the Wyoming Dinosaur Center, USA. The short preorbital region of the skull and contact of the squamosal with
the parietals, jugal and postorbital, allow for identification to the genus. However, large amounts of the skull and mounted
skeleton are reconstructed so description is difficult in all but a cursory way. In recent years, the popularity of Moroccan
material among collectors and the general public has increased the number of display specimens in museums across the
world. As plesiosaurs remain very scarce in these rich marine vertebrate strata, any information expanding our
understanding of the osteology of a known species warrants study. The new specimen of Zarafasaura provides
information on postcranial elements which were previously unknown, along with additional information regarding a fairly
complete skull that preserves elements that are missing in the holotype specimen.
INTRODUCTION
Plesiosaurs are an extinct group of predatory
marine reptiles which have been discovered from the
uppermost Triassic (Rhaetian) to the uppermost
Cretaceous (Maastrichtian; see discussion in Smith
and Vincent, 2010). Remains of plesiosaurs have
been found on all continents, with some key
discoveries made in North America, Europe and
Australia. They achieved a worldwide distribution
during the Jurassic. Specifically, the elasmosaurids
flourished during the Maastrichtian (Gasparini et al.,
2003). In spite of this, the record of plesiosaurs from
Africa is relatively poor, with only five taxa currently
recognized, the latest of which being Zarafasaura
oceanis (Vincent et al., 2011). A phylogenetic
analysis of plesiosauroids compiled by Vincent et al.
(2011) indicates that Zarafasaura shares close
affinities with elasmosaurids from the Late
Cretaceous of North America and Japan. This newly
described elasmosaurid represents the first valid
plesiosaur from the latest Cretaceous of Africa
(Vincent et al., 2011). The scarcity of plesiosaur
remains in Morocco has remained somewhat a
mystery. In spite of such a low number of reported
plesiosaurian discoveries in Africa, the continent has
produced some excellently preserved and
scientifically important fossils. Such remains have
helped to develop a better and more complete
understanding of the paleoenvironments and extinct
flora and fauna of the region. Specifically the
Cretaceous-Paleogene phosphatic deposits of
Morocco are famed for their rich fauna of marine
vertebrates, such as crocodyliformes, birds,
chelonians and a variety of mosasaurs (Bardet et al.,
2010; Vincent et al., 2011).
The previous description of isolated material
from this formation assigned it to the “wastebasket”
taxon Plesiosaurus mauritanicus (Arambourg, 1952).
They include the holotype tooth (MNHN PMC 22)
and a referred specimen consisting of thirteen
articulated posterior cervical to anterior dorsal
vertebrae (MNHN PMC 23; Arambourg, 1952).
Several isolated teeth and one vertebra (MNHN PMC
24, 25, 26-28) were also referred by Arambourg
(1952). However, P. mauritanicus was regarded as
anomen dubium by Vincent et al. (2011) because: 1)
the holotype and paratype specimens are composites,
derived from two different individuals from two
distinct localities; 2) both the holotype and paratype
do not exhibit any autapomorphies; 3) the holotypic
tooth is morphologically similar to many other Late
Cretaceous elasmosaurids; and 4) the genus
Plesiosaurus is restricted to the Lias. Based on the
geographical location and scarcity of other
plesiosaurs, much of the material collected from this
region is often attributed to this taxon. As isolated
remains, including teeth, are deemed
morphologically similar to P.mauritanicus, the
overview of fragmentary material from this region
reduces their usefulness. More recently Vincent et al.
(in press) reported several new plesiosaur specimens
from the same locations within the Maastrichtian
Phosphates of Morocco. This material comprises
numerous specimens, including partial skeletons (e.g.
OCP-DEK/GE 115), limb elements, and isolated

98 LOMAX AND WAHL—PLESIOSAUR ZARAFASAURA FROM MOROCCO
series of cervical/dorsal vertebrae. All of the
described material was determined to belong to
Elasmosauridae gen. et sp. indet and could not be
referred to Zarafasaura, as no cranial elements were
found associated. Thus at present, isolated elements
cannot be identified taxonomically. But more
complete specimens, such as WDC CMC-01, will
help solve the issue of identifying fragmentary or
isolated material, such as that described by Vincent
et al. (in press).
As the majority of plesiosaur remains from
Africa are often rather fragmentary or of poorly
constrained provenance, the importance of recording
more complete or diagnostic material is vital.
Vincent et al. (2011) reported on a dorsoventrally
crushed skull (holotype), and a partial mandible (the
paratype) belonging to a new elasmosaurid,
Zarafasaura oceanis. No post cranial elements are
associated with the type material. Here we describe
a new, more complete specimen of Z. oceanis from
the Upper Cretaceous phosphatic mines of Oued-
Zem, Morocco. The specimen includes a nearly
complete skull and numerous postcranial elements
including limbs, pectoral and pelvic material, and
vertebrae. Cranial features of WDC CMC-01
identify this individual as belonging to the taxon.
This paper provides a report on this relatively
complete individual, although the reconstruction of
the material and mount of the specimen makes a
thorough description difficult.
Abbreviations Used—FFHM, Fick Fossil
Museum, Oakley, Kansas (USA); MNHN Muséum
national d’Histoire naturelle, Paris (France); OCP,
Office Cherifien des Phosphates Khouribga
(Morocco); UCMP, University California Museum
of Paleontology, Berkeley (California) USA; WDC,
Wyoming Dinosaur Center, Thermopolis (Wyoming)
USA.Materials and Methods—WDC CMC-01 was
collected from the type location of OCP-DEK/GE
315, the holotype of Zarafasaura oceanis. The
location of Sidi Daoui is near the city of Oued Zem,
situated within the Khouribga Province of the NE
Oulad Abdoun Basin. Here the Maastrichtian
phosphatic series is very condensed. The series is
considered to be late Maastrichtian in age and
comprises from base to top: basal grey phosphatic
limestone named ‘basal bone-bed’; soft yellow
phosphates named ‘Couche III inférieure’ (Lower
CIII level); soft grey phosphates with brown streaks
named ‘Couche III supérieure’ (Upper CIII level);
and yellow marls (Vincent et al., 2011). WDC
CMC-01 derived from the Upper CIII level, Point A
(Figure 1). The specimen was discovered and
excavated in April 2004, by collector Mohamed Atid.
It was primarily collected in five large plaster
jackets, with numerous smaller blocks containing the
phosphate matrix and isolated bones. The soft grey
phosphatic blocks contained numerous elements,
including several articulated sections of the vertebral
column, which were largely associated with ribs
(Figure 2).
The studied specimen was fully prepared and
mounted for display (Figure 3). To ensure no
information was lost and the bones and orientation of
elements were correct as possible, each fragment of
bone was individually marked and numbered before
preparation. The majority of the posterior caudal
vertebrae are entirely carved, with other parts of the
postcranial skeleton reinforced with plaster, and
sculpted. Initially the skull was found dorsolaterally
flattened (Figure 4), and later taken apart piece-by-
piece and constructed in three-dimensions for display
and public auction. With the collection and
preparation of the studied specimen being rendered
for display, it is important to note that the entire
reconstruction was based upon specimen UCMP
33912, Hydrotherosaurus alexandrae (Welles,
1943), an elasmosaurid from the Maastrichtian of
California. The specimen was subsequently acquired
by the WDC in 2006.
DESCRIPTION
The following description of the studied
specimen is largely based on the original photos of
the articulated skeleton and crushed skull, although
some of the descriptions are derived from the
tentative reconstruction, and are so stated.
Skull and Mandible—Due to the nature of the
original preservation and subsequent reconstruction
of the skull, certain cranial features such as sutures
and individual bones are difficult to distinguish with
certainty. As the skull has been reconstructed in
three-dimensions, its original morphology cannot be
determined. Additionally, the lack of many
distinctive cranial features could also be due to
distortion, which may have slightly deformed the
general pattern and shape of the skull. Careful
preparation of the specimen allowed the skull to be
reconstructed with several autapomorphies intact,
including the short snout and wide palate. Some
carved elements of the skull have slightly changed its
original shape, but not dramatically. For this study,
the descriptions of the skull are derived from
photographs of the unprepared skull, where possible.
In describing the morphology of individual bones of
the skull, the use of original photographs were vital.
The skull was found partially articulated with
both the fused atlas-axis, and at least fourteen partly
articulated cervical vertebrae. It was compressed
dorsolaterally and slightly crushed, somewhat similar
to the holotype specimen that is crushed
dorsoventrally (Vincent et al., 2011, fig. 2a-c). After
preparation, the skull was rebuilt with each cranial
element pieced back together (Figures 5, 6),
assuming a typical elasmosaurian shape, ogival in
dorsal aspect and slightly rectangular in lateral
aspect. As reconstructed, the skull of WDC CMC-01
is the same as the holotype, bearing a shorter

PALUDICOLA, VOL. 9, NO. 2, 2013 99
FIGURE 1. Geographical and stratigraphical location of Zarafasaura oceanis. A. Main phosphatic basins of Morocco. B. Detail of the Oulad
Abdoun Basin, with indication of the Sidi Daoui area where Zarafasaura oceanis has been found. C. Stratigraphical column of t he Maastrichtian
phosphatic levels in Sidi Daoui area, with indication of the level where the holotype and paratype specimens of Zarafasaura oceanis have been
found. Abbreviations Ph, phosphates; Ma, marls; Li, limestones. From Vincent et al. (2011), reproduced with permission.
____________________________________________________________________________________________________________________
preorbital (about 28% of skull length) and postorbital
section, with a long relatively low parietal crest. The
front half of the skull is almost entirely original, with
very little reconstruction. The premaxillae are both
present and give a short snouted appearance to the
skull, as suggested for the taxon by Vincent et al.
(2011). Like the holotype, the anteriormost portion
of the snout is rounded, suggesting an oval cross
section at the rostrum. Due to the short snout of the
animal, the nares are close to the anterior tip of the
skull, as in Libonectes (Carpenter, 1999). The
premaxillae are complete and fused into a short, stout
bone. They form a rounded apex that gives the
anteriormost section a ‘soft’, rugose and triangular
appearance, emphasized by the dorsal surface, which
has multiple tiny nutritive foramina. The
anteriormost portion of the premaxillae is missing in
the holotype specimen. Between the premaxillae in
WDC CMC-01, a long, original midline suture
contact extends up to the level of the external nares,
where the premaxillae constitute the anterior and
medial margins of the external naris. The facial
process is preserved, though it cannot be
distinguished in the three-dimensional recon-
struction. It can, however, be seen on the original
crushed skull where it contacts the parietal
posteriorly (Figure 4). Maxillae are almost complete,
and both the right and left are original. The maxilla is
a long rectangular bone, contacting the premaxillae
anteriorly, below the nares. It forms most of the
anterolateral margin of the orbit, as in the holotype.
A suture line between the maxillae and premaxillae
can be seen in both the reconstruction and in the
images of the original skull (Figures 4, 5, and 6). The
left maxillary segment is near complete but the
posterior end is missing and broken, although it has
been rebuilt in the reconstruction.
Prefrontals are both preserved, and they are
fairly long and located between the premaxilla and
maxilla. The prefrontals form the dorsomedial
margin of the orbit. Looking at the original skull, the
prefrontal forms a fraction of the external naris, but
in the reconstruction it forms a larger amount. The

100 LOMAX AND WAHL—PLESIOSAUR ZARAFASAURA FROM MOROCCO
FIGURE 2. Two of the original five large field jackets. A. The large block displays the posterior cervicals, pectorals and dorsal vertebrae; along
with the pectoral girdle. B. This large block follows the first, it consists solely of dorsal vertebrae, the left humerus and numerous ribs, note the
long and ‘squared off’ neural spines. Scale bar = 20 cm.
____________________________________________________________________________________________________________________
FIGURE 3. Complete mounted skeleton of WDC CMC-01, including reconstructed skull. Scale bar = 20 cm.
___________________________________________________________________________________________________________________
right prefrontal is fairly complete and contacts what
appears to be the frontal, posteriorly. The left is also
preserved, but partly broken and reconstructed. The
frontals are difficult to distinguish in the original
photo of the skull. In the reconstruction, preserved on
the left side of the skull is a triangular-shaped bone,
possibly the postfrontal, although it appears to have
been placed next to the anterior section of the parietal
rather than in its original position. On the inner side
of both the prefrontals and contact with the maxilla
are an arrangement of relatively large foramina, but it
is unclear what they are. It is possible that they
could be an artifact of preparation.
The parietal is relatively complete, though
partially crushed and distorted. It is a rather narrow
bone. The suture line cannot be distinguished with
confidence. The anteriormost section of the parietal
seems to contact the long facial process of the
premaxillae, though this is difficult to determine. The
center is genuine, but is damaged by fractures and
crushing. It is rebuilt at the contact between the
posterior parietal and anterior squamosal. At the

PALUDICOLA, VOL. 9, NO. 2, 2013 101
level of half the temporal fenestra length, the
squamosal unites with the parietal dorsoposteriorly.
In the reconstruction, however, the contact between
the anterior squamosal and posterior parietal is
almost entirely rebuilt, whereas the original
photographs display only matrix and bone fragments.
_________________________________________
F
IGURE 4. Original skull of WDC CMC-01 post preparation and
prior to reconstruction. A. Skull in dorsolateral view. Note the
sutures between the premaxillae and maxillae (lower arrow) and
the bone numbered ‘21’, which may be that protruding into the
temporal fenestra (upper arrow). B. Skull in ventral view. Note
the atlas-axis and basioccipital (arrow). Scale bar = 10 cm.
_____________________________________________________
The squamosal morphology of Zarafasaura is
an autapomorphy of the taxon, forming a distinctive
contact with the parietals, jugal, and anterior contact
with the postorbital. It allows the squamosal to form
the whole of the lateral temporal fenestra margin
(Vincent et al., 2011), and, as a whole, dominates the
posterior of the skull (Figures 5, 6). In the original
skull, it is difficult to determine confidently the
contact of the squamosal with the jugal and
postorbital, although this can be seen in the
reconstruction. The dorsal ramus of the squamosal is
relatively elongate and thin. In the original skull, the
jugals cannot be identified, although there are several
large sections of bone which likely belong, in part, to
the jugals. For the reconstructed skull, when viewed
in left lateral view (Figures 5A, 6A), a large
rectangular bone is present. The jugal is rectangular
shaped in most plesiosauroids and in the holotype
specimen of Zarafasaura. We are not confident that
this element is the jugal, though its large size and
positioning would suggest it as a likely candidate.
There appears to be no suture lines present on this
bone and it is bounded anteriorly by the postorbital,
though the most anterior section is reconstructed.
Quadrates are indeterminable in the original skull,
but appear to be part of the reconstruction. Although
the right is relatively more complete, few details can
be derived. Both are surrounded by the reconstructed
posterior edge of the squamosal.
The palate is largely rebuilt, though part of the
right and left is real with a suture separating the two.
There is no suture delineation of the vomer. As the
anteriormost section of the palate is missing, it gives
a triangular shaped appearance to the palate. The
braincase is present but based on the reconstruction
is difficult to describe. The basioccipital is original,
with a well-rounded occipital condyle. The ventral
surface of the brain case is almost entirely
reconstructed (Figures 4B, 5D and 6D). The
parabasisphenoid, pterygoid and epipterygoid appear
to be present, though the pterygoid is rebuilt distally
towards the interpterygoid vacuities. The orbits are
rounded, well preserved, and easily distinguished. In
the original skull, the orbits are seen in dorsal aspect
and appear slightly more oval than rounded, as
reconstructed. This may be due to flattening of the
skull, during diagenesis. Like in the holotype
specimen, the orbits are bordered by the prefrontals
and frontals medially, the maxillae and ?jugals
laterally, and the postorbitals posteriorly (Vincent et
al., 2011). The maxilla comprises the majority of the
orbital margin. The temporal fenestrae are relatively
large and occupy >55% of the reconstructed skull.
This figure is much higher than the 45% of the
holotype (Vincent et al., 2011). The preservation of
the parietal and squamosals in the unprepared skull
(Figure 4) indicates oval-shaped temporal fenestrae,
shorter than in the reconstructed skull. A dorsally
projecting triangular process on the left ?squamosal
protrudes into the temporal fenestra and appears to
be broken or distorted. It is unclear whether this is
original, or an artifact of preparation and/or
reconstruction. The unusual triangular process may
be that numbered ‘21’ (Figure 4A), but it does not
appear to be in its original position, and may be
incorrectly placed on the reconstruction.
The mandible of WDC CMC-01 is mostly
original. The dentary is a long and straight bone,
which is almost complete. The lateral[?] surface
bears a longitudinal groove which contains an
ornamentation of tiny foramina. The dentary has a
diastema between the sixth and seventh pair of teeth.
In relation to the length of the skull, the mandibular
symphysis is short and is covered with relatively
strong, pitted ornamentation. The diamond shape
between the two rami of the mandible (Vincent et al.,
2011, p. 8) is unclear in WDC CMC-01, as this area
is partly rebuilt. As in the holotype, a slightly
elevated crest is present where the rami meet the

102 LOMAX AND WAHL—PLESIOSAUR ZARAFASAURA FROM MOROCCO
FIGURE 5. Reconstructed three-dimensional skull of WDC CMC-01 viewed in different aspects. A. Left lateral view. B. Right lateral view. C.
Dorsal view. D. Ventral view. Lighter colored areas are reconstructed. Scale bar = 10 cm.
FIGURE 6. Interpretive illustrations of the three-dimensionally reconstructed skull of WDC CMC-01 viewed in different aspects. A. Left lateral
view. B. Right lateral view. C. Dorsal view. D. Ventral view. Note that white = original bone; grey = bone from the opposite side of the skull;
black = holes and/or foramina; and lines = reconstruction. An isolated question mark indicates that the particular section may or may not be
original and cannot be identified. Abbreviations: Bo: Basioccipital; F: Frontal; Ju: Jugal; Mx: Maxilla; N: External naris; Orb: Orbit; Pal:
Palatine; Par: Parietal; Pf: Postfrontal; Pmx: Premaxilla; Po: Postorbital; Prf: Prefrontal; Q: Quadrate; Sq: Squamosal; Tf: Temporal fenestra. A
question mark with an abbreviation indicates a tentative identification. Illustrations by Reece Davies.
____________________________________________________________________________________________________________________
mandibular symphysis. The dentary bears sixteen
pairs of teeth. The meckelian canal extends the
length of the dentary to the boundary of the angular.
On the inner side of the left dentary are an
arrangement of nearly vertical marks, perhaps a sign
of scavenging or a feature of preservation or
preparation (Figure 7). Both the angular and
surangular are almost complete. The lateral surface
of the surangular is ornamented by strong ridges and
is slightly reminiscent of heavy muscle attachment.
Both articulars are original, but perhaps slightly
deformed. They were reconstructed for display, as
the right does not match the left and appears more
bulbous in ventral aspect.

PALUDICOLA, VOL. 9, NO. 2, 2013 103
FIGURE 7. Vertical, serration-like marks preserved on the inner
side of the left dentary. Lighter colored area is reconstructed.
Scale bar = 10cm.
_____________________________________________________
FIGURE 8. The odd, strong rugosity that surrounds the base of the
posterior premaxillary teeth and is preserved atop the meckelian
crest. Scale bar = 4cm.
_____________________________________________________
As many as 12 partial teeth were discovered in
situ in their original position in the jaw (Figure 5).
Very few were preserved with a complete apex; the
majority were broken or eroded. From information
supplied from the preparator, every tooth, except for
a few tips that are slightly protruding from the
mandible, were replaced with other isolated teeth
from the same formation. The same tooth type,
which was described for Plesiosaurus mauritanicus,
are found in all the phosphatic Maastrichtian levels
of Morocco (Vincent et al., 2011). Of the original
teeth, the crown is rather slender and linguolabially
compressed and recurved, forming a thin cone-
shaped morphology with a pointed apex. Fine
longitudinal enamel ridges are present on both the
lingual and labial surfaces, and they run the length of
the tooth. This general tooth morphology is typical
for plesiosauroids (Vincent et al., 2011) and
considered a piercing tooth form by Massare (1987).
There are a total of 6 premaxillary teeth and 10
maxillary teeth. The largest teeth of WDC CMC-01,
as in the holotype, occur around the premaxilla-
maxilla suture as the sockets are larger at this point
than elsewhere. In this area, an unusual rugosity
surrounds the tooth alveoli and the base of the tooth
itself (Figure 8). This can also be seen in the anterior
of the dentary surrounding the teeth above the
meckelian canal.
Axial Skeleton—The vertebral column was
preserved in five large sections (Figure 9). For the
mount of WDC CMC-01, numerous copies of
vertebrae were created, based upon
Hydrotherosaurus, in order to complete the vertebral
column. The most prominent reconstructed sections
of the vertebral column were the cervicals and
caudals. Including the casts, there are a total of 121
vertebrae (counting the atlas and axis as two)
associated with WDC CMC-01: 56 cervicals (35
original); 4 pectorals (all original); 26 dorsals (all
original); 3 sacrals (all original) and 30 caudals (7
original).
Atlas-Axis and Cervical Vertebrae: The first
block, which contained the skull, included the fused
atlas-axis complex, which is rather elongated and
partially reconstructed in the center (Figure 4B). The
atlas-axis was preserved on the same block as at least
fourteen cervical vertebrae. An additional cast
vertebra was inserted after the atlas-axis. We feel
that this is incorrect and do not include the inserted
cervical in the following description (or cervical
count, above). The cervical centra are fairly circular,
slightly wider than they are tall, and display a
prominent rim. The first cervical (not the cast) is
partially reconstructed. Cervicals 1, 2 and 3 are
moderately smaller than the rest. Both c6 and c7 are
partly reconstructed, especially c7, as the majority of
the vertebra was poorly preserved and broken on one
side. Cervicals 8, 9, 10, 11 and 12 are well
preserved. Cervical 13 is also largely rebuilt and was
badly broken when collected. Cervical 14 consists of
the neural spine and a very poorly preserved
centrum. After c14 is the insertion of 20 cast
vertebrae into the cervical series. There is a
distinctive difference between the last original
anterior cervical and the next original posterior
cervical, thus suggesting that vertebrae are missing.
The preparator opted to include an additional series
of vertebrae at this point. The next original vertebra
(c35) is relatively robust and preserves the majority
of the centrum and a near complete neural spine,
which is also far more robust than those of the
previous cervicals. The next fourteen vertebrae are
similarly preserved to that of c35, most missing
sections of the cervical ribs, which have been
reconstructed. Cervicals 51, 52 and 53 are relatively
complete, though like the previous are missing
fragments of the ribs. The only feature that can be
noted is that the ribs are very rugose on both sides,
which is probably not diagnostic. All cervical ribs
are fused to the centra, suggesting an adult age for
WDC CMC-01.

104 LOMAX AND WAHL—PLESIOSAUR ZARAFASAURA FROM MOROCCO
FIGURE 9. Original in situ sketch of entire specimen as discovered, with individual blocks placed together. The arrow indicates missing
cervical vertebrae. Numbers do not include the casts of cervical and caudal vertebrae. Scale bar = 10 cm.
____________________________________________________________________________________________________________________
From the atlas-axis to the insertion of the casted
cervicals (at cervical 16), the neural spines have the
same height from the post-zygapophysis to the top of
the neural arch. The neural spines are inclined
caudally at an angle of 45° from the horizontal, much
the same as Libonectes (Buchy, 2005). Several
neural arches from the posterior cervicals (post c35)
retain a bulbous edge at the top of the spinous
process. This bulbous edge was not reconstructed in
the corresponding vertebrae casts or on reconstructed
neural arches post c35. This bulbous feature appears
natural and may have been a bone lesion associated
with tendon attachment. Also post c35, the neural
arches are “squared off” at the tip. This feature is
consistent to the caudals, regardless of whether it is
real bone or a cast.
Pectoral Vertebrae: In the total series
(including atlas-axis) vertebra 58 is the first pectoral
vertebrae, the following three appear to be pectorals
also. The majority of the pectorals have some
restoration. However, the neural spines are relatively
well preserved and display a long rectangular shape,
the base of which is quite wide and gradually
narrows, becoming thinner towards the spinous
process. Pectoral ribs are only partly preserved and
are not fused with the transverse process. The
proximal portion of the shaft of the pectoral ribs
exhibits a caudally directed crest, which is positioned
well before the taper to the lateral shaft.
Dorsal Vertebrae: Vertebra 62 of the series is
likely the first dorsal, although the reconstruction
makes this determination difficult. The dorsal centra
are oval to fairly circular in shape. The difference in
shape from one dorsal to the next could be due to
slight flattening from taphonomic processes (Figure
2). Ten of the dorsals are identified by their
transverse processes and the separation of ribs.
These are mounted at the position above the pelvic
girdle, where the column corresponds with the ribs.
The dorsal vertebrae are relatively well preserved, far
more robust than the cervicals, and more complete.
The cross-sectional shape of the centra is unclear due
to some of the reconstruction. As there are some
breaks across the vertebrae, there is more
reconstruction of the neural arches and zygapophyses
of the dorsals and sacrals than on the pectorals. The

PALUDICOLA, VOL. 9, NO. 2, 2013 105
dorsal vertebrae are reconstructed with the fan-
shaped neural arches, as in the cast cervicals c16 to
c34/35.
Sacral Vertebrae: Vertebra 88 is likely the first
sacral vertebra. The following two vertebrae are
almost certainly sacrals. The sacral vertebrae have a
large, kidney-shaped rib facet, partially reconstructed
on the center of the lateral face of the centrum. The
dorsal and sacral centra appear to be sub-cylindrical
and slightly higher than they are long. Notably, both
the dorsal and sacral vertebrae were found
articulated. The neural spines, like their
predecessors, are partially rebuilt although
quadrangular shaped at the edge. Some of them are
associated with ribs.
Caudal Vertebrae: Almost every posterior
caudal vertebra is a cast, except for 6-7 partial. The
caudals consist solely of isolated centra with no
processes preserved. All of the chevrons are
reconstructed.
Ribs and Gastralia: Elements of the gastral
basket are reconstructed as the typical spindle shape
with a shallow longitudinal groove on the ventral
surface where the gastralia fit together. Preserved
ribs are relatively small fragments with the odd
medium-sized to large chunk, but most are missing
and replaced by the casts. The ribs that are preserved
are quite robust and convex anteriorly, as noted in
Libonectes (Buchy, 2005).
Pectoral Girdle—The pectoral girdle of WDC
CMC-01 is relatively complete with only a few parts
reconstructed (Figure 10). The entire girdle is 70 cm
long, 54cm wide at the lateral margin of the glenoid,
and 68 cm wide across the posterior cornuae. The
posterior edge of the right coracoid is largely rebuilt,
but this section is almost fully preserved in the left
coracoid, which displays a fan-shaped curvature to
the posterior part. The coracoid measures 56 cm
long by 33 cm wide. The most striking feature of the
coracoid is the unusually deep mesial notch or
intercoracoid vacuity. The notch is relatively long
(13cm) and narrow (4.5cm), and protrudes
anterolaterally about half way through the coracoid
(Figure 10). This area is missing in the right
coracoid and has been rebuilt to match the left. The
glenoids are relatively long, and robust, the center of
_______________________________________________________________________________________________________________
FIGURE 10. Complete pectoral girdle of WDC CMC-01 in ventral view. Note the unusual narrow notch of the coracoid (left set of bones). As
the photos were taken prior to the mount of the specimen (and we could not access them) the positioning of the scapulae and orientation could
not be changed. Light colored regions are reconstructed. Scale bar = 10 cm.

106 LOMAX AND WAHL—PLESIOSAUR ZARAFASAURA FROM MOROCCO
FIGURE 11. Right humerus of WDC CMC-01. A. Dorsal view. B.
Ventral view, note the interesting depression situated towards the
base of the shaft. Compare the size of the radius facet to the ulna
facet. Scale bar = 10 cm.
_____________________________________________________
which has a fairly strong ornamentation within the
cavity. Only a small section of the glenoids are
rebuilt. The pectoral bar of the coracoid forms a
triangular-shaped anterior process, although part of
the right pectoral bar is reconstructed. Both scapulae
are preserved. The left scapula was articulated with
the coracoid when found. The right scapula
measures 33 cm long by 26 cm wide; the left
measures 32 cm long by 27 cm wide, but they are
both partly reconstructed. The clavicles and
interclavicles appear to be missing but the facets are
present on the respective scapulae. The facets are
prominent in length, extending 13 cm from the
partial midline symphysis to the furthest anterior
edge of the scapulae, and the depth of the facet
ranges from 5 cm to a maximum of 7.5 cm. The
scapulae are not mounted with the left and right in
articulation with each other, and so the “keel” and
the corresponding angle of the elements could not be
determined. The pectoral fenestra, as formed by the
posterior edge of the scapulae and the anterior edge
of the coracoids, are oval in medial-lateral view.
Forefins—Both forefins are partially preserved.
The humerus of the right forefin measures 38 cm
long by 28 cm wide, and is nearly complete, with
restoration of the head only (Figure 11). The radius
is complete but only a fragment of the ulna is
preserved. The radiale is near complete and the
intermedium may partly be preserved. The rest of
the right forefin has been cast. As for the left forefin,
the humerus measures 38.5 cm long by 31 cm wide
and is complete, but only anterior fragments of both
the radius and ulna are preserved. The rest has been
cast. The propodials are very robust and short, and
display a very low length to width ratio, with a thick
shaft and an oval cross-section at the center of the
shaft. The heads of both humeri appear to have been
flattened taphonomically though the reconstruction
does not allow the extent to be noted. The proximal
shafts are thick with several prominent muscle
attachments. On the ventral side (in the center dorsal
surface) of the right humerus is a large sub-circular
depression measuring 10 cm by 7 cm. This
depression appears to penetrate several layers of
bone with a portion of the intact bone having been
pushed inward (Figure 11). This may be evidence of
predation, but no corresponding mark on the dorsal
side is present. Alternatively this may suggest a form
of differential distortion, though other large flat
bones are not damaged in the same way.
The distal ends of both humeri are very wide.
Both facets are preserved and the radius facet is
almost double the length of the ulna facet (17 cm and
9.5 cm, respectively), as seen in other elasmosaurids
such as Elasmosaurus platyurus and Styxosaurus
snowi (Welles, 1943). Compared to the ulna, the
facet for the radius is relatively shallow, though the
ulna facet bears a 'hook' at the posterior trailing edge.
The leading edge of the humerus and corresponding
radius, radiale and metacarpal appear to comprise a
thin edge, as seen in Styxosaurus (Welles and Bump,
1949). This could be a factor of the proportional
growth of the limb to the body, as noted in the
subadult elasmosaurid FFHM 1974-823 (Everhart,
2006).
Pelvic Girdle—WDC CMC-01 has an almost
complete pelvic girdle, with some sections partially
restored (Figure 12). The pelvic girdle is roughly 1/5
smaller than the pectoral girdle. The right side of the
pelvis is more complete than the left. The pubis is a
large flat quadrangular shaped bone. Both pubes are
partially reconstructed. The left pubis is largely
reconstructed at the distal end. The right pubis is
26cm long and about 33 cm wide, with the maximum
width at the anterior of 56 cm and the minimum
width at the posterior of 23cm. The midline
symphysis at the contact of the pubis is 14 cm long
and at the contact of the ischia the posterior
symphysis is 10cm long. However, as the pelvis
was mounted flat, these contacts may have produced
a “keel”. Both ischia are relatively complete. The
ischium is a rather flat bone, which is widely
expanded at the distal end, although the proximal end
is robust. The right ischium is estimated at 24 cm
long and 23 cm wide. Together the pubes and the
anterior process of the ischia form a sub-circular
pelvic fenestra. The pelvic fenestrae are partially
bordered by the pelvic bar of the ischia and otherwise
by the pubis and ischia respectively. The fenestrae
are 9.5 cm by 10 cm and the pelvic bar extends
anteriorly, separating the fenestra by 7cm from the
medial extension of the ischial “neck”. However, the
furthest anterior edges of the pelvic bar are
reconstructed. The ilia are long and thin bones that
are both estimated at 17 cm in length, at the
boundary of the ilial blade (where it contacts the
rebuilt sacral ribs) they expand from 4 cm to 6.5 cm
in width. Likewise both ilia are partly restored. The
acetabulum facet is compact and oval in lateral view,
5 cm high by 12 cm long.

PALUDICOLA, VOL. 9, NO. 2, 2013 107
FIGURE 12. Complete pelvic girdle of WDC CMC-01 preserved in dorsal aspect. The photographs were taken prior to the mount of the
specimen (and we could not access them), thus they are loosely arranged in rough relative position. Light colored regions are reconstructed.
Scale bar = 10 cm.
_________________________________________________________________________________________________________________
Hindfins—Both hindfins are partially
complete. The right femur measures 29 cm long by
18.5 cm wide and is nearly complete, with
restoration of the distal portion of the shaft. The
right femur appears bulbous and more pronounced
than the left, and may reflect a pathologic condition.
The tibia and fibula are both preserved and partially
reconstructed, as are the three mesopodials and some
of the metapodials. A total of 25 phalanges appear to
be original, with digit IV being the most complete.
The left femur measures 29 cm long by 19 cm
at its widest point, and is complete with a minimum
of restoration. The left tibia, fibula, mesopodials
and metapodials are well preserved, with little
restoration. As for the digits, each has some
complete and restored phalanges. Digit III is the
most complete with at least 8 original phalanges. A
total of 33 phalanges are complete to partially
restored (Figure 13). A phalangeal formula for either
hindfin cannot be determined as the hindfins were
largely disarticulated and partly scattered when
found.
The femora are smaller than the humeri, with a
much more rounded and thinner shaft, the head of
which is robust and elongate. The shaft quickly
expands into a fan-shaped posterior giving a
relatively wide distal end. Both facets of the tibia
and fibula are only slightly sub-equal in length. The
intermedium is relatively hexagonal in shape and the
fibulare more robust then the tibiale.
CONCLUSION
The description of the skull of Zarafasaura
(Vincent et al., 2011), in which the taxon was
established, has enabled this description of WDC
CMC-01. Vincent et al. (2011) described the
discovery of Zarafasaura as key to understanding the
expansion of Late Cretaceous plesiosaur
paleobiodiversity. As part of the general review of

108 LOMAX AND WAHL—PLESIOSAUR ZARAFASAURA FROM MOROCCO
FIGURE 13. Left hindfin of WDC CMC-01 in dorsal view. Note the relatively complete anterior elements including the femur. Light colored
regions of elements have been reconstructed. Stippled elements have been entirely reconstructed. Scale bar = 10 cm
____________________________________________________________________________________________________________________
marine reptiles from the Late Cretaceous Phosphate
beds of Jordan, Bardet and Pereda-Suberbiola (2002)
described isolated teeth and limb bones of
elasmosaurid plesiosaurs, but could not confidently
assign them to a taxon. In addition, as mentioned
previously, isolated material found in the vicinity of
the phosphate beds of Morocco were previously
considered to belong to Plesiosaurus mauritanicus,
although this taxon is now regarded as a nomen
dubium by Vincent et al. (2011). For this, WDC
CMC-01 may allow for a review and identification of
some of the isolated plesiosaur elements that have
been collected in the area, potentially recognizing
their taxonomic and scientific value (e.g. the material
described by Vincent et al., in press). Thus there lies
the prospect of identifying specific elements,
previously considered to be P. mauritanicus or
Plesiosauria indet., as Zarafasaura oceanis, helping
to develop a better understanding of the morphology
of the taxon. ACKNOWLEDGMENTS
We would like to thank the Wyoming Dinosaur
Center and Dr Burkhard Pohl for allowing access to
study this specimen. We also thank Dr Adam Stuart
Smith for offering fantastic advice in order to
progress this manuscript, and an anonymous
reviewer who offered good points which were
helpful in restructuring the paper. We also thank the
preparator of WDC CMC-01, who wished to remain
anonymous. He supplied key background
information which would otherwise have been lost.
Additionally we are grateful to Dr Nathalie Bardet
who supplied helpful comments for this study.
Finally we are very thankful to Reece Davies for
carefully creating, with guidance, the illustrations of
the reconstructed skull.
LITERATURE CITED
Arambourg, C. 1952. Les vertébrés fossiles des
gisements de phosphates (Maroc–Algérie–
Tunisie). Notes et Mémoires du Service
géologique du Maroc 92:1-372.
Bardet, N. and X. Pereda Suberbiola. 2002. Marine
reptiles from the Late Cretaceous
phosphates of Jordan. Palaeobio-
geographical implications. Geodiveritas
24:831-839.
Bardet, N., X. Pereda-Suberbiola, S. Jouve, E.
Bourdon, P. Vincent, A. Houssaye, J. C.
Rage, N. E. Jalil, B. Bouya, and M.
Amaghzaz, 2010. Reptilian assemblages
from the latest Cretaceous—Palaeogene
phosphates of Morocco: from Arambourg to
present time. Historical Biology 22:186-
199.
Buchy, M. C. 2005. An elasmosaur (Reptilia:
Sauropterygia) from the Turonian (Upper
Cretaceous) of Morocco. Carolinea 63:5-28.
Carpenter, K. 1999. Revision of North American
elasmosaurs from the Cretaceous of the
western interior. Paludicola 2:148-173.
Everhart, M. J. 2006. The occurrence of
elasmosaurids (Reptilia: Plesiosauria) in the
Niobrara Chalk of Western Kansas.
Paludicola 5:170-183.
Gasparini, Z., N. Bardet, J. E. Martin, and M.
Fernández, 2003. The elasmosaurid
plesiosaur Aristonectes Cabrera from the
latest Cretaceous of South America and
Antarctica. Journal of Vertebrate
Paleontology 23:104-115.
Massare, J. A. 1987. Tooth morphology and prey
preference of Mesozoic marine reptiles.
Journal of Vertebrate Paleontology 7:121-
137.
Smith, A. S. and P. Vincent. 2010. A new genus of
pliosaur (Reptilia: Sauropterygia) from the
Lower Jurassic of Holzmaden, Ger-
many. Palaeontology 53:1049-1063.
Vincent, P., N. Bardet, , X. Pereda Suberbiola, B.
Bouya, M. Amaghzaz, and S. Meslouh.
2011. Zarafasaura oceanis, a new
elasmosaurid (Reptilia: Sauropterygia) from
the Maastrichtian Phosphates of Morocco
and the palaeobiogeography of latest
Cretaceous plesiosaurs. Gondwana Research
19:1062-1073.

PALUDICOLA, VOL. 9, NO. 2, 2013 109
Vincent, P., N. Bardet, A. Houssaye, M. Amaghzaz,
and S. Meslouh. (In Press). New plesiosaur
specimens from the Maastrichtian
Phosphates of Morocco and their impli-
cations for the ecology of the latest
Cretaceous marine apex predators,
Gondwana Research
http://dx.doi.org/10.1016/j.gr.2012.11.011
Welles, S. P. 1943. Elasmosaurid plesiosaurs with a
description of the new material from
California and Colorado. University of
California Memoirs 13:125-254.
Welles, S. P. and Bump, J. 1949. Alzadasaurus
pembertoni, a new elasmosaur from the
Upper Cretaceous of South Dakota. Journal
of Paleontology 23:521-535.