Earth Sciences
Pterodactyloid pterosaur bones from Cretaceous deposits of the Antarctic Peninsula
1Laboratory of
Systematics and Taphonomy of Fossil Vertebrates, Departamento de
Geologia e Paleontologia, Museu Nacional/Universidade Federal do Rio de
Janeiro, Quinta da Boa Vista s/n, São Cristóvão, 20940-040 Rio de
Janeiro, RJ, Brazil
2Laboratório de
Paleontologia, Departamento de Ciências Biológicas, Centro de Ciências
Humanas e Naturais, Universidade Federal do Espírito Santo, Avenida
Fernando Ferrari, 514, 29075-910 Vitória, ES, Brazil
3Laboratório de
Paleontologia de Vertebrados e Comportamento Animal/LAPC, Centro de
Ciências Naturais e Humanas/CCNH, Universidade Federal do ABC, Rua São
Paulo, s/n, Jardim Antares, 09606-045 São Bernardo do Campo, SP, Brazil
4CENPALEO - Centro Paleontológico da Universidade do Contestado, Universidade do Contestado, 89306-076 Mafra, SC, Brazil
5Departamento de
Biologia, Universidade Federal do Espírito Santo, Alto Universitário,
s/n, Guararema, 29500-000 Alegre, ES, Brazil
6Laboratório de
Paleobiologia e Microestruturas, Centro Acadêmico de Vitória,
Universidade Federal de Pernambuco, Rua do Alto Reservatório, s/n, Bela
Vista, 55608-680 Vitória do Santo Antão, PE, Brazil
7Programa de
Pós-Graduação em Geociências, Universidade Federal de Pernambuco/UFPE,
Centro de Tecnologia e Geociências, Av. da Arquitetura, s/n, Cidade
Universitária, 50740-550 Recife, PE, Brazil
8Technical University of Munich, Chair of Soil Science, Emil-Ramann-Strasse 2, 85354, Freising-Weihenstephan, Germany
Fossil vertebrates from Antarctica are considerably rare, hampering
our understanding of the evolutionary history of the biota from that
continent. For several austral summers, the PALEOANTAR project has been
carrying out fieldwork in the Antarctic Peninsula in search for fossils,
particularly Cretaceous vertebrates. Among the specimens recovered so
far are two bones referable to Pterosauria, more specifically to the
Pterodacyloidea, the first volant reptiles from Antarctica to be fully
described. MN 7800-V (part and counterpart) was recovered from a moraine
at the Abernathy Flats (Santa Marta Formation, Lachman Crags Member,
Santonian-Campanian) on James Ross Island. It is interpreted as the
distal articulation of a first phalanx of the wing finger, representing
an animal with an estimated wingspan between 3 and 4 m. The second
specimen (MN 7801-V) comes from Vega Island (Snow Hill Island Formation,
Maastrichtian) and is identified as a wing metacarpal IV of an animal
with an estimated wingspan from 4 to 5 m. These occurrences show that
pterodactyloids inhabited the Antarctic Peninsula at least during the
Upper Cretaceous and demonstrate that large pterosaurs were widespread
through all parts of the planet during that period.
Key words Antarctica; Antarctic Peninsula; Pterosauria; PALEOANTAR; Cretaceous
INTRODUCTION
Pterosaurs were volant reptiles that existed for most of the Mesozoic
Era, and became extinct at the end of the Cretaceous (e.g.,
Kellner 2006,
Barrett
et al. 2008). Their record is quite irregular, with a few deposits
contributing to the majority of the recovered specimens and taxonomic
diversity (
Kellner 1994).
Until now, the only record of these flying reptiles reported from
Antarctica is a humerus that was recovered from the Early Jurassic Falla
Formation, in the Central Transantarctic Mountains (
Hammer
and Hickerson 1994). It was regarded as representing a
“rhamphorhynchoid” (= non-pterodactyloid) and tentatively referred to
the Dimorphodontidae (
Hammer and Hickerson 1996). This specimen was subsequently figured but without a detailed description (
Hammer and Hickerson 1999, fig. 5).
In 2006/2007, a team led by the Museu Nacional/UFRJ developed the
project PALEOANTAR aiming to search for fossils at distinct locations of
Northern Antarctic Peninsula (
Kellner
et al. 2011). All Brazilian scientific activities in Antarctica are
funded by a special program of the Conselho Nacional de Desenvolvimento
Científico e Tecnológico (CNPq) called PROANTAR (Programa Antártico
Brasileiro - Brazilian Antarctic Program) and are developed in
collaboration with the Brazilian Navy. Due to limitations of funding,
only in the austral summer of 2015/2016 could the PALEOANTAR project
resume, this time including several other institutions. During that
field season, several researchers worked collaboratively at James Ross
Island (the largest undertaking of PROANTAR), which included seven
researchers of PALEOANTAR, five researchers of TERRANTAR (examining
Antarctic soils), and three alpinists. Hundreds of fossils were
collected, the majority housed at the Museu Nacional/UFRJ, including
ichnofossils, several plant materials (e.g., logs, leaves, one pinecone,
charcoal), invertebrates (e.g., ammonites, bivalves, corals, lobsters),
teeth (sharks, bony fishes, plesiosaurs), and isolated bones (fish and
reptiles). Among the specimens recovered during that field season is one
bone divided into two parts (MN 7800-V) that is here referred to
Pterosauria. Two years later, in the austral summer of 2017/2018,
members of the PALEOANTAR project participated in field activities at
Vega Island under the TERRANTAR project. Although this and subsequent
expeditions to that island did not recover as many fossils as the
2015/2016 field season, there is one incomplete element that represents a
second pterosaur bone from the James Ross Archipelago (MN 7801-V). Both
specimens are described here, showing the presence of flying reptiles
in the Antarctic Peninsula during the late Cretaceous (
Fig. 1).
They are housed in the paleovertebrate collection (Departamento de
Geologia e Paleontologia) of the Museu Nacional (MN) - Universidade
Federal do Rio de Janeiro (UFRJ).
GEOLOGICAL SETTING
The area prospected by PALEOANTAR is situated in the eastern portion
of northern Antarctic Peninsula, more specifically at the northwestern
part of the James Ross Archipelago where outcrops of the Larsen Basin
are exposed (Fig. 1). The Larsen Basin is subdivi
ded into the James Ross and Larsen sub-basins (
Macdonald et al. 1988,
Del
Valle et al. 1992). The James Ross sub-basin was formed during the
breakup of Gondwana and has one of the most complete sedimentary
sequences from the Cretaceous to the Paleogene of the southern
hemisphere (e.g.,
Crame et al. 1996). These deposits are divided into three groups: Nodenskjöld, Gustav and Marambio (
Feldmann et al. 1993,
Hathway 2000,
Hathway and Riding 2001,
Riding and Crame 2002), the latter being of interest in the present study (
Crame et al. 2004, Fig. 2). Some new geologic data of the Snow Hill Island and the López de Bertodano formations are included here.
The Marambio Group crops out in several islands of the James Ross Archipelago, including James Ross and Vega islands (
Crame
et al. 1991). It comprises siliciclastic deposits, mainly sandstones,
siltstones and mudstones that are intercalated with levels of coquina
deposited under storm conditions on the internal and external platform
(Crame et al. 2004). Four formations are recognized in the Marambio
Group: Santa Marta, Snow Hill Island, López de Bertodano and Sobral (
Fig. 2).
The Santa Marta Formation is the basal unit of the group and thought
to have been deposited between the Santonian and Campanian (
Olivero et al. 1986,
Marenssi
et al. 2001; Fig. 2). Sandstones, siltstone, mudstone, volcanic tuffs
and rare coquinas are the predominant lithologies, regarded as
volcanoclastic deposits formed in a deltaic system (
Scasso et al. 1991,
Olivero 2012). In the northwestern portion of the James Ross Island,
the Santa Marta Formation is divided into the Lachman Crags and the
Herbert Sound members (Crame et al. 1991).
The Lachman Crags Member (Santonian-Campanian) is also composed of
siliciclastic rocks with rare conglomerates, reaching a thickness of
about 850 m (Crame et al. 1991). It is overlain by the Herbert Sound
Member (Campanian-Maastrichtian), which consists of a 250 m section
composed by fine sandstones with cross-stratification and coquinas
(Crame et al. 1991,
Olivero et al. 1992,
Olivero and Medina 2000).
The Snow Hill Island Formation (Maastrichtian) shows massive
sandstones with parallel bedding and cross lamination on its base.
Occasionally thin layers of dark and red bioturbated shales are found at
a decimetric scale associated with layers with calcium carbonate cement
that are up to 30 cm thick. Carbonate nodules are very common and occur
in two levels at the medium portion of the section where invertebrate
fossils are concentrated. According to Olivero (2012), these deposits
were formed in a progradation deltaic system. Coarse to medium grained
sandstones occur on the top of the sequence, followed by bioturbated
sandstones and siltsones, also containing levels with fossiliferous
concretions, which are interpreted as tidal deposits formed in a
regressive environment (
Pirrie et al. 1997).
The López de Bertodano Formation (Maastrichtian) crops out in the
elevated areas of the Vega, Snow Hill and Seymour islands (Crame et al.
2004). On Vega Island, its contact with the underlying Snow Hill Island
Formation is gradational and mostly covered by soil. It comprises fine
to coarse grained massive sandstones, with dispersed rounded pebbles.
The sandstones reach a thickness up to 3 m, with a thin layer (~20 cm)
of carbonaceous mudstone at their top. On the lower portion of this
unit, close to the contact with the Snow Hill Island Formation, fossils
were collected by the project TERRANTAR in the field season of
2017/2018. This area, revisited by the PALEOANTAR project in 2018/2019,
is characterized by alternating bands of fine sandstones and friable
mudstones with wavy lamination. They grade to lenticular deposits of
different scales and represent tidal channels within a tidal-dominated
estuarine environment (
Olivero et al. 2007). This unit preserved the contact between the Late Cretaceous and the Paleogene (
Milanese et al. 2019).
The last unit of the Marambio Group is the Sobral Formation
(Paleocene) that crops out mostly at Seymour Island. It predominantly
comprises fine sandstones and mudstones that were deposited during two
transgressive-regressive cycles (Marenssi et al. 2012).
Both specimens studied here (MN 7800-V and MN 7801-V) were not collected
in situ,
but were found eroding on the surface, a common feature of fossil bones
in the James Ross Archipelago. The specimen MN 7800-V is formed by two
parts regarded as belonging to the same bone (see Description) and was
found at the Abernathy Flats (James Ross Island) by one of the authors
(CWM) while collecting soil in the frame of the TERRANTAR project. He
informed the occurrence to AWAK and LCW who collected the specimen. The
Abernathy Flats consists of an extended flattened area between the Crame
Col and the Paso San José, in front of the Brandy Bay. The area where
the specimen was found (S63º53´35,9” / W57º55´47,3”) is a moraine
consisting of unlithified sediments of glacial origin with no outcrops
nearby (
Fig. 3).
According to geologic maps, the rocks in that area belong to the
Lachman Crags Member of the Santa Marta Formation (Fig. 2). It came from
the middle portion of this unit, which suggests a Santonian-Campanian
age. No other fossils were found in a radius of ~200 m.
The second specimen (MN 7801-V) was also the result of surface
collecting, this time at Vega Island. It was found during field activity
of the TERRANTAR project at the Cape Lamb area in 2017/2018, on the
beach (Thales Nunes da Silva, personal communication). Outcrops of two
main stratigraphic units occur in this area: the Snow Hill Island and
the López de Bertodano formations, both Maastrichtian in age (Olivero et
al. 1986, Marenssi et al. 2001). Field activity of the PALEOANTAR
project in the austral summer of 2018/2019 established that most of the
fossiliferous layers in the Cape Lamb area are located close to the
contact between the Snow Hill and the López de Bertodano formations
(S63º51´53.0” x W57º34´20.0”). Based on the available geologic map, the
region close to the beach are part of the Snow Hill Island formation,
from where MN 7801-V most likely came.
MATERIALS AND METHODS
A portion of the counter part of the specimen MN 7800-V was sampled
for paleohistological analysis. While in the field, this bone was
covered by thermoplastic resin (Paraloid B72) for protection purposes.
The specimen was subsequently measured and photographed according to the
protocol proposed by
Lamm (2013) and the thin section was prepared using standard fossil histology techniques (
Chinsamy
and Raath 1992, Lamm 2013). The sample was embedded in clear epoxy
resin RESAPOL T-208, catalyzed with BUTANOX M50, and cut with a
diamond-tipped blade on a saw (multiple brands). The mounting-side of
the section was wet-grounded using a metallographic polishing machine
(AROPOL-E, Arotec LTDA) with Arotec abrasive papers of increasing grit
size (60/P60, 120/P120, 320/P400, 1200/P2500) until a final thickness of
30-60 microns was reached.
The osteohistological structures were observed with an optical
microscope in transmitted light mode. Parallel/crossed nicols and
fluorescence filters were used to enhance birefringence. Histological
images were taken using an AxioCam digital sight camera (Zeiss Inc.,
Barcelona, Spain) mounted to an Axio Imager.M2 transmitted light
microscope (Zeiss Inc. Barcelona, Spain). Images were taken at 56x and
106x total magnification.
DESCRIPTION
The specimen MN 7800-V from James Ross Island consists of two
elements that were found close to each other in unlithified sediments
with several large pebbles of different lithologies (Figs. 3,
4). One portion is preserved in three dimensions, an uncommon condition in pterosaurs (e.g.,
Bantim et al. 2014) whose bones are generally found very flattened (e.g.,
Cheng
et al. 2017). The second part consists of a thin surface of bone
preserved in a dark grey matrix formed by sandstone. Both pieces are
about the same size and although the contact surface does not show a
100% fit because bone is missing, we regard them as representing the
same element based on the overall morphology and the particular
conditions where they were found.
The main preserved portion of MN 7800-V has a maximum length of 67.8 mm (
Figs. 5a, b).
Several parts of the external bone surface were lost, but where
measurable, the bone cortex is thin (~1 - 1.5 mm), thus congruent to
pterosaur bones. It comprises a short shaft that expands forming a
boot-shaped articulation, which is typical of the distal end of
pterosaur wing phalanges (e.g.,
Wellnhofer 1991a). The maximum width (mediolaterally) of the articulation is 52.1 X 25.4 mm.
A slice of the counterpart of MN 7800-V was taken for histological
analysis (Fig. 5c). The osteohistological features are consistent with
those observed in pterosaurs (see Steel 2008 for a review). The bone
cortex is thin and presents a fibrolamellar tissue type (Fig. 5d). The
vascular canals are longitudinal and have similar diameters, with no
sign of anastomoses and composed only by primary osteons. As far as
observations are possible, the endosteal cavity does not present
internal lamellae. A film of thermoplastic resin applied on the specimen
in the field, hinders the visualization of this part of the bone. A
thin layer of avascular periosteal lamellae covers the outer surface. No
growth marks such as zones, annuli or lines of arrested growth were
observed. These structures are rarely preserved in pterosaurs due to the
extensive endosteal resorption for the maintenance of the medullary
cavity (
Steel 2008,
Prondvai
et al. 2012), preserving only the most recent histological events. No
resorption zones or secondary osteons were observed, which indicates no
active remodeling in this individual at time of death.
The presence of avascular periosteal lamellae (External Fundamental
System - EFS) has been used before to infer somatic maturity in sampled
individuals, which indicates the ending of the development of the
periosteal circumference (
Cormack 1987,
Starck and Chinsamy 2002,
Ponton et al. 2004). However, bones can still become more robust (
Woodward et al. 2011,
Kellner et al. 2013,
Lee and O’Connor 2013,
Andrade
et al. 2015). The EFS has already been recorded in mammals (e.g.
Cormack 1987), birds (e.g. Ponton et al. 2004), non-avian dinosaurs (
Erickson et al. 2004,
Horner and
Padian
2004, Padian et al. 2004), and crocodylomorphs (Woodward et al. 2011,
Andrade et al. 2015). In pterosaurs, the EFS is so far restricted to a
few specimens (
Sayão 2003, Steel 2008,
Kellner et al. 2013), being rare in non-pterodactyloid taxa with only two records (
Gross
1934, De Ricqlès et al. 2000) and, so far, absent in many derived
pterodactyloid groups such as azhdarchids. In all known occurrences of
the EFS, the cortex is marked by the presence of LAGs, which were not
seen in MN 7800-V, a feature that should be investigated further.
Despite the absence of other microstructural features, we consider that
the specimen was an adult that has or was about to complete its somatic
growth.
Comparisons with osteohistological sections of other tetrapods
reported from the same deposit, such as ornithopod dinosaurs and
plesiosaurs, show important differences to MN 7800-V. Ornithopods have a
more robust cortex, with weakly woven to parallel-fibered bone
composing the primary bone and an outer cortex almost exclusively
lamellar (
Horner et al. 2009,
Werning 2012).
This clade also shows an inner cortex heavily remodeled by secondary
osteons, extending into the mid-cortex with the total absence of
original primary remains (Werning 2012). These osteohistological
features differ from the fibrolamellar bone presented by MN 7800-V that
also lacks a remodeled cortex. Another difference is the presence of a
cyclic growth in ornithopods characterized by multiple LAGs.
Plesiosaur osteohistology is very peculiar with bone matrix filling
almost completely the medullar region, composing a dense
osteosclerotic-like skeleton (
Salgado et al. 2007,
Ossa-Fuentes 2017). This differs remarkably from the condition of MN 7800-V that presents a thinner cortex with a free medullary cavity.
The second specimen (MN 7801-V) described here comes from the Vega Island (
Fig. 6). It consists of a long bone with a preserved length of 169.2 mm and lacks any articulation (
Fig. 7).
Originally, it was collected in one piece, but was further broken into
three parts due to the fire of the Museu Nacional (Kellner 2019). MN
7801-V is a naturally compressed element, with one side flattened (Fig.
7a), the other convex (Fig. 7c), and rounded margins. The shaft is
slightly curved (Figs. 7b, d) and the transverse section at the middle
portion is D-shaped (37 x 22 mm, Fig. 7e). The external bone surface is
reasonably preserved and shows a very thin cortex (~ 0.8 - 1.2 mm),
tending to be thicker at the edges, a common condition in pterosaurs
(Sayão 2003).
DISCUSSION
Despite the efforts of several paleontological research groups (e.g.,
Molnar et al. 1996,
Case et al. 2000,
Novas et al. 2002,
Salgado and Gasparini 2006,
Reguero and Gasparini 2007,
Reguero et al. 2013, Kellner et al. 2011,
Cerda et al. 2012,
Coria et al. 2013,
Roberts et al. 2014,
Lamanna
et al. 2019), our knowledge on the Antarctic vertebrate fauna is still
in its infancy. Due to the harsh conditions that predominate throughout
the year, fieldwork is confined to the austral summer, when working
conditions are more suitable but still very challenging. When found, the
vast majority of fossil vertebrates is very incomplete, fragmentary and
in many instances difficult to interpret. This is also the case off the
material from James Ross and Vega islands. Unfortunately, no bone beds
have been identified in both islands.
The specimen MN 7800-V from James Ross Island is a good example of
the fragmentary nature of vertebrate fossils from Antarctica. It was
recovered from a moraine, quite an unlikely place for any fossil bone to
be found, especially that of a pterosaur. Notwithstanding, based on the
overall expanded boot-shaped articulation and the thickness of the
cortex, we interpret this specimen as belonging to the distal end of a
pterosaur wing phalanx (
Fig. 8).
The osteohistological features are consistent with this interpretation.
Although the boot-shaped expansion is a feature common to all distal
ends of pterosaur wing finger phalanges (e.g., Wellnhofer 1991a, b,
Kellner and Tomida 2000), we regard MN 7800-V as belonging to the first
wing finger phalanx mainly due to size. If this assignment is correct,
when compared to other specimens the wingspan of the James Ross
pterosaur would range between 3 and 4 m. Such large wingspans are not
seen in non-pterodactyloid pterosaurs, but are quite common within the
Pterodactyloidea (e.g., Wellnhofer 1978, 1991a). Among pterodactyloids
there are several clades represented by large flying reptiles (e.g.,
Kellner 2003), with the Pteranodontoidea or the Azhdarchoidea the best
candidates to which MN 7800-V might be classified. The age of the main
deposits of the Lachman Crags Member (Santonian-Campanian) is consistent
with this suggestion.
The identification of MN 7801-V from Vega Island as a pterosaur is
less problematic. Despite the lack of articulations, the bone-surface is
better preserved in this specimen, with a very thin bone cortex that is
an uncontroversial pterosaurian feature. Its large size indicates that
it is referable to the Pterodactyloidea as well. Based on size, we also
preclude its referral to the Archaeopterodactyloidea, which comprises
much smaller animals (Kellner 2003).
Regarding which part of the skeleton MN 7801-V represents, there are
not many options to be considered. It is a long bone whose size and
overall morphology leads to the exclusion of the hindlimb. Although size
does not rule out that it could represent a phalanx from the wing
finger, these bones tend to show a triangular to subtriangular
transverse section (e.g.,
Wellnhofer 1978,
1991b) or are T-shaped (
Martill and Frey 1998,
Averianov
2014), differing from the D-shaped condition observed in MN 7801-V
(Fig. 7e). The pterosaur ulnae are also quite long and somewhat
anteroposteriorly compressed (e.g., Wellnhofer 1978, 1985, 1991b,
Kellner and Tomida 2000,
Bennett
2001), but their transverse section is more oval or rounded, differing
from the Vega Island specimen. The metacarpal IV, however, is elongated
and several of these elements show a D-shaped cross-section (e.g.,
Wellnhofer 1985, figure 15), quite similar to MN 7801-V. Therefore, we
interpret this bone as being a wing metacarpal, possibly from the right
side (Fig. 8).
The size of the Vega Island specimen is also consistent with either a
pterosaur from the Pteranodontoidea or Azhdarchoidea clades, which have
quite distinctive wing configurations (see Kellner 2003). The age of
the Snow Hill Island Formation, however, is Maastrichtian, when the most
predominant flying reptiles were azhdarchoids, for all Azhdarchidae
(e.g., Witton and Naish 2008, Averianov 2014,
Longrich et al. 2018). Azhdarchoids have proven to be quite diverse in size (e.g.,
Martin-Silverstone et al. 2016) and ecological niche, with potentially different feeding strategies (e.g.,
Langston 1981,
Witton and Naish 2008, Averianov 2014,
Kellner and Calvo 2017,
Bestwick
et al. 2018, Kellner et al. 2019), and a cosmopolitan distribution.
Although at this point no further taxonomic decision on MN 7801-V can be
made, it can be said to represent a large pterosaur with an estimated
wingspan between 4-5 m based on comparisons with other pterodactyloids
(e.g., Wellnhofer 1991a, b, Kellner and Tomida 2000, Bennett 2001,
Kellner 2003).
The sole other pterosaur specimen reported from Antarctica is a bone
identified as a humerus that was briefly mentioned (Hammer and Hickerson
1994, 1996) and subsequently figured (Hammer and Hickerson 1999, fig.
5), but no anatomical description was presented so far. The published
picture is difficult to interpret. If this specimen is complete and
indeed represents a pterosaur humerus, this bone is shown from the
lateral rather than from the dorsal view. The proximal articulation is
expanded and the portion of the bone that might have been interpreted as
the deltopectoral crest is triangular. Based on the anatomy of
Dimorphodon humeri (e.g., Wellnhofer 1978,
Padian 1983,
Unwin 1988), the shaft is curved, differing from the straight bone reported from Antarctica. In
Dimorphodon,
the deltopectoral crest is longer, does not extend as deep on the shaft
and is not triangular. If the reported bone from Antarctica is indeed
pterosaurian, it differs from
Dimorphodon and, based on the available information, cannot be assigned to any specific pterosaur clade.
CONCLUSIONS
Although very fragmentary, like most vertebrate fossils recovered
from Antarctica, the specimens MN 7800-V and MN 7801-V indicate that
large pterodactyloid pterosaurs were present in the Antarctic Peninsula
region at least from the Santonian to the end of the Cretaceous. To our
knowledge, these specimens are also the first from Antarctica to be
fully described, demonstrating that at least during the Late Cretaceous,
large flying reptiles could reach all parts of the planet. Continuous
fieldwork in this region will certainly reveal additional and
better-preserved material of these flying reptiles, filling in one of
the gaps of our knowledge of their evolutionary history.
On a further note, we would like to point out that MN 7801-V is the
first specimen recovered from the debris of the Museu Nacional after the
fire that affected this institution in 2018 (e.g., Kellner 2019) to be
studied. Several important specimens were recovered so far, and we are
sure that more will follow during the recovery and rebuilding activities
of this worldwide important institution.
