Taphonomy of two Holocene penguin taphocoenoses in Potter Peninsula, South Shetland Islands, Antarctica

ABSTRACT Abandoned rookeries belonging to any of the Pygoscelis species are very particular taphocoenosis, usually preserved due to the accumulations of bones, pebbles, and guano year after year in the same nesting area, that condense together with other biogenic material and form ornithogenic sediments. The results of the excavation, sieved and analyses of sediment and materials found in two rookeries formed during the last 8000 years, named Pingfo I and Pingfo II and found in the 25 de Mayo/King George Island (South Shetland Islands, West Antarctica) are presented here. The analyses of the bones and eggshells, taxonomic and ontogenetic composition, elements representation, transport degree, predation, and scavenging marks, together with the micropalaeontological content provided informative tools for the reconstruction of the nesting areas, suggesting that bones were transported and accumulated in Pingfo I that would represent high energy beach, whereas a breeding colony was settled in Pingfo II. The preservation of hexactinellid spicules and several arthropod remains are particularly informative for the levels at Pingfo I. The preservation of a Culicidae (Diptera) wing could be the first middle Holocene record of the group in Antarctica, the only continent without extant mosquitoes. Graphical Abstract


Introduction
Penguins (Aves, Sphenisciformes) are philopatric birds that normally nest in colonies close to the beachline (Ainley et al. 1995;Williams 1995), on shoreline ice, rocks or even on rocky hillsides.They usually return to reproduce in their natal colony and occupy the same place year after year, maintaining for several seasons the same partner.Of the 18 living species (Winkler et al. 2020), five reproduce in Antarctica, including the three pygoscelid penguins, which constitute 70% of the total bird Antarctic biomass (Everett 1976).The three species Pygoscelis adeliae Hombron and Jacquinot, 1841 (Adelie penguin), P. papua (Forster, 1781) (Gentoo penguin), and P. antarcticus (Forster, 1781) (Chinstrap penguin) live in sympatry constituting large and mixed circumpolar colonies in Antarctica and subantarctic islands (Williams 1995).
The settlement of 'Pygoscelid penguins' called rookeries requires ice-free terrain for nesting along beachlines and has been documented for thousands of years in the Antarctic continent and the Sub-Antarctic Islands.These rookeries are part of a rich biocoenosis, in which also participate other seabirds (e.g.skuas, giant petrels, wandering albatross), marine mammals (e.g.crabeater seals, elephant seals, whales), fishes, crustaceans, annelids, insects, hairgrass, moss and algae, besides penguins.Evidence of ancient nests from the Holocene have been found in areas where nesting colonies continue to settle, but also in completely abandoned areas.
The Adelie penguin, the most abundant species in Antarctica, is the main rookeries builder (Emslie et al. 2014), given their habit of collecting pebbles to build or rebuild nests.The Gentoo and Chinstrap penguins were also identified as rookeries builders in abandoned colonies, but in a secondary proportion compared with the Adelie penguins.This could be due to a late colonisation of the areas by these species or to a bias in the record, given that Gentoo and Chinstrap penguins nest in steep areas where sediment accumulates less abundantly (Emslie et al. 2014).
A higher amount of organic matter is accumulated in the rookeries due to the daily activity of the penguins.The influx of nutrients received mainly from penguin guano and excreta, carcases of dead penguins and food debris among other organic remains, locally modified the ice-free soil.Consequently, phosphatic soils specific for these Antarctica and Sub-Antarctic regions are formed (see for example Tatur 1989 and an extensive list of references in: Emslie et al. 2014).Differences in the chemical composition of these soils depend on the geography and the distance of the rookery to the coast, the climate, the substrate preferences of the species, the topography, and the distance to the colony.The term 'ornithogenic soils' was proposed by Syroechkovsky (1959) to designate organic soils modified by the activity of bird colonies on the continental Antarctic, where the low temperature caused no chemical reactions between guano and the underlying rocks (Campbell and Claridge 1966;Ugolini 1972).The ornithogenic soils are formed in situ and around the colony.In the South Shetland Islands and many other places of West Antarctica, these soils are compositionally different due to the constant wash by the marine waters and the less severe climatic conditions.This difference in the soil formed by birds in marine and continental Antarctica has led to different interpretations of the term 'ornithogenic soils'.In this sense, Everett (1976) for example, restricted it to those areas devoid of organic matter other than bird droppings, that in West Antarctica would only include active rookeries.On the other hand, a more inclusive use of this term refers to any soil under the impact of solutions coming from the colony, including abandoned colonies and vast surfaces of soils surrounding penguin colonies (Tatur and Myrcha 1984).To overcome this problem, another distinction has been proposed, naming ornithogenic soils to the bird-formed soils characteristic of penguin colonies, and ornithogenic sediments to the sediments accumulated in nearby neighbouring areas (lakes, topographically low sites) that preserve biological and geochemical evidence of ancient breeding colonies (Emslie et al. 2014).In this contribution, we fully adhere to the latter proposal which clearly explains the results obtained here regarding the genesis of the sequences excavated.
The objectives of this work are focused on the study of two sites known as Pingfo I and Pingfo II, located in the Potter Peninsula (Isla 25 de Mayo/King George Island), South Shetland Islands (Figure 1).The study focused on a taxonomic, ontogenetic, and taphonomic analysis of the recovered remains emphasising a strong stratigraphic control of the taphocoenosis.The excavated sites were screened, allowing the analysis of microscopic organisms.The new specimens collected (including penguins, other vertebrates, invertebrates, and plants) were systematically and anatomically determined.Penguin bones were analysed considering the parts and ages representation, as indicators of the transportation degree of the bodies and the closeness to the reproductive colony area.The information provided here also constitute an important input for the reconstruction of the coastline, and the restriction of the deglaciation age of the coastal areas in the Potter Peninsula.

The study area
The geological sequence is mainly composed of Palaeogene levels corresponding to a stratiform volcanic succession of basaltic-andesite lavas and tuffs, rocks that also dominate the geological composition of the Potter Peninsula (Figure 1), King George Island (Figure 1B), South Shetlands Islands (Figure 1A) in the Antarctic Peninsula (Del Valle et al. 2002;Figure 1C), to which is added Quaternary neoglacial moraines and marine sediments (Birkenmajer 1998a, b).
The ice volume retractions and regional isostatic uplift of lands during the Holocene allowed the formation of raised marine beaches on coastal areas of the peninsula (Del Valle et al. 2002, 2007).Penguins that already lived on high cliffs by those times, descended towards the lower coastal areas that emerged progressively (Tatur et al. 1997).Raised marine beach deposits elevated above sea level, preserved ancient assemblages of bones and other biological remains, dominated by penguin remains.The analysis of new subfossil specimens recovered around a penguin colony named Pingfo I (Figure 1D,) and in the abandoned Pingfo II (Figure 1D,), located in the Potter Peninsula (Figure 1A-D) are the focus of the present study.
The locality 'Pingüi' could not be relocated during our fieldwork in 2016-17 following the coordinates (62°15ʹ00"S, 58°36ʹ48") given in Montalti et al. (2009) since they are placed in the Warszawa ice field.It seems more probable that there is a mistake in the GPS data since the known retreatment experimented by the Warszawa ice field (see Lagger et al. 2018).
Pingfo I (62°15ʹ26.483"S,58°37ʹ08.530"W)or 'Pingfo' in Del Valle et al. (2002) is a site corresponding to a marine beach elevated 17.3 m.a.s.l. and located on the south-eastern coast of Potter Peninsula, precisely at the north-western of Stranger Point (Figure 1B,).Pingfo I is in the ASPA N° 132 given that the coastal areas, made of raised beaches, host important bird colonies, marine mammal breeding areas, and diverse vegetal species (Figure 1D,).
Previous dates of the penguin bones of the basal level 1 of Emslie et al. (2020), estimate the maximum age of Pingfo I in 5925-5665 and 7425-7215 cal BP based on a humerus and tibiotarsus, respectively (Emslie et al. 2020).Level 3 was estimated in 5615-5315 cal BP using a penguin bone (Del Valle et al. 2002), in 5850-5590 cal BP using a proximal end of humerus, and in 5030-4715 cal BP using a feather calamus (Emslie et al. 2020).Level 4 was dated in 5550-5265 and 5555-5305 cal BP based on two penguin bones (Emslie et al. 2020), and finally, level 5 was estimated in 5560-5280 cal BP using a penguin bone (Del Valle et al. 2002).
Materials from these levels were previously reported by Del Valle et al. (2002).From these remains collected in Pingfo I, the penguin assemblage was examined more in detail in Montalti et al. (2009) and sub-fossil bones of skuas were studied in Acosta Hospitaleche et al. (2010).Other important contributions about abandoned rookeries, including Pingfo I and other recently discovered mounds, can be consulted in Emslie et al. (2020).
Pingfo II (62°14 20.73"S, 58°40 ́21.00"O) is an ascended marine beach located 2,17-2,77 m.a.s.l. on the Southern coast of the Potter Peninsula (Figure 1B,), discovered some years ago by Del Valle et al. (2007).The site is between the scientific station Alejandro Carlini and Mirounga Point (Figure 1D,) and is frequented by mammals and birds nesting in the area.Unlike in Pingfo I, no penguin breeding colonies are settled in the surrounding area.
Previous relevant contributions include the dates of penguin bones collected in level 3, whose results show estimated ages around 7.562 cal BP and 7.414 cal BP (Del Valle et al. 2007), and the preliminary analysis of the first bones collected there (Montalti et al. 2009)

Materials
The materials, collected in the field and separated from the sediment in the laboratory, are housed in the Antarctic Repository of Palaeontologic and Geologic collections of the Argentine Antarctic Institute (IAA).Comparative skeletons belong to the Ornithological Section of the Vertebrate Zoology Department and the Vertebrate Palaeontology Division of La Plata Museum (Argentina).

Collection methods
The stratigraphic profiles of Pingfo I (Figure 2) and Pingfo II (Figure 3) were raised, and a first collection of the material was made on the surface and the exposed face of each level.After that, grids of one square metre were excavated in each profile to access the lower levels and systematically sample them.The degree of consolidation and the lithology of Pingfo I and Pingfo II determined different techniques in each case.When possible, the sediment was separated in the field with sieves of different mesh sizes (i.e. 3 mm, 1.5 mm, and 0.5 mm) obtaining granulometrically selected material that was later examined in the laboratory under a binocular microscope.Besides, consolidated blocks were extracted from levels 2 and 6 of Pingfo I for disaggregation in the laboratory in the search for organic material and geological proxies.
Most of the macroscopic materials of Pingfo II, identified as birds and mammal bones, were found spread on the surface and labelled as remains from superficial deposits.A few other specimens from the deeper levels were collected, and the sieved sediment was transported to the laboratory for examination under magnification.Bivalves in life position were found in situ in some levels of Pingfo II.In those cases, the depositional conditions were registered in our field notebook and are provided in the individual information.

Pingfo I
The sediment of level 1 was divided into fractions using a sieve with 1 mm 2 mesh.The coarse fraction was examined by naked eye, separating bony remains and seaweed packs.The consolidated samples of level 1 were divided into three and each one submerged in: (A) 1.5% H 2 O 2 at Normal Pressure and Temperature Conditions (CNPT), (B) 3% H 2 O 2 at CNPT, and (C) distilled water at CNPT; to prepare samples to search different kinds of microscopic sub-fossils.Each sample after being strained and dried was examined under magnification.Then, sample (A) was heated at 100°C in 3% H 2 O 2 for 20 minutes to disaggregate the still consolidated fractions.This sediment experienced visible changes allowing the better separation of calcareous remains under a binocular microscope.
Consolidated sediments of level 2 were immersed in distilled water for a complete week and then separated by fractions with two sieves of 0.10 and 0.15 mm 2 meshes.The coarse fraction was heated at 100°C in 3% H 2 O 2 for 40 minutes, and, when cooling, the sediment was filtered with a 100 µm sieve.The dried solid residual sediment was examined under a binocular microscope Arcano ZTX Zoom (10-40X).
The sieved sediments of levels 3-7 were examined with naked eye and under magnification.Besides, a sample of 25 ml of each level was separated, heated at 100°C in 3% H 2 O 2 for 20 minutes, and filtered with a 100 µm sieve.The finest fraction was observed under the binocular microscope to separate microscopic structures by picking.Other consolidated blocks (from levels 2 and 7 mainly) were mechanically disaggregated with a variable speed rotary Dremel tool and the small portions were observed under magnification without any chemical treatments.

Pingfo II
The sediment samples sieved and granulometrically selected were observed directly under a binocular microscope.Chemical treatment was not required given that the entire sample was unconsolidated and clean.

Analysis of the material
Avian anatomy terms follow Baumel and Witmer (1993).The ontogenetic stage of each remain was assigned according to ossification and fusion degree plus the textural ageing.Bone specimens were divided in four main categories: small chicks (with a bone surface characterised by a fibrous texture, and covered with furrows and dimples as a consequence of the incomplete ossification), large chicks (with thinner furrows covering the bone surface, and elements still unfused -shaft and ends in long bones, synsacral elements, pygostyles, metatarsals, etc.-), juveniles (almost reaching the adult size, but with a rough texture on the bone surfaces, and the sutures still visible in some cases), and adults (with bones completely fused, without visible sutures, and a silky texture on the surface) (Acosta Hospitaleche et al. 2017; Acosta Hospitaleche and Picasso 2020).
Bones were also grouped according to the skeleton element to calculate the parts representation (Cruz and Savanti 1999;Cruz 2003).The widely used approach of Behrensmeyer (1978) to evaluate weathering damage is here discarded in favour of another recent categorisation, based on seabird skeletons that evaluate similarly the damage caused on the bones subaerially exposed on cold beaches (Muñoz and Savanti 1998).A classic taphonomic approach (see Lawrence 1968; Seilacher 1973 for example) is adopted for the analysis of the general biostratinomic processes.The eggshells were examined under a binocular microscope and directly compared with modern eggshells of the Antarctic Pygoscelis species.Seaweeds were identified in the laboratory of the Argentine Antarctic Institute (IAA).
For the discussion, we included data from vertebrates (penguins and mammals), feathers, and eggshells previously collected by members of the IAA, which are currently housed at the IAA and MLP (Del Valle et al. 2002, 2007;Montalti et al. 2009;Acosta Hospitaleche et al. 2010;Emslie et al. 2020).Unfortunately, repository numbers to identify each remain are not provided in Del Valle et al. (2002) and Emslie et al. (2020).

Pingfo I
The profile is exposed at the western limit of a north-east terrace of 9 metres long, in which the eroded layers can be followed in three blocks partially displaced from each other (Figure 1D,).Seven levels, named 1 to 7 from the bottom to the top (partially equivalent to 0-6 in Del Valle et al. 2002) are recognised based on the sedimentological and palaeontological content (Figure 2).There are some differences between the thickness of the levels here recognised and those previously reported (Valle et al. 2002;Emslie et al. 2020; and see Figure 4).These subtle discrepancies are probably due to the partial displacement of each of the three blocks and the consequent recognition of the top and bottom of each bed, are subtle.Each bed was delimited following a sedimentological criterion and measured in north-south direction.In the sequence described by Del Valle et al. (2002), seven levels have been described, including a cero level or bedrock of basaltic andesitic lavas.Emslie et al. (2020) identified five beds for the same locality, since levels 1 and 2 of the former were considered as part of the same unit.The differences in levels regarding the present study rest in our recognition of a more detailed subdivision of level 5 of Del Valle et al. (2002) or 4 of Emslie et al. (2020), performed for practical reasons during sampling (see below).Despite these minor differences, there is a clear correlation between the three profiles (Figure 4).
The basal level 1 is around 0.40 metres thick and composed of sands with large and angular basaltic-andesitic clasts which are comparable to the Palaeogene volcanic rocks cropping out in nearby hills on the peninsula.The overlying level 2 is exposed in two of the blocks that complete the sequence and consists of sands and a few pebbles.In the southernmost block, level 2 is together with level 1 and reaches 0.40 metres thick, whereas in the block, level 2 is under level 3 and develops only 0.30 metres thick.Level 3 is 0.46 metres thick and is composed of sands and large and angular clasts.This level is partially exposed under level 3 but continues also in the third laterally displaced block constituting a layer of 0.76 metres thick together with part of level 2. Level 4 reaches 0.20 metres thick and is composed of fine and dark sands.Level 5 is also 0.20 metres thick and constituted by large and angular clasts with bones.The overlying level 6 is only 0.16 metres thick and consists of little or no consolidated sands with small clasts and bones.Level 7, at the top, consists of 0.60 metres thick of consolidated coarse sands with poorly selected clasts from angular to rounded and a mud-rich matrix.Penguins are represented through complete bones, well preserved but fragile.This is the most surficial unit and therefore the most exposed to subaerial erosion.

Materials from Pingfo I
In contrast to previous works where only macrofossil remains were collected and examined (Del Valle et al. 2002;Montalti et al. 2009;Emslie et al. 2020), the sieve of sediments in each level, allowed to recover a collection with less size bias, including small bone elements but also specimens of invertebrates not previously recorded.As a result, a valuable collection of 474 bones assigned to penguins (Pygoscelis spp.) and elephant seals (Mirounga sp.) was made.Most of the recovered material was assigned to penguin (Sphenisciformes) bones, feathers, and eggshell fragments, although other taxa were also present.Specimens of flying birds (probably Charadriiformes), fishes (digested bone), krill (Euphausiacea), sponges (Hexactinellida), sea urchins (Echinoidea), culicid (Diptera), and ticks (Ixodidae), seaweed (Desmarestiales), and lichens (Ascomycota), were also identified.
Level 1.Most of the remains belong to penguin bones, although two invertebrates were also recovered.From a large chick, skull fragments (Pingfo I-7) consist of an interorbital region with the nasal gland sulcus (Figure 4A) and part of the cranial roof, a piece of the rostrum, and part of the jugal arch.The inner surface of the cranial roof presents a weak layer of periosteal bone covered by small foramina homogeneously distributed.The outer surface texture is neither rough nor fibrous, but neither is smooth and shiny like in the oldest juveniles and adults of penguin A femur (Pingfo I-1) characterised by the subtle torsion and curvature of the shaft and the incipient development of the fossa poplitea is assigned to a chick of about two months old given the lack of ossification of the condyles in the distal end, and the head and femoral trochanter at the proximal end, plus the texture of the surface (Figure 4B).The bone surface is not homogeneous along the shaft and the ends.The ends present elongated pores that merge and form open furrows converging towards the shaft, which is a little more ossified at least in some areas (mostly at the caudal face) that present a thin layer of bone covering the surface.Pores are smaller and less densely distributed in comparison with the ends, where endolithic lichens grow taking advantage of the rough texture.Small seaweeds are also attached to the proximal end.
Two humeri of different sizes (Pingfo I-13) characterised by a fibrous surface with incomplete ossification of the distal, detached pieces of periosteal bone and longitudinal fissures, are assigned to large chicks (Figure 4C-D).The humerus Pingfo I-2 belongs also to a juvenile since the proximal and distal ends are completely developed but it presents signs of incomplete ossification (Figure 4E).The fossa tricipitalis, bipartite in all the living penguin species, is still not divided, and although the ossification of the surface is partial, the muscle attachment scars are already developed.The fibrous texture and the presence of isolated pores indicate the immature condition.The surface of the shaft presents longitudinal fibres with isolated pores widened by the action of lichens apothecia that grow into the more superficial layers of bone.Erosive damage is also occasioned by the grains of sediments, some of them still attached to the periosteal bone.
Pingfo I-3 consists of a set of cervical, thoracic, and caudal vertebrae assigned to large chicks, in which the processes are ossified but other elements remain still unfused (Figure 4F).This condition prevents their assignment to free vertebrae or a composite structure.
The typical articular texture appears also in the unfused distal tibiotarsus (Pingfo I-5) of a large chick (Figure 4G).Similarly, two metatarsals with different development degrees (Pingfo I-4) are assigned to a large chick due to their rough texture with tight furrows and pores along almost the entire surface, plus small and homogeneously distributed pores at the ends (Figure 4H).The pedal phalanges Pingfo I-6 present a homogeneous surface with a tight longitudinal striation and small foramina.Towards the ends, the striation disappears, and the foramina is more evident, as the typical texture present in small chicks.Other bone fragments (Pingfo I-8) also belong to large chicks.
Besides, two fragments of feathers tail calami (Pingfo I-9, Figure 4I) small piece of penguin eggshell (Pingfo I-12) were recovered.The wider calamus (Pingfo I-68, Figure 4J) presents a longitudinal small fracture and modern epilithic lichen on its surface that easily degrades the bone.
Apart from the penguin remains, an arthropod moult (Pingfo I-11) that probably belongs to a hard tick (Figure 5A), and dehydrated macroalgae assigned to Desmarestia sp (Pingfo I-10) were separated from the sediment sieved and packed in the field (Figure 4L).
Finally, a cylindrical fragment of sea urchin spine (Pingfo I-14) with a rounded and button-shaped tip could belong to a Camarodonta or a Diadematoidea (Figure 5B-C).The spine is characterised by the lack of external ornamentation (like those of Camarodonta and some Diadematoidea) and a hollow interior (like those of Diadematoidea and some Camarodonta).
Level 2. Pingfo I-39 is the only element of the skull, identified as a quadrate of a large penguin, with a rough surface covered by fibrous bony tissue and abundant pores.
Several femora were identified for this level, including a complete femur of a juvenile (Pingfo I-23), and other femora of large chicks (Pingfo I-18, Figure 4M) and Pingfo I-36 with a central fossa poplitea incipiently developed.All these remains are badly preserved, with the outer surface rough and fragile, and the inner part of the shaft filled with trabecular and barely ossified tissue.Pingfo I-161 is also a fragmentary femur of a chick with the periosteal surface covered by longitudinal sulci and elongated foramina that constitute poorly defined walls.The ends are not preserved, and the bony tissue is fibrous and splintered.
The fragments of tibiotarsal shafts (Pingfo I-15 and Pingfo I-16) are assigned to large chicks due to the fibrous external texture, particularly at one of the sides.The ends of the shaft are rounded and poorly ossified, with wide sulci in-between the fibres and large oval dimples.Unfused ends of tibiotarsi (Pingfo I-27) were also recovered.A fibula (Pingfo I-39) which probably belongs to a juvenile due to its relatively small size was also identified.Small sediment grains are attached to the surface obscuring the texture surface.
Specimen Pingfo I-24 corresponds to a patella (Figure 4N), and Pingfo I-34 consists of two metatarsals still unfused, belonging to large chicks (Figure 4O-P).Three non-associated pedal phalanges Pingfo I-22 are assigned to large chicks (Figure 4Q-R).The surface presents a thick layer of bone with longitudinal and tightened fibres, that interrupt at the proximal and distal ends, where the surface is smoother and covered by small and homogeneously distributed foramina.
A scapula of a small (young) chick (Pingfo I-17) with an incomplete ossification, especially at the level of the ends, and three other scapulae (Pingfo I-20) with different weathering degrees, rounded ends not completely ossified, and a tightly striated periosteal bone (Figure 4S), reminds the surface of large chick wing elements.Specimens Pingfo I-25 correspond to four coracoids of large chicks (Figure 4T), and Pingfo I-29 to a slender and delicate furcula, without one of the ends.Although it is smaller than the adult furcula, the surface is more ossified than the other bones assigned to large chicks.The texture is smoother, and pores are absent.Pingfo I-50 consists of sternal fragments and Pingfo I-30 corresponds to unfused elements of a chick pelvis (Figure 4U-V).
From the wing, three complete ulnae (Pingfo I-21) with a striated surface and small and elongated pores are assigned to large chicks.Pingfo I-37 cannot be assigned unquestionably to an ulna or a radius since their immature morphology but is clear a zeugopod bone of a small chick a few days old.The incipiently ossified trabecular bone is exposed on the surface.Another undetermined wing element is Pingfo I-26, a flat element, like all the wing elements including the phalanges and the carpometacarpus, but with a general morphology that fits mostly with ulna/radius.However, the lack of ossification, a condition that allows its assignment to a small chick, prevents a more accurate determination.The texture is characterised by longitudinal furrows that diverge towards the end, where large and poorly defined foramina are dispersed in the areas under relief.
Spine elements are also frequent.Three vertebrae of chicks (Pingfo I-28, Figure 4W), vertebra belonging to a synsacrum of a large chick still covered by sediment (Pingfo I-35), a cervical vertebra of a large chick (Pingfo I-38, Figure 4X) with a spongy surface texture determined by the lack of ossification of the periosteal bone, and other vertebral fragments (Pingfo I-55) not completely ossified and eroded, probably belonging to small chicks, were collected.
Specimen Pingfo I-43 corresponds to a wing of a Diptera that was assigned to a Culicidae indeterminate on the basis of the veins pattern (Figure 6).The dented wing was found in situ in the sediment during picking work.To check the morphology of this insect wing, the material was stretched with a brush bristle on a slide.The wing is deteriorated and most of the scales and the fringe are not preserved.The membrane of most of the cells is particularly wrinkled and generates, on one hand, the thinning of the general contour of the wing, and on the other, that many of the veins are closer to each other or even giving the idea of an abnormal pattern, as for example a fake contact between the medial and cubital vein clearly seen in the photo as an 'X', (Figure 6A) which was not observed during the wing preparation and stretching.Despite these difficulties, the veins anatomy is mostly conserved and here tentatively identified.The six main longitudinal veins (costal, subcostal, radial, medial, cubital, and anal) were identified, being the subcostal vein (Figure 6B: Sc) only preserved in the medial sector of the wing.Through the distal part of the wing, the costal vein is closer to the radial vein, so all the cells between them are collapsed.The radial vein R 4+5 is broken and divided into two segments.The medial part of the radial vein is faint and not easily identified since it is also closer to the medial vein.The medialcubital cross vein is well preserved, but no other cross vein could be identified without doubt.The anal lobe is distorted in outline, and the anal vein is very tenuous.In particular, the A 2 vein looks like it is broken in its trajectory.
Level 3. Distal end of a humerus (Pingfo I-60) with a fibrous and fragile periosteal surface damage by subaerial exposition and a general morphology is consistent with a late juvenile or even with an adult morphology.Proximal half of a radius/ulna (Pingfo I-66), with patches of longitudinally striated bone that has detached in some areas exposing a spongy bone covered by small foramina, corresponds to a chick, and two presumably unfused metacarpals (Pingfo I-67), with opened furrows and elongated foramina, indicate their belonging to chicks.
Pingfo I-56 corresponds to an almost complete coracoid that already acquired the adult morphology, and a smooth surface (Figure 4C').However, the cracking and incipient peeling of the periosteal bone suggests an incomplete ossification consistent with a juvenile specimen.Specimen Pingfo I-64 corresponds to a fibula of a penguin chick (Figure 4D').The periosteal bone is only developed at the proximal end, and the trabecular bone is exposed along the shaft.A femur shaft (Pingfo I-57) with the surface highly weathered and an incipient development of the fossa poplitea belongs to a large chick (Figure 4E').Cracks and thin fissures run longitudinally, and pieces of periosteal bone are detached from the cranio-medial surface.A bioerosive trace, cranial to the fossa poplitea, evidences larval activity probably associated with the decay of tissues.Pingfo I-58 consists of two patellae, with a porous texture evidencing the incomplete ossification of the periosteal bone characteristic of large chicks.An unfused metatarsal bone (Pingfo I-59) is assigned to a large chick due to the ossification and fusion degree.Small and dense foramina covered homogeneously the surface, indicating the complete absence of compact bone.A small pedal phalanx (Pingfo I-61) and a tiny ungual phalanx (Pingfo I-65) subtly arched (Figure 4 F'), and with a neurovascular sulcus already developed, are assigned to large chicks due to the fibrous and porous texture.
Five fibrous vertebral fragments (Pingfo I-62), and an unfused portion of pelvis (Pingfo I-63), with the periosteal bone covered by small foramina loosely distributed and cracks occasioned by subaerial exposure (Figure 4 G'), are assigned to large chicks.
A set of bone fragments (Pingfo I-69) are also assigned to chicks given the characteristics of the surface texture and the incomplete ossification.Small pieces of seaweed are attached to the porous surfaces (Pingfo I-70).Other bone fragments (Pingfo I-71) are highly eroded.
Level 4. Two small cranial fragments (Pingfo I-77) with a fragile structure and a rough texture belong to an immature specimen.Two wing bones (Pingfo I-83) of different sizes would belong to a humerus and a radius or ulna, although the ossification state does not allow greater precision.In both cases, the general morphology, and the rough texture, characterised by parallel and tight furrows with elongated pores indicate their belonging to small chicks.Pingfo I-80 is an incomplete carpometacarpus assigned to a small juvenile, since the incipient fusion of the metacarpals and the striated texture with elongated pores Pingfo I-73 consists of an incomplete tarsometatarsus assigned to a juvenile (Figure 4I').The metatarsalia are partially fused and the sutures between them are still visible.The surface is notably fibrous and a thin layer of periosteal bone, covered by small foramina heterogeneous in size and form.The plantar surface is a little more ossified, and as a result, the striation is tight, and the texture appears a little smoother.An isolated metatarsal (Pingfo I-75) is assigned to a large chick of penguin given the lack of fusion of the metatarsalia and the striated surface of the bone.The ossification is less developed towards the ends, where the foramina are larger, and the striation turns into open furrows.A still unfused tarsometatarsal proximal end (Pingfo I-76) characterised by a rough texture at the proximal surface and a more porous one at the caudal surface, which is fused with the metatarsals in the juveniles and adults, is assigned to a large chick.
Three small pedal phalanges (Pingfo I-72) exhibit a certain weathering degree mixed with incomplete ossification of the periosteal bone, that occasion a heterogeneous and rough texture.Poorly defined striation running along the shaft with small foramina scattered between the sulci and that became denser towards the articular areas located at the proximal and distal ends.Two ungual phalanges (Pingfo I-74) present a smooth surface and isolated dimples (Figure 4 J'-K').The ossification is incomplete, especially at the proximal end and in the neurovascular sulcus area, which is developed but barely delimited.These features allow their assignment to large chicks.The five elements are assigned to large chicks.
Pingfo I-78 consists of rib fragments of chicks with a striated surface with elongated porous in the less ossified areas (Figure 4 L').A cervical (Pingfo I-79) and another undetermined vertebra (Pingfo I-84), together with small fragments of eroded bones are assigned to chicks, they present a rough texture, and the articular faces are covered by small foramina, like in the set of fragmentary and indeterminate bones (Pingfo I-81), and other chicks.Pingfo I-82 consists of several calami from penguin tail feathers (Figure 4K).
Finally, Pingfo I-85 was identified as a seabird tick (Figure 5E).Seabird ticks are classified into the families Ixodidae (hard ticks) and Argasidae (soft ticks), both included in the Ixodida (Arachnida).The hard ticks are characterised by the presence of a scutum, which is absent in the Argasidae, and a prominent gnathosoma, concealed beneath the body in the Argasidae.Based on the presence of a scutum and a gnathosoma visible in dorsal view, specimen Pingfo I-85 is assigned to the Ixodidae Ixodes a genus of hard tick widely distributed in the Antarctic Peninsula including the South Shetland Islands (Vanstreels et al. 2020).Although four species of Ixodes are present in seabirds, Pygoscelis populations are commonly infested with I. uriae White, 1852 (Barbosa et al. 2011), and numerous reports include the Antarctic penguin inhabitants of the Potter Peninsula (Barbosa et al. 2011;Vanstreels et al. 2020).
Level 5. Few materials were recovered from this layer, including a caudal vertebra (Pingfo I-86) of a chick (Figure 7A), bone fragments with a porous and rough surface, an unfused tibiotarsal trochlear end (Pingfo I-85, Figure 7B,C), and small pieces of fibrous bone (Pingfo I-88) of penguin chicks.Besides, a small fragment of digested trabecular bone (Pingfo I-89) probably belonging to fish was also separated from the sediment.
Level 6.An element of the wing (Pingfo I-92) is assigned to a radius or ulna of a large chick by the presence of a rough surface.The material is too fragmentary for a more accurate assignment.
Three fragmentary shafts (Pingfo I-96) of large chicks are preliminarily assigned to the femora of large chicks.The ends are rounded and non-defined, and the surface is not completely ossified.Pingfo I-94 consists of a tarsometatarsus fragment (Figure 7F).Despite the small size, the preserved trochlea exhibits a partial fusion, and a porous corresponding to the periosteal bone, that indicates a large chick-small juvenile age.Pingfo I-95 is a free metatarsal belonging to a penguin chick with both ends covered by small foramina homogeneously distributed (Figure 7D), whereas Pingfo I-97 is also a free metatarsal but assigned to a young juvenile due to the visible signs of a partial previous fusion (Figure 7E).It corresponds to the fourth left metatarsal and presents a more porous proximal end in comparison with the rest of the bone, characterised by a rough texture.Eight pedal phalanges (Pingfo I-91 and Pingfo I-104) are assigned to chicks (Figure 7H), whereas the three ungual phalanges (Pingfo I-99), with a striated surface and the neurovascular sulcus poorly defined, are identified as juveniles (Figure 7G).
Pingfo I-90 consists of six vertebrae not associated and including elements of the cervical and caudal regions.The fibrous surface of the bones, partially covered by small foramina, allows its assignment to large chicks, like the set of small fragments of bones (Pingfo I-100).
Fragments of highly deteriorated feather calami (Pingfo I-93 and Pingfo I-102) are dehydrated and present longitudinal fractures.They would belong to penguins, like the eggshell fragments Pingfo I-105.Seaweeds (Pingfo I-101) are contained within a small block of consolidated sediment, and small pieces of trabecular bone presumably of fish (Pingfo I-103 and Pingfo I-106) present signs of chemical digestion.Level 7. The elements of the skull recovered from this level correspond to immature specimens.Pingfo I-131 consists of the nasal gland sulcus area that is easily separated from the rest of the cranium even in medium-to-large chicks, in which the fusion of the elements is still incomplete, and is preserved without fractures (Figure 7J).Pingfo I-128 corresponds to the rostral fragment of a mandible with wide and open furrows and large foramina (Figure 7N).These features indicate a deficient ossification consistent with the young age of a chick.An articular end of a mandible (Pingfo I-125) with a silky surface and covered by longitudinal fissures belongs to an older specimen, probably a juvenile (Figure 7M).The material is incompletely ossified and highly eroded, mainly at the level of the cotyles.
A complete articulated left wing (Pingfo I-151) of a large juvenile penguin was extracted from a consolidated block of sediment (Figure 7O,P).The wing is entirely articulated, and preserves the humerus, radius, ulna, radiale, ulnare, carpometacarpus, and the three phalanges of the second and third digit.The elements are weakened due to the incomplete ossification of the periosteal bone, which acquires a rough and irregular texture along all the surfaces.
Other isolated elements of the wing include specimen Pingfo I-121 which consists of two poorly preserved proximal ends of humeri, assigned to juveniles.The tricipital fossa and the scar for the insertion of the main wing muscles are incompletely developed.The ossification degree of the bone is advanced, acquiring a silkier texture in comparison with younger individuals, although covered by longitudinal thin fractures that make the surface rougher.Also assigned to large chicks-young juveniles are the complete humeri Pingfo I-122 (Figure 7K,L) and Pingfo I-135 (Figure 7R).The surface is fibrous, and the periosteal bone is highly deteriorated due to the presence of longitudinal fractures and the detachment of pieces of periosteal bone.The caput humeri, the fossa tricipitalis, and the distal ends are not completely ossified.Two other humeri (Pingfo I-123) are only partially ossified and have undefined ends.The surface is striated like in two-month-old chicks, and the periosteal bone is cracked and detached in some areas.
A radius (Pingfo I-124) completely ossified presents a texture consistent with a sub-adult specimen.The surface is rougher and barely silky, texture partially erased by the longitudinal fractures that covered the entire bone.Irregular and shallow channels, probably caused by larvae, are excavated in the medial edge of the ventral face.
Two complete femora (Pingfo I-119) with the caput femori poorly defined, the distal end incompletely ossified, and a striated surface are consistent with young juveniles.Pingfo I-130 (Figure 7V) and Pingfo I-152 correspond to two femora of large chicks with the caput femori still not defined, the most external layer of bone completely cracked, and with some pieces detached.The shaft Pingo I-215 seems more curved than normal, although the severe damage on the periosteal bone prevents a more precise evaluation of the material looking for old traumatic fractures.Pingfo I-153 is a femur of a penguin chick with a bioerosive trace probably caused by a larva (Figure 7W), and Pingfo I-134 corresponds to four fragments of chicks femoral shafts.The ossification is incomplete, and the ends lack definition.Pronounced longitudinal fractures facilitate the detachment of periosteal bone in some areas.The distal ends of the femora (Pingfo I-111 and Pingfo I-136) are assigned to large chicks based on the unfused condition and the textural ageing.The ossification degree indicates an advanced age among chicks.
An unfused proximal end (Pingfo I-108) and three trochlear ends of tibiotarsus (Pingfo I-107 and Pingfo I-139) are assigned to large chicks due to their unfused condition regarding the shaft and the porous texture.Pingfo I-120 consists also of an unfused distal end of tibiotarsus which belongs to a juvenile.The articular area is incompletely ossified, and the periosteal bone is cracked by dehydration.Pingfo I-109, Pingfo I-110, and Pingfo I-149 are 13 tibiotarsal shafts of large and medium chicks with the sulcus extensorius incipiently developed and a rough and porous surface.Pingfo I-112 is an isolated patella.
Pingfo I-114 are three free metatarsals without signs of previous fusion of the elements that together with the rough and porous texture of the incompletely ossified periosteal bone, allow their assignment to small chicks.Pingfo I-132 represents a free metatarsal of a large penguin chick with a fibrous texture that facilitates the attachment of the surrounding sediments.Two larger metatarsals (Pingfo I-113) are assigned to large chicks or young juveniles due to the presence of a small scar indicating a previous incipient metatarsal fusion.Pingfo I-144 is a metatarsal belonging to a juvenile with an extended scar of the previous fusion with the neighbouring metatarsals.
Three pedal phalanges (Pingfo I-115) characterised by a fibrous surface and covered by irregular foramina are assigned to large chicks.A small pedal phalanx (Pingfo I-143) with a porous surface would belong to a penguin chick.
Pingfo I-116 corresponds to a coracoid of a large chick-young juveniles.Although the material is still included in the matrix, a fibrous and porous texture indicating an incomplete ossification is visible.Longitudinal fractures covered the surface.Pingfo I-133, Pingfo I-147, and Pingfo I-148 represent coracoids of large penguin chicks with incomplete ossification of the omal and sternal ends, that together with the fibrous texture characterised by open furrows and the partial detachment of the periosteal bone, allows the ontogenetic estimation.Likewise, two omal ends (Pingfo I-117) and a distal end of coracoids (Pingfo I-118) with the sternal articulation still undefined, are assigned to large chicks.A tunnel probably caused by a larva is in the dorsal surface of Pingfo I-118 (Figure 7I).Pingfo I-156 is a scapula barely ossified belonging to a small chick.
Nine incomplete vertebrae of different spine regions (Pingfo I-126) were found isolated and non-associated with each other.The elements are poorly ossified, mainly concerning the processes and arches, indicating they belonged to immature individuals, probably large chicks.Two complete vertebrae (Pingfo I-150 and Pingfo I-142), plus two partially fused synsacra and other associated vertebrae Pingfo I-140 (Figure 7U) and Pingfo I-141 (Figure 7S,T), belong also to large chicks.Other fragments of bones (Pingfo I-127, Pingfo I-137, Pingfo I-145, and Pingfo I-154) also correspond to chicks.
Three highly modified penguin feathers (Pingfo I-159, Figure 7X) probably belonging to dorsal contours, and a downy feather (Pingfo I-160, Figure 7Y)), were separated from the sediment.Size and morphology are consistent with the Pygoscelis plumage.
Also from birds, the ulna of a flying marine species (Pingfo I-129) was found (Figure 7Q).The surface is covered by longitudinal and parallel fractures and the broken distal end presents signs of diagenetic damage.The proximal end is lacking, but the bone is consistent in size and robustness with an adult skua.Pingfo I-157 corresponds to a small vertebra of fish (Figure 7Z).The surface is eroded, and no other relevant feature was preserved to provide a more accurate determination.Finally, seaweed fragments mixed with unconsolidated sediment (Pingfo I-145), and isolated thallus fibres Pingfo I-158 were found (Figure 7A').

Penguin ages and parts representation
All the materials of our sample collected at Pingfo I correspond to immature specimens, most of them belonging to large chicks, and a few to small chicks and juveniles (see supplementary file).The level 5 is the poorest in penguin bone representation with only two isolated chicken bones and was not represented in the pie charts of the supplementary file.Being as thick as level 4 (0.20 m) which has a decreasing proportion of bones assigned to large chickens, small chickens, and juvenile bones respectively, the scarcity at the fifth level could be more related to taphonomic differences rather than a sample bias.The rest of the levels could be characterised by the predominance of large chicken bones, followed in number by small chicken and always a few juveniles, except the level 1 where they are completely absent.Comparisons with the materials previously collected in Pingfo I are not possible since bones were not identified by ontogenetic stages in Del Valle et al. (2002).Among the large chicken category, the vertebrae are particularly very well represented in most of the levels.Probably their massive structure facilitates the preservation of these elements.

Remarks on erosive damage
At first sight, the whole sample of Pingfo I could be assigned to stages 1 (short and shallow longitudinal cracks) and 2 (porous surface with more evident cracks) of the five groups proposed by Muñoz and Savanti (1998).However, as we establish above, chicks are the predominant ontogenetic stage in the sample, and their periosteal bone is still not completely ossified.Consequently, chicks present a porous and rough surface due to bone immaturity and not to weathering (see Acosta Hospitaleche and Picasso 2020 for details).For that reason, real weathering damage might be lower than it appears.
Besides, bioerosive traces presumably caused by larvae activity were found in bones from levels 3 (Pingfo I-57) and 7 (Pingfo I-118, Pingfo I-124, and Pingfo I-153), which add to the sternum with predation marks attributed to skuas previously reported by Montalti et al. (2009).

Pingfo II
The profile (Figure 4) is a mound similar to the surrounding sediment in appearance, which is represented by five levels named as 1 to 5 from base to top, that constitute a continuous sequence with the site identified as 'PC' (Del Valle et al. 2007 drawn from data reported by John and Sugden 1971), currently destroyed by the meltwater circulation that runs from the highest areas to the sea and the anthropic action (Figure 4).
The base of the sequence exposed at Pingfo II starts with level 1 of 0.60 m of muddy sands containing Laternula elliptica (King and Broderip, 1832) in life position together with its escape traces.The overlying level 2 is 0.40 m thick and consists of glaciomarine deposits that acquire a cuneiform shape.
Level 3 is only 0.60 m thick but lithologically more complex.Most of the level, and particularly the base, consists of sands with crushed shells of Laternula elliptica.Beds of fine-grained gravel with penguin bones and seaweed-rich sands are intercalated towards the top.L. elliptica appears in life position and oblique to the boundaries of the layer.Del Valle et al. (2007) inform that all the sub-fossil vertebrates collected in Pingfo II (penguins and marine mammals) come from this level, partially correlated with the equivalent level III of the disappeared 'PC' sequence, from which only bivalves were mentioned (John and Sugden 1971).
Level 4 is 0.80 m thick and consists of glacial till deposits, and finally, at the top, level 5 consists of a thin and surficial layer of 0.10 m thick of marine beach gravels.

Materials from Pingfo II
Pingfo II was previously excavated by Del Valle et al. (2007) and a set of macrofossils were found during the lithofacies analysis of the sequence.In that case, penguin bones were collected only from level three together with a few marine mammal bones, seaweed, and bivalve fragments (Figure 8).Additionally, complete specimens of the bivalve Laternula elliptica found in life position together with their escape traces were reported from the lowest level (Del Valle et al. 2007).
Unlike previous investigations (see Del Valle et al. 2007), we found a small number of vertebrates along the excavated sequence.It could be due to more intensive work in the past and to the destructive processes operating in the area, which is closer to the Station activity (see below) and furrowed by melt streams.The information about the skeletal element, systematic and ontogenetic assignment is included in Supplementary Table 1.The preservational attributes are described below.
Level 1. a set of stem ossicles of crinoids (Pingfo II-52, Figure 8A) were separated from the sediment.
Level 3. A distal end of humerus (Pingfo II-2), a radius and small fragments of bivalve shells (Pingfo II-3), and a carpometacarpus (Pingfo II-4) are assigned to juvenile penguins (Figure 8D) based on the textural ageing and the advanced ossification degree.Besides, bone splinters (Pingfo II-5), and fragments of small bivalve shells (Pingfo II-6 and Pingfo II-7, Figure 8E), complete the sample of this level.
Level 4. A single piece of bone consistent in a proximal end of a penguin coracoid (Pingfo II-8), together with a Laternula elliptica shell (Pingfo II-9, Figure 8F) found in life position (Figure 8G), were collected.
Surficial deposits.Two isolated bones were assigned to elements of the skull, a bill fragment (Pingfo II-10) of a juvenile penguin (Figure 8H), and an articular end of mandible (Pingfo II-11) of a small chick.Both remains are highly deteriorated by subaerial exposition and the bone is fragile and brittle.
Based on the ossification and the textural ageing, a cranial fragment of sternum preserving the keel (Pingfo II-12) is assigned to an adult (Figure 8I), whereas another fragmentary cranial fragment of sternum (Pingfo II-13) is interpreted as a chick remains.The keel is lacking, and the surface is covered with cracks.Small pieces of periosteal bone are detached.
Two cervical and caudal vertebrae (Pingfo II-14) belonging to juvenile penguins present a silky surface and reach almost the adult size.A fragment of synsacrum (Pingfo II-15, Figure 8J) fragment of pelvis (Pingfo II-16) are also assigned to large chicks based on the relatively smooth texture and the ossification of the periosteal bone.
Ribs (Pingfo II-17 and Pingfo II-18) could belong to chicks, although beyond the size, no other features could be observed.Two larger vertebral segments of ribs (Pingfo II-19) could belong to a juvenile penguin.
From the wing, a complete humerus (Pingfo II-20) slightly smaller than those of the adults (Figure 8K), five proximal ends (Pingfo II-21), and three distal ends of humeri (Pingfo II-22) of juvenile penguins with the surface completely weathered and pieces of periosteal bone detached, were collected.Although the ossification seems advanced, all the materials are highly weathered, three of them showing also signs of dehydration and a splintered surface.Another proximal end (Pingfo II-23) belongs to a younger specimen, probably a large chick, although its surface is barely damaged by weathering.Two ulnae (Pingfo II-24) incompletely ossified, and a rough surface are assigned to large chicks and a carpometacarpus (Pingfo II-25) of a small chick with a weathered surface covered by wide fractures (Figure 8L).One of the faces is colonised by black and orange lichen apothecia.Coracoids of adult penguins are represented by the proximal end (Pingfo II-26) and probably Pingfo II-27 which consist of seven complete coracoids that could also belong to highly weathered large juveniles with the periosteal texture obscured by subaerial exposition.An incomplete and barely damaged coracoid (Pingfo II-28) assigned to a juvenile penguin presents the periosteal bone completely cracked and splintered.Also, the six coracoids (Pingfo II-29) are assigned to large chicks due to the fibrous texture of the surface that presents a splintered periosteal bone with some pieces detached.Endolithic and fruticose lichens are attached to the surface of one of the coracoids.
Two distal ends of coracoids (Pingfo II-30) covered by fine longitudinal cracks and with small pieces of periosteal bone detached are interpreted as belonging to a large penguin chick, like the scapula (Pingfo II-31) which is barely damaged and completely covered with apothecia of endolithic lichens.
Pingfo II-32 consists of two complete femora of juvenile penguins (Figure 8M), with the surface rough and weathered by incomplete ossification and subaerial exposure, whereas Pingfo II-33 and Pingfo II-34 belong to a proximal end of femora assigned to large penguins based on the cracked and splintered surface, weakened by the incomplete ossification of the periosteal bone.
Proximal ends of tibiotarsi are assigned to large chicks (Pingfo II-35 and Pingfo II-36), in which the cnemial crests are not developed yet.Distal ends of a juvenile (Pingfo II-37) with the periosteal bone fractured and splintered and belonging to an adult (Pingfo II-38) are also preserved.Pingfo II-39 consists of an isolated patella of a juvenile.
A tarsometatarsus (Pingfo II-40) with the metatarsalia completely fused and a rough surface is assigned to a juvenile (Figure 8N), whereas an incomplete free metatarsal (Pingfo II-41) is determined as a large penguin chick by the precarious fusion of elements.Pingfo II-42 is a pedal phalanx of a juvenile penguin.
Flying birds are represented by a fragment of synsacrum and pelvis belonging to a non-penguin bird (Pingfo II-43) due to its smaller size and the fusion between the pelvis and the synsacrum (Figure 8O).Bioerosive traces probably caused by larval activity are on the external surface of the pelvis.Besides, Pingfo II-44 consists of two splintered elements, with detached pieces of periosteal bone, which might belong to the ulnae of flying birds.Besides, a feather (Pingfo II-45) is preliminarily assigned to a giant petrel Macronectes giganteus (Figure 8B).

Penguin ages and parts representation
In our sample, not all the levels allow the recovery of specimens; penguin bones were collected from levels 3, 4 and the surface.The most abundant collection of Pingfo II corresponds to the superficial deposits; however, previous contributions report abundant bones from level 3, most of them presumably corresponding to adult penguins as their ages were not reported (Del Valle et al. 2007).The sample of the surficial levels is composed of penguin bones assigned to chicks of different ages, juveniles, and adults.

Remarks on erosive damage
The taphonomic damage of the elements recovered in Pingfo II includes anthropic disturbance and contamination of the surficial deposits; two cigarette filters were mixed with the bones collected at the top of the sequence.
Most of the bones exhibit moderate erosive damage that can be assigned to the first (bones with longitudinal and shallow cracks) and second (deeper fractures and a porous general surface) stages of Muñoz and Savanti (1998).Besides, endolithic and fruticose lichens are attached to the surface of some bones (e.g.Pingfo II-29 and Pingfo II-31), whereas other bioerosive traces probably produced by larval activity were found on other bones (i.e.Pingfo II-43).

Discussion and conclusions
Whereas previous studies focus the attention on the age estimates of the samples and the eustatic levels changes observed in the examined profiles (Del Valle et al. 2002, p. 2009, see however, Montalti et al. 2009), we did it on the composition of each level in a macroscopic and microscopic level, the preservational state, the ontogenetic and the skeletal parts representation.First, we raised the profiles of Pingfo I and Pingfo II, whereas PC was discarded for new research given its disturbed current state.Fossil/sub-fossil materials were collected and samples of sediments and/or consolidated rocks were taken to process in the laboratory.On the other hand, the 'Pingui' locality needs to be relocated, since the geographic coordinates seem to be mistaken in the main publication, and therefore we could not work there.
The presence of both, wing and legs elements, in a higher proportion than the axial elements, has been interpreted as an indicator of the natural origin of the assemblage (Emslie 1995;Cruz and Savanti 1999;Cruz 2003) when foot propelled marine divers are analysed (Cruz 2006).Like in actualistic approaches (Cruz and Savanti 1999;Cruz 2003) and other Antarctic rookeries analysed on surface and excavated (Emslie 1995), all the limb bones are represented in a high proportion in most of the levels of Pingfo I which could be interpreted as representing a nesting area (Supplementary Table 1).Particularly when each bone element is considered, axial elements (i.e.vertebrae) are very abundant in levels 1, 3, 4, 6, and 7 for the large chicken category.Although they only have the highest number of bone representations at level 1, vertebrae seem to be very well preserved in all the sequence (Supplementary Table 1 and Supplementary Figure 1).The differences with other collections made at Pingfo I, where less vertebral elements were found (Del Valle et al. 2002;Montalti et al. 2009) maybe due to the sampling methods here used (i.e.sediment sieving and fractionation of whole rock blocks in the laboratory).The vertebrae body is more massive, exposing less surface per volume to any possible dissolving or weakening agents in the sediment, than certain long bones.Vertebrae presence and proportion in nesting areas, which are characterised by low, or none transport, and a constant sedimentation rate as a consequence of the biocoenosis activity, could produce rapid burials of these elements.A large sampling in other similar taphocoenosis, applying the same methods here used, is needed to evaluate the previously less proportion of axial elements (Emslie 1995;Cruz and Savanti 1999;Cruz 2003).
The analysis of ages representation is supported by several actualistic studies on breeding colonies.For instance, observations of extant biocoenosis of Spheniscus magellanicus (Forster, 1781) indicate that large accumulations of chick carcases only happen within the colony (see Cruz 2003).Skeletons of smaller chicks are more easily destroyed and tend to disappear, whereas large chicks are abundant and preserve better, and juveniles preserve well but are scarce within the colony.This matches with our results for almost all the levels (see results and Supplementary files).
At present, numerous variables cause the death of hicks and juveniles every year.Bodies are deposited within the nidification site due to their nidicolous and semi-altricial condition (Starck and Ricklefts 1998), plus their philopatric behaviour.All the remains recovered from Pingfo I correspond to sub-adult individuals assigned to different aged chicks, indicating the settlement of a breeding colony in this site during the time of deposition of the sequence.
Another parameter to evaluate the transportation degree is the weathering damage.Muñoz and Savanti (1998) proposed the categorisation into five groups, from 0 (absence of weathering) to 4 (splintered bone that falls apart), but only based on adult specimens.Its implementation without the proper pondering of the ossification degree proper of each ontogenetic stage, would result in an overestimation of the transportation degree in the sample of Pingfo I.
However, another clear evidence of the presence of nests in this locality is the finding of a considerable amount of penguin eggshells.These elements are easily fragmented due to the trampling within the colony, and the small pieces are usually destroyed by the taphonomic processes in situ, not supporting the minimum transport.The proximity to the coastline, where nests are usually built, is reinforced by the abundance of the seaweed recale Desmarestia.Besides, the undigested pieces of krill found in Pingfo I could have accumulated when the parents carried them to the nests to feed their chicks, and the digested trabecular bone could come from the penguin faeces.
The tick of the genus Ixodes collected in Pingfo I is also consistent with our interpretation of the site as a nesting area.Hard ticks are characterised by low mobility restricted to metres and their dispersion depends on hosts, in this case, penguins.Despite larvae, nymphae, and adults feed on different hosts, all the stages are found on penguins (see Randolph 1989 for further details).
The Culicidae wing separated from the sediment of Pingfo I deserves a particular comment.The wing undoubtedly corresponds to a mosquito which is not present in Antarctic latitudes today.Although it is impossible to discard the sample contamination during transport or in laboratory works, the nature of the finding should be considered.A large rock sample was taken by two of us (CAH and JNG) as a unique consolidated block of around 15 kg.It was transported completely wrapped and within a hermetic bag from the collection site (Pingfo I) to the laboratory in the Museo de La Plata, where it was processed five years later, during an entire year.The rock was manually disaggregated, separating macroscopic specimens (e.g.bones) and checking the sediment under a binocular microscope.Part of the sediment was also prepared for microscopic specimens sampling analysis.The wing, found by one of the authors (BQ) folded in on itself and covered by fine sediment, was in the unconsolidated sediment.Arguments against contamination will always be considered ad-hoc statements to support a positive interpretation, so we are going to avoid them.We prefer to be cautious, use the wing specimen as a working hypothesis, and wait for future sampling to reinforce, or reject this specimen assignment.
The southernmost natural population of parasite culicid in South America is represented by Aedes albifasciatus (Macquart, 1838) recorded in Tierra del Fuego, Argentina (Rossi 2015), around 1000 km north from South Shetland Islands.Mosquitoes are cosmopolitan and are present worldwide except in Iceland, a few extremely isolated Arctic and South Pacific islands, and the Antarctic continent and their surrounding islands (Rueda 2008) where there is no fossil or ancient record either.In this sense, the specimen here described could be the first ancient record of the group, dated around the same age of the level where it was found.Level 3 of Del Valle et al. (2002) or level 2 in Emslie et al. ( 2020) and the present work, was temporally constrained between dates around 5,850-5,590 and 5,925-5,665 cal BP.This suggests the presence of mosquitoes in the South Shetland Island during the Northgrippian Age (Holocene) partially coincident with the Holocene Climatic Optimum (HCO).
According to the analysis of ice cores from coastal and central sites in East Antarctica, the warmest period was recorded in Antarctica roughly 11,500 to 9,000 cal BP, closer to and after the end of the last ice age (Masson et al. 2000).Other proxies, particularly for the Antarctic Peninsula in the Lallemand Fjord, indicated the presence of an open water assemblage between 7,890-3,850 cal BP, suggesting that the sea ice cover was at a minimum level, and associating them with climatic warming, which characterises much of the Antarctic Peninsula during middle Holocene (Taylor et al. 2000(Taylor et al. , 2001)).Despite several differences related to the nature of the information used for climatic inferences in Antarctica (see a summary in Florindo et al. 2021) the 'mummified' mosquito wing in Pingfo I sequence, allows a wide range of possibilities concerning the climatic conditions.
If mosquitoes were present in the South Shetland Islands by the middle Holocene in coincidence with the HCO, a close relation with the warmest climatic conditions than today seems to be inevitable to infer.At least temperatures and other climatic variables are similar to those for the southern mosquitoes distribution in South America, for example in Tierra del Fuego.But also, it should be considered that extant populations of parasitic Culicidae are today present in even higher latitudes in the Arctic, for example in Hazen Camp 81°49' N., 71" 18' W, Ellesmere Island (Corbet and Danks 1973).In contrast to the wide range of terrestrial vertebrates present as potential hosts in the Arctic, no terrestrial vertebrates are present in Antarctica.If mosquitoes were present in West Antarctica during the middle Holocene, the fossil record suggests that their seek of blood meal will be restricted to seabirds, and the marine mammals which breed on land.Also, several adaptations to face blood scarcity should be considered.Aedes nigripes (Zetterstedt, 1838), probably the most abundant and widespread Arctic mosquito (Vockeroth 1954) presents an autogeny adaptation, to be able to reproduce without a blood meal (Corbet 1964(Corbet , 1967)).These actualistic approaches suggest that neither temperature or other polar climatic conditions, nor the absence of possible hosts, are real constraints for the presence of Culicidae during the middle Holocene in West Antarctica.
Unlike in Pingfo I, the superficial deposits of Pingfo II are composed of different skeletal elements, without a clear predominance of limb bones.It might be related to the genesis of the deposits, not corresponding with the observed in other accumulations formed in situ.Beyond the transport suffered by these bones, the main problem we found during the analysis of the remains is the contamination with recent elements of variable procedence.
The lack of a clear predominance of chicks in Pingfo II, also prevents the assumption of an ancient penguin rookery settled down in the same area.These bones could have been transported from a nearby coast.The closer locality for further comparisons is the non-longer existent 'PC', described in detail by John and Sugden (1971) which was probably eroded.Although vertebrates were not mentioned in that contribution, one skull and several long bones assigned to penguins were later reported from there (Del Valle et al. 2002).Although these bones were not radiocarbon dated, the seaweed closely associated with them was dated by John and Sugden (1971) in c. 6400 reservoir-corrected 14C cal BP when correction of −1300 years (Berkman et al. 1998) is applied.
The taxonomic diversity of the sub-fossil fauna reported at Pingfo II in previous research (Del Valle et al. 2007) and the current research is quite similar.Although penguins are the dominant elements of the assemblage (Supplementary Table 2), bones and feathers of other birds, and mammals are also present.On the contrary, eggshells and invertebrates remains were not found in this profile.Among the surficial deposits of Pingfo II several shafts of indeterminate flying birds, a feather probably belonging to a giant petrel Macronectes giganteus due to the size and the calamus diameter, and some large bones of immature marine mammals were recognised in this sample.The deeper levels are composed of bivalve shells, some of them in life position and even associated with escape traces, and a few isolated bones.In Pingfo II, seaweeds are not abundant like in Pingfo I.
In summary, according to all the parameters evaluated above, and the information previously published, we agree with the proposal of two different depositional environments for Pingfo I and Pingfo II.According to the ages and parts representation (Supplementary Table 1 and Supplementary Figure 1), Pingfo I corresponds to a breeding colony of pygoscelid penguins for the middle Holocene (mentioned as early to middle in Emslie et al. 2020), but despite this, no deposits of ornithogenic soils of that age are present here.The topography of the locality and the solifluction of soils over time have already been postulated to explain their absence (Emslie et al. 2020).The parts and ages representation, plus the presence of recale seaweeds, abundant eggshell fragments, and other microscopic organisms related to the colony daily activity point consistently to the same depositional environment.Pygoscelid penguins nested in an ice-free beach, topographically lower than it appears today, during the deposition of the Pingfo I sequence.Vast breeding colonies dominated by P. adeliae and P. papua are settled down around Pingfo I today.
On the contrary, Pingfo II is not so informative regarding the proximity of the nesting area.The age representation (Supplementary Table 2), and the absence of eggshells and other elements related to the daily activity of the colony, suggest that Pingfo II would not correspond with the settlement place of an ancient breeding colony.The preservational state of the bones, and the associated fauna, indicate a minimum transportation degree, supporting that the excavated profile could represent an old beach where bones transported from a nearby site were finally deposited.This is relevant because Pingfo II corresponds to one of the oldest penguin assemblages of King George Island.The radiocarbon dating on penguin bones yielded ca.7414 yr BP (Del Valle et al. 2007).Another penguin colony dated ca.6700 yr BP [7.4-5.8 cal ka BP] was mentioned for Ardley Island, in the southwest end of King George Island, close to Fildes Peninsula (Roberts et al. 2017).Other penguin rookeries indicate the penguin occupation in 5800-5300 cal BP (Barsch and Mäusbacher 1986) and 6401-6486 cal BP (Pallàs et al. 1997) for King George Island.The data provided here constitute an important input for the reconstruction of the coastline and the restriction of the deglaciation age of the coastal areas in the Potter Peninsula.

Figure 1 .
Figure 1.Map of the area under study: A, South Shetland Islands; B, detail of the King George Island, located at the west of the Antarctic Peninsula; C, Close-up of the Potter Peninsula; D, location of Pingfo I, Pingfo II and the penguin and elephant seal breeding colonies (see legends for references).The pictures below correspond to the excavated rookeries; E, Pingfo II; and F, Pingfo I.

Figure 2 .
Figure 2. Geologic profile of Pingfo I with a correlation of the levels recognised in previous contributions.The modern breeding colony corresponds to Pygoscelis species.

Figure 5 .
Figure 5. Part of the invertebrates collected in Pingfo I: A, Pingfo I-11 arthropod moult probably belonging to a hard tick (Ixodidae), the schematic lines indicate the recognised parts of the body; B-C, Pingfo I-14 sea urchin spine (Echinoidea), showing in B, the lack of ornamentation and in C, the hollow interior (hi); D, Pingfo I-45 spicule of a siliceous sponge (Hexactinellida); E, Pingfo I-85 hard tick (Ixodidae) with the mouthparts (mp) and the coxa of legs (cl) indicated in the picture and the scheme.Scale bars represent 0.01 mm.

Figure 7 .
Figure 7. Part of the materials collected in Pingfo I (levels 5-7): A, Pingfo I-86 caudal vertebra of penguin chick; B, Pingfo I-87 unfused distal end of tibiotarsus in lateral and; C, distal views; D, Pingfo I-95 unfused metatarsal of penguin; E, Pingfo I-97 metatarsal of a large penguin chick; F, Pingfo I-94 fragment of tarsometarsus of a juvenile penguin; G, Pingfo I-99 ungual phalanx of a juvenile penguin; H, Pingfo I-104 pedal phalanx of penguin; I, Pingfo I-118 coracoid of penguin with a bioerosive trace magnified into the square; J, Pingfo I-131 cranial fragment of a penguin chick; K-L, Pingfo I-122 two humeri of juvenile penguins; M, Pingfo I-125 articular end of mandible probably of a juvenile penguin; N, Pingfo I-128 rostral end of mandible of a chick; O, Pingfo I-151 complete and articulated penguin still into the sedimentary matrix; P, Pingfo I-151 extracted from the matrix; Q, Pingfo I-129 ulna of a flying bird; R, Pingfo I-135 humerus of a penguin chick; S-T, Pingfo I-141 two synsacra of large penguin chicks; U, Pingfo I-140 synsacrum of penguin chick; V, Pingfo I-130 femur of a large penguin chick; W, Pingfo I-153 femur with a bioerosive traces magnified into the square; X, Pingfo I-159 contour feather of penguin; Y, Pingfo I-160 downy feather of penguin; Z, Pingfo I-157 small vertebra of fish; A', Pingfo I-158 seaweed.Scale bars represent 10 mm, close-ups areas, X, Y, and Z have their own scales equivalent to 0.1 mm.

Figure 8 .
Figure 8. Part of the materials collected in Pingfo II: A, Pingfo II-52 stem ossicle of crinoid; B, Pingfo II-45 feather of a giant petrel Macronectes giganteus; C, Pingfo II-1 small bivalve; D, Pingfo II-4 carpometacarpus of a juvenile penguin; E, Pingfo II-6 and Pingfo II-7 fragments of a bivalve shell; F, Pingfo II-9 Laternula elliptica; G, L. elliptica in life position before its extraction from the sediment in the field; H, Pingfo II-10 fragment of the bill of a juvenile penguin; I, Pingfo II-12 cranial fragment of a sternum; J, Pingfo II-15 synsacrum of a large chick; K, Pingfo II-20 complete humerus of a juvenile penguin; L, Pingfo II-25 carpometacarpus of penguin chick covered by apothecia; M, Pingfo II-32 femur of a juvenile penguin; N, Pingfo II-40 tarsometatarsus of a juvenile penguin; O, Pingfo II-43 synsacrum and partial pelvis of a flying bird; P, Pingfo II-47 femur diaphysis of a marine mammal; Q, Pingfo II-48 rib of marine mammal; R, Pingfo II-50 fragment of calcareous seaweed.Scale bar represents 10 mm; B and E have their own scale.