Late Pleistocene-Early Holocene faunal remains from the Waterfall Bluff Rock Shelter in the Mpondoland region of Eastern Cape Province, South Africa.

A number of coastal archaeological sites across Southern Africa provide an excellent record of faunal and Hominin interactions during Pleistocene interglacial periods and the Holocene, but a much poorer record of the same during glacial periods, when sealevels were lower, with the effect that most coastal settlements during these intervals would be below modern sealevels. The Waterfall Bluff Rock Shelter in the Mpondoland region of Eastern Cape Province, appears to be an exception to this, with an archaeological record which begins during Marine Isotope Stage 3 (roughly 39 000 to 29 000 years ago) to the Early Holocene (about 8000 years ago), and includes the Last Glacial Maximum, and the glacial/interglacial transition at the end of the Pleistocene. This site has yielded the remains of Fish, Shelfish, and Marine Mammals, which demonstrate that the hunter-gatherer populations using the rock shelter were utilizing coastal resources even during the Last Glacial Maximum.

In a paper published in the South African Journal of Science on 4 December 2024, Sandee Oster and Jerome Reynard of the School of Geography, Archaeology and Environmental Studies at the University of the Witwatersrand, Hayley Cawthra of the African Centre for Coastal Palaeoscience at Nelson Mandela University, and the Minerals and Energy Unit at the South African Council for Geoscience, Irene Esteban, also of the African Centre for Coastal Palaeoscience at Nelson Mandela University, and of the Archaeological and Archaeometric Research Unit and Institute of Archaeology at the University of Barcelona, Justin Pargeter of the Department of Anthropology at New York University, and the Rock Art Research Institute at the University of the Witwatersrand, and Erich Fisher, again of the African Centre for Coastal Palaeoscience at Nelson Mandela University, and of the Evolutionary Studies Institute at the University of the Witwatersrand, the Interdisciplinary Center for Archaeology and the Evolution of Human Behavior at the University of Algarve, and the Institute of Human Origins at Arizona State University, present a preliminary analysis of the fauna at the Waterfall Bluff Rock Shelter, and discuss the implications of this for the local palaeoenvironmental and palaeoecological conditions, and how this would have affected the subsistence activities of hunter-gatherer populations in the area.

Eastern Mpondoland, defined as the area between the mouths of the Mthatha and Umtamvuna rivers, has a diverse landscape with a number of deeply incised plateaus, and areas of sourveld grassland, forest vegetation and bushveld, with patches of Southern Coastal Forest, Southern Mistbelt Forest and Scarp Forest. This area has a diverse fauna including woodland species such as bushbuck, Tragelaphus scriptus, Blue Duiker, Philantomba monticola, and Grey Rhebok, Pelea capreolus, grassland species such as Southern Reedbuck, Redunca arundinum, Oribi, Ourebia ourebi, and Bontebok, Damaliscus pygargus,  as well as mixed environment species such as Eland, Tragelaphus oryx.

Waterfall Bluff Rock Shelter is located 24 m above sealevel, close to the Mlambomkulu River Waterfall. It has yielded an assemblage of lithic tools, predominantly made from hornfels (metamorphically altered sandstone), as well as Plant remains which have been used to demonstrate that all the vegatation types found in the area have been present since the End Pleistocene. More than 17 000 items have been recovered from the site to date, including tools, Plant, and Animal remains.

Two broad stratigraphic layers, termed Stratigraphic Aggregates, or 'StratAggs', have been determined from the rock shelter. These are somewhat homogenised internally, but apparently distinct from one-another. The oldest is the Light Brown Coarse Sands, which has been dated to between 37 600 and 12 500 years before the present. Overlying this is the Shell-Rich Clayey Sands, dated to between 11 000 and 10 500 years before the present, both of which can be sub-divided into a number of discrete sub-units, termed sub-Stratigraphic Aggregates, or 'SubAggs'.

The ages and locations of SubAggs at Waterfall Bluff. Oster et al. (2024).

The material used in Oster et al.'s study was excavated in 2016 under permit from the Eastern Cape Provincial Heritage Resources Authority, and with the support of King Zanozuko Tyelovuyo Sigcau, Nkosi Mthuthuzeli Mkwedini, and the Lambasi AmaMpondo community. Specimens were compared to reference material in the collection of the Ditsong National Museum. Where Bovid remains could not be identified they were split into four size classes, with (1) being the smallest, and including species such as Oribi, and (4) is the largest, including species such as Eland. Where the species could be determined, ungulates were categorised as grazers, browsers or mixed feeders.

The majority of the specimens came from the Early Holocene Shell-Rich Clayey Sands, with a smaller amount from the Late Pleistocene Light Brown Coarse Sands. The majority of the material was extremely fragmentary, typically less than 2 cm on the longest dimension. About 10% of the material showed signs of Human modification, such as cut or percussion marks, though only one of these modified bones could be assigned to species level, a charred astragalus with cut marks from the Early Holocene determined to come from a Blue Duiker.

The most common Animals in the sample were Bovids (the group that includes Antelopes), while the most easily identifiable remains were those of Rock Hyraxes, Procavia capensis. A single Seal-tooth was found in a Pleistocene layer dated to between 22 560 and 19 430 years before the present. This was determined to have come from a Leopard Seal, Hydrurga leptonyx, a species which today is found on Antarctic and sub-Antarctic islands, and the waters which surround these, although occasional visitors to South Africa are recorded. The presence of such a tooth in a Pleistocene layer at the Waterfall Bluff Rock Shelter could imply that during this period, when sea temperatures would have been about 3° lower that today, Leopard Seal colonies were found further north, on the South African coast, or simply that individuals occasionally visited the area, as they do today. Three Seal bone specimens were also found, a vertebra and two bone fragments, from both the Pleistocene and Holocene layers, although these could not be identified more precisely.

Leopard Seal tooth from Waterfall Bluff Rock Shelter (#CN47208, Lot 303). Oster et al. (2024).

Only two Bovid specimens could be identified from the Late Pleistocene deposits, a Common Duiker, which is an obligate woodland species, and an Eland, which is environmentally adaptable. No obligate grassland species could be identified, but with the very small sample size, it would be problematic to place much emphasis on this.

Both Fish and Seal remains were found in Late Pleistocene layers along with marine Shelfish and possible Barnacle fragments, suggesting that the area was within practical range of the coast even during periods of glaciation. This is not completely surprising as the coastal shelf is narrow on this part of the coast (i.e. at the edge of the continental shelf the seafloor is steep, so that the sealevel can drop a long way vertically without moving far horizontally), although even at this time it was probably about 8 km from the rock shelter.

In the Early Holocene, Bushbuck and Reedbuck were both common. Since these species favour riverine environments, suggesting that the Mlambomkulu River had continued to run at this time; this supports earlier findings of incisions on the (now submerged) continental shelf, which also support the presence of a river. Reedbuck also favour grassland environments, as do African Buffalo and Bontebok/Blesbok, which were also fairly common in the Holocene layers, suggesting a grassland environment was definitely present by this time. The presence of Blue Duiker and Vervet Monkeys, which are obligate woodland species, indicate that woodland was still present, hinting at an environment not much different from today.

See also...

Genetic analysis of individuals from the Oakhurst Rockshelter suggest 9000 years of genetic isolation in South Africa.

Modern Southern African populations contain genetic diversity which records the deepest branching events known in the genetic history of extant Humans. The region also has a long archaeological record, with archaic Homo sapiens first appearing here around 260 000 years ago and Anatomically Modern Humans around 120 000 years ago. Because of this, the ancient Human populations of Southern Africa have been the subject of numerous archaeological and palaeogenomic studies, by scientists hoping to gain insights into population structures during the later stages of Human evolution. However, this interest in the most ancient Human genomes recoverable has left somewhat of a gap in the study of more recent, Holocene populations.

The Holocene has seen significant changes in technology and culture within Southern Africa. During the last 2000 year new populations have migrated into the area, bringing with them pastoralism and crop-farming. This began with the arrival of herders from East Africa, and was followed by farming populations from West Africa, who also brought the Bantu language group to the region. As well as  setting up  new communities with new ways of living, both of these groups contributed to the genetic structure of the original populations, so that all  extant San and Khoe populations draw at least 9% of their genetic material from outside Southern Africa. 

In a paper published in the journal Nature Ecology & Evolution on 19 September 2024, Joscha Gretzinger of the Department of Archaeogenetics at the Max Planck Institute for Evolutionary Anthropology, Victoria Gibbon of the Division of Clinical Anatomy and Biological Anthropology at the University of Cape Town, Sandra Penske also of the Department of Archaeogenetics at the Max Planck Institute for Evolutionary Anthropology, Judith Sealy of the Department of Archaeology at the University of Cape Town, Adam Rohrlach, also of the Department of Archaeogenetics at the Max Planck Institute for Evolutionary Anthropology, and School of Computer and Mathematical Sciences at the University of Adelaide, Domingo Salazar-García of the Department of Geological Sciences at the University of Cape Town, and the Departament de Prehistòria, Arqueologia i Història Antiga at the Universitat de València, and Johannes Krause and Stephan Schiffels, again of the Department of Archaeogenetics at the Max Planck Institute for Evolutionary Anthropology, present the results of a study in which they obtained genetic samples from a series of individuals from the Oakhurst rockshelter in South Africa, and compared these to other genetic samples from historic and living populations in South Africa.

The Oakhurst Rock Shelter is located 7 km from the southern coast of South Africa, close to the town of George in Western Cape Province. It was first excavated in the 1930s, and has yielded a remarkable sequence of archaeological remains, now known to represent about 12 000 years of accumulation. The Early Holocene layers here have yielded an assemblage of macrolithic tools which has been named the 'Oakhurst Complex' in reference to the site, which has been discovered at many sites across South Africa. Around 8000 years ago, this Oakhurst Assemblage was replaced by a set of microlithic tools, which have been named the Wilton Assemblage, which persisted throughout the remainder of the Middle and Late Holocene, with minor variations. Around 2000 years ago, ceramics also begin to appear at the site. 

As well as the numerous cultural artefacts, the Oakhurst Rock Shelter has also yielded 46 sets of Human remains, adult and juvenile, deposited throughout the archaeological sequence, including the oldest dated set of Human remains to have yielded DNA in South Africa, which are 10 000 years old. Gretzinger et al. obtained genetic material from 13 individuals from the Oakhurst Rock Shelter, all of which have been radiocarbon dated from their bone collagen, yielding ages of between 10 000 and 1300 years; nine of these dates are from previous studies, while four are new dates obtained by Gretzinger et al.. The generic sex was determined for all thirteen individuals, with the mitochondrial haplogroup obtained for nine individuals and the Y chromosome haplogroup for five.

Because mitochondrial DNA is found in the mitochondria, organelles outside the cell nucleus, it is passed directly from mother to child without being sexually recombined each generation, enabling precise estimations of when individuals shared common ancestors, at least through the female line. It is also possible to trace direct ancestry through the male line, using DNA from the Y chromosome, which is passed directly from father to son without sexual recombination.

Gretzinger et al. next created a haplotype population tree including ancient DNA from the nine Oakhurst individuals with mitochondrial DNA haplotypes, as well as samples from other archaeological sites in Africa, and modern populations. Most of the samples used were from previous studies, and are publicly available, however, some of the sequences were obtained from San skeletal material held by the University of Cape Town, and used only with permission of the San communities from which they were obtained. Access to this data is only available to other researchers with the permission of the University of Cape Town Skeletal Repository Committee and the relevant San communities. 

This recovered the Oakhurst individuals as being on the deepest branching limb of the living Human tree, which also includes living San populations, but closest to other ancient individuals from South Africa than to any living population. They also note that they recovered an ancient divide between San populations living north and south of the Kalahari, and that all the ancient South African populations, including the Oakhurst individuals, are on the same branch as the San populations from south of the Kalahari.

Maximum likelihood tree showing genetic affinities between ancient and present-day southern Africans, generated using TreeMix of genome sequences from present-day and ancient populations, excluding populations with evidence of asymmetrical allele sharing with non-Africans indicative of recent gene flow. Branches of ancient individuals/groups are truncated for better readability. Gretzinger et al. (2024).

Looking at the wider genomes, Gretzinger et al. found that San and Khoekhoe populations split into three principle groups, with the Kx`a-speaking Ju|’Hoan and !Xuun forming a northern cluster, Khoe-Kwadi-speaking Nama, and Tuu-speaking ‡Khomani and Karretjiemense forming a southern cluster (Karretjiemense is an Afrikaans word meaning 'people of the cart', but is how these people self-identify), while the Tuu-speaking Taa, Kx`a-speaking ǂHoan, and Khoe-Kwadi-speaking Gǀui and Gǁana form a central group. Eight of the Oakhurst individuals lie within the southern cluster, as do four other Later Stone Age skeletons with published genomes from South Africa, although the oldest individual in the dataset show a slightly greater affinity for the northern cluster. Notably, within the southern cluster, the Oakhurst individuals showed the greatest affinity to populations still living close to the area today, with the youngest individual, OAK007, dated to 1344 years before the present, showing the greatest affinity for living populations, sharing more and longer identical by descent segments with the Karretjiemense and ‡Khomani than with any other tested population.

Comparison of the genomes of the Oakhurst individuals to other, previously published, ancient African genomes, Gretzinger et al. found that all South African Later Stone Age genomes were closer to one-another that to those of any other ancient African. The youngest individual, OAK007, was most closely related to two other Later Stone Age individuals, from St. Helena and Faraoskop, both of which have been dated to about 2000 years before the present. Together, these three individuals form a sister group to two further individuals from Ballito Bay on the eastern coast of KwaZulu-Natal, thought to be of similar age. Older genomes from Oakhurst become steadily less closely related to these individuals as they get older, as well as less closely related to the genomes of historical San samples from Sutherland, Western Cape Province. However, the genome of a 1200-year-old pastoralist from South Africa clustered with Later Stone Age genomes from Malawi, while those of four Iron Age farmers from South Africa clustered most closely with Later Stone Age genomes from Cameroon. 

Next Gretzinger et al. looked for potential ingression of non-San genetic sequences into the Oakhurst individuals, finding no trace of affinity to populations in either East of West Africa, and a consistent grouping with southern rather than northern San groups, the genetic gulf between which groups appears to have been widening steadily since their split around 20 000 years ago, before the drying of the Lake Makgadikgadi palaeo-wetland, which once covered most of central Botswana.

All of the Oakhurst individuals dating to between 10 000 and 1344 years before the present, form part of a single clade (group with shared common ancestry), which also includes individuals from St. Helena, Faraoskop, and Ballito Bay dating to between 2200 and 1300 years before the present. However, the genomes of individuals from South Africa from between 1300 and 1200 years before the present show a significant discontinuity with earlier individuals, with a second discontinuity observed between 1200 and 400 years before the present. Gretzinger et al. attribute these discontinuities to the influxes of first pastoralists from East Africa and then farmers from West Africa into the region. However, they find no trace of West African ancestry in three individuals from Sutherland dating to the late nineteenth century, while about 11% of their genome appears to be of East African ancestry, a proportion similar to that seen in living ‡Khomani individuals from the Northern Cape Province, who typically have genomes comprising about 9% East African genetic material.

Based upon this, Gretzinger et al. observe that no evidence of any genetic influx from outside of modern South Africa recorded at Oakhurst Rock Shelter between 10 000 and 13 000 years before the present, a remarkable period of genetic continuity lasting almost 9000 years. Despite this, the Oakhurst individuals show no signs of being genetically isolated, The level of conditional nucleotide diversity (the  extent to which each member of a pair of chromosomes differs from its partner, used as a measure of inbreeding within a population( maintained within the Oakhurst samples is lower than that found in  Later Stone Age population from Malawi, Kenya and Cameroon, but comparable to other Later Stone Age populations from Western Cape and KwaZulu-Natal, and greater than is seen in ancient hunter gatherer populations from Serbia, Japan, and Brazil, as well as modern San and Khoe populations. This is non consistent with a model of long-term isolation, instead indicating to the presence of a much larger population of Later Stone Age hunters in South Africa before about 1300 years before the present, when other groups are generally accepted to have begun to arrive in the region, and a subsequent dramatic reduction in the size of that population.

Reconstructing the demographic history of South Africa over the past 2000 years is complicated, with at least two significant prehistoric population influxes, and substantial genetic exchange with both other parts of Africa and other continents following the establishment of the first European settlements in about 1650. To try to address this, Gretzinger et al. created a model using genomes from Later Stone Age hunter-gatherers in South Africa, the Luxmanda archaeological site in Tanzania, which has been dated to about 3000 years before the present, and modern Mende populations from West Africa.

They then developed a best-fit model which enabled them to group populations into primarily West Africa or Primarily East African (excluding populations with a substantial amount of genetic material from both sources), in order to estimate dates for the admixtures of the West and East African components. They found that San and Khoe populations began to absorb genes from East Africa substantially before those from West Africa, with an estimated date of 1068 years before the present. This is consistent with the East African ancestry recovered the 1200-year-old pastoralist remains from Kasteelberg, on the southwest coast of South Africa near St. Helena Bay, and the estimated date of admixture of 1228 years before the present recovered from the nineteenth century Sutherland material. 

The arrival of West African genes in South Africa appears to have been considerably more recent, with living Bantu-speaking groups such as the Herero, Tswana, and Kgalagadi, producing an estimated admixture date around 808 years before the present, while 400-year-old remains attributed to Iron Age farmers from KwaZulu-Natal yielded an estimated admixture date around 832 years before the present. Living San and Khoe groups yielded a more recent estimated admixture date, of about 578 years before present. Gretzinger et al. suggest that this may reflect either several waves of West African arrivals, or a continuous flow, with an initial admixture of San and Khoe genetic material into the ancestors of modern Bantu-speaking groups and a subsequent flow of West African genes into the ancestors of modern San and Khoe populations.

All groups show considerably more Later Stone Age ancestry on their X chromosomes than on their autosomal (non-sex determining) chromosomes, with this signal being stronger in San and Khoe populations than the Bantu-speaking groups. This implies that in most cases, the contribution from Later Stone Age hunter-gatherers was from the female side. The extent to which this is true appeared to vary between living populations, with the living Damara (a Khoekhoe-speaking people from northwestern Namibia) having had about 1.4 female Later Stone Age hunter-gatherers in their ancestry for each male, the ǂHoan (a Kxʼa language-speaking group from Botswana) having about 2.28 Later Stone Age hunter gatherer females per male in their ancestry, the Shua (a Khoe-speaking group from central Botswana) having about 4 Later Stone Age hunter gatherer females per male in their ancestry, the Haiǁom (a Khoekhoe speaking group from Namibia) having about 5.2 Later Stone Age hunter gatherer females per male in their ancestry. This also applies to South Africa Bantu-speaking groups (for whom the overall contribution of Later Stone Age hunter gatherer genetic material is lower), with about 2.1 females per male having contributed genetic material to the extant population. 

This female bias can also be seen in the historical Sutherland genomes and the 1200-year-old pastoralist remains from Kasteelberg, although, surprisingly, not to the four Iron Age KwaZulu-Natal individuals, who have a higher proportion of Later Stone Age hunter gatherer genetic material on their autosomal chromosomes than on their X chromosomes, indicating a higher proportion of male Later Stone Age hunter gatherer ancestors than female ones. This is different to the situation seen in all other groups in South Africa and Botswana for which a trend could be determined, and may reflect a change in the way different groups were integrating in the past 400 years compared to the nature of such interactions during the arrival of the first farmers into the region.

Gretzinger et al. finally note a recent admixture of male northwest European DNA into San/Khoe and mixed groups from Colesberg and Wellington. The estimated date for these ingressions is 199 years before the present, despite the known arrival of Dutch and British migrants into the region from the mid-1600s onwards, something which led to a collapse in San and Khoe genetic, linguistic and cultural diversity. In addition to severely disrupting existing population structures, the European arrivals introduced a range of new populations into the region, all of which have contributed to modern population structures to some extent. As an example, Gretzinger et al. note that mixed-ancestry South Africans from Colesberg drew an average of 24.4% of their ancestry from South Asia, 2.8% from East Asia, 8.2% from Northern Europe, and about 35.5% from Later Stone Age hunter gatherers. Some San and Khoe groups also have a significant proportion of European ancestry, with the Karretjiemense drawing an average of 5.61% of their ancestry from Europe, the ‡Khomani on average 9.45%, and the Nama on average 6.83%. This suggests that southern San populations were particularly affected by intermixture with Europeans, with these groups having a higher proportion of European ancestry that other San or Khoe populations, comparable to that of the mixed-ancestry South Africans sampled. Thus, the modern populations most closely related to the Oakhurst individuals appear to be particularly affected by genetic ingression from other populations.

Demographic changes in the San and Khoe populations of southern Africa: Summary of the inferred population history of the San and Khoe in southern Africa. Sex symbols indicate male- and female-biased reproduction. Note that pastoralism and farming both appeared in present-day South Africa at about the same time, 2,000 years ago. Gretzinger et al. (2024).

The question of population continuity within Later Stone Age communities in Southern Africa has engaged archaeologists for over a century. In the past two decades, the application of genetic methodology to archaeological problems has helped to unravel the demographic histories of Stone Age populations in Europe, Asia, and North Africa, revealing episodes of large-scale migration in these regions, during which indigenous populations were either replaced by or absorbed into the new population. These biological replacements of populations also appear to have been vectors for the spread of new technologies. In South Africa, in contrast, there appears to have been a surprisingly long period of genetic continuity, with on detected influx of genetic material from elsewhere for at least 9000 years, from the beninning of the Holocene till around 1200 years ago, during which time the Southern San remained isolated from Northern and Central San populations as much as from other populations elsewhere in Africa.

This implies that the cultural changes seen at the Oakhurst Rockshelter, such as the transition from the Oakhurst to the Wilton technocomplex, were a result of entirely local inovation. It has previously been observed that there have been slight fluctuations in craniofacial size in Later Stone Age populations in coastal South Africa, something which has been interpreted as a sign of genetic discontinuity, something which Gretzinger et al.'s results contradict. Since the population was not a small isolated one which might be subject to strong effects from genetic drift, it seems likely that these variations were driven by changes in the local environment.

The 9000 years of genetic and cultural isolation experienced by Later Stone Age hunter gatherers in South Africa seems to have ended quite abruptly, with the spread of herding communities from East Africa shortly followed by the arrival of farming communities from West Africa. Most parts of South Africa do not record any genetic trace of these arrivals before about 1300 years ago. However, there is evidence for changes in settlement patterns and other cultural behaviours in this coastal South Africa from about 2000 years ago, which have been interpreted as a response to the arrival of herding in the area. In Europe, a similar cultural shift is seen at the Neolithic-Mesolithic boundary, with a genetic admixture between the incoming farming population and the extant hunter gatherer population not being recorded for about 2000 years after the cultural shift. This implies that in Europe at least, farming and hunter gatherer populations were able to live alongside one-another for a long period of time before beginning to intermix, something which may also have been true in South Africa. Alternatively, pastoralism may have been culturally transmitted from East Africa to Southern Africa long before the spread of East African populations into the region.

However, from about 1200 years ago onwards, there has been substantial migration into Southern Africa from other regions, and substantial ingression of new genetic material into all populations, with the effect that all living San and Khoe populations are admixed with one or both of East African Pastoralist and West African Farmer ancestry. This process was accelerated by the arrival of European settlers in the mid-seventeenth century, which led to widespread population collapse among hunter gatherer populations in Southern Africa. Combined with a loss of oral traditions, these events have greatly obscured the prehistoric population structure of southern Africa. 

Genetic methods such as those used by Gretzinger et al. provide a way to study these ancient population structures, showing that the San and Khoe inhabitants of South Africa are the direct decendants of the Early Holocene inhabitants of the region, despite considerable disruption to their lifestyle and population structure by later migrants to the region.

See also...

The oldest known rock art in Patagonia, and what it tells us about the people who made it.

Patagonia lies at the southern tip of the Americas, and was one of the last areas to have been settled by Humans, who did not arrive there until the end of the Pleistocene. The area has a unique environment, which would have presented challenges to the people settling there. It also has a climate particularly suitable for the preservation of archaeological remains, making it particularly interesting to archaeologists. The area has extensive rock art, but to date little of this has been accurately dated.

In a paper published in the journal Science Advances on 14 February 2024, Guadalupe Romero Villanueva of the Instituto Nacional de Antropología y Pensamiento Latinoamericano, Marcela Sepúlveda of the Department of Social Sciences at the Universidad de Tarapacá, José Cárcamo-Vega of the Laboratorio de Espectroscopía Vibracional at the Universidad de Chile, Alexander Cherkinsky of the Center for Applied Isotope Studies at the University of Georgia, María Eugenia de Porras of the Instituto Argentino de Nivología, Glaciología  y  Ciencias  Ambientales, and Ramiro Barberena of the Centro de Investigación, Innovación y Creación at the Universidad Católica de Temuco, and the Instituto Interdisciplinario  de  Ciencias Básicas at the Universidad Nacional de Cuyo, present dates for four pieces of rock art from the Cueva Huenul 1 archaeological site in Neuquén Province, in the northern part of Argentinian Patagonia, and discuss the implications of this for the early peopling of the region.

The Cueva Huenul 1 archaeological site is located a kilometre above sealevel, to the east of the Andes amid the inland deserts of northern Patagonia. The Andes present a major topographic barrier, preventing the prevailing westerly winds from carrying moisture from the Pacific to the South American Arid Diagonal, where the Cueva Huenul 1 site is located. Precipitation in the region is between 150 mm and 200 mm each year, 75% of which falls in the winter. Rainfall tends to be higher in the west, closer to the Andes, and dryer to the east. with more abundant vegetation in areas with higher rainfall.

Location  of  Cueva Huenul 1,  other  sites  with  rock  art  in  northern  Neuquén  Province (Argentina),  and  palaeoecological  sites  from  northwestern  Patagonia. María Eugenia de Porras in Villanueva et al. (2024).

The Cueva Huenul 1 site is a cave with a habitable area of 620 m³, formed by the erosion of ignimbrites of the Tilhué Formation beneath an overlying basalt layer of the El Puente Formation, which have not eroded and now form the ceiling of the cave. Excavations within this cave have produced over 5500  lithic artifacts, principally flaked stone objects, and 8800 bone specimens, mostly Guanaco, Lama guanicoe. The site also has a long, and well-defined dating sequence, spanning 12 000 years, with a number of discrete phases of activity identified. The microenvironment within the cave appears to have remained stable and dry over this period, allowing for the excellent preservation of items such as Animal dung, and plant remains. 

Cueva Huenul 1 environment and landscape. (A) emplacement of Cueva Huenul 1 (yellow arrow) in a volcanic landscape within the Monte desert. (B) to( D) Views of the cave’s geology and topography. (E) View from Cueva Huenul 1 of the volcanic landscape of northwestern Patagonia. Guadalupe Romero Villanueva in Villanueva et al. (2024).

Notably, the Cueva Huenul 1 site has yielded a remarkable quantity of art-related materials, including  perforated  shell  beads, decorated Guanaco bones, and pyro- engraved  gourds. Also found inside the cave was a pit-structure containing a large number of twigs from the desert shrub Senna  aphylla, which have been stained with red ochre. A large amount of pigments, of various colours and in varying states of preparation.

Cueva Huenul 1 site plan and special findings. (A) excavation units at ch1. (B and C) General and detailed view of pit structure filled with vegetal remains of Senna  aphylla stained with red ochre. (D) Pyro-engraved gourd. (E) Perforated shell bead. (F) decorated guanaco (Lama guanicoe) bone. (G) Pigments. each image has an individual metric scale. Guadalupe Romero Villanueva and Ramiro Barberena in Villanueva et al. (2024).

The site has a panoramic view of the surrounding landscape, although it is not visible from any other known archaeological site in the region, nor can any such site be seen from the cave. Other sites in the region include a series of smaller caves and rockshelters, most of which only appear to have been used within the last 2000 years. Some of these, such as El Ciénego and Paso de las Tropas, also have rock art, although it is less diverse in technical style and less formalised than the art at Cueva Huenul 1. The Cueva Yagui site, to the north of Cueva Huenul 1, also records a long timeline, in this case about 8 500 years, and appears to have been more intensely occupied, on the basis of stone tools, abundant ceramics, and faunal remains, with this occupation being particularly intense over the past 2000 years. Both the stratigraphic sequences and the styles of rock art suggest that Cueva Huenul 1, Cueva Yagui, and other sites appear to have been linked.

Cueva Huenul 1 hosts one of the most impressive collections of rock art in northwestern Patagonia, with central portion of the cave’s internal wall and part of the ceiling covered by 895 discrete pieces of rock art, which have been grouped into 466 identifiable motifs. Most of these motifs are geometric shapes rather than pictures, with strokes, dots, circles, and lines being common, and parallel lines, reticulates, polygons, and cruciforms also present, as well as some Human silhouettes and a face, and silhouettes of Guanaco and Choique, Rhea pennata, and some representations of dynamic group  activities. A range of colours are used in the cave art, although a haematite-derived red is the most common, along with different hues of white, yellow, and black.

Examples of the rock art of Cueva Huenul 1. Each tracing has a 10cm scale bar. Guadalupe Romero Villanueva in Villanueva et al. (2024).

A large number of motifs are superimposed over earlier artworks, and there appear to be three distinct degrees of weathering, as well as distinct artistic phases. The majority of the art is presumed to be of Late Holocene origin, based upon similarities to styles of art used at other localities. However, the long history of occupation at the site combined with presence of a clear artistic sequence, raises the possibility that some art at the site may be much older.

Four artworks at the Cueva Huenul 1 site were chosen to be dated. These were all classified as 'comb-shapes' based upon a shared basic morphology of a perpendicular horizontal line with several parallel vertical lines extending downwards from it. Within this simple pattern, however, the comb motifs can be divided into simple and complex forms. Three of the motifs examined, UT3- M48, UT5- S4- M7, and UT3- M37, were of the simple type, while the fourth, UT5-S2-M19, was of the complex form. All of the comb motifs are executed in a reddish black pigment. Most are isolated from other artworks, although one (UT5-S2-M19) is part of a complex series of superimposed images from different periods. 

Dated rock art paintings from Cueva Huenul 1. (A) Motif Ut3- M37. (B) Motif Ut3- M48. (C) Motif Ut5- S4- M7. Guadalupe Romero Villanueva in Villanueva et al. (2024).

By careful examination of the motifs, Villanueva et al. were able to establish that there was no potentially contaminating organic matter  either  on,  within,  or  below  the  paint  layer. Three layers were found to be present, the bedrock, the pigment layer, and an overlying layer of translucent particles forming a patina. The pigment layer was identified spectrographically as amorphous carbon; this was found not to contain any significant amount of phosphates, making it unlikely it was derived from carbonized bone.

Dated rock art motif UT5- S2- M19 from Cueva Huenul 1. (A) Original photograph and digital enhancement with DStretch of the complete rock art panel. (B) Original photograph and digital enhancement with DStretch of the dated black comb-shaped motif. (C) digital tracing of the complete rock art panel showing the dated black comb-shaped motif underlaying a series of superimpositions. Guadalupe Romero Villanueva in Villanueva et al. (2024).

Further examination of the samples by Scanning Electron Microscopy and Energy Dispersive X-Ray Analysis demonstrated the presence of plant cells within the pigment layer of all four examined motifs, as well as a composition consistent with a carbonaceous material mixed into an aluminosilicate matrix (i.e. a mixture of charcoal and clay), with the overlying layer of material rich in calcium and sulphur, probably indicating some form of salt. The precise origin of the wood used to make the charcoal was impossible to determine, although it is likely to have been one of the woody shrubs known to have been growing in the region in the Middle Holocene, such as Prosopis spp., Larrea sp., or Schinus sp..

Cross section microphotography of sample CH1-AMS1 embedded in resin showing three differentiated layers. From the bottom, the layers which can be distinguished are; the bedrock support, the black pictorial layer, and a thin layer of patina or varnish. Marcela Sepúlveda and Guadalupe Romero Villanueva in Villanueva et al. (2024).

Based upon this, Villanueva et al. conclude that the black pigment was formed by the incomplete burning of Plant matter, something which should lead a carbon¹⁴ signature. Since there are no long-lived trees in the region, material from which can give misleading date information, the charcoal can be assumed to have come from a short-lived woody C₃ shrub and/or a Cactus (which have their own Crassulacean acid metabolism resulting in a distinctive carbon isotope signature). It was possible to recover sufficient carbon from three of the motifs to be confident that the isotope signature recovered was accurate, while one sample, taken from motif UT5-S2-M19, yielded a much lower amount of carbon, raising concerns that contamination from later sources (this is the motif which is partially overlain by later artworks), leading to this data being excluded from the remainder of the study. The remaining three motifs were found to be between 7728 and 7565 years old (UT3- M48), between 6271 and 6239 years old (UT5- S4- M7), and between 5643 and 5629 years old (UT3- M37).

As well as the dates obtained for the rock art motifs, Villanueva et al. obtained 16 dates from archaeological remains at the site, in order to build up a stratigraphic sequence. This led them to conclude that there had been four stages of occupation at the site, over a period of about 18 000 years. 

The first phase is calculated to have lasted approximately 4683 years, from about 17 407 to about 12 934 years before the present. During this phase the cave was occupied by Giant Sloths, with no signs of Human activity. 

The second phase is calculated to have lasted approximately 1620 years, from about 11 721 to about 10 162 years before the present, and shows evidence of the first Human activity in the area, including Guanaco bones with cut marks, hearths with charcoal, and a grass bedding structure. 

There is then a significant hiatus in activity, with the third phase starting about 8171 years ago and lasting for approximately 3246 years, till about 5074 years before the present. This phase includes the emergence of rock art at the site, with all three dates obtained for the comb motifs falling within this interval, as well as the ochre covered Plant remains. Assuming that the average Human generation time was about 25 years, this would suggest a cultural tradition using similar symbolism which lasted for about 130 generations. 

The final phase of activity at the cave includes much more intensive activity, includuing the majority of the rock art, as well as cultural similarities to other sites in the region, and spans about 1500 years in the Late Holocene.

Climatically, the area had a sharper east-west variation in moisture during the Late Pleistocene and Early Holocene, until about 10 000 years ago, with the western Andes being wetter than today, while the eastern Andes were drier. The region to the east of the Andes, inclding the Cueva Huenul 1 site, was significantly more arid than today between about 10 400 and 9 400 years ago. From about 10 000 years ago onwards both the Andes and Eastern Patagonia became extremely arid.

The area reached peak aridity in the Middle Holocene, with widespread deserts and only patchy, fragmented areas of habitable land, which could have acted as stepping stones for the first Humans entering the environment. Large areas would have either too dry for occupation, or too unstable to be entered on more than a temporary basis. Nevertheless, Humans did enter the landscape during this time, probably relocating frequently, and needing to maintain social contact over large distances, while at the same time coming up with innovative technologies for subsistence.

By assembling a comprehensive database of radiocarbon dates for Human activity across the South American Arid Diagonal region, Villanueva et al. conclude that during the period 14 000 to 10 400 years before present the first Human population appeared and rapidly grew, expanding to occupy new niches. From about 10 800 to 7000 years ago a period of cultural stasis appeared, combined with a slowly declining population. The oldest rock art at Cueva Huenul 1 is slightly less that 8000 years old, coinciding with the later part of this period of apparent cultural stasis. The population is also thought to have remained fairly static or shrank across South America between about 9000 and about 5500 years ago. 

This suggests that during the Middle Holocene northwestern Patagonia was probably home to a small and scattered population of highly mobile hunter-gatherers, coping with an extremely arid climate with occasional wetter spells. This population was static or shrinking slightly, widely scattered, and having to cope with frequent but unpredictable extreme weather conditions.

The dating of the comb motifs in the rock art of Cueva Huenul 1 gives a date for the origins of rock art in northwestern Patagonia. The repeated nature of these motifs makes it unlikely that these marks were random, with similar marks being repeated several times over a period of about 3000 years, suggesting it was linked to a system for passing information between generations.

Transmission of knowledge can become linked to particular sites, which eventually become key locations for a culture, where people meet to re-enforce cultural identities and maintain extended social networks. Villanueva et al.  suggest that Cueva Huenul 1 first became such a culturally important site during the Late Pleistocene, being used regularly over a period of about 1400 years across the End Pleistocene and Early Holocene. Visits to the site continued into the Middle Holocene, though the behaviour of the visitors changed, with the appearance of activities such as marking the walls. During this phase, there is little sign of non-ritual activities, such as food-processing or tool-making. Villanueva et al. suggest that the transition of the space to a ritual centre where these profane activities were not carried out probably implies that the site was not, as previously assumed, abandoned for long periods during the Middle Holocene, but rather underwent a change of purpose connected to its new, sacred status. 

Villanueva et al. suggest the emergence of sites such as Cueva Huenul 1 which helped hold widely scattered cultures together would have been key to Human survival in the arid landscape of northwestern Patagonia. The emergence or rock art was probably a way of re-enforcing knowledge transfer across generations, building upon an earlier oral tradition. 

The population of many areas in South America apparently struggled to cope with conditions in the Middle Holocene, as increasing aridity fragmented the available liveable spaces, leading to a slowly dwindling population. The first rock art at Cueva Huenul 1 coincides with this period, possibly aiding social cohesion and helping  people to survive a particularly harsh period, before populations began to recover between 7000 and 5000 years ago. 

The dating of the rock art at Cueva Huenul 1 gives an insight into the context in which such art first appeared in Patagonia. Here, a style of art appears around 8000 years ago and persists for over 3000 years. This happened at a time when the climate was much drier and less predictable than today, at a time when Human populations were at best maintaining stasis, and were probably suffering frequent crashes. It this marginal, sparsely populated environment the rock art apparently helped to preserve collective knowledge across multiple generations. 

See also...

A possible prehistoric hunting wall off the Baltic coast of Germany.

The seafloor is shaped by geological, biological, and to a certain extent anthropological forces, yet to a large extent is hidden from Human view. It has for some decades been possible to map larger seafloor structures with ship- borne multibeam echosounder systems, and to explore areas where there were known sites of interest with both submersible vehicles and Human divers. Nevertheless, huge areas of the seafloor remain effectively unexplored, hiding much information about the geology, biology, and even Human history of the planet.

In a paper published in the journal PNAS on 12 February 2024, Jacob Geersen of the Institute  of  Geosciences at Kiel University, and the Leibniz Institute for Baltic Sea Research Warnemünde, Marcel Bradtmöller of the Heinrich Schliemann-Institute of Ancient Studies and the Interdisciplinary Faculty, Department WKT at Rostock University, Jens Schneider von Deimling, also of the Institute  of  Geosciences at Kiel University, Peter Feldens, also of the Leibniz Institute for Baltic Sea Research Warnemünde, Jens Auer of the Landesamt für Kultur und Denkmalpflege Mecklenburg- Vorpommern, Philipp Held, Arne Lohrberg, and Ruth Supka, again of the  Institute  of  Geosciences at Kiel University,  Jasper Justus Lutz Hoffmann of the Department of Coastal Geology at the Alfred- Wegener- Institute, and the Department of Geosciences, Marine Geology and Seafloor Surveying Group at the University of Malta, Berit Valentin Eriksen of the Schleswig-Holstein State Museums Foundation Schloss Gottorf of the Centre for Baltic and Scandinavian Archaeology, Wolfgang Rabbel, again of the Institute of  Geosciences at Kiel University, Hans-Jörg Karlsen, also of the Heinrich Schliemann-Institute of Ancient Studies at Rostock University, Sebastian Krastel, once again of the Institute  of  Geosciences at Kiel University, David Brandt and David Heuskin of the Institute for the Protection of Maritime Infrastructures at the German Aerospace Center, and Harald Lübke, also of the Schleswig-Holstein State Museums Foundation Schloss Gottorf and the Leibniz  Centre  for  Archaeology of the Centre for Baltic and Scandinavian Archaeology, describe a linear stone structure at a depth of 21 m, in the Bay of Mecklenburg, about 10 km northwest off Rerik in the Rostock District of Mecklenburg-Western Pomerania, Germany, which they interpret as a pre-historic, Human-made wall.

The Bay of Mecklenburg was shaped during the Weichselian Glaciation (i.e. the most recent glacial interval), and since the retreat of the ice has been significantly uplifted by isostatic rebound (the weight of the ice pushed the solid lithosphere down into the liquid mantle, and when removed it bobbed up). At the same time, the sea level in the area has been calculated to have risen significantly, as water locked up in glacial ice was released, from about 40 m below its current level to about 20 m below, between 13 300 and 12 700 years before present, and from about 28 m below the current level to about 10 m below between 8570 and 8000 years before the present. Since the water in the Bay of Mecklenburg current maximum depth of 28 m, this implies an extended period between the retreat of the ice and the area being covered by the sea. The seafloor over much of the bay is covered by Holocene mud and sand (sandier towards the shore), overlying a glacial till, which is exposed in places.

The seafloor of the Bay of Mecklenburg has a high density of archaeological sites, many of which were documented by the Sinking Coasts project between 2002 and 2009/ Much of this work concentrated on the period after the Littorina Transgression (i.e. following the first establishment of a sea within the Baltic Basin), with 23 sites discovered which date to between 8500 and 5000 years ago, which currently lie in waters between two and eleven meters deep. Older sites are known from the shores of Scandinavia and the Eastern Baltic, but most of the known older sites in Germany are further to the south, away from the Baltic Coast. The putative wall discovered by Geersen et al. lies at a depth of about 21 m, and presumably therefore predates the Littorina Transgression.

During recent decades, the seafloor of the Bay of Mecklenburg has been extensively explored by hydroacoustic surveys.  In 2021 one of these surveys recorded a linear structure almost a kilometre long and slightly less than a metre in height. This structure, which Greersen et al. refer to as the Blinkerwall, is located on the southern, landward, side of a northeast–southwest trending bathymetric ridge. At its northeast end, the structure abuts a conical mound which rises to 13.5 m below sealevel. The Blinkerwall runs for 971 m, sinking from a depth of 21 m at the eastern end, to 21.5 m in the west. 

The study area in the Baltic Sea. (A) Overview map of the Western Baltic Sea. (B) Detailed structure of the Bay of Mecklenburg including the location of the Blinkerwall. Greersen et al. (2024).

The majority of the Blinkerwall is less than a metre high, and visits by divers and autonomous underwater vehicles have shown that it is made of a succession of individual stones. 1673 individual stones were identified, with a combined volume of 52.75 m³, and a calculated combined mass of 142 473 kg (based upon the assumption that the average density of the rocks is equivalent to that of granite). Most of the stones have calculated masses below 100 kg, but 288 were identified as likely to be heavier, with the largest, towards the centre of the wall, having a calculated weight of 11 389 kg. This rock marks a turn in the direction of the wall, which runs east-west  to the west of this rock, and southwest-northeast to the east. The next three largest rocks, with masses of 2083 kg, 2506 kg, and 5792 kg are all located towards the western end of the wall, with the largest one marking the end of the wall. The ten heaviest stones are all associated with slight changes in the direction of the wall.

Morphology of the southwest–northeast trending ridge that hosts the Blinkerwall and the adjacent mound. (A) Multibeam bathymetry collected with RV ALKOR (2021) and FK Littorina (2023). (B) Multibeam bathymetry collected with an autonomous underwater vehicle. (C) Side- scan image (bright colors equal high backscatter) collected with RV Elisabeth Mann Borgese in 2020. White arrows point at the Blinkerwall. Greersen et al. (2024)

The heights of the rocks were measured with several different techniques, producing consistent results, which increased Greerson et al.'s confidence in their findings. There are almost no stones of similar size within 10 m on the northward side of the wall, particularly towards its western extent, and the density of such stones is depleted compared to the rest of the bay further north than this, not reaching typical levels till about 50 m north of the wall. No stones of similar size could be found south of the wall.

Bathymetric scans show that the wall runs along a bathymetric ridge, with sedimentary basins to the north and south. Within the basin to the south there are two sedimentary units, the lower of which is 3,5 m thick and has an erosional upper surface, while the upper unit is about 2.5 m thick. A sediment core sunk in this area recorded an upper layer 1.8 m thick, which was comprised of Holocene mad, over a thin peat layer, then an eroded surface from which some wood fragments were recorded. The wood was dated to 10 578 years before the present, leading Greersen et al. to suggest that the eroded layer probably corresponds in time to the Yoldia to early Ancylus Lake lowstand, about 11 700 years ago; similar unconformities have been recorded at this time from other locations in the region. Another piece of wood, recovered from the upper layer of sediment, was dated to 9886 years before the present. Based upon this, Greerson refer to the upper layer as Holocene sediments and the lower layer as Baltic Ice Lake sediments.

North- south- oriented sediment echosounder profile across the east–west trending ridge that hosts the Blinkerwall and the sedimentary basin to the south. The erosional unconformity of the Yoldia to early Ancylus Lake lowstand land surface is located at 1 to 3 m depth in the basin and crops out slightly south of the Blinkerwall. The peat layer (dated to 10 662 to 10 494 years before present) that hindered a deeper penetration of gravity core is visible as a layer with low reflectivity located above the erosional unconformity. The latter is underlain by the sediments of the Baltic Ice Lake. Greersen et al. (2024).

The eroded layer reaches the surface slightly to the south of the Blinkerwall, with till sediments on the  ridge not covered by Holocene muds. This also occurs at the southern margin of the basin, towards the shore at Rerik and the islands of Wustrow and Poel. The bathymetric results suggest that at the Yoldia to early Ancylus Lake lowstand, if  the Blikerwall already existed, then it would have run along an east-west running promontory, connected to the German mainland to the southeast.  To the south of the wall would have been a 5 km wide basin with a maximum depth of 25 m.

The Blinkerwall is a remarkable  topographic feature, unlike anything else which has previously been detected beneath the Baltic Sea, located upon  reworded  till deposits  associated with the Weichselian  Glaciation. The Weichselian ice retreated from the area about 17 000 years ago, leaving an exposed till deposit, which was covered by floodwaters and then exposed several times before the final development of the Baltic Sea in the early Holocene. The Blinkerwall is currently 21 m below sealevel, but would have been exposed above the surface during the Baltic Ice Lake phase, around 12 800 years ago, and the Yolida Sea phase, about 11 700 years ago, both of which fall within the Younger Dryas interval. The wall would also have been exposed on the surface during the Early Holocene Ancylus Lake regression, about 9500 years ago, before finally being covered by the Baltic Sea during the Littorina trans­gression, between 8600 and 8000 years ago.

The deposition of Holocene muds has led to a thick sediment cover over much of the Baltic seafloor, burying many topographic features. The Blinkerwall lies upslope of a basin now largely infilled with stratified muds. Greersen et al. reason that the lowest sediment layers in the basin predate the Yoldia to early Ancylus Lake lowstand unconformity, and probably therefore date from the time of the Baltic Ice Lake or Baltic Ice Lake II. Exposure and erosion of the till layers would have led to the concentration of rocks upon the surface, although south of the Blinkerwall these are now largely covered by subsequent sediments. However, these processes are likely to have contributed to the presence of at least some of the rocks that make up the 971 m long Blinkerwall.

Autonomous underwater vehicle multibeam data from sections along the Blinkerwall. Greersen et al. (2024).

Stones can be moves and emplaced by natural processes. The largest tsunamis are capable of shifting boulders with diameters of several metres, and depositing them in strewn fields. No tsunami, however, has ever been recorded placing rocks in a neat line along a promotory, and it is difficult to conceive of one ever doing so. 

Glaciers are also capable of moving rocks. The retreat of the Weichselian Ice Sheet left a variety of landforms across northern Europe, which are still largely responsible for the landscapes seen today. These include a variety of moraines, linear structures caused by rocks and till being pushed into place by the movement of glaciers, but these take the form of ridges of unconsolidated debris, including gravels and clays as well as rocks and boulders. 

Eskers are linear structures formed beneath glaciers, from material deposited in streams running beneath the glacier. In the Baltic region eskers with high-stone compositions up to 7 m high and a few hundred metres in length have been recorded, and in other parts of the world, eskers up to 50 m high, 150 m in width, and several tens of kilometres long have been recorded. All eskers, however, are comprises of a mixture of sand, gravel, and rocks, and they do not contain neat linear arrangements of rocks. 

The deposition of material by floating ice has been cited as a cause for beach ridges elsewhere around the Baltic, although these are again generally made up mostly of sand and gravel. Furthermore, a large beach ridge would require a stable coastline over a long period of time, so that the material was always placed in the same position. The Blinkerwall is located at a depth of 21 m, and while beaches at 21 m below sealevel undoubtedly existed during the Younger Dryas, the sealevel is known to have fluctuated significantly over this interval, making it unlikely that a particularly large beach ridge would have built up during this interval.

The most plausible natural cause for the Blinkerwall would be ice thrusting associated with the Baltic  Ice Lake, about 13 000 years ago. This could conceivably have formed a shoreline at -21 m, and occurred at a time when the glacial till was not covered by significant amounts of sediments. However, a wall formed at this time would have had to endure several subsequent emersions in high-energy coastal waters during the subsequent transgressions and regressions of the sea, and it seems unlikely that the Blinkerwall, less than 2 m wide and made mostly of rocks weighing less than 100 kg, would have survived this as intact as it is. Nor is the form of the Blinkerwall consistent with such an origin, as such thrust ridges tend to be several, often tens of, metres wide, and contain significant proportions of sand and gravel, whereas the Blinkerwall is less than 2 m wide, and made entirely of large rocks and boulders.

Taken together, these lines of evidence suggest that it is highly unlikely that the Blinkerwall is a natural phenomenon, and that therefore it was most likely formed by the activity of prehistoric Humans.

The Blinkerwall is located at a depth of 21 m, in a location which would have been covered by the waters of the Baltic for a final time between 8570 and 8000 years ago, giving a minimum possible age for its construction. Preserved wood in the basin to the south of the Blinkerwall has been dated to 9100 years before the present, indicating that this basin must have flooded after that time. This is long before the appearance of agrarian societies in Europe, at a time when the Baltic Basin was known to have been inhabited by hunter gatherer societies living in small bands and forming non-permanent camps. The Blinkerwall is made up of 1385 smaller stones with masses of less than 100 kg, which could conceivably have been moved by small groups of Humans, connecting 288 larger stones, which it is unlikely that Humans of the time would have been able to move with the tools available to them. All of the rocks are emplaced on the southern flank of a ridge made of glacial till, where rocks of this size would have been readily available; the absence of such rocks in the zone to the north of the wall supports the idea that the rocks were sourced from close by and deliberately emplaced, and that most of them were moved downslope to get them into position. It has been estimated that about 1281 people were living in Germany and Poland during the latest glacial interval, in bands of 40-45 people. For a group of this size to construct a stone wall almost a kilometre in length would not be impossible, but the project would have had to have been of considerable benefit to them to have been considered worthwhile.

No similar structures are known from the region. Late Mesolithic communities are known to have made extensive use of coastal resources, including fishing with static wooden weirs. These structures would have been at most a few hundred metres in length, significantly smaller than the Blinkerwall, and are generally emplaced on rivers or on coastal sections with a strong tidal range, so while stone Fish weirs have been recorded in other parts of the world, it is unlikely that this was the intended purpose here. 

Groins for sea defence purposes have been recorded in the Neolithic of the Middle East, though the Blinkerwall appears too small to have served such a purpose. Nor is it likely to have formed part of an artificial harbour, since it predates the documented use of boats in the region significantly, making it likely that even if boats did exist, they would have been small structures easily dragged up a beach to safety. Furthermore, both a groin and an harbour wall are structures likely to be built only by people with a permanent claim to the land, which is contrary to our understanding of roaming Mesolithic bands of hunter gatherers, Rather, for such a structure to have purpose, the people who made it would have needed to have been able to leave it unattended for much of the year without any maintenance, and to have been confident it would have been ready to use at any time when they did return.

The purpose that Greersen et al. suggest for the wall is as a piece of hunting architecture, used to control the movements of herds of Ungulates as they were driven by hunters. While we do not today associate hunter gatherer societies with large architectural projects, recent studies in the Middle East, where an arid climate tends to favour preservation of such structres, have shown the presence of numerous long stone walls, and evidence of such stuctures has also been uncovered in western Greenland and the Great Lakes Region of North America. These structures vary in form, and often include features such as multiple drive lanes, enclosures for trapping Animals, and blinds where hunters could have hidden. In the Middle East these walls are presumed to have been used to hunt Gazelles or other Antelopes, and while such Animals would not have been present in the Baltic region, such a wall would have been equally effective in controlling the movements of Bison or Reindeer.

The dating of large stone walls has proven problematic, beyond stating that they are clearly prehistoric, although walls apparently used for hunting Reindeer and now submerged by changes in sea-and/or-lake level, such as the Blinkerwall and the 'Drop 45' structure on the Alpena-Amberley Ridge beneath Lake Huron in North America can be presumed to date from the Late Glacial/Early Holocene transition. The Blinkerwall and 'Drop 45' structures share a number of features, including a location near the top of a slope, but below the crest, a subparallel trending marsh/lakeshore on one side, construction on bedrock, and solid and continuous construction. The 'Drop 45' site is also presumed to be a drive lane used in the hunting of Ungulates, something which in this case has been supported by the recovery of numerous stone tools associated with hunting by divers, which were concentrated at points interpreted as hunting blinds. It is possible that the large erratic blocks which served as part of the Blinkerwall could also have served as hunting blinds, giving a potential target for future archaeological surveys.

In other ways, however, 'Drop 45' differs significantly from the Blinkerwall, comprising two stone lines which narrow towards a funnel through which the hunted Animals were presumably driven, and covering a total area of only about 30 m by 100 m. It is possible that another line of stones is present to the south of the Blinkerwall, which has now been covered by Holocene mud, or it is possible that the second limb of the trap was formed by the former lake in the basin to the south of the wall. Reindeer will enter water, and are fairly good swimmers, but are much slower in water than they are on land, which would make it easy to target them from nearby boats.

Most ancient drive lane traps are believed to have been placed on the migration routes of Ungulates, effectively allowing a herd to be targeted twice a year, and significantly increasing the yield of meat achieved in those hunts. The 'Drop 45' site is thought to have been associated with Caribou, and the most likely candidate for the target of the Blinkerwall would have been Eurasian Reindeer (which are the same species). Reindeer are known to have been targeted as a prey species since at least the Middle Palaeolithic, and would certainly have been a familiar prey species to any hunting population living in the Baltic region. Such hunters would have been familiar with the behaviour of Reindeer, and able to predict how they would react to a structure like the Blinkerwall. Furthermore, the Reindeer population of the period is thought to have migrated east-to-west, making a location on am east-west oriented ridge an ideal location for such a trap.

As the Holocene progressed, the area would have become more forested, changing the seasonal climatic patterns, and causing the Reindeer herds to shift their migrations northward. By around 9800 years ago, they would have been absent from the region, which implies that if the Blinkerwall was used for hunting Reindeer, then it is probably Pre-Boreal in origin, making it one of the oldest known hunting structures in the world, and possibly the oldest known megastructure in Europe. However, direct dating of such structures is notoriously difficult, as datable remains such as carcasses were generally cleared away by the site's users. The oldest known structure for which a direct date exists is a desert kite in Jordan, which has been dated to 10 000 years before the present. The Blinkerwall is potentially older than this structure, though direct dating of it is likely to be extremely problematic.

3D model of a section of the Blinkerwall adjacent to the large boulder at the western end of the wall. The scale bar at the Top- Right edge of the image is 50 cm. Philipp Hoy & Jens Auer in Greersen et al. (2024).

Based upon analysis of sedimentological, geophysical, and archaeological data, Greersen et al. suggest that the Blinkerwall dates from the End Pleistocene or very earliest Holocene. There is an absence of similar structures from this period in Europe, but that does not indicate that they were not being built, rather it is likely that they have simply not survived the to modern times on the densely populated European continent, or that other surviving structures have not been identified. In offshore environments, the most likely environment for such a structure to survive, the identification of walls is at the limit of capability of hydroacoustic surveys. Furthermore much of landscape which would have been exposed at the end of the Pleistocene has subsequently been covered by Holocene mud, so structures would only be visible if they were located on highgrounds which still remain exposed. There are, nevertheless, a number of areas beneath the Baltic where moraines associated with the Weichselian Glaciation are known to still be exposed, at depths of 20 m or more, providing locations which could provide further insights into Late Glacial and Mesolithic cultures during phases of rapid advance of the Baltic Sea.

See also...