Base map redrawn from Orquera 1987:Figure 1

Late Pleistocene and early Holocene hunter-gatherers of the Pampas and Patagonia, Argentina and Chile


The archaeology of the late Pleistocene and early Holocene of the Southern Cone of South America is a robust and fascinating record. Since the end of the 19th century, there have been hints at the rich Pleistocene paleontological record of the Pampas and Patagonia. Within the last 70 years, the archaeological record has shown to be just as remarkable, with major advances in the excavation and interpretation of the record of the Southern Cone. Hrdlicka once noted (1912:386) that his dismissal of skeletal remains from the Pampas of Argentina

...should not be taken as a categorical denial of the existence of early man in South America, however improbable such a presence may now appear; but the position is maintained, and should be maintained, it seems, by all students, that the final acceptance of the evidence on this subject can not be justified until there shall have accumulated a mass of strictly scientific observations requisite in kind and volume to establish a proposition of so great importance.

There is little doubt that enough evidence has been accumulated thus far to discuss the early inhabitants of the region. The purpose of this paper is to describe the notable late Pleistocene and early Holocene sites from the region, compare their assemblages, and make general conclusions about the organization of these hunter-gatherers of the Southern Cone. The study focuses solely upon the eastern side of the Andes, in the Pampas and Patagonia, as these regions are more environmentally similar to one another than they are to the western side of the Andes.
Climate zones redrawn from Orquera 1987:Figure 2

Topography and Modern Climate


The Pampas are a large and flat grassland that makes up the central and western portions of Argentina. Three small sets of hills provide topographic relief to the Pampas, the Tandilia, Ventania, and Lihué Cahel ranges.

The Pampas are usually defined as extending from the Rio Colorado on the south, east to somewhere between the Rio Salado/Rio Paraná and Uruguay (the Argentinean Mesopotamia), to the west towards the Andean foothills, and to the north towards the Chaco.

The climate of the Pampas follows an east to west gradient of decreasing moisture, with the eastern Pampas (in modern conditions) receiving approximately 800 mm of rainfall per year, whereas the grasslands to the west begin to turn to semi-arid grasslands with annual rainfalls of 200-400 mm (Orquera 1987:340). This precipitation gradient provides a "grassy treeless steppe in the east and a xerophytic scrub in the west" (Orquera 1987:340). This vegetation distinction gives rise to a dichotomy between a "humid pampa" to the east and a "dry pampa" to the west. This pattern of grassland variability is very similar to that expressed on the Plains of North America, where a tall grass prairie gives away to a short grass prairie on a decreasing precipitation gradient of east to west.

Surface water is reportedly scarce and not of the best quality on the Pampas. Several rivers and streams drain the Pampas from the north to south, however, and were likely heavily utilized during prehistory.

The Pampas are situated within a very large region of sand and loess accumulation (1.1 x 103 km2) that is known as the Pampean loess formation (Clapperton 1993:198). The formation has an average thickness of 30-40 m. The loess and sand are derived from volcanic eruptions, dry river beds, and the Altiplano to the west and northwest of the Pampas. Large sand dunes have also formed on top of this thick loess deposit, covering an area of 300,000 km2 in the southern part of the Pampas (Clapperton 1993:200-201).


The Patagonian steppe begins to the south of the Pampas, extending east to west from the Andes to the coast of the Atlantic Ocean, and north to south from the Rio Colorado to the Straits of Magellan. Patagonia is a dry semi-desert steppe, with rainfall averaging between 100 and 300 mm a year (Orquera 1987:341).

The terrain is much more uneven than the flat Pampas, made up of broken mesas and rolling terrain. Large portions of Patagonia are covered by gravel deposits, much of which are derived from glacial outwash (Clapperton 1993:480-485). These gravels likely provided abundant raw materials for hunter-gatherers in the region.

The most environmentally productive region of Patagonia is a transitional zone, 50 to 100 km wide, running along the foothills of the Andes (Orquera 1987:341). This area has an abundance of trees, and water is available from streams originating in the Andean uplands. Moving eastward towards the coast from this belt, however, the water availability decreases, as does arboreal vegetation. Along the coast, the cold Malvinas Current hampers vegetation growth, with minimal availability of trees (Orquera 1987:342). Salt flats are common along the coast, with little freshwater available.

Patagonia also has large numbers of paleolakes, many of which are dry today. Archaeological sites are located along the edges of these dry lakes, however, documenting fluctuating water levels throughout the Holocene (Clapperton 1993:509-511).

Southern Patagonia is more extreme in the area of the Magellan Strait and Tierra del Fuego, as the high latitudes are encountered. Towards the south, the precipitation begins to increase as the mean annual temperature drops (McCulloch et al. 1997:16). Moving to the west, and into the Andes, forests and Magellanic moorlands are encountered, along with significantly higher precipitation.

Thus, there is a graded variability between the three regions in terms of landforms and the availability of water, vegetation, and animals. The differences between the regions are more gradual, however, than the sharp differences in climate encountered when crossing the Andes Mountains.

Pleistocene and Early Holocene Climates

Pollen data has been used to reconstruct the late Pleistocene and Holocene environments of the Pampas (Preito 1996:73). At the end of the Pleistocene (around 10,500 rcybp1), the central Pampas were a herbaceous grassy steppe whereas the southwestern portion of the Pampas contained a mixed xerophytic woodland and steppe vegetation. These plant communities indicate a subhumid-dry climate with rainfall estimated to be 100 mm less than modern conditions. During the early Holocene, and up until about 8000 rcybp, the area contained ponds and swamps, indicating an increase in local precipitation and a probable decrease in evapotranspiration.

Iriondo and Garcia (1993) come to similar conclusions based upon data collected primarily from buried soils. During the late Pleistocene and into the early Holocene (18,000-8500 rcybp), the Pampas were "arid and cool, with aeolian sand and loess deposition" (1993:209). They characterize the fauna of the region as being similar to modern Patagonian species.

In Southern Patagonia, the environment at the end of the Pleistocene was one greatly influenced by the adjacent ice fields in the Andes. It appears that two glacial advances took place at the end of the late Pleistocene, one at 15,000-14,000 rcybp, and another readvance at 12,000-10,000 rcybp (Coronato et al. 1999). The climate during this period was characterized by a tundra environment, which gave way to a steppe with some pockets of forest sometime after 10,000 rcybp.

History of Research

The early peopling of the Southern Cone has been of considerable research interest for well over a hundred years now. The historical importance of this research is a fascinating topic, but beyond the scope of this paper. A brief introduction to the history of research is presented below.

The most notable early investigations were conducted by Florentino Ameghino, who created the first regional paleontological and archaeological sequences with his investigations of the Pampas (Ameghino 1879, 1880). Not surprisingly, he found skeletons in his fieldwork -- some of which were deeply buried, implying at least some antiquity to the remains, and some of which were associated with Pleistocene fauna.

Hrdlicka made a trip to the Pampas in 1908 and reviewed the skeletons and their geological associations, as well as the stone tools of the region (Hrdlicka 1912). He made two important conclusions (1912:385), first that the geological evidence did not support antiquity for the specimens, as the skeletons were intrusive burials. Second, so-called primitive traits exhibited on the crania were not in themselves indicative of great antiquity because they were within known modern human variation.

At the same time as Ameghino was making headlines, there was also the start of the great "Mylodon Rush" at the end of the 19th century, with the discovery of large quantities of Mylodon listai skin and bones in the caves of Ultima Esperanza (Southern Patagonia, Chile), most notably at Cueva del Mylodon. The remarkable preservation of claws, hair, and skin (one piece reportedly 1.5 m in length) led some people to believe that ground sloths might still be alive in this remote part of South America, and led to expeditions in search of large ground sloths (Bird 1988). There was a considerable interest in these discoveries, noted in at least 48 articles since the initial cave discovery (Bird 1988:222). Several notable researchers conducted excavations in the cave, including F.P. Moreno, O. Nordenskiold, and R. Hauthal, all between 1895-1900. Specimens from the cave reached world wide audiences, even Teddy Roosevelt was given a piece of hide and some dung from the site (Bird 1988:235). No shit.

Beginning in the late 1930's and continuing through at least the 1970's, Argentinean archaeology focused upon finding index fossils (i.e. diagnostic tool classes) and proposing Bordes-style cultures and industries for the archaeological deposits. Early and important research in the Patagonian region was completed by Junius Bird, of the American Museum of Natural History. His pioneering research at Fell's Cave and other early sites in the area, was the only research to be conducted on Paleoindian archaeology for many years.

However, the principle work conducted during this period (1930's-1960's) was completed by Argentinean researchers (Orquera 1987), principally that of O.F.A. Menghin, whose work in the Los Toldos area also began to yield evidence of late Pleistocene/early Holocene populations in Patagonia. His followers were some of the more notable archaeologists of the 1970's, including A. Cardich and C.J. Gradin. Also of importance during this period was the work of the Frenchmen J. Emperaire and his wife, A. Laming-Emperaire, who continued work at Fell's Cave, and in the Tierra del Fuego region of Patagonia.

Beginning in the 1970's, according to Orquera (1987:345), more work began to focus on technological studies, with less effort focused upon typological classification of type specimens and industries. At this time, Cardich, Gradin, and Palanca were all publishing accounts of late Pleistocene/early Holocene deposits in Argentina.

The majority of the data presented in this paper was derived from excavations completed after the 1970's, when Argentinean archaeologists shifted to more precise methods, a better understanding of lithic technology, and a focus on interdisciplinary work. Most of these recent archaeologists have also begun placing Argentinean archaeology into a wider, and more theoretical context than any previous work has ever attempted.

Pleistocene/Holocene site overviews

In this section, I review the known late Pleistocene and early Holocene sites of the Pampas and Patagonia. The majority of these sites are located within Argentina, although a few Southern Patagonian sites are located within Chile. A basic description of each site is presented, including the site location, lithics recovered, the types of raw materials present, the recovered fauna, and the dates from the sites. Some sites have better detailed coverage than others, and this is reflected by the amount of details I present on each site. In what follows, the sites are arranged by geographic region; at present, there are four main research areas identified, the Pampas, the Andean foothills, the Deseado River basin, and the area of Southern Patagonia (Chico Basin, Ultima Esperanza, and Tierra del Fuego).

Las Pampas

Cerro la China (Flegenheimer 1987, Zárate and Flegenheimer 1991, Flegenheimer and Zárate 1997)

Cerro La China is located in Loberia county at an elevation of 450 m above sea level, in the province of Buenos Aires. The site is made up of three localities located within 300 m of one another, on a low quartzite hill situated in the Tandilia Sierras. Site 1 is a small rockshelter and Sites 2 and 3 are open air sites. Springs are present in the area near Site 2, whereas a spring is located at Site 3.

Two levels have been identified at Site 1, the lowermost of which (level L2) is relevant to the late Pleistocene/early Holocene. A fluted stem of a projectile point (possible preform) and a Fell's Cave stemmed point preform were recovered from this level. Faunal preservation was very poor, but did include extinct armadillo (Eutatus seguini). The lower unit (B) of Site 2 contained a low artifact density, with two complete Fell's Cave stemmed points recovered. Site 3 has a more dense occupation in the lower stratigraphic unit (B), containing biface reduction debitage, although no tools were recovered.

The raw materials utilized at the site show a mix of local and perhaps semi-local (?) sources. The dominant raw material was a non-local quartzite, although a smaller amount of local quartzite was used as well. Chalcedony, opal, quartz, and basalt were also used in small frequencies. Given that Cerro La China is located within the mountains/foothills region (and atop a mesa of Paleozoic quartzites) (Flegenheimer 1985-86), most of the raw material sources are likely fairly local in origin.

Table 1:
Artifacts recovered from the Cerro la China site 2, stratigraphic unit 3 (Zárate and Flegenheimer 1991:286)
2 Fell's Cave Stemmed points1 core
1 graver71 flakes/fragments
6 unidentified tools/fragments 

Table 2:
Artifacts recovered from the Cerro la China site 3, stratigraphic unit 3 (Zárate and Flegenheimer 1991:287)
6 side scrapers6 bipolar cores
3 double side scrapers7 cores
3 double convergent scrapers16 core fragments/blanks
3 end scrapers949 flakes/fragments
10 unidentified scrapers/fragments 
2 scraper planes 
2 knives 
2 cutters 
2 notches 
4 denticulates 
7 gravers 
1 burin 
11 piece esquilles 
35 unidentified tools/fragments 

Eight radiocarbon dates have been obtained for the three Cerro La China localities, five from Site 1, two from Site 2, and one from Site 1. They range across the three localities from 10,525 ±75 to 11,150 ± 135 rcybp (Appendix 1). After calibrating these dates, Flegenheimer and Zárate are not sure whether this broad band of occupation dates documents either dating errors (calibration curve or otherwise) or multiple, repeated occupations of the site (1997:27).

Cerro el Sombrero (Flegenheimer 1987, 1995; Flegenheimer and Zárate 1989, 1997)

Cerro el Sombrero is located within the Buenos Aires province, in the Tandilia Sierras. Like Cerro La China, the site is located on top of a Paleozoic quartzite mesa (Flegenheimer and Zárate 1989). Surface collection of approximately 12000 m2 has revealed a substantial Paleoindian occupation of the hilltop. At least 29 fragments of Fell's Cave stemmed points were known as of 1987, from both Flegenheimer's surface collection, as well as Madrazo's collections in the early 1970's. Since Flegenheimer's project began at the site, several excavations have been completed in various areas around the site.

Small excavations of a highly dense area named Sector 12 (16 m2 in 1987) have revealed 454 flaked tools (including 18 Fell's Cave stemmed projectile points and/or fragments; also an abundance of side scrapers), 1 core, approximately 2,500 flakes, and 4 groundstone tools (similar to discoidal stones seen elsewhere in the Southern Cone). The vast majority (94%) of the artifacts are made of a fine-grained quartzite, sourced 30-60 km distant. Local raw materials such as a coarse-grained quartzite, opal, chalcedony, and igneous rocks, make up the rest of the assemblage.

Additional excavations (12 m2) were placed in a gridded fashion along the surface of the hilltop, in order to assess the site boundaries and density. Once again, a very large assemblage was recovered. The assemblage includes 1 Fell's Cave stemmed point, 2 Fell's Cave stems, 3 whole tools, 28 biface fragments, 46 uniface fragments, 865 flakes, and 3 pieces of ochre. The assemblage contains worn out tools, as well as late stage biface manufacturing debris. Thus, it appears that at least some retooling and manufacture was taking place on the hilltop.

A rockshelter has also been excavated (6 m2 as of 1989) at the Cerro el Sombrero site. The locality, named Abrigo 1, has yielded 12 tools (including Fell's Cave stemmed point fragments), 4 cores, and 205 flakes, as well as pieces of ochre. Like the main localities of Cerro el Sombrero, semi-distant sources are prevalent: 50% of the assemblage was exotic quartzite, 4% local quartz, 41% local quartzite, and 5% other raw materials.

Five radiocarbon dates have been reported for Cerro el Sombrero (from Abrigo 1?) (Appendix 1). They range in age from 8,060±140 to 10,745±75 rcybp, although the late date appears in error, as the other four dates cluster in the mid-10 kya range.

Clearly, Cerro el Sombrero is an amazing archaeological site, with an abundant surface and sub-surface record. The incredible amount of finished tools and manufacturing debris will provide incredible insight into the technological organization of hunter-gatherers inhabiting the area at the end of the Pleistocene.

La Moderna (Politis and Olmo 1986, Politis and Salemme 1990)

La Moderna is located in the province of Buenos Aires, exposed along the Arroyo Azul. The site is famous, as it was the first site on the Pampas to conclusively demonstrate the association of Paleoindian hunter-gatherers with extinct fauna. Excavations were carried out by F. Palanca in 1972-73 and continued by G. Politis between 1982-84.

The Paleoindian level (Unit a´) contains is a glyptodon (Doedicurus clavicaudatus) bonebed, interpreted as a kill site on the edge of a swamp. Additional fauna recovered include scutes of Glyptodon sp. and Sclerocalyptus sp., dermal bones of Mylodontinae, and a few bones of guanaco (Lama guanicoe), tuco-tuco (Ctenomys sp.), mataco (Tolypeutes sp.), rata-nutria (Holochilus brasiliensis), otter (Myocastor coypus), and ñandú (Rhea americana). Politis believes that humans are responsible for the deposition of the Doedicurus assemblage, however, the additional fauna are likely natural deaths.

A large lithic assemblage was associated with the fauna, consisting primarily of minimally retouched flakes. Palanca's excavations yielded 258 flakes, 690 chips, and 1000+ pieces of waste debris (shatter?). The assemblage was made entirely of quartz. Politis' renewed excavations recovered 4 secondary flakes with retouch and 35 waste flakes, all made of quartz, as well as 2 quartzite flakes and 1 chert flake.

Five dates have been run from the site, all from Doedicurus bone (Appendix 1). They range in age from 6,550±160 to 12,330±370 rcybp. Four of the five are AMS dates, although they show considerable range; the additional standard date was the most recent, at 6,550 rcybp. The dates from the site, therefore, remain ambiguous as to revealing the age of the deposits.

Arroyo Seco 2 (Politis et al. 1987, Politis and Salemme 1991)

Arroyo Seco 2 is located on the plains between the Tandilia and Ventania hills of the Pampas, in the province of Buenos Aires. It is an open site located on a loess ridge between Seco Creek and a small lagoon. The site was first excavated in the 1970's. Levels "Y, S, and Z" have been identified as the Paleoindian levels, although the material appears to be coming from a unit which is 35-50 cm thick.

The lithic assemblage recovered from the lower component contains 83 artifacts showing either primary and/or secondary flaking, as well as an "abundance" of flaking debris. Bipolar percussion and bifacial retouch have been identified in the assemblage. Tools include 5 types of scrapers, a single piece esquille, one half of a bola stone, and two pieces of groundstone.

Faunal remains recovered from the lower units include the extinct giant ground sloth (Megatherium americanum), ground sloth (Glossotherium robustus), ground sloth (cf. Mylodon), glyptodon (Glyptodon sp.), macrauchenid (Macrauchenia patachonica), llama (Paleolama cf. wedelli), horse (cf. Equus (Amerihippus) sp.), horse (Hippidion or Onohippidium), and armadillo (Eutatus seguini). Extant species were recovered, including guanaco (Lama glama guanicoe), Pampean deer (Ozotoceros bezoarticus), two species of armadillo (Zaedyus pichiy and Chaetophractus villosus), tuco-tuco (Ctenomys sp.), mara (Dolichotis patagonum), vizcacha (Lagostomus cf. maximus), Canidae, rhea (Rhea americana), and huron (Lyncodon cf. patagonicum). Sixteen human skeletons have also been recovered from the late Pleistocene/early Holocene levels (Unit Z). These skeletons, for the most apart, appear to underlie the lithic assemblage and faunal assemblage (Units Y and S).

Ten dates have been obtained at Arroyo Seco 2, all of which are on bone (Appendix 1). They range in age from 5,250±110 to 11,590±90 rcybp.

Cueva Tixi (Mazzanti 1996)

Cueva Tixi is a rockshelter located in the province of Buenos Aires. The shelter is located in La Vigilancia range, which is a smaller portion of the larger Tandilia Hills area of the Pampas. Excavations were begun at Cueva Tixi in 1988, an area of 15 m2 has been excavated to a depth of 80 cm (as of 1996). Five strata were identified in the cave deposits, which have yielded around 80,000 faunal specimens.

The late Pleistocene/early Holocene level is identified from two dates at 10,045±95 and 10,375±90 (Appendix 1), which were collected from two burned areas associated with fauna and lithic debris. Both extinct and extant fauna were recovered from the early deposits from the site and include Lama guanicoe, Ozotoceros bezoarticus, Myocastor coypus, Dolichotis patagonum, Zaedius pichiy, Eutatus seguini, Canis avus, and other faunal remains. Tools recovered include an anvil, hammerstones, cores, and low and medium grade scrapers, as well as some ochre. There was a spatial clustering of the debitage, which along with the artifacts recovered, suggests primary reduction and last stages of tool manufacture in the shelter. Raw materials represent both local (quartz and quartzite) and distant sources (basalt and flint), although no quantification was presented as to proportions or distances.

Paso Otero 5 (Martínez 1997)

Paso Otero 5 is an open site located within the valley of the Rió Quequén Grande, in the province of Buenos Aires. The site was first discovered in 1994, a total of 16 m2 have been excavated as of 1996.

Faunal remains recovered include giant ground sloth (Megatherium cf. M. americanum), American horse (Equus neogeus), Glyptodon sp., Toxodon, Hemiauchenia, Lama guanicoe, and probably ground sloth (Scelidotherium sp. or Mylodon). The assemblage consists of 32 complete and fragmented bones of both extinct and modern fauna, as well as hundreds of smaller bones and fragments. Some of the faunal assemblage is burnt.

Lithic recovered include two tools and a small amount of debitage. The first tool appears to be a unifacial knife, which has been retouched on several occasions. It appears that the knife was hafted at one time. The other tool is a multipurpose tool, with a graver (?) point, two notches, and has marginal retouch. Seventeen pieces of debitage (eight flakes, nine shatter) were also recovered. All the lithic artifacts are of quartzite, except the unifacial knife/point, which is made from a red chalcedony. The raw material sources for these artifacts is unknown, but the site is in close to other quartzite dominated assemblages in the Tandilia area, where quartzite sources are somewhat abundant.

A single radiocarbon date (10,190±120 rcybp) was obtained from a bone sample (Appendix 1).

Surface finds of Fell's Cave stemmed projectile points from the Southern Cone (Flegenheimer and Báyon 1996, Politis 1991)

Flegenheimer and Báyon have reported six Fell's Cave stemmed points from surface sites in the Pampas. Two specimens came from inland locations. One point (white quartzite) was associated with recent artifacts on a surface site. The other point (quartz), was found with two scrapers and a bifacial blank in a plowed field. Six others came from near the ocean coast. Four of the six are made of a yellow quartzite, and all appear to be resharpened. The other two points were made of igneous rocks (likely basalt and rhyolite). The points were manufactured from raw material sources that occur between tens and hundreds of kilometers away, and all appear to have been modified through resharpening.

There has been limited professional reporting of Paleoindian artifacts from surface sites in both the Pampas, but in Patagonia as well. It is not known if this is either a product of rather limited collecting and reporting by landowners or lack of visible early Holocene/late Pleistocene surfaces, but Flegenheimer and Báyon suggest that the high number of pieces from the coast reflects an area that avocational archaeologists frequently visit, as well as areas that have higher amount of eolian erosion.

Politis (1991) also presents information of the distribution of Fell's Cave stemmed points for the Southern Cone (Appendix 3). Additional specimens have been reported from sites on the Pampas such as Arroyo Pinto, Tacuarembo, Cerro Largo, and Rio Sauce Chico and sites in Uruguay such as Lobos, San Cayeta-no, Cabo Polonio, Valizas, and Santa Teresa. Other reported (not verified by Politis) or atypical Fell's Cave stemmed points have come from Chile at sites such as Aysen and Temuco; from Brazil at sites such as Rio Grande do Sul and Itapiranga; and other parts of Argentina, such as in Southern Patagonia at Abrigo de los Pecadores, Los Toldos, Caleta Olivia, and El Ceibo; in the Andes foothills at Rio Limay and La Crucecuta; and in central Argentina at Rio Tercero.

Andean Transition

Agua de la Cueva-Sector Sur (García 1992, 1995; García et al. 1999; Gil et al. 1997)

Agua de la Cueva is a rockshelter located in the province of Mendoza, in the mountains at an elevation of 2,900 m above sea level. The site is located close to a spring (García 1995:13). García believes the site was occupied seasonally during the warm months, since the site would be covered in snow during the winter. The site contains a rich stratigraphic record documenting hunter-gatherer use of the Pre-cordillera uplands, beginning in the late Pleistocene and continuing into the Holocene.

Stratum 2b has been dated to the late Pleistocene. Several hearths were recovered in these lower levels (García 1992:13; 1995:13). A diverse group of species has been recovered from level 2b (Gil et al. 1997:139). Camelid species dominate (Lama sp.), although birds and smaller species are present in minor abundance. The remains include large Aves, Rheidae, indeterminate mammal and large mammals, Lama sp., Lama (Vicugna sp.), Lama guanicoe, Chaetophractus vellerous, Lagidium sp., and Rodentia.

The tool assemblage from the shelter is very large, with 732 artifacts and 16,071 flakes being recovered from the 2b unit (García et al. 1999:49). Unfortunately the analysis that has been published is rather nondescript; however, there are several features worth mentioning. First, the artifacts are "big or medium-sized", they are mostly scrapers or knives exhibiting mostly unifacial retouch (bifacial retouch is present), and many of the artifacts appear to have multiple edges that could have been utilized. One projectile point fragment was recovered (possibly lanceolate), but it is missing the proximal portion.

Rhyolite, quartz, and "flint" are the predominate raw materials in the lithic assemblage, which are all locally available. More distant raw materials were brought to the rockshelter, however, including obsidian, opal, basalt, and quartzite (García 1992).

Six radiocarbon dates have been presented for level 2b from Agua de la Cueva, all on charcoal. They range from 9210±70 to 10,950±90 rcybp (Appendix 1).

Cueva Epullán Grande (summarized in Nami 1996)

Cueva Epullán Grande is a rockshelter located in the province of Neuquén. The shelter is approximately 50 m2 in size. A lower level from the cave yielded a hearth and a small number of flakes. These artifacts were associated with camelid bones of an extinct species. Mylodon sp. remains were also recovered from this lower level. A single radiocarbon date of 9,970±100 was obtained from the hearth (Appendix 1).

Arroyo Corral or Cueva del Manzano (summarized in Nami 1996)

Arroyo Corral is a cave located near Corral stream, in the province of Neuquén. An upper level in the cave deposits contained obsidian, silica, and basalt triangular projectile points, associated with modern faunal remains (Lama guanicoe and Ctenomys sp.). Nami (1996:262) believes this level is similar to the 9 kya level at Traful 1. Below this level, horse bones as well as some Mylodon dermal bones were recovered from a ochre-orange colored layer. Several flakes and a "scraper-like" artifact were also recovered near the faunal remains in this level, but the association remains unclear.

Cueva Traful 1 (summarized in Nami 1996)

Cueva Traful 1 is located along the bank of the Traful River, in the province of Neuquén. Over 3 m of deposits have been excavated in the cave, and the lowermost two (?) levels are Paleoindian in age.

Traful 1 is a level identified as containing triangular projectile points and modern fauna. This level has been dated to 9285±313 rcybp (Appendix 1; but see Nami 1996 for another date). A hearth was discovered below this level; a core and 19 flakes were associated with the firepit. Some of the flakes appear to be debitage from bifacial reduction. Unfortunately, no dates have been run on the hearth. Mylodon sp. bones have also been recovered from the site, but not in association with the cultural remains.

Gruta del Indio (D'Antoni 1983, García and Lagiglia 1999, Long et al. 1998, Long and Martin 1974)

Gruta del Indio is a rockshelter, at 660 m above sea level, in the province of Mendoza. The rockshelter has yielded a substantial amount of ground sloth dung over the years (Long and Martin 1974). Although human artifacts have been found in the cave deposits, the association is not quite conclusive. The site has four archaeological levels, the oldest of which, Atuel IV, is of interest. In this layer, large amount of ground sloth dung were found associated with charcoal fragments and possible firepits.

A large array of radiocarbon dates, on both charcoal and dung, have been run from the site deposits, ranging in age from 8045±55 to 30,800±700 rcybp (Appendix 2). Renewed stratigraphic and radiocarbon analyses by Long (Long et al. 1998) show that there is only rather weak evidence for the overlap of human occupation in the shelter with extinct faunal remains. Stratigraphic trenches did not confirm overlap of occupations, and the only radiocarbon dates which overlap (between the charcoal and sloth dung) were the samples with large standard errors.

Deseado Basin (South-Central Patagonia)

Piedra Museo (Miotti 1992, 1995; Miotti and Cattáneo 1997)

The Piedra Museo site is located in the province of Santa Cruz. The Alero EP-1 rockshelter is a locality of the Piedra Museo site first excavated by Miotti in 1990. The area surrounding Piedra Museo has many workshops and quarries, with Paleoindian materials on the surface (Miotti 1992). The lowest archaeological level at AEP-1 is Stratum 6. The level contained several lithics and faunal remains. Extinct fauna were recovered, including horse (Hippidion saldiasi), ground sloth (Mylodon sp.), ñandú (Rhea cf. americana) and guanacos (Lama gracilis). Modern fauna were recovered as well, including guanaco (Lama guanicoe) and ñandú (Pterocnemia pennata). Lithics recovered (28 pieces) include 3 large bifacial thinning flakes, 5 knives, 11 "secondary" flakes, and 9 pieces of knapping debris (likely shatter). The artifacts were made of chert, and several types of chalcedony.

Strata 4 and 5 also contained Paleoindian materials. These levels contained broken fluted projectile points and/or Fell's Cave stemmed points, unifacial scrapers, endscrapers, and knives. The fish-tailed projectile point was made of raw material available within 6 km of the site. The surface archaeology of the area, combined with excavation data, suggests that procurement of all raw materials at the site occurred within 6-10 km (Miotti 1995).

Two radiocarbon dates have been obtained for the lower levels at Piedra Museo, both on charcoal (Appendix 1). Stratum 6 has been dated at 12,890±90 rcybp, whereas Strat 4 and 5 have been dated to 10,400±80 rcybp.

El Ceibo Cave (summarized in Nami 1996, Borrero and Franco 1997)

El Ceibo Cave is located in the province of Santa Cruz. The site was excavated by Cardich between 1979 and 1983. A deep layer in the cave, Level 12, yielded a component similar to the one identified as Nivel 11 at Los Toldos 3. The assemblage includes mostly unifacial artifacts such as end-scrapers, lateral scrapers, and utilized flakes, although one bifacial artifact was recovered. Faunal remains recovered include horse (Hippidion saldiasi) and guanaco (Lama guanicoe and Lama gracilis). There is some confusion as to the dating of the site, as Nami (1996:261) reports a date of 12,600±600 (FRA-98) (Appendix 1) for the site, although this is potentially the same date from Nivel 11 at Los Toldos 3. Thus, it appears that there are no dates available for El Ceibo Cave.

Los Toldos 3 (Bird 1970, Cardich 1978, Cardich et al. 1987, Borrero and Franco 1987)

Los Toldos is a series of rockshelter/caves located along the "south middle" portion of the Deseado River in the province of Santa Cruz. Most of the caves appear to have been at least partially investigated. Menghin originally excavated the site in the early 1950's, although Cardich began to re-excavate the site in the early 1970's (Cardich 1978, Cardich et al. 1987). Los Toldos 3 yielded two late Pleistocene/early Holocene levels, which have been interpreted as representing two different culture complexes.

Nivel 11 (Level 11) contained large and thick retouched flakes, some smaller retouched flakes, large end scrapers, side scrapers, retouched knives, unifacial triangular "Mousteroid" points. Cardich did not recover any bifacial evidence from Nivel 11. Faunal remains recovered include guanaco (Lama guanicoe), horse (Parahipparion), Lama gracilis, Dusicyon avus, rodent bones, as well as some unidentified bone fragments.

The Toldense complex, from Levels 9/10 contained side and end scrapers, bifacial sub-triangular points, fishtail point fragments, bifacial knives, biface core fragments, a discoidal stone, and several bone tools including awls and spatulas. Faunal remains include camelids (Lama guanicoe and Lama gracilis), horse (Parahipparion), several birds (Eudromia and Rhea).

Only two dates have been reported from Los Toldos 3, charcoal samples from each of the levels (Appendix 1). The top of level 9, or the end of the Toldense level, was dated to 8,755±480 rcybp. Nivel 11 was dated to 12,600±600 rcybp. A potential problem is that dates provided in an earlier Cardich publication (1978) don't match those provided in a later publication; both measurements are different (8,755 versus 8,750; 12,600 versus 12,650) as well as the lab number for the controversial Nivel 11 date (FRA-98 versus FRA-96).

Another problematic detail is that Bird (1970) reports a discoidal stone and two Fell's Cave stemmed points from Menghin's excavations of Los Toldos Cave 2. These same artifacts were recovered in Los Toldos 3, and reported by Cardich (1978). Looking at the photo of the discoidal stone presented in Bird (1970:206), it is clearly visible that the discoidal stone is labeled as having come from Cave 2.

Arroyo Feo and Cueva de las Manos (summarized in Borrero et al. 1998)

Arroyo Feo and Cueva de las Manos are two sites located near the Pinturas River, in the Deseado Basin. Both sites were investigated by Gradín and associates. The lowest components in each cave were characterized by the Toldense industry, of Los Toldos 3. Triangular projectile points, marginally retouched artifacts, and modern fauna were recovered (mostly guanaco). Two dates were obtained from Arroyo Feo's lower level, Layer 11 Toldense, dated to 9330±80 and 9410±70 rcybp (Appendix 1). Two dates were also collected from Cueva de las Manos from the Layer 6 Toldense occupation, 9300±90 and 9320±90 rcybp (Appendix 1).

Southern Patagonia (Chico Basin)

Cueva Fell (Bird 1938, 1988)

Fell's Cave is located along Río Chico, in the province of Magallanes y Antartica Chilena, Chile. Fell's Cave is very famous, as it provided the first definitive evidence for the co-association of Pleistocene fauna with humans in southern Patagonia, and in-general, for all of South America. Bird's fieldwork in Patagonia, mainly completed in 1936-37 (as well as trips in the late 1960's) became the stuff of legend, in part due to the lack of comprehensive publication on his findings. His fieldwork was published posthumously in the late 1980's (Bird 1988), finally giving credit to his pioneering work fifty years earlier.

Fell's Cave measures 36 feet wide, by 28 feet deep, and 11 feet tall from the original floor (the floor forming the earliest occupation) (Bird 1938:269). The deepest layer, Layer 5, represented the first occupation of the cave. It was very thin, only 3-9 inches thick, and contained lithic artifacts and faunal remains. Bird estimates that 76.5 cubic feet were removed from the layer (Bird 1938:270). Four hearths were identified, each about 5 inches deep and about 2 feet across. The hearths contained fine black powder (ash?); burned bones of horse, sloth, and guanaco; and stone flakes.

Table 3:
Artifacts recovered from the lowest level of Fell's Cave, 1936-37 excavations (Bird 1988:145)
Fell's Cave -- Level 5Artifact type
1 triangular, concave base pointhafted implement
15 Fishtail pointshafted implement
3 knife fragments knives
26 single edge, rough flake scrapersscraper
1 two-edge, large scraperscraper
2 two-points scrapersscraper
6 large, rough, circular edge scrapersscraper
1 reversed-edge scraperscraper
6 end scrapersscraper
2 circular rubbing stone (discoidal stones)discoidal stones
63 total artifacts 

More than 300 flakes/chips were also recovered from Layer 5, as well as 5 bone artifacts. Bird states that there were "enough fragmentary bones to fill a 20-gallon container (Bird 1938:270). Faunal remains recovered include Aves, Canidae, guanaco (Camelidae), horse (Onohippidium), and giant ground sloth (Mylodontidae).

Table 4:
Bird remains recovered from the lowest level of Fells's Cave (Humphrey, Péfaur, and Rasmussen 1993, cited in Borrero and Franco 1998:233)
Cueva Fell -- Period I
Chloephaga pictaGoose1
Chloephaga rubidicepsGoose 1
Podiceps cf. P. occipitalisGrebe 1
Theristicus melanopsisIbis 1
Anas sibilatrixDuck 1
Anas sp.Duck 1
Tachyeres patachonicusDuck 1
Buteo polyosomaHawk 2
Buteo ventralisHawk 3
Buteo sp.Hawk 1
Attagis malovinusSeedsnipe 1

Six radiocarbon dates have been reported from the lower levels of Fell's Cave (Appendix 1). They range in age from 10,080±160 to 11,000±170 rcybp for the lowermost deposit (Magellan I or Bird I) and from 8480±135 to 9030±230 rcybp for the Magellan II or Bird II component.

Cueva Palli Aike (Bird 1938, 1988, Borrero and Franco 1997)

Palli Aike Cave is located on a lava hill, in the province of Magallanes y Antartica Chilena, Chile. Palli Aike was also discovered and excavated by Junius Bird in the late 1930's (Bird 1938). The cave is roughly 20 feet wide, 46 feet deep, and between 6-13.5 feet high.

The lowest deposits had stone and bone tools mixed in with the broken bones of horse and ground sloth, although the stratigraphy is anything but straight-forward. Fauna recovered include guanaco (Lama guanicoe), horse (Hippidion saldiasi), ground sloth (Mylodon sp.), and fox (Dusicyon avus). One Fell's Cave stemmed point was recovered from the site. Three cremated human remains were also recovered in the deep deposits of the cave. Buried in the rear of the shelter, under a pile of rocks was a "sloth cache" as Bird (1988:107) described it, parts of at least 7 ground sloths were recovered (Bird 1938).

One date has been reported from the site from the lowest component (Appendix 1), at 8639±450, although this is interpreted by most researchers as being contaminated, and should be treated as a minimum date.

Cerro Sota and Palli Aike human remains (Soto-Heim 1994, Turner 1992)

Human burials were also found at Cerro Sota, a site very near to Fell's Cave. Unfortunately, osteological analysis of these human burials was completed before their ages were independently assessed. Despite the association with extinct fauna, the human bones from Cerro Sota have been dated to 3,900 rcybp (Hedges et al. 1992 cited in Borrero and Franco 1997:222). Thus, not much faith may be placed in Turner's proud statement (1992:15) that "these specimens are probably the most securely dated of all presumed Paleoindian teeth from South America, and uniquely provide a population sample." There is no consensus as to the radiocarbon age of the Palli Aike human bone, as no independent dating has been performed.

Cueva las Buitreras (Sanguinetti de Bórmida and Borrero 1984, Torres 1986)

Cueva las Buitreras is a rockshelter located within the province of Santa Cruz. Five coprolites were recovered from the shelter. They were associated with lithics and faunal remains. The deposits have not been radiocarbon dated, but the Paleoindian level is overlain by a volcanic ash that provides a minimum age of 9,100 rcybp for the coprolites.

Of the five coprolites, one was of human origin, containing charcoal and unidentified plant remains. The four other coprolites came from an omnivorous carnivore, containing digested beetles, rodent molars, and plant fragments. The insects included Taurocerastes patagonicus and Caenorhynus lineatus, both found today on the rainforest and steppe of southern Patagonia. The rodent molars were from Akodon xanthorhinus, a field mouse also found on the rainforest and steppe of southern Patagonia. The coprolites, unfortunately, do not provide an environmental reconstructions that differ from modern ecological conditions in southern Patagonia.

Fauna also recovered from site deposits, include extinct species such as ground sloth (Mylodon listai), fox (Dusicyon avus), and horse (Hippidion-Onohippidion), as well as extant species, including guanaco (Lama guanicoe), small rodents (Reithrodon and Ctenomys), and dolphin (Delphinidae).

Southern Patagonia (Ultima Esperanza)

Cueva del Mylodon (Bird 1988)

Mylodon Cave is an enormous shelter located in the Ultima Esperanza Inlet of Chile. The cave is approximately 400 ft wide by 680 ft deep (height was not reported). As discussed earlier, the discoveries at Mylodon Cave fueled some of the first archaeological/paleontological research into the Patagonian region. In the 1930's, Junius Bird visited the site, hoping to discover whether there were any intact deposits left after the many years of prospecting in the cave.

Faunal remains recovered from the site include guanaco (Lama guanicoe), horse (Hippidion saldiasi), ground sloth (Mylodon sp.), jaguar (Panthera onca mesembrina), fox (Dusicyon avus), and Macrauchenia sp. As mentioned, an untold (no doubt tremendous) amount of sloth skin, hair, bones, and dung have been removed from the cave over the years, mostly from the late 19th century. The collections are spread the world over, including specimens in Rome, Berlin, Stockholm, Copenhagen, London, La Plata, Magallanes, Santiago, New York, and Los Angeles (Bird 1988:226).

At least 23 radiocarbon dates have been reported from skin, bones, and dung of mylodons from the cave (Appendix 2). They range in age from 10,200±400 to 13,560±180 rcybp.

Cueva del Medio (Nami 1987, Nami and Casé 1988, Nami 1989, Menegaz and Nami 1994, Borrero and Franco 1997)

Cueva del Medio is located near the Ultima Esperanza Inlet, in the Province of Ultima Esperanza, approximately one kilometer from Mylodon Cave. The cave had previously been quarried for paleontological specimens in the late 19th century (during the same time Mylodon Cave was being "excavated"), although modern archaeological research was begun there in the mid-1980's by Nami. The cave formed within a conglomerate, and measures about 40 meters wide, 90 meters deeps, and about 6 meters high. Two cultural components have been identified, Bird I and Bird III, the former of which is late Pleistocene in age (Nami 1987). At least 80 m2 had been excavated from Cueva del Medio as of 1994 (Menegaz and Nami 1994).

Faunal remains recovered from the Paleoindian levels include extinct species such as horse (Hippidion saldiasi), ground sloth (Mylodon listai), fox (Dusicyon cf. culpaeus and Dusicyon sp.), Dolichotys patagonum, Rheidae, llama (Lama owenii, Lama guanicoe, Lama gracilis) and other remains (Nami 1989, Menegaz and Nami 1994). The jaguar (Panthera onca mesembrina), Felis sp., and Cervidae recovered are from deposits below this dated level (Borrero and Franco 1997:231). Three hearths have been identified from the Paleoindian levels.

Raw materials utilized at the site come from three sources, all of which are local in origin (Nami and Casé 1988). Cobbles are available from within the cave, in both primary and secondary deposits of the conglomerate within which the cave was formed. Raw materials are also available from glacial deposits surrounding the site. The rocks available are rhyolites, dacites, and devitrified tufas.

Eighteen dates have been collected for Cueva del Medio, on both charcoal and bone (Appendix 1).

Cueva del Lago Sofía 1 (summarized in Nami 1996, Borrero and Franco 1997)

Cueva del Lago Sofía is a cave located on a terrace of the Rivas River, in the Province of Ultima Esperanza. The site is located across the valley from Cueva del Mylodon and Cueva del Medio. The cave is 5 m wide and 25 m deep. A level has been identified within the cave with the same extinct species as those found in Cueva del Medio, as well as evidence of human occupation. Guanaco (Lama guanicoe), horse, (Hippidion saldiasi), and ground sloth (Mylodon sp.) were recovered. A hearth was also discovered, surrounded by this extinct fauna. Bifacial and unifacial flakes, other debitage (?), and a bird bone awl were associated with the hearth. Two dates have been obtained from charcoal from the site, ranging between 11,570±60 to 12,990±241 (Appendix 1). Given the limited reporting of this site, the context of the dates might be questionable, although these doubts have not been voiced in the Argentinean literature.

Cueva del Lago Sofía 4 and Dos Herraduras 3 (summarized in Borrero et al. 1998, Borrero and Franco 1997)

Cueva del Lago Sofía 4 and Dos Herraduras 3 are paleontological sites in Ultima Esperanza. Sofía 4 is an extinct carnivore den, dated between 13,000 to 12,000 rcybp, which has yielded guanaco (Lama guanicoe), horse (Hippidion saldiasi), ground sloth (Mylodon sp.), and saber-toothed tiger (Smilodon sp.). Herraduras 3 is a ground sloth natural death site, dated between 12,000 to 11,300 rcybp, which has yielded horse (Hippidion saldiasi) and ground sloth (Mylodon sp.).

Southern Patagonia (Tierra del Fuego)

Tres Arroyos (summarized in Nami 1996)

Tres Arroyos is a rockshelter located on De Los Onas hill, in the Camen Sylva Range of the province of Tierra del Fuego. The site was investigated during the early to mid-1980's. Two layers, at 60-80 and at 125 cm (Levels Va and Vb), have yielded Paleoindian occupations.

End and lateral scrapers, retouched knives, a core, and two projectile point fragments (as well as debitage; assemblage size=213) have been recovered from these levels. The points are thought to be Fell's Cave type points. An oval-shaped hearth (40 cm x 30 cm x 12 cm) was also recovered, surrounded by extinct fauna. Fox (Canis (Dusicyon) aves), guanaco (Lama guanicoe, Lama sp.), and ground sloth (Mylodon sp.) have been recovered, as well as some snails. Horse (Hippidion saldiasi) was recovered from below the dated level.

Three radiocarbon dates, all on charcoal, have been obtained from Tres Arroyos Paleoindian level. They range in age from 10,280±110 to 11,880±250 rcybp, although two of the three are in the 10,300 range.

Marazzi Rockshelter (summarized in Borrero et al. 1998)

Marazzi Rockshelter is also located on Tierra del Fuego. The site dates to 9590±210 rcybp (another discrepancy in reported date (Borrero et al. 1998 versus Orquera 1987)) (Appendix 1) and documents a unifacial tool industry and modern fauna.

Comparison of Sites

In this section, I will summarize the available archaeological data from the Pampas and Patagonia. First, I will discuss the timing of the initial peopling of the Southern Cone, as well examine some limited data on Pleistocene extinctions. Second, I will describe patterns in site locations and faunal diversity and processing. Third, I will summarize aspects of the technological organization of hunter-gatherers in the Pampas and Patagonia at the Pleistocene/Holocene boundary.

Radiocarbon Dates

Eighty-one radiocarbon dates are available from the Pampas and Patagonia that document the early peopling of the Southern Cone (Appendix 1). I will be examining the timing of early occupations by each individual region, then I will combine the data to examine the entire date list. In addition, I will briefly examine radiocarbon evidence for Pleistocene extinction of ground sloths with 54 additional radiocarbon dates from non-cultural deposits in Mylodon Cave (Ultima Esperanza) and Gruta del Indio (Andes foothills) (Appendix 2).

I have not calibrated the radiocarbon dates for several reasons. Calibration would provide calendrical corrections for these dates and likely stretch and/or compress parts of the patterns seen here. However, I believe that in order to do the analysis more thoroughly, in addition to calibration, much more data would have to be collected for each sample, such as the exact stratigraphic association of the date, as well as firmer control of the material dated, so that C13/12 corrections could be made properly.

Another point worth mentioning is that these dates are not averaged by site or by each site component; but this would be a much more accurate assessment of the record. However, this is also difficult given that some sites have a tremendous number of dates -- often spread over millennia -- which are interpreted as repeated occupations of the same site. Other sites, however, have only a single date. Making all these corrections and checks to the dataset is clearly beyond the scope of this paper.

Thus, I argue by enumeration, and accept the majority of dates that conform to the pattern, and seriously question dates that are slightly older than what is expected given the observed pattern. This is definitely a biased perspective. But unfortunately, unless more dates are run on different types of materials, from both paleontological and archaeological contexts, we will continue to have a biased perception of the Pleistocene timing and the environment being colonized. That said, what follows is a regional investigation of the radiocarbon record.


There are 32 radiocarbon dates available from the Pampas from six different sites (Appendix 1). A surprising number of the dates are on bone (17 or 53%), whereas there are 10 dates (31%) of unknown material (I was not able to determine what was dated), and 5 charcoal dates (16%). Several patterns are evident. There is a clear plateau of dates between 10,000-10,800 rcybp, which clearly indicates that by this time, populations are present in the region. Four dates are older than this, from Cerro la China, Arroyo Seco and La Moderna, three of which are on bone.

At Arroyo Seco, Politis (Politis et al. 1995:195) states that the spread of dates from the site -- over 4000 radiocarbon years -- represents repeated occupations by hunter-gatherers. But, the stretch of dates between 5000 to 9000 (many of which are from Arroyo Seco) are almost exclusively bone dates. These dates do document subsequent occupations of the Pampas, but I believe that at least some of these bone dates should be considered more as minimum dates.

The early date from La Moderna is certainly questionable, given the other evidence from the Pampas, and the majority of the other dates from La Moderna, which hover around 7000 rcybp. Politis believes the sample to be contaminated, likely due to secondary carbonate intake (Politis et al. 1995:196). The cluster of dates at La Moderna, around 7000 rcybp, associated with the glyptodon kill, might indicate a Holocene presence of this species, however, given that all the dates obtained from the site have been on bone, and that we already have some groundwater carbonate problems, leads me to conclude that these are also minimum dates, with the timing of glyptodon extinction to be tested elsewhere.

Andean foothills

Moving to the west from the Pampas, nine radiocarbon dates have been obtained from five sites in the Andean pre-Cordillera (Appendix 1). All nine dates are on charcoal. They have a gradual spread, from right after 9000 rcybp to slightly before 11,000 rcybp. They are in general agreement with the dates from the Pampas, with little evidence for hunter-gatherer occupation before 11,000 rcybp. There are no anomalous dates in the sequence.

Deseado Basin

Moving south from the Andes, into the Deseado Basin region, nine radiocarbon have been reported from five sites, all from charcoal. Much like the other two regions, we have clear evidence from right before 9000 rcybp to about 10,400 rcybp. The two remaining dates, from Piedra Museo and Los Toldos 3, are of question, as they do not appear to match the pattern. The early date from Los Toldos 3 (12,600 rcybp) has been questioned by researchers, as the sample was collected from dispersed pieces of charcoal (Borrero 1996:341, Lynch 1990:22), and there is a question of the El Ceibo date, which I think is the same date published for Los Toldos Cave 3. The other early date is from Piedra Museo (12,890 rcybp), and does not appear to have any criticism thus far by Argentinean researchers.


The Magellan region, comprised of the Chico River Basin and the Ultima Esperanza Inlet, is very well dated, with 27 radiocarbon samples from four different sites. Again, bone dates dominate the sample list, with 14 (52%), although there are 7 charcoal dates (26%) and 6 dates of unknown material (22%). Viewing the patterns, the dates document a clear presence in the Magellan from approximately 8500 rcybp to 11,100 rcybp. The radiocarbon evidence suggests that the Magellan region was occupied by hunter-gatherers before the Pampas, the pre-Cordilleran Andes, and possibly the Deseado Basin. The early occupation evidence comes from both Fell's Cave and Cueva del Medio. Four dates from Cueva Lago Sofía and Cueva del Medio (both charcoal and bone) are even earlier than the overall sequence. The earliest date from Cueva Lago Sofía, at 12,990 rcybp, has been stated to not come from cultural levels (Borrero et al. 1998:195), although there appears to be no controversy over the 11,570 rcybp date from Cueva Lago Sofía. At Cueva del Medio the earliest date at 12,720 rcybp and is again not from a cultural level (Borrero et al. 1998:195), but the next earliest date (12,390 rcybp) appears anomalous, because additional samples from the same feature dated to the 10,300 to 10,800 rcybp range (Borrero et al. 1998:196).

Tierra del Fuego

There are only four radiocarbon dates published for Tierra del Fuego, which is located to the south across the Strait of Magellan. All four dates are from charcoal and came from two sites. Again, a plateau of occupation is evident from 9590 rcybp to 10,420 rcybp. The earliest date of 11,880 rcybp is from Tres Arroyos and is thought to be a close approximation of site occupation (Stern 1992), but this is clearly in question (Borrero et al. 1998:196).

Tierra del Fuego is an interesting region, as the area was connected to the mainland at the time of these early occupations (McCulloch et al. 1997:25). Prior to 14,000 rcybp, the Straits of Magellan were filled with a large meltwater river, but there were times from 14,000-13,000 rcybp and again from 10,000-8000 rcybp when the area would have been a dry land bridge between the Magellan area and Tierra del Fuego. From 12,000 to 10,000 rcybp, the Strait would have been filled with an episodic meltwater river, much more of a localized obstacle. After 8000 rcybp, however, the Straits were filled with ocean water. Thus, it appears that the area was first occupied during the period when the Straits were filled by an intermittent river.

Pleistocene peopling of Southern South America

Thus, comparing the Southern Cone radiocarbon data from the eastern side of the Andes, the region appears to have been occupied in the south first, conservatively stated at shortly before 11,000 rcybp. A few sites might prove to be older, such as Piedra Museo, Tres Arroyos, and Cueva Lago Sofía, but the predominate evidence points to a later occupation date. Later occupations are evident to the north, next in the Pampas, right around 11,000 rcybp and shortly thereafter, in the Andes foothills at mid-10,000 rcybp.

Ground sloth dates from Gruta del Indio and Mylodon Cave

Fifty-four radiocarbon dates have been obtained from two ground sloth den sites, Gruta del Indio in the Andean foothills and Mylodon Cave on the Ultima Esperanza inlet. The dates are from a variety of materials, including the bone, dung, and hide of ground sloths, as well as associated wood and charcoal. The majority of dates are from dung (34) and charcoal (13).

There are several interesting trends in the data from the two caves. The Andes data has a much older sequence, with several dates in the mid-20kya range and low-30kya. A radiocarbon hiatus is evident between 23,000 and 14,000 rcybp. The majority of dates then fall between 14,000 and 8000 rcybp. In general, the Mylodon Cave dates are older than those of Gruta del Indio.

Comparing this data to the archaeological dates, it is interesting to note that the earliest cultural dates were from the Magellan region around Mylodon Cave. On the contrary, one of the last regions to be occupied by hunter-gatherers were the Andes foothills in the vicinity of Gruta del Indio. If ground sloths were slowly going extinct, it might appear that they went extinct first in the south, in the area around Mylodon Cave. As human populations moved to the north, and into even more remote regions, such as the Andes foothills, the last of the ground sloth went extinct in areas such as Gruta del Indio around the time of Pleistocene/Holocene boundary.

Pleistocene/Holocene site comparison

Laura Miotti recently summarized data for 31 sites from the Southern Cone (Miotti and Salemme 1999), spanning the late Pleistocene and early Holocene and into the middle Holocene.

Table 5:
Southern Cone sites used in the Miotti and Salemme (1999) study
Tres ArroyosMagellan late Pleistocene/early Holocene (13-8.5 kya)
Fell Cave IMagellan late Pleistocene/early Holocene (13-8.5 kya)
Fell Cave IIMagellan late Pleistocene/early Holocene (13-8.5 kya)
Palli AikeMagellan late Pleistocene/early Holocene (13-8.5 kya)
Cerro SotaMagellan late Pleistocene/early Holocene (13-8.5 kya)
Cueva del Medio IMagellan late Pleistocene/early Holocene (13-8.5 kya)
Cueva del Medio IIIMagellan late Pleistocene/early Holocene (13-8.5 kya)
Mylodon CaveMagellan late Pleistocene/early Holocene (13-8.5 kya)
Las BuitrerasMagellan late Pleistocene/early Holocene (13-8.5 kya)
El Ceibo 7Deseado Basin late Pleistocene/early Holocene (13-8.5 kya)
Los Toldos 3, Level 11Deseado Basin late Pleistocene/early Holocene (13-8.5 kya)
Los Toldos 3, ToldenseDeseado Basin late Pleistocene/early Holocene (13-8.5 kya)
Los Toldos 2, BDeseado Basin late Pleistocene/early Holocene (13-8.5 kya)
AEP 1, lower componentDeseado Basin late Pleistocene/early Holocene (13-8.5 kya)
Arroyo Feo 11Deseado Basin late Pleistocene/early Holocene (13-8.5 kya)
Manos PintadasDeseado Basin late Pleistocene/early Holocene (13-8.5 kya)
Arroyo Seco 2, lower componentPampa late Pleistocene/early Holocene (13-8.5 kya)
La Moderna, lower componentPampa late Pleistocene/early Holocene (13-8.5 kya)
Tixi Cave, lower componentPampa late Pleistocene/early Holocene (13-8.5 kya)
Tunel 1, lower componentMagellan middle Holocene (8.5-4.5 kya)
Fell Cave IIIMagellan middle Holocene (8.5-4.5 kya)
Cerro Sota, upper componentMagellan middle Holocene (8.5-4.5 kya)
AEP 1, upper componentDeseado Basin middle Holocene (8.5-4.5 kya)
Los Toldos 3, CasapedrenseDeseado Basin middle Holocene (8.5-4.5 kya)
La Marita, lower componentDeseado Basin middle Holocene (8.5-4.5 kya)
Arroyo Feo 8/9Deseado Basin middle Holocene (8.5-4.5 kya)
Alero CardenasDeseado Basin middle Holocene (8.5-4.5 kya)
Arroyo Seco 2, middle componentPampa middle Holocene (8.5-4.5 kya)
Tixi Cave, middle HolocenePampa middle Holocene (8.5-4.5 kya)
La Toma, lower componentPampa middle Holocene (8.5-4.5 kya)
La OllaPampa middle Holocene (8.5-4.5 kya)

Thirty-eight variables were coded for each of the 31 sites, in terms of the presence or absence of certain traits (see below). I modified their database in several ways to examine regional variability in the Southern Cone. First, I separated the sites into three regions including the Pampas, Deseado Basin, and Magellan. Second, I divided the traits into three classes of data (site settings, tool technology, and faunal processing). Finally, the presence of each trait was then summed for all sites within each particular region. Thus, in the following tables, the cell value refers to the percent of all cases from that region/period exhibiting the particular trait in question.

Site Setting

Site setting is composed of 6 categories, and except for the cave/open air categories, the categories are not mutually exclusive of one another.

Table 6:
Site setting (Miotti and Salemme 1999)
N (sites)9 7 3 3 5 4
cavern100 100 33.3 66.7 100 0
open air0 0 66.7 33.3 0 75
cliff0 28.6 33.3 0 20 0
topographic trap66.7 85.7 100 33.3 100 75
lagoon22.2 0 33.3 0 0 0
river terrace0 71.4 33.3 0 80 50

Quite strikingly, caves are the predominate setting for both the Magellan and Deseado areas throughout the sequence. Proximity to water is also common, with high percentage of Deseado and Pampa sites being near river valleys, and Magellan and Pampas sites being near lagoons (ponds/lakes?). Topographic traps are a subjective category, but it appear that sites from all three regions were commonly located near traps.

The presence of such large numbers of cave sites speaks to the amazing faunal preservation in Patagonia, especially in the Magellan region, where ground sloth hides were preserved in dry cave climates. Most Pampa sites are open air sites, although this is not surprising given that the Pampas are mainly grasslands with a few low-lying foothill ranges. In the Pampas, it is likely that most well preserved sites will be found in deeply buried river valleys, although these are clearly more difficult to locate and excavate.

During the Holocene, the site settings remain fairly constant, with the noticeable disappearance of sites near lagoons and cliffs. The shift away from lagoons might be related to a Holocene drying of lakes and ponds. The decrease in cliff sites is also interesting, but it should be noted that there are no mass kill sites (especially those at the bottom of cliffs) known for the Southern Cone; sites with large faunal assemblages are all thought to be accretional accumulations (Borrero 1996:349). Thus, it is doubtful that this site setting shift is related to changes in hunting strategies.

Tool Technology

Fourteen categories are represented in this class, mostly related to certain stone tool classes, as well as lithic reduction stage.

Table 7:
Tool assemblages (Miotti and Salemme 1999)
N (sites)9 7 3 3 5 4
awl55.6 100 66.7 33.3 100 75
blades/knives22.2 14.3 0 0 20 0
projectile points77.8 100 66.7 66.7 0 0
scrapers44.4 100 33.3 0 40 25
tools without edges66.7 85.7 33.3 100 100 75
utilized flakes/knives33.3 71.4 33.3 33.3 60 50
utilized bone0 14.3 0 0 20 0
weapons11.1 14.3 0 0 0 25
nodules/hammerstones11.1 14.3 0 0 20 0
cores11.1 0 0 0 20 0
flake debris22.2 14.3 33.3 0 0 0
natural edges?55.6 85.7 66.7 66.7 60 75
no cortex11.1 0 0 0 0 0
hearth0 0 0 0 20 0

Awls, projectile points, and scrapers are common tools present in the three regions. There is a general decrease in the presence of these formal tools during the mid-Holocene. "Weapons", blades/knives, cores, hammerstones and utilized bone were not well represented during either period.

Concerning technological organization, it appears that most flaking was done off-site, as flake debris was not recovered from most sites. Yet, the off-site reduction was not extensive, as most flakes retain cortex. Most flaking was probably informal and expedient, as most sites have flakes and tools with natural edges, as well as tools without edges. These patterns in technological organization appear to remain constant from the early sites and into the mid-Holocene sites (although no flake debris was recovered from any Holocene sites?!).

Only one hearth, from a mid-Holocene site, was identified in the Miotti study. This is generally surprising, given that hearths were identified in many sites discussed earlier in this paper. Clearly, the Miotti and Salemme data must be taken as generalizations of the actual data.

Faunal Processing

Fourteen variables were coded for patterns observed in faunal remains. Most variables concern the presence of anatomical units, although disarticulation and processing are also addressed.

Table 8:
Faunal Processing (Miotti and Salemme 1999)
N (sites)9 7 3 3 5 4
complete carcasses66.7 85.7 66.7 100 100 50
bone concentrations33.3 14.3 0 0 20 0
articulated long bones11.1 0 0 0 0 0
articulated ribs100 100 100 66.7 100 75
fragments of skulls11.1 0 66.7 0 0 0
tusks/antlers77.8 57.1 33.3 66.7 80 50
mandibles0 0 33.3 0 0 0
vertebrae55.6 57.1 66.7 66.7 100 75
ribs44.4 57.1 66.7 33.3 60 75
pelvis and/or scapula66.7 100 100 33.3 80 75
autopodials88.9 85.7 100 66.7 80 50
fragments of limb bones100 100 100 100 100 75
cutmarks22.2 28.6 66.7 0 20 0
green bone breaks88.9 57.1 33.3 66.7 60 75

Complete carcasses are fairly common for all regions and all periods. Processing is more intense on the limb bones, with only 1 site showing articulated long bones, whereas most sites contain carcasses with articulated ribs.

Relatively few crania or mandibles were generally recovered, although all other elements have high representations, including tusks/antlers. This suggests either differential transport of select elements, a higher attrition rate on skulls and mandibles (highly unlikely due to their bone density), or perhaps faunal accumulation by carnivores inhabiting these caves.

Cutmarks and green bone breaks are common in all three regions, with high percentages of green bone breaks. Whether these green bone breaks are humanly produced is probably debatable, as most of these sites are cave sites that were likely serving as carnivore dens. Cutmarks are more commonly encountered during the Pleistocene, rather than the Holocene, which suggests more intensive processing. This is quite surprising, given the opposite pattern in North America, where Paleoindian populations rarely practiced extensive processing.

Faunal Presence

Forty-two species were recovered from the sites. I divided the fauna into three general classes for comparison large herbivores, birds, and carnivores and assorted other animals.

Table 9a:
Fauna recovered (Miotti and Salemme 1999)
  N (sites) 8 7 3 3 5 4
Taxa Common
Megatheriumgiant ground sloth 0 0 33.3 0 0 0
Mylodon sp.ground sloth 87.5 0 66.7 0 0 0
Glossotherium sp.ground sloth 0 0 33.3 0 0 0
Toxodon sp.toxodon 0 0 33.3 0 0 0
Glyptodon sp.glyptodon 0 0 66.7 0 0 0
Doedicurus sp.glyptodon 0 0 33.3 0 0 0
Sclerocalyptusglyptodon 0 0 33.3 0 0 0
Eutatusarmadillo 0 0 66.7 0 0 0
Tolypeutesarmadillo 0 0 33.3 0 20 0
Chaetophractus villosusarmadillo 0 0 66.7 0 0 50
Zaedyus pichiyarmadillo 0 0 66.7 0 20 75
Equus (Amerhippus)horse 0 0 33.3 0 0 0
Hippidion 87.5 71.4 33.3 0 0 0
Lama gracilis camelid 0 57.1 0 0 0 0
Lama guanicoe camelid 87.5 100 100 66.7 100 100
Lama sp. camelid 37.5 0 66.7 0 0 0
Macrauchenia Macrauchenia 0 0 33.3 0 0 0
Ozotoceros bezoarticus deer 0 0 66.7 0 0 75
Cervidaedeer 12.5 0 33.3 0 0 25

There is remarked diversity in the fauna present in each region during the early sites. In the Pampas, the rich faunal diversity is represented by the Arroyo Seco site. Two species of ground sloth, three species of glyptodons, four species of armadillos, toxodons, Macrauchenia, and Pampas deer appear solely in the Pampas. Horse, guanacos, and mylodons were also recovered in the Pampas and the regions to the south.

The Deseado region has relatively few species represented, with only horse and guanacos present. Whereas, in the Magellan region, horses and guanacos are common, in addition to the large numbers of ground sloth (Mylodon). It is interesting to note that ground sloths have not been recovered from the Deseado2; there must be a remarked shift in paleoenvironments between the two regions, as the sites from both regions are caves, where ground sloths are expected to den.

During the middle Holocene, the faunal diversity drops remarkably, with the extinction of large numbers of species. Guanacos, armadillos and deer remain common, which matches later ethnographic studies from the Pampas and Patagonia.

Table 9b:
Birds recovered (Miotti and Salemme 1999)
  N (sites) 8 7 3 3 5 4
Taxa Common
Rheidaerhea 37.5 71.4 33.3 0 40 75
Rhea americanarhea 0 28.6 66.7 0 0 25
Pterocnemia pennatapartridge 0 28.6 0 0 20 0
Eudromia eleganspartridge 0 28.6 33.3 0 20 0
Nothura maculosapartridge 0 0 33.3 0 0 0
Fulica sp.coot 0 28.6 0 0 0 0
Speotyto cuniculariaburrowing owl 0 0 33.3 0 0 0
Avesgeneralized birds 75 42.9 33.3 100 40 25

The bird data is also very interesting, as the Deseado Basin has a large diversity of birds, whereas the Magellan region is lacking. Most of these birds are likely natural occurrences, with perhaps the exception of the rhea, which was likely hunted.

During the mid-Holocene, the bird diversity drops, with the exception of rheas and undifferentiated birds.

Table 9c:
Additional fauna recovered (Miotti and Salemme 1999)
  N (sites) 8 7 3 3 5 4
Taxa Common
Pantera onca mesembrinabig cat 0 0 0 0 0 0
Felis 37.5 0 0 33.3 0 0
Felis concolorcat 12.5 42.9 33.3 33.3 0 0
Canis (D.) avusfox 25 14.3 33.3 0 0 0
Canidaedog/wolf 62.5 57.1 66.7 66.7 40 25
Canis (C.) familiarisdomesticated dog 25 0 0 33.3 20 0
Dolichotis patagonumPatagonian hare 0 0 33.3 0 0 25
Lagostomus maximusplains viscacha 0 0 33.3 0 0 0
Lagidium sp.mountain viscacha 0 28.6 0 0 0 0
Ctenomys sp.tuco-tuco 37.5 28.6 66.7 33.3 80 50
Piscesfish 0 14.3 0 33.3 0 0
Adelomedon brasiliensis  12.5 0 33.3 0 0 25
Volutidaemollusc 12.5 28.6 33.3 66.7 40 25
Pinnipediaseal/sea lion 25 0 0 33.3 0 25
Lutra sp.otter 0 0 0 0 20 0

A large number of carnivores were recovered from the cave sites in Patagonia, including various Felis, foxes and wolves. These species likely account for some of the accumulation and attrition to the bone recovered from the caves.

Many of the smaller species appear to be represented in habitat-specific settings, such as the hare, the viscachas, and the sea lions. Interestingly, the tuco-tuco was recovered from several sites in each region, probably the result of diversified hunting of smaller body sized mammals.

Table 10:
Regional faunal diversity, with cell values indicating the number of sites (Miotti and Salemme 1999). The median diversity values for each region are in red.
# species Pleistocene
00 0 0 0 0 0
10 0 0 0 1 0
20 0 0 0 0 0
30 0 0 0 1 0
41 1 0 1 1 0
51 1 0 1 0 1
61 1 0 1 0 2
72 2 0 0 1 0
82 1 1 0 1 1
91 1 0 0 0 0
100 0 1 0 0 0
260 0 1 0 0 0

Examining the faunal diversity per region shows some interesting patterns as well. First, there is a higher species diversity for each region during the early sequence as compared to the later sites. Pleistocene extinctions removed a number of species from the landscape.

Second, the Pampas have a larger species diversity throughout the sequence. The Arroyo Seco 2 site is a very diverse assemblage, but other sites in the Pampas have large faunal inventories as well. In the better preserved settings of caves in rockshelters of the Deseado and Magellan areas of Patagonia, the faunal diversity shifts slightly between the two periods.

Tool and raw material organization

The final comparison of sites is of tool and raw material organization. The Miotti and Salemme data already hinted at the relatively informal tool technology employed at many sites.

The availability, abundance, and quality of raw materials affect the resultant technology employed at both the site and regional level. In Patagonia, where raw material sources are common, the tools show the use of such local resources. For example, Borrero and Franco (1997) recently summarized the raw material sources for sites in the Deseado and Magellan areas.

Table 11:
Raw material sourcing from select sites in Patagonia
(modified from Borrero and Franco 1997: Table 2)
Slope Region Site Level Local
Atlantic slopeDeseado River El Ceibo level 12 907
(99.89% (total sample))
1 obsidian flake
(0.12% (total sample))
Atlantic slopeDeseado River Los Toldos, Cave 3 level 11 mostly local a few obsidian flakes 473
Atlantic slopeDeseado River Piedra Museo Lower level mostly local no data no data
Atlantic slopeDeseado headwaters Arroyo Feo 1 level 11 no data 10 obsidian flakes (2.38% (total sample)) 421
Atlantic slopeDeseado headwaters Cueva de las Manos level 6; base and intermediate section no provenance date obsidian (17.1% (tool sample))of intermediate section. No data for base section) 1005
Pacific slopeUltima Esperanza Sound Cueva del Medio   mostly local chalcedony not available in the immediate vicinity no data (partial description)
Pacific slopeUltima Esperanza Sound Cueva Lago Sofía   37 (100% (total sample)) 0 37

There is almost a complete reliance upon locally available raw materials, with the exception of Cueva de las Manos, which has about 1/5 of the assemblage from more distant sources. In the Andean foothills, sites such as Agua de la Cueva also show a reliance upon local materials.

In the Pampas, the data show a shift to more distant sources, but the distribution of raw material sources is much more patchy in the Pampas. At sites such as Cerro el Sombrero, raw materials appear to be coming from 30-60 km distant, although more local sources were also utilized. Cerro la China also shows the use of semi-local and local sources, but this site is located on top of a raw material source. At Cueva Tixi, the reliance is upon both local and distant sources. The isolated Fell's Cave points from the Pampas, located on the plains and coast (not in the foothills), are made of raw materials located between tens and hundreds of kilometers distant.

Formal tools are relatively uncommon, especially those made on biface blanks. Additional data from Borrero and Franco shows that very little work was put into bifacial tool production.

Table 12:
Presence of bifacial tools for select sites in Patagonia (Borrero and Franco 1997: Table 6)
  Biface Preforms Biface reduction flakes Biface projectile points
Los Toldos 3
Level 11
no no data no
Los Toldos 2
Level 10
no no data no
El Ceibo no 1 (0.2% (flake sample)) no
Piedra Museo no data no data 1
Cueva de las Manos 1 (0.82% (tool sample) ) no data 2 (1.65% (tool sample))
Arroyo Feo
Level 11
1 (2.85% (tool sample) ) no data no
Cueva Fell no data no data yes
Cueva del Medio 1 yes yes
Cueva Lago
Sofía 1
1 (16.67% (tool sample) ) 3 (13.4% (flake sample) ) no
Tres Arroyos no yes 2 (18.18% (tool sample) )

In fact, the occurrence of the Fell's Cave stemmed points over such a large region appears anomalous. This projectile point type shows up in all three regions, at sites such as Cerro la China, Cerro el Sombrero, Piedra Museo, Los Toldos 3, Fell's Cave, Palli Aike, Tres Arroyos, in addition to isolated points recovered over the Southern Cone.

Many of these sites are nearly contemporaneous in age. From a North American perspective, arguments would be made that the groups making these tools were highly mobile foragers, ranging over great space. However, the raw material organization and tool production shows reliance upon local sources, and so-called expedient technology. Many of the points are made on unifacial blanks (Politis 1991).


The archaeology of the late Pleistocene and early Holocene of the Pampas and Patagonia is a very rich record, and there is large amount of data available for comparative analysis. One of the most obvious similarities in many Southern Cone sites is the widespread presence of Fell's Cave stemmed points. At one time, North American researchers, in particular, envisioned these tool users to be highly mobile colonizing populations.

However, despite the Fell's Cave projectile point as a true horizon marker, there is little evidence for extensive mobility, for exclusive big-game hunting, or for a highly curated lithic technology. Most sites show use of locally available resources.

The timing of the peopling of the Southern Cone appears first in the extreme south, and then moves to the north, and finally into the Andes, during the interval from slightly before 11,000 rcybp to mid-10,000 rcybp. There are several sites with even older dates that do not appear to be controversial, and might push the sequence further back. The lithic technology from most sites is a mix of unifacial and bifacial reduction, with simple flake tools being more common. The lack of investment in complex bifacial forms might signal the lack of need for transport of toolkits over long distances. Given that most of the fauna that commonly appear in sites are ubiquitous in their individual environments, it is doubtful that these hunter-gatherers were necessarily tracking game across the landscape. Most hunting was probably opportunistic intercept hunting.

The faunal data should be more thoroughly studied, as not all the site fauna is necessarily cultural in origin, nor is the observed attrition necessarily the result of human processing. Nevertheless, the remarkable diversity evident in multiple regions of Argentina is very encouraging, in that it allows for animal community reconstructions, as well as demonstrating the high degree of preservation available in many sites in the Southern Cone.

In conclusion, I have tried to document the rich record of the Southern Cone, and discuss some patterns evident in the data. Hopefully future research can be conducted in areas outside of the four that I have discussed, so that a truly representative sample can be drawn. Regardless, there is a high degree of patterning within the Argentinean data that has been published thus far.

Proceed to:
   Appendix 1
   Appendix 2
   Appendix 3
   Foot Notes
or, return to:
   Opening/Abstract Page
   Paleoindian & Other Stuff