Environmental Setting, Resources and Archaeological Visibility

The Fuegian steppe in the northern sector of the Isla Grande de Tierra del Fuego (Figure 2a, b, c) extends from the Magellan strait to the low hills of Viamonte Cape, south of the Grande river (Cabrera Reference Cabrera1976; Collado Reference Collado, Godoy Manríquez and Paileleo2007; Coronato Reference Coronato, Godoy Manríquez and Paileleo2007). The climate is cold-temperate, with oceanic influence, and annual rainfall ranges between 300 and 350 mm (Cogliati and Cuello Reference Cogliati, Cuello, Godoy Manríquez and Paileleo2007; Coronato Reference Coronato, Godoy Manríquez and Paileleo2007). Such low precipitation, an average annual temperature of 5.5 °C, and a persistent and strong wind regime generate poor conditions for water preservation in summer (Iturraspe and Urciolo Reference Iturraspe and Urciuolo2002, Reference Iturraspe, Urciuolo, Godoy Manríquez and Paileleo2007). Persistent west-northwest winds generate a dry effect in summer and cool temperatures in winter. The annual mean wind speed is 23 km/h, though it can occasionally reach 140 km/h, particularly in the spring and summer seasons (Iturraspe and Urciolo Reference Iturraspe, Urciuolo, Godoy Manríquez and Paileleo2007).

Figure 2.

Fuegian steppe environment. (a), (b), and (c): landscape; (d), (e), and (f): detail of the cover vegetation where Festuca gracillima dominates. Photos by Jimena Oría.

The area is marked by numerous deflation hollows defined as “pans” (Villarreal and Coronato Reference Villarreal, Coronato and Rabassa2017), formed in dry lands mainly as a result of aeolian deflation. Most of these pans are the depositional centers of endorheic basins that receive short and ephemeral streams. In some periods many of them sustain a shallow lake, locally affected by wave action as a consequence of permanent and intense action of the western winds. This generates significant erosion in the eastern coastal areas (Villarreal and Coronato Reference Villarreal, Coronato and Rabassa2017).

Local vegetation is characterized by herbaceous steppe flora, including Festuca gracillima (Figure 2d) and Empetrum rubrum. In some sectors, shrubs of Chilliotrichum diffusum are also present (Figure 2e, f). Grasslands or meadows develop along the streams, due to more humid conditions (Collado Reference Collado, Godoy Manríquez and Paileleo2007). From a physiographical viewpoint, this sector of the island is influenced by several geomorphological and environmental factors:

• • low ranges and hills, with altitudes up to 300 m asl

• • low-density channel network

• • alternated events of flooding and desiccation of river channels and lake basins

• • steppe vegetation

• • low temperature and precipitation, with strong winds

• • soil types developed from sandy materials, unfavorable for a dense vegetation cover

• • freezing soils

All of these factors, in addition to the trampling and herding, lead to unstable conditions in certain surface sections, giving rise to hollows or hydro-aeolian depressions. This surface denudation, however, increases archaeological visibility, which is otherwise quite low along the Fuegian steppe (see Figure 2). Altogether, these conditions result in a differential preservation of the record.

A look upon the present landscape shows little variation; however, hunter-gatherer populations that inhabited the steppe at different moments of the Holocene used to have a solid knowledge (sensu Rockman Reference Rockman, Rockman and Steele2003) of the location and spatial distribution of significant resources, including water, wood, lithic raw materials, fauna, and shelterFootnote ¹.

From a regional perspective, lithic raw materials (the topic of this article) were extracted from secondary sources, primarily inland glaciofluvial deposits and/or gravels that are available in different environments of the island. The distribution of raw material has been described as “heterogeneous” for the San Sebastián bay (Figure 1a) area. Although access to cobbles might have been limited to a few sectors of the marine coast to the south of the bay (Franco and Borrero Reference Franco, Borrero, Aschero, Korstanje and Vuoto1999), there are large areas of several square kilometers where cobbles are completely absent, particularly along the northern bay coast (Borrazzo Reference Borrazzo2010, Reference Borrazzo, Oría and Tivoli2014).

At the inland area, nearby Lake Amalia, the raw material sources are glaciofluvial deposits (Figure 3). Gravel deposits are recorded in diverse points of the landscape. Even though gravel levels do not yield continuous exposure, outcrops are not too distant, and there are no known limitations to accessibility. As opposed to the heterogeneity proposed for the sea coast, inland lithic raw materials are an available resource used in the past (see Santiago, Bujalesky, and Salemme Reference Santiago, Bujalesky, Salemme, Morello, Martinic, Prieto and Bahamonde2007). Consequently, this resource should not be considered as scarce or of limited accessibility.

Figure 3.

Lithic raw material at a regional level. (a) stratigraphic position of cobbles used for knapping; (b), (c), and (d): gravel outcrops; (d) and (e) boulders in different sectors of southern (d) and eastern (e) coasts of Lake Amalia; (f): cobbles in the dry bottom of Lake Arturo. Photos by Jimena Oría.

Theoretical Framework

Distributional archaeology and landscape ecology (Dunnell and Dancey Reference Dunnell and Dancey1983; Foley Reference Foley1981; Stafford and Hajic Reference Stafford, Hajic, Rossignol and Wandsnider1992) are suitable approaches for the study of archaeological finds of this sector in the Fuegian steppe (Belardi Reference Belardi2005; Belardi and García Reference Belardi and García1994; Belardi et al. Reference Belardi, Fernanda García, Campan and Borrero1998; Borrero and Lanata Reference Borrero, Lanata, Borrero and Lanata1992). Relatively continuous, though low in density, surface records and limited stratified contexts are often described in the area (Oría Reference Oría2012; Santiago Reference Santiago2010; and bibliography cited there).

Regarding this issue, Stafford (Reference Stafford1995) focused on the theoretical relationship between geomorphological analysis and models explaining the links between settlement patterns and landscape structure. Landscape ecology considers landscape structure as the way its elements are arranged, defining these elements as discrete ecosystems that represent a space in which an attribute—landform, soil, vegetation, etc.—is homogeneous (Stafford Reference Stafford1995; Stafford and Hajic Reference Stafford, Hajic, Rossignol and Wandsnider1992). The distribution of these elements in the landscape could have influenced hunter-gatherers’ mobility strategies. Similarly, their frequency and spatial distribution would have conditioned space-use patterns and, at least partially, the spatial arrangement or structure of the archaeological record (Stafford Reference Stafford1995).

Distributional archaeology is suitable for the model presented here in terms of the spatial scale and analysis of the archaeological record of a whole region rather than individual sites (Schlanger Reference Schlanger, Rossignol and Wandsnider1992). In fact, distributional studies that encourage the artifact as the analytical unit (i.e., nonsite archaeology, Dunnell and Dancey Reference Dunnell and Dancey1983; Foley Reference Foley1981; Thomas Reference Thomas and Mueller1975) provide a clear view of the structure at the regional level. From this perspective, the archaeological record was categorized as isolated findings, concentrations and sites based on the density of recorded artifacts sensu Borrero and colleagues (Reference Borrero, Lanata, Ventura, Borrero and Lanata1992). The analysis of variations in artifact density reveals the relationship between the structure of the archaeological record and certain elements of the landscape, such as lakes (shallow lakes, in this case), hills, river plains, etc.

Additionally, Schlanger (Reference Schlanger, Rossignol and Wandsnider1992) coined the term “persistent place,” a concept defined as a place repeatedly used for a long time on a regional scale. The integration of isolated findings, minor concentrations, and sites highlights the potential of the archaeological record to analyze space use and reuse during long-term occupations. This focus broadens the scale of the analysis, including not only the individual loci, but also the isolated findings and minor concentrations scattered along a locality such as Amalia. This theoretical framework supports the idea that past populations organized recurrent visits to this locality or to a particular sector of it, where raw material and faunal resources were available.

Several occupation episodes are recorded in many of the sites in the area: (a) Tres Arroyos 1 rock shelter yielded the earliest human occupation for Tierra del Fuego, circa 10,500 ¹⁴C years BP (Massone Reference Massone2010; Massone and Prieto Reference Massone and Prieto2004; Massone et al. Reference Massone, Alfredo Prieto, Gabriel Cárdenas and Cárdenas1999), as well as late Holocene and historical times’ records in the upper levels. (b) Discrete superimposed occupations in shell middens were dated between circa 5500 and 3200 ¹⁴C years BP (Morello et al. Reference Morello, Borrero, Torres, Massone, Arroyo, McCulloch, Calas, Lucero, Martínez, Bahamonde, Salemme, Santiago, Álvarez, Piana, Vázquez and Mansur2009; Prieto et al. Reference Prieto, Elisa Calás and Torres2007; Salemme and Santiago Reference Salemme, Santiago, Vázquez, Elkin and Oría2017; Salemme, Santiago and Oría Reference Salemme, Santiago and Oría2014; Salemme et al. Reference Salemme, Bujalesky, Santiago, Morello, Martinic, Prieto and Bahamonde2007; Santiago Reference Santiago2013; Santiago, Oría, and Salemme Reference Santiago, Oría and Salemme2007). (c) A large number of archaeological sites were radiocarbon dated to the last two millennia, suggesting a significant demographic growth during this period. As stated by Morello and others (Reference Morello, Borrero, Massone, Stern, García-Herbst, McCulloch, Arroyo-Kalin, Calás, Torres, Prieto, Martínez, Bahamonde and Cárdenas2012:76), “Redundancy and reoccupation in the use of sites is frequent, and all the available environments on the island appear to be occupied.” Until absolute dating is available, two chronological alternatives may be considered to interpret the record at Lake Amalia: (a) a “long count” for human presence at the Fuegian steppe (ca. 10,500 BP), or (b) a more probable occupation scenario dated to the middle and late Holocene.

Radiocarbon dates were obtained from interbedded soil, ash, and aeolian sequences at other lakes: Arturo (Coronato, Fanning et al. Reference Coronato, Patricia Fanning, Jimena Oría and Ponce2011; Coronato, Salemme et al. Reference Coronato, Salemme, Moretto, Oría, Ponce, Orgeira, Onorato, Vásquez, Osterrieth, Fernández, Laprida and Benvenuto2011); O'Connor (Oría, Villarreal et al. Reference Oría, Villarreal, Coronato, Salemme and López2014); Las Vueltas 1, and Perro (Santiago Reference Santiago2013; Santiago and Salemme Reference Santiago and Salemme2016). These offer a chronological framework that contributes to the archaeological picture for the area, though the provenience of dated remains is sometimes hard to identify. In Lake Amalia, several edaphized levels were dated; the first is in Amalia 4, on the southern coast of the lake, that directly overlies tertiary substrate (3020 ± 60 years ¹⁴C BP - LP-3068; Oría [Reference Oría and Mena2016]; Oría, Villareal et al. [Reference Oría, Villarreal, Coronato, Salemme and López2014]), the second in Amalia 5, located on the eastern coast (360 ± 30 years ¹⁴C BP - Beta-346885; Oría [Reference Oría and Mena2016]; Oría, Villarreal et al. [Reference Oría, Villarreal, Coronato, Salemme and López2014]). In Amalia 6, radiocarbon dates obtained from edaphized levels suggest climate amelioration and stabilization of the surfaces, in at least nine opportunities since the middle Holocene (ca. 5000 years BP; Coronato et al. Reference Coronato, Mónica Salemme and Mari2017). This shows that, at least in some sectors of the lake, early Holocene sediment was either not deposited or not preserved; this may point to a relatively later age for occupation of this locality.

Discussion: The Regional Scale

A significant database has been compiled in recent years (Oría Reference Oría2012; Santiago Reference Santiago2010) reflecting the spatial distribution and characteristics of the archaeological record from the area between the Chico and Grande river basins (Northern Tierra del Fuego, Argentina). Santiago (Reference Santiago2010) described the lithic contexts from nine sites either in stratigraphy or on the surface: La Arcillosa 2 (LA2), Río Chico 1 (RCh1), Las Vueltas 1 (LV1), Perro 1 (Pe1), Herradura 1 (H1), Avilés 1 and 3 (A1 and A3), and Pozo Tierra del Fuego 2 and 3 (PTF2 and PTF3; Figure 1a). Core frequencies vary widely at these sites (Table 3) and are better represented at Pe1, H1 and A1.

Table 3.

Frequency of Cores, Hammerstone, and Anvil from Sites in the Study Area.

Recorded cores and nodules, sometimes tested or used as hammerstones, show relatively high frequencies and sizes directly linked to the sources of raw material exploited (oval glaciofluvial cobbles smaller than 150 mm at their maximum axes). They were portable pieces that could be easily replaced; cobbles are readily available.

LV1, located in the eastern coast of Lake Las Vueltas, was reoccupied at least three times during the late Holocene. It is interpreted as a kill and butchering site, where the main prey were guanacos (Santiago and Salemme Reference Santiago and Salemme2016). Cores seem to have been obtained from pebbles collected in the glaciofluvial deposits close to the site, next to the lake shoreline (Santiago et al. Reference Santiago, Pal and Salemme2009), but there is no reference to their size. As expected, due to the interpreted functionality of LV1, there is a low representation of cores and a very high frequency of debris (Santiago Reference Santiago2013); this is typical of task places where edges have to be constantly retouched (Binford Reference Binford1978). In Lake Las Vueltas, archaeological evidence of workshop or quarry is absent. Its classification as a specific hunting place without incidence in lithic raw material acquisition practices is acceptable.

In the case of the surface sites Pe1, H1, A1, A3, PTF2, and PTF3, mentioned by Santiago (Reference Santiago2013), cobble cores are well represented. The most frequent shapes are amorphous, bipolar, and globular cores. In the Pe1 site, a hammerstone and an anvil were also recovered; the anvil, however, could have been used for marrow extraction (Santiago Reference Santiago2013). Bipolar technology is well represented in the H1 site, in terms of cores and nodules; most of them were tested and discarded because of bad quality or the presence of internal flaws. Although there is no reference to cobble size, it seems clear that they were transported into the site and tested there. These cores’ types differed from those recorded in Lake Amalia (Supplemental Table 1). In the sites Amalia 4 and 5 the “tested nodule” type is not the most representative.

Similar characteristics of Pe1 offered the record from A1, A3, PTF2, and PTF3 sites. At A1 there is solid evidence of the bipolar technique, whereas the high rate of cortex in the pieces from the other three sites is a consequence of a reduced number of flakes knapped from each core (Santiago Reference Santiago2013). At the PTF3 site, nodules were reduced through bipolar percussion, and both rough and retouched edges were used. No information about nodule weight and size was published.

The LA2 and RCh1 sites are both stratified shell middens dated ca. 5000 ¹⁴C years BP (Salemme, Oría et al. Reference Salemme, Jimena Oría and Turnes2014; Salemme, Santiago, and Oría Reference Salemme, Santiago and Oría2014; Santiago, Oría, and Salemme Reference Santiago, Oría and Salemme2007). Both sites are located in the ancient sea shoreline of the middle Holocene. The former, close to the shallow Lake La Arcillosa, not only yielded lithic items but also abundant faunal remains of vertebrates and invertebrates. They were interpreted as campsites.

The lithic collection from the LA2 site consists of 929 pieces (Salemme, Oría et al. Reference Salemme, Jimena Oría and Turnes2014). At RCh1, the total number of lithic items amounted for 1602 (Table 3). Refitting was possible for five core fragments. At these sites, six cores were used as hammerstones (2 in the LA2 site, and 4 in RCh1). This evidence suggests that knapping tasks took place at both sites, along with other activities. Cores (both complete and fragmented) were weighed, allowing a comparison with the information from other sites and localities (Figure 6a). Only four cores from the LA2 site weigh more than 1 kg (1067.9 g, 1450 g, 1817.1 g, and 2096.8 g, respectively); they are the outliers shown in Figure 6a.

Figure 6.

(a) Weight and (b) size of cores and nodules from different localities and settings. The items in Lake Amalia include cores, nodules, hammerstones, and anvils from Amalia 4, Amalia 5, and Yar Hayen sites. To compare the weights, a larger sample includes two stratified sites (LA2 and RCh1).

The inland archaeological survey, taking into account the main landforms—identified as hills, fluvial valleys, and shallow lakes (Oría Reference Oría, Salemme, Santiago, Alvarez, Piana, Vázquez and Mansur2009, Reference Oría2012)—followed the aforementioned distributional archaeology approach. Cores were recorded as isolated findings but also in concentrations of materials and at sites. Core frequency varied; the weight and size of the cores and nodules chosen for knapping were measured. In each different landform, 50% of the elements have a similar range (see boxes in Figure 6a and b). In the case of the stratified sites, a similar profile has been verified only in terms of weight.

Considering hills, 16 isolated findings are cores, nodules, and hammerstones. However, none of them are larger than 130 mm, and only a core and a hammerstone weigh more than 1 kg. In valleys, 18 items were recovered at the Chico river locality: one is an isolated finding, and the rest are distributed in two concentrations (C2 and C4) and two sites (RCh2 and RCh3). Among all of these findings, the largest core is 120 mm, but none weighs more than 1 kg.

In shallow lakes, aside from the Amalia locality, 18 items were found as isolated findings, nine in two concentrations (TM C1 and C2) and 73 at the sites known as Arturo 1 (Coronato, Fanning et al. Reference Coronato, Patricia Fanning, Jimena Oría and Ponce2011; Oría and Salemme Reference Oría and Salemme2016), Arturo 2 (Oría and Salemme Reference Oría and Salemme2016), and Tres Marías (Oría, Salemme et al. Reference Oría, Salemme, Santiago and Montes2010). From these 100 items, only 10 cores weigh more than 1 kg, with the largest measuring 145 mm.

Over 80% of the cores (both complete and fragmented) present more than 50% of cortex on the surface. The exploitation intensity of these cores is similar to that observed by Santiago (Reference Santiago2013) for the pieces in the nearby area.

The comparison presented in Figures 6 and 7 shows a slight difference with the data observed in the Amalia locality. Weight and MLD values in this locality show higher means compared to the others, but medians and percentiles are similar. However, it is important to highlight that these graphics do not take into account the refitting at Yar Hayen. Also, the frequency of core fragments in the sites at Amalia locality is significantly higher (Tables 2 and 3). At least part of the cores considered in the boxes of Amalia—equivalents in ranges with those from other localities or landforms—are not complete pieces.

Figure 7.

Maximum linear dimension (MLD) of cores and nodules in different localities and settings. MLD values were plotted directly (a), and then a logarithmic transformation was applied (b) to increase homogeneity in the distribution of the data.

Numerous gravel outcrops were locally recorded, with a notable difference in grain size from the cobbles and gravels exploited as raw material in the Amalia locality. Neither the available cobbles nor the ones definitely used in other localities (Figures 7 and 8) coincide with some nodule types selected in Lake Amalia (see outliers for this locality in Figures 6a and 7a and the example in Figure 8). Without taking into account the refitting masses already mentioned, at Lake Amalia the maximum values, as well as the coefficient of variation, are higher, showing a different use of raw material and the probability of a wider range of sizes and weights available. In the case of the Yar Hayen site, refitting indicates that cobbles up to 9 kg were knapped using a hammerstone and an anvil that weighed 4 kg each. This data may be interpreted in two different ways: a) people were working in situ, having moved to the lake shoreline; or b) they shared a deep knowledge of the availability of a different resource that was restricted to Lake Amalia and uncommon in other areas.

Figure 8.

Cobbles and pebbles at the bottom of Lake Amalia. Photos by Jimena Oría.

Following Rockman (Reference Rockman, Rockman and Steele2003), environmental knowledge considers three types of information: locational, limitational, and social. These sets of information are not exclusive, but rather mutually reinforcing. The Amalia locality can be “thought about” in these terms and their connections: “Locational knowledge includes information relating to the spatial and physical characteristics of particular resources… It also includes the ability to relocate such resources after their discovery” (Rockman Reference Rockman, Rockman and Steele2003:4). In this sense, the recurrent visits to the Lake Amalia would not only have economic implications but also social meaning. The former would include those resources that acted as “pull” factors (sensu Rockman Reference Rockman, Rockman and Steele2003), whereas the latter is derived from the need to share some degree of social interaction and information.

Leaving aside the peculiarity of the Yar Hayen workshop, the contexts from the different sites and concentrations in the Amalia locality suggest recurrent use of this space. Borrero (Reference Borrero1985) proposed a pattern to describe the spatial occupation of the northern Tierra del Fuego, which is characterized by high residential mobility with logistical components, particularly in areas lacking certain resources such as raw materials in several sectors of the San Sebastián bay (Borrazzo Reference Borrazzo, Oría and Tivoli2014; Borrazzo et al. Reference Borrazzo, Martin, Barberena, Borrero, Borrero and Franco2008). This model is visible in the archaeological record from the marine coast and surrounding area of the depression San Sebastián-Inútil bays. The archaeological record of this area for the last two millennia shows not only differences at the technological level but also in raw materials used (Borrazzo et al. Reference Borrazzo, Martin, Barberena and Borrero2007, Reference Borrazzo, Martin, Barberena, Borrero, Borrero and Franco2008). Sites farther from the marine coast present lower artifact densities. Additionally, they yield a higher frequency of tools and raw materials of good knapping quality. Borrazzo and others (Reference Borrazzo, Martin, Barberena, Borrero, Borrero and Franco2008) proposed a curated technology for these areas, as opposed to the more expeditious ones present in coastal localities such as Los Chorrillos. Considering that the depression is large enough for the movements of a hunter-gatherer band, the aforementioned differences, among other lines of evidence, support the idea of certain logistical behaviors to exploit the resources available at each space. The model of high mobility is still appropriate, but the significant record heterogeneity found at the depression San Sebastián-Inútil bays suggests a more complex use of the space.

High mobility generates a record characterized by small sites, such as the ones widely observed in the Fuegian steppe. For the northern coast of Tierra del Fuego, the archaeological record shows “both economic and use of space strategies that include frequent visits of a short or medium duration, which may be more frequent in some places than in others” (Horwitz Reference Horwitz, Borrero and Barberena2004:49, translation by author). In San Sebastián bay, the record at the Los Chorrillos locality is interpreted as the result of repeated occupations (Horwitz Reference Horwitz1995), due to the quantity of stratigraphic loci in the dunes and surface artifact density. Although the evidence points towards a recurrent occupation, specific redundancy is low. This means that “the mobility cycles included periodic occupations, but not necessarily of the same locus” (Horwitz Reference Horwitz, Borrero and Barberena2004:45; translation by author).

Southwards, in the area between the Chico and Grande rivers where Lake Amalia and the other localities included in this study are located, the archaeological record adjusts to the high mobility model. In contrast, moving inland, the archaeological record does not fit the logistic elements in the model. At the Amalia locality, samples with high frequencies of tools are recorded. They were knapped from the best quality raw material found at a regional scale, including green obsidian and silicified tuff from the Miraflores locality (see Borrazzo et al. Reference Borrazzo Karen, Borrero, D'Orazio, Etchichury, Massone and Angelis2015 for a discussion about this source of raw material). At the same time, other contexts, such as the Yar Hayen site, exemplify in situ primary exploitation of raw material. Thus, this heterogeneity of the record may be manifested even at the same locality.

Broadly speaking, the inland area shows a scenario similar to the one described by other researchers along the northern coastal sector: the record points to a general use of the space, with ephemeral settlements and very low specific reoccupation. Lake Amalia is probably the only case in which a recurrent visit pattern can be verified at the locality level (Oría Reference Oría2012, Reference Oría, Oría and Tivoli2014; Oría, Salemme et al. Reference Oría, Salemme, Coronato and Vázquez2016; Oría, Salemme, and Fernández Reference Oría, Salemme and Fernández2014; Vázquez et al. Reference Vázquez, Oría and Salemme2016).

According to the record available, hunter-gatherers used this shallow lake intensively throughout the Holocene. In several lakes nearby, isolated findings were also recovered, but their frequency is considerably lower than at Lake Amalia (Oría Reference Oría and Mena2016). The diversity of the record indicates a rather varied and repeated use of the lake area, as exemplified by two relevant sites: Yar Hayen (Oría and Pal Reference Oría and Pal2011) described here, and Yowen Ko, a place that presents a high density of guanaco bones with clear anthropic marks (i.e., cutmarks, thermal alteration), together with some charcoal pieces (Vázquez et al. Reference Vázquez, Oría and Salemme2016). The high number of tools (quite unusual for this area, even in surface contexts) at the Amalia 4 and Amalia 5 loci suggests the reoccupation of the sites or, alternatively, recurrent visits to the area. The evidence for reactivation after fracture in some of the artifacts from Amalia 5 further suggests this interpretation (Oría, Salemme, and Fernández Reference Oría, Salemme and Fernández2014).