Microwear and Plant Residue Analysis in a Multiproxy Approach from Stone Tools of the Middle Holocene of Patagonia (Argentina)

1 Introduction

The studies carried out to identify the functionality of archaeological lithic artifacts covered two lines of analysis that were developed historically in parallel, use-wear analysis and the study of residues of plant and animal origin. For the latter, it was previously pointed out that rigorous methodological validations were rarely applied during the past century, and it was warned about the bias errors that can be generated when drawing conclusions from indicators derived from the application of singular approximations (Anderson-Gerfaud, 1980; Grace, 1989; Lombard & Wadley, 2007; Mansur, 1984; Shanks et al., 2005; Zurro & Gadekar, 2024, among others). However, in the last two decades, the approaches that combine both mentioned lines, together with other methodologies to accurately evaluate the use of artifacts in plant processing have increased (Ciampagna, Cueto, Lema, & Capparelli, 2020; Clemente, Risch, & Zurro, 2002; Hayes, Fullagar, Mulvaney, & Connel, 2018; Lombard & Wadley, 2007).

We consider that the study of plant management in hunter-gatherer societies forms a key line for evaluating changes and trends in collection, transportation, exchange, and consumption practices, including conditioning and transformation of lithic artifacts. In continental and insular Argentine Patagonia, all these subjects have been addressed from archaeobotany based on the study of macro-remains, artifacts, and filling structures made from plants and micro-residues (Berihuete Azorín, 2010; Caruso Fermé, 2012; Ciampagna, 2015; Cueto & Andreoni, 2016; Pérez de Micou, 2002; Piqué i Huerta, 1999). Particularly in the continental sector of this region, the density of archaeobotanical evidence decreases from north to south. The province of Santa Cruz has the least amount and variability of plant remains, with its lowest expression in the Central Plateau. This could be due, on the one hand, to a lower environmental availability of plants since the vegetation decreases in density, diversity, and height both from north to south, and from the Atlantic coast and the mountain range towards the interior of the plateau (Ciampagna, 2015). Additionally, methodological issues since archaeobotanical studies in the province are relatively recent (Caruso Fermé & Aschero, 2020; Ciampagna, Molares, Ladio, & Capparelli, 2021). Also, it could be attributed to social factors that imply differences in the post-harvest practices (Ciampagna, 2015).

The few archaeolobotanical macro-remains recovered from continental Patagonia made it possible to identify, up to the present, a restricted range of use categories that are predominantly for fuel, but also for food, medicinal, and technological uses (Caruso Fermé, 2012; Ciampagna, 2015; Cueto & Andreoni, 2016). There has been a recent emphasis on the study of micro-residues from lithic milling artifacts, ceramic sherds, and coprolites allowed for the broadening of the discussion on use categories (Ciampagna et al., 2021).

The studies that combine the analysis of use-wear traces and the examination of vegetal micro-remains on archaeological artifacts in continental and insular Argentine Patagonia, although scarce, constitute a valuable source of evidence to increase such understanding of the modes of use of plants, the selection criteria, techniques involved in the different stages of their conditioning and transformation, and the contexts in which they were carried out (Álvarez et al., 2009; Cueto, 2015; Mansur, 1984). The functional analysis of lithic artifacts has a long history in the region, both in the continental portion (Castro, 1994; Cueto, Frank, & Castro, 2017; Lynch, 2016; Mansur, 1984; Paunero, Castro, & Reyes, 2007) and in the insular segment (Álvarez et al., 2009; De Angelis, 2015). If we focus on the continental Central Plateau of Santa Cruz, general trends and particular characteristics were recognized for some archaeological periods, registering changes and continuities in the way of using instruments. Such characteristics arise from the use-wear analysis of archaeological instruments from different sites (Cueto, 2015; Cueto et al., 2017; Lynch, 2016). These indicate that the processing of wood was inferior to that of other substances, such as leather, meat, and bone, during the final Pleistocene first peopling of the southern tip of the continent. Along the early Holocene an increase in its processing is recorded, reaching values next to those of leather, the most exploited substance (Cueto, 2015; Lynch, 2016). For later occupations assignable to the middle Holocene, wood is among the most worked material together with bone and leather. Here, the first traces attributable to the hafting of compound instruments generated by intermediaries made of wood also appeared (Cueto et al., 2017; Paunero et al., 2007). Finally, in the late Holocene occupations, wood occupies the second place behind bone (Cueto, 2015).

Consequently, we wonder if it is possible to identify which plants were processed with lithic artifacts, for what purpose, and in what way these works were done, if wood was used to haft the instruments, also when and in what areas these technological practices occurred. To answer these questions, we are developing a line of work aimed at identifying both use-wear traces and micro-residues of plant origin on lithic artifacts. The first step was to design an experimental program aimed at identifying and characterizing the use-wear traces and the plant residues adhered to the rim, edges, and active surfaces of the pieces used to process wood in order to generate a reference base. Methodology and protocols of this program were published by Ciampagna et al. (2020), and a brief synthesis is given here. From this reference base, we are able to deepen the examination of the production and consumption strategies of lithic artifacts as well as the practices of collection and use of plant species implemented by the hunter-gatherer societies that inhabited the Central Plateau.

Our objective in this article is to use the experimental program as a reference base, to evaluate the functionality of an archaeological assemblage of artifacts corresponding to the mid-Holocene occupation of the La Mesada site located in La María archaeological locality at the Central Plateau of Santa Cruz (Figure 1a) through a multi-proxy perspective that combines the use of wear analysis and plant residue identification.

Figure 1

(a) Study region. Sites: LT: Los Toldos 2, 3, 13; CM: Cueva Maripe; AEP1: Alero El Puesto 1; LM4: La Martita 4; VQ: Viuda Quenzana; EV: El Verano; LM: La Mesada; LL: La Lavandería; LV: La Ventana; CT: Cueva Túnel; CDM1: Casa del Minero 1; LG2: La Gruta 2. (b) La Mesada. (c) Landscape, plateaus, canyons with caves, shallows, and lagoons.

1.2 Experimental Program as a Reference Base for Patagonia

We elaborated an experimental collection of lithic artifacts on reddish-brown flint from the La María locality: a rock widely available in the Santa Cruz Central Plateau and representative of most of the archaeological artifacts recovered there. We worked logs of two woody species Berberis thunbergii DC and Neltuma chilensis (Molina) C.E. Hughes & G.P. Lewis. We made a sample of disaggregated reference material composed of a dissociated cell and tissue remains diagnostic at the genus level, from which the anatomical characterization of the residues recovered from the artifacts was carried out (Ciampagna et al., 2020).

The surface of the artifacts was characterized, using a stereoscopic microscope (SM) and metallographic microscope (MM), prior to their use and consequently free of plant residues. Variables such as appearance, gloss, grain, and homogeneity/heterogeneity were considered (Figure 2a, b, and f). With the artifacts, debarking and scraping work was carried out on fresh wood logs. A scraper was not used and remained as a control artifact. The work carried out on the wood involved transverse movements to the active edge, with different angles of attack and pressure, for a total of 30 min, by manual pressing on logs (Table A1). After 30 min of use, the initial scraping (IS) was carried out, which consisted of the extraction of residues from one-half of the worked edge. The extracted material was analyzed using the optical transmission microscope.

Figure 2

Flint. (a) Cortex: thin, fine grain, wavy topography, smooth, silky appearance and intense shine. (b) and (f) internal surface: smooth and soft appearance, matte shine, very fine and homogeneous grain. (c) and (d) Artifacts No 10 and No 2 with plant residues. SM microphotos: (e) and (f) Neltuma bark, 90× and 10×, artifact No 4. Berberis 60×: (g) Xylem, artifact No 8. (h) Bark – white arrow, stains – black arrow, artifact No 3.

Cleaning procedures were then applied – NaOH at 20%, to remove organic residues; HCl at 10%, to remove carbonates and some inorganic residues; and combining both – to evaluate the incidence of each one on plant residues. After these washes, the final scraping (FS) was performed, taking a sample of residues from the other half of the worked edge (Table A1). The quantification of residues adhered to the pieces before and after the chemical treatment was calculated on the basis of the slide area occupied by the residues removed after the scraping of the edges. The entire sample was subjected to a final wash with a neutral soap solution, and subsequently, the functional analysis of the artifacts was carried out.

As a first result of the experimental program, the anatomical characterization of the dissociated cell types allowed us to recognize the differences between both genera, and after the IS and FS (Table A2). All the artifacts presented residues and plant adhesions on the edges and adjacent internal parts after working for 15 and 30 min. At the macroscopic and SM level, tissue fragments and fiber bundles were identified, especially in Neltuma bark, together with adhesions coating the surface of artifacts, stains, and isolated particles (Figure 2c–h).

At the microscopic level, a greater amount of plant residues was observed in artifacts that processed Berberis xylem than those that processed Neltuma xylem. Among the artifacts that processed bark, in the IS and FS, those that processed Neltuma had a higher number of residues compared to Berberis. In both species, from bark and xylem, disintegrated cellular elements and tissue fragments were recovered.

The three cleaning procedures partially removed the residues from both species; however, the washing with HCl caused a greater removal of plant residues than the other types of cleaning. None of these cleaning procedures eliminated the totality of the residues, which was observed when identifying the use-wear traces, immediately after FS (Figure 3a and b). Because of that, we removed the rest of the adherences with a final washing with a soap solution before using wear analysis. Regarding plant micro-remains analysis, we removed adherences before cleaning techniques were applied. This is more convenient due to the alterations caused by washing (degraded cell walls, fragmentation) that are similar to those of technological processing of wood with lithics.

Figure 3

Polishes and residues associated (150×), then 30 min work, of (a) Berberis xylem, artifact No 8. (b) Neltuma bark, artifact No 4. (c) and (d) Micropolish by scrape Neltuma bark – artifact No 10 – and xylem – artifact No 7 –, 300×, from Ciampagna et al., 2020.

After the final washing with soap solution, diagnostic and intermediate diagnostic micropolishes were recognized on all the artifacts. The latter are slightly shinier in appearance and in some cases very shiny, mostly intermediate in thickness with instances of thick development, and rough and smooth in appearance. These traces are distributed as a band parallel to the edge and on the cusps of the microtopography, extending marginally. The polishes that appeared in more sectors and that were more developed were formed in the artifacts used to work Neltuma bark and xylem (Figure 3c and d).

2 Regional Setting: Characteristics and Evidence of Mid-Holocene Occupation of the Central Plateau of Santa Cruz

The Central Plateau of Santa Cruz is a region of volcanic origin (Figure 1a–c). It presents a landscape interrupted by canyons, hills, valleys, and endorheic basins. Modern climate is arid and vegetation corresponds to a steppe with shrub and sub-shrub steppe. There is a diversity of fauna, in which guanaco (Lama guanicoe) and rhea (Pteronemia pennata) populations stand out. It has a large availability of lithic raw materials with a homogeneous distribution (Cueto, 2015). The region presents large formations of acid ignimbrites with caves and rock shelters. These preserve evidence of hunter-gatherer societies from the final Pleistocene to recent historical times. The plateau has hundreds of sites with rock art and open-air deposits (Cueto, 2015; Miotti & Salemme, 2004).

The mid-Holocene in this region was characterized by an arid climate. Contexts in stratigraphy attributable to this period are distributed in 14 sites – Figure 1a (Cueto, Iparraguirre, & Paunero, 2020). It has been proposed that hunter-gatherers had reduced residential mobility. Subsistence was based on hunting guanaco, indicating a specialized strategy. It has also been claimed that consolidation of regional social networks occurred, facilitating the long-distance movement of goods and information (Miotti & Salemme, 2004).

3 Materials and Methods

For the purpose of this work, we took a sample of six scraper-class artifacts from the mid-Holocene occupation of the La Mesada site to be analyzed under a multiproxy perspective (Table 1, Figure 4). The scrapers were selected based on integrity, technical, morphological, and functional criteria. The six were unbroken artifacts whose design was compatible with that proposed as suitable for hafting or on which evidence of this type of device was found, in contemporary assemblages. These instruments have a short distal edge, a laminar module, and retouching on one or both lateral edges. Also, the most representative raw materials of the assemblage (flint and chalcedony) were prioritized in the selection criteria, together with the presence of adhered sediments.

Figure 4

Archaeological studied artifacts, dorsal view.

For residue analysis, two samples per artifact were extracted before washing them, in order to prevent plant remains removal and/or alteration. The first was obtained from the distal portion, which involves a blade and adjacent surfaces on the dorsal and ventral sides. The second was obtained from the proximal and middle portion, considered the manual grappling or hafting sector. This involves the lateral edges/rims (right and left, in dorsal view of the artifact) and the adjacent surfaces on the dorsal and ventral face, and the butt sector. Both portions were scraped for 20 s, with a new plastic stick washed with 100% sodium hypochlorite. For each subsample, an average of 25.21 µL of dry material was extracted (Table 2), suspended in 50% glycerol in a centrifuge tube, and then spread over microscope slides with a micropipette with disposable tips and sealed with a cover slip. Two control samples of the site occupation sediment were analyzed (Table 2). One was close to the cave mouth, the other to the innermost grid of the site. Chemical treatment of the sediment samples was avoided, as suggested by Coil, Korstanje, Archera, and Hastorf (2003) in order to allow recovery of many microfossils with the least possible damage. These samples were suspended and treated in the same way than the residue samples. All the slides were scanned under a light microscope with polarized light Leica DM/LM, and photographed with a Toup 5MgPx camera. Sample recovery and slide mounting were carried out in a room free of any food substances (Mercader et al., 2017) and inside an acrylic box specially arranged for the extraction of microremains. All manipulation of the fragments was performed without gloves – since industrial maize starch is frequently found in this kind of accessories in Argentine – but with washed hands. Disposable laboratory supplies were first tested to prevent cross-contamination, they were washed with distilled water, and the sample was scanned under a light microscope to ensure that it was starch free.

The residues obtained were compared with reference material from the herbarium, xylotheque and micro-remains collection of the GIADA (Archaeobotany Group of the Archaeology Division of the Natural Sciences La Plata Museum), directed by Dr. Capparelli. The diagnostic features were recorded following standard protocols and the nomenclature of the IAWA List of Microscope Features for Hardwood and Softwood Identification (Richter, Grosser, Heinz, & Gasson, 2004; Wheeler, Baas, & Gasson, 1989), the ICPT et al. (2019), and the ICSN (2011). Hardwood residues were tested against quali and quantitative characters of the woody plants of the area descripted by Franch, Ciampagna, Mansur, Zubimendi, and Capparelli (2022); Bambusoideae (from now on canes), following Luo et al. (2019); and softwood residues, against the eight native species of Argentina (Zuloaga & Morrone, 1996), from which Fitzroya cupressoides and Austrocedrus chilensis (Roig, 1992) are the only ones where latewood thick-walled tracheids are present. Other published diagnostic characters from Ciampagna et al. (2020) and Giovannetti, Lema, Bártoli, and Capparelli (2008) were considered. Bos taurus tendons were cut to have references of the fresh material. Animal and plant tissue (carbohydrates) were differentiated by staining the slide with an Iodine-potassium iodine solution (Lugol).

The functional study of the scrapers was carried out with a Union ME-D MM and SM, and they were photographed with a Motic S6 camera. Edges/rims and dorsal and ventral surfaces of the distal, middle, and proximal portions were examined. The lithic artifacts were washed manually with a neutral soap solution to remove the remaining residues. The presence/absence of macro and micro use-wear traces – rounding and/or retraction of the edge, scars, striae – were recorded. For their identification as well as for the recognition of the polishes, various characteristics were considered – e.g., brightness, regularity, thickness, distribution, surface features – suggested by Cueto (2015) and Mansur (1984). In addition, the presence/absence of macroscopic – scars, striae, and rounding – and microscopic – polishing, scars, striae and bright spots – hafting traces were recorded, following the proposal of Rots (2008) and own experimental references (Cueto, 2021). Their distinction is based on precise traits, on the importance of a particular type of wear, on its recurrence and distribution, and the association between different traces – e.g., scars and shiny spots or striae, shiny spots and striae. These are clear hafting indicators, which are restricted to the part of the tool opposite the edge used, with a marked limit between both sectors (Rots, 2008). The residual and functional analyses were carried out as a blind test.

4 Results

The multiproxy examination made it possible to recognize use-wear traces exclusively in the distal portion of the scrapers. In contrast, in the proximal and middle portion, alterations and use-wear marks of the polished type, bright spots, and scars were recognized. Scarcely presence of animal and abundant plant residues, some with possible signs of anthropic modification, were found as follows (Table 2).

4.1 Examination of the Distal Portion

The use-wear traces and the residues of the edges of the distal portion of all the scrapers indicate that they would have been used to process mainly substances of animal origin such as leather (n = 3) and of vegetable origin such as wood (n = 2), while an edge was used to process an indeterminate hard substance.

There are three artifacts (342, 365, and 611) that have leather processing use-wear traces. These show a development with a transverse orientation to the edge. It is distributed as a band parallel to the edge, although it is also found in isolated points, on cusps of the microsurface, and within the scar negatives. This polishing has a rough appearance with sectors characterized by parallel semi-spherical holes and/or surface grooves, perpendicular to the edge, features, that provides a discontinuous appearance to its surface. These traces present an intermediate thickness, a slightly reflective appearance that varies between a dull shine in the chalcedony artifact (365) to a slightly shiny one in the flint artifacts. At SM, these edges show continuous and discontinuous scars that are very small with the retraction of some segments of the edge and the characteristic rounding of the work in transverse actions.

These three edges have exclusively plant microresidues – except for one with egg microresidues, in piece 611. The remains belong mainly to softwood (n = 26), although vessels and fibers assignable to canes in pieces 342 (Figure 5n) and 611 (n = 3) and one cf. Neltuma sp. vessel in piece 611 were present (Figure 6a). Softwood remains are present in the three pieces and consist on entangled balls, together with ripped and fragmented sections (part of the body and tapering pointed endings) of thin and thick-walled early and latewood tracheids and compression wood of Fitzroya cupressoides and Austrocedrus chilensis; furthermore, three fragments of a thread made of two elements of thick-walled latewood tracheids of F. cupressoides or A. chilensis, Z and S twisted/wrapped were found in piece 611.

Figure 5

Artifacts 589 and 342, type of evidence and location. (a) polishing point by hard substance, 150×. (b) entangle balls of thin and thick-walled late (compression) wood Trch. of Fitzroya cupressoides or Austrocedrus chilensis. (c) and (d) Neltuma sp. starch grain. (e) bright spots from contact with wood, 300×. (f) polish possibly from friction with mineral particles, 300×. (g) thread of thin and thick-walled latewood Trch. of F. cupressoides or A. chilensis Z twisted, (h) Neltuma sp. stomatal apparatus, (i) mesocarp tissue, (j) A. chilensis Trch. associated with sieve tubes, (k) vessel cane, (l) indeterminate polishing associated with scar, 150×; (ll) resin polishing, 300×; (m) Poaceae rondel silicophytolith, (n) fiber cane; (ñ) wooden polishing, 300×; (o) thread made of two elements F. cupressoides or A. chilensis Trch. (b–d, g, m–n, and o, scale 50 µm).

Figure 6

Artifacts 611 and 731. Vessels of (a) cf. Neltuma sp., (b) cf. Nothofagus sp., and (c) cf. Adesmia/Berberis (scale 50 µm).

Two other artifacts (357 and 731) have traces attributable to wood processing (Figure 7a and b). These microtraces present sectors with diagnostic development and orientation transversal to the edge. They are distributed as a thin band on the edge and also manifest in more internal points, on ridges, and associated with small scars. Portions of the 731-scraper edge exhibit numerous scars and microfractures, rounding, and shrinkage. The polishing has a smooth and thick appearance, and forms lobes. These are very bright clear reflective surfaces. At SM, these edges show continuous and isolated scars, some of them thick and with abrupt endings.

Figure 7

Artifact 357, type of evidence and location. (a) and (b) Polish by scraping wood, 200×. (c) Latewood gymnosperm Trch. (d) cf. Bambusoideae vessel, (e) tendon fascicle, (f) and (g) Neltuma sp. starch grain, (h) Indeterminate generalized polishing, 150×, (i) Scar, 200×, (j) cf. Asteraceae vessel, and (k) thin- and thick-walled late (compression) wood Trch. of F. cupressoides or A. chilensis, (l) sub-parallel striations, 75×, (ll) rounding and polishing on the dorsal ridge, 200×, and (m) crossing field pitting of Austrocedrus chilensis (c–g, j–k, m scale 50 µm).

The edges mainly record plant microresidues, except for two connective fiber remains in piece 357. The microresidues of both pieces are softwood (n = 12) and hardwood (n = 2), while vessels and plates assignable to cane were present only in piece 357 (n = 2) (Figure 7d). Softwood remains consist of parenchymatic cells and single thin and thick-walled late (compression) wood tracheids of F. cupressoides or A. chilensis – some S or Z spun, very ripped, fragmented, or only tips present – (Figure 7c); hardwood remains to cf. Asteraceae (piece 357, Figure 7j) and cf. Nothofagus (piece 731, Figure 6b) vessels. In artifact 357 a Neltuma sp. simple starch grain was also present (Figure 7f and g).

Artifact 589 presents indeterminate processing traces for the substance, with a transverse arrangement to the edge. These are located within flaking negatives and on ridges. In one sector it presents continuity as a band parallel to the edge. In other sectors, the traces develop on very reflective clear minerals, making their characterization difficult. There are a few polished points that have a thick and slightly shiny appearance, features that suggest similarities with polishes originating from the work of hard substances such as wood (Figure 5a). The edge exhibits trapezoidal and elongated scars with a simple or tapered termination. This artifact has exclusively plant microresidues. Softwood remains predominate (n = 11), and consist of entangled balls, tips, and single thin and thick-walled late (compression) wood tracheids of F. cupressoides or A. chilensis (Figure 5b). A cane vessel plate and a Neltuma sp. simple starch grain (Figure 5c and d) were recovered.

5 Discussion and Conclusions

The multiproxy approach, applied in a segmented manner on two portions of the instruments, requires integrating its results to evaluate the technological operations of the manufacturing processes and the use of these lithic artifacts. These are linked to the design of artifacts, the conservation of residues, and the strategies of supply and use of wood.

The use-wear traces identified in the distal portion indicate that three scrapers (342, 365, and 611) were used to process the skins. These substances would have been obtained from the prey that entered the site or the camps possibly located at the entrance of the La María Quebrada canyon, among which the guanaco predominates. With these artifacts, work related to cleaning, kneading, and scraping skins was carried out, which implies a transverse movement. These tasks are associated with the initial and final stages of preparation of the skins for the manufacture of goods. It is possible that these scrapers were used recurrently in these tasks, this is due to the development that the polishes present in terms of extension, location, and thickness, and the magnitude of the macroscopic alterations. The three artifacts present abundant plant microresidues from which softwood (at least Fitzroya cupressoides and/or Austrocedrus chilensis) predominate, followed by remains assignable to canes commonly used in Patagonia to make spear’s or arrow shafts (Caruso Fermé et al., 2011; Ciampagna & Capparelli, 2012). The presence of entangled balls of ripped and very fragmented fibers (similar to the observed in the experimental program) let us suppose that these remains are the result of brush/polish on this kind of wood, although no microtraces were identified that indicate this work. This could be due to the fact that very soft tissues (low densities of softwood in the study by Richter et al., 2004; Roig, 1992) or immature tissues in the case of canes (softer than the mature ones) were processed, which possibly did not produce microwear; but also because leatherworking could have hidden the previous work with this type of wood. Neltuma sp. vessel in piece 611 would suggest the contact of the edge with hardwood logs for a very short time since it does not record microtraces of hardwood (as those presented in the experimental program). Although there is no evidence of debarking activities from the plant microresidues, it is probable that the egg recovered could have been contained in bark rugosity. These edges were likely used to process softwood or occasionally contacted softwood when used to remove ropes or twisted tendon cords on hafted artifacts. It is interesting to note that the chalcedony artifact (365) only shows softwood work from the microresidues, which could be related to the greater fragility of this rock with respect to flint.

The use-wear traces identified in the distal portion of two other scrapers (357 and 731) indicate their recurrent use in wood processing. This is inferred by the polish development and by the continuity and morphology of the macroscopic alterations. The presence of cane, soft, and hardwood microresidues on the edges is consistent with the substances to which the use-wear traces refer. In particular, softwood is represented in both artifacts by xylem at least of F. cupressoides and/or A. chilensis. The vessel element of cf. Asteraceae and the starch grain cf. Neltuma in 357 suggests the processing of xylem, and probably bark, of these taxa – taking into account that starch grains are common in latewood and bark (Castro, 2009). This fact is supported by the diagnostic development and the high gloss of their polishes, similar to those recorded in the experimental artifacts used to process Neltuma chilensis logs. In artifact 731, the presence of a vessel element of cf. Nothofagus sp. stands out.

With these artifacts, the circumference of the log would have been modified longitudinally by means of scraping, with a movement transverse to the edge. In general terms, these practices can be linked to the modification of logs for technological purposes, as was documented for Neltuma and Nothofagus in Patagonia, used for making hafts, bows, instruments, and cradles (Caruso Fermé et al., 2011). Although Asteraceae is the family with the most mentioned historical uses in Patagonia, mainly medicinal (Ciampagna & Capparelli, 2012), there are no reports up to now about its wood use. The presence of animal residues in the distal sector of 357 pieces, especially tendon fibers, could be due to remains of the fixings used for their hafting. They would have reached this sector due to breakage or wear.

Scraper 589 shows microwear evidence assignable to the processing of dry wood, related to the lower intensity of gloss that the polishes of specific sectors have. Plant microresidues are similar to those of the distal part of piece 357. Softwood is represented by entangled balls, ripen fiber, and tips of A. chilensis and/or F. cupressoides thin and thick-walled late (compression) wood tracheids. One cf. Bambusoideae plates and one Neltuma starch grain were also recovered. The assemblage suggests softwood and cane working together with hardwood debarking. Although the edge has other indeterminate substance traces that could correspond to the same work depending on its location and distribution, there are aspects of them that suggest that they would be the product of the work of animal soft tissues. Thus, the indeterminacy of these traces would be due to the longer work time required for their diagnostic development against more resistant substances and to the low brightness of this type of polishing (Cueto, 2015).

The integral evaluation of the proximal and middle portion of the artifacts allows us to infer that a subset of four scrapers was hafted (342, 357, 611, and 589), and that another two, with disparate evidence, could have been hafted. These are assigned to two distinguishing degrees of probability, a major (artifact 365) and a minor (artifact 731).

The first subset of artifacts featured hafts made of wood. This is inferred from the traces on both lateral edges and on some of the dorsal and ventral surfaces, or next to the butt, and as a result of the associated plant residues: pieces of softwood and fragments of threads made from thick-walled late (compression) wood tracheids of A. chilensis or F. Cupressoides.

The diagnostic polishes would have been generated by friction or pressure with the substances that made up the hafting device – wood, leather, resin – or that correspond to the detachment of the lithic artifact – mineral particles. The polishing points would be attributed to contact with wood, leather, and others undetermined. It stands out that polishing points are usually associated with scars or receding rim segments. Also, they register other alterations – scars, bright spots, and rounding – that by themselves and because they appear isolated are less diagnostic for hafting.

In addition, the four artifacts have xylem residues. In 357 and 589 A. chilensis, wood was used to make the hafts and in artifacts 342 and 611 A. chilensis or F. cupressoides, while the fruit tissue of Neltuma sp. could come from the distal edge (589). Poaceae phytolith 342 of the middle/proximal part might be a migration resulting from green cane cleaning in the distal part. The hafting devices on 357, 611, and 589 would have had fixations made of tendons. They also have other animal residues such as wool fibers and hair, some assignable to Lama guanicoe.

Another subset (365, 731) combines evidence suggesting the likelihood of the scrapers hafting, although they present less quantity and variety of traces and with a lower degree of diagnosis.

For 365, the possibility is configured from a series of traces on its rims. They are bright spots, some with a defined orientation, isolated scars, generalized alteration of the edge and retraction of the ridge. Here, we associate a fragment of A. chilensis with a possible wooden haft and fibers similar to the threads of tracheids but just folded which are difficult to identify as anthropic modification. Scraper 731 is less likely to have been hafted. Although its rims have traces that could be linked to this device, such as wood polishing, striations, and association with scars, including several associated with bright spots. Some scars can be attributed to taphonomic alterations, the dorsal face ridges do not show significant damage, and the residue corresponds to vessel elements assignable to cane and cf. Adesmia/Berberis, and pieces of wood were absent.

The results of the multiproxy examination allow us to discern some technological aspects. Within the dominant class of instruments, scrapers, those made on blades have design elements at the level of the active edge – distal – that distinguish them from the rest of the artifacts. These tend to be short and, based on the analyzed sample, with high angles abrupt and very oblique. Instead, for these characters, their lateral edges/rims match the general patterns of the assemblage. The length of the active edge is attributed to the restriction imposed by the width of the chosen base shape, but the choice of the distal end, over the lateral edges, would be linked to the search for an artifact design that facilitates its gripping. Likewise, this search would be evidenced in the irregular finish of the lateral edges, discontinuous in some cases, or combined with natural edge segments, and with lower angles than the active edge. These would have been modified in this way to adapt them to the haft and reduce the impact that the natural rims with a very sharp bevel could generate on leather or tendon bindings, or on the leather cloths used to cushion contact with the wood of the haft. We also consider that an attempt was made to adapt the general shape of these scrapers, configuring instruments with a length/width module: long and medium, and analogous in terms of width/thickness module: medium, except for one (589) which is thin. This would be explained by an extraction carried out on the dorsal face of the base form, carried out to thin the support. Similar treatment to that of three other scrapers (342, 365, and 611). All refer to the adequacy of the proximal and middle portion of the base form, for its connection with the haft.

This research suggests that a class of scrapers of a particular design was developed and used, that was hafted or had this possibility. These artifacts required a manufacturing sequence with greater labor investment, a fact that distinguishes it from the regional pattern for the period. However, it has similarities with the patterns followed regarding the kind of support used, the type of raw material, and the secondary modification on one of its faces, by flaking techniques. The identification of this sequence confirms the hypothesis that suggests the hafting of distal scrapers made on the blade, at least for a segment of them. This corresponds to one of the ways in which blade technology has been taken advantage of based on the versatility and flexibility that it offers.

We attribute the proper preservation of plant and animal residues in the artifacts to the environmental conditions in which the site is located, where there is little annual rainfall and a very dry and arid climate. These conditions favor the preservation of even starch grains (Ciampagna et al., 2021). The presence of charcoal and the state of conservation of the bone remains in the layer reinforce the premise that it was well preserved. However, the presence of antifungal substances and resins like those present in A. chilensis and F. cupressoides (Castro, 2009; Olate, Vélez, Greslebin, & Schmeda-Hirschmann, 2015) might have favored preservation and the adherence to the lithic pieces and them to be overrepresented against cane and hardwood.

Regarding the origin of the plants identified, we consider that they could correspond to the local and non-local areas according to the vegetation units inferred for the study period. Thus, the groups would have selected Berberis and Adesmia logs that came from spaces near the site in the framework of their daily activities. These, in turn, processed logs of Neltuma, a shrub that would be absent from the central plateau during the middle Holocene, and that would come from a non-local area attributable to the shrubby steppe of the San Jorge Gulf, near the Atlantic coast. This is located at a minimum distance of 195 km from the site. In addition, they used logs of canes (i.e., Chusquea culeou), A. chilensis and F. cupressoides, characteristics of Andean forests that developed between 39°50′ and 43°30′ south latitude during this period (Miotti & Salemme, 2004). The two latter correspond to a non-local landscape located at a minimum distance of 500 km. The redundant presence of compression softwood could account for the use of flag trees like those of the face of the Andean mountains rather than of those from the proper Valdivia rainy forest (Roig, 1992). We consider that the circulation of non-local species towards the central plateau would have occurred through mechanisms such as exchange and, to a lesser extent, direct supply. Both could have developed in the context of population growth and an increase in temporary/annual mobility circuits among hunter-gatherer societies. Logs or other organs could have been transported from these plants, as in the case of Neltuma and Chusquea, or already configured hafts integrating composite artifacts or ready to haft an artifact, as in the case of Austrocedrus or Fitzroya, whose wood is between light to medium soft. Its choice stands out given that near the site there are species such as Schinus or Berberis, which were used to make hafts, according to ethnographic data. Although the whole wooden hafts were not recovered in La Mesada, neither in other continental nor insular sites of Argentine Patagonia (except for one), this could be due to the fact that they were part of the personal equipment of some members of the groups, constituting a preserved artifact, conceived to last after the replacement of the lithic segment and easy to transport to other places. They could also have been easily discarded in the hearth when no longer useful. Non-favorable environmental conditions, such as in the case of humid forest areas of Tierra del Fuego (see Franch et al., 2022; Piqué I Huerta, 1999, for more detail), might also have affected the preservation of wood from archaeological hafts in insular Patagonia. Instead, the first evidence of archaeological wooden hafts for mid-Holocene Patagonia from plant residues analyses presented here stands out and is consistent with the first record of hafts, made in Berberis, recently reported from the early Holocene occupation of Cerro Casa de Piedra 7 site (Caruso Fermé & Aschero, 2020).

In sum, this approach allowed the discerning about the functionality and hafting of a particular scraper design. Even recognizing the limitation of the sample in terms of the number of scrapers evaluated, the richness and diversity of the results obtained are high and singular. It allowed us to clarify the production sequence of the artifacts and identify that it has distinctive aspects regarding regional technological trends, for the period. Practices linked to plant management by hunter-gatherers were recognized. The use of wood from vegetation units distant from the landscape where the site is located, stands out. This invites us to reflect on the mobility circuits and exchange mechanisms that these societies maintained. Furthermore, the results will be contrasted with the examination of other instruments and experimental program applied in softwood and canes, to recognize patterns of use, manipulation, and design, and to evaluate technological change.