Living in the Mountains. Late Mesolithic/Early Neolithic Settlement in Northwest Portugal: Rock Shelter 1 of Vale de Cerdeira (Vieira do Minho)

4.1 Objectives and Stages of Study

As mentioned in Section 1, one of the objectives of this article is the description of the methodological options and technical procedures adopted in the study of raw materials. Thus, the study of raw materials was developed with three mains objectives in mind:

• The classification and characterisation of different raw materials (of a local and exogenous nature) exploited by the prehistoric communities;

• To recognise raw material management practices;

• To identify, if possible, exploited areas and potential sources of raw material.

For this purpose, a study schedule was established and divided into three distinct phases:

• Analysis of the raw materials that make up the archaeological assemblage;

• Conducting field surveys to comprehend the geological context and recall geological samples;

• Execution of analytical procedures to create associations between archaeological and geological specimens.

The analysis of raw materials was based on those of local origin, consisting of different varieties of quartz. This decision was made bearing in mind that local resources are dominant in both SU-1 and SU-2. For other raw materials, including those of exogenous origin, we have not established a specific study plan given their low significance in quantity and percentage. Their classification and provenance proposal was made by Thierry Aubry, whom we take this opportunity to thank for his valuable contribution.

4.2 Quartz Classification

As stated, quartz is the dominant raw material in the lithic collection recovered in Rock Shelter 1 of Vale de Cerdeira. By consulting the geological mapping that coincides with the location of the shelter (Geological Sheet 6-C, Cabeceiras de Basto, from the 1:50,000 Geological Map of Portugal) (Noronha, 1992) and the neighbouring areas, we can understand that this mineral is an abundant resource on a regional scale, available through both fault-associated veins and pegmatite geological bodies.

In previous studies by Meireles (2009, 2010, 2013), quartz was divided into two major categories: hyaline quartz or rock crystal and other quartz varieties, similar to Mourre’s proposed division (1996, 1997) between automorphic and xenomorphic quartz, referring to the crystal habit and the diverse environmental conditions of formation that enhance the emergence of different varieties. In the analysis that has been conducted recently, the hyaline quartz category was maintained, adopting for the other types a chromatic-based typology, comprising the following varieties: milky, translucent, rose, grey, and smoky.

Being conscious of the limitations of the use of chromatic classifications, rightly recognised and criticised by several authors (e.g. de Lombera-Hermida, 2020; Driscoll, 2010; Rodríguez Rellán, 2010), it was nonetheless decided to use them, given the use of identical designations in mineralogical studies but also for providing a first distinction of an expedient nature, easy to apply and which can also be related to different geological environments. In addition, in some recent works, which have quartz lithic industries as their object of study, the colour corresponds to one of the criteria used in the classification of the different raw materials (e.g. Aubry et al., 2016a; Gaspar et al., 2016).

In addition to the colour-based typology and to overcome its limitations, it was also decided to follow the methodology proposed by Llana Rodríguez for the quartz and quartzite from the Upper Palaeolithic of Galicia and Asturias (Llana Rodríguez, 1990; Martínez Cortizas & Llana Rodríguez, 1996). This macroscopic classification allows us to verify the presence or absence of two variables that seem to have more influence on the quartz’s aptitude for knapping: texture and joints/internal effects/crystallisation planes/fractures.

In this sense, four morphostructural groups were established based on the presence (S) or absence (N) of the two indicated variables – NN, SN; SN and SS – with the first letter referring to the grain and the second to the internal planes: NN (no grain/no plane); NS (no grain/plane); SN (grainy/no plane) and SS (grainy/plane) (Figure 3). By relating quartz formation processes to its mechanical properties, this classification may help to uncover raw material acquisition and selection criteria and to identify some technical constraints associated with each group (de Lombera-Hermida, 2008).

Figure 3

Quartz morphostructural groups.

Concerning the assemblage from “Rock Shelter 1 of Vale de Cerdeira,” even though this analysis is still ongoing, it is relevant to point out that the quartz from category “NN,” which reveals a great aptitude for knapping, is frequently used to produce bladelets or elongated blanks and to manufacture microlithic tools. In another sense, fieldwork made it possible to understand that the quartz existing in the immediately surrounding area of the shelter, which occurs in hydrothermal veins, displays a fine-grained texture fitting the SN and SS categories of the mentioned classification. Nevertheless, most of the quartz types present in the two stratigraphic units under study are of type “NS,” which, as far as our fieldwork has shown, occurs in other sites located at a greater distance from the archaeological site. This fact reveals a selective behaviour on the part of the prehistoric communities of Serra da Cabreira concerning the supply of the raw material.

4.3 Fieldwork

To recognise some possible raw material supply locations, a series of archaeological surveys were carried out. For this task, geological mapping is of particular importance, as well as other resources, namely the webpage SIORMINP – Information system of Portuguese mineral occurrences and resources (http://geoportal.lneg.pt/pt/bds/siorminp/).

The varieties of quartz present in the Vale de Cerdeira collection can be encountered in quartz veins and pegmatite formations. The study of the geological cartography and the initial fieldwork made us understand that it would be impossible, from the logistical and time perspectives, to characterise all the geological bodies existing on the periphery of the shelter. Thus, to make this survey more feasible, it was necessary to direct the search towards rarer types, namely rose, smoky, hyaline, and hyaline or translucent quartz with inclusions. The prospecting programme carried out to date, particularly focused on the detection of pegmatites positioned up to 15 km away from the archaeological site, has recorded more than 200 spots of possible archaeological and geological interest.

Despite occasional constraints, such as accessibility or the destruction and fragmentation of certain geological bodies, some of the detected points are worth highlighting, especially the site known as “Muro Alto,” located about 5 km from the archaeological site. This is a pegmatite body, providing different types of quartz, including rose, smoky, translucent, and with inclusions. Nowadays, the place is very much modified from what would be its natural configuration, because of intense mining during the 40 and 50 s (pers. com. C. Leal Gomes). Also, according to C. Leal Gomes and based on the geological analysis of the available information, before the mining exploitation, the site could have presented an imposing quartz body, over 6 m high, constituting a true landmark in the surrounding landscape (which could be related to the place name “Muro Alto,” which is Portuguese for “High Wall”).

At present, the site has already been studied in the context of our work, and some geological samples have been collected. However, the dense vegetation and the lack of ground clearance have momentarily made it impossible to continue the work safely. We intend to collect some more samples and perform analyses and laboratory procedures that may permit comparison with the archaeological artefacts and associate or discard this geological body as a source of raw material for the Cabreira mountain prehistoric hunter–gatherers.

4.4 Analytical Procedures

The scarcity of provenance studies on quartz can be explained, to some extent, by the fact that quartz is generally considered an immediate and local resource and that this type of study has focused mainly on exogenous raw materials because of the information they can provide on exchange networks and circulation routes of goods and people (de Lombera-Hermida, 2020). Indeed, on this topic, Bracco (1997, p. 287) states that in regions where quartz is not present on a regional scale, it is not exploited, a fact that reveals its low circulation and its exclusion from exchange networks.

Just as relevant is the nature and chemical composition of this mineral, essentially based on silica, making it quite challenging to discriminate between different sources, even of close origin (Cousseran, 1999; Rodríguez Rellán, 2010). In comparison to other raw materials, namely flint, quartz reveals a low internal variability, which translates into an immense difficulty in determining provenance from macroscopic analyses (Pereira et al., 2015).

In recent years, different approaches have been developed to overcome these constraints, including those from the area of geochemistry. For example one can mention isotope and trace element analyses, whose principle is based on the identification and tracing of elements present in the quartz as silica substitutes or retained in the interstitial impurities (Meighan, Simpson, Hartwell, Fallick, & Kennan, 2003), and secondary ion mass spectrometry, used in the characterisation and sourcing of pegmatite quartz artefacts from the Boreal period of central Canada, suggesting the existence of displacements on the order of 200 km (ten Bruggencate, Fayek, Brownlee, Milne, & Hamilton, 2013; ten Bruggencate, Fayek, Milne, & Brownlee, 2014).

To characterise the different quartz from the collection, as well as discuss working hypotheses about the location of the supplying sites, some laboratory techniques used in Earth Sciences have been applied.

Full preference was given to non-destructive techniques from an early stage of the study, to ensure the inviolability of the artefacts (Garrison, 2006). This decision was based on the heritage and cultural value that we recognise in the archaeological artefacts, but also due to the expectation that, in the future, other procedures and research projects may be developed on the same materials, possibly involving the use of more accurate techniques.

The non-destructive concept has a double meaning: first, in the obvious fact of avoiding any alteration or damage to the piece submitted to the analytical procedure and, second, in the sense of being reproducible, allowing the repetition of tests or the performance of different analyses on the same artefact (Xavier & Alves, 2019). Some techniques, although destructive, requiring the cutting or fragmentation of the artefact, enable, on the contrary, the performance of several sets of tests, avoiding continuous damage to the archaeological material.

The set of laboratory techniques tested included (Figure 4): scanning electron microscope examination to obtain chemical characterisations and information concerning textural and structural analysis; RAMAN analyses, useful for the understanding of the mineral chemical composition; X-ray fluorescence studies, equally important in the chemical characterisation, both global and pointwise and, hyperspectral imaging analyses, which, through the quantitative evaluation of chromatic coordinates, allow, in an objective manner, to determine the colour of a certain piece.

Figure 4

Analytical procedures.

Some of the difficulties experienced in these tests were related to the irregular surface of the artefacts, restricting the applicability of some procedures and the scarce penetration capacity of some of the techniques and equipment, an aspect that is particularly important in the analysis and detection of inclusions, both fluid and mineral, described in Section 4.5.

4.5 Study of Inclusions

As mentioned, the great chemical and crystallographic homogeneity between different varieties of quartz make it exceedingly difficult to detect the sources of supply (Cousseran, 1999; Cousseran, Pècher, & Bintz 2000). A possible approach to the provenance sites of these materials lies in the identification of inclusions in automorphic quartz artefacts, which can provide insights regarding the formation environment and, consequently, its origin (ibidem). Inclusions correspond to materials retained inside minerals and can be classified according to their nature as solid (when dealing with other minerals) or fluid, if in a gaseous or liquid state (Figure 5); they are also categorised according to their formation, being protogenetic when formed before the crystal, syngenetic, when facing a simultaneous development, and epigenetic, constituted only after the integral development of the crystal (Sachanbiński, Girulski, Bobak, & Łydzba-Kopczynska, 2008, p. 1013). This methodology requiring the use of analytical techniques, including microthermometry and RAMAN spectrometry, has been successfully applied, notably in S. Cousseran’s studies for the Mesolithic and Neolithic of the Western Alps and Lombardy (Cousseran, 1999, 2000a, 2000b; Cousseran et al., 2000) and also those of Sachanbiński for Prehistory in the SW of Poland (Sachanbiński et al., 2008).

Figure 5

Study of inclusions.

In the scope of our study, observation under the petrographic microscope of artefacts in hyaline and translucent quartz allowed the recognition of inclusions. A substantial part of the effort invested in the laboratory techniques mentioned in the previous point was devoted to the determination of the constituent material of the inclusions. As regards the mineral inclusions, the studies carried out suggest that they may correspond to schorlite and elbaite; as for the fluid ones, the results have been unclear because of the already mentioned limited capacity of penetration into the pieces by the laboratory equipment and to the circumstance that the analyses were done on the complete artefacts. Nevertheless, the identification of the mineral inclusions has provided possible guidelines for fieldwork.

5.1 Stratigraphic Unit-1

Amongst the raw materials, the vast majority, as will be demonstrated below, correspond to different varieties of quartz. However, and although representing a minor role, there are other lithologies, namely flint from the West Portuguese Meso-Cenozoic Border, the vast majority of which is attributed to the Bajocian, with a very small contribution of flint from the Tajacian (SU-1) and Cretaceous (SU-2). This classification, as already mentioned, was carried out by Thierry Aubry, who also made another valuable contribution by identifying a small group of micro-quartz silicifications that were provisionally classified as flint. According to the procurement studies carried out in the Côa valley, NE Portugal (e.g. Aubry, Chauvière, Mangado Llach, & Sampaio, 2003; Aubry et al., 2016a,b), these silicifications, of epi and hydrothermal formation, appear associated with milky quartz and chalcedony veins and are considered to be of local and/or regional origin, having been detected in primary position, at a distance between 20 and 50 km from the archaeological sites. C. Leal Gomes had also pointed out the probable existence of these silicifications in the geological context where the “Rock Shelter 1 of Vale de Cerdeira” is located (C. Leal Gomes, pers. comm., August 2017).

This small group of artefacts shows some variability, revealing different origins; however, within the scope of this work, they will be treated as a whole, without any individualization. Under the designation of “Others,” there is also a group of very diverse geological resources, such as schists, siltstones, or other undetermined resources, whose grouping is only justified because they constitute, in their totality, a rather reduced quantity of about two dozen pieces.

Thus, regarding the raw materials, in the general context of the collection (Table 1), one can observe the preponderance of xenomorphic quartz, namely milky, followed by translucent; the hyaline quartz represents 13.3% and the flint not even 1%. In what concerns the morphostructural groups, the NS variety, fundamentally associated with milky and translucent quartz, is the most relevant, whereas the NN, generally related to the rock crystal, stands at 22.8%. The quartz with granular texture is not very significant in general terms, despite being available less than 5 m from the archaeological site. These proportions show, however, some variations between different structural categories.

Amongst the bladelets, there is a greater preference for the use of rock crystal (33%); at the same time, NN-type quartz also makes up 29% in these blanks, whereas flakes do not make up more than 16%. In this artefact category, the most frequent type of quartz is milky of the “NS” group, with a total of 40%. It is in the microlithic instruments where a greater use of flint and microquartz silicifications is observed (the management of these two resources is always carried out similarly) because 10% of these pieces were made in these resources. It is equally amongst these instruments that we can see great use of the hyaline quartz (>53%) and the quartz from the morphostructural group NN, in the order of 80%. In the opposite direction, in what corresponds to the common tools, the distribution of the raw material is identical to the non-retouched blanks, prevailing the quartz with milky and translucent tonality from the “NS” group. As a matter of fact, in this particular case, only different varieties of quartz were used in the fabrication of the tools.

The dominant structural category in SU-1 is that of knapping debris (chunks, chips, and indeterminate fragments), representing more than 75% of the collection. The second place is taken by debitage products (flakes and bladelets) close to 19%, whereas cores represent no more than 2.1%. Equally low is the percentage of tools, just above 2% (Table 2).

The high percentage of knapping debris and indeterminable fragments is partly explained by the fracture mechanics of quartz highly subject to the occurrence of fractures and is in line with what has been described by several authors (e.g. Bracco, 1997; Driscoll, 2011; Mourre, 1996, 1997; Tallavaara, Manninen, Hertell, & Rankama, 2010).

For a brief description of the blanks, it can be first noted that the great majority of the butts are plain (more than 50%), in both subcategories (flakes and bladelets). Contrary to what other studies and experiments on quartz have revealed (e.g. Driscoll, 2011), more than half of the debitage products show well-pronounced percussion bulbs (61% in flakes and 57% in bladelets). Regarding the morphology of the blanks, if in the flakes it is not possible to detect a preference for a certain shape, in the case of the bladelets, there seems to be a clear preference for parallel edges (36%). The occurrence of fractures is quite significant reaching 55%. Amongst the most common fractures are Siret (accidental break), more than 30%, and transversal ones, around 25%. However, if in the case of flakes the difference between Siret and transversals is not very pronounced, amongst bladelets, the transversal fracture represents half of the total and the Siret type no more than 17%.

The most common reduction strategies observed in cores are polyhedral (89 pieces – 29.7%), bipolar-on-anvil (77 pieces – 25.7%), and prismatic (40 pieces – 13.3%). The bipolar-on-anvil method, commonly mentioned regarding lithic industries in quartz, was applied to smaller volumes of raw material; as for the prismatic method, it was usually practised on quartz prisms, taking advantage of the morphostructural characteristics of the crystalline rhombohedron for the production of bladelets and elongated blanks (Ramil Rego & Ramil Soneira, 1997).

Despite the raw material abundance in the local geological context, this does not prevent that more than 40% of the cores are intensively explored and 16% are even exhausted. We should note the collection of a set of unexplored raw material volumes, amongst which 49 quartz prisms can be understood as the constitution of a raw material reserve by the hunter–gatherers of Serra da Cabreira.

The typological features of the lithic industry show a robust microlithic character because these tools represent 72% of the total, as opposed to the so-called “common tools” that use flakes as blanks and which hold a share of 27%; the remaining 1% integrates the miscellaneous category.

Amongst the “common tools” (Figure 6), a total of 93 units, the four most represented groups, in descending order, are borers, notches, retouched flakes, and burins. Other types are also present, although in significantly smaller amounts, such as end-scrapers, denticulates, scapers, and truncatures.

Figure 6

Common tools SU-1 (percentage and absolute amounts).

As for the microlithics (242 pieces) (Figure 7), points are the predominant type, followed by segments and trapezes. The list is completed by retouched bladelets, followed by truncatures, micro-borers, and indeterminates.

Figure 7

Microlithic tools SU-1 (percentage and absolute amounts).

Regarding microlithic points, triangular cross-section and morphology predominate, although a division must be established amongst those with a simple (non-retouched) (31 units) and retouched base (20 units). Equally framed in the triangular points, however, featuring a particularly original solution, is another subtype, made from a removal usually taken from the apex of the quartz prism. The dorsal surface of these blanks is thus devoid of previous extractions, rather displaying the original crystallographic face of the quartz prism. From then on, taking advantage of the usually triangular and pointed morphology of the blanks and through a discontinuous but very precise retouch – sometimes direct and marginal practised on one of the edges and, on other occasions, inverse and invasive, thinning the base – a pointed armature is shaped (Figure 8, no. 12 and 14). To this particular type of point and because, in our understanding, its conception results from the fact that quartz is the main raw material, from now on we would like to name it “Vale de Cerdeira microlithic point.” This point variety reflects a great understanding of the characteristics of the raw material, i.e. the physical and mechanical behaviour of quartz, perhaps as a consequence of a long-term and solid territorial implantation. In SU-1, there are 19 of these points.

Figure 8

Tools from stratigraphic unit 1.

In a minority of cases, but equally present, are fusiform points (seven pieces), pedunculated points (five units), and points with bilateral convergent retouching (two pieces). Of particular importance to understanding the production method of these armatures was the identification of five pre-forms, i.e. points not fully configured, but in the middle of the production process.

The segments are mostly symmetrical (38 pieces); the asymmetrical and elongated segments present identical values, with a slight primacy for the former, 11 and 7 accordingly. Four pre-forms of this typology were also identified. In turn, amongst trapezes, there is a small advantage of the asymmetric ones (17) concerning the symmetric (13) and, also, in this case, pre-forms were reported (4). No triangle was identified amongst geometric tools.

There is a remarkable homogeneity between the geometrics manufactured in quartz and other raw materials, as far as their sub-morphologies, dimensional averages, and retouching mode are concerned. The retouching mode, except for minor variations, is identical in both “common tools” and microliths: direct, continuous, abrupt, marginal, and parallel or sub-parallel. In microliths, it is also often done on an anvil. Also, in this domain, it should be noted that the double bevel retouching technique was observed only at one point of the SU-1, not extending into the following stratigraphic unit.

In summary, about the lithic production system, there is complete agreement with what was previously defined by Meireles (2010, 2013). Thus, the strategy is geared towards the obtainment of two debitage products. With a greater preponderance, flakes of a small module are produced through operating chaînes opératoires from polyhedral, bipolar, and parallelepiped cores, consisting of so-called random methods of debitage. In smaller quantities, there is also the manufacture of bladelets and micro-bladelets from prismatic cores, essentially unipolar, standing out a specific chaîne opératoire based on the exploitation of quartz prisms, which takes advantage of the structural characteristics of the crystalline rhombohedron, using the natural facets of the crystals as guiding-ridges to define the debitage surfaces. A significant part of the microlithic instruments in the collection are made from these blanks.

5.2 Stratigraphic Unit-2

In the domain of raw materials, broadly speaking and compared with SU-1, there is an increase of milky quartz, which concentrates more than half of the entirety, as well as of flint and micro-quartz silicifications. On the contrary, the rock crystal shows a global decrease, although with a more directed application in the production of bladelets and configuration of microlithic tools (Table 3).

There is a slight increase in micro-quartz silicifications as well as flint. In SU-1, we found Bajocian and Tajocian flint (two artefacts) from the West Portuguese Meso-Cenozoic Border, more specifically from Cantanhede and Anadia localities, respectively. In SU-2, Tajocian flint disappears, being replaced by Cretaceous flint (about six artefacts), probably from the Rio Maior region. Amongst the “Others,” besides schists, silcretes, and other unknown lithologies, another resource not present in the previous unit was identified, more specifically hornfels (a bladelet and a chip).

The quartz with granulated texture (morphostructural groups SN and SS) are only a small part of the resources manipulated, which may be related, more than to its lower or higher availability in the local geological environment, to the reduced aptitude for knapping that they seem to present. Within the cores explored under the prismatic method (23 units), it is worth mentioning a higher concentration of quartz from the “NN” morphostructural group (15 pieces), characterised by a higher textural and structural homogeneity, which are important requirements in the application of this reduction strategy.

The difference between the total number of artefacts between SU-1 and SU-2 is just over 3,000 pieces, the latter counting exactly 11,474 artefacts (Table 4). There are considerable similarities, in various aspects, between both stratigraphic units, starting with the percentage value of some structural categories. Thus, in SU-2, knapping debris is more than 80% and debitage products around 16%. However, there was a reduction in the percentage of cores as well as tools that, this time do not exceed 1.7%. Amongst these, the microlithics remain the great majority.

The number of debitage products is lower than in SU-1, but their characteristics are quite similar. For instance, the percentage of fractured flakes and bladelets settles again at values close to 50%. There is also another similarity in the predominant type of butt (plain) – 57% in flakes and 44% in bladelets – as well as in the well-developed bulbs – 61% in flakes and 62% in bladelets. Parallel-edged bladelets with prominent arris on the dorsal side correspond to the type of support most wanted by prehistoric knappers (35.8%).

Although the main reduction strategies are maintained from one stratigraphic unit to the other, there is a small difference in their order. This time, the main strategy is based on bipolar percussion on an anvil (29.7%), followed by polyhedral (21%) and prismatic (19%) percussion, once again, mainly from unipolar cores. The percentages of intensively exploited and exhausted cores remain high, 35.6 and 24.6%, respectively. In fact, amongst the possible causes for abandonment of the cores – and even though in more than 37% of the situations it was not possible to point out an objective reason – exhaustion had a more predominant role than, for example, raw material defects (13.8%) and the occurrence of fractures (15%). It is precisely amongst the cores exploited by the prismatic method where greater exploitation of the volumes is observed.

If the blocks or fragments are the types of support on which the polyhedral method is developed, as far as the bipolar and prismatic methods are concerned, these are practised fundamentally on quartz prisms. Similarly to what was mentioned for SU-1, intact quartz crystals were also recovered in SU-2, even in a larger quantity, counting 59 units.

In what concerns the tool structural category, the microlithic component remains more substantial, now representing 71%; the “common tools” represent 28% and the remaining 1% integrates the miscellaneous category.

In the case of common tools, with a total of 48 units (Figure 9), three morphologies can be distinguished, more specifically: retouched flakes, borers, and notches. Other types and morphologies go through a great decrease in their absolute values in this stratigraphic unit, such as burins, denticulates and scrapers. Once again, there is no evidence of particular management of lithological resources in the production of these tools, in contrast to what can be seen in the microlithic subcategory. Nevertheless, and contrary to what was described for SU-1, there are three tools not made in quartz, namely a quartzite “bec” and, in flint, a small thumb-nail end-scraper and a retouched flake.

Figure 9

Common tools SU-2 (percentage and absolute amounts).

As for the microlithic tools, this time segments are the predominant type (62 pieces), followed by points (40 pieces), even if, compared with SU-1, they register a remarkable decrease. The third most represented category is, for the time being, retouched bladelets (28), which, in percentage terms, are more than double that of SU-1. There is a notable decrease in trapezoidal armatures (only two units), also highlighting the presence of a triangle, a geometric sub-type not identified in the previous stratigraphic unit (Figure 10).

Figure 10

Microlithic tools SU-2 (percentage and absolute amounts).

In layer 2, although not very significant (<15%, nine pieces), a new variant of a segment is observed. Such pieces are on the borderline between segment and microlithic point, as one of the ends is slightly angled or transverse, as if it were a small base. As in SU-1, symmetrical segments continue to predominate, representing more than 40% (26 units).

One issue that was not addressed in the previous section and that cannot go unmentioned concerns the mode of production of geometrics. In the lithic collection of Vale de Cerdeira, there is no evidence of the use of the microburin blow technique. Thus, the realisation of microlithic armatures, namely geometrical and points, followed two technical procedures. The first one comprehends the shaping, through retouching, of blanks of reduced modules, such as bladelets, flakes, or even, occasionally, chips; the second corresponds to the transversal rupture technique consisting in obtaining a fracture surface perpendicular to the blank axis through bending or, more rarely, percussion, from which a fragment is selected to manufacture the piece. This technique, known for its expeditious nature, was applied, amongst other possible examples, in the cave of El Espertín, in León, Spain (Fuertes Prieto, 2000–2001). Its recognition, in the Vale de Cerdeira case, was possible because of the identification of a set of bladelets proximal/mesial fragments, systematically fractured by bending, showing that this was not the result of some accidental circumstance but, on the contrary, of intentional behaviour.

The non-application of the microburin blow technique in Vale de Cerdeira can be, in our view, seen in two ways. If we assume that this circumstance is due to the raw material, and accept the often uncritical evaluation of quartz as a raw material of poor suitability for knapping, we could subscribe to the reflection of Araújo (1995–1997, p. 137) for whom this technique presents a reduced efficiency in low-quality lithic resources. On the other hand, there is the perspective of Eren, Lycett, Roos, and Sampson (2011) who state that the quality of the raw material should not necessarily be understood as a determining or constraining factor, but as an opportunity for prehistoric knappers to adapt, often with good results, to the characteristics of the raw material.

As far as the Vale de Cerdeira collection is concerned, not all quartz can be classified as having a low aptitude for knapping. The great majority of the hyaline quartz and some milky and translucent quartz, particularly those that fit into the morphostructural group NN, can effectively be considered good resources. However, other characteristics of this raw material (anisotropy? fracture mechanics?), other than its quality, can make the microburin blow technique not very suitable. Finally, we must also bear in mind the technical and “cultural traditions” of the human group, which might not have envisaged this mode of manufacturing microlithic tools. The truth is that, although in residual quantities, other lithological resources such as flint and micro-quartz silicification are present, both particularly directed towards the production of bladelets and microlithic instruments, without representing the adoption of significantly different technical procedures.

In the category of microlithic points, there is a decrease in the number of pieces, now counting 40 items. Continuity of the main types can be observed, as well as the primacy of points with triangular cross-section and morphology in the three aforementioned variants: simple base (7), retouched base (10), and “Vale de Cerdeira” (7) (Figure 11, no. 7 and 8). Points with fusiform morphology and points with peduncle (6) or bevelled ends (2) are also registered, in a similar proportion to the previous stratigraphic layer.

Figure 11

Tools from stratigraphic unit 2 (Photographs by Adriano Borges).

Nevertheless, some differences or innovations can also be observed. A particularly significant aspect is the appearance, in the various types and subtypes of these tools, of curved backs, a technical feature that does not occur in layer 1, where oblique or rectilinear backs prevail. Another differentiating element is the appearance of two transverse-edged points; in turn, the bilateral convergent-edged points, whose presence was already not very representative in the SU-1 context, are now reduced to a single piece.

It seems particularly relevant that only different varieties of macrocrystalline quartz have been used in the making of microlithic points (in both stratigraphic units). Flint, as well as micro-quartz silicifications or the few silcretes, is used in other classes of artefacts, such as bladelets, retouched or not, geometrics, basically segments, and even micro-borers. We do not have any evidence to explain this behaviour, which is unlikely to be the work of chance. Even as speculation, we can put forward some working hypotheses to be discussed in the future. Could the points that take part in composite tools different from those in which the geometric or retouched elongated blanks would be used? Or would the difference lie in the type of hunting prey, given that these types of tools, despite being suitable for other purposes (Clarke, 1978; Pignat & Plisson, 2000), are essentially used in hunting activities? Perhaps, we should not exclude cultural or identity criteria that may be underlying this decision.

On the other hand, the quartz used in the manufacture of points covers a great variety, both in what concerns the chromatism and in what relates to its attachment to a particular morphostructural group, being also worth mentioning the use of qualities with a notoriously lower aptitude knapping.

Importantly, there are no differences between the production system of SU-2 and SU-1, whereas the same strategy orientations of the previous set tool are followed. So, the most distinctive aspect of the second stratigraphic level is the presence of a small set of polished stone materials, constituted by a small sillimanite/fibrolite axe, one fragment (butt) of another axe also in sillimanite/fibrolite and four other miscellaneous ones. Moreover, 20 sherds of ceramics were recovered, all of manual technique, most of them without any decoration. Amongst the four decorated fragments, two have a simple incised decoration and the other two impressions and dragged impressions, also known as “boquique type” (Figure 12). As a whole, these ceramics encountered parallels with Early Neolithic productions found in the archaeological site of Prazo, located in Alto Douro, NE Portugal, and, in the absence of radiometric dating, might provide a chronological basis for the occupation materialised in this stratigraphic unit (Monteiro-Rodrigues, 2011).

Figure 12

Decorated ceramic sherds and polished axe from SU-2 (adapted from Meireles, 2010, p. 93).