Vale Boi (Algarve, Portugal) and the Solutrean in southwestern iberia

Uned. Espacio, Tiempo y Forma. Serie I, Nueva época. Prehistoria y Arqueología, t. 5, 2012. 455 Espacio, Tiempo y Forma Serie I, Nueva época Prehistoria y Arqueología, t. 5, 2012 pp. 455-467 ISSN: 1131-7698 eISSN: 1111-1354 doi: 10.5944/etf i.5.5376 (1) Núcleo de Arqueologia e Paleoecologia. FCHS. Universidade do Algarve. Campus Gambelas. P-8005 Faro (Portugal). [email protected] (2) Facultad de Geografía e Historia. Universidad de Sevilla. Sevilla (España). (3) Institución Milà i Fontanals. CSIC. Barcelona (España). (4) School of Social Science. The University of Queensland. Brisbane St Lucia, QLD 4072 (Australia) VALE BOI (ALGARVE, PORTUGAL) AND THE SOLUTREAN IN SOUTHWESTERN IBERIA Vale Boi (Algarve, Portugal) y el Solutrense en el suroeste de la Península Ibérica João Cascalheira 1, Nuno Bicho 1, João Marreiros 1, Telmo Pereira 1, Marina Évora 1, Miguel Cortés 2, Juan Gibaja 3, Tiina Manne 4, Frederico Regala 1, Célia Gonçalves 1 y Patrícia Monteiro 1 Recibido el 11 de marzo de 2013. Aceptado el 17 de junio de 2013 Abstract. Located at the crossroads of two rather different ecological and cultural worlds (Mediterranean Spain and Portuguese Atlantic), the site of Vale Boi (Algarve, Portugal) is a crucial element in understanding the economic and social traits of the communities that inhabited Southwestern Iberia during the Last Glacial Maximum (LGM). Combining an open-air with a rockshelter component, Vale Boi presents a lengthy Solutrean record starting with a Proto-Solutrean phase followed by a set of occupations in the 25 to 20.3 ka cal BP time-span. The very rich and well preserved assemblages proved that the site was treated, throughout, as a seasonal residential camp and although a striking combination of exogenous cultural traits has been identified, regional adaptive idiosyncrasies are quite evident. This paper focuses on the results of the lithics, fauna, beads and portable art analysis from Vale Boi, and their impact on the comprehension of the LGM ecodynamics in Southwestern Iberia. Keywords: Vale Boi, Solutrean, Portugal, Last Glacial Maximum. Resumen. Localizado en el marco de dos contextos diferentes desde el punto de vista ecológico y cultural (el Mediterráneo español y el Atlántico portugués), el yacimiento de Vale Boi (Algarve, Portugal) es un lugar fundamental para comprender la organización económica y social de las comunidades que habitaron el sudoeste de la Península Ibérica durante el Último Máximo Glacial (LGM). Situado en una zona en la que se combinan ocupaciones al aire libre y en abrigo, Vale Boi presenta un amplio registro solutrense que comienza con el Proto-Solutrense y a la que le siguen un amplio número de ocupaciones entre el 25 y el 20,3 ka BP. El importante y bien preservado conjunto demuestra que este asentamiento funcionó como un campamento residencia estacional. Aunque han sido identificados diversos caracteres culturales de origen exógeno, son también evidentes los elementos adaptativos idiosincráticos. El presente artículo se centra en los resultados de los análisis del utillaje lítico, la fauna, las cuentas ornamentals y los objetos de arte mueble de Vale Boi y su impacto en la comprensión ecodinámica del LGM en el sudoeste de la Península Ibérica. Palabras clave: Vale Boi, Solutrense, Portugal, Último Máximo Glacial. João Cascalheira, Nuno Bicho, João Marreiros, Telmo Pereira, Marina Évora, Miguel Cortés, Juan Gibaja, Tiina Manne, Frederico Regala, Célia Gonçalves y Patrícia Monteiro 456 Uned. Espacio, Tiempo y Forma. Serie I, Nueva época. Prehistoria y Arqueología, t. 5, 2012. 1. INTRODUCTION Geographically limited to the westernmost regions of Europe and with a relatively short chronological range when compared to other pan-european technocomplexes, the Solutrean has been traditionally regarded as one of the first examples of human refugia during the Late Pleistocene (Jochim 1987; Straus 1991; Straus and Clark 1986). Clusters of sites in particular regions of SW Europe seem to reveal that, similarly to what has been recently argued for animal and vegetal species (Carrión et al. 2008; Gomez and Lunt 2007), Solutrean settlement patterns form a scenario of isolated refugia, that may have contracted and expanded their cultural influence as climate fluctuated (Banks et al. 2009; Cascalheira and Bicho in press; Zilhão 2013). Solutrean settlement in Iberia is, indeed, marked by very clear concentrations of sites, most of them located within ca. 50 km of the present coast line, in areas which were, presumably, repeatedly preferred for settlement due to the existence of diverse and reliable resources (Straus, et al. 2000). Although differences have been long-recognized between these regions, particularly in the later phases of the technocomplex in Iberia, through the preferential distribution of specific types of lithic weaponry (Jordá Cerdá 1955; Smith 1966; Straus 1977, 1990; Tiffagom 2006; Zilhão 1997), divergences have rarely been evaluated based on the detailed adaptive patterns of human populations to the characteristics of each region. Among these clusters there are frequently geographical hiatus where the complete absence or the presence of very few sites do testify human passage but not necessarily intensive occupation (Aura et al. 2009b). During the last two decades those blank areas have progressively being punctuated with new sites, whose evidence have contributed to a better understanding of Solutrean settlement and cultural dynamics, clearly emphasizing the importance of peripheral regions. The southwesternmost region of Iberia (Algarve, southern Portugal) was, until very recently, one of these geographical gaps, for which no evidence of Solutrean occupation were known. Actually, by the mid 1990s available data on the Upper Paleolithic occupation of Algarve were very scarce. Most of the information came from the survey projects carried out in the remote 1940s (Breuil et al. 1943; Viana 1947), the elaboration of broad regional archaeological maps (Marques, 1992, 1995; Gomes and Silva 1987; Gomes et al. 1995), or from mitigation projects in the context of road and public infrastructure constructions (Quelhas ▲ Figure 1. Location of Vale Boi in southwestern Iberia. VALE BOI (ALGARVE, PORTUGAL) AND THE SOLUTREAN IN SOUTHWESTERN IBERIA Uned. Espacio, Tiempo y Forma. Serie I, Nueva época. Prehistoria y Arqueología, t. 5, 2012. 457 and Zambujo 1998). As a consequence, the region remained for several years an archaeological void between the wellknown Upper Paleolithic sites of Southern Spain and those of Central Portugal. Only recently (1996-2001) a large, extended, survey of the Algarve for Paleolithic and Mesolithic sites was carried out as a project funded by the Portuguese National Science Foundation (FCT) (Bicho 2003). This resulted in the identification of some sixty sites with chronologies spanning from the Mousterian to the Calcolithic (Bicho 2004). Among those, six were associated to the Upper Paleolithic but only one —Vale Boi— presented a long stratigraphic sequence, with large amounts of well preserved materials, including thousands of stone tools, marine shells, terrestrial fauna, portable art, body adornments and habitat features (Bicho et al. 2007, 2012a). Located close to the southwesternmost point of Iberia (37° 05’ 4.02” N; 08° 48’ 5.41” W GDE; Fig. 1) Vale Boi has been since 2002, target of several projects that have revealed the multi-component nature of the site, as well as a fairly complete Upper Paleolithic sequence, with important levels of Magdalenian, Solutrean, Proto-Solutrean and Gravettian. Regarding the Solutrean, Vale Boi still is, with exception of the site of Vala, whose chronological attribution to a Solutreogravettian phase is poorly based in the presence of two small backed shouldered points and none absolute dates (Zambujo and Pires 1998), the only known site with securely dated and fairly well preserved Last Glacial Maximum (LGM) deposits in Algarve. Its location at the crossroads of two rather different ecological and cultural worlds (Mediterranean Spain and Portuguese Atlantic) makes it a crucial element in understanding the economic and social traits of the communities that inhabited Southwestern Iberia from ca. 25 to 20.3 ka cal BP. In this paper, we present and discuss the available chronometric, technological, subsistence and symbolic data on the Solutrean occupation of Vale Boi and its impact on the traditional perspectives of LGM hunter-gatherer ecodynamics in Southern and Western Iberia. 2. VALE BOI: AN OVERVIEW OF THE SITE, EXCAVATION AND STRATIGRAPHIC SEQUENCE Vale Boi is located in the eastern side of a wide valley, 2.5 km inland from the present-day coastline, in limestone ▲ Figure 2. Schematic stratigraphy of the different areas of Vale Boi. João Cascalheira, Nuno Bicho, João Marreiros, Telmo Pereira, Marina Évora, Miguel Cortés, Juan Gibaja, Tiina Manne, Frederico Regala, Célia Gonçalves y Patrícia Monteiro 458 Uned. Espacio, Tiempo y Forma. Serie I, Nueva época. Prehistoria y Arqueología, t. 5, 2012. country (Bicho et al. 2007). The site occupies a wide portion of a relatively steep slope, marked by the presence of a 10-15 m high limestone cliff on the top. The slope presents a sequence of staggered platforms formed by the geological phenomena that affected the area over time. The limits of human occupation at Vale Boi appear well marked in the eastern and western extremes of the slope: in the first case by the limestone cliff at the top, in the second, by the alluvial plain of the Vale Boi stream. The northern and southern boundaries are much less perceptible in the current landscape, but the series of test pits carried out to understand the spatial dispersion of the occupations showed a scattering of artifacts on an area that exceeds 10 000 m2. The initial test pits started in 2000 with the excavation of units G25 and Z27 located on the mid-slope section, where large concentrations of surface materials were identified, and where the topography appeared more favorable to the preservation of in situ archaeological levels (Cascalheira et al. 2008). Those test pits were progressively extended, most notably the G25, where the constant presence of materials throughout a stratigraphic sequence of 2.5 m justified the excavation of an adjacent area of 8 m2. In 2003, another set of test pits (J20, S1 and S2) were opened, precisely in the west and east ends of the archaeological site. These pits revealed very interesting results, and once again, required, during the following years (to the present), the excavation of contiguous areas of 22 m2 in the case of the J20, and of 21 m2 in the S2. Since then, archaeological work at Vale Boi have mainly focused in three distinct loci —the Rockshelter (near the cliff face), the Terrace (in the lower section above the alluvial plain) and the Slope (in between)— all with unique characteristics regarding stratigraphy, chronology and, most certainly, space functionality (Cascalheira et al. 2008). The general stratigraphic correlations of the exposed profiles of Vale Boi are shown in Figure 2. The Slope locus covers over two thirds of the site, presenting, within the various excavated sectors, heterogeneous patterns regarding site formation processes. Some areas are, in fact, remarkably well preserved, while in others (e.g. top layers of Z27 sector) traces of washing, size sorting, and loss of faunal material were identified (Manne et al. 2012). Stratigraphically, the Slope is marked by a sequence of very fine silts, containing clasts of limestone whose size and frequency vary significantly throughout the layers. The top of the sequence comprises two highly mixed litostratigraphic units (Fig. 2) that are followed by 3 layers encompassing occupations from the Magdalenian, Solutrean, Proto-Solutrean and Gravettian. Noteworthy is the presence of a geological discontinuity between layers 2a and 2b that, similar to what happens in the Terrace area (see below), marks the onset of the LGM and of the Solutrean occupation in Vale Boi. No habitation features were identified in any of the Slope sectors, and based on the type and preservation conditions of artifacts and faunal remains, Bicho et al. (2010) argued that this would have been an area used mostly for garbage dumping. The Terrace presents the longest and most complete archaeological sequence at Vale Boi. While the top litostratigraphic unit is strongly affected by agricultural work, Layer 2 contained an early Neolithic occupation with well preserved habitat features, including a hut pavement, postholes and hearths (Carvalho 2008). The sequence continues with a partially mixed layer (Layer 3) with the majority of the Solutrean materials deposited, most certainly, by colluvial action, separated from the remaining sequence by a thin layer of small limestone blocks. This discontinuity tops a Proto-Solutrean occupation, as well as a series of well preserved Gravettian horizons, chronologically bracketed between ca. 32 ka and 29 ka cal BP (Bicho et al., submitted). Although the Gravettian sequence from the Terrace is one of the richest at all levels, the most striking feature is the presence, in Layer 5, of a very specific type of weaponry technology - the double pointed and double backed bladelets – whose specificity and chronological exclusivity lead Marreiros et al. (in press) to the identification of a regional early Gravettian facies: the Vicentine Gravettian. A single fire structure was identified in Layer 4 indicating that, together with the wide variety of lithics, fauna, body adornments and portable art, the Terrace was, most probably, an area with residential function (Bicho et al. 2010). The same type of space functionality has also been addressed to the third and final locus: the Rockshelter. Stratigraphically, 4 major litostratigraphic units with archaeological remains were identified, containing two partially eroded horizons of Magdalenian (Layer Z) and Gravettian (Layer D) age, and a remarkably well preserved series of Solutrean occupations (Layers A, B and C) whose preliminary spatial analysis revealed at least three separable archaeological horizons. At the bottom of the sequence, Layer E is sterile in archaeological materials and composed of very fine sediment coming from bedrock disintegration. This area corresponds to a collapsed rockshelter. The collapse must have taken place right after the LGM as it is attested by the large collapsed-roof limestone boulders that overlay the Solutrean sequence. Throughout the stratigraphy faunal remains are exceptionally well preserved. Solutrean layers are the richest, providing a huge diversity of lithic materials, bone industry and portable art (see below). As for the Gravettian occupation, the most relevant fact is the presence of a small set of bladelet projectiles, identical to the ones recovered in Layer 5 of the Terrace locus. The radiocarbon dates obtained from each of these two layers clearly attest that those projectiles integrate the earliest dated Gravettian occupations in Iberia (Marreiros et al. in press). Very recently, a new area of 8 m2 was opened some 10 meters away from the original Rockshelter excavation, whose VALE BOI (ALGARVE, PORTUGAL) AND THE SOLUTREAN IN SOUTHWESTERN IBERIA Uned. Espacio, Tiempo y Forma. Serie I, Nueva época. Prehistoria y Arqueología, t. 5, 2012. 459 sequence is, so far, similar in the top ca. 2 meters. A considerable amount of Solutrean materials were already recovered from this new sector, with particular emphasis to a set of bifacial stemmed and winged parpalló-type points and a couple of small schist plaquettes with evidences of engraving. 3. SOLUTREAN SEQUENCE AND CHRONOLOGY Presently, there are a total of 15 AMS radiocarbon dates for the Proto-Solutrean and Solutrean occupations at Vale Boi (Table 1). Twelve of these dates were obtained from unidentified charcoal samples and only three were measured from either shell or animal bones. Some of these results are, however, problematic. In the specific case of the samples coming from Terrace’s Layer 3 (Wk-13685, Wk-24761, AA63305, AA-63310), a consistent age-pattern is revealed, with all radiocarbon results being close to 10.9 ka cal BP. Although these ages were extracted from non-identified charcoal remains, two additional samples of identified species were already sent to the radiocarbon laboratory, whose results will hopefully shed some light on this apparent problem. Yet, the classification of both charcoals as Olea sp. may be, by itself and with the right caution, somehow instructive in understanding the problems with the previous samples. Indeed, recent work by Carrión et al. (2010) have demonstrated that, when present and directly dated in Upper Paleolithic layers from Iberia sites (e.g. Buraca Grande, Cova de les Cendres), Olea sp. samples exclusively yielded Holocene ages. The explanation for their presence in Late Pleniglacial contexts has, thus, been always attributed to the existence of post-depositional processes responsible for sample percolation from Holocene deposits. Since the Solutrean sequence of the Terrace area is in direct contact with an early Neolithic occupation from where some remains of Olea sp. were recovered (Figueiral and Carvalho 2006) this seemed, at a first glance, a parsimonious hypothesis. The problem is that, even if we accept some kind of deposit perturbation, the ca. 10.8 ka cal BP dates are incompatible with the ages obtained for the Neolithic horizon, all of ca. 7.0 ka cal BP (Carvalho 2008). Moreover, the total absence of other Neolithic markers (e.g. ceramics) within the Solutrean levels makes it very difficult to believe in the existence of severe mixing processes between the two litostratigraphic units. On the other hand, in the chronological context of the first Holocene human occupations in Western Iberia, the AMS results from Terrace’s Layer 3 are in full agreement with others coming from Epipaleolithic sites of Central Portugal (Bicho, 1994). Hopefully, the still ongoing analysis of the lithic materials would shed some light on the possibility of an, so far undetected, Epipaleolithic occupation at Vale Boi. With respect to the other areas of the site, the reasons behind the problems in the chronology of the dated samples associated with Solutrean materials seem to be of another nature. While in the case of the Slope the two problematic samples (AA-63307 and AA-63308) fit perfectly well in the chronology of the Magdalenian occupation of Vale Boi, evidencing, as discussed above, the problems of the top layers of the Slope section; in the case of the Rockshelter (Wk-17840), some kind of charcoal contamination is the most plausible reason behind the apparent chronological inversion. Overall, the calibration of the accepted dates clearly shows that the Solutrean occupation at Vale Boi is bracketed between ca. 25.0 ka cal BP and 20.3 ka cal BP (1 sigma range). Identification of specific occupational phases within this time span is possible by the existence of four different clusters of dates (Fig. 3), centered around 24.5 ka cal BP; 23.3 ka cal BP; 22 ka cal BP; and 21 ka cal BP. A fifth and Area Layer Lab Age Material Age cal BP (2 sigma) Notes* Rockshelter B Wk-31088 20340±160 Charcoal 24786-23845 R Rockshelter B Wk-17840 18859±90 Shell 22338-21566 A Rockshelter C Wk-24765 19533±92 Charcoal 23720-22684 A Rockshelter C Wk-26800 20620±160 Charcoal 25045-24196 A Rockshelter C Wk-26802 20570±158 Charcoal 25020-24119 A Slope 2 AA-63307 11840±280 Charcoal 14821-13131 R Slope 2 AA-63308 15710±320 Charcoal 19584-18155 R Slope 2 Wk-12131 17634±110 Bone 21405-20518 A Slope 2 Wk-12130 18410±165 Bone 22357-21505 A Slope 3 AA-63306 21420±170 Charcoal 26106-25063 A Slope 3 Wk-16415 21830±195 Shell 26164-25038 A Terrace 3 Wk-13685 8749±58 Charcoal 10116-9548 R Terrace 3 Wk-24761 8886±30 Charcoal 10171-9906 R Terrace 3 AA-63305 8825±57 Charcoal 10160-9684 R Terrace 3 AA-63310 8696±54 Charcoal 9888-9541 R * R – rejected; A - accepted ▲ Table 1. AMS radiocarbon dates for the Proto-Solutrean and Solutrean of Vale Boi. João Cascalheira, Nuno Bicho, João Marreiros, Telmo Pereira, Marina Évora, Miguel Cortés, Juan Gibaja, Tiina Manne, Frederico Regala, Célia Gonçalves y Patrícia Monteiro 460 Uned. Espacio, Tiempo y Forma. Serie I, Nueva época. Prehistoria y Arqueología, t. 5, 2012. oldest cluster is also represented in Figure 3, clearly identifying the Proto-Solutrean occupation of the Slope sector, dated to ca. 25.5 ka cal BP, but which will not be herein referred since it was already addressed elsewhere (Cascalheira and Bicho in press). The Rockshelter presents the most complete Solutrean sequence, covering all dated moments, except for the most recent result of ca. 21 ka cal BP. Nevertheless, given the chronological correspondence between Layer B of the Rockshelter and the bottom of Layer 2 of the Slope sector, it is highly probable that the most recent date coming from the top of Slope’s layer 2 may also correlate with the industries recovered from layer A of the Rockshelter (to which no chronometric evidence is available), marking the final solutrean occupation of the Rockshelter and at Vale Boi. When compared with the currently accepted chronological boundaries for the various Solutrean phases in southern and western Iberia, the results of Vale Boi seem to be slightly displaced. The main reason is not the ages per se, but the materials associated with them. Traditionally, the Solutrean in these regions has been divided into several phases based on the presence or absence of very specific weaponry types. Stemmed and winged points, for example, are thought to be an Upper Solutrean marker, and are recognized in the Portuguese and Spanish records from ca. 23 ka cal BP onwards (Tiffagom 2009; Zilhão 1994, 2013). At Vale Boi, however, this particular type of point appears in the Rockshelter from as early as 24.5 ka cal BP and continues until the end of the Solutrean occupation. Their presence throughout the sequence is not uniform and variations in size have been recognized between the different layers (Cascalheira 2010; Gibaja and Bicho in press). While for the bottom layer (C), Parpalló points present a mean width and thickness of 25.93 mm and 5.33 mm respectively, for the top of the sequence (layer A) a considerable reduction in those dimensions (to 15.85 and 3.83 mm) is rather evident. More interesting is the fact that layer B exhibit average values that fit between the top and bottom moments. Being clearly a chronological shift, it is reasonable to propose that, over time, a successive refinement of this type of point was accomplished at Vale Boi. A trend that accompanies a certain reduction in size of the lithic assemblages as well as the progressive increase in the importance of mediterranean shouldered points (Cascalheira 2010; Bicho et al. 2012). ▲ Figure 3. Plot of Solutrean and Proto-Solutrean accepted radiocarbon dates from Vale Boi with the three moments referred in the text. Calibration curves are IntCal09 and Marine09 (Reimer et al. 2009). VALE BOI (ALGARVE, PORTUGAL) AND THE SOLUTREAN IN SOUTHWESTERN IBERIA Uned. Espacio, Tiempo y Forma. Serie I, Nueva época. Prehistoria y Arqueología, t. 5, 2012. 461 In sum, combining both chronological data with the general vertical distribution of the Solutrean tools group (see Table 3), at least three key-moments in the Solutrean sequence of Vale Boi seem to stand out (Fig. 3): 1. Between ca. 24.5 and 23.3 ka cal BP: first Solutrean occupations of the site with hunting tool-kits dominated by wide and thick stemmed and winged points and rare shouldered projectiles; 2. At ca. 22 ka cal BP: wide variety of lithic points are present in balanced proportions, including stemmed and winged projectiles, laurel leafs and shouldered points with abrupt and plan retouch; 3. At ca. 21 ka cal BP: final Solutrean occupation at the site with a clear predominance of small backed shouldered points and when parpalló-type points, though reduced in number, are much narrower and thinner. 4. LITHIC INDUSTRIES As became clear from the description of the radiocarbon results, the Rockshelter presents the most complete Solutrean sequence of the entire site. Luckily, this is also the locus from where a complete analysis of the lithic assemblages is available. Data presented below, although only referring to the Rockshelter materials, are, therefore, likely representative of the general techno-typological patterns of the Solutrean occupation at Vale Boi. Also, previous works on this topic (Cascalheira 2008, 2010) have revealed that at the technological level, with very few exceptions, no major differences can be identified throughout the stratigraphy, and thus reduction sequences and general technological patterns can and will be presented as a whole. More than 22.000 lithic artifacts were recovered from the Solutrean layers of the Rockshelter (Table 2). The most common raw material is chert (ca. 65%), followed by quartz (ca. 23%), greywacke (ca. 9%), and only 3% of the assemblage is composed of other rocks, like quartzite, schist or chalcedony. All these materials are available regionally, the main three from a radius of no more than 15 km from the archaeological site. In the case of quartz and greywacke, the most probable source was the nearby riverbed. Both these materials were knapped through very expeditious orthogonal reduction sequences, to produce mostly flakes and rare bladelets, not often transformed into formal tools such as notches, scaled pieces and end-scrapers. Simultaneously, greywacke and quartz were also used in very specific tasks: the first, being a variety of a hard sandstone that appears at Vale Boi mostly in the form of large slabs, was frequently used as anvils with the likely task of fracturing animal bones as well as for the preparation of pigments (Bicho et al. 2012; Manne et al. 2012); quartz, on the other hand, and particularly the nodules with both less water-worning and internal homogeneity, was most likely associated with activities involving heating, such as stone-boiling for bone-grease rendering (Stiner 2003; Manne and Bicho 2009). Contrary to other areas in Vale Boi, such as the Slope from where large quantities of fire-cracked quartz were recovered, the Rockshelter presents much lower frequencies of these elements, perhaps due to differences in the functionality of spaces (Cascalheira 2010). Such apparent direct relation between raw materials and tasks are congruent with observations done among modern hunter-gatherer societies showing that raw material selection is more driven by its suitability for the intended task rather than by its flaking properties (Stout 2002; Arthur 2010). This pattern is also congruent with the pattern for almost all the Upper Paleolithic sites in Portugal (Pereira et al. 2012). Schist was mostly used as a blank for mobile art plaquettes (see above) and chalcedony, on the other hand, Layer A Layer B Layer C Total % N % N % N End-scrapers 12.50% 6 13.77% 19 15.94% 11 14.12% 36 Multiple tools 0.00% 0 2.17% 3 1.45% 1 1.57% 4 Perfurators 0.00% 0 1.45% 2 0.00% 0 0.78% 2 Burins 0.00% 0 6.52% 9 2.90% 2 4.31% 11 Backed tools 16.67% 8 2.90% 4 2.90% 2 5.49% 14 Truncations 2.08% 1 2.90% 4 0.00% 0 1.96% 5 Retouched blades 0.00% 0 0.00% 0 1.45% 1 0.39% 1 Solutrean tools 25.00% 12 16.67% 23 17.39% 12 18.43% 47 Domestic tools 27.08% 13 31.88% 44 30.43% 21 30.59% 78 Bladelet tools 2.08% 1 6.52% 9 2.90% 2 4.71% 12 Others 14.58% 7 15.22% 21 24.64% 17 17.65% 45 Totals 100.00% 48 100.00% 138 100.00% 69 100.00% 255 ▲ Table 3. Tool groups counts for Rockshelter. João Cascalheira, Nuno Bicho, João Marreiros, Telmo Pereira, Marina Évora, Miguel Cortés, Juan Gibaja, Tiina Manne, Frederico Regala, Célia Gonçalves y Patrícia Monteiro 462 Uned. Espacio, Tiempo y Forma. Serie I, Nueva época. Prehistoria y Arqueología, t. 5, 2012. was knapped to produce few flakes but it is also present in the form of a bifacial preform. Chert is present in a wide range of chromatic and grainsize varieties. Although the identification of each specific source is still in progress, a large number of primary positioned chert deposits have already been identified in the region (Boix et al. 2011; Verissimo 2004, 2005), whose macroscopic characteristics of the collected samples seem to match some of the types present in Vale Boi (Santos 2005). This correlation between known sources and artifacts indicate an exploitation of the westernmost cliff borders that delimitate the present dry land. Nonetheless, the existence of non-yet-identified sources seems to open the possibility of exploitation of territories located further north, in the Algarve mountainous regions and eastern adjacent areas. If confirmed by ongoing surveys and analysis, these sources might help to shade light on the relation between foodstone acquisition, namely to understand if chert was acquired within dedicated expeditions or if it was gathered when encountered during the execution of other tasks. Both nodular and tabular volumes were used, always very small in size and with frequently natural defects. These characteristics must have greatly influenced the general technological options, namely the overall dimension of the debitage products and tools (Cascalheira 2010). Nodules were, most certainly, introduced at the site with none or very little preparation since decortication is well attested by high frequencies of blanks and tools with total or large percentages of cortex in their dorsal surfaces (Cascalheira 2008). Chert reduction sequences were fairly simple and involve mostly unidirectional or bidirectional prismatic strategies. Core maintenance products are frequent, particularly crested pieces and trimming elements. Flakes are the most common debitage element (ca. 62%), followed by bladelets (ca. 23%) and blades (ca. 6%). The former was also the favorite blank type to the manufacture of formal tools, while blades and bladelets occupy the second and third options, respectively. Typologically, the chert assemblage is dominated by end-scrapers, notches, splintered pieces and unevenly retouched blades/flakes (Table 3). Differences between layers in the typological composition of the assemblages are, as already mentioned, mostly seen in the elements coming from the Solutrean tools group and backed shouldered points. Finally, use-wear analysis of the lithic assemblage revealed that a wide range of activities were accomplished (Gibaja and Bicho 2006). The use of retouched and non-retouched edges for working wood, hide and other soft materials (probably meat, fresh skin, etc.) has been identified. Some of these pieces, either retouched or not, were also used in the manufacture of bone and antler tools (Évora 2007). 5. FAUNAL RECORD Tables 4 and 5 present the NISP (Number of Identifiable Species) and MNI (Minimum Number of Individuals) counts/percentages for respectively vertebrates and shellfish remains recovered from the Solutrean levels of each area of Vale Boi. It should be noted that the values presented for Layer A Layer B Layer C Total % N % N % N Chips 61.55% 2356 60.68% 7858 59.10% 3432 60.42% 13646 Fragments 17.55% 672 19.42% 2515 17.15% 996 18.52% 4183 Flake 8.39% 321 8.59% 1112 9.90% 575 8.89% 2008 Flake proximal frag. 1.52% 58 1.41% 183 1.79% 104 1.53% 345 Flake distal frag. 1.54% 59 1.92% 249 2.60% 151 2.03% 459 Thinning flake 0.26% 10 0.11% 14 0.60% 35 0.26% 59 Blade 0.50% 19 0.48% 62 0.52% 30 0.49% 111 Blade proximal frag. 0.29% 11 0.26% 34 0.22% 13 0.26% 58 Blade distal frag. 0.50% 19 0.44% 57 0.50% 29 0.46% 105 Bladelet 2.06% 79 1.84% 238 1.96% 114 1.91% 431 Bladelet proximal frag. 0.97% 37 0.86% 111 0.53% 31 0.79% 179 Bladelet distal frag. 2.38% 91 1.88% 243 1.77% 103 1.93% 437 Burin Spall 0.13% 5 0.08% 10 0.05% 3 0.08% 18 Core 0.76% 29 0.62% 80 0.98% 57 0.74% 166 Core maint. products 0.29% 11 0.17% 22 0.22% 13 0.20% 46 Anvils 0.13% 5 0.13% 17 0.21% 12 0.15% 34 Manuports 0.03% 1 0.03% 4 0.07% 4 0.04% 9 Tools 1.18% 45 1.09% 141 1.81% 105 1.29% 291 Total 100.00% 3828 100.00% 12950 100.00% 5807 100.00% 22585 ▲ Table 2. Solutrean lithic assemblage of Vale Boi - Rockshelter. VALE BOI (ALGARVE, PORTUGAL) AND THE SOLUTREAN IN SOUTHWESTERN IBERIA Uned. Espacio, Tiempo y Forma. Serie I, Nueva época. Prehistoria y Arqueología, t. 5, 2012. 463 the Terrace are incomplete given the ongoing state of the analysis. Similarly to the patterns observed from the Gravettian and Magdalenian occupations (Manne and Bicho 2009; Manne 2010; Manne et al. 2012), Solutrean vertebrate assemblage is dominated by rabbit, red deer and horse. Although in minor quantities, other species are also present, like wild ass, aurochs and ibex. Carnivores and birds were Taxon Observed Rockshelter Slope Terrace NISP NISP% NISP NISP% NISP NISP% Ungulates Bos primigenius 73 2.4 32 1.4 3 0.4 Equus caballus 276 9.0 193 8.5 56 8.1 Equus sp. (possibly hydruntinus) 27 0.9 48 2.1 10 1.5 Cervus elaphus 371 12.1 813 35.9 149 21.6 Sus scrofa 1 0.0 1 0.1 Capra/Ovis 3 0.1 5 0.2 3 0.4 Small Mammals Oryctolagus cuninculus 2274 74.4 1147 50.7 457 66.3 Microtus sp.7 0.2 Carnivores Vulpes vulpes 4 0.1 2 0.1 Canis lupus 2 0.1 1 0.0 Canis spp.3 0.1 3 0.1 Panthera leo 1 0.0 2 0.3 Lynx pardina 1 3 0.1 1 0.1 Felis sylvestris 6 5 0.2 2 0.3 Mustela sp.1 1 0.0 Martes sp.3 0.1 Aves Small bird 1 0.0 1 0.1 Medium bird 4 0.1 1 0.0 Large bird 1 0.0 1 0.1 Totals 3052 2261 686 ▲ Table 4. Solutrean vertebrate NISP and NISP% for Vale Boi site areas (Terrace, Mid Slope and Rockshelter). Taxon Observed Terrace Slope Rockshelter MNI MNI% MNI MNI% MNI MNI% Mytilus edulis 21 5.1 5 3.7 Pecten maximus 3 0.7 1 0.7 Acanthocardia sp.2 0.5 1 0.7 Cerastoderma edule 13 9.6 Cardiidae 2 1.5 Callista chione Ruditapes decussatus 1 33 9 2.2 2 1.5 Veneridae 2 0.5 Patella sp.2 67 374 90.8 106 77.9 Cerithiidae Lacuna parva 1 0.7 Littorina 4 2.9 Naticidae Gibbula sp.1 0.2 Pollicipes pollicipes 1 0.7 Totals 3  412  136  ▲ Table 5. MNI counts and % MNI of shellfish (and Pollicipes pollicipes ) for Terrace, Slope and Rockshelter assemblages.