In the Vale Boi Rockshelter an approximate area of 20m2 has been excavated, in which three consecutive levels are attributed to the Gravettian, Solutrean and Magdalenian. The Magdalenian materials are very scarce and probably the human occupation extends to the southern side off the excavation area (Mendonça 2009). Both Gravettian and Solutrean (Cascalheira 2010) deposits are sealed bellow the limestone boulders that, around the Last Glacial Maximum, collapsed from the shelter cover (Bicho et al. in press). The Gravettian collection is fairly small with flakes, cores and double backed points, and marked by the presence of faunal remains well preserved. The dispersion of the artifacts and fauna, however, seem to suggest some kind of erosive process took over the deposit.
There are two Gravettian levels in the Terrace section of the site. The remains from the upper occupation are associated with a hearth, that contained charcoal dated to c. 24 K RCYBP. The lower level contains a high concentration of lithic and faunal remains, with a date on charcoal of c. 28 RCYBP. Although there is a thick hiatus and sediment changes between the occupations, from the technological perspective both industries are similar, even though double backed points are only associated with the lower level.
The Proto-Solutrean is present in the Slope and Terrace areas. In the case of the Terrace where the lithic assemblages have been analyzed, materials are present in the top artificial levels of layer 4. Unfortunately, there are not radiometric dates at the moment. Like the underlaid Gravettian, the most common raw materials are chert, quartz and greywacke. Their technological traits are, however, distinctive. Quartz is the most common raw material in the assemblage like what is known in Portuguese Estremadura (Almeida, 2000). For the first time, there are exotic raw materials: jasper, quartzite, probably both exogenous to the area, and chalcedony, a local mineral. Flakes dominate the quartz blank assemblage, while elongated products are rare. Quartz flakes were used as blanks for notches, endscrapers, sidescrapers and scaled pieces.
The tool kit is mostly represented by tools made on chert. It seems evident that certain classes and retouched tool types such as endscrapers, elongated blanks with continuous retouched in one or both edges, denticulates, were intentionally made on local chert.
Jasper was exclusively used to make Vale Comprido Points. The closest source for this his exogenous rock can be found in secondary position in the riverbanks of the Guadiana River, close to 100 km from the site (Francisco Almeida, personal communication). The Vale Comprido points are known for the Portuguese Estremadura as type-fossils for the Proto-Solutrean (Zilhão 1997), the transition culture between the Gravettian and the Solutrean.
Although some points à face plan have been identified in the early Solutrean levels in the Slope area of Vale Boi, the existence, in Southern Portugal, of a Middle Solutrean occupation, in its traditional definition, still cannot be confirmed. Thus, only Upper Solutrean lithic technology can be clearly described, based at the moment, exclusively on the materials recovered from the Rockshelter section.
While in techno-typological terms the Solutrean lithic industries are naturally different from the Gravettian assemblages, the choice and use of raw materials was essentially the same. Local chert, quartz and greywacke were the most used materials. The differences can be found in much lower percentages of quartz than in the previous phase, and in the sporadic use of chalcedony for the manufacture of bifacial implements. Quartz and greywacke maintain, also, their functional dichotomy either as knappable materials for, mainly, flake production or as pounding elements associated with activities involving heating, such as stone-boiling for bone-grease rendering (Manne et al, in press).
In the case of chert, the patterns of exploitation indicate favored debitage sequences, relatively simple, through which it would be obtained, by a well-structured management of the nodules, all the blanks needed for the manufacture of both bifacial and non-bifacial tools.
Though still within a context of flake dominated reduction sequences, a greater focus on blade extraction than in the previous technocomplexes is evident in the assemblage. The blanks were then transformed into a great variety of tools including small à cran backed points and the rarer and larger Atlantic à cran points with invasive retouch. Other projectile points, such as the traditional Solutrean laurel and false willow leaf bifacial points, as well as the stemmed Parpalló points were also produced on-site (attested by the presence of a good quantity of bifacial thinning flakes and bifacial preforms) and within the sequences mentioned, mostly by the progressive bifacial thinning of thick and large flakes. Heat treatment has been identified (Gibaja and Bicho, in press), but is only observed from final production phases of bifacial elements, when meticulous retouch by pressure was applied to finish the pieces (Cascalheira 2010).
Stemmed projectiles appear at Vale Boi in a great variety of subtypes, including the so-called “pedunculated arrow” (Zilhão, 1997) present in Portugal only at the Solutrean assemblages of Salemas cave. There is also a unique, very small, flattish type, stemmed only by one notch in each side and semi-abruptly retouched on both sides to make it pointed. On the other hand, and within this diversity, a gradual assimilation of the concept of the stemmed point was possible to identify. In fact, the differentiation between layers for this type of projectile seems to reveal a first phase of its adoption, around 20 000 BP. It is represented by the presence of wider, thicker, and coarser retouched elements, passing through an intermediate stage in which the morphologies are more varied and the improvement of the bifacial technology is well-evident. This phase culminates in a decrease in the percentage of these points, but the projectiles became the most standardized in terms of technical detail. Furthermore, the gradual disappearance of the stemmed bifacial elements is accompanied by a tendency of increased microlithization through time, with a propensity for the miniaturization of weaponry (Cascalheira, 2010; Gibaja & Bicho, in press).
Data on the Magdalenian of southern Portugal are still relatively scarce, although there are a few sites where it is known (Vale Boi, Vale Santo 4, Ponta Garcia, Cruz da Pedra, Lagoa do Bordoal and Praia da Galé). The new information comes only from Vale Boi, since no other sites was object of either field work or more extensive analyses (Aubry and Bicho, 2006). In fact, and although there is only one absolute dating by OSL to the Magdalenian of Algarve (c. 14,800 cal BP), a recent technology analysis carried out seems to point to the presence of at least two internal chronological phases: a Magdalenian early/middle and final Magdalenian (Mendonça 2009). The existence of a Solutreo-Gravettian has been suggested by Zilhão (1997), but its presence was never confirmed.
There seems to be a tendency for microlithization. However, this trend is mostly marked in a later phase of the Magdalenian. Alrthough, an increase in back bladelets seems to be present in the Magdalenian (Mendonça, 2009), is never comparable to that seen either in the Valencian region, the traditional area where the Solutreo-Gravettian is better known (Cascalheira, 2010), or in Portuguese Estremadura.
In the regional Magdalenian technological model the production of elongated elements is clearly marginal, with a few bladelets and very rare blades. Naturally, this fact led to a composition of the tool kit based on flake tools (scaled pieces, denticulates, notches, endscrapers and rarer burins) although bladelet tools are present, but in very low frequencies (Mendonça, 2009). The truth is that in the Upper Paleolithic of Algarve the backed elements are quite rare, so this feature may be more than a Magdalenian technology gap. On the other hand, scaled pieces are very abundant during the whole Upper Paleolithic in Algarve.
Although not comparable to the numbers of Gravettian and Solutrean, there are mollusc remains (limpets and cockles) in Vale Boi pointing to the use of the coastal system (both rocky and sandy bottoms), as in the previous periods. The Mammalian faunas are still marked by the presence of the same previous main species, red deer and rabbits, which seemed to have increased in the Magdalenian, probably associated with improved climatic conditions (Manne and Bicho, 2009). Albeit in small numbers, the aurochs and the horse continued to be hunted, but there is no trace of wild boar, goat or any equids. This absence may be linked to the insignificant size of the faunal sample. Anyway, it seems clear that, as in the Portuguese Estremadura, the overall picture is of diversification and intensification of the exploitation of food resources.