Late glacial prehistory of central and southern Portugal.

Bicho, Nuno Ferreira

Very little is known of the Upper Palaeolithic of Portugal, although it has been assumed to have the same general characteristics as elsewhere in southwestern Europe. New evidence suggests clear technological distinctions between Portugal and other areas of southwestern Europe after the Last Glacial Maximum, c. 18,000 (uncalibrated) years ago, and allows an initial synthesis for Portuguese Late Glacial prehistory, 16,000-8500 b.p.

Introduction

Western Europe saw important transformations in hunter-gatherer lifeways after the Last Glacial Maximum (LGM). Among these changes, faster and more significant after the Holocene began, are the lack of a bifacial lithic technology such as that of the Solutrean, an increase in microlithic point production and more diverse and complex lithic assemblages (Audouze 1987; Burdukiewicz 1986; Campbell 1986; Straus 1986a; 1986b). Along with these lithic changes, new hunting techniques (Straus 1987) and economic systems exploiting a larger range of animals and plants (Clark & Straus 1986) were developed. These transformations were, sometimes, directly or indirectly caused by environmental changes. Rising sea level and increased global temperature altered river valleys and lacustrine environments, introducing animals and plants as well as expanding and contracting their range. Though, in a certain way, similar in all western Europe, these changes were neither uniform nor contemporaneous across the region. After the LGM, central and southern Portugal, and most likely southern Spain, felt a faster and earlier palaeoenvironmental transformation than northern Iberia or France.

New data on palaeoenvironment and tardiglacial Portuguese prehistory show significant differences between Portugal and the traditional core areas of western Europe. This paper is based on recent analysis by this author of material from the Bocas rock-shelter and from the Estremadura project (directed by Drs Marks and Zilhao), as well as other published material.

The 1993 radiocarbon calibration extends the period where determinations can be calibrated back to the LGM; but here, uncalibrated determinations in years 'b.p.' are used, as has been customary.

Brief history of Portuguese Palaeolithic archaeology

Research on the Palaeolithic period began early in Portugal. Cave excavations carried out in 1866 by the National Geological Survey (created in 1848; Zilhao in press a) were directed by the geologists Carlos Ribeiro and Joaquim Filipe Nery Delgado, participants in the major 19th-century debate over eoliths and Tertiary humans (Grayson 1986). They were also, without knowing it, pioneers in work on site formation processes and taphonomic problems (Zilhao in press b). Despite these promising independent origins, Portuguese Palaeolithic research has, since then, been strongly influenced by French research.

In 1918, Abbe Breuil published a paper in a Portuguese journal describing Palaeolithic finds from the Lisbon peninsula, pointing out the similarities with the French Palaeolithic. Twenty years later, Breuil returned to Portugal to expand his first 'Impressions', and survey fluvial and coastal Pleistocene terraces with Georges Zbyszewski (Breuil & Zbyszewski 1942; 1945). They found large river valleys, as well as coastal cliffs, very rich in Palaeolithic artefacts. Naturally, the lithic artefacts were divided following the French chrono-cultural classification of Breuil (1912).

During some 20 years, beginning in the late 1930s, Manuel Heleno, director of the National Museum of Archaeology, excavated in the Rio Maior (inland) and Torres Vedras (coastal area north of Tagus river) areas of the Portuguese Estremadura. Unfortunately, these excavations did not follow modern procedures. Entire sites were excavated in long trenches with very poor vertical control. The results were not published beyond a few short notes labelling the assemblages by Breuil's classification (Heleno 1944; 1956; cf. Zilhao 1985), which brought 'Aurignacian', 'Gravettian', 'Solutrean' and 'Magdalenian' into the Portuguese Palaeolithic. Heleno was not an archaeologist, but an historian interested only in establishing the antiquity and background of the 'Portuguese nation'. Like Breuil, he saw type-fossils as indicative of whole industries. As a consequence, he assigned wrong classifications to some assemblages (Zilhao 1985; 1987).

In the decade after Heleno's 1956 note, Jean Roche began research on the Portuguese Palaeolithic (Roche 1964; 1979; 1982). Breuil's classification was again used to identify lithic assemblages from various cave and open-air sites, and again, re-evaluation of cultural attributions has proved some wrong (Zilhao 1987; 1988).

Recent work by Zilhao and Araujo has greatly advanced the study of the Final Upper Palaeolithic and early Mesolithic. They have studied tardi-glacial and early Holocene sites (Araujo in press; Araujo & Zilhao in press; Zilhao et al. 1987), and helped to correct cultural attributions of assemblages excavated prior to 1980 (Zilhao 1985; 1987; 1988). Yet, except for Zilhao's (1987; 1990; 1991) study of the Solutrean, there is no general synthesis of the Portuguese Late Upper Palaeolithic.

In 1987, Marks and Zilhao began in the Portuguese Estremadura to build a chronological, environmental and cultural scheme for the Upper Palaeolithic and Mesolithic of Portugal (Marks et al. in press). These new data indicate that such periods as the Magdalenian or the Azilian, well defined and represented in Cantabrian Spain and western France, are not directly applicable to the terminal Pleistocene occupations of Portugal (Bicho 1992a). The term Magdalenian, applied to Portuguese assemblages, can be defined as a lithic techno-complex based on flake and bladelet productions, covering the time-span from after the Late Glacial Maximum (c. 18,000 b.p.) to Holocene times (c. 9000 b.p.). It can be isolated from the Portuguese Solutrean by the lack of a bifacial technology, a general decrease in tool size and an increase in backed bladelets and micropoints (Bicho 1992a). After Solutrean times, major technological and typological changes occurred only after 8500 b.p., with the introduction of the Mesolithic economy and the large-scale production of geometrics using the micro-burin technique (Arnaud 1987; 1990; Vierra 1992). Although the Portuguese 'Magdalenian' is different from that in northern Iberia, it shows the same general variability as in the rest of western Europe. This wide variability both at the micro- and macro-regional scale contrasts with the Solutrean, which is more uniform across space. In Portugal, the term Magdalenian is, essentially, a temporal term, while Solutrean has temporal, technological, geographical and cultural meaning.

Palaeoenvironment of central Portugal: the Estremadura region

Portuguese Estremadura is located between c. 40 |degrees~ and 38 |degrees~ 30|prime~ latitude north. A coastal area, Estremadura is bordered by the river Mondego to the north and by the Sado and Tagus river valleys to the south and east. The environment of central Portugal is deeply marked by the wide and plain valleys of the Tagus and the Sado and their tributaries, as well as by the great topographic variability of the limestone massif: the mountains of Serra d'Aire, Candeeiros and Montejunto all rising above 600 metres. The coast, characterized by a sinuous line of peninsulas, capes and river estuaries, is marked by sand dunes dating to the Late Pleistocene and Holocene, separating enclaves of much older high cliffs. The present climate, influenced by Atlantic and Mediterranean elements, has a high gradient from north to south in both temperature and rainfall (Ribeiro et al. 1988).

The traditional scenario for the Portuguese palaeoenvironment during the LGM, based on geomorphology and the pollen sequence from northern Spain and southern France (between 20,000 b.p. and 18,000 b.p.), is of a treeless tundra vegetation above 700 m in the mountains of Estrela, Cabreira, Geres and Peneda in central and northern Portugal (Daveau 1971; 1973; 1980; 1986). Lower altitudes were characterized by an open pine forest, thicker in the river valleys. The polar front located around the 42 |degrees~ parallel, with its strong, salty winds, eliminated the vegetation cover and deposited sand dunes, creating a 'coastal desert' as far south as Lisbon (Daveau 1986; Zilhao 1987). This scenario is believed to have been stable between 15,000 b.p. and c. 10,000 b.p., except for the contraction of the glaciers in the lowlands around the mountains. By 11,000 b.p., a cold and dry period (corresponding to Dryas III pollen zone sequence) caused an expansion of the glaciers of Estrela and the regression of the pine forests to more southern latitudes. Although palaeoenvironmental information is still incomplete, new data from Rio Maior, various Estremaduran caves and two lacustrine pollen sequences from the southern coast suggest instead that climatic amelioration in central and southern Portugal occurred earlier and more rapidly than in the rest of western Europe. The replacement of Atlantic by Mediterranean species most likely occurred immediately after the regression of the polar front located at 42 |degrees~ latitude, around the Gulf of Biscay (CLIMAP 1976; McIntyre & Kipp 1976; Ruddiman & McIntyre 1981). High frequencies of such microfaunal species as Apodemus sylvaticus, Eliomys quercinus and Terricola duodecimcostatus and reduced Microtus agrestis occur in a stratum (Eb) of Caldeirao cave dated to between c. 15,000 b.p. and 10,000 b.p. (Povoas et al. 1992). The increase of these three rodent species indicates the development of a Mediterranean climate, probably with pine, olive and oak forests, in a relatively high area (160 m) characterized today by a pre-Atlantic climate. The stratum lacks such colder climate macrofaunal species as ibex (Capra pyrenaica) and chamois (Rupicapra rupicapra), very important components of the earlier Solutrean occupation of the cave (Povoas et al. 1991; Zilhao 1991), and other Solutrean sites (Cardoso 1992). In the Magdalenian level of Caldeirao and at the Algar de Cascais and Algar Joao Ramos (the latter two are not anthropic assemblages: Cardoso 1992) Capra pyrenaica and Rupicapra rupicapra were replaced by wild boar (Sus scrofa) and roe deer (Capreolus TABULAR DATA OMITTED capreolus), two species adapted to forested areas, as well as by red deer (Cervus elaphus), aurochs (Bos primigenius) and horse (Equus), all species adapted to more open temperate environments and humid in the case of aurochs (Cardoso 1992). At Cabeco de Porto Marinho, wood charcoal from an archaeological level dated to 11,200 b.p. (Figueiral in press) indicates a Mediterranean forest with pine (Pinus pinea and Pinus pinaster), evergreen and deciduous oaks (Quercus ilex/suber), birch (Fraxinus angustifolia), wild strawberry tree (Arbutus unedo) and olive (Olea europea ?sylvestris). Olive and wild strawberry trees suggest a warm and dry climate, similar to that of today.

Pollen sequences are known from very few sites, all dated to Holocene times. Lagoa Comprida, located in the Serra da Estrela at a high elevation of 1600 m, provides a sequence dated between c. 9200 b.p. and 850 b.p. (Janssen & Woldringh 1981; Van den Brink & Janssen 1985). The lowermost zone is characterized by pine pollen (Pinus sylvestris), with some non-arboreal pollen, suggesting an open pine forest. Around 9000 b.p., the pollen assemblage exhibits a high frequency of oak pollen and low frequencies of pine and non-arboreal pollens. This oak forest lasted for about 800 years;. it was replaced around 8300 b.p. by a mixed deciduous oak and birch forest that covered the landscape until deforestation, around 5000 years ago, probably due to human action (Van den Brink & Janssen 1985; Chester & James 1991). Another pollen sequence from Lagoa Travessa, south of Lisbon, is dated between 7580 |+ or -~ 70 b.p. and present times (Mateus 1985). After a coastal pine forest (Pinus pinaster) dated to 7600-6500 b.p., there was an increase in oak and alder pollens, as well as coastal scrub and salt marsh species, corresponding to a marine transgression and the retreat of the coastline.

These data show for the tardi-glacial and early Holocene in central Portugal a gradual increase of vegetational density and diversity, with a mix of Atlantic and Mediterranean species in the lowlands and river valleys. This gradual increase was caused by the northern expansion of the Mediterranean biotic communities that lived in southernmost Iberia during the LGM. After the LGM, the tree and shrub vegetation mantle of the mountains (cold-adapted pine species and some steppe vegetation) slowly migrated upwards. Faunal communities that had been using the same areas slowly disaggregated, with cold-adapted ibex and chamois moving towards higher altitudes, while species like red deer, roe deer, wild boar, horse and aurochs shared the lowlands.

Regional geomorphology was altered by the cut and fill of rivers, caused by the Atlantic transgression during the last deglaciation, culminating around 5000 b.p., with high sea level and lacustrine environments in the Tagus and Sado basins. This cut and fill opened areas for localized eolian movement of sands that would affect human settlement, fauna and flora, while the coast was a permanent area of beach and eolian action, as in the present.

In summary, the climate would have been moderate and relatively dry after 16,000 b.p. with a progressive increase through time in both temperature and moisture. The location and abundance of the different trees were more a local consequence of soil composition, ground water availability and atmospheric availability than global climatic conditions.

The sites: dating and location

There are only 18 sites in Portugal dated between 16,000 b.p. and c. 8500 b.p., although some have multiple levels. At Carneira there are four different areas: Carneira, excavated by Heleno some 50 years ago; Olival; Pinhal; and Carneira II (Bicho 1992a; Marks et al. in press). Cabeco do Porto Marinho (CPM) also has multiple areas and archaeological levels spanning the Aurignacian to the Bronze Age, with 11 levels dating between 16,000 b.p. and 9000 b.p. (Bicho 1992a; Marks et al. in press). Lapa do Suao, a cave site, contains Upper Palaeolithic and Epipalaeolithic occupations among others (Roche 1979; 1982). Finally, the rock-shelter of Bocas yielded five archaeological levels, of which the three lowest are dated to c. 10,000 years ago. Most occupations (site area, or archaeological level) are dated to after 12,000 b.p. Although the overall number of sites is very low, their location suggests that the coast was only rarely used as a settlement area before 10,000 years ago. Marine shells and fish bones from Caldeirao cave (Zilhao 1992a; 1992b) show human movement between inland and coastal areas. In Pre-Boreal and Boreal times both coastal and inland areas were widely used by the hunter-gatherer communities. Sites tend to concentrate in discrete areas, around the Rio Maior and Mira rivers, or the coastal area north of the Tagus River. Shell-middens are present only after 10,000 b.p.; both coastal (Pedra do Patacho) and inland sites (e.g. Bocas and Casal Papagaio) have many marine and estuarine shells, which indicate regular movement to and from the coast. While the coastal site, Pedra do Patacho, is an open settlement, the two inland shell-middens are in caves. All cave sites are inland, suggesting that present coastal caves were not used during the terminal Upper Palaeolithic, as they were during the Solutrean.

Open-air sites are located at low altitude (generally below 100 m a.s.l.) in sand deposits, frequently dunes, within a few hundred metres of a stream or near the confluence of two rivers. They also overlook water from high points on the landscape; the southern sites, Palheiroes do Alegra and Pedra do Patacho, overlook the Atlantic and, in the case of the latter, the estuary of the Mira River.

Features and site function

In the Rio Maior area, the sites are on south-facing gentle slopes, on the higher sand ridges between the rivers, sunnier locations that offered protection from the predominant north winds. Hearths from the Rio Maior sites have rocks and cobbles which would have provided protection from the wind, but two Areeiro III hearths did not contain rocks. They consisted only of large chunks of wood charcoal, piled against each other in a concentric manner and protected by a natural erosional cut of the sand dune (Bicho 1991). The stone hearths found at some Rio Maior sites, as well as other sites, such as Ponta da Vigia (Zilhao et al. 1987), have semi-circles of angular, broken rocks and large cobbles, as well as large, fractured quartzite cores. The large cobbles are mostly quartz and quartzite, while the angular rock fragments are quartz, quartzite, sandstone and basalt. Around these hearths the artefact density is very high (Bicho 1992a; Marks 1991). At Palheiroes do Alegra, the various hearths are 'cuvettes', circular in shape with a few rocks and slightly elevated at the centre, where the sand is very dark due to fire (Raposo et al. 1989). They are, most likely, the same kind of hearth as at Areeiro III, but the charcoal has disappeared as Palheiroes do Alegra is a surface site. The Rio Maior archaeological levels are characterized by high frequencies of different sized broken rocks and gravel, sometimes burned and very abundant, although very rarely do they resemble the Magdalenian-age pavements found in open-air sites in the Isle Valley of the Perigord (Gaussen 1980). Only one feature suggests a purposeful shape, a rectangle c. 3 x 2 m. In general, these rock concentrations are denser in the occupations dated around 12,000 to 11,000 b.p. than in those older or younger.

Apparently, there are different types of sites. The most common type is very large, with large areas of occupation (more than 100 sq. m in area), very large lithic collections and features such as hearths and rock concentrations. Such sites could possibly represent what Binford (1980) has called base camps. The best examples of these occupations are some later levels of Cabeco do Porto Marinho, Palheiroes do Alegra and, possibly, Areeiro III (although this last might correspond to a successive number of occupations during a few hundred years). The only large occupation in a cave is represented by the levels 'Fundo' and '1' of the rock-shelter of Bocas I, with large concentrations of lithic materials, including many large quartzite and quartz cores, and thick deposits. In level '1', a thick shell-midden marked by an abundance of estuarine shells, further indicates a residential base camp. Short term open air camps, extending over only a few square metres and, in general, concentrating around a single hearth, are found in some levels of Cabeco do Porto Marinho and some Carneira areas. This type, may in fact, represent remains of specific activities, where there is low tool richness (cf. Chatters 1987; Thomas 1988; 1989), smaller assemblage size, and fewer and less

diversified features, as in the level dated to 10,940 b.p. of the 'CPM III Trench' area.

At Caldeirao cave (Zilhao 1992b) and in level '0' of Bocas I, although the assemblage size is relatively large, the few tool classes (i.e. low artefact class richness), high frequency of a very specialized type of microlithic backed point, and a very narrow range of lithic raw materials suggest possible logistical camps. The lack of meat-bearing bones (Binford 1978; Chatters 1987; Davidson 1983) such as vertebrae, ribs or proximal epiphyses of long bones, suggest a dismemberment and butchering camp, consistent with the presence of numerous aurochs and red deer teeth, as well as distal fragments of tibia and aurochs astragalus. The location of Bocas is relevant, since the site is in the middle of a deep and narrow canyon of the Rio Maior, which provides the only easy communication between the eastern and western parts of the region. The entrance of this canyon would be the perfect place to ambush large game like aurochs and red deer, as well as species of more forested areas such as wild boar and roe deer.

The bone and lithic assemblages: reduction strategies, technology and typology

Bone tools are rare in the Portuguese Palaeolithic, most likely due to the low number of sites with preserved faunal remains. Bone tools, mostly fragments, dated to after the LGM are present only at Caldeirao cave (Zilhao pers. comm. 1992) and Bocas rock-shelter. These fragments are, in general, distal portions of sagaies (bone joints). An unpublished fragment recovered in the late 1930s by Manuel Heleno in Bocas, is decorated with short, oblique parallel lines. There is no evidence for harpoons.

Lithic raw materials largely indicate dependence upon local sources with a strong preference for cherts. In the south, the local raw material is greywacke, available on beaches as cobbles, and the assemblages are mostly composed of it. Flint, not readily available locally, occurs in very low frequencies, used intensively for the manufacture of retouched tools. In the Rio Maior area and in the coast north of the Tagus, quartzite and quartz are immediately available. At Rio Maior, flint is also available at a distance of 1-3 km, depending on the site (Marks et al. 1990), and assemblages are mainly of flint. There were three different economic strategies for the different raw materials. The strategy for quartz and quartzite was expediency (Binford 1979); time consumption for lithic procurement and transportation was practically non-existent, shaping and preparation of the core very simple, core maintenance very rare or absent. Also, the retouched tools on these raw materials are very rare and, frequently, of very simple form. This strategy is seen in Rio Maior (Bicho 1992a; in press a), Palheiroes do Alegra (Raposo et al. 1989; Vierra 1992), Caldeirao (Zilhao n.d.) and Ponta da Vigia (Vierra 1992; Zilhao et al. 1987). The second strategy also present at these three latter sites, was TABULAR DATA OMITTED curation, used for the silicious raw materials (Binford 1979), since the flint, jasper and chalcedony were brought from non-local sources. The third strategy for flint and other silicious raw materials was expenditure, time-consuming in transport of nodules, as well as in their shaping, preparation and maintenance, which produced large amounts of debitage and blanks for manufacture of formal retouched tools. These tools were manufactured, used and discarded frequently since the raw material was abundant. Different reduction sequences were used for different raw materials (Bicho in press b). While flint was largely used in all strategies, quartz and quartzite were used mostly in unidirectional flaking; rarely quartz and quartzite cores exhibit evidence for bi-directional and multi-directional strategies. Shaping and preparation of the core was very different among the raw materials. Flint cores were extensively prepared and the cortex removed from all faces, while the cortex of quartz cores was rarely extensively removed. This pattern was more extreme in the case of quartzite, where the original core faces were rarely and never very extensively prepared (Bicho 1992a).

The ratio of flint to non-silicious raw materials changed through time and was different between the two facies ('Carinated' and 'Rio Maior'); the variability, both diachronic and synchronic, is related to the increase in bladelet production. Although bladelet cores were made in quartz, there are very few complete quartz bladelets. Most complete bladelets and all retouched bladelet tools are on flint, suggesting that flint was preferred for bladelet production because it fractured less. As bladelet production increased, the frequency of non-silicious raw materials decreased.

The assemblages from the Rio Maior area clustered into various groups (Bicho 1992a; 1992b) based on technological attributes, that indicate both diachronic change and synchronic diversity. The two technological facies (synchronic variability) can be easily separated by high and low frequencies of the carinated reduction strategy. Within each technological facies there are temporal divisions. The 'Carinated' facies was present after 11,000 b.p., while the 'Rio Maior' facies was present, at least, since 16,000 b.p.

The 'Rio Maior' facies, the most common, is dated, in a first phase to between 16,000 b.p. and 15,000 b.p.. Quartz and quartzite expedient technology was common. Assemblages dated to 14,000 b.p. are marked by a high frequency of quartz materials, in the form of very large cores, as well as an abundance of large flakes. Flint material is essentially similar to that of the older assemblages. From 12,000 b.p. and 9000 b.p., two different phases (12,000 b.p. to 10,500 b.p. and 10,500 b.p. to 9000 b.p.) share an increase in platform facetting, an increase in bladelet production and in flint use, this increase depending on the distance to the source (Bicho 1992a).

The 'Carinated' facies is characterized by the extensive use of carinated technology, although other strategies (unidirectional, opposed, multi-directional) were also largely used. Non-silicious raw materials were rare. Bladelet size is significantly different from those of the other assemblages, while the bladelet tends to be twisted and pointed. The 'Carinated' facies also shows some temporal diversity. Pinhal da Carneira, dated to c. 11,000 b.p., is still very similar both in technology and typology to most other assemblages dated to between 11,000 b.p. and 9000 b.p., but the younger assemblage, Areeiro III dated to c. 8500 b.p., is very different. This places the change from Palaeolithic to Mesolithic later than elsewhere, since it is only after 8500 b.p. that we see the material evidence (fauna, lithic material, site location, settlement pattern) corresponding to the two social-economic systems, change significantly.

Conclusions: subsistence and economy of the tardi-glacial hunter-gatherers of central and southern Portugal

Tardi-glacial human occupations are known mostly from inland open air sites in Estremadura, with only two sites located south of the Tagus, close to the Mira River. This is a major contrast with the earlier period, the Solutrean, when both cave and open air sites were frequently used at inland and present coastal areas (Zilhao 1992b). Although the distances from Solutrean and Magdalenian sites to the present coast are different, distances to contemporary shore-lines may have been the same. There is a significant difference in absolute areas and range of occupation. After the LGM, coastal resources were more frequently used as shown by marine shell, used in adornments, and the increase of fish remains at Caldeirao cave (Zilhao 1992a; n.d.). There is earlier evidence for shellfish consumption at Figueira Brava Cave, south of Lisbon, during the final Mousterian c. 30,000 b.p. (Antunes et al. 1989; Cardoso 1992). The human diet after the LGM was based on the large herbivores such as red deer, roe deer, wild boar, horse, and aurochs. From 10,000 b.p. there was a clear change in diet. Shell-middens, both inland and coastal, indicate intensive use of estuarine and coastal environments. There was regional variation in human subsistence. Hilly areas, around the Caldeirao and Casal Papagaio caves, were probably relatively dense forests, where the main prey species was red deer, although rabbit and hare were also largely exploited (Zilhao 1992b), and roe deer and wild boar less so (Zilhao 1987; 1992b). This faunal assemblage contrasts sharply with that of the LGM, with ibex and chamois (Zilhao 1992b). It is also different from other assemblages also dated to after the LGM but located in flat plains such as the Rio Maior area. In these flatter, lower areas, the river banks were forested with Mediterranean and Atlantic tree types and contained animal species such as red deer and wild boar. Further from the river, on the plains, aurochs and horse lived in open Mediterranean woodland with oaks, pines and olive trees. At Bocas, close to flat plains and forested river banks, the prey species were red deer and aurochs. There is indirect evidence that tardi-glacial hunter-gatherers were exploiting plant resources; grinding stones (including hand-stones and milling-stones), and charcoal from tree species which provide edible nuts and fruits such as olives, acorns and wild tree strawberries. The grinding stones could have been for ochre-processing, but there is no macroscopic evidence for this. It is reasonable to think that they were used for plant-processing. After 16,000 b.p., the subsistence of the Tardi-glacial human groups was based on terrestrial mammals, mostly red deer and wild boar or aurochs. Rabbits and hares were part of the diet, mostly in hilly areas. Plant foods were available and were most likely exploited. Shellfish, present at sites through and after the LGM, became important only c. 10,000 b.p.. The inland shell-middens suggest groups frequently moving between coastal and inland areas, using coastal, estuarine, inland plain and inland mountain zones. During this period both bladelet and micropoints became more common and smaller in size. The increase of micropoints is possibly related to the development of tip/barb for bow-and-arrow technology, as well as to change in mobility, also suggested by different and more complex site location. The increase in mobility was most likely related to the availability of new and more diversified resources, that made possible a new scheduling of economic activities, depending more on the seasonal rotation of the natural resources such as plants, small mammals, fish, and birds. In the case of Rio Maior, the coastal-inland movement may be a consequence of the presence of good quality flint, although flint is found at a few coastal points such as Lisbon and Nazare. For the southern sites, presently known only on the coast, the likely sources for silicious raw material are the inland mountains (Vierra 1992), suggesting, again, movement between coastal and inland areas.

The change in the lithic technology, possibly an indirect consequence of change in vegetation density from open to closed forests, culminated around 6,500 years ago during the Flandrian transgression, with extensive use of microburin technique and mass production of geometrics, seen in the shell-middens of the Tagus and Sado basins (Arnaud 1990), and on the Estremadura and Alentejo coasts (Vierra 1992).

The data presented here indicate palaeo-environmental changes earlier in central Portugal than in the rest of Europe, influencing technology, hunting strategies, and economic systems. This technology, more settled and apparently more reliable, lasted until later, 8,500 b.p., than in other areas of Europe, with a stability clearly different from the traditional 'Magdalenian' seen elsewhere. Acknowledgements. I thank the National Science Foundation (Grant BNS-8803798 and BNS-9107144), the Portuguese Government (Junta Nacional de Investigacao Cientifica e Tecnologica, and Department of Archaeology of the Instituto Portugues do Patrimonio Cultural), and the Institute for the Study of Earth and Man, Southern Methodist University (Seed Grant) for funding the Estremadura research. The dates from Bocas I were obtained through the good offices of the Museu Nacional de Arqueologia. I am very grateful to Dr A.E. Marks for his help and comments during these last five years of work. I also wish to thank Dr David Meltzer for comments on an earlier version of this paper, and my wife, Dr Maria Masucci, who helped in the editing. In addition, helpful comments were offered by an anonymous reviewer. The responsibility for the lithic analysis, interpretation, conclusions and any possible errors are mine alone. References

ANTUNES, M.T., J.M. PEIXOTO CABRAL, J.L. CARDOSO, A. MONGE SOARES. 1989. Paleolitico Medio e Superior em Portugal: datas 14C, estado actual dos conhecimentos, sintese e discussao, Ciencias da Terra 10: 127-38.

ARAUJO, A.C. In press. O concheiro Epipaleolitico do Cabeco do Curral Velho (Cambelas, Torres Vedras), O Arqueologo Portugues.

ARAUJO, A.C. and J. ZILHAO. In press. O Magdalenense Terminal de Vale da Mata (Cambelas, Torres Novas), O Arqueologo Portugues.

ARNAUD, J. 1986. Post-glacial adaptations in southern Portugal, World Archaeological Congress, Southampton, 1986. (Preprint).

1987. Os concheiros Mesoliticos dos Vales do Tagus e do Sado: Semelhancas e diferencas, Arqueologia 15: 53-64.

1990. The Mesolithic communities of the Sado Valley, Portugal, in their ecological setting, in C. Bonsall (ed.), The Mesolithic in Europe: 614-31. Edinburgh: John Donald.

ARNAUD, J. & J.D.B. BENTO. 1988. Caracterizacao da ocupacao pre-historica da Gruta do Cassal Papagaio (Fatima, Vila Nova de Ourem), Boletim da Sociedade Portuguesa de Espeleologia 2: 27-34.

AUDOUZE, F. 1987. The Paris basin in Magdalenian times, in O. Soffer (ed.), The Pleistocene Old World, Regional Perspectives: 183-200. New York (NY): Plenum Press.

BICHO, N. 1991. Areeiro III, an open air site dated to 8850 BP (Rio Maior, Portugal), Mesolithic Miscellany 12(2): 1-10.

1992a. Technological change in the final Upper Paleolithic of Rio Maior, Portuguese Estremadura. Unpublished Ph.D dissertation, Southern Methodist University, Dallas (TX).

1992b. Technological variability during the Pleistocene-Holocene boundary in the Portuguese Estremadura. Paper presented at the 57th annual meeting of the Society for American Archaeology, Pittsburgh, April 8-12.

In press a. Magdalenian flint technology at the site of Cabeco de Porto Marinho, Rio Maior, Portugal, in Proceedings of the VI Flint International Symposium, Madrid, Bilbao, Granada 1-5 October 1991.

In press b. The role of quartz and quartzite in the Magdaleninn of Cabeco de Porto Marinho, Rio Maior, Portugal, in N. Moloney, L. Raposo & M. Santonja (ed.), The role of quartzite and other non-flint raw materials in the Iberian Palaeolithic. Oxford: British Archaeological Reports. International series.

BINFORD, L. 1978. Nunamiut ethnoarchaeology. New York (NY): Academic Press.

1979. Organization and formation processes: looking at curated technologies, Journal of Anthropological Research 35(3): 255-73.

1980. Willow smoke and dogs' tails: hunter-gatherer settlement systems and archaeological site formation, American Antiquity 45: 4-20.

BREUIL, H. 1912. Les subdivisions du Paleolithique superieur et leur signification, 14th Congres International d'Anthropologie et Archeologie Prehistoriques, Geneve I: 165-238.

1918. Impressions de voyage paleolithique a Lisbonne, Terra Portuguesa III: 34-9.

BREUIL, H. & G. ZBYSZEWSKI. 1942. Contribution a l'etude des industries paleolithiques du Portugal et de leurs rapports avec de la geologie du Quaternaire, Comunicacoes dos Servicos Geologicos de Portugal 23.

1945. Contribution a l'etude des industries paleolithiques des plages quaternaires de l'Alentejo littoral, Comunicacoes dos Servicos Geologicos de Portugal 27.

BURDUKIEWICZ, J.M. 1986. The Late Pleistocene shouldered point assemblages in Western Europe. Leiden: E.J. Brill.

CAMPBELL, J.B. 1986. Hiatus and continuity in the British Upper Palaeolithic; a view from the Antipodes, in D.A. Roe (ed.), Studies in the Upper Palaeolithic of Britain and Northwest Europe: 7-42. Oxford: British Archaeological Reports. International series 296.

CARDOSO, J.L. 1992. Contribuicao para o conhecimento dos grandes mamiferos do Plistocenico Superior de Portugal. Unpublished Ph.D dissertation, Universidade Nova de Lisboa, Lisbon.

CARVALHO, E., L. STRAUS, B. VIERRA, J. ZILHAO, A.C. ARAUJO. 1989. More data for an archaeological map of the county of Torres Vedras, Arqueologia 19: 16-33. CHATTERS, J.C. 1987. Hunter-gatherer adaptations and assemblage structure, Journal of Anthropological Archaeology 6:3 36-75.

CHESTER, D.K. & P.A. JAMES. 1991. Holocene alluviation in the Algarve, southern Portugal: the case for an anthropogenic cause, Journal of Archaeological Science 18: 73-87.

CLARK, G. & L.G. STRAUS. 1986. Synthesis and conclusions Part I: Upper Palaeolithic and Mesolithic hunter-gatherer subsistance in northern Spain, in Straus & Clark (1986): 351-66.

CLIMAP. 1976. The surface of the Ice Age Earth, Science 191(4232): 1131-7.

DAVEAU, S. 1971. La glaciation de la Serra de Estrela, Finisterra 6(11): 5-40.

1973. Quelques exemples d'evolution quaternaire des versants au Portugal, Finisterra 8(15): 5-47.

1980. Espaco e Tempo. Evolucao do ambiente geografico de Portugal ao longo dos tempos pre-historicos, CLIO 2: 13-37.

1986. Les bassins de Lousa et d'Arganil. Recherches geomorphologiques et sedimentologiques sur le massif ancien et sa couverture a l'est de Coimbra. Lisboa: Centro de Estudos Geograficos. Memorias 8.

DAVEAU, S. A.R. PEREIRA & G. ZBYSZEWSKI. 1982. Datation au 14C du site archeologique de la Plage de Magoito (Portugal), Clio 4: 133-7.

DAVIDSON, I. 1983. Site variability and prehistoric economy in Levante, in G. Bailey (ed.), Hunter-gatherer economy in prehistory: a European perspective: 79-95. Cambridge: Cambridge University Press.

FIGUEIRAL, I. In press. Cabeco de Porto Marinho: une approche paleoecologique. Premiers resultats, in Actas da VIII reunion Nacional sobre Cuaternario, Valencia, September 1991.

GAUSSEN, J. 1980. Le Paleolithique superieur de plein air en Perigord: secteur Mussidan-Saint Astier, moyenne vallee de l'Isle. Paris: CNRS. Gallia Prehistoire supplement 14.

GRAYSON, T. 1986. Eoliths, archaeological ambiguity, and the generation of 'Middle-Range' research, in D.J. Meltzer, D.D. Fowler & J.A. Sabloff (ed.), American archaeology, past and future: 77-133. Washington (DC): Smithsonian Intitution Press.

HELENO, M. 1944. O problema Capsence; Contribuicao Portuguesa para a sua revisao. Comunicacao ao Instituto de Arqueologia na Sessao de Abril de 1944 (Lisboa).

1956. Um quarto de seculo de investigacao arqueologica, O Arqueologo Portugues II-III: 221-37.

JANSSEN, C. & R. WOLDRINGH. 1981. A preliminary radiocarbon dated pollen sequence from the Serra da Estrela, Portugal, Finisterra 16: 299-309.

MARKS, A.E. 1991. The Upper Palaeolithic at Cabeco do Porto Marinho in Portuguese Estremadura. Paper presented at the 12th Congress of UISPP, Bratislawa, September 1991.

MARKS, A.E., J. SHOKLER, & J. ZILHAO. 1990. Raw material usage in the Paleolithic: the effects of local availability on selection and economy. Paper presented at the International Symposium on Raw Material of Prehistoric Hunter-gatherers. University of Kansas, Lawrence, Kansas.

MARKS, A.E., J. ZILHAO, N. BICHO & R. FERRING. In press. Upper Palaeolithic prehistory in Portuguese Estremadura: preliminary results, Journal of Field Archaeology.

MATEUS, J.E. 1985. The coastal lagoon region near Carvalhal during the Holocene; some geomorphological aspects derived from a palaecological study at Lagoa Travessa, in Actas da I Reuniao do Quaternario Iberica 2: 237-49. Comissao Lisboa: Organizadora da I Reuniao do Quaternario Iberico.

MCINTYRE, A. & N. KIPP. 1976. Glacial North Atlantic 18,000 years ago: a CLIMAP reconstruction, Geological Society of America Memoir 145: 43-76.

MORALES, M.G. & J. ARNAUD. 1990. Recent research on the Mesolithic in the Iberian peninsula: problems and perspectives, in P.M. Vermeersch & P. Van Peer (ed.), Contributions to the Mesolithic in Europe: 451-62. Leuven University Press, Leuven.

POVOAS, L., J. ZILHAO, J. CHALINE & P. BRUNET-LECOMTE. 1992. La faune de rongeurs du Pleistocene Superieur de la Grotte de Caldeirao (Tomar, Portugal), Quaternaire 3(1): 40-47.

RAPOSO, L., C. PENALVA & J.P. PEREIRA. 1989. Noticia da descoberta da estacao Mirense de Palheiroes do Alegra, Cabo Sardao (Odemira, Portugal), in Actas da |2.sup.a~ Reunion del Cuaternario Iberico, Madrid September 1989: 25-29.

RIBEIRO, O., H. LAUTENSACH & S. DAVEAU. 1988. Geografia de Portugal II. O Ritmo Climatico e a Paisagem. Lisboa: Edicoes Sa da Costa.

ROCHE, J. 1964. Le Paleolithique superieur portugais. Bilan de nos connaissances et problemes, Bulletin de la Societe Prehistorique Francaise 31: 41-7. 1979. Le Magdalenien Portugais, in D. Sonneville-Bordes (ed.), La fin des temps glaciaires en Europe: 753-8. Paris: CNRS.

1982. A gruta chamada Lapa do Suao (Bombarral), Arqueologia 5: 5-18.

RUDDIMAN, W.F. & A. MCINTYRE. 1981. The north Atlantic Ocean during the last deglaciation, Palaeogeography, Palaeoclimatology, Palaeoecology 35: 145-214.

STRAUS, L.G. 1986a. Human adaptations across the Pleistocene-Holocene boundary in southwest Europe, in A. ApSimon & S. Joyce (ed.), The Pleistocene perspective: 2: 324-33. London: Allen & Unwin.

1986b. The end of the Palaeolithic in Cantabrian Spain and Gascony, in L.G. Straus (ed.), The end of the Palaeolithic in the Old World: 81-116. Oxford: British Archaeological Reports. International series 284.

1987. Hunting in late Upper Palaeolithic western Europe, in M. & D. Nitecki (ed.), The evolution of human hunting: 147-76.

STRAUS, L.G. & G.A. CLARK (ed.). 1986. La Riera Cave: Stone Age hunter-gatherer adaptations in Northern Spain. Tempe (AZ): Arizona State University. Anthropological Research Papers 36.

THOMAS, D.H. 1988. The archaeology of Monitor Valley 3: Survey and additional excavations New York (NY): Museum of Natural History. Anthropological papers 66 (Part 2).

1989. Diversity in hunter-gatherer cultural geography, in R.L. Leonard & G.T. Jones (ed.), Quantifying diversity in archaeology: 85-91. Cambridge: Cambridge University Press.

VAN DER BRINK, L. & C.R. JANSSEN. 1985. The effect of human activities during cultural phases on the development of montane vegetation in the Serra da Estrela, Portugal, Review of Paleobotany and Palynology 44: 139-215.

VIERRA, B. 1992. Subsistence diversification and the evolution of microlithic technologies: a study of the Portuguese Mesolithic. Unpublished Ph.D dissertation. University of New Mexico, Albuquerque (NM).

ZILHAO, J. 1985. Donnees nouvelles sur le Paleolithique superieur du Portugal, in Actas da I Reuniao do Quaternario Iberico, Lisboa 2: 101-12.

1987. O Solutrense da Estremadura portuguesa. Uma proposta de interpretacao paleoantropologica. Lisboa: Departamento de Arqueologia do Instituto Portugues do Patrimonio Cultural. Trabalhos de Arqueologia 4.

1988. The early Upper Palaeolithic of Portugal, in J.F. Hoffecker & C.A. Wolf (ed.), The early Upper Palaeolithic: evidence from Europe and the Near East: 135-55. Oxford: British Archaeological Reports. International series 437.

1990. The Portuguese Estremadura at 18,000 BP: the Solutrean, in O. Soffer (ed.), The World at 18,000 BP. High Altitudes: 109-25. London: Unwin Hyman.

1991. Le Solutreen du Portugal: environnement, chronologie, industries, peuplement, origines, in Les Industries a pointes foliacees du Paleolithique superieur europeen, Krakow 1989 (ERAUL 42): 485-501. Liege.

1992a. Gruta do Caldeirao, O Neolitico Antigo. Lisbon: Instituto do Patrimonio Arquitectonico e Arqueologico, Departamento de Arqueologia. Trabalhos de Arqueologia 6.

1992b Estrategias de Povoamento e Subsistencia no Paleolitico e no Mesolitico de Portugal, in Elefantes, ciervos y ovicaprinos: 149-62. Santander: Universidade de Cantabria.

In press a. The Upper Palaeolithic of Portugal: past research and current perspectives, in R.E. Webb (ed.), Recent research on the European Palaeolithic. In press b. As origens da Arqueologia Paleolitica em Portugal e a obra metodologicamente precursora de J.F. Nery Delgado, in Volume comemorativo dos 125 anos da Associacao dos Arqueologos Portugueses.

N.d. The use of non-flint raw materials in a Late Palaeolithic assemblage from Gruta do Caldeirao (Tomar, Portugal), in N. Maloney, L. Raposo & M. Santonja (ed.), The role of quartzite and other non-flint raw materials in the Iberian Palaeolithic. Oxford: British Archaeological Reports.

ZILHAO, J., E. CARVALHO & A.C. ARAUJO. 1987. A estacao Epipaleolitica da Ponta da Vigia, Arqueologia 16: 8-18.