THE PALAEOLITHIC COLONIZATION OF EUROPE : AN ARCHAEOLOGICAL AND BIOGEOGRAPHIC PERSPECTIVE

Current knowledge on the Lower Palaeolithic peopling of Europe is synthesized, using toolmaking repertoires, geochronological. biogeographical and palaegeographic evidence. The oldest traces belong to a non-Acheulian horizon, dating between 0.90 to 0.55 my. Tracing hominid geographic origin and dispersal routes into Europe identifies three alternatives: through the Levant corridor, across Gibraltar strait or more remotely, from a Central Asia filter route. No conclusive proof exists for any of these but there is more coherent empirical support for Central Asia where a non-Acheulian industry is also present. In this alterna ti ve, ancient hominid colonization of Europe was synchronous with a major Pleistocene mamalian turnover, the end-Villafranchian/Galerian dispersal event, originating in Central Asia. RESUMEN Sintetizamos el conocimiento actual sobre el poblamiento de Europa durante el Paleolítico inferior, empleando datos de repertorios líticos, así como evidencia geocronológica, biogeográfica y paleogeográfica. Los vestigios humanos más antiguos pertenecen a un horizonte no-Achelense con una edad entre 0.90 y 0.55 ma. Buscando el origen geográfico y los caminos de dispersión hasta Europa, identificamos tres alternativas: a través del corredor del Levante, del estrecho de Gibraltar O, más improbablemente, por un camino de tipo filtro desde Asia Central. No existen pruebas definitivas para ninguna de esas alternativas pero hay un apoyo empírico más coherente para la hipótesis de Asia Central. donde existe también un repertorio no-Achelense. Según esta alternativa, la colonización de Europa coincide con un importante cambio en la fauna pleistocénica, la dispersión del final del Villafranquiense/Galeriense, la cual también se originó en Asia Central.

Fossil hominids reach back to the Pliocene, nearly 4.0 my at the latest.They show the presence of several hominid lines: genus AUSlralopithecus, beginning with A. afarensis and its robUSlus and africanus successors; genus Horno with H. habilis evolving into ereclus by about 1.6 my.Only H. erectus expanded beyond Africa.
The sequential order of these two Lower Palaeolithic horizons in Subsaharan Africa, as successive technological developments, has important implications for discussions on the peopling of Eurasia: a) both originated in Africa prior to hominid dispersal; b) the Subsaharan African Acheulian itself precedes by a substantial amount of time any similar occurrences in North Africa and Eurasia.In lhe case of Europe, this may be by a coefficient of three (Alimen, 1977;Isaac, 1978).Given that the earliest known archaeological traces for a human presence beyond Africa cannot be much older than 1.25 my e.g. the Acheulian site of 'Ubeidiya, in Israel, as a datum line, any claims for nonhandaxes or so-called «Pebble-Culture» in Eurasia requires conclusive proofs and a coherent explanation.
Acheulian assemblages display considerable internal variability.Careful consideration must therefore be given to possibilities of sampling problems, before making a diagnosis.Consequently, the null hypothesis is that the Lower Palaeolithic repertoire possessed by the early hominids who colonized Eurasia was the Acheulian.
If Subsaharan Africa was the cradle of mankind, it is also there that the basic biocultural parameters that made hominid dispersal beyond that continent possible, such as bipedalism, toolmaking skills and a meat-eating propensity involving procurement strategies, were also developed (Bartholomew and Birdsell, 1957;Bunn, 1981;Gable, 1980).Issues about the respective roles of predation and scavenging are still debated (Binford, 1981;Bunn and Kroll, 1987;Blumenschine, 1987;Potts, 1984;Shipman, 1986) but there is no doubt that andent hominids integrated early within a complex web of more or less competitive or cooperative relationships with other carnivore species.
The palaeontological record points to a major faunal transformation at the beginning of the lower Pleistocene in Subsaharan Africa, by 1.6 my, with the appearance of modern e1ements (Maglio, 1975(Maglio, : 461-462, 1978: 612;: 612;Martin, 1986: 380) and the local replacement of archaic carnivores e.g.sabertooth cats, by spedes with more developed social capabilities e.g.spotted hyena, lion, leopard, cheetah (Turner, 1982(Turner, : 234-245, 1985)), a behavioural trait also found in hominids, and correlated with larger brains (Hemmer, 1982).Until this early Pleistocene turnover, hominids (H.habilis) had achieved a symbiotic scavenging relationship with sabertooth cats, in a closed habitat community (Marean, 1989).The tumover coincides with the onset of drier conditions and more open habitats (Kappelman, 1984;Bonnefille, 1979).Hominids (H.ereclus stage) then became involved in confrontational relationships with other carnivores.Their success in this modified relationship would apparently rank them high in the competitive hierarchy (Eaton, 1979: 16).
Tumer (1982,1985) argues that these hominid/ carnivores interactive networks constituted one major initial condition for their simultaneous expansion into Eurasia.
1) expansion beyond African boundaries to incorporate gradually aH of Eurasia more than doubled the realm already occupied by a single primate species.Subcontinental regions such as Europe, India, Southeast Asia or China may aH fit within the African continent (Clark, 1981: fig. 4) but the latter fits easily inside Eurasia; 2) moving from a tropical continent into a bioclimatically varied one required coping with increasingly unfamiliar conditions, as hominids began to settle middle and higher latitudes.The initial colonization of subtropical portions of the Palaearctic region and of the Oriental region tropics provided habitat continuity.Moving northwards into temperate Palaearctic zones, however, meant coming to terms with «time stressors,.(Torrence, 1983) such as sharply defined seasonal contrasts with long winters, protracted dark nights, reduced food biomass, at the same time as meeting higher calorie requirements.The basic physiological adaptations of andent hominids in the tropics (Montagu, 1964) did not predispose them for such a transition.
Nevertheless, their descendants were already leaving traces of their presence in mid-Iatitudes, from the North European plain to Korea by 200-300 ky.Their success in achieving this resulted from combining physiological, behavioural and technological traits.Artificial means, gradualIy mastered until post-Pleistocene times, included fire-making tecniques, clothing, constructed shelters, animal traction or snowshoes for transportation and specialized toolkits and weaponry; 3) this ongoins dispersal throughout huge areas during the Pleistocene, however, did not lead to speciation or adaptive radiation.This is due to one basic factor, shared with other larger carnivores, namely eurytopic and exogenous capabilities, with reliance on a carnivorous diet, enabling widespread geographical expansion without profound genotypic modifications or ecological specializations (Foley, 1987: 263-267).Several other African carnivore species (lion, leopard, spotted hyena) succeeded in dispersing throughout Eurasia but unlike hominids who consolidated their conquests of new environments, their realms were drastically reduced or they retreated into Africa after the Pleistocene; 4) an alternative consequence of expanding ranges and territories by hominids was that technoeconomic and social repertoires diversified and became more elaborate over time.This increasing scope for ethnic and cultural differentiation is shown by archaeological and ethnological evidence (Braidwood, 1960: fig. 1;Lévi-Strauss, 1958: 10) while peripheral gene flow maintained the human species' essential biological homogeneity.

The InUlal homlnld dispersa} Into Eurasla
The scenario outlined by Howell (1960: 225) for earliest hominid movements beyond Africa remains essentially plausible, though still largely circumstantial and supported by only scant direct evidence.It involved first the Levant, with the Acheulian of 'Ubeidiya, datable to 1.25 my (Techernov, 1987(Techernov, , 1988)), followed by expansion along lowland and less arid areas of Southwest Asia such as the Mediterranean zone, extending then further inland prior to mid-Pleistocene tectonic upheavals (Horowith, 1979: ch. 3), the rim of the Fertile Crescent with the Assyrian steppe, the Mesopotamian lowlands (much more extensive during glacio-eustatic episodes) and perhaps, the lowermost portions of the outer Taurus and Zagros piedmonts.This would then lead to movements along the southern edge of the Iranian platcau and its mountain barriers, although direct evidence for this again is absent, scarce and mostly undated (Singer and Wymer, 1978: 24;Smith, 1986: 16).
Evidence becomes abundant upon reaching the Indian subconlinent and roughly datable to mid-Pleistocene times, from associated faunal remains.AH of it is Acheulian, often accompanied by numerous pebble-tools.Older evidence is either eroded away or deeply buried (Misra, 1989).lt should be mentioned that the subcontinent was not part of the Oriental faunal regio n during the Pleistocene.Its fauna was highly endcmic, with added Palaearctic e1ements (Kretzoi, 1961-64;Azzaroli, 1986).
With Southeast Asia and the Oriental faunal region, we encounter a distinct Lower Palaeolithic repertoire, associated with Humo ereclus remains in Java and China, defined and labeHed the «Chopper-Chopping Tool Complex~ by Movius (1944Movius ( , 1948) ) for the entire Far East.Many have criticized sorne of Movius' conclusions (amorphous artifactual assemblages, technological stagnation) and terminology and even questioned its reality (Freeman, 1977;Boriskovskii, 1978;Chung, 1976Chung, , 1984;;Yi and Clark, 1983), partly owing to new discoveries of Acheulian-Iike implements such as handaxes and cleavers in China, Java and Korea.Movius failed to identify a cause for the origins of this East Asian Lower Palaeolithic complex (1978).
The reality of a Lower Palaeolithic repertoire dichotomy comprising a «westerm handaxe and cleaver technocomplex (Acheulian) and the non-handaxe East Asian complex as an issue, has considerable bearing on the problem of the peopling of Eurasia. A. Sieveking (1962) proposed an explanation, when reporting on the Kota Tampan si te in Malaya: the East Asian non-handaxe complex, renamed the «Pebble and Flake,.tradition, was the outcome of a rapid and widespread initial hominid dispersal from Africa into Asia, who manufactured an unspecialized «Pebble-Culture» industry, derived from the Oldowan, which could diffuse into varied environments, during the late lower Pleistocene.Conditions in the Indo-Malaysian tropical forests of Southeast Asia favoured simple technologies, and inhibited the subsequent spread of the more specialized Acheulian, developed under conditions of savanna or savanna-open woodlands.This interpretation has become redundant due partly to the unsatisfactory Kota Tampan evidence and superseded by other interpretations.
Sufficient reasons remain, nevertheIess, to retain the notion of a separate Lower Palaeolithic repertoire in East Asia (Ikawa-Smith, 1978;Aigner, 1978;Ayres and Rhee, 1984): new and dated discoveries in Java, Thailand and China; the inconclusive nature of so-called handaxe and cleaver finds in these regions.Most are atypical, very low in relative or absolute frequencies (this writer's data show that several single Acheulian sites in Western Europe or Africa contain more handaxes and cIeavers than the en tire collections from East Asia).Their dating is upper Pleistocene times, suggesting a self-contained development.
The most reasonable explanation for the appearance of this East Asian non-handaxe complex is that it represents an adaptive modification of the Acheulian when hominids colonized the dense Indo-Malaysian tropical forests of Southeast Asia.This induced a shift towards greater dependence on plant foods, solitary or smaHer game (Watanbe, 1985) and widespread exploitation of bamboo as an all-purpose raw material, requiring nothing more than a rudimentary lithic toolkit, in environments where supplies of cryptocristaline rocks are scarce (Hayden, 1978;Hutterer, 1978;Pope, 1984Pope, , 1988;;Solheim, 1970).
This modified repertoire, involving the abandonment of handaxe and cIeaver manufacture, persisted when the subtropical and cool temperate latitudes of China became settled, although the need for an expanded lithic equipment was resumed.This was due to geographic isolation and long distances from the nearest area of handaxe/ eleaver manufacture Le.India, as weH as continuing bamboo availability.Fossil and archaeological evidence for hominid presence in East Asia reaches back to over 1.0 my (Pope el al, 1987).In China, archaeological occurrences of lower Pleistocene age with local Villafranchian faunal associations in elude Xihoudou (Lanpo, 1980), Dunggutuo and Xiaochangliang, in the Nihewan basin (Lanpo and Qi, 1987;Qi, 1988), aH without handaxes or cleavers but sorne with heavy duty or large cutting tool equipment, as well as finely made flake implements.

Obstacles to dlrect or early movements lnto Europe:
Conditions for penetrating Europe during the Lower Palaeolithic were less favourable because more contingent on geographic factors than latcr on, more rudimentary technological means, and lower hominid population densities restricting habitat diversit~, and home ranges.Cyelical glacial episodes amplified or modified environmental barriers constraints.These inelude: 1) physical barriers such as a) major high mountains and plateaus ranges, stretching West to East across Eurasia, with the Taurus, Zagros, Armenian Knot, the Caucasus, Elburz, Kopet Dagh and especially, further East, the massive system linking more or less continuously the Paromisus, Hindu Kush, Pamirs, Karakorum, Tien Shan, Hymalaya, Kunlun, Altyn Tagh, Nan Shan and Qinling ranges.While not all entirely impenetrable, they inhibited early humans and their limited expansion means; b) aquatic barriers, with the Mediterranean in the West.Glacio-eustatic sea leveI fluctuations could improve direct crossing conditions from Africa into Europe, or around the Black Sea and across the Bosphorous (Chepalyga, 1984) but the Caspian Sea expanded considerably to the North, narrowing passage areas already reduced by glaciallake barriers in Westem Siberia (Alekseev et al, 1984: map 3); 2) climatic barriers relating to extended permafrost zones, modified colder air circulation pattems (Velichko, 1984: figs. 25.10, 25.15) during ice ages, constituted significant limitations for a tropical species.The 10° isothem of the coldest month of the year always remained a major impediment to human movements (Whiting el al., 1982); 3) ecological barriers, as hominids moved North of the 30-40° latitude in Eurasia, encountering cold boreal forests and steppes, taiga and cold deserts from West to East, and loess deposition belts restricting vegetation growth during glacials.The archaeological record shows that prehistoric communities coped with such unfamiliar and unattractive conditions gradually, e.g. the Middle Palaeolithic for steppic continental environments in Central Europe (Bosinski, 1982), the Siberian taiga during the Holocene.

Emlgration routes lnto Europe
Europe could be reached 1) direclly from the Maghreb through Gibraltar or across the strait separating Tunisia from Sicily.Sea level regressions may have created narrow landbridge strips or reduced water barriers (channels).Altematively, crossings may have been achieved at any time through artificial means (Alimen, 1975;Bordes and Thibault, 1977;Chard, 1963); 2) indirectIy, through the Levant and into Anatolia and Southeastem Europe.The Taurus/Zagros/ Armenian Knot, however, presented topographic and bioclimatic barriers for movements beyond which were overcome only later (Howell, 1960).Movements from the East, across the Kazakh and Russian steppes would imply a more precocious adaptation to cold, continental steppes than is evidenced directIy.We shall examine, further on, another possibility from the East, with a route from Central Asia through a filter corridor along the northem edge of the Iranian platea u, the Transcaucasus and northem Anatolia or, perhaps, the presentIy submerged northern rim of the Black Sea.
In summary, despite Europe's proximity to Africa, ancient hominids could have encountered difficulties in expanding there, whatever the route taken.

Age of earliest anthropic evldence in Europe
Viewpoints conceming human antiquity in Europe have varied and have opposed proponents of short and long chronologies.By the early 1960s, after the «eoliths» controversy and the Piltdown forgery had been settled, it was considered that a Mindel/EIster date was probably the oldest limit, based on the Abbeville (Porte du Sois) archaeological finds and the Mauer mandible (Howell, 1960).By 1966, new evidence (Vallonet cave, L'Escale, Fenne de Gr:'lce) required pushing this boundary further back i.e.Günz (Howell, 1966: 177-178;Bordes, 1968: 48-49).Other archaeological finds included the Chilhac finds, whose 1.7 my estimated date and provenience were controversial.
Recent claims for anthropic occurrences of Pliocene age from the Massif Central (Bonifay, 1989;Bonifay el al, 1989) have been received skeptically (Delson, 1989;Lumley el al, 1988: 591-596).The equally controversial cranium from Orce, Venta Micena, claimed to be hominid (Agustí et al., 1983) also appears questionable (Moyá Solá and Agustí, 1990).Despite conflicting claims or taphonomic arguments suggesting that the scarcity of anthropic finds in Europe is mure apparent than real and has more tu do with evidence los s than with later peopling (Dennell, 1983: 35-37;Bordes, 1968: 49-50), steadily increasing discoveries and better chronological resolution indicate an age around 0.90 my for earliest humans in Europe.
While there is growing acceptance that most of the earliest archaeological traces in Europe appear to be non-handaxe occurrences (Bordes and Thibault, 1977;Lumley el al, 1988;Pipemo el al, 1983), there is surprisingly little discussion about the implications of chronological overlap between handaxe and non-handaxe industries (see Collins, 1969 and comments).Issues inelude díscriminating between environmental and behavioural (function, tradition) factors possibly involved and the origins of non-handaxe repertoires.
The dominant view has been to regard older non-handaxe industries as part of an archaic and widespread «Pebble-Culture» honzon throughout Eurasia, derived directly from the Oldowan (Bordes, 1968: 134;Bordaz, 1970: 18-20;Howell, 1966: 178;Kretzoi and V értes, 1965;Rolland, 1977: 134;Sieveking, 1962).We shall review reasons why this needs revision.The question of discontinuous non-handaxe repertoires in different parts of Eurasia, establishing their chronological positions, are part of evidence and issues to be considered when assessing the origins and dispersal routes followed by hominids into Europe.

Was the colonlzatlon of Europe a separate event?
HoweU or Sieveking treated the peopling of Europe as a later event, because of environmental barriers.There has been since a tendency to lump it as part of the synchronous colonization of Eurasia, aroun~ 0.7 my or Brunhes/Matuyama geomagnetic boundary, proposed as a lower/middle Pleistocene boündary (Isaac, 1974(Isaac, : 512, 1978) )

Dlfficultles of synthesls
Reasons for a virtual lack of extensive, in-depth, and up to date syntheses about Lower Palaeolithic peopling of Europe include 1) its multidisciplinary scope which means keeping abreast of developments in a range of disciplines, partieularly palaeoanthropology, Palaeolithic archaeology and Quatemary research.It also involves gathering, sifting and evaluating a voluminous body of publications (journal papers, reports, chapters and monographs) in several languages, often difficuIt of access, 2) the fact that this published information constitutes only a portion of a larger body of research results, reaching professionals with a time lag.New empirical findings or new conceptual developments may also lead to complete revisions of accepted conclusions.Synthesis requires firsthand acquaintance with much of this primary evidence, published or not, to evaluate it critieally, 3) most of the archaeological information is the outcome of field researches designed with other objectives in mind.The considerable demands of work and time that fieldwork involves leave principal investigators with little additional time for large-scale syntheses.Few projects are explicitly devoted to the present topie, in strategically-defined regions, 4) diffieult resolution in topics relating to earlier Pleistocene and Palaeolithic horizons.Major information gaps persist concerning the archaeologieal record, chronology and regional coverages, despite a growing volume and quality of primary evidence.Generalizations remain provisional, leaving many questions unanswered.

Lines of evldence
The topie requires considering a wide array of evidence: anthropie (fossil hominid remains, Palaeolithic sites and remains), palaeobiological (animal palaeontology, palaeoecology), Quaternary research, and geochronologieal.The most crucial for our purpose relate to biogeography and Lower Palaeolithie repertoires, followed by palaeogeography and geochronology.
Biogeography.Historical biogeography or the study of dispersal events involving land mammals (Simpson and Beck, 1965: ch. 28) is especially relevant.Population movements of this kind are distinct from migration i.e. cyclical or seasonal and regulated movements within home ranges (1965( : 733, Rouse, 1986: 9): 9).Colonization is a kind of emigration or less regulated movement, with expansion into ranges hitherto devoid of members of the same species, and adaptive processes, over several generations.Causes such as demographie pressure or environmental change apply to human movements as well (Darlington, 1957: 621-646;Lack, 1970: 242-243).
The late lower Pleistocene colonization of Europe by one primate species emigrating from the Ethiopian into the Palaearctie regio n was an event whieh could be studied as part of a larger process including other animal species as prey or as competitors (Turner, 1982).
Biogeographers classify dispersal routes or «dispersal probabilities» as follows (Simpson, 1940(Simpson, , 1962;;Simpson and Beck, 1965): a) corridors where chances for large-scale spread-are high, such as the Eurasian corridor within the Palaearctie region, connecting Europe with China, or the Levant corridor between the Levant and Africa; b) ¡ilters allowing the spread between regions or continents NICOLAS ROLLAND of only eertain species or populations because of barriers' sdectivc effects e.g. the episodic erossings of the Sahara, along narrow stretches of oasis areas, during wetter palaeoclimatic phases, e) sweepslake routes where crossings occur during brief periods of opportunity and are very improbable for most elements beca use of effective barriers such as channels, straits, sea barriers.Only single emigrant groups sueceed e.g. the hominid peopling of Australasia during the upper Pleistoeene or perhaps, the crossing of Gibraltar at an carlier time.
Late Cenozoic mammalian dispersal events.Palaeontologists have identified several Plio-Pleistoeene events for the Holoarctic regions (Repenning, 1967;Lindsay el al., 1980).Major faunal turnovers such as during thc early Pleistoeene in Africa or late lower Pleistocene in Eurasia (Maglio, 1975) could have been related with early hominid dispersals.

Llnkages wlth hornlnld colonlzatlon events
We summarized earlier Tumer's hypothesis about sueh eonnections with respect to emigrations out of Africa.Dispersal events identified for the Plio-Pleistocene in Europe are summarized in table l.The most significant, as well as relevant to the Lower Palaeolithic peopling of Europe, is the end-Villafranchian/Galerian dispersal event (Guthrie, 1984) during later lower Pleistocene times.No attempts have been made so far to eonneet the initial hominid expansion into Europe with this event originating in Central Asia -although M. F. Bonifay (1983: 2) notes a significant hominid presenee then.
Most pre-existing Eurasian camivores (Homotherium, Magantereon, Felix toscana, Hyaena brevirostris) were gradually replaced by incoming more social African carnivores and the wolf, from the Near East.The cave líon of Europe assumed giant proportions (Kurtén, 1968).
This major faunal transformation happened rapidly, coinciding in time with the Jaramillo normal event, between 1.0 and 0.9 my and into the late Matuyama reversed epoch (Azzarolí, 1983).It began in Central Asia and spread throughout the Palaearctic region.Villafranchian holdovers lingered longer in Western Europe.Its causes are not related to connections by intercontinental landbridges but with protracted environmental changes, beginning with major tectonic upheavals in Central and South Asia (Azzaroli, 1983: 135;Azzaroli and Napoleone, 1982), then with significant marine and continental climatic shifts, reflected in changes of atmospheric circulation patlerns and more seasonal vegetation pattems.
Lower Palaeolithic repertoires.Subsuming Palaeolithic archaeology within palaeontology (Turner, 1982: 236) would add a new, meaningful dimension of hominid behaviour studies but entails the risk, if camed out unilaterally, of blurring specifie and significant information contents within toolmaking repertoires which may provide evidence no les s decisive than socio-ecological insights about issues such as dispersa!routes.The artifactual composition and time-space distributions of Lower Palaeoliihic complexes vary non-randomly.It would be shortsighted to ignore this.These inter-assemblage variations can be sorted descriptively under two major assemblage types: 1) handaxel cleavers assemblage type.These all refer to the Acheulian technocomplex.The   et al. 1988et al. , Bonifay 1979et al. , 1980et al. , 1983)  nomcnclature has becn simplified.It is found throughout Africa, Westcrn Eurasia and the Indian subcontinent.Its most specialized and diagnostic dements are the large cutting tools known as handaxes and eleavers but Acheulian assemblages display a marked degree of polythetic variability in tenns of modal frequencies, ranges and dominance patterns, not confined to large cutting tools and applying to other implement fonns such as pebble-tools, scrapers, denticulates, note hes.The term «Abbevillian-, while still used, should be abandoned.Its distinetiveness has been overstated.The label has eonnotations of an archaic handaxe manufacture stage, at variance with its contents.It is comparatively late, when compared with the Acheulian in Africa.
2) non-handaxe assemblage types.Their apparent simplicity results more from generalized and technological1y restricted characteristics.These assemblages constitute a heterogenous group, rather than a single complex such as the so-called «Pebble-Culturc)l.Terminologies also remain unsatisfaetory.They inelude the Oldowan, the «Pebble-Culturc)l of North Africa, the non-handaxe Lower Palaeolithic of Central Asia, the cePre-Acheulian)l of Europe and the Chopper-Chopping TooI complex of East Asia.
The Oldowan is better defined.The «Pebble-Culture- (Biberson, 1961 b) would be a parallel or phyletically related to it but its reality is now in doubt.The East Asian Complex, originally thought •to be derived from the Oldowan, is better seen as a perrnanently modilied version -but not a facies or atypical variant-of the Acheulian.80th non-handaxe assemblage types from Central Asia and Europe require an interpretation and appropriate nomenc\atures.The Clactonian and other later mid-Pleistocene industries of Europe are left out.Their age makes them less relevant to questions relating to the peopling of Europe.
Geochronology.It remains essential to have a degree of control over the time coordinate, to establish the relative anteriority, posteriority or penecontemporaneity between Lower Palaeolithic complexes, mammalian dispersal events, as well as palaeoclimatic and palaeogeographic circumstances.Predsion with respect to earlier Pleistocene horizons remains unavoidably broad.It could not allow detailed plottings of time-space distributions documenting exactly dispersion rates of human or mammal species, as for instance with the spread of Neolithic economies into Europe.At any event, certain expansion events were so rapid that they would escape detection on a geological tirne-scale, e.g. at rates of 1000 km per century (Kurtén, 1957).
Palaeogeography.These have already been mentioned: topographic and water barriers; latitudinal and ecological barriers; palaeoclimatic changes influencing positively or restricting animal or human movements.These factors would al so interplay in a complex fashion with andent hominid demographic parameters (local group or regional densities, reproductive networks) (see Dennell, 1983: 37-38;Wobst, 1976).Unfortunately, it is difficult to make these important processes operational.

THE EARLIEST LOWER PALAEOLITHIC HORIZON IN EUROPE
After outlining main sources of evidence, we survey the anthropic record for the initial occupation horizon of Europe ie.a time-span from late lower to early mid-Pleistocene times or 0.90 to 0.55 my, to establish its characteristics and patterns, and whether it contains archaeological indications about the origins and directions of hominid movement into Europe.One issue is whether this early Lower Palaeolithic horizon is a handaxe (or Acheulian) or non-handaxe repertoire and what are the implieations of this diagnosis.

General condltlons
1) the quality of this evidence varies in terms of sampling sizes, dating prccision and associated remains.The list of 68 occurrences (66 archaeological) ranges from findspots with single or few artifacts to rich collections in primary archaeological contexts.Sorne may find the coverage too generous; others, too restrictive.Virtually all occurrences, however, mee! minimum requirements of provenience control within a broad geological framework, artifact identification and chronology, and exclude controversial finds e.g.Massif Central.Venta Micena, Rumania and Macedonia; 2) a large portion of the si tes listed were from chance or incidental discoveries, or during geological or palaeontological investigations, without further fieldwork.Much of the information remains therefore provisional.Nevertheless, this cvidence is largely new and in need of synthesis.Systematic, archaeologically oriented excavations are increasing as discovery prospects are rewarded; 3) the paucity, variable quality and often rudimentary information content of early hominid traces in Europe has much to do with preservation or visibility conditions, which are less favourable in mid-latitudes, due to geomorphological changes related to glacial/interglacial cycles, as well as to the effects of millenia of intense agro-pastoralist exploitation, than in privileged regions such as East or Southem Africa (Bordes, 1968: 49-50;Dennell, 1983: 35, 37).Discoveries of the last 15-20 years, however, show that sorne evidence has survived and that certain regions' conditions approach those of Africa.

Presentation of evidence
A detailed discussion of the archaeological record is beyond this paper's scope.Coverage is limited to main features of the evidence, partly summarized in table 2, and identifying issues.
High altitude and a continental climate made the Massif Central a region of only episodic Palaeolithic occupation which was probably marginal during the earlier Pleistocene.Discovery pace is slowed down by scarce exposures or stratigraphic «windows», due to a moist climate and lush vegetation.
It is worth noting that siles are more numerous and oflen richer in lhe Mcditerranean or lemperale mid-Ialilude areas of Europe lhan further North, whieh may inform about ancient hominid adaptive eapabilities.

Llthlc assemblage characterlstlcs
Study melhods.No uniformally agreed methods or descriptive norms are available for the early Palaeolithic of Europe, although sorne degree of consensus exists.This is part of a wider problem concerning Lower Palaeolithic artifact studies, excluding analyses of more standardized implement forms sueh as handaxes or cleavers (Bordes, 1961;Isaac, 1977;Roe, 1964;Tixier, 1957).Pebble-tool categories of Africa (Ramendo, 1963) or Europe (Carbonell et al, 1984;Collina-Girard, 1978) are also better described.The main stumbling block concerns flake implements (Movius, 1949;Bordes, 1971;Isaac, 1977), because of low specification or attribute cohesion degrees.This ((amorphousness», however, has been overstated (see Freeman, 1978 for Lower Palaeolithic China), and does not apply to the flint artifacts of Visogliano or Venosa-Loreto.Attribute analysis, e.g. the Buda industry analysis (Vertes, 1962), or lithic reduction sequence analysis, e.g. the study ~f the East African Oldowan (Toth, 1985), are future directions to follow.The best example in our site list is in the El Aculadero monograph (Querol and Santonja, 1983a).
Raw material properties.Their shapes, sizes or textures can be determinant, given the limited range of flaking techniques available during the Lower Palaeolithic.Raw material exploitation varied according to local circumstances, aIthough there are few tabulated comparisons or sourcing studies: flint, shale, obsidian, quartzite, quartz, phtanite, andesite, basalt, gneiss, hornstone, limes tone, lydite.Certain sites display considerable heterogeneity e.g.Korolevo, where black andesite was common (Adamenko el al, 1989: 18).
Local materials were usually exploited aIthough flint (Soleihac), quartz or quartzite (Stráská Skála) could be imported.Flint quality could be good (La Pi neta, Visogliano, Trzebnica) or poor (Ca' Belvedere).Quartz from volcanic regions Le.Massif Central.could make artifact identification problematic.Less tractable materials such as Iydite (Prezletice), hornstone (Stránská Skála) or limestone (Visogliano, Colle Marino, Fontana Liri), gave artifacts an amorphous appearance.Limestone flakes still retain recognizable flaking attributes (butts, points of impact, percussion bulb) but their porous texture made them liable to geochemical disintegration, unless buried by sediments (Colle Marino).This local material was frequently used for making pebble-tools.
Lithic reduction sequence.Most occurrences contain variable frequencies of artifact classes diagnostic of successive lithic reduction stages: manuports, nodules, roughouts, discards, flaking debris, waste flakes, preforms, shaped tools.Certain sites (El Aculadero, Pointe aux Oies, Bibbona, Néa Skála) contain mostly preliminary stages attifacts.Primary f1aking is unspecialized, more or less exhaustive, with nodular pieces bearing unifacial or bifacial scar patterns.Small flakes were often obtained by shattering tabular f1int and used the fragments as cores, at La Pineta.Flakes are usually thick, with plain butts and wide percussion angles, or otherwise cortical.Trimming is al so unspecialized, applied unilateralIy or multilaterally, less often bifacially.Working edges patterns show scrapingl cutting, serrated or notched attributes.Flint assemblages from Visogliano or Venosa-Loreto, however, show greater standardization and, perhaps, resharpening.
Early Lower Palaeolithic assemblages of Europe show internal and inter-assemblage variability analogous to the polythetic patterns of non-handaxe/eleaver components of the Acheulian.Pebbletools often dominate but many assemblages show flake tools dominance (La Pineta, Visogliano, V enosa-Loreto, Trzebnica).

Chronology
A detailed discussion of the European early Palaeolithic chronology, while desirable, would exceed the paper's scope.We concentrate instead on summarizing the main findings.
Certain regions (Massif Central, Central and Southern Italy, parts of Central Europe) benefit from the converging contributions of several methods.Others depend on lithostratigraphy alone, such as the provisionally dated findspots from the Aquitaine.The Lower Palaeolithic sensu lato from the Spanish Meseta, on the other hand, retain the advantage of a broadly reliable relative chronology, based on careful, detailed reconstructions of terrace sequences, incorporating tectonic factors and avoiding palaeoclimatic presuppositions or correlations with the Alpine sequence (Delgado, 1977;Pérez González, 1983 a, b;Santonja and Villa, 1990).Several of the early non-handaxe occurrences come from higher (and older) terraces, while the Acheulian is found in younger ones.
Results.Table 3 lists the better dated occurrences within a broad time framework.Section A contains those dated with several methods, ineluding age calibrations (radiometric and/or geomagnetic), whereas section B provides for time-placements with wider margins of imprecision.All fit within a timespan beginning from the Jaramillo normal polarity event (or 0.90 my), through late Matuyarna reversed polarity and up to early Brunhes normal polarity epoch (until 0.60-0.55my).We cornment briefly on the most important results: 1) Vallonel cave.Dated by coastal lithostratigraphy, biostratigraphy (microtine and an end-Villafranchian/early Galerian faunas), geomagnetism and ESR dating (Chaline, 1988;Lurnley, 1976;Lumley el al, 1988;Yokoyama el al, 1988).This evidence places the artifact-bearing layers B, B2, C during late Jaramillo normal polarity event, around 0.89-0.91my and correlated with lower biozone 20 (Guérin el al, 1983: table 1).
2) Soleihac.The anthropic layers (units B, C, D) belong to a fIuvio-lacustrine series deposited during a normal polarity phase, initially placed at the beginning of the Brunhes epoch (Bout and Goi!r de Heme, 1976), but since found to be interpolated between the E and A reversed polarity  3) Isernw.La Pineta.Evidencc appearcd conflicting, at first.The microtine and large mammal fauna is exclusively Galerian (Sala, 1984), without Villafranchian holdovers, fitting within Cordy's (1982) zone V and Guérin's zone 22.The prcsence of D. hemitoechus and Arvicola would be compatible with a 380-480 ky dating (Cordy, 1982: tables 1, 2).Calibrations, however, show reversed polarity (McPherron and Schmidt, 1983), and repeated K/ Ar dating, using biotite and then sanidine sampling materials, gives consistently an age over 0.73 my, in agreement with geomagnetism (Sevink et aL, 1981; Delitala el aL, 1983).The possibility that the tuff deposits containing the samples could have been redeposited (Segrc, 1982: 587) seems ruled out by laboratory verifications.This dating, if accepted, implies that thc lower time boundaries of ccrtain early mid-Pleistoccne mammals found in c1assic localities, e.g.Stránská Skála, Süsscnborn, Voigstedt, must be pushed back (Azzaroli el aL, 1987: 80).An earlier than expectcd presence of A rvicola in the Levant ('Ubeidiya site, Israel) suggests a precocious and extra-European origin (Tchernov, 1988b: 855), and is more compatible with a long chronology.
4) Olher localities in Central and Southern Italy.These provide further valuable chronological indications for the early Lower Palaeolithie in Europe, thanks to well-established and calibrated, correlated sequences, from fluvio-Iacustrine series and tephrochronological datum Iines (Biddittu, 1984;Biddittu and Segre, 1982;Biddittu el al, 1979;Segre, 1978Segre, , 1984;;Segre el al, 1984).Crossdated Latium occurrences show that Colle Marino, Fontana Liri and Arce predate the Latium volcanic eruptions (K/ Ar dated at 706 ky), while Castro dei Volsci, Valchetta Cartoni postdate it slightly (700 ky).Further South, the stratigraphic and geomagnetic position (Balssas, 1980) of the archaeological layers A and B of Venosa-Loreto would be later than the K/ Ar dated (811 to 830 ky) Vulture eruption.Combining this with the early Galerian megafauna could place layers A and B around 600 ky (Angelelli el al, 1978; Biddittu et al, 1979: 67).Ihe in situ Irsina artifact find predates stratigraphically the Vulture eruption.
Summary.Relative chronologies and calibrations concur in showing that the oldest anthropic remains of Europe, with a cluster of dates from across Europe up to the Transcausus, document a hominid peopling of Europe during later lower Pleistocene times, instead of the mid-Pleistocene or early Brunhes datum line proposed previously.Ihis nevertheless confirms that no human presence can be traced to Villafranchian times sensu slriclo, that is before the Jaramillo event -the 1.7 my date postulated for the Chilhac artifacts is no longer tenable.The fauna used for dating is not assodated with them (Iexier, 1985).The synchronous appearance in Europe of both the first hominids and the Galerian fauna is potentially significant.It may merely be a case of chronological coinddence or prove to be a significant linkage.
The colonization of Europe apparently took place during moderate c1imatic conditions, equivalent with marine isotopic stage 23 and the Bavelian North Sea basin phase, and with Cordy's biozone I and Guérin's biozone 20.It was followed by the first European cold episode (isotopie stage 22) of late Matuyama, correlated with loess cycle K in Central Europe and datable to 782 ky (Bonifay, 1980: figs.1,2,3; Kukla, 1987: 42).

Diagnosis
The object is to establish the identity of the early Lower Palaeolithic of Europe: whether these occurrences belong to handaxe (Acheulian) or non-handaxe repertoires.A diagnosis may indicate a geographic origin and ancient hominid dispersal routes into Europe.Ihese depended on palaeogeographic factors but archaeological diagnosis may point to specific alternatives.
Since the Acheulian is already present in Subsaharan Africa by 1.5 my, one assumes that this repertoire probably accompanied andent humans when they colonized Western Eurasia, 200 or 300 ky and Europe, 900 ky latero Diagnosis of Lower Palaeolithic occurrences beyond Subsaharan Africa therefore treat an Acheulian wrdict as the null hypothesis, or that mode 2 (handaxe/cleaver) assemblage types constitute the oldest Palaeolithic horizon throughout Eurasia.The burden of proof rests on the alternative verdict: establishing the existence of a non-handaxe entity will requirc unequivocal empirical support and a coherent explanation.
The issue goes beyond a fonnal or contrived exercise, generated by an arbitrary dichotomy reminiscent of type-Fossil naiveties.A populational and polythetic approach to lithic assemblage analysis should not detract from recognizing technological ~mutations» (Harrison, 1954) e.g.mode 1 to 2 and growing specialization in Palaeolithic repertoires.
Published descriptions -a long with personal examination of much of the West European portion of the documents-of the early Lower Palaeolithic in Europe show a complete laek of handaxes and cleavers, the Porte du Bois locality excepted.This rcmains insuffieient for a fonnal diagnosis -though, many specialists reach a si miliar conclusion e.g.Bordes and Thibault, 1977;Santonja, 1983;Peretto and Piperno, 1985;Lumley el al, 1988;Segre el al., 1982-until submitted to stringent criteria and alternative interpretations.
Even accepting the Porte du Bois or Kent's Cavern dating (perhaps 600-700 ky) and a coexistence with the later portion of the non-handaxe horizon, would still leave the remainder without handaxe evidence.
. Dating the initial Acheulian in Europe remains unsettled.The earliest date is from Fontana Ranuccio, at 458 ky (Segre and Ascenzi, 1984), much later th;;tn the Colle Marino non-Acheulian (700 ky), also from the Anagni basin.The rich Pinedo site, in the Spanish Meseta, is associated with a fully Galerian fauna but without radiometric or geomagnetie dates.In Central Europe, Series II of the Staré Mesto locality, undated but perhaps of «inter-Mindel» age, contains crude-looking bifacial artifacts, at least one of which could be an atypical quartz handaxe (fig.14, Chlachula, 1990Chlachula, , 1991) ) (3) The underlying Series 1, without such implements, is late Matuyama.
2) Sampling arguments.These could inelude: a) taphonomic factors, such as negative consequences of repeated erosion cydes on archaeological preservation.This undoubtedly impoverished the record but the late Matuyama to early Brunhes evidence has yielded enough in situ and rich assemblages to suggest that at least a few handaxe finds should have t urned up by now; b) Slali5ticaL Handaxe/ cleaver frequencies range from nil to high values within Acheulian asscmblages but these implcments can be found in numerically smaU assemblages.In fact, our data fail to show any covariations between assemblage sizes and handaxe/ c1eaver frequency variations.Single finds of these implements are al so common in Africa or Eurasia.Several early non-Acheulian assemblages contain hundreds and thousands of artifacts but handaxes/ cleavers still remain absent (El Aculadero, La Pineta, Ca' Belvedere, Venosa-Loreto).The steady accumulation of these occurrences throughout Europe suggests increasingly that handaxe manufacture was not taking place then; c) activity facies couId introduce sampling biases because butchering or lithic quarrying did not r~quire using handaxes or cleavers.Infonnation about techno-economic organization during the early Palaeolithic horizon in Europe is poor.Where these activities are identified, large non-handaxe shaped implements are present (El Aculadero, Pointe aux Oies, Bibbona, CaseUa di Maida) but in similar Acheulian occurrence types (Aridos, Cagny-Ia-Garenne, Markkleeburg), handaxes (or bifacial trimmers) are found.
3) Raw nwterial properties.Texture, shape and sizes of lithic resources could restrict handaxe manufacture.Coarse-grained materials such as lydite, quartz, are untractable and if used, would produce atypical implements.The raw material diversity encountered throughout Europe for the early Palaeolithic horizon, however, transcends such contingencies or circumstances.Furthermore, Fontana Ranuccio contains a typical limestone handaxe and many Acheulian, particularly cave (La Chaise, Pech de l'Aze, Combe-Grenal) sites, recognizable handaxes made of poor quality materials.
The shape e.g.cobbles or pebbles, did not impede handaxe/ cleaver manufacture (Spanish Meseta, Upper Garonne, Central Italy).By contrast, the earlier La Pineta contained suitably-shaped tabular flint in addition to limestone, yet without any traces of handaxe-making.Small nodules or pebbles could impede effectively the manufacture of such implements, as was probably the case at V értesszóllas or Bilzingsleben, but experiments at El Aculadero establish that handaxes could be made from medium and moderately small pebbles, if intended.Many early non-Acheulian sites contain heavy-duty tools which could have been shaped into handaxes.
In summary, the early Palaeolithic horizon contains too many instances where raw material conditions aUowed a mode 2 technology.

ImpllcatloDs
The most economical verdict would seem that the existence of a non-handaxe horizon in Europe is more apparent than real and that such occurrences represent a facies or atypical variant of the Acheulian technocomplex.Nevertheless, we observe that a) no demonstrable chronological overlap exists between non-handaxe occurrences and the earliest Acheulian and b) sampling biases or influences of discrete drcumstance must be ruled out because this early horizon shows too much internal and contextual variability, as well as geographical dispersion across Europe.
The weight of evidence indicates rather that it constitutes a separate, polythetic technocomplex.The interrelatiQn of morphological traits (mode 1 technology), non-random chronological (late Matuyama to early Brunhes) and geographical (Europe) dumping authorizes rejecting the null hypothesis that this entity fits within the range of variation of the Acheulian technocomplex, beca use mode 2 technology is consistently absent.It suggests that andent hominids occupying Naming this entity «Pre-Acheulian" (Biddittu, 1984: 35), «Archaic Palaeolithic» (Santonja, 1982;Segre et al, 1982: 199), «Pebble-Culture,.(Kretzoi and V értes, 1965) or early «Clactonian» (Collins, 1978: 55), raises interpretative issues: a) «Pre-Acheuliam., while culture-stratigraphically correct inside Europe, remains relative and provisional, since the Acheulian outside Europe is as old or older; b) «Archaic •• or «Pebble-Culture,., beyond a descriptive usage, implies linear affinities with the chronologically distant Oldowan; c) «Clactoniam.may bestow an exclusively European identity but Clactonian sensu strieto is confined to scarce fully mid-Pleistocene occurrences in England, while existing analytical methods cannot discriminate techno-typological differences between it and the Oldowan.For these reasons, we shall refer henceforth to this technocomplex as the European Early Palaeolithie Horizon, provisionally.
We face the following issues: 1) finding its origin ur derivation and 2) how this may indicate hominid penetration routes into Europe.
The idiosyncrasies of the European Early Palaeolithic Horizon could be derived internally or outside Europe.The first altemative means that the Acheulian repertoire was modified to meet environmental circumstances within Europe, analogous to those invoked for East Asia The horizon's time duration is substantial and contains several cyclical palaeoclimatic cycles but this is not correlated with concommitant changes in toolmaking repertoires.It should be added that the European Pleistocene environment retained the kind of vegetation mosaic that was compatible with andent hominid food-getting strategies involving camivorous exploitation and Iithic technology, developed in Africa, maintained in Eurasia, except for Southeast Asia.
It seems reasonable to conclude that the characteristics of this technocomplex were derived outside Europe.This happened either by retaining, under unspecified isolating conditions, an Oldowan-like or mode 1 technology, or by reduction-segregation of the Acheulian repertoire.

ANCIENT HOMINlD DISPERSAL ROUTES INTO EUROPE
This section tries to identify colonization routes into Europe by examining: 1) three independent and possibly concurring Iines of evidence, namely locating non-Acheulian antecedent repertoires outside Europe; mammalian dispersal events, i.e.African camivore expansion into Eurasia and Galerian dispersal; palaeogeographic conditions delimiting population movements; 2) three a1ternative hominid expansion routes: the Levant corridor, a conspicuous starting point for movements into Southeastern Europe; the oft-mentioned crossing routes across Gibraltar or the Siculo-Tunisian strait to Iberia or Italy; another, more roundabout, hitherto Iittle considered, from a Central Asia staging area, through a filter route along the Palaearctic region's southern margino Southwest Asia: the Levantlne corrldor This constitutes the most economical hypothesis.The region is a natural crossroad between Africa and Asia, with converging vegetation belts (Horowitz, 1979: ch. 2;Zohary, 1962), and a logical bottleneck for faunal exchanges with Northeast Africa, with dry land intercontinental connection (Bar-Yosef, 1987;Tchemov, 1987).It offers a gradual transition from tropical to subtropical lones.Acheulian occurrences from South to North (Goren, 1981;Hours, 1975;Muhesen, 1981;Yalcinkaya, 1981), are concentrated along fossil drainage systems (Gilead, 1975: figs. 1,2) in grassland habitats with gregarious game, reminiscent of African landscapes familiar to andent hominids and carnivores.How the Levant was settled remains unclear and beyond the scope of this paper but evidence now favours a lower Pleistocene Bab al Mandib landbridge crossing from the Hom into the Western Arabian peninsula.1. Lower Palaeolilhic evidence.Most of it is recent Acheulian.The issue is to ascertain whether non-Aeheulian occurrences exist in the region, sufficiently early to be geographical anteeedents for their European eounterparts.The record, though slender, is adequate for a verdict.lt includes findspots, with or without handaxes in Lebanon (Borj Qinnarit), Northem Syria (Sitt Markho, Jebel Idriss, Sheikh Muhamad, Sharia, Khattab, Muharde 11: Besan<;on el al, 1977;Hours, 1981;Hours andSanlaville, 1972: Muhesen, 1981).'Ubeidiya, Central Jordan Valley, Israel, remains the most informative site (Bar-Yosef. 1975, 1988).
Daling.AlI oceurrences date to lower Pleistocene phases and roughly peneeontemporaneous.Dating, 'Ubeidiya excepted, relies on lithostratigraphy and seant fauna.'Ubeidiya's age, though better substantiated, has been debated.Preliminary faunal analysis and K/ Ar dates indicated around 700 ky (Bar-Yosef and Tchemov, 1972;Haas, 1973;Siedner and Horowitz, 1973), a datum tine for the peopting of Eurasia.New findings push this further baek in time: a) a faunallist revision led Fejfar and Repenning (1982) to a 2.5-1.9 my dating, implying also an extra-African origin for Horno ereelUS and the Acheulian, a view challenged (Jaeger, 1983;Opdyke el al, 1983); b) definitive fauna analyses (Eisenmann el al, 1983;Tchernov, 1987Tchernov, , 1988) ) indicate rather a 1.4 to 1.0 my age; c) geomagnetic and new K/ Ar datings of the Yannuk and Naharayim basalts, overlying the 'Ubeidiya fonnation, show 850 and 840 ky and reversed polarity, thus a post-Olduvai and pre-Jaramillo age, r.25 my being a reasonable approximation.
The early Lower Palaeolithic from the Levant, 'Ubeidiya particularly, documents the oldest anthropic traces in Eurasia.Its Acheulian identity, however, rules it out as an antecedent for the European Early Palaeolithic Horizon.
3. Palaeogeography.Hominids moving beyond the Levant had to negotiate physico-ecological barriers (Howel1, 1960), with a string of mountain ranges (Kurd, Amanus, Anti -and Central Taurus), leading Northeast to the Annenian Knot, a high elevation barrier with a harsh, cold climate ( Fisher, 1963: 330-332).Several of these mountains were not impenetrable e.g.thc Cilidan Gate, but piedmont areas oceupation remained apparently marginal for small, disperseo ano technologically Iimited early hominids, until they developed a transhumant settlement systcm, adapted to altitudinal biozonation (Smith, 1986: 10), later in the Pleistocene.Reaching Europe by crossing the Anatolian plateau meant settling in an arid, low primary productivity regio n -exeept for inland depressions-then moving along a more favourable northem Anatolian tectonic trough and valley system (Fisher, 1963: 316-317;Redman, 1978: fig. 2.5).Lower Palaeolithic evidence on the plateau coincides with later Aeheulian findspots (Minzoni-Deroche and Sanlaville, 1988;Yalcinkaya, 1988).Circumventing the central platea u by moving Northeast meant entering the inhospitable Eastern Anatolian highlands.Following the narrow Mediterranean coastal strip up to the Antalya Plain would be impeded further North or West, by another isolating barrier, the Western Taurus (Fisher, 1963: 322) whose role is eonfirmed by the area's Quatemary fauna endemism.
In summary, the Levant corridor seems a natural eolonization route toward Europe (e.g.Foley, 1987: fig.10.1), but closer scrutiny of the evidence does not support this prediction: significant physico-ecological barrier, delaying human implantation; the early Acheulian, an unlikely antecedent for the Early Palaeolithic Horizon in Europe.

The Ibero-Moroccan and Slculo-Tunlslan stralts
We review possible hominid dispersal routes aeross the West Mediterranean straits, with emphasis on Gibraltar.
Alleged Pleistocene connections between Africa and Europe through Mediterranean islands landbridges, Sicily and Tunisia particularly, were refuted by studies of insular dwarf elephant species (Vaufrey, 1929), amplified by other findings on dwarfism and gigantism (Thaler, 1973).
The rationale for direct connections between the Maghreb and Iberia or Italy has rested on arguments of Lower Palaeolithic assemblage similarities.This distributional and geographic contiguity criterion brings up the question of Pleistocene dry land or water crossing, especially for Gibraltar, revived by recent bathymetric findings.Biogeography, by contrast, seems of marginal or negative significance (Agustí el al, 1983: 20).
Physical barriers.The narrowest water barrier (15 km) between North Africa and Europe is a stretch from Tarifa to Punta del Carnero, Spain, and from Alcazarseguer to Punta Leon, Moroeco. Alimen (1975) estimates, from bathymetric studies and sea-floor mapping by echo soundings, submarine surface photos and deep-sea eores (Stanley, 1972), that dry land crossing became feasible when seal levels dropped to -290 m, creating a South-North narrow (23 km) isthmus from East of Tangier and West of Tarifa.Current action has prevented Pleistoeene sea-floor sediment aecumulation (Ketting and Stanley, 1972: 494-496) and there is no evidence for significant neotectonic uplift.
Glacio-eustatic lowering favouring dry land or water erossings of narrow channels, however, existed only during Riss glacial much later than the hominid peopling of Europe.
Water crossing at any time (Bordes and Thibault, 1977;Chard, 1963), on the other hand, implied watercraft navigation, difficult to document directly, and did not eliminate the problem of strong current velocity, increasing with narrow channels.Findings also indicate intense pelagic transport during glacio-eustatic phases (Huang and Stanley, 1972: 552).Circumstantial support for movements across Gibraltar is therefore inconclusive.
The Levant and Maghreb thus document a simultaneous appearanee of hominids and C. crocula beyond Subsaharan Africa, before or after 1.0 my, and perhaps the Hon and leopard in the Maghreb in North Africa by 700 ky.This evidence is too slender to show felid or hominid dispersal routes into Europe but a Gibraltar crossing is not entirely ruled out (Geraads, 1980), by a «sweepstake» dispersal movement (Agustí el al, 1983: 20).North African gerbils and amphibians are reported from Cueva del Higuerón and Las Yedras, Southem Spain (López Martínez, 1972;López Martínez and Ruiz Bustos, 1977), perhaps by land crossing or introduced by prey birds (Santonja, 1983: 9).Another North African species (P.miocaenicus) is known in Spain since the Miocene but could have reached both sides of the Mediterranean from its Asiatic homeland separately.These exceptions emphasize the lasting effectiveness of the Gibraltar barrier.
Lower Palaeolithic evidence.A case for hominid dispersal through Gibraltar stands or falls, ultimately, on archaeological proofs.
1) The Acheulian.Occurrences concentrated of both sides of Gibraltar were regarded as a positive indication (Chard, 1963) for a distributional argument until nullified by discoveries from around the entire Mediterranean basin e.g.Palaiokástron, Greece (Higgs, 1963).Specific teehnotypological trait discriminants e.g.handaxe and cleaver manufactured on sidestruck flake preforms in Africa, or handaxe typological attributes, also reported in Iberia (A limen, 1975;Riet Lowe, 1932, 1945: 59;Freeman, 1975;Rolland, 1986) have also been invoked in favour of connections.Several of these actually recur in Iberia and Southwest France -flint handaxes or cleavers made on sidestruck flakes at San Isidro, or Les Pendus (Bergeracois).
These traits may provide clues about diffusionary movements out of North Africa, beeoming more diluted as they spread deeper into Westem Europe, and perhaps even about the European early Acheulian's (e.g.Pinedo) origin, but need closer investigation.The model may prove more plausible than the debatable (see Villa 1983) notion of a «Southem Acheulian» province (Bordes, 1971).In any case, it would have taken place too late to be a colonization event, an issue left without real discussion in AHmen (1975).
Detailed but schematic subdivisions within protracted Early and Evolved stages, served to identify gradual techno-typological developments, leading to an in situ transition into the Acheulian.The sequence, thought to parallel the East African Oldowan to Acheulian succession, has also been used for long-distance crossdating within Africa (Balout, 1981: 569) and Europe (Bordes and Thibault, 1977).An apparent concentration of artifact-bearing loealities similar to El Aeuladero, a10ng the southwestem coast of Spain, in addition to other finds from the peninsula (Santonja, 1983), would not be incompatible with the notion of direct connections with North Africa, although we note that reliable «Pre-Acheulian» occurrences from Iberia, providing a eounterpart repertoire for the 4<Pebble-' culture» across Gibraltar remained unreported (Freeman, 1975: 733;Zbyszewski, 1976) until recentIy.
The Moroccan Pebble-Culture sequence nevertheless contains flaws: a) lack of biostratigraphic, geomagnetic and radiometric dating, b) absencc of fossil hominids, whose identity, given the postulated relative chronology, could be AuslralopiLhecus ur Hamu ha bilis.H. ereclus is known from mid-Pleistocene sites in the region and at Tighenif, c) accuunting for the local Acheulian's late appearance (Alimen, 1977), not earlier than late Matuyama or isotopic stage 22 (Texier el al, 1986: tables 1, 2), despite a proposed local linear transition from the Pebble-Culture, and implying a cultural lag (McBumey, 1975: 413).New researches by Bordeaux University's Institut du Quatemaire and Moroccan institutions have revamped radically this Atlantic Morocco Pleistocene sequence and its palaeoclimatic interpretation (Raynal el al, 1986a, b;Texier el al, 1985).Furthermore, critical evaluations of the provenience, retrieval conditions and identity of the area's Pebble-Culture raise serious doubts about its reality in North Africa (Butzer, 1982: 43).This negative verdict has been confirmed since by Bordeaux University researchers (4).
Removing the controversial Maghreb Pebble-Culture has fundamental implications: 1) it resolves chronological and developmental discrepancies between that region's Palaeolithic and the East African master sequence, where the Acheulian emerges early, linked with H. ereclus; 2) it supersedes the concept of an early and widespread diffusion of an Oldowan or ~Pebble-Culture» Palaeolithic substratum across Africa and Eurasia and its postulated «persisten ce» in certain regions; 3) it eliminates the last criterion left for considering a hominid peopling of Europe directly out of Northwest Africa; 4) the Acheulian constitutes the earliest Palaeolithic horizon in North Africa and is penecontemporaneous with the European Early Palaeolithic Horizon.These conclusions, therefore, add to the series of inconclusive claims made over the years about Ibero-Mor<>ccan prehistoric contacts, prior to the Neolithic (see Ripoll López and Cacho Quesada, 1990;Villaverde Bonilla and Fumanal García, 1990, conceming the Solutrian in Spain).
The Siculo-Tunisian úmdbridge.Recent reports of Acheulian (Graziosi, 1968) and Pebble-Culture in Sicily (Bianchini, 1969) have revived the question of Pleistocene connections with Tunisia.Pebble tool finds need confirmation and connections will remain conjectural until systematic geochronological and palaeogeographic researches will be carried out (Segre el al, 1982: 183-184, 192).They must also take due account of recent findings conceming the existence of a Pebble-Culture in the Maghreb.
The Italian peninsula's configuration with its narrow, compartmentalized structure, surrounded by the Mediterranean and bound by the Alps, was markedly modified during Pleistocene ice ages.Its major geographical links then were to the East, with the emergence of the Adriatic-Ionian shelf into a shallow marshy plain (Mussi, 1990: 139 and fig. 7.1).The density of Early Palaeolithic occurrences, their presence across the Adriatic-Ionian seas (Visgliano, Sandalja, Nea Skalá) and Southeast (Yanm Burgaz) may hold clues about a more significant palaeogeographic orientation of hominid movements.

The Central Asia migration filter
Lower Palaeolithic, biogeographic and palaeographic evidence fail to establish conclusively a hominid dispersal route through the Levant or the Maghreb into Europe.The search for nonhandaxe repertoires, as antecedents for the European Early Palaeolithic Horizon, must therefore move beyond the Southern and Eastem Mediterranean basin, whether such a technocomplex constitutes a linear continuation of the Oldowan or one derived from the Acheulian.Africa is ruled out since the Oldowan is confined to Eastem and Southern Africa, from the Horn to the Transvaal.The only alternative left, inside Eurasia, is the remo te Chopper-Chopping Tool Complex in the Far East.It was concluded earlier that this represented a permanently modified version of the Acheulian rather than an isolated persistence of mode 1 (Oldowan) technology.
1.The Eurasian biogeographic corridor.It illustrates connections and continuities of mammalian spccies between Westcm Europc and China, through Mongolia, Siberia and Eastem Europe, fostered by common bioclimatic and ecological factors (Simpsun and Beck, 1965: 734), including the Pleistocene (Alekseev. 1970;Kurtén, 1975;Sher. 1975).Corridors, like barriers, are never absolute and favour or impede certain species.Horno erectus' technological development was probably insufficient for overcoming bioclimatic obstacles of the Eurasian corridor: the cold Siberian taiga with its low animal and food resources (Chard. 1974: 5), during interglacials; the cold, continental steppe biome, to which hominids adapted only by the Lower to Middle Palaeolithic transition.during glacials.Movements into the westem portion of the corridor were also more restricted by large glacial lake barriers in Westem Siberia and the Khvalyn depression.
2. The Central Asia filter.This region comprises Soviet Central Asia, Afghanistan and Western China (Xinjiang/Uighur region.Oinghai and Inner Mongolia).It is a landmass completely cut off from oceanic influences, bound to the North by the Siberian taiga, to the South by a string of high mountains and plateaus.while retaining narrow internal communication lines or funnels (Hambly, 1969: 4-5).It is regarded as inhospitable, unattractive for human settlement beca use of its arid, harsh clirnate, scarce in food supplies and difficulties of access (Ranov, 1972).
It attracted first the attention of palaeogeographers and palaeontologists as a conjectural hominoid primate evolution centre and even of human origins, since de Ouatrefages, Osborn, Black and sorne Russian scholars, although this has been abandoned since, in the absence of any evidence (see Okladnikov. 1972;Teilhard de Chardin, 1953).
Eventually, several Lower Palaeolithic discoveries have pushed back steadily the date of initial human settling of the region (Ranov, 1984).These documents benefit from a reconstructed regional Pleistocene sequence, based on river terraces.loess deposits, tectonics and mammalian biostratigraphy, calibrated by geomagnetic and TL dating (Dodonov andRanov, 1977, 1984;Ranov and Davis, 1979).
Chronology.The detailed loess and palaeosols stratigraphies from which thcse occurrences carne, complemented by geomagnetic and TL dating, provide a roughly reliable time framework.The record shows Kul'dara as the only late Matuyama anthropic site, so far, in Central Asia but as is the case for 'Ubeidiya, concerning the Levant and Eurasia as a whole it constitutes a significant datum line for Central Asia.Its archaeological contents have also implications because of a complete absence of handaxes (unlike 'Ubeidiya), making it a plausible landmark between Europe and the Far East.
Another observation of interest is that although fauna at the site was scarce and unidentified, the site is close to and belongs to the same time-stratigraphic unit as the Lakhuti 2 faunal locality (Ranov el al, 1987: 69-70).Lakhuti 2 contains a characteristic Galerian fauna of late Matuyama age (Azzaroli, 1983).Another Galerian faunal site in the region is Koshkurgan, of early Brunhes ages (Azzaroli el al, 1988: fig.6).Sel'ungur's basallayers apparently contain similar elements (E stenonis) though no artifacts.It is worth stressing this simultaneous occurrence of the region's earliest Lower Palaeolithic and first Galerian fauna, as well as recalling that this region was al so the Galerian's initial evolution and dispersal centre.
Diagnosis and origins.None of the aboye mentioned assemblages contain any traces of an Acheulian or mode 2 technology, the attribution of the Sel'ungur industry to a «Southern Acheulian» is inappropriate: the term is used exclusively for Southwestern Europe; the Sel'ungur artifact illustrations are not recognizably Acheulian.This absence is attributed partIy to raw material constraints (we know, however, that Acheulian handaxes are made of poor materials unless minimal sizes or certain untractable materials would become an impediment), partly to a local tradition (Ranov, 1982).Central Asia does show evidence for a continuing use of pebble materials, up to the Neolithic.
Scattered handaxe finds have actually been reported, as predicted by Ranov and Davis (1979: 255) but these are c1ustered in the westernmost margin of Central Asia (Turkmenia, western Kazakhstan).While undated, most appear to be long to the Middle Palaeolithic and a few to a late Acheulian probably synchronous with the Caucasus cave Acheulian (Vyshnyatsky, 1989a).Unless new discoveries reverse this generalization, the Lower Palaeolithic of Central Asia adds to evidence suggesting that certain portions of Eurasia were occupied by populations without a mode 2 repertoire (Vyshnyatsky, 1989b).
The genesis of Lower Central Asia's Lower Palaeolithic is unclear although Kul'dara has stimulated new discussions.Mentioned affinities with Southern and Eastern Asia, implying direct connections with South Asia, are problematic (Ranov and Davis, 1979: 255).These become even more questionable because the Sohanians's reality is doubtful.Furthermore, the presence since the Miocene, of major mountain barriers has impeded northward movements (Heintz and Brunet, 1982).
Kul'dara's antiquity led sorne researchers to shift a search for phyletic affinities directly with East Africa and the Oldowan, supported by a string of early occurrences in Western Eurasia: 'Ubeidiya, Azykh, Korolevo, Sandalja, Becov, Vallonet (Gladilin and Ranov, 1986), thus reaffirming or implying the notion of an Oldowayen repertoire precocious dispersal.A closer look at the evidence cited reveals that these Lower Palaeolithic occurrences constitute a lumping of distinct Lower Palaeolithic repertoires: a) 'Ubeidiya is early Acheulian and also the earliest dated site in (5) Artifacts found by 1. Mochanov in the Lena River area, are claimed to be over 1.0 rny (Davis, 1987: 3(0)  A legitimate altematíve, though requiríng a more detailed hypothesis and needing archaeological support, would be to look again for direct connections inside Asia but within the Palaearctic region, in the direction of Northem China.Recently dated sites there indica te a lower Pleistocene hominid occupation (Xihoudou, the Nihewan basin sites), around 1.1-1.0my.Ihis date, combined with their non-Acheulian artifactual pattern, make them a credible antecedent for a similar industry in Central Asia and perhaps ultimately of the European Early Palaeolithic Horízon.
These affinities between Central and East Asia remain hypothetical, arríved at by deduction (by eliminating South Asia or East Afríca) and circurnstantial.and without known Lower Palaeolithic in the intermediary, Western China, region.Iwo observations are made regarding this: 1) research experíence in Central Asia has shown that Lower Palaeolithíc discoveríes are diffícult to make, requiríng intensive, repeated fieldwork, and wiIl remain few in number; 2) Quatemary research in Westem China indícates that hominid settlement conditions were more favourable duríng the lower Pleistocene than this presently inhospitable region would indícate.
Palaeoclimatíc reconstructions, however, indicate that although this basic pattern was already in place by the lower Pleistocene, the trend increased until Neothermal times, underlying climatic fluctuations.The Qaidam basin and YeIlow River areas had a milder, wetter climate then, becorning dríer and colder from the mid-Pleistocene onwards (Wang Kelu et al, 1988;Zhang Lansheng, 1988;Zhen Benxiang, 1988;Jiang De-Xin, 1988).These findings provide palaeogeographic support for a possible westward hominid dispersal around 1.0-1.1 rny, along narrow ríver valleys (Sulo He, Tarím) and palaeolakes rnargins, ultimately reaching Central Asia (Kirghizia, Uzbekistan, Tadjikistan, Afghanistan) by a filter route between the Tien Shan and the Pamirs.
3. Dispersal route into Europe.In this section, we identify; 1) routes available to andent hominids dispersing from Central Asia toward Europe, along the southern Palaearctic region, 2) Lower Palaeolithic occurrences linking both regions and 3) localities with end-Villafranchian/Galerían fauna.The purpose is to examine the extent of overlap between these lines of evidence.Positive findings would gíve initial support for a hypothesis a) linking the non-Acheulian of Europe with that of Central Asia into a single repertoire, b) that processes underlying such a connection could relate to a dispersal event involving both horninids and the Galerían mammal spedes.
Dispersal route.The most natural one, west of Central Asia, would involve avoiding the northern deserts (Kizil Kum, Kara Kum), rnoving south of the Kopet Dagh and Elburz, along the more attractive alluvial fans habitats at the edge of these mountains and flanking the northern inland deserts of the Iranian plateau (Spooner, 1972).Ihere is evidence for Pleistocene fossil or enlarged lakes, especially in the Northeast (Smith, 1986: 11).Contacts to the south across deserts probably never took place until late prehistoríc times and the advent of animal transport and long-distance trade.The southem deserts would be especially forbidding.Beyond this, movements would lead westward into the favourable habitats of the Transcaucasus and then, along the north Anatolian tectonic trough towards the Bosphorous and into Southeastern Europe.Alternatively (or additionally), glacioeustatíc loweríng of the Black Sea would liberate substantial areas of presently submerged land in the Sea of Azov and west of Crimea.
Lower Pakieolithic landmarks.Ihese remain sparse and scattered widely, outside Europe, particularly in Iran where Palaeolithíc research has be en marginal: the Kashafrud basin finds, in Khorassan, Northeastern Iran, Azykh cave (Azerbaidjan), Gerassimovka (Azov coast, Russia), and  Yanm Burgaz (Turkish Rumdia).The Kashafrud finds (Abravan, Baghbaghu), several of which are in stratigraphy or with known provenience (Ariai and Thibault, 1975-77), come f rom fossillacustrine contexts.Artifactual inventories are similar to those from Europe and Central Asia.Dating is imprecise, based on geology but the finds' position in the terrace stratigraphic sequence indicates considerable antiquity.Their location, adjacent to Central Asia. is also interesting, lending credence to suggestions that the earliest occupants of the Iranian plateau conceivably carne from that region (Smith, 1986: 14).Another potentially important though poorly dated occurrence comes from Iranian Azerbaidjan (in Smith, 1986) would also fit the dispersal trajectory proposed here.(based on Azzaroli, 1983;Azzaroli et al., 1988;&mifay, 1979, 1980). o Galerian mammalian fauna occurrences.Europe (based on Howell, 1960;Oakely, 1964;Whithing el al., 1982) --+ --+ --il1itial dispersal/rom Easl Africa up /0 Southeast Asia; cOl1tinuing dispersal through the Southeast Asinn tropical forest filter route, then along the temperate, broad-lea/ forest habitat of coastal China, and through the arid continental arid zone filter of Western China.
. Linkage with dispersal events.The end-Villafranchian/Galerian fauna spread from its Central Asia centre in several directians, mainly northcast into Siberia and to the West into Europe (Azzaroli e/ al, 1988: fig.2).Europe's greater si te concentration reflects more intensive coverage.Figure 1 plots Galerian fauna sites and non-handaxe Lower Palaeolithic sites, from Central Asia to Europe.A noticeable geographical overlap emerges, including many sites where both lines of evidence are associated.This overlap is partial, since mammals disperscd in multiple directions across their continental realm whereas hominids followed an apparently linear pattcrn.The latter illustrates, of course, only one species' dispersal, which in the case of Hamu eree/lIs probably confined itself to more familiar home ranges.The Siberian faunal sitcs are entirely without artifacts.
This broadly cohesive time-space palterning in Western Eurasia of early Palaeolithic occurrences with the Galerian fauna suggests a broader event, analogous to that earlier one proposed by Tumer for hominids and camivores out of Africa.Azzaroli identifies the probable causes (tectonics, climatic deterioration) for the large-scale late lower Pleistocene end-Villafranchian/ Galerian tumover across temperate Eurasia.The hominid populations who were beginning to people Central Asia undoubtedly feh the repercussions of these profound perturbances, such as shifting home ranges pressures to adjust to game animal dispersions.
The hypothesis contains weak and strong points.The fonner include: 1) a sparse often poorIy dated archaeological evidence, outside Europe; 2) a major archaeological documentary gap between Central Asia and North China; 3) the geographical overIap of Lower Palaeolithic and Galerian sites may be without significance if hominid movements proceeded out of Europe, in a direction opposite to the Galerian fauna dispersa\.This objection, however, cannot be tested because chronological data are too broad to detect age-area pattems but if considered seriously, it would leave unanswered the question of the European EarIy Palaeolithic Horizon's origino The strong points are: 1) it provides the only plausible explanation left for the earIy European non-handaxe technocomplex; 2) the overIapping pattems require an interpretation; 3) the model is the only one without conflicting evidence; 4) its «long joumey» scenario for the peopling of Europe, shown on Figure 2, takes into consideration the physiographic configuration of much of Eurasia and its East to West trending, which conditioned movements and isolation.This orogenic barrier al so accounts for the persistence in isolation, from North China to Europe, of an initially Acheulian repertoire, modified permanently once hominids moved into the tropical Indo~Malaysian forest; 5) the hypothesis is open to refutation by empirical evidence, such as if earIy Acheulian sites were discovered in Europe, Central Asia or China.
As it stand s, this hypothesis is based on goodness of fit of evidence and merely represents an exploratory model in need of a larger data base.It may not be the most economical explanation but prehistory and biogeography both contain ample instan ces of protracted, reticulate events.

CONCLUSION
The paper synthesizes current knowledge about the Palaeolithic colonization of Europe.Like any surveys of this kind, it can become rapidly obsolete, as new evidence is added, sorne of which may require radical revisions of viewpoints presented he re.The main focus was on the question of the nature and origin of the earliest Palaeolithic horizon in Europe and along with it, of population movement routes followed by andent hominids.
Three main altematives were considered, with respect to colonization routes: through the Levant, across Gibraltar, out of Central Asia and ultimately, northem China.None have been established conclusively yet but the latter, linking hominid and Galerian faunal dispersal, is more cohesive and without data incompatibility.It is a preliminary testable hypothesis, but there is a pressing need to improve the quantity and quality of its data 'base, particularIy in regions such as Southeastern Spain, the loess regions of Central Europe, Central Asia and especially, Western China.New empirical findings, however, may invalidate entirdy this third alternative.What has been covered represents only one aspect of the theme of Palaeolithic colonization of Europe.Other key topics would inelude reconstructing ancient hominid socio-ecology, adaptive processes under earlier Pleistocene bioclimatic conditions in temperate Eurasia, the role of predation, of fire making, settlement systems but all would require a much richer documentation.
FIG. 2.-Hypothetical early hominid dispersal routes out o/ A/rica, through southern Eurasia, and wes/ward in/o the southern Palaearctic filter route, from Central Asin into Europe, /00 C mean temperature isotherm durint coldest month in the Northern hemisphere.Pleistocene glacio-eustatic coastal con tour of Eurasia.The large-scale loop shown on the map traces hominid population dispersal route circumventing the major Bast-West trending mountain barrier of Eurasia.T. P., 1992, nI! 49 (c) Consejo Superior de Investigaciones Científicas Licencia Creative Commons 3.0 España (by-nc) http://tp.revistas.csic.esTHE PALAEOLITHIC COLONIZATION OF EUROPE: AN ARCHAEOLOGICAL ANO ...

TABLE 3 .
Chronological Summary 01 Archaic (Pre-Acheulúm) Lower Palo.eolithicOccurrences in Europe T. P., 1992, n Q 49 but little detailed information has reach this writer yet. .Eurasia; b) the others, from Kul'dara to Vallonet, are non-Acheulian -but not «belated Oldowan»and a liule later than 'Ubeidiya.It is recalled that the Acheulian supersedes the Oldowan by 1.4 my in Subsaharan Afríca.