Cour. Forsch.-Inst. Senckenberg

256

305–314

4 Figs, 1 Tab.

Frankfurt a. M., 15. 11. 2006

Afro-Eurasian mammalian dispersal routes of the Late Pliocene and Early Pleistocene, and their bearing on earliest hominin movements

With 4 figs, 1 tab. Hannah O’REGAN, Laura BISHOP, Sarah ELTON, Angela LAMB & Alan TURNER

Abstract Mammalian migrations provide an excellent context for the dispersal of hominins in the Plio-Pleistocene. Here we present a discussion of a variety of different techniques that can be used to elucidate migration patterns in the fossil record, focussing on the Gibraltar Straits as an example. There are three main mechanisms for studying such patterns – 1) recent mitochondrial DNA analyses on a variety of taxa, 2) modern circumMediterranean biogeography and 3) the biogeography of fossil taxa. The advantages and disadvantages of each are considered, and the circum-Mediterranean palaeobiogeography of three genera: Theropithecus, Hippopotamus and Homo are discussed. We consider that, with the exception of a few small bodied animals, there is little evidence of across Gibraltan dispersal since the Messinian, making this an unlikely route for Hominin migration. The most likely route of dispersal for both fossil and modern large mammal migrations appears to be the Levantine corridor. We then conclude this paper with a discussion of the wider theoretical context for such large-scale palaeobiogeographic studies.

Key words:

biogeography, phylogeography, Gibraltar Straits, Quaternary, Theropithecus, Hippopotamus, Homo eschweizerbartxxx sng-

Introduction The number, timing and direction of earliest extra-African dispersal events in human evolution has long been unclear and contentious (ROLLAND 1998, TURNER 1999a), although substantiated dates for hominin appearance in Eurasia set a minimally early date for the original dispersal from Africa. Early Pleistocene assemblages from ‘Ubeidiya in Israel are now generally accepted to date to around 1.5 million years (BELMAKER et al. 2002), while quartzite flake artefacts have also been reported in deposits from northern Pakistan considered to date to 1.9 Ma and perhaps even earlier (DENNELL 1998). However, the earliest extra-African hominin material generally given credence is that from Dmanisi in Georgia with a claimed age of 1.75 Ma (GABUNIA et al. 2001, VEKUA et al. 2002) based on dating volcanic deposits that immediately underlie the hominin-bearing level, magnetostratigraphy and the presence of an undoubtedly Early Pleistocene mammalian

fauna. The current consensus view, insofar as one can be identified, would seem to favour a shorter chronology for more extensive and intensive occupation of Europe in particular and of Eurasia in general – that is after around 0.5 Ma – with a tail of more sporadic appearances back to and perhaps even prior to the Plio-Pleistocene boundary (TURNER 1999a, ROEBROEKS 2001). Whether these earliest dated occurrences represent one or more discrete dispersal events is a matter for debate (AGUIRRE & CARBONELL 2001, DENNELL 2003), and they are unlikely to represent the first actual appearance of the animals, but rather the first appearance in the fossil record (WHITE 1995). In our view, efforts to assess the reality and patterning of the very earliest events must consider not only the archaeological and human skeletal evidence but also the larger context of human dispersal as one facet of the evolutionary history of the terrestrial mammalian fauna of the Plio-Pleistocene (TURNER 1999a, b). That context must include some view of the theoretical underpinnings

Authors’ addresses: Dr. Hannah O’REGAN (corresponding author), Laura BISHOP & Alan TURNER, School of Biological and Earth Sciences, Liverpool John Moores University, Liverpool, L3 3AF, UK, ; Sarah ELTON, Hull-York Medical School, University of Hull, Hull, HU6 7RX, UK; Angela LAMB, NERC Isotope Geosciences Laboratory, British Geological Survey, Keyworth, Nottingham, NG12 5GG, UK. © E. Schweizerbart’sche Verlagsbuchhandlung (Nägele u. Obermiller), 2006, ISSN 0341-4116

O’REGAN et al.: Mammalian dispersal in Plio-Pleistocene Afro-Eurasia

of the origin and dispersal of species, and our starting point is the notion that new taxa originate in allopatry and remain within their preferred habitats until fresh conditions provoke a fresh pattern of range fragmentation. Species, in other words, have a localised place of origin and attain a wider range by dispersal (TURNER & PATERSON 1991). The fossil record is best suited to the study of the resultant patterns and the inference of processes that took place over long time periods, across a range of taxa and across wider geographical areas. The approach that we advocate differs from those proposed by ANTÓN et al. (2002) and MITHEN & REED (2002) for investigating hominin dispersals. Both sets of authors offer a modelling technique, the former to estimate diffusion coefficients and the latter to give a computer simulation of migration rates and likely timetabling of movements. Such modelling approaches have considerable heuristic value, and certainly meet the call for a theoretical underpinning to investigations, but they also present some problems. Both require some first appearance datum (FAD) on which to calibrate the model and both quite reasonably select the site of Dmanisi for this role. In the case of the computer simulation offered by MITHEN & REED (2002), the need to calibrate the model in this way is clearly a difficulty, since it relies on accurate dating of the FAD. Moreover, while it is certainly possible to build environmental factors such as glacial conditions, or sea-level changes into the simulation, the model can only take account of so much, forcing the authors to conclude that historical contingency may have had an immense influence on the actual outcome. Therefore, even when guided by a model derived from a simulation, it is only by looking at the fossil evidence itself that one can decipher the sequence of events and attempt to reconstruct the evolutionary history of taxa. As MITHEN & REED (2002) point out, their model seeks to go beyond generalisations about the likely impact of proposed barriers and land bridges on human dispersal by using quantification and simulation modelling. We too ask: what is the likely timetable of earliest hominin movements from Africa and what were the constraints and obstacles operating at various times. We believe that by looking at the wider palaeontological context we can discern when other terrestrial taxa were moving. This allows reconstruction of the availability of migration routes across a wider section of the biota, permitting investigation of the history of hominin migrations to go a step further, a point also made by ROLLAND (1998: 211). eschweizerbartxxx sng-

Routes out of Africa There are several potential routes out of Africa in the Pleistocene, including the Gibraltar and Bab-el-Mandeb Straits, through the Levant or a sea route from Tunisia to Italy via Sicily. One thing we can be certain of is that any pattern of movement from Africa occurred against a background 306

of climatic change. Late Pliocene and Pleistocene climatic history may be characterised as a general cooling trend towards the present day, interrupted by stepped changes and periods of increasingly intense glacial cycles, commencing with the first northern hemisphere ice sheets at around 2.6 Ma (SHACKLETON et al. 1984, SHACKLETON 1995, DENTON 1999). There is a small step towards averagely colder conditions at around 1.8 Ma, followed after 0.9 Ma by a further step and the onset of the intense glaciations that characterise the Middle and Late Pleistocene. Various studies, led by the pioneering work of Elisabeth VRBA (1995a, b, 1999), strongly suggest a link between faunal turnover events, including dispersals, and major climatic steps in Africa and Eurasia (TURNER 1995, 1999b, TURNER & ANTÓN 1999), although recent temporally controlled studies of African mammalian faunas have found complicated patterns of change, making climatic forcing less clear cut (BEHRENSMEYER et al. 1997, BOBÉ et al. 2002). For many, the most probable route for human movements out of Africa has lain through the Arabian Peninsula and the Levant, either across the Sinai Peninsula or the Bab-el-Mandeb Straits at the south of the Red Sea. This is indicated by the mixed Afro-Eurasian nature of the fauna of the region since the later part of the Pliocene and in particular by the Early Pleistocene at the site of ‘Ubeidiya (TCHERNOV 1992, TURNER 1999a, BELMAKER et al. 2002) as discussed below. The dispersal represented by these extra-African occurrences appears to have been part of a larger faunal turnover in Eurasia, which AZZAROLI et al. (1988) termed the Elephant-Equus event in Europe since it was thought to be marked by the first appearance of Mammuthus and the true horse, Equus, although the first appearance of these genera now seems to have been somewhat earlier (RADULESCO & SAMSON 1990, 2001). The larger scale nature of the turnover is seen in Africa by 2.3 Ma (TURNER 1995, 1999b, VRBA 1995a, b, 1999), while in Asia, African bovids appear as immigrants in the Siwaliks (VRBA 1995a). The Early Pleistocene sites of both Dmanisi and ‘Ubeidiya contain a small number of African faunal elements. At Dmanisi (GABUNIA et al. 2001) this consists of a hominin that can really only be of African ancestry, whilst the ostrich, Struthio dmanisensis, may be African but this is not certain. At ‘Ubeidiya the African affinities are much clearer with Homo sp., the bovid Pelorovis oldowayensis, the hippo Hippopotamus gorgops, a suid Kolpochoerus olduvaiensis and the spotted hyaena, Crocuta crocuta (TCHERNOV 1992). Of course if ‘Ubeidiya is indeed around 300,000 years younger than Dmanisi then one could argue that the hominin at the former is as likely to have come from Georgia as from Africa, although the Acheulean industry at ‘Ubeidiya is unknown at Dmanisi and does appear to have developed in Africa by 1.5 Ma (ASFAW et al. 1992). Other Early Pleistocene localities in Europe have relatively few African species. Much has been made of the appearance of the cercopithecoid Theropithecus cf. oswaldi

Cour. Forsch.-Inst. Senckenberg, 256, 2006

at Cueva Victoria in southeastern Spain (see below) and of the possibly African machairodont cat, Megantereon whitei, at Venta Micena and in the fauna at Dmanisi and at the Greek locality of Apollonia (MARTÍNEZ-NAVARRO & PALMQVIST 1995, 1996). However, while Theropithecus may indicate African links, the genus Megantereon is a well-established member of the Afro-Eurasian carnivore fauna since around 3.0 Ma (TURNER 1987) so that claims for its significance in the question of dispersals require further investigation. Of rather more obvious importance is the presence of Hippopotamus, which makes an early appearance at ‘Ubeidiya (TCHERNOV 1992) and Venta Micena (MARTÍNEZ-NAVARRO 1992) and is clearly a member of the African fauna. There is a Late Pliocene record of Hippopotamus sp. at the site of Valea Graunceanului in Romania (BOLOMEY 1965), but this occurrence has not been mentioned in later syntheses of the Romanian fauna (RADULESCO & SAMSON 1990, 2001), and is therefore not discussed further here. The question of movements across other parts of the Mediterranean does however recur in the literature, most commonly in the case of the Gibraltar Straits. ROLLAND (1998) argues that conditions during glacial maxima may have reduced the seaway through the Straits to a mere 8 km and that the lowered sea-level would not have increased surface currents. In his view, while the general absence of evidence for movement of other species underlines the effectiveness of the barrier to movement, it becomes possible to argue for sweepstake-like movements, especially during MI Stages 12 and 16. FLEMMING et al. (2003) argue that the Strait itself would not have narrowed significantly during sea-level falls, although they point out that now submerged areas to the west of the Strait would have formed substantial islands that might have provided “stepping stones”. This possibility is discussed further by COLLINA-GIRAUD (2001), who describes and illustrates four islands in the Gibraltar Straits that would have been visible at the Last Glacial Maximum (and presumably also in the preceding glacial periods). So far as other routes across the Mediterranean are concerned, particularly between Tunisia and Sicily, FLEMMING et al. (2003) reach no conclusion. To investigate the Gibraltar route we offer an assessment of likely movements across the Straits, beginning with a theoretical consideration of the actual evidence likely to offer support for, or against, hypothesised movements across such a region. eschweizerbartxxx sng-

Gibraltar: theoretical patterns of movement and evidence in support The Mediterranean area may be seen as essentially a water body dividing Africa from Eurasia, with a path around one side (the Levant), and a gap at the other where the Gibraltar Straits exist. However both sides of the Gibraltar Straits are considered to be part of the Palaearctic region, today

(e.g. DOBSON 1998) and as far back as the Early Pleistocene (O’REGAN et al. 2005). If we consider a hypothetical population that originated in the Caucasus and subsequently dispersed to North Africa and Iberia (see fig. 1a), then it could have migrated via one of three routes: 1) Through Europe to Iberia and then across the Gibraltar Straits to North Africa (fig. 1b). 2) Through the Levant and North Africa, crossing the Gibraltar Straits from South to North, and ending in Iberia (fig. 1c). 3) Two population movements – one through Europe to Iberia and the other through the Levant to North Africa, with the Gibraltar Straits forming a permanent barrier to dispersal (fig. 1d). If the founding population were then to become extinct, these three routes would leave superficially the same distribution pattern, an apparent population in North Africa and Iberia, with the possibility of a migration across the Gibraltar Straits. How then could someone studying either the modern or fossil populations tease these three possibilities apart? There are three techniques that can be used to infer past migrations – modern biogeography, phylogeography and palaeobiogeography. Each of these has its own advantages and disadvantages, and these will be discussed in the light of recent work on fauna of North Africa and Iberia. Modern biogeography In a recent study of extant and Holocene mammals in the Mediterranean region, it was found that there were no terrestrial mammal species present in North Africa and Iberia that were not also present in the Levant (DOBSON & WRIGHT 2000). Species that are found in both North Africa and Iberia include the wild boar (Sus scrofa), red deer (Cervus elaphus), otter (Lutra lutra) and the red fox (Vulpes vulpes), suggesting that these animals have taken a circum-Mediterranean route rather than crossing the Gibraltar Straits (equivalent to fig. 1d) (DOBSON 1998). However, most mammals are capable of swimming, so the possibility of dispersal across the Straits cannot be completely ruled out. Flying mammals (Chiroptera) exhibited a slightly different pattern, with the same species often being found on both sides of the Straits and further eastwards in Europe, but not recorded in the Levant or elsewhere in North Africa. DOBSON & WRIGHT (2000) therefore suggested that such distributions offered strong support for movements across the Gibraltar Straits, but see the section below on phylogeography. One of the problems associated with studying modern distributions is the difficulty of accounting for human interference – both in causing or assisting range contractions and in aiding dispersal of particular species. For example, of the 17 terrestrial mammal species inhabiting North Africa today, only four are considered to have a natural circum-Mediterranean distribution by DOBSON (1998), Vulpes vulpes, Sus scrofa, Lutra lutra and Felis 307

O’REGAN et al.: Mammalian dispersal in Plio-Pleistocene Afro-Eurasia

Fig. 1: Possible migration routes around the Mediterranean to Europe and North Africa illustrated by the dispersal of a hypothetical population from the Caucasus. 1a) Initial dispersal of Caucasus population to North Africa and Iberia. 1b) The northern route – dispersal from the Caucasus through Europe to Iberia and then across the Gibraltar Straits from North to South, ending in North Africa. 1c) The southern route – dispersal through the Levant and North Africa, crossing the Gibraltar Straits from South to North, and ending in Iberia. 1d) The pincer route – two population movements, one through Europe to Iberia and the other through the Levant to North Africa, with the Gibraltar Straits forming a permanent barrier to dispersal. eschweizerbartxxx sng-

sylvestris. The rest are thought to be recent introductions, such as the Barbary macaque (Macaca sylvanus) brought to Iberia from North Africa, and the rabbit (Oryctolagus cuniculus) taken to North Africa from Europe (DOBSON 1998). The extinction of three more species (Bos primigenius, Ursus arctos and Cervus elaphus) in the Holocene may also indicate human interference. The examples above fall mostly within the time of historical records, but other changes may have occurred much earlier, perhaps pre-Holocene, and would not therefore be picked up in a study that looked at Holocene distributions. If Homo sapiens and their dogs were capable of dispersing over substantial bodies of water to reach Australia in the Late Pleistocene (BOWLER et al. 2003), then it is also possible that they may have been moving around the Mediterranean and transporting animals prior to the Holocene. Mus spretus (the Algerian mouse) is a good example of an animal that initially appears to have an excellent across-Gibraltan distribution since it is only found in Iberia, southern France and the Maghreb. However, it is only found in the fossil record of North Africa (DOBSON 1998), and mitochondrial DNA analysis indicates 308

that it is either a very recent introduction or a Pliocene migrant to Europe (BOURSOT et al. 1985). Phylogeography In the last few years there have been several mitochondrial DNA analyses of small vertebrate populations such as bats, amphibians and reptiles to look at the patterns of dispersal to the Canary and Mediterranean islands, as well as across the Gibraltar Straits. These studies are summarised in table 1 and discussed below. In the case of the Barbastelle bats (Barbastellus barbastellus) contact does indeed appear to have occurred. The migration route for these animals is thought to have been across the Straits since there are no zoological records or fossil remains for this species elsewhere in North Africa, although it is present in Turkey (JUSTE et al. 2003). However, the suggestion that the presence of particular species of bat on either side of the Straits necessarily represent animals that are or have recently been in contact with each other (DOBSON & WRIGHT 2000) has been countered by recent genetic analyses of the animals.

Cour. Forsch.-Inst. Senckenberg, 256, 2006

For example, the mitochondrial DNA of the North African mouse-eared bat is so divergent from that of the Iberian Myotis myotis that it can be regarded as a different species (CASTELLA et al. 2000). The suggested divergence date for these two populations is the Pliocene, but the authors of the study point out that the population may have evolved elsewhere, and only recently arrived at its present across Gibraltar distribution (CASTELLA et al. 2000). Such a dispersal pattern with an ultimately circum-Mediterranean distribution would be in agreement with fig. 1d (above). Something similar is seen in the genus Plecotus, which is found in Egypt, Turkey, Greece, Iberia and North Africa. However, in this case there is a divergence between the eastern and western Mediterranean populations thought to relate to migration patterns from glacial refugia (JUSTE et al. 2004). The North African bats are most closely related to those from the Canary Islands, Turkey and the Balkans, whilst there are high levels of divergence between the North African and Iberian bats. Although occasional vagrant individuals are recorded, they do not appear to have contributed to the gene pool, and the authors suggest that the new arrivals may be outcompeted by the indigenous animals (JUSTE et al. 2004). In the case of the ribbed salamander, Pleurodeles, one haplotype occurs on both sides of the Straits, indicating that there has been at least one successful crossing since the Messinian (VEITH et al. 2004). This could be due to human-aided dispersal (although the African haplotype is widespread in Spain, perhaps indicating some antiquity for the arrival in Iberia) or natural rafting, although many amphibians are intolerant to salt water. The wall lizard (Podarcis spp.) is also found on both sides of the Strait, and in this case the divergence has been dated through calibration with the fossil record to two separate events, one at 3.5 Ma and again at 1.5 Ma (HARRIS et al. 2002). These are both too early for humans to have aided in their migration and must therefore be the result of natural rafting. The conclusion to be drawn from these genetic studies is that even closely related species have different dispersal patterns, so the phylogeography of one cannot be used to predict the biogeography of another. It is also apparent that Gibraltar forms a barrier for some animals and not for others, and that some such as amphibians do appear to have crossed since the end of the Messinian, despite being intolerant of salt water. It would be interesting to do further work on large mammal species such as the wild boar (Sus scrofa), otter (Lutra lutra) and the red fox (Vulpes vulpes) to test the prediction from biogeography that they have taken a circum-Mediterranean migration route and have not crossed the Gibraltar Straits. This however brings us to one of the problems of the genetic studies – they are often done on mitochondrial DNA and use relatively small sample sizes. For example, if a male wild boar were to cross from North Africa to Spain it might interbreed with the local animals, but would not be detected as a new mitochondrial lineage since that only reflects female ancestry. Even if a female crossed the Straits, it is possible that eschweizerbartxxx sng-

those genes would be ‘lost’ within the population, which means that occasional vagrants would not be identified in limited sampling regimes. Only if that species were not present would we see the arrival of the new individuals as a migration event; information that can already be picked up from the fossil record. Genetic studies are also not independent of the fossil data, since migration events are usually calibrated by the first appearance datum of that species in the fossil record. This of course is open to many errors, including misidentification, poor dating of sites or poor provenance. Such errors could ultimately lead to wrongly calibrated migration or speciation events. In addition, only in cases when a founder population is extant, possessing greater genetic diversity than subsequent populations, can DNA tell us where a species originated. On a similar note, DNA can only be extracted from extant or recently extinct species, and therefore the further back in time we go the less helpful it is. Palaeobiogeography – fossil distributions We conclude the discussion of how the modern techniques and extant species can be used to infer migration and dispersal by considering three fossil taxa. The first of these is the primate genus Theropithecus (see fig. 2). The genus Theropithecus is best known from Africa, where it has an extensive distribution (PICKFORD 1993, ELTON et al. 2003), including several findspots in North Africa where the earliest record is in Late Pliocene deposits at Ahl al Oughlam (GERAADS et al. 1998). Theropithecus material has also been found outside Africa, although the total amount is very small. A fragment of right maxilla with two teeth was identified, albeit controversially, from deposits of later Early to Middle Pleistocene age at Mirzapur in India (DELSON 1993). In 1995 Theropithecus was reported among an Early Pleistocene assemblage from Cueva Victoria in Spain (GIBERT et al. 1995), and although this find consisted of a single tooth it was suggested that it indicated an across Gibraltar dispersal from North Africa to Iberia. More recently, three cervical vertebrae of Theropithecus have been reported from Pirro Nord in Italy (ROOK et al. 2004) and a juvenile calcaneum from ‘Ubeidiya in Israel has also been assigned to this genus (BELMAKER 2002). Unfortunately, the identification of this latter specimen is by no means certain. However, if we accept these identifications as correct the distribution of Pleistocene Theropithecus changes from that in fig. 1c to that of 1d, where the animals have migrated around the Mediterranean. This would parallel the pattern seen in the modern fauna as discussed by DOBSON & WRIGHT (2000). The second is Hippopotamus, an African genus that was present throughout Europe in the later Early Pleistocene (see fig. 3 for an indication of the range of Hippopotamus at this time). It is present in both North Africa at Thomas Quarry 1, level L (dated to 1.0Ma) (GERAADS 2002) and in the Levant at ‘Ubeidiya and Evron Quarry 309

O’REGAN et al.: Mammalian dispersal in Plio-Pleistocene Afro-Eurasia

Fig. 2: European and North African Late Pliocene, Early and Middle Pleistocene sites with Theropithecus spp.

eschweizerbartxxx sng-

Fig. 3: Indicative distribution of European Early Pleistocene sites with Hippopotamus spp. (not all sites shown), also shown are North African Early and Middle Pleistocene sites with H. sirensis.

(TCHERNOV 1992, TCHERNOV et al. 1994). Therefore it could plausibly have crossed the Straits, or migrated through the Levant to reach Europe. It is present at sites in the Levant and Europe which are earlier than the finds in North Africa, suggesting that it followed a circum-Mediterranean route. However, the hippo is a water dwelling animal and is capable of swimming some distance. Therefore it is possible that it may represent a migration event from North Africa to Europe (fig. 1c), or from the Levant to both North Africa and Iberia (fig. 1d), or two dispersal events, one within Africa to North Africa, and another out of Africa via the Levant. The third example involves members of the genus Homo, which have been found throughout the Mediterranean area (see fig. 4) at various times from the beginning of the Early Pleistocene at Dmanisi in the Caucasus (VEKUA et al. 2002) to ‘Ubeidiya in Israel, Ceprano (ASCENZI et al. 1996) and Atapuerca in Europe (BERMUDEZ DE CASTRO et al. 1997) and at Thomas Quarry 1, Level G, in Morocco (GERAADS 2002). In this case there is possible evidence for hominins in Iberia at archaeological sites such as Fuente Nueva 3 (MARTÍNEZ-NAVARRO et al. 1997), much earlier than that from North Africa (the Thomas Quarry 1 Homo mandible only dates to 0.5 Ma), which could imply a migration across Gibraltar from North to South (fig. 1b). However, they may also have approached from the Levant, which would be migration like that seen in fig. 1d. There are of course problems with using only fossil distributions to assess migrations, since the syntheses we produce are only as good as the data that we have at any one time (e.g. Theropithecus, above) and include all of the taxonomic and chronological problems inherent in palaeontology. We are also unlikely to find as a fossil the first or only individual of a species that was present in a particular area, so finding an individual suggests that it was there as part of a breeding population. And, as discussed with reference to the genetic evidence, the first occurrence of a species in the fossil record is not necessarily found where or when it originated.

Discussion and Conclusions Routes from Africa

Fig. 4: European and North African Early and Middle Pleistocene sites with Homo spp.

310

Three points deserve emphasis. First, although it is possible that the Gibraltar Straits were the scene of later Pliocene or Pleistocene movements, this is not supported by our analyses of modern biogeography and phylogeography. Second, therefore the Levant/Arabian Peninsula area is currently the only established route, and was the scene of a two-way faunal movement between continents during the later Pliocene. Third, the diversity of elements dispersing across the Levant points to the relatively hospitable nature of the area in the later Pliocene. Hominin dispersals into Eurasia during the later Pliocene are suggested in the literature from time to time (BONIFAY & VANDERMEERSCH

Cour. Forsch.-Inst. Senckenberg, 256, 2006

1991, BOITEL et al. 1996), and while these are unsupported by some critical assessments of the evidence the possibility of migration at this time cannot be dismissed as impossible or even unlikely. These conclusions parallel some of those reached by MITHEN & REED (2002) in their computer simulation of dispersals and are stressed elsewhere (DENNELL 1998). It is clear from the evidence of climatic change that the Levantine filter would have begun to operate more strongly in the period after 2.0 Ma, as increased cold was accompanied by greater aridity and an overall harshening of conditions. This strongly suggests that the picture of dispersal events that we see in the palaeontological record is a robust one. Taken overall, the Early Pleistocene evidence suggests the effect of a strong filter on dispersals across the Levant during that period, particularly in contrast with the Late Pliocene period. The earliest agreed traces of human presence in Eurasia are unlikely to record the earliest movement from Africa, and there is a strong prima facie case for seeing the later Pliocene as a more likely period of dispersal. While recent investigations confirm the continuity of African and Eurasian Homo erectus (ASFAW et al. 2002), there is even the possibility of movements in both directions (CLARKE 2001). We must therefore consider at least the past 3.0 Ma as a possible time frame for actual and potential biotic contact and movements of early hominins between Africa and Eurasia, based on an understanding of mammalian palaeontology, palaeoclimatic reconstruction and evolutionary theory as it applies to the origin and subsequent history of species. eschweizerbartxxx sng-

Wider implications of biogeographic synthesis On a wider note, dispersal is but one facet of evolutionary change – the others being speciation, extinction and within-species changes – potentially provoked by some level of change in the physical environment. Without such a larger perspective, any given change in the distribution of hominins or indeed of any other species remains a single event explicable only in terms of its particular details, such as direction, and can only be interpreted on the basis of such details and vague notions about a will to roam or disperse into new territories or population pressure. Such interpretations have no evolutionary significance; we would know nothing of the context in which events took place, whether climatic changes occurred, whether tectonic events coincided, whether other species also moved or whether the observed phenomenon was unique to the species in question. In contrast, a dispersal that takes place in conjunction with movements of other taxa has an immediate context. It points to the general availability of dispersal routes and, when we examine the types of other taxa involved, can inform about the constraints and obstacles operating at various times. It allows us to tackle not only the question of why dispersals happen at given times and places but also the equally important question of why they do not

happen at other times and places, because we are able to ask whether there are identifiable periods when, or reasons why, dispersal is unlikely to have been possible. The example of Gibraltar is a case in point. In the context of palaeoanthropology, a wider perspective allows us to assess how likely it is that hominins moved from Africa even prior to the Pleistocene boundary, as frequently suggested in the literature, and it establishes very clearly whether the processes underlying early human dispersal are likely to be the same as those operating across the biota. It does this in terms of the correlated patterns between taxa, but it does it perhaps more clearly still if these patterns in turn correlate with discrete climatic events. Then we have clear indications that early human migrations are driven by the very motors of evolution, unless we accept simplistic alternative answers to the implications of cross-taxon synchronicity in movement patterns such as one suggesting the humans were merely following the animals. If we ignore such trivial interpretations, we have at that point achieved the combination of the theoretical and the empirical, the combination as argued by ROSENZWEIG (1995: 5) with “the ultimate power to convince us of scientific reality”. So far as sampling and the treatment of small data sets from small site numbers is concerned, a frequently raised objection to discerning patterns in the fossil record, such matters always present problems for palaeontological synthesis. Synthesis is in our view an essential and necessary component of investigation in any historical discipline if we are going to advance thought and hypotheses within that field. The fossil record is never as good as it is going to be, and if one is going to argue that the existing data rule out synthesis then we simply go back to shopping lists of sites and fossils and then fulfil some critics’ accusations of palaeontology as stamp collecting. Although the sample of localities is small, that is all we have, and we cannot pretend to have data sets that might be preferred in other disciplines. That issue besets the study of human evolution in particular, which is why we argue strongly that it is better to look at human dispersal in the context of wider biogeographic patterning. It is surely better to have a good idea of what is happening in a subset of all localities than no coherent clue, because throwing up our hands in the face of small sample size without even looking for patterns is counter productive. At the very least we should be able to start to rule out or rule in some options, and again the Gibraltar example serves as a case in point. The foregoing bears on questions of distinguishing “events” from patterns due to uneven sampling in space and time or from noise, including of course taphonomic biases. We emphasise this very point, stressing that inferred dispersal episodes or other evolutionary events such as speciations and extinctions correlated with climatic and habitat changes must be shown to relate logically to the climatic excursions in terms of direction and timing of the movements if we are to get beyond mere ‘shopping lists’ of species at localities. On the question of assessing randomness versus non-randomness in the fossil record 311

O’REGAN et al.: Mammalian dispersal in Plio-Pleistocene Afro-Eurasia

we must acknowledge that pattern recognition is notoriously difficult to achieve through statistical methods, the very reason we stress the logic rather than simply the coincidence of proposed relationships. Such logical relationships are one method of assessing patterning, and that methodology would include the question of logically related time lags between cause and effect and determination of a sequence of events, without which the questions of cause and effect are meaningless. Logical relationships can point to likely responses in the terrestrial biota, such as differing patterns of responses of stenobiomic versus eurybiomic taxa, of cold- versus warmth-loving or tolerant taxa and of grazers versus browsers, to any given pattern or patterns of climatic and habitat change, outlined for example in the “traffic light” model of inter-continental migrations proposed by VRBA (1995a, b). Finally, the problem of assessing randomness and pattern recognition bears directly on the question of analysis itself. As VRBA (1995b) has pointed out, analyses of data to do with evolutionary events and testing of predictions must not assume that the phenomena in question are rigidly identical but recognise the fact that “sub-events” will occur in a complex relationship. When seeking patterns in palaeontological data we must first recognise the limitations of ‘historical’ compared with experimental data, whilst not abandoning the primary purpose of finding order and palaeoecological meaning in broad scale, patchily sampled information.

It is clear to us that setting the search for earliest hominin dispersals within the larger context of movements within the mammalian fauna offers a much better understanding of events and their relationship to changes in the physical environment. In the specific context of movements from Africa, a wider survey of the evidence for contact across the Gibraltar Straits shows that, with the exception of a few very small animals that are capable of rafting, there is no evidence for cross-Gibraltar dispersal. Instead the data indicate long-range dispersal using a circum-Mediterranean route. While we can never exclude the possibility of the odd hominin being washed ashore on the Costa del Sol, we therefore believe that the Gibraltar route can be ruled out as a means of establishing breeding populations of hominins and other large mammals until much later in the Pleistocene when water craft became available.

Acknowledgements Much of the research incorporated in this paper has been carried out within the framework of NERC grant NER/T/ S/2002/00431 awarded to the authors. Site co-ordinates and faunal distributions shown in figs. 2-4 were largely obtained from the Paleobiology Database (www.paleodb.org).

eschweizerbartxxx sng-

Table 1: Summary of mitochondrial DNA studies of animals that are found on both sides of the Gibraltar Straits. Animal

Common name

Crossed the Straits?

Estimated timing of crossing

Method of Dispersal

Direction of crossing

Reference

Mus spretus

Algerian mouse

Yes

Recent or Pliocene

Natural or human introduction

North Africa to Iberia

BOURSOT et al. (1985)

Barbastellus barbastellus

Barbastelle bat

Yes

Plecotus

Long-eared bat

No

Pleurodeles waltl

Greater mouse-eared bat Ribbed salamander

Podarcis spp.

Wall lizard

Myotis myotis

No Yes Yes

?Holocene, frequent crossing But occasional vagrant* But occasional vagrant* Since Messinian Twice – 3.5 Ma and 1.5 Ma

JUSTE et al. (2003)

Flying

Rafting/?human introduction Rafting

CircumMediterranean

JUSTE et al. (2004)

CircumMediterranean

CASTELLA et al. (2000)

North Africa to Iberia Iberia to North Africa

VEITH et al. (2004) HARRIS et al. (2002)

*Occasional vagrant bats are recorded in Europe which have arrived from North Africa, but they do not appear to enter the mitochondrial gene pool in both the Plecotus and Myotis genera.

312

Cour. Forsch.-Inst. Senckenberg, 256, 2006

References AGUIRRE, E., & CARBONELL, E. (2001): Early human expansion into Eurasia: the Atapuerca evidence. — Quaternary International, 75: 11–18; Oxford. ANTÓN, S. C., LEONARD, W. R., & ROBERTSON, M. L. (2002): An ecomorphological model of the initial hominid dispersal from Africa. — Journal of Human Evolution, 43: 773–785; Oxford. ASCENZI, A., BIDDITTU, I., CASSOLI, P. F., SEGRE, A. G., & SEGRE-NALDINI, E. (1996): A calvarium of late Homo erectus from Ceprano, Italy. — Journal of Human Evolution 31: 409–423; Oxford. ASFAW, B., BEYENE, Y., SUWA, G., WALTER, R. C., WHITE, T. D., WOLDEGABRIEL, G., & YEMANE, T. (1992): The earliest Acheulean from Konso-Gardula. — Nature, 360, 732–735; London. ASFAW, B., GILBERT, W. H., BEYENE, Y., HART, W. K., RENNE, P. R., WOLDEGABRIEL, G., VRBA, E. S., & WHITE, T. D. (2002): Remains of Homo erectus from Bouri, Middle Awash, Ethiopia. — Nature, 416: 317–320; London. AZZAROLI, A., DE GIULI, C., FICCARELLI, G., & TORRE, D. (1988): Late Pliocene to Early Mid-Pleistocene mammals in Eurasia: faunal succession and dispersal events. — Palaeogeography, Palaeoclimatology, Palaeoecology, 66: 77–100; Amsterdam. BEHRENSMEYER, A. K., TODD, N. E., POTTS, R., & MCBRINN, G. E. (1997): Late Pliocene faunal turnover in the Turkana Basin, Kenya. — Science, 278: 1589–1594; Washington. BELMAKER, M. (2002): First evidence of the presence of Theropithecus sp. in the southern Levant. — Israel Journal of Zoology, 48: 165; Jerusalem. BELMAKER, M., TCHERNOV, E., CONDEMI, S., & BAR-YOSEF, O. (2002): New evidence for hominid presence in the Lower Pleistocene of the southern Levant. — Journal of Human Evolution, 43: 43–56; Oxford. BERMUDEZ DE CASTRO, M., ARSUAGA, J. L., CARBONELL, E., ROSAS, A., MARTINEZ, I., & MOSQUERA, M. (1997): A hominid from the lower Pleistocene of Atapuerca, Spain: Possible ancestor to Neanderthals and modern humans. — Science, 276: 1392–1395; Washington. BOBÉ, R., BEHRENSMEYER, A. K., & CHAPMAN, R. E. (2002): Faunal change, environmental variability and Late Pliocene hominin evolution. — Journal of Human Evolution, 42: 475–497; Oxford. BOITEL, F., DÉPONT, J., TOURENQ, J., LORENZ, J., ABELANET, J., & POMEROL, J. (1996): Découverte d’une industrie lithique dans le Pliocène supérieur du sud du bassin parisien (Formation des sables et argiles du Bourbonnais). — Comptes Rendus de l’Académie des Sciences, Paris, 322: 507–514; Paris. BOLOMEY, A. (1965): Die Fauna zweier villafrankischer Fundstellen in Rumänien. — Berichte der geologischen Gesellschaft der DDR, 10: 77–88; Berlin. BONIFAY, E., & VANDERMEERSCH, B. (Eds.) (1991): Les Premiers Européens. — 1–319; Paris (Editions du C.T.H.S). BOURSOT, P. JACQUART, T., BONHOMME, F., BRITTON-DAVIDIAN, J., & THALER, L. (1985): Différenciation géographique du génome mitochondrial chez Mus spretus Lataste. — Comptes Rendus de l’Académie des Sciences, Paris, Série III, 301: 161–166; Paris. BOWLER, J. M., JOHNSTON, H., OLLEY, J. M., PRESCOTT, J. R., ROBERTS, R. G., SHAWCROSS, W., & SPOONER, N. A. (2003): New ages for human occupation and climatic change at Lake Mungo, Australia. — Nature, 421: 837–840; London. CASTELLA, V., RUEDI, M., EXCOFFIER, L., IBÁÑEZ, C., ARLETTA, R., & HAUSSER, J. (2000): Is the Gibraltar Strait a barrier to gene flow for the bat Myotis myotis (Chiroptera: Vespertilionidae)? — Molecular Ecology, 9: 1761–1772; Oxford. CLARKE, R. (2001): Out of Africa and back again. — International Journal of Anthropology, 15: 185–189; Firenze. COLLINA-GIRAUD, J. (2001): L’Atlantide devant le détroit de Gibraltar? Mythe et géologie. — Comptes Rendus de l’Académie des Sciences, Sciences de la Terre et des Planètes, 333: 233–240; Paris. DELSON, E. (1993): Theropithecus fossils and taxonomy. — In: JABLONSKI, eschweizerbartxxx sng-

N. G. (Ed.): Theropithecus: The Rise and Fall of a Primate Genus. — 157–189; Cambridge (Cambridge University Press). DENNELL, R. W. (1998): Grasslands, tool making and the hominid colonization of southern Asia: a reconsideration. — In: PETRAGLIA, M., & KORISETTAR, R. (Eds.): Early Human Behaviour in Global Context. — 280–303; London (Routledge). DENNELL, R. (2003): Dispersal and colonisation, long and short chronologies: how continuous is the Early Pleistocene record for hominids outside East Africa? — Journal of Human Evolution, 45: 421–440; Oxford. DENTON, G. (1999): Cenozoic climatic change. — In: BROMAGE, T., & SCHRENK, F. (Eds.): African Biogeography, Climatic Change and Early Hominid Evolution. — 94–114; New York (Oxford University Press). DOBSON, M. (1998): Mammal distributions in the western Mediterranean: the role of human intervention. — Mammal Review, 28 (2): 77–88; Oxford. DOBSON, M., & WRIGHT, A. (2000): Faunal relationships and zoogeographical affinities of mammals from north-west Africa. — Journal of Biogeography, 27: 417–424; Oxford. ELTON, S., BARHAM, L., ANDREWS, P., & SAMBROOK SMITH, G. (2003): Pliocene femur of Theropithecus from the Luangwa Valley, Zambia. — Journal of Human Evolution, 44: 133–140; Oxford. FLEMMING, N. V., BAILEY, G. N., COURTILLOT, V., KING, G., LAMBECK, K., RYERSON, F., & VITA-FINZI, C. (2003): Coastal and marine palaeo-environments and human dispersal across the AfricaEurasia boundary. — In: BREBBIA, C. A., & GAMBIN, T. (Eds.): Maritime Heritage. — 61–74; Southampton (WIT Press). GABUNIA, L., ANTÓN, S. C., LORDKIPANIDZE, D., VEKUA, A., JUSTUS, A., & SWISHER III, C. C. (2001): Dmanisi and dispersal. — Evolutionary Anthropology, 10: 158–170; New York. GERAADS, D. (2002): Plio-Pleistocene mammalian biostratigraphy of Atlantic Morocco. — Quaternaire, 13: 43–53; Paris. GERAADS, D., AMANI, F., RAYNAL, J. P., & SBIHI ALAOUI, F. Z. (1998): La faune de Mammifères du Pliocène terminal d’Ahl al Oughlam, Casablanca, Maroc. — Comptes Rendus de l’Académie des Sciences, Sciences de la Terre et des Planètes, 326: 671–676; Paris. GIBERT, J., RIBOT, F., GIBERT, L., LEAKEY, M., ARRIBAS, A., & MARTÍNEZ, B. (1995): Presence of the Cercopithecid genus Theropithecus in Cueva Victoria (Murcia, Spain). — Journal of Human Evolution, 28: 487–493; Oxford. HARRIS, D. J., CARRANZA, S., ARNOLD, E. N., PINHO, C., & FERRAND, N. (2002): Complex biogeographical distribution of genetic variation within Podarcis wall lizards across the Strait of Gibraltar. — Journal of Biogeography, 29:1257–1262; Oxford. JUSTE, J., IBÁÑEZ, C., MUÑOZ, J., TRUJILLO, D., BENDA, P., KARATAS, A., & RUEDI, M. (2004): Mitochondrial phylogeography of the long-eared bats (Plecotus) in the Mediterranean Palaearctic and Atlantic Islands. — Molecular Phylogenetics and Evolution, 31(3): 1114–1126; Amsterdam. JUSTE, J., IBÁÑEZ, C., TRUJILLO, D., MUÑOZ, J., & RUEDI, M. (2003): Phylogeography of Barbastelle bats (Barbastella barbastellus) in the western Mediterranean and the Canary Islands. — Acta Chiropterologica, 5: 165–175; Warszawa. MARTÍNEZ-NAVARRO, B. (1992): Revision sistematica de la fauna de macromamiferos del yacimiento de Venta Micena (Orce, Granada, Espana). — In: GIBERT, J. (Ed.): Proyecto Orce-Cueva Victoria (1988–1992): Presencia humana en el Pleistoceno inferior de Granada y Murcia. — 21–85; Granada (Ayuntamiento de Orce). MARTÍNEZ-NAVARRO, B., & PALMQVIST, P. (1995): Presence of the African machairodont Megantereon whitei (Broom, 1937) (Mammalia, Carnivora, Felidae) in the Lower Pleistocene site of Venta Micena (Orce, Granada, Spain), with some considerations on the origin, evolution and dispersal of the genus. — Journal of Archaeological Science, 22: 569–582; London. MARTÍNEZ-NAVARRO, B., & PALMQVIST, P. (1996): Presence of the African saber-toothed felid Megantereon whitei (Broom, 1937) (Mammalia, Carnivora, Machairodontinae) in Apollonia-1 (Mygdonia

313

O’REGAN et al.: Mammalian dispersal in Plio-Pleistocene Afro-Eurasia Basin, Macedonia, Greece). — Journal of Archaeological Science, 23: 869–872; London. MARTÍNEZ-NAVARRO, B., TURQ, A., BALLESTER, J. A., & OMS, O. (1997): Fuente Nueva-3 (Orce, Granada, Spain) and the first human occupation of Europe. — Journal of Human Evolution, 33: 611–620; Oxford. MITHEN S., & REED, M. (2002): Stepping out: a computer simulation of hominid dispersal from Africa. — Journal of Human Evolution, 43: 433–462; Oxford. O’REGAN, H. J., BISHOP, L. C., LAMB, A., ELTON, S., & TURNER, A. (2005): Large mammal turnover in Africa and the Levant 1.0–0.5Ma. — In: HEAD, M. J., & GIBBARD, P. (Eds.): Early-Middle Pleistocene Transition: The Land-Ocean Evidence. — 231–249; London (Geological Society). PICKFORD, M. (1993): Climate change, biogeography and Theropithecus. — In: JABLONSKI, N. G. (Ed.): Theropithecus: The Rise and Fall of a Primate Genus. — 227–243; Cambridge (Cambridge University Press). RADULESCO, C., & SAMSON, P. (1990): The Plio-Pleistocene Mammalian Succession of the Oltet Valley, Dacic Basin, Romania. — Quartärpaläontologie, 8: 225–232; Berlin. RADULESCO, C., & SAMSON, P. (2001): Biochronology and evolution of the Early Pliocene to the Early Pleistocene mammalian faunas of Romania. — Bolletino della Societé Paleontologica Italiana, 40: 285–291; Modena. ROEBROEKS, W. (2001): Hominid behaviour and the earliest occupation of Europe: an exploration. — Journal of Human Evolution, 41: 437–461; Oxford. ROLLAND, N. (1998): The lower Palaeolithic settlement of Eurasia, with special reference to Europe. — In: PETRAGLIA, M., & KORISETTAR, R. (Eds.): Early Human Behaviour in Global Context. — 187–220; London (Routledge). ROOK, L., MARTÍNEZ-NAVARRO, B., & HOWELL, F. C. (2004). Occurrence of Theropithecus sp. in the Late Villafranchian of Southern Italy and implication for Early Pleistocene “out of Africa” dispersals. — Journal of Human Evolution, 47: 267–277; Oxford. ROSENZWEIG, M. L. (1995): Species Diversity in Time and Space. — 1–430; Cambridge (Cambridge University Press). SHACKLETON, N. J. (1995): New Data on the Evolution of Pliocene Climatic Variability. — In: VRBA, E. S., DENTON, G. H., PARTRIDGE, T. C., & BURCKLE, L. H. (Eds.): Paleoclimate and Evolution, with Emphasis on Human Origins. — 242–248; New Haven, London (Yale University Press). SHACKLETON, N. J., BACKMAN, J., ZIMMERMAN, H., KENT, D. V., HALL, M. A., ROBERTS, D. G., SCHNITKER, D., BALDAUF, J. G., DESPRAIRIES, A., HOMRIGHAUSEN, R., HUDDLESTUN, P., KEENE, J. B., KALTENBACK, A. J., KRUMSIEK, K. A. O., MORTON, A. C., MURRAY, J. W., & WESTBERG-SMITH, J. (1984): Oxygen isotope calibration of the onset of ice-rafting and history of glaciation in the North Atlantic region. — Nature 307: 620–623; London. TCHERNOV, E. (1992): Eurasian-African biotic exchanges through the Levantine corridor during the Neogene and Quaternary. — Courier Forschungsinstitut Senckenberg, 153: 103–123; Frankfurt a. M. TCHERNOV, E., HORWITZ, L. K., RONEN, A., & LISTER, A. (1994): The faunal remains from Evron Quarry in relation to other lower Paleolithic hominid sites in the southern Levant. — Quaternary Research, 42: 328–339; New York. eschweizerbartxxx sng-

314

TURNER, A. (1987): Megantereon cultridens from Plio-Pleistocene age deposits in Africa and Eurasia, with comments on dispersal and the possibility of a New World origin (Mammalia, Felidae, Machairodontinae). — Journal of Paleontology, 61 (6): 1256–1268; Ithaka. TURNER, A. (1995): Plio-Pleistocene correlations between climatic change and evolution in terrestrial mammals: the 2.5 Ma event in Africa and Europe. — Acta zoologica cracoviensia, 38: 45–58; Kraków. TURNER, A. (1999a): Assessing earliest human settlement of Eurasia: Late Pliocene dispersions from Africa. — Antiquity, 73: 563–570; Oxford. TURNER, A. (1999b): Evolution in the African Plio-Pleistocene mammalian fauna. — In: BROMAGE, T., & SCHRENK, F. (Eds.): African Biogeography, Climatic Change and Early Hominid Evolution. — 369–399; New York (Oxford University Press). TURNER, A., & ANTÓN, M. (1999): Climate and evolution: implications of some extinction patterns in African and European machairodontine cats of the Plio-Pleistocene. — Estudios Geologicos, 54: 209–230; Madrid. TURNER, A., & PATERSON, H. E. H. (1991): Species and speciation: evolutionary tempo and mode in the fossil record reconsidered. — Geobios, 24: 761–769; Lyon. VEITH, M., MAYER, C., SAMRAOUI, B., BARROSO, D. D., & BOGAERTS, S. (2004): From Europe to Africa and vice versa: evidence for multiple intercontinental dispersal in ribbed salamanders (genus Pleurodeles). — Journal of Biogeography, 31: 159–171; Oxford. VEKUA, A., LORDKIPANIDZE, D., RIGHTMIRE, G. P., AGUSTI, J., FERRING, R., MAISURADZE, G., MOUSKHELISHIVILI, A., NIORADZE, M., PONCE DE LEON, M., TAPPEN, M., TVALCHRELIDZE, M., & ZOLLIKOFER, C. (2002): A new skull of early Homo from Dmanisi, Georgia. — Science 297: 85–89; Washington. VRBA, E. S. (1995a): The fossil record of African antelopes (Mammalia, Bovidae) in relation to human evolution and paleoclimate. — In: VRBA, E. S., DENTON, G. H., PARTRIDGE, T. C., & BURCKLE, L. H. (Eds.): Paleoclimate and Evolution, with Emphasis on Human Origins. — 385–424; New Haven, London (Yale University Press). VRBA, E. S. (1995b): On the connections between paleoclimate and evolution. — In: VRBA, E. S., DENTON, G. H., PARTRIDGE, T. C., & BURCKLE, L. H. (Eds.): Paleoclimate and Evolution, with Emphasis on Human Origins. — 24–45; New Haven, London (Yale University Press). VRBA, E. S. (1999): Habitat theory in relation to the evolution in African Neogene biota and hominids. — In: BROMAGE, T., & SCHRENK, F. (Eds.): African Biogeography, Climatic Change and Early Hominid Evolution. — 19–34; New York (Oxford University Press). WHITE, T. D. (1995): African omnivores: Global climatic change and Plio-Pleistocene hominids and suids. — In: VRBA, E. S., DENTON, G. H., PARTRIDGE, T. C., & BURCKLE, L. H. (Eds.): Paleoclimate and Evolution, with Emphasis on Human Origins. — 369–384; New Haven, London (Yale University Press).

Manuscript submitted 2005–10–04 Revised manuscript accepted 2006–07–04