- Email: [email protected]
The Water Optimisation Hypothesis and the human occupation of the mid-latitude belt in the Pleistocene
Quaternary International 300 (2013) 22e31
Contents lists available at SciVerse ScienceDirect
Quaternary International journal homepage: www.elsevier.com/locate/quaint
The Water Optimisation Hypothesis and the human occupation of the mid-latitude belt in the Pleistocene Clive Finlayson a, b, * a b
The Gibraltar Museum, 18-20 Bomb House Lane, Gibraltar Department of Social Sciences, University of Toronto at Scarborough, Canada
a r t i c l e i n f o
a b s t r a c t
Article history: Available online 30 March 2013
A hypothesis e the Water Optimisation Hypothesis e is proposed to predict the occupation of sites by Homo. The hypothesis proposes that Homo occupied intermediate positions in the humidity spectrum. A consequence is that many suitable areas coincided with semi-arid to sub-humid rainfall regimes. Homo should therefore have located its sites in relation to water sources. Analysis of 357 sites occupied by Homo between 200,000 and 10,000 years ago confirms the prediction and reveals no difference in behaviour between Homo sapiens sapiens and the Neanderthals (H. s.neanderthalensis), indicating that the water-attachment of sites is a universal feature of the genus. It is proposed that a belt of similar characteristics, sometimes severed by climate, stretched from south-west Iberia to south-east Australia and was the cauldron of modern human evolution. Geographical expansion of Homo is predicted to have followed river courses and, only occasionally, coastlines. Ó 2013 Elsevier Ltd and INQUA.
1. Introduction The aim of this paper is to establish whether there have been, in the course of the evolution of Homo sapiens and its contemporaries, particular regions of the Earth e hotspots e that have been extensively and repeatedly occupied. If we are able to establish that there have been, then the subsidiary aim is to determine whether these hotspots had features in common. We will then be in a position to ask if those features were of the universal kind (LéviStrauss, 1963) that characterised all humans. The role of water has received relatively little attention (Nicholas, 1998; Finlayson et al., 2008, 2011) in comparison to studies of diet and technology. Yet it is of prime consideration when understanding the survival of humans, particularly in arid tropical and subtropical environments where water is at a premium. For this reason, in this paper I will place particular emphasis on the importance of water as a limiting factor in human dispersion and dispersal. 2. Theoretical framework e the Water Optimisation Hypothesis The working hypothesis behind this paper, which I have termed the Water Optimisation Hypothesis (WOH), is that humans, have * The Gibraltar Museum, 18-20 Bomb House Lane, Gibraltar. E-mail address: jcfi[email protected]. 1040-6182/$ e see front matter Ó 2013 Elsevier Ltd and INQUA. http://dx.doi.org/10.1016/j.quaint.2013.03.040
been constrained by physiological and energetic factors. On a humidity gradient, humans would be best suited to occupying intermediate portions of the gradient, optimality falling on either side but more rapidly towards the humid than the arid end (Fig. 1). On the hyper-arid side of the gradient, humans would have been constrained by physiology in situations of high water deficit and by energetics, where food sources would have been severely limited (Whittaker, 1975). The situation in the hyper-humid end of the gradient would have constrained humans for reasons of food availability and accessibility in situations dominated by dense forests, which characterise areas with high rainfall regimes (Le Houérou, 2009). I suggest that the conditions allowing access to resources would drop more rapidly towards the hyper-humid end as forest closed access up. The broad spectrum of rainfall regimes falling in between the two extremes would have offered ample opportunities for humans. These would have included arid and semi arid portions of the gradient. The further to the left of the xaxis in Fig. 1, the greater the need to have the necessary skills to tap ephemeral, seasonal and highly variable (often hidden) water sources (H. sapiens sapiens; Finlayson, 2004). The further to the right, the greater the need to have skills that allowed for the exploitation of energetic resources in situations of increasing cover (Neanderthals; Finlayson, 2004). I have highlighted sapiens and neanderthalensis because they represent two opposite poles in terms of ability to move over large distances on account of their morphology (Raichlen et al., 2011). Such differences would have
C. Finlayson / Quaternary International 300 (2013) 22e31
23
Fig. 1. The Water Optimisation Hypothesis. Homo is expected to have occupied an optimal zone in a humidity gradient ranging from Eremitic to Hyperhumid. Suitability of sites is expected to drop away from the optimal zone. In the context of this paper the Eremitic category would be defined as having mean annual rainfall under 50 mm and effectively 0 days of rainfall/year and Hyperarid would have 50e100 mm/year and <5 rain days/year. Hyperhumid, at the other end, would be defined as having mean annual rainfall >1500 mm, with >270 rainy days in the year. The optimal zone would be marked by a sub-humid climate having a mean annual rainfall between 600 and 1000 mm and between 150 and 240 days of rain/year. Sub-optimal zones would fall at either end of the optimal zone but it is proposed that the width of the suboptimal zone on the arid side (Semi Arid, rainfall 400e600 mm/year, 120e150 rain days/year; Arid, rainfall 100e400 mm/year, 90e120 rain days/year) would be greater than that on the humid side (Humid, rainfall 1000e 1500 mm/year, 210e270 rain days/year) (Le Houérou, 2009).
been critical in the exploitation of widely scattered water sources. It follows that we would expect H. s. neanderthalensis to have been more susceptible to water shortage than H. s. sapiens. 3. Methods 357 Palaeolithic sites from across the world were selected for this study (Supplementary Table 1), covering the period 200 to 10 ka (thousand years ago), within the period of existence of H. sapiens and its contemporaries and prior to the advent of agriculture. The sites were chosen if they presented evidence of persistent occupation or were sites that revealed intensive occupation at a particular moment. The reason for such choice was to try and identify sites that could be regarded as source populations, in metapopulation biology terminology (Hanski and Gilpin, 1997). When considering archaeological sites it is assumed that human presence at a site indicates that the location was a favourable one but this need not be the case (Finlayson, 2004). A site might have held a sink population, which was maintained by immigration from a source population; in such a scenario this population would have been at risk of extinction. So mixing source with sink sites could obscure, even distort, patterns. Here, I assume that the sites chosen were source populations on the basis of the selection criteria used. A series of variables was recorded for each of the sites: cave/open air site; proximity of site to lake, river, marsh, coast or other wetland (within 5 km); recorded presence of large terrestrial mammals, small terrestrial game, rocky habitat mammals, marine/freshwater mammals, or marine/freshwater small game (Supplementary Table 1). The latter category included molluscs. In the case of coastline, the distance was estimated to the coastline when the site was occupied. 4. Results The sites in Supplementary Table 1 may be conveniently divided into general regions, from which we are able to identify areas with
particularly high densities of sites (Fig. 2). It is recognised that sampling intensity may bias the observed pattern although largely in the direction of poorly represented regions as those with high site density must reflect a high level of human occupation. The important result is that areas of site concentration are identifiable, in keeping with the idea that hotspots existed during the Pleistocene time frame under discussion. Twelve hotspots are identified, accepting that others may emerge in the future. These are: 1. southern Africa; 2. The Levant; 3. south-western Iberia; 4. Central Mediterranean; 5. North-west Iberia/south-west France; 6. CircumAlpine region; 7. Northern Caucasus/Black Sea; 8. southern Siberia; 9. southern Australia; 10. Coastal California; 11. Coastal Peru; and 12 Coastal Chile. The 357 sites were grouped by broad geographical zones. Some zones encompass more than one hotspot while others are single hotspots. Number of sites was a criterion determining whether or not hotspots were clustered. The broad geographical areas were: 1. Eastern Africa, including the Levant (AFR: including hotspots 1 and 2 and sites in between); 2. south-western Iberia (SWIB); 3. Central Mediterranean (CMED); 4. North-west Iberia/south-west France (PYRE); 5. Circum-Alpine region (ALPS); 6. Eurasian Plain (SIBE) including hotspots 7 and 8; 7. Australasia (AUST); 8. Pacific Rim (PAC) including hotspots 10, 11 and 12; and 9. North American Plain (NAME). Were there regional differences in the characteristics of sites occupied by humans? Sites were frequently caves in many areas. The frequency of caves sites was much lower on the Eurasian and North American Plains (Fig. 3) suggesting that human occupation of these vast high latitude regions may have required alternative sources of shelter. The circum-Alpine Region occupied an intermediate position suggesting that independence from caves may have originated here (or in another mountain-plains ecotone) and spread afterwards onto the plains. Coastal sites dominated in southern regions: AUST, AFR, PAC, SWIB and CMED (Fig. 4). The proportion of coastal sites decreased in PYRE and was low in ALPS,
24
C. Finlayson / Quaternary International 300 (2013) 22e31
Fig. 2. Geographical hotspots of Homo occupation: 1. southern Africa; 2. The Levant; 3. south-western Iberia; 4. Central Mediterranean; 5. North-west Iberia/south-west France; 6. Circum-Alpine region; 7. Northern Caucasus/Black Sea; 8. southern Siberia; 9. southern Australia; 10. Coastal California; 11. Coastal Peru; and 12 Coastal Chile.
SIBE and NAME, even though these regions do have coasts. In contrast, the proportion of wetland sites (including rivers, inland marshes and lakes) was highest in NAME, SIBE, ALPS and PYRE (Fig. 5). There was little variation in proportion of sites that incorporated large mammals, other than proportionately lower numbers
in AUST and PAC (Fig. 6). Small terrestrial game sites exceeded 50% in southern sites: AUST, AFR, SWIB and CMED but not PAC. These sites were frequent, but at or below 50%, elsewhere (Fig. 7). Sites with rocky habitat terrestrial mammals occurred where mountainous terrain dominated and therefore reflected habitat. CMED
Fig. 3. Geographical distribution of Homo occupation of cave and open air sites.
C. Finlayson / Quaternary International 300 (2013) 22e31
25
Fig. 4. Geographical distribution of Homo occupation of coastal sites.
and PYRE, and to a lesser degree ALPS and SWIB, stood out as the main regions for rocky terrestrial mammals (Fig. 8). Marine or freshwater mammal sites were never dominant and were always associated with southern regions, being absent from PYRE, ALPS, SIBE and NAME (Fig. 9). Sites with marine or freshwater small game were more frequent than those with marine or freshwater mammals and also predominated in southern over northern sites (Fig. 10).
5. Discussion 5.1. The importance of water in the distribution of Homo The 357 sites analysed were close to either the coast or to wetlands, in a number of cases both (Supplementary Table 1). Inland wetlands were close to human occupation sites in 276 cases (77.3%); the remaining 81 sites (22.7%) were all close to the coast.
Fig. 5. Geographical distribution of Homo occupation of riverine sites. This includes lakes, marshes and other wetlands.
26
C. Finlayson / Quaternary International 300 (2013) 22e31
Fig. 6. Regional association of Homo with large terrestrial mammals.
Such coastal sites would have had a high probability of access to sources of freshwater, particularly in cases of lowered sea levels (which would have been frequent in the Pleistocene) when coastal oases would have dotted the emerged coastal shelf (Faure et al., 2002). These results therefore confirm that human geographical distribution in the Pleistocene was intimately connected to, and
presumably limited by, sources of freshwater. The limitations would have become greater the more arid the situations encountered, promoting increased mobility as observed in sapiens. It follows that geographical expansion would have followed sources of freshwater. The linear nature of rivers and coasts would have made them ideal Leitlinie (leading lines, Geyr von Schweppenburg, 1963).
Fig. 7. Regional association of Homo with small terrestrial game.
C. Finlayson / Quaternary International 300 (2013) 22e31
27
Fig. 8. Regional association of Homo with terrestrial rocky habitat mammals.
Fig. 9. Regional association of Homo with marine and fresh-water mammals.
In reverse, coasts and rivers would have acted as Heligoland Traps (Hunemorder, 1995), that is funnels which would concentrate human populations within refugia. 5.2. General characteristics of clusters If we cluster the various regions, which we have defined previously in accordance to their characteristics (Figs. 3e10), we obtain a dendrogram that indicates their affinities (Fig. 11). Four broad
clusters emerge. On the one hand the northern, largely plains, regions cluster together: ALP and SIBE are very close and they, in turn, cluster with NAME. This cluster is quite distinct from the remaining regions. From the remaining regions, one cluster (intermediate) separates CMED and PYRE from all others while PAC forms a third cluster on its own. This leaves a fourth (southern) cluster, of regions that are geographically distant from each other: SWIB, AUST and AFR (Fig. 11). The characteristics of these particular clusters are summarised in Table 1. They are:
28
C. Finlayson / Quaternary International 300 (2013) 22e31
Table 1 Importance of variables associated with sites by cluster.***¼>66.7% of sites; **¼between 33.3 and 66.7% of sites; *¼<33.3% of sites; e ¼ no sites Cluster
Caves
Coastal
Wetland
Large terrest.
Small terrest.
Rocky
Large marine
Small marine
Northern Intermediate Southern Pacific
*/** *** *** **
* **/*** **/*** ***
*** */*** **/*** *
*** *** *** **
** ** **/*** **
-/** *** -/** e
e e */** *
* * **/*** ***
5.2.1. Northern cluster A dominance of open air sites over cave sites; few coastal sites and correspondingly low representation of marine animals; high representation of inland wetlands and large terrestrial mammals; good representation of small terrestrial fauna and variable representation of rocky mammal fauna, reflecting site variability within the regions of the cluster. 5.2.2. Intermediate cluster A dominance of cave sites over open air sites; the highest representation of rocky terrestrial mammals of all regions, unsurprisingly given the mountainous characteristics of the regions making up the cluster. Coastal site representation is variable, reflecting the inland-coastal nature of the cluster’s regions, with no large marine game sites and only a weak representation of marine/freshwater small game sites; variable presence of inland wetlands and a high representation of large terrestrial mammals and a good representation of small terrestrial game. 5.2.3. Southern cluster A dominance of cave sites over open air sites; high representation of coastal sites with the highest representation of marine animals (large and small) of any of the clusters; high proportion of inland wetlands and terrestrial game (large and small) but variable representation of rocky habitat animals. 5.2.4. Pacific cluster A unique cluster characterised by a very high representation of coastal sites, high in small marine game but low in large
marine game; low representation of inland wetlands and relatively poor representation of terrestrial game (large and small) in comparison to other clusters; no rocky terrestrial mammals represented. 5.2.5. Homo survival strategy in different contexts The four clusters provide contrasts that are suggestive of the Homo survival strategy in different contexts. The southern cluster indicates concentrations of sites around coastal and inland wetland areas, sometimes in close proximity, for example where rivers formed marshes, deltas or estuaries as they entered the sea. The northern cluster, on the other hand, reveals little coastal site concentration in spite of the vast stretches of coastline available. The reason for this may be that human occupation of the northern plains was late in the history of geographical expansion of Homo (Finlayson, 2004) and the high latitude of the coasts would have made them ice-bound for most of the time. The human response on the northern plains appears to have been to concentrate activities around rivers and other inland wetlands. The Pacific cluster is in complete contrast with the others, with the majority of sites concentrated on the coast with few along other wetlands. In terms of the type of site occupied (cave or open air) or food resources exploited, the data suggest that context played a critical role. 5.2.6. The mid-latitude belt, hotspots and human range expansions A surprising result (Fig. 11) is the clustering of SWIB with AFR and AUST and not with geographically closer European regions. This is not so surprising when we consider the characteristics of SWIB
Fig. 10. Regional association of Homo with small marine and fresh-water game.
C. Finlayson / Quaternary International 300 (2013) 22e31
29
Fig. 11. Dendrogram derived from cluster analysis of geographical regions using site variables. Geographical regions: 1. Eastern Africa, including the Levant (AFR: including hotspots 1 and 2 and sites in between); 2. south-western Iberia (SWIB); 3. Central Mediterranean (CMED); 4. North-west Iberia/south-west France (PYRE); 5. Circum-Alpine region (ALPS); 6. Eurasian Plain (SIBE) including hotspots 7 and 8; 7. Australasia (AUST); 8. Pacific Rim (PAC) including hotspots 10, 11 and 12; and 9. North American Plain (NAME).
(Finlayson et al., 2008) e seasonal and inter-annually variable rainfall regimes in the arid to sub-humid range in warm climates. These characteristics would have been shared for long periods across the southern flanks of the Mid-latitude Belt (MLB e Finlayson, 2004) and,
during warmer, earlier, periods of the Pleistocene and Pliocene also along its northern flank (Fig.12). These climates would have provided suitable conditions for hominins across large areas e Savannahstan e of southern Eurasia, the Sahara, Arabia and India (Dennell and
Fig. 12. The southern mid-latitude belt (green) with the northern mid-latitude belt (blue). The connection between south-east Asia and the rest of the southern MLB opened and closed with the contraction and expansion of the rain forest. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
30
C. Finlayson / Quaternary International 300 (2013) 22e31
Roebroeks, 2005). I argue here that these climatic regimes were precisely those predicted as optimal human living conditions under the WOH. It follows that regions with these climatic characteristics were the crucibles of human evolution, and not simply corridors between other regions (Drake et al., 2011). Thus, the WOH predicts that vast areas of the Saharan (van Peer et al., 2003; Rose, 2004), Arabian (Groucutt and Petraglia, 2012) and Indian Deserts (James and Petraglia, 2005; Petraglia et al., 2012) of today would have been important areas within the optimal life zones of Pleistocene humans (e.g. Petraglia et al., 2011). Further east, the connection with Australia would have been sporadic depending on times when the rainforest barriers opened up during dry climatic phases (Bird et al., 2005). Thus, once colonised by humans, biological and cultural changes in Australia would have proceeded in relative isolation. In the Late Pleistocene, south-eastern Australia, South Africa and south-western Iberia would have represented the extremes of this vast world (Fig. 12). While providing optimal conditions (hot spots) for human life, their geographical location, all of them touching the shores of oceans, would have made them cul-de-sacs (the Heligoland Traps referred to earlier in this discussion) in times of stress. It is not surprising that south-western Iberia was a refuge of the Neanderthals (Finlayson et al., 2006) in which water played a prominent role (Finlayson, 2008; Jennings et al., 2011) and that water shortage was a key factor in their extinction (Blain et al., 2013; Jimenez-Espejo et al., 2013). It is unlikely that they were ever epicentres of human dispersal, as has been suggested for South Africa (e.g. Marean, 2011), given their remoteness. The hot spots that led to major human dispersals in the Late Pleistocene are likely to have been within the central parts of this region, in the central and eastern Sahara, north-east Africa, the Arabian Peninsula and northern India. And the main Leitlinie, supported by the results presented here, would have been inland river systems (for example the vast networks in the Sahara; Drake et al., 2011) supplemented by coasts where they occurred and where they provided suitable sources of fresh water. In my view excessive importance has been placed on the coast, on its own, as the prime conductor in the expansion of modern human populations (e.g. Stringer, 2000). While not rejecting their role in particular geographical and temporal contexts, coasts alone do not explain human geographic expansion. 5.3. Universality of features common to human occupation sites One question remains to be answered: were the features that characterised sites of human occupation universal? In other words were they shared by different human taxa or were they specific to H. s. sapiens? To answer this question I compared sites occupied by H. s. neanderthalensis with those occupied by H. s. sapiens (from Supplementary Table 1) within the same geographical area. Comparison of cave and open air occupation revealed proportionately more open air sites in H. s. sapiens compared to H. s. neanderthalensis (c2 1 ¼ 5.181; p < 0.05). The result may indicate the late expansion of H. s. sapiens onto the Eurasian Plain where there would have been few caves (Finlayson, 2004). Comparison of occupation of coastal against inland wetlands revealed no significant difference (c2 1 ¼ 0; p > 0.05). Comparison of the different food categories also indicated no significant difference (c2 4 ¼ 5.776; p > 0.05). These results lead to the conclusion that the features observed in this paper e attachment to water sources, diversity of food categories taken e were universal features that characterised the genus Homo and were not exclusive to particular lineages within the genus. Differences between sites therefore reflected context and not the abilities or preferences of particular lineages.
6. Conclusions This paper confirms the generality of the Homo adaptive strategy and indicates that observed differences between taxa reveal the contexts of the sites occupied and not significant differences in behavioural strategies adopted by the different taxa. The sites studied have been assumed to have been population sources and not sinks, based on a set of criteria; even if some were not sources, the link between Homo and the characteristics of these sites should not be downplayed. H. s. neanderthalensis and H. s. sapiens were effectively identical in behaviour and responded to different contexts in a similar manner. The greater mobility of sapiens compared to neanderthalensis, however, seems to have permitted the former to access Homo-suitable sites where the distribution of water was patchier than normal. At this scale of observation the difference may have allowed sapiens to penetrate deeper into the aridity gradient than neanderthalensis. Crucial to the Homo adaptive strategy, predicted by the WOH, was fresh-water. Homo occupation sites were practically always in semi-arid to sub-humid contexts and close to fresh-water e those that were not, were coastal and, in all likelihood, had access to fresh-water. Thus rivers, lakes, wetlands and coasts acted as focal points for Homo; they were also leading lines in times of geographical expansion and funnel traps in times of geographical recession. An omnivorous strategy characterised Homo, not surprisingly as it would have permitted flexibility in food-type selection within the context of each site. Freshwater would have had the added advantage of concentrating potential prey in times of water shortage. The southern MLB emerges as a vital region in human evolution. Its limits are defined by south-western Iberia (in the north-west), South Africa (in the extreme south) and south-eastern Australia (in the south-east). Finally, it is proposed that the epicentre of Homo evolution was the southern MLB and that areas of the Sahara, Arabia and India e currently underrepresented by archaeological sites e were central to its evolution and expansion. Acknowledgements I am grateful to James Blinkhorn and Huw Groucutt for inviting me to speak at the Oxford conference and to present a paper to this volume. I thank Geraldine Finlayson for discussion of the contents of this paper and to Mike Petraglia and two anonymous referees for helpful comments that improved the manuscript. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.quaint.2013.03.040. References Bird, M.I., Taylor, D., Hunt, C., 2005. Palaeoenvironments of insular Southeast Asia during the Last Glacial Period: a savanna corridor in Sundaland? Quaternary Science Reviews 24, 2228e2242. Blain, H.-A., et al., 2013. Climatic conditions for the last Neanderthals: herpetofaunal record of Gorham’s Cave, Gibraltar. Journal of Human Evolution. http://dx.doi. org/10.1016/j.jhevol.2012.11.003. Dennell, R., Roebroeks, W., 2005. An Asian perspective on early human dispersal from Africa. Nature 438, 1099e1104. Drake, N., et al., 2011. Ancient watercourses and biogeography of the Sahara explain the peopling of the desert. Proceedings, National Academy of Science USA 108, 458e462. Faure, H., Walter, R.C., Grant, D.R., 2002. The coastal oasis: ice age springs on emerged continental shelves. Global and Planetary Change 33, 47e56. Finlayson, C., 2004. Neanderthals and Modern Humans. An Ecological and Evolutionary Perspective. Cambridge University Press, Cambridge. Finlayson, C., 2008. On the importance of coastal areas in the survival of Neanderthal populations during the Late Pleistocene. Quaternary Science Reviews 27, 2246e2252.
C. Finlayson / Quaternary International 300 (2013) 22e31 Finlayson, C., et al., 2006. Late survival of Neanderthals at the southernmost extreme of Europe. Nature 443, 850e853. Finlayson, C., et al., 2011. The Homo habitat niche: using the avian fossil record to depict ecological characteristics of Palaeolithic Eurasian hominins. Quaternary Science Reviews 30, 1525e1532. Finlayson, G., et al., 2008. Dynamics of a thermo-Mediterranean coastal environment e the Coto Doñana National Park. Quaternary Science Reviews 27, 2145e2152. Geyr von Schweppenburg, H.F., 1963. Zur Terminologie und Theorie der Leitlinie. Journal of Ornithology 104, 191e204. Groucutt, H., Petraglia, M.D., 2012. The prehistory of the Arabian Peninsula: deserts, dispersals, and demography. Evolutionary Anthropology 21, 113e125. Hanski, I., Gilpin, M.E. (Eds.), 1997. Metapopulation Biology. Ecology, Genetics, and Evolution. Academic Press, San Diego. Hunemorder, Ch, 1995. Ornithology on the island of Heligoland and the role of the biologische anstalt up to the foundation of the separate “Vogelwarte”. Helgoland Marine Research 49, 125e134. James, H.V.A., Petraglia, M.D., 2005. Modern human origins and the evolution of behavior in the later Pleistocene record of south Asia. Current Anthropology 46, S3eS27. Jennings, R., et al., 2011. Southern Iberia as a refuge for the last Neanderthal populations. Journal of Biogeography 38, 1873e1885. Jimenez-Espejo, F.J., Rodriguez-Vidal, J., Finlayson, C., et al., 2013. Environmental conditions and geomorphologic changes during the MiddleeUpper
31
Paleolithic in the southern Iberian Peninsula. Geomorphology 180-181, 205e216. Le Houérou, H.N., 2009. Bioclimatology and Biogeography of Africa. Springer, Berlin. Lévi-Strauss, C., 1963. Structural Anthropology. Basic Books. Marean, C.W., 2011. Coastal south Africa and the coevolution of the modern human lineage and the coastal adaptation. In: Bicho, N.F., Haws, J.A., Davis, L.G. (Eds.), Trekking the Shore. Changing Coastlines and the Antiquity of Coastal Settlement. Springer, New York, pp. 421e440. Nicholas, G.P., 1998. Wetlands and hunter-gatherers: a global perspective. Current Anthropology 39, 720e731. Petraglia, M., et al., 2011. Middle Paleolithic occupation on a marine isotope stage 5 lakeshore in the Nefud desert, Saudi Arabia. Quaternary Science Reviews 30, 1555e1559. Raichlen, D.A., Armstrong, H., Lieberman, D.E., 2011. Calcaneus length determines running economy: Implications for endurance running performance in modern humans and Neandertals. Journal of Human Evolution 60, 299e308. Rose, J., 2004. The role of the Saharo-Arabian arid belt in the modern human expansion. Act. IV Congress Pen, pp. 57e67. Stringer, C.B., 2000. Palaeoanthropology: coasting out of Africa. Nature 405, 24e26. van Peer, P., et al., 2003. The early to Middle Stone age transition and the emergence of modern human behaviour at site 8-B-11, Sai island, Sudan. Journal of Human Evolution 45, 187e193. Whittaker, R.H., 1975. Communities and Ecosystems. MacMillan, New York.
Contents lists available at SciVerse ScienceDirect
Quaternary International journal homepage: www.elsevier.com/locate/quaint
The Water Optimisation Hypothesis and the human occupation of the mid-latitude belt in the Pleistocene Clive Finlayson a, b, * a b
The Gibraltar Museum, 18-20 Bomb House Lane, Gibraltar Department of Social Sciences, University of Toronto at Scarborough, Canada
a r t i c l e i n f o
a b s t r a c t
Article history: Available online 30 March 2013
A hypothesis e the Water Optimisation Hypothesis e is proposed to predict the occupation of sites by Homo. The hypothesis proposes that Homo occupied intermediate positions in the humidity spectrum. A consequence is that many suitable areas coincided with semi-arid to sub-humid rainfall regimes. Homo should therefore have located its sites in relation to water sources. Analysis of 357 sites occupied by Homo between 200,000 and 10,000 years ago confirms the prediction and reveals no difference in behaviour between Homo sapiens sapiens and the Neanderthals (H. s.neanderthalensis), indicating that the water-attachment of sites is a universal feature of the genus. It is proposed that a belt of similar characteristics, sometimes severed by climate, stretched from south-west Iberia to south-east Australia and was the cauldron of modern human evolution. Geographical expansion of Homo is predicted to have followed river courses and, only occasionally, coastlines. Ó 2013 Elsevier Ltd and INQUA.
1. Introduction The aim of this paper is to establish whether there have been, in the course of the evolution of Homo sapiens and its contemporaries, particular regions of the Earth e hotspots e that have been extensively and repeatedly occupied. If we are able to establish that there have been, then the subsidiary aim is to determine whether these hotspots had features in common. We will then be in a position to ask if those features were of the universal kind (LéviStrauss, 1963) that characterised all humans. The role of water has received relatively little attention (Nicholas, 1998; Finlayson et al., 2008, 2011) in comparison to studies of diet and technology. Yet it is of prime consideration when understanding the survival of humans, particularly in arid tropical and subtropical environments where water is at a premium. For this reason, in this paper I will place particular emphasis on the importance of water as a limiting factor in human dispersion and dispersal. 2. Theoretical framework e the Water Optimisation Hypothesis The working hypothesis behind this paper, which I have termed the Water Optimisation Hypothesis (WOH), is that humans, have * The Gibraltar Museum, 18-20 Bomb House Lane, Gibraltar. E-mail address: jcfi[email protected]. 1040-6182/$ e see front matter Ó 2013 Elsevier Ltd and INQUA. http://dx.doi.org/10.1016/j.quaint.2013.03.040
been constrained by physiological and energetic factors. On a humidity gradient, humans would be best suited to occupying intermediate portions of the gradient, optimality falling on either side but more rapidly towards the humid than the arid end (Fig. 1). On the hyper-arid side of the gradient, humans would have been constrained by physiology in situations of high water deficit and by energetics, where food sources would have been severely limited (Whittaker, 1975). The situation in the hyper-humid end of the gradient would have constrained humans for reasons of food availability and accessibility in situations dominated by dense forests, which characterise areas with high rainfall regimes (Le Houérou, 2009). I suggest that the conditions allowing access to resources would drop more rapidly towards the hyper-humid end as forest closed access up. The broad spectrum of rainfall regimes falling in between the two extremes would have offered ample opportunities for humans. These would have included arid and semi arid portions of the gradient. The further to the left of the xaxis in Fig. 1, the greater the need to have the necessary skills to tap ephemeral, seasonal and highly variable (often hidden) water sources (H. sapiens sapiens; Finlayson, 2004). The further to the right, the greater the need to have skills that allowed for the exploitation of energetic resources in situations of increasing cover (Neanderthals; Finlayson, 2004). I have highlighted sapiens and neanderthalensis because they represent two opposite poles in terms of ability to move over large distances on account of their morphology (Raichlen et al., 2011). Such differences would have
C. Finlayson / Quaternary International 300 (2013) 22e31
23
Fig. 1. The Water Optimisation Hypothesis. Homo is expected to have occupied an optimal zone in a humidity gradient ranging from Eremitic to Hyperhumid. Suitability of sites is expected to drop away from the optimal zone. In the context of this paper the Eremitic category would be defined as having mean annual rainfall under 50 mm and effectively 0 days of rainfall/year and Hyperarid would have 50e100 mm/year and <5 rain days/year. Hyperhumid, at the other end, would be defined as having mean annual rainfall >1500 mm, with >270 rainy days in the year. The optimal zone would be marked by a sub-humid climate having a mean annual rainfall between 600 and 1000 mm and between 150 and 240 days of rain/year. Sub-optimal zones would fall at either end of the optimal zone but it is proposed that the width of the suboptimal zone on the arid side (Semi Arid, rainfall 400e600 mm/year, 120e150 rain days/year; Arid, rainfall 100e400 mm/year, 90e120 rain days/year) would be greater than that on the humid side (Humid, rainfall 1000e 1500 mm/year, 210e270 rain days/year) (Le Houérou, 2009).
been critical in the exploitation of widely scattered water sources. It follows that we would expect H. s. neanderthalensis to have been more susceptible to water shortage than H. s. sapiens. 3. Methods 357 Palaeolithic sites from across the world were selected for this study (Supplementary Table 1), covering the period 200 to 10 ka (thousand years ago), within the period of existence of H. sapiens and its contemporaries and prior to the advent of agriculture. The sites were chosen if they presented evidence of persistent occupation or were sites that revealed intensive occupation at a particular moment. The reason for such choice was to try and identify sites that could be regarded as source populations, in metapopulation biology terminology (Hanski and Gilpin, 1997). When considering archaeological sites it is assumed that human presence at a site indicates that the location was a favourable one but this need not be the case (Finlayson, 2004). A site might have held a sink population, which was maintained by immigration from a source population; in such a scenario this population would have been at risk of extinction. So mixing source with sink sites could obscure, even distort, patterns. Here, I assume that the sites chosen were source populations on the basis of the selection criteria used. A series of variables was recorded for each of the sites: cave/open air site; proximity of site to lake, river, marsh, coast or other wetland (within 5 km); recorded presence of large terrestrial mammals, small terrestrial game, rocky habitat mammals, marine/freshwater mammals, or marine/freshwater small game (Supplementary Table 1). The latter category included molluscs. In the case of coastline, the distance was estimated to the coastline when the site was occupied. 4. Results The sites in Supplementary Table 1 may be conveniently divided into general regions, from which we are able to identify areas with
particularly high densities of sites (Fig. 2). It is recognised that sampling intensity may bias the observed pattern although largely in the direction of poorly represented regions as those with high site density must reflect a high level of human occupation. The important result is that areas of site concentration are identifiable, in keeping with the idea that hotspots existed during the Pleistocene time frame under discussion. Twelve hotspots are identified, accepting that others may emerge in the future. These are: 1. southern Africa; 2. The Levant; 3. south-western Iberia; 4. Central Mediterranean; 5. North-west Iberia/south-west France; 6. CircumAlpine region; 7. Northern Caucasus/Black Sea; 8. southern Siberia; 9. southern Australia; 10. Coastal California; 11. Coastal Peru; and 12 Coastal Chile. The 357 sites were grouped by broad geographical zones. Some zones encompass more than one hotspot while others are single hotspots. Number of sites was a criterion determining whether or not hotspots were clustered. The broad geographical areas were: 1. Eastern Africa, including the Levant (AFR: including hotspots 1 and 2 and sites in between); 2. south-western Iberia (SWIB); 3. Central Mediterranean (CMED); 4. North-west Iberia/south-west France (PYRE); 5. Circum-Alpine region (ALPS); 6. Eurasian Plain (SIBE) including hotspots 7 and 8; 7. Australasia (AUST); 8. Pacific Rim (PAC) including hotspots 10, 11 and 12; and 9. North American Plain (NAME). Were there regional differences in the characteristics of sites occupied by humans? Sites were frequently caves in many areas. The frequency of caves sites was much lower on the Eurasian and North American Plains (Fig. 3) suggesting that human occupation of these vast high latitude regions may have required alternative sources of shelter. The circum-Alpine Region occupied an intermediate position suggesting that independence from caves may have originated here (or in another mountain-plains ecotone) and spread afterwards onto the plains. Coastal sites dominated in southern regions: AUST, AFR, PAC, SWIB and CMED (Fig. 4). The proportion of coastal sites decreased in PYRE and was low in ALPS,
24
C. Finlayson / Quaternary International 300 (2013) 22e31
Fig. 2. Geographical hotspots of Homo occupation: 1. southern Africa; 2. The Levant; 3. south-western Iberia; 4. Central Mediterranean; 5. North-west Iberia/south-west France; 6. Circum-Alpine region; 7. Northern Caucasus/Black Sea; 8. southern Siberia; 9. southern Australia; 10. Coastal California; 11. Coastal Peru; and 12 Coastal Chile.
SIBE and NAME, even though these regions do have coasts. In contrast, the proportion of wetland sites (including rivers, inland marshes and lakes) was highest in NAME, SIBE, ALPS and PYRE (Fig. 5). There was little variation in proportion of sites that incorporated large mammals, other than proportionately lower numbers
in AUST and PAC (Fig. 6). Small terrestrial game sites exceeded 50% in southern sites: AUST, AFR, SWIB and CMED but not PAC. These sites were frequent, but at or below 50%, elsewhere (Fig. 7). Sites with rocky habitat terrestrial mammals occurred where mountainous terrain dominated and therefore reflected habitat. CMED
Fig. 3. Geographical distribution of Homo occupation of cave and open air sites.
C. Finlayson / Quaternary International 300 (2013) 22e31
25
Fig. 4. Geographical distribution of Homo occupation of coastal sites.
and PYRE, and to a lesser degree ALPS and SWIB, stood out as the main regions for rocky terrestrial mammals (Fig. 8). Marine or freshwater mammal sites were never dominant and were always associated with southern regions, being absent from PYRE, ALPS, SIBE and NAME (Fig. 9). Sites with marine or freshwater small game were more frequent than those with marine or freshwater mammals and also predominated in southern over northern sites (Fig. 10).
5. Discussion 5.1. The importance of water in the distribution of Homo The 357 sites analysed were close to either the coast or to wetlands, in a number of cases both (Supplementary Table 1). Inland wetlands were close to human occupation sites in 276 cases (77.3%); the remaining 81 sites (22.7%) were all close to the coast.
Fig. 5. Geographical distribution of Homo occupation of riverine sites. This includes lakes, marshes and other wetlands.
26
C. Finlayson / Quaternary International 300 (2013) 22e31
Fig. 6. Regional association of Homo with large terrestrial mammals.
Such coastal sites would have had a high probability of access to sources of freshwater, particularly in cases of lowered sea levels (which would have been frequent in the Pleistocene) when coastal oases would have dotted the emerged coastal shelf (Faure et al., 2002). These results therefore confirm that human geographical distribution in the Pleistocene was intimately connected to, and
presumably limited by, sources of freshwater. The limitations would have become greater the more arid the situations encountered, promoting increased mobility as observed in sapiens. It follows that geographical expansion would have followed sources of freshwater. The linear nature of rivers and coasts would have made them ideal Leitlinie (leading lines, Geyr von Schweppenburg, 1963).
Fig. 7. Regional association of Homo with small terrestrial game.
C. Finlayson / Quaternary International 300 (2013) 22e31
27
Fig. 8. Regional association of Homo with terrestrial rocky habitat mammals.
Fig. 9. Regional association of Homo with marine and fresh-water mammals.
In reverse, coasts and rivers would have acted as Heligoland Traps (Hunemorder, 1995), that is funnels which would concentrate human populations within refugia. 5.2. General characteristics of clusters If we cluster the various regions, which we have defined previously in accordance to their characteristics (Figs. 3e10), we obtain a dendrogram that indicates their affinities (Fig. 11). Four broad
clusters emerge. On the one hand the northern, largely plains, regions cluster together: ALP and SIBE are very close and they, in turn, cluster with NAME. This cluster is quite distinct from the remaining regions. From the remaining regions, one cluster (intermediate) separates CMED and PYRE from all others while PAC forms a third cluster on its own. This leaves a fourth (southern) cluster, of regions that are geographically distant from each other: SWIB, AUST and AFR (Fig. 11). The characteristics of these particular clusters are summarised in Table 1. They are:
28
C. Finlayson / Quaternary International 300 (2013) 22e31
Table 1 Importance of variables associated with sites by cluster.***¼>66.7% of sites; **¼between 33.3 and 66.7% of sites; *¼<33.3% of sites; e ¼ no sites Cluster
Caves
Coastal
Wetland
Large terrest.
Small terrest.
Rocky
Large marine
Small marine
Northern Intermediate Southern Pacific
*/** *** *** **
* **/*** **/*** ***
*** */*** **/*** *
*** *** *** **
** ** **/*** **
-/** *** -/** e
e e */** *
* * **/*** ***
5.2.1. Northern cluster A dominance of open air sites over cave sites; few coastal sites and correspondingly low representation of marine animals; high representation of inland wetlands and large terrestrial mammals; good representation of small terrestrial fauna and variable representation of rocky mammal fauna, reflecting site variability within the regions of the cluster. 5.2.2. Intermediate cluster A dominance of cave sites over open air sites; the highest representation of rocky terrestrial mammals of all regions, unsurprisingly given the mountainous characteristics of the regions making up the cluster. Coastal site representation is variable, reflecting the inland-coastal nature of the cluster’s regions, with no large marine game sites and only a weak representation of marine/freshwater small game sites; variable presence of inland wetlands and a high representation of large terrestrial mammals and a good representation of small terrestrial game. 5.2.3. Southern cluster A dominance of cave sites over open air sites; high representation of coastal sites with the highest representation of marine animals (large and small) of any of the clusters; high proportion of inland wetlands and terrestrial game (large and small) but variable representation of rocky habitat animals. 5.2.4. Pacific cluster A unique cluster characterised by a very high representation of coastal sites, high in small marine game but low in large
marine game; low representation of inland wetlands and relatively poor representation of terrestrial game (large and small) in comparison to other clusters; no rocky terrestrial mammals represented. 5.2.5. Homo survival strategy in different contexts The four clusters provide contrasts that are suggestive of the Homo survival strategy in different contexts. The southern cluster indicates concentrations of sites around coastal and inland wetland areas, sometimes in close proximity, for example where rivers formed marshes, deltas or estuaries as they entered the sea. The northern cluster, on the other hand, reveals little coastal site concentration in spite of the vast stretches of coastline available. The reason for this may be that human occupation of the northern plains was late in the history of geographical expansion of Homo (Finlayson, 2004) and the high latitude of the coasts would have made them ice-bound for most of the time. The human response on the northern plains appears to have been to concentrate activities around rivers and other inland wetlands. The Pacific cluster is in complete contrast with the others, with the majority of sites concentrated on the coast with few along other wetlands. In terms of the type of site occupied (cave or open air) or food resources exploited, the data suggest that context played a critical role. 5.2.6. The mid-latitude belt, hotspots and human range expansions A surprising result (Fig. 11) is the clustering of SWIB with AFR and AUST and not with geographically closer European regions. This is not so surprising when we consider the characteristics of SWIB
Fig. 10. Regional association of Homo with small marine and fresh-water game.
C. Finlayson / Quaternary International 300 (2013) 22e31
29
Fig. 11. Dendrogram derived from cluster analysis of geographical regions using site variables. Geographical regions: 1. Eastern Africa, including the Levant (AFR: including hotspots 1 and 2 and sites in between); 2. south-western Iberia (SWIB); 3. Central Mediterranean (CMED); 4. North-west Iberia/south-west France (PYRE); 5. Circum-Alpine region (ALPS); 6. Eurasian Plain (SIBE) including hotspots 7 and 8; 7. Australasia (AUST); 8. Pacific Rim (PAC) including hotspots 10, 11 and 12; and 9. North American Plain (NAME).
(Finlayson et al., 2008) e seasonal and inter-annually variable rainfall regimes in the arid to sub-humid range in warm climates. These characteristics would have been shared for long periods across the southern flanks of the Mid-latitude Belt (MLB e Finlayson, 2004) and,
during warmer, earlier, periods of the Pleistocene and Pliocene also along its northern flank (Fig.12). These climates would have provided suitable conditions for hominins across large areas e Savannahstan e of southern Eurasia, the Sahara, Arabia and India (Dennell and
Fig. 12. The southern mid-latitude belt (green) with the northern mid-latitude belt (blue). The connection between south-east Asia and the rest of the southern MLB opened and closed with the contraction and expansion of the rain forest. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
30
C. Finlayson / Quaternary International 300 (2013) 22e31
Roebroeks, 2005). I argue here that these climatic regimes were precisely those predicted as optimal human living conditions under the WOH. It follows that regions with these climatic characteristics were the crucibles of human evolution, and not simply corridors between other regions (Drake et al., 2011). Thus, the WOH predicts that vast areas of the Saharan (van Peer et al., 2003; Rose, 2004), Arabian (Groucutt and Petraglia, 2012) and Indian Deserts (James and Petraglia, 2005; Petraglia et al., 2012) of today would have been important areas within the optimal life zones of Pleistocene humans (e.g. Petraglia et al., 2011). Further east, the connection with Australia would have been sporadic depending on times when the rainforest barriers opened up during dry climatic phases (Bird et al., 2005). Thus, once colonised by humans, biological and cultural changes in Australia would have proceeded in relative isolation. In the Late Pleistocene, south-eastern Australia, South Africa and south-western Iberia would have represented the extremes of this vast world (Fig. 12). While providing optimal conditions (hot spots) for human life, their geographical location, all of them touching the shores of oceans, would have made them cul-de-sacs (the Heligoland Traps referred to earlier in this discussion) in times of stress. It is not surprising that south-western Iberia was a refuge of the Neanderthals (Finlayson et al., 2006) in which water played a prominent role (Finlayson, 2008; Jennings et al., 2011) and that water shortage was a key factor in their extinction (Blain et al., 2013; Jimenez-Espejo et al., 2013). It is unlikely that they were ever epicentres of human dispersal, as has been suggested for South Africa (e.g. Marean, 2011), given their remoteness. The hot spots that led to major human dispersals in the Late Pleistocene are likely to have been within the central parts of this region, in the central and eastern Sahara, north-east Africa, the Arabian Peninsula and northern India. And the main Leitlinie, supported by the results presented here, would have been inland river systems (for example the vast networks in the Sahara; Drake et al., 2011) supplemented by coasts where they occurred and where they provided suitable sources of fresh water. In my view excessive importance has been placed on the coast, on its own, as the prime conductor in the expansion of modern human populations (e.g. Stringer, 2000). While not rejecting their role in particular geographical and temporal contexts, coasts alone do not explain human geographic expansion. 5.3. Universality of features common to human occupation sites One question remains to be answered: were the features that characterised sites of human occupation universal? In other words were they shared by different human taxa or were they specific to H. s. sapiens? To answer this question I compared sites occupied by H. s. neanderthalensis with those occupied by H. s. sapiens (from Supplementary Table 1) within the same geographical area. Comparison of cave and open air occupation revealed proportionately more open air sites in H. s. sapiens compared to H. s. neanderthalensis (c2 1 ¼ 5.181; p < 0.05). The result may indicate the late expansion of H. s. sapiens onto the Eurasian Plain where there would have been few caves (Finlayson, 2004). Comparison of occupation of coastal against inland wetlands revealed no significant difference (c2 1 ¼ 0; p > 0.05). Comparison of the different food categories also indicated no significant difference (c2 4 ¼ 5.776; p > 0.05). These results lead to the conclusion that the features observed in this paper e attachment to water sources, diversity of food categories taken e were universal features that characterised the genus Homo and were not exclusive to particular lineages within the genus. Differences between sites therefore reflected context and not the abilities or preferences of particular lineages.
6. Conclusions This paper confirms the generality of the Homo adaptive strategy and indicates that observed differences between taxa reveal the contexts of the sites occupied and not significant differences in behavioural strategies adopted by the different taxa. The sites studied have been assumed to have been population sources and not sinks, based on a set of criteria; even if some were not sources, the link between Homo and the characteristics of these sites should not be downplayed. H. s. neanderthalensis and H. s. sapiens were effectively identical in behaviour and responded to different contexts in a similar manner. The greater mobility of sapiens compared to neanderthalensis, however, seems to have permitted the former to access Homo-suitable sites where the distribution of water was patchier than normal. At this scale of observation the difference may have allowed sapiens to penetrate deeper into the aridity gradient than neanderthalensis. Crucial to the Homo adaptive strategy, predicted by the WOH, was fresh-water. Homo occupation sites were practically always in semi-arid to sub-humid contexts and close to fresh-water e those that were not, were coastal and, in all likelihood, had access to fresh-water. Thus rivers, lakes, wetlands and coasts acted as focal points for Homo; they were also leading lines in times of geographical expansion and funnel traps in times of geographical recession. An omnivorous strategy characterised Homo, not surprisingly as it would have permitted flexibility in food-type selection within the context of each site. Freshwater would have had the added advantage of concentrating potential prey in times of water shortage. The southern MLB emerges as a vital region in human evolution. Its limits are defined by south-western Iberia (in the north-west), South Africa (in the extreme south) and south-eastern Australia (in the south-east). Finally, it is proposed that the epicentre of Homo evolution was the southern MLB and that areas of the Sahara, Arabia and India e currently underrepresented by archaeological sites e were central to its evolution and expansion. Acknowledgements I am grateful to James Blinkhorn and Huw Groucutt for inviting me to speak at the Oxford conference and to present a paper to this volume. I thank Geraldine Finlayson for discussion of the contents of this paper and to Mike Petraglia and two anonymous referees for helpful comments that improved the manuscript. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.quaint.2013.03.040. References Bird, M.I., Taylor, D., Hunt, C., 2005. Palaeoenvironments of insular Southeast Asia during the Last Glacial Period: a savanna corridor in Sundaland? Quaternary Science Reviews 24, 2228e2242. Blain, H.-A., et al., 2013. Climatic conditions for the last Neanderthals: herpetofaunal record of Gorham’s Cave, Gibraltar. Journal of Human Evolution. http://dx.doi. org/10.1016/j.jhevol.2012.11.003. Dennell, R., Roebroeks, W., 2005. An Asian perspective on early human dispersal from Africa. Nature 438, 1099e1104. Drake, N., et al., 2011. Ancient watercourses and biogeography of the Sahara explain the peopling of the desert. Proceedings, National Academy of Science USA 108, 458e462. Faure, H., Walter, R.C., Grant, D.R., 2002. The coastal oasis: ice age springs on emerged continental shelves. Global and Planetary Change 33, 47e56. Finlayson, C., 2004. Neanderthals and Modern Humans. An Ecological and Evolutionary Perspective. Cambridge University Press, Cambridge. Finlayson, C., 2008. On the importance of coastal areas in the survival of Neanderthal populations during the Late Pleistocene. Quaternary Science Reviews 27, 2246e2252.
C. Finlayson / Quaternary International 300 (2013) 22e31 Finlayson, C., et al., 2006. Late survival of Neanderthals at the southernmost extreme of Europe. Nature 443, 850e853. Finlayson, C., et al., 2011. The Homo habitat niche: using the avian fossil record to depict ecological characteristics of Palaeolithic Eurasian hominins. Quaternary Science Reviews 30, 1525e1532. Finlayson, G., et al., 2008. Dynamics of a thermo-Mediterranean coastal environment e the Coto Doñana National Park. Quaternary Science Reviews 27, 2145e2152. Geyr von Schweppenburg, H.F., 1963. Zur Terminologie und Theorie der Leitlinie. Journal of Ornithology 104, 191e204. Groucutt, H., Petraglia, M.D., 2012. The prehistory of the Arabian Peninsula: deserts, dispersals, and demography. Evolutionary Anthropology 21, 113e125. Hanski, I., Gilpin, M.E. (Eds.), 1997. Metapopulation Biology. Ecology, Genetics, and Evolution. Academic Press, San Diego. Hunemorder, Ch, 1995. Ornithology on the island of Heligoland and the role of the biologische anstalt up to the foundation of the separate “Vogelwarte”. Helgoland Marine Research 49, 125e134. James, H.V.A., Petraglia, M.D., 2005. Modern human origins and the evolution of behavior in the later Pleistocene record of south Asia. Current Anthropology 46, S3eS27. Jennings, R., et al., 2011. Southern Iberia as a refuge for the last Neanderthal populations. Journal of Biogeography 38, 1873e1885. Jimenez-Espejo, F.J., Rodriguez-Vidal, J., Finlayson, C., et al., 2013. Environmental conditions and geomorphologic changes during the MiddleeUpper
31
Paleolithic in the southern Iberian Peninsula. Geomorphology 180-181, 205e216. Le Houérou, H.N., 2009. Bioclimatology and Biogeography of Africa. Springer, Berlin. Lévi-Strauss, C., 1963. Structural Anthropology. Basic Books. Marean, C.W., 2011. Coastal south Africa and the coevolution of the modern human lineage and the coastal adaptation. In: Bicho, N.F., Haws, J.A., Davis, L.G. (Eds.), Trekking the Shore. Changing Coastlines and the Antiquity of Coastal Settlement. Springer, New York, pp. 421e440. Nicholas, G.P., 1998. Wetlands and hunter-gatherers: a global perspective. Current Anthropology 39, 720e731. Petraglia, M., et al., 2011. Middle Paleolithic occupation on a marine isotope stage 5 lakeshore in the Nefud desert, Saudi Arabia. Quaternary Science Reviews 30, 1555e1559. Raichlen, D.A., Armstrong, H., Lieberman, D.E., 2011. Calcaneus length determines running economy: Implications for endurance running performance in modern humans and Neandertals. Journal of Human Evolution 60, 299e308. Rose, J., 2004. The role of the Saharo-Arabian arid belt in the modern human expansion. Act. IV Congress Pen, pp. 57e67. Stringer, C.B., 2000. Palaeoanthropology: coasting out of Africa. Nature 405, 24e26. van Peer, P., et al., 2003. The early to Middle Stone age transition and the emergence of modern human behaviour at site 8-B-11, Sai island, Sudan. Journal of Human Evolution 45, 187e193. Whittaker, R.H., 1975. Communities and Ecosystems. MacMillan, New York.