The Journal of Island and Coastal Archaeologypeople.rses.anu.edu.au/lambeck_k/pdf/264.pdf · The...

This article was downloaded by:[Australian National University Library]On: 8 November 2007Access Details: [subscription number 773444841]Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

The Journal of Island and CoastalArchaeologyPublication details, including instructions for authors and subscription information:http://www.informaworld.com/smpp/title~content=t716100767

Coastlines, Submerged Landscapes, and HumanEvolution: The Red Sea Basin and the Farasan IslandsGeoff N. Bailey a; Nic C. Flemming b; Geoffrey C. P. King c; Kurt Lambeck d;Garry Momber ae; Lawrence J. Moran a; Abdullah Al-Sharekh f; Claudio Vita-Finzi ga Department of Archaeology, University of York, York, UKb National Oceanography Centre, University of Southampton Waterfront Campus,Southampton, UKc Laboratoire Tectonique, Institut de Physique du Globe, Paris, Franced Research School of Earth Sciences, The Australian National University, Canberra,Australiae Hampshire and Wight Trust for Maritime Archaeology, National OceanographyCentre, Southampton, UK

f Department of Archaeology, College of Tourism and Archaeology, King Saud University, Riyadh, Saudi Arabiag The Natural History Museum, London, UK

Online Publication Date: 01 July 2007To cite this Article: Bailey, Geoff N., Flemming, Nic C., King, Geoffrey C. P., Lambeck, Kurt, Momber, Garry, Moran,Lawrence J., Al-Sharekh, Abdullah and Vita-Finzi, Claudio (2007) 'Coastlines, Submerged Landscapes, and HumanEvolution: The Red Sea Basin and the Farasan Islands', The Journal of Island and Coastal Archaeology, 2:2, 127 - 160To link to this article: DOI: 10.1080/15564890701623449URL: http://dx.doi.org/10.1080/15564890701623449

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf

This article maybe used for research, teaching and private study purposes. Any substantial or systematic reproduction,re-distribution, re-selling, loan or sub-licensing, systematic supply or distribution in any form to anyone is expresslyforbidden.

The publisher does not give any warranty express or implied or make any representation that the contents will becomplete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should beindependently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings,demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with orarising out of the use of this material.

http://www.informaworld.com/smpp/title~content=t716100767

http://dx.doi.org/10.1080/15564890701623449

http://www.informaworld.com/terms-and-conditions-of-access.pdf

Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7

Journal of Island & Coastal Archaeology, 2:127–160, 2007Copyright © 2007 Taylor & Francis Group, LLCISSN: 1556-4894 print / 1556-1828 onlineDOI:10.1080/15564890701623449

ARTICLES

Coastlines, SubmergedLandscapes, and HumanEvolution: The Red SeaBasin and the FarasanIslandsGeoff N. Bailey,1 Nic C. Flemming,2 Geoffrey C. P. King,3

Kurt Lambeck,4 Garry Momber,1, 5 Lawrence J. Moran,1

Abdullah Al-Sharekh,6 and Claudio Vita-Finzi71Department of Archaeology, University of York, York, UK2National Oceanography Centre, University of Southampton Waterfront

Campus, Southampton, UK3Laboratoire Tectonique, Institut de Physique du Globe, Paris, France4Research School of Earth Sciences, The Australian National University,

Canberra, Australia5Hampshire and Wight Trust for Maritime Archaeology, National

Oceanography Centre, Southampton, UK6Department of Archaeology, College of Tourism and Archaeology, King Saud

University, Riyadh, Saudi Arabia7The Natural History Museum, London, UK

ABSTRACT

We examine some long-standing assumptions about the earlyuse of coastlines and marine resources and their contribution tothe pattern of early human dispersal, and focus on the southernRed Sea Basin and the proposed southern corridor of movement

Received 8 May 2007; accepted 6 June 2007.Address correspondence to Geoff N. Bailey, Department of Archaeology, King’s Manor, University ofYork, York YO1 7EP, UK. E-mail: [email protected]

127

Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7

Geoff N. Bailey et al.

between Africa and Arabia across the Bab al-Mandab Straits. Wereconstruct relative sea levels in light of isostatic and tectoniceffects, and evaluate their paleogeographical impact on thedistribution of resources and human movement. We concludethat the crossing of the Bab al-Mandab posed little significantor long-lasting physical or climatic barrier to human transitduring the Pleistocene and that the emerged continental shelfduring periods of low sea level enhanced the possibilities forhuman settlement and dispersal around the coastlines of theArabian Peninsula. We emphasize the paleogeographical andpaleoenvironmental significance of Pleistocene sea-level changeand its relationship with changes in paleoclimate, and identifythe exploration of the submerged continental shelf as a highpriority for future research. We conclude with a brief descriptionof our strategy for underwater work in the Farasan Islands andour preliminary results.

Keywords paleoclimate, sea-level change, shell mounds, tectonics, underwater archaeol-ogy

INTRODUCTION

The idea that coastal environments andmarine resources were an importantfactor in early human development hashad a long and fitful history. Over ahundred years ago, Lewis Henry Morganprophetically remarked on the univer-sal distribution of such resources, theirindependence from climatic variation,and their value in facilitating humanexpansion beyond the boundaries ofa primordial terrestrial habitat (Morgan1877:21). Carl Sauer (1962) eloquentlyelaborated the theme and saw the at-tractions of seacoasts as significant bothin human origins and subsequent dis-persal. Sir Alister Hardy, inspired bySauer, lent notoriety to the idea byarguing for a biological phase of spe-cialized aquatic adaptation as a nec-essary condition of emergent tool-use,meat-eating and bipedalism by hominins(Hardy 1960; see also Morgan 1982).Paleoanthropologists have rejected sucha notion on the grounds that the paleon-tological, archaeological, and genetic ev-

idence argues for a straightforward lineof descent from plant-eating arborealapes to meat-eating, ground-dwellinghumans. Archaeologists, in their turn,have tended to discount coastlines andmarine resources in overviews of worldprehistory (cf. Gamble 1993; Klein1999).

In recent years, there has been arevival of interest in the deeper antiquityof coastal adaptations, resulting from avariety of mostly indirect considerations.General assessments of the ecology andrelative attractiveness of coastal andaquatic resources and increased finds ofmollusc shells and other aquatic food re-mains in Paleolithic deposits (Erlandson2001), simulation of dispersal patterns(Mithen 2002), modelling of genetic his-tories (Macaulay et al. 2005; Templeton2002), and recognition of the over-whelming bias against survival of coastalevidence because of sea-level changeand coastal erosion, have persuadedmany that exploitation of coastal envi-ronments and marine resources couldhave been an attractive option from the

128 VOLUME 2 • ISSUE 2 • 2007

Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7

Red Sea Basin and Farasan Islands

earliest period, and that coastlines andseaways may have acted as gatewaysrather than barriers to human movement(Bailey 2004a, 2004b; Bailey and Milner2002; Erlandson 2001; Erlandson andFitzpatrick 2006; Flemming et al. 2003;Mannino and Thomas 2002; Stringer2000; Turner and O’Regan 2007; Walteret al. 2000).

Evaluation of the coastal factor isof particular importance in the largerpattern of human dispersal because theinterface between the major continentsis primarily a marine and coastal one.Between Africa and Eurasia there isonly one certain land corridor that haspersisted throughout the past 3 millionyears across the whole 5000 km arcfrom Morocco to Somalia, and that isthe Sinai Peninsula. By convention thishas been treated as the only pathwayout of Africa, and the notion of a singlenarrow exit from Africa, intrinsicallyvulnerable to blockage by physical, cli-matic or ecological factors, also providesan attractive geographical rationale forbottlenecks in patterns of biological andcultural evolution (Derricourt 2005). Asea channel at least 10 km wide persistedduring the period under review at theGibraltar Straits in the west. Whetherthe Bab al-Mandab Straits at the southernend of the Red Sea, currently 29 kmwide, remained open throughout theQuaternary is less certain, but variationsin salinity as recorded in the isotopicrecord of deep-sea marine cores suggestthat the Red Sea remained narrowlyopen to exchange of seawater with theIndian Ocean throughout at least thelast 450,000 years and perhaps furtherback in time (Siddall et al. 2003). Hence,any broadening of the zone of contactbetween Africa and Eurasia would haverequired sea crossings by swimming,rafting or use of boats (Flemming et al.2003). Between Asia and North Amer-ica, a land passage across the Bering

Strait and southwards beyond the NorthAmerican ice sheet was available onlyintermittently (Mandryk et al. 2001),while movement from Southeast Asiainto Australia and New Guinea wouldalways have involved sea crossings ofsome sort (Allen 1989; Allen and Hold-away 1995; Lourandos 1997; Robertset al. 1994).

In all these examples a case has beenmade, or can be made, for the signif-icance of coastlines, marine resources,and water crossings in the pattern of dis-persal (Bowdler 1977, 1990; Erlandson2002; Stringer 2000). The possibilitythat maritime seal-hunters might havefollowed the edge of the pack ice acrossthe North Atlantic from the IberianPeninsula to the eastern seaboard of theAmericas during the Last Glacial Maxi-mum has also been proposed (Stanfordand Bradley 2002), but the plausibility ofsuch a dispersal across more than a thou-sand kilometers and the archaeologicalevidence cited in its support have beenvigorously contested (Straus et al. 2005).Some technological capacity to make seacrossings was certainly present in someparts of the world after 50,000 years ago,given the evidence for the colonizationof Australia, but how much furtherback in time such skills can be tracedremains highly uncertain. The presenceof early artefacts dated at ca. 800,000years on the island of Flores in theIndonesian archipelago (Bednarik 2003;Morwood et al. 1999) and of claimedLower and Middle Paleolithic materialon Mediterranean islands (Cherry 1990;Vita-Finzi 1973) provides hints of earliersea crossings, but the distances involvedare mostly small (∼10–20 km), the pre-cise paleogeographical configuration ofpaleoshorelines rarely attempted in lightof local isostatic and tectonic effects (butsee Lambeck 1996a for the Aegean), andthe dates of the Mediterranean artifactsunconfirmed.

JOURNAL OF ISLAND & COASTAL ARCHAEOLOGY 129

Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


Between Africa and Eurasia, particu-lar interest has focused on the ‘southerndispersal route’ (Lahr and Foley 1994),across the southern end of the RedSea and the Arabian Peninsula to theIndian Subcontinent, with the attrac-tions of coastal and marine resources onthis route hypothesized as a particularlypotent factor in the rapid dispersal ofanatomically modern humans (Field andLahr 2005; Field et al. 2007; Stringer2000; Walter et al. 2000). The tech-nological similarities of Middle StoneAge industries in East Africa with thoseof southern Arabia, and the contrastswith material of broadly similar age inthe Levant (Beyin 2006; Rose 2004) aresuggestive, but not decisive. Similaritiesin stone-tool industries and mammalianfaunas on either side of sea channelssuch as Bab al-Mandab and Gibraltardo not provide decisive evidence foror against sea crossings or the exis-tence of temporary land bridges, sincethe populations in question could haveended up on opposite shores by takinga longer and more circuitous route byland. The greater time that might beinvolved in movement via the longerroute is insignificant on the time scales ofobservation that are generally applicablein the Pleistocene.

Given all the above uncertainties,the significance of identifying the par-ticular pathways of human dispersalmay be questioned. On one level orscale, it might be argued that the factand timing of human dispersals fromone region or continent to the nextmay be the significant achievement ofinterest rather than detailed knowledgeof the particular pathways by whichthey were accomplished. Our view isthat a focus on dispersal routes isimportant for two reasons. First, theease of passage, and whether it wasconstrained to a single narrow corridoror possible at many points across a

broader geographical range, clearly hasimplications for the availability, extentand duration of cultural contact andgene flow between the regions in ques-tion, and is therefore highly relevant toa broader understanding of large scalecultural and biological variability. Manyof the interpretations currently beingoffered to explain such variability arebased on untested assumptions aboutclimatic and paleogeographic change,and human cognitive and technical abili-ties, or on quite specific inferences frompaleontological and genetic data thatrequire corroboration against indepen-dent sources of evidence (cf. Forsterand Matsumura 2005; Macaulay et al.2005; Thangaraj et al. 2005). Secondlyand perhaps more importantly, a focuson pathways of dispersal highlights ourignorance about the habitat preferencesand paleodiets of the populations inquestion, the availability of terrestrialand marine resources in different re-gions and at different times, and theattractiveness of particular landscapesunder paleoclimatic and paleogeograph-ical conditions that were clearly verydifferent from those of today. Derricourt(2005) recently made a persuasive casefor the land route via the Sinai Peninsulaas the key to dispersal beyond Africa,on the grounds that this provides themost parsimonious explanation of theavailable data. However, there are goodreasons for supposing that the availabledata are not only incomplete but seri-ously biased.

Throughout these debates, argu-ments have relied on largely untestedassumptions about human paleoe-conomies and the nature and persis-tence of physical and environmental bar-riers to dispersal, rather than on directevidence. The problem of differentialsurvival and visibility of evidence posesparticular difficulties. All shorelines aresubject to peculiarly active geological

130 VOLUME 2 • ISSUE 2 • 2007

Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


processes of erosion and sedimentationleading to high risks of removal orburial of evidence and the transforma-tion of shoreline conditions relative tothe present day. This situation is aggra-vated by the fact that for most of humanexistence sea levels have persisted atlevels much lower than the present, byat least 40 m and for shorter periods byover 100 m. Such a large drop in sealevel has obvious and well-recognizedadvantages for prehistoric populationsin narrowing sea channels and creatingland bridges for human dispersal onfoot. But the same processes of sea-level change have ensured that most ofthe relevant coastlines, much of theirimmediate hinterlands, and any associ-ated archaeology that might contradictland-based views of human develop-ment and dispersal, are now submerged,eroded away, or buried under marinesediments. Changes of sea level havealso almost certainly been accompaniedby considerable changes in marine andcoastal ecology compared to present dayconditions. If we consider that manycoastal environments are likely to haveprovided relatively attractive conditionsfor human settlement, with improvedwater supplies (see Faure et al. 2002)and greater richness and diversity ofresources on the landward side as well asthe addition of marine resources at theshore edge, especially in relation to therelatively arid conditions that generallyprevailed during glacial periods, thenwe may be missing a very large part ofthe Pleistocene record, and perhaps themost significant part. Reconstructions ofdispersal routes and barriers that takeno account of the very different con-ditions of local topography, hydrology,and environment that obtained in coastalregions on the exposed continental shelfduring periods of lowered sea level mustbe considered at best misleading and atworst quite wrong.

Here we focus on the southern endof the Red Sea Basin as a dynamiccoastal landscape and a potential migra-tion corridor involving a crossing of theBab al-Mandab Straits. We pay particularattention to the evidence of sea-levelchange during glacial cycles, the impacton shoreline reconstructions of isostaticand tectonic processes, and the effectsof these changes on the possibilitiesof human settlement and dispersal. Weevaluate critically the current archaeo-logical evidence for or against an earlyuse of coastal and marine resources, andwe present new information on a projectdesigned to explore submerged coast-lines in the region and our strategy forlocating relevant material underwater.

THE RED SEA BASIN: GEOLOGICAL ANDENVIRONMENTAL CONTEXT

The Red Sea separates Arabia from Africa(Figure 1). It is nearly 2000 km longwith a maximum width of about 280km, and extends from 12.5◦ N to 30◦N (Braithwaite 1987; Head 1987). In thenorth it branches into the Gulf of Suezand the Gulf of Aqaba. Its sole naturallink to the world ocean at its southernentrance, the Bab al-Mandab, currentlyhas a maximum depth, at the Hanish Sill,of 137 m. In winter, warmer and fresherwater flows into the Red Sea near thesurface, while cooler and saltier waterflows into the Gulf of Aden at depth.In summer the surface flow is reversed,and intermediate water from the Gulf ofAden flows into the Red Sea betweenthe two outflowing layers (Siddall et al.2002). The eastern shore is bordered formuch of its length by a steep and locallyhigh escarpment. In the south there isa coastal plain, the Tihamah, with anaverage width of about 60 km backedby the Asir Highlands, which includethe highest mountains of the Arabian


Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


Figure 1. General map of the Arabian Peninsula and adjacent regions showing major tectonicfeatures and Paleolithic sites. See text for sources of site information (compiled anddrawn by G. Bailey and C. Vita-Finzi).

escarpment. On the western, Africanside there are plains and tablelands tothe north and the mountains of Sudan,Ethiopia, and Eritrea to the south, withthe Afar Depression in the southwestcorner. Both sides of the growing riftare composed of crystalline basementand sedimentary rocks locally cappedby basalt flows. These are bordered bycoastal sediments, evaporites, and coralreefs.

Climatic conditions throughout theregion are arid. Rainfall in coastal andlowland regions rarely exceeds 180 mm

per year, with semi-desert vegetationthat supports a sparse fauna of rodentsand antelopes (Head 1987; Sanlaville1992). Inland and at higher elevationrainfall is much higher. In the AsirMountains, annual rainfall ranges be-tween about 300 and 1000 mm, andthis is the wettest region in the wholeof the Arabian Peninsula. To the east,the Indian Ocean Monsoon brings sum-mer rains and relatively fertile condi-tions to coastal areas facing the Gulfof Aden. The northern part of the RedSea is influenced by the North Atlantic

132 VOLUME 2 • ISSUE 2 • 2007

Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


oscillation, which leads to periods ofgreater or lesser aridity every six yearsor so (Felis et al. 2000).

Coastal regions throughout the Ara-bian Peninsula provide access to pro-ductive marine environments. In theRed Sea, coral reefs, sea grasses, andintermittent mangroves support a richmarine biota of reef and pelagic fish,sea mammals such as dugongs andcetaceans, and marine molluscs, espe-cially on the more extensive areas ofshallow continental shelf in the southernpart of the basin. The shallow plankton-rich waters around the archipelagos ofDahlak and Farasan support the mostabundant fisheries in the Red Sea today.

The Red Sea was originally a north-ern extension of the African Rift, associ-ated with modest opening and overallextension, though the onset of thisprocess is poorly determined. A majorphase of activity came later when Arabiastarted to split from Africa, and the Gulfof Aden began to open, allowing the seato enter the continent. Opening of theGulf of Aden seems to have started inthe east and extended to the west tothe present day Afar, and it has beenargued that the Red Sea has also beenassociated with a propagation of activityfrom the Afar to the north (Bonatti1985; Omar and Steckler 1995). Noneof this tectonic activity is well dated anddifferent dates have been attributed todifferent parts of the boundary of theArabian Plate. However, opening of theGulf of Aden had begun by 13 Myr (e.g.,Hubert-Ferrari et al. 2003; Manighetti etal. 1997), with approximately contem-poraneous activity in the Caucasus (e.g.,Hubert-Ferrari et al. 2002) and the DeadSea Rift (Garfunkel and Ben Avrahem1996; Garfunkel et al. 1981). Girdlerand Styles (1974) suggested two majorphases of sea-floor spreading within theRed Sea proper, the first 41–34 Myr ago,corresponding to activity of the proto-

Red Sea Rift when it was still a landbasin, and the second during the last4–5 Myr, involving opening of an alreadyestablished Red Sea marine basin.

Rifting is the result of thinning andseparation of the Earth’s crust and isaccompanied by volcanism and faulting,subsidence of the rift floor, and pro-gressive uplift of the rift flanks. In EastAfrica the result is complex landformscomprising near-vertical fault scarps,lake basins, numerous volcanic cones,and extensive lava flows of considerablesignificance in understanding patternsof human evolution (King and Bailey2006). In the Red Sea, as a consequenceof the separation of Arabia, the processof rifting has proceeded much furtherto form an incipient ocean. Continentalextension is associated with the creationof a deep axial trough in the center ofthe basin (the Rift or Graben) and upliftof the rift margins to form the mountainescarpments. The highest mountains arein the south in Yemen and Ethiopia, withelevations exceeding 3000 m, whilethe deepest part of the axial trough is2850 m.

During the Miocene, from about 25Myr to 5 Myr, the Red Sea Basin shows lit-tle direct evidence of continued rifting,but high rates of evaporation in a semi-enclosed basin resulted in the formationof thick deposits of salt or evaporites(Girdler and Whitmarsh 1974). Thesehave resulted in different and dramatictectonic processes on a more localizedscale. Because evaporites are less densethan overlying deposits, they tend topush upwards to create salt domes ordiapirs. The salt can also be dissolvedby water action, resulting in localizedsalt withdrawal and formation of deepdepressions. The region of the FarasanIslands and Jizan on the adjacent main-land is located above an uplifting saltdome, while the offshore topography ofthe Farasan Islands has a number of deep


Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


depressions resulting from salt solution(Bosence et al. 1998; Plaziat et al. 1998;Warren 1999).

Both types of tectonic movementhave a significant bearing on the assess-ment of paleogeographical conditionsrelevant to human settlement during thePleistocene, rifting because it has prob-ably altered the width and geometry ofthe mouth of the Red Sea at its southernend, and salt tectonics because it hascaused localized warping and distortionof shorelines, which complicates thetask of paleoshoreline reconstruction ata local scale.

THE ARCHAEOLOGICAL CONTEXT

In spite of early discoveries of Paleolithicimplements in the south of the ArabianPeninsula towards the middle of the lastcentury (Caton-Thompson 1953), theregion has suffered from comparativeneglect in terms of its archaeologicalpotential. Difficulties of access, limitedsurveys, lack of dated sites, and anassumption that the Arabian Peninsulawas an arid and inaccessible cul-de-saccut off by the Red Sea, have resulted ina greater emphasis on the Levant andthe Fertile Crescent with their longerhistory of investigation and apparentlyricher and better-watered environments.In fact there are many Paleolithic sitesin Arabia. The Comprehensive Archaeo-logical Survey Program of Saudi Arabia,conducted in the late 1970s, discovereda large number of sites of broadly Pa-leolithic date, together with discoveriesby American expeditions led by NormanWhalen, and recently summarized byPetraglia (2003) and Petraglia and Al-Sharekh (2003), to which can be addedmaterial discovered in the Yemen (Al-Ma’mary 2002; Amirkhanov 1991) andmore recently discovered finds in Saudi

Arabia (Al-Sharekh 2006) and the Yemenand Oman (Rose 2004).

One general handicap in assessingthe Arabian evidence is the rarity ofchronometric dates. Currently, only onesite has any radiometric dates, and thatis Saffaquah, near Dawadmi (Figure 1).Here Uranium series dates of calciteconcretions on the underside of theartifacts have given two different setsof determinations, at c. 100,000 and200,000 years (Whalen et al. 1984).These are, of course, minimum dates,since the calcite formation postdatesby an unknown interval the depositionof the artifacts themselves. Otherwise,attempts have been made to date sites bymeans of typological comparisons withAfrican and European assemblages. Atleast three categories have been identi-fied: Oldowan, Acheulean, and MiddlePaleolithic. Oldowan material is iden-tified primarily on the basis of simpleflaked stones, Acheulean by the pres-ence of bifacially flaked handaxes, andMiddle Paleolithic by the presence offlakes produced from prepared cores.However, it is far from certain thatthe so-called Oldowan material neces-sarily indicates an early age. Acheuleanand Middle Paleolithic are more reli-able categories that can reasonably beassumed to represent a broad chrono-logical sequence, but the earliest dateand time span of the Acheulean, andthe earliest date of the Middle Pale-olithic are unknown, though the latteris believed to date back at least to MISStage 5e on the basis of associationswith raised marine terraces along theRed Sea of presumed Last Interglacialdate. Evaluation is further complicatedby the fact that almost all the knownartifacts are surface finds on deflationsurfaces with no depth of deposit andno stratigraphic integrity. Consequentlythere is no guarantee that collectionsof material found in the same location

134 VOLUME 2 • ISSUE 2 • 2007

Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


are coherent assemblages rather thanpalimpsests representing aggregationsof artifacts coincidentally deposited inthe same place on different occasionsover thousands or tens of thousands ofyears.

Nevertheless, the overall distribu-tion of the material provides some gen-eral indications (Figure 1). The most no-ticeable feature is that Paleolithic findsare widely distributed across the ArabianPeninsula. Material is found in coastal lo-cations, in the broad sense of that term,as well as inland and in desert regionswhich today lack sufficient water forhuman survival, and especially arounddried up lake basins. That feature ofthe archaeological distribution, togetherwith geological indications of water flowin the form of stream deposits, tufas, andlake deposits in the arid interior (Sanlav-ille 1992), also shows that climatic con-ditions in the Peninsula were generallywetter than today at certain periodsduring the Pleistocene. We discuss thechronology and duration of these wetterperiods below, but it is clear in broadterms that human populations took ad-vantage of such conditions from an earlyperiod and established settlement farinto the interior of the Peninsula.

Two other factors have attractedgreater attention to this region. The firstis the suggestion that the southern endof the Red Sea may have posed much lessof a barrier to dispersal between Africaand Arabia than once thought, and mayeven have been cut off from the IndianOcean by a land connection at some pe-riods. The second is the greater interestin the potential of marine resources andespecially marine molluscs at the shoreedge to provide an added attractionand a means of subsistence that mayhave facilitated viable human settlementand dispersal, especially in otherwisearid regions. As in other parts of theworld there are many indications of

maritime adaptations dating from aboutmid-Holocene times onwards, with shellmounds and coastal settlements clearlyindicating well-organized and intensivefishing and shellgathering activities, es-pecially on the eastern side of thePeninsula (Beech 2004; Biagi and Nis-bet 2006). However, as elsewhere, theappearance of this material at about7,000 to 6,000 years ago coincideswith the establishment of modern sealevel, and the absence of earlier mate-rial of similar type may simply reflectsubmergence or loss of evidence ac-cumulated during periods of loweredsea level. Thus particular interest at-taches to the evidence of coastal sitesassociated with the high sea levelsof the Last Interglacial, and to theprospects of discovering underwatershorelines and associated archaeologyformed at the lowered sea levels thatpersisted during the period betweenabout 125,000 years and 7,000 yearsago.

SEA-LEVEL CHANGE

For localities far from former ice sheets,a first approximation of the impact ofsea-level change on coastal paleogeogra-phy is to take global models of eustaticsea-level variation and apply them toselected bathymetric contours (e.g., VanAndel 1989a, 1989b). This can giveuseful general results, depending on thelevel of accuracy required, but is limitedby a number of uncertainties and poten-tial errors. Navigational charts providebathymetry that is quite accurate downto depths of about 30 m, sufficient toallow ships to navigate safely, but belowthat depth soundings give only the mostgeneral indication and a false sense ofsmooth topography that may disguiselocal complexity. Sediment cover, thegrowth of coral reefs, erosion associated


Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


Figure 2. Global sea-level change over the past 140,000 years according to the isotope record ofdeep-sea cores, showing the correction indicated by measurements of elevated marineterraces. Horizontal bars indicate approximate date and depth of submerged shorelines,which may be re-worked at successive sea-level stands after initial formation. Visibility ofmarine resources shows the periods when archaeological sites on or close to the moderncoastline were close enough to the paleoshoreline to accumulate remains of marinefood resources (data from Chappell and Shackleton 1986; Lambeck and Chappell 2001;Shackleton 1987; Van Andel 1989a; drawn by G. Bailey).

with transgressive phases, and otherfactors can also affect the accuracy ofpaleoshoreline and coastal landscape re-constructions. Also, different proxies formeasuring sea-level change give some-what different results, together with asubstantial margin of error on the datingof particular sea-level stands, so that

even if one can establish with confi-dence where the shoreline was whensea level was, say, 40 m below present,the dating of that event may be subjectto a margin of error of several thousandyears. Movements of the Earth’s crustresulting from loading and unloading ofice, water, or sediments (glacio-hydro-

136 VOLUME 2 • ISSUE 2 • 2007

Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


Figure 3. Global sea-level change over the past 140,000 years according to the Red Sea isotoperecord. Gray and dashed gray lines show the sea-level curves of Figure 2 (data fromSiddall et al. 2003; drawn by G. Bailey).

isostatic effects) and tectonic move-ments introduce additional variables thatmay differ considerably from region toregion. The standard measure of eustaticsea-level change is the deep-sea record ofchanges in oxygen isotope composition(Figure 2), which forms the basis for themarine isotope stratigraphy widely usedas a framework for Quaternary studies.Since variations in δ18O reflect variationsin the isotopic composition of the seawater with differential removal of thelighter 16O isotope to feed the growthof the continental ice sheets, this curveis in effect a measure of changes inglobal ice volume and hence a measureof sea-level change (Shackleton 1987).However, dated elevations of marine

beaches on rapidly uplifting coastlinessuch as the Huon Peninsula of northernNew Guinea demonstrate that the iso-tope record during a large part of the lastglacial period is amplified by a drop insea temperature, which exaggerates theamount of sea-level fall if uncorrected(Figure 2). A third sea-level curve hasbeen derived from deep-sea cores fromwithin the Red Sea (Figure 3), in whichthe oxygen isotope record is convertedinto sea-level variation via a more com-plex model of salinity changes. In thismodel, the salinity of the Red Sea in-creases as sea level drops because ofreduced exchange of seawater with theIndian Ocean through the narrow andshallow Bab al-Mandab, combined with


Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


high rates of evaporation (Siddall et al.2003). The effect of temperature changeis factored into the model and the resultis a high-resolution record on a 100-yeartime scale back to about 450,000 yearsago, compared to the coarser 1,000-yeartime scale but longer time span of theoceanic deep-sea records. These differ-ent measures show broad agreementover the glacial-interglacial cycle of thepast 140,000 years, even if they differ indetail. One point that is clearly evidentis that periods of high sea level havebeen relatively short-lived, and that formost of the intervening millennia sealevels have oscillated within a range of40–60 m below present, with occasionalsea-level stands as low as –130 m andothers at about –20 m. Moreover, sea-level change appears to have followeda very similar amplitude and periodicityover the past 800,000 years. Before that,the amplitude of variation in the isotopesignal is lower, but it suggests thatsea levels significantly lower than thepresent have been the norm throughoutthe earlier Pleistocene, even if the ex-treme lowstands associated with morerecent glacial maxima were not reached.These facts reinforce the point madeearlier that the archaeological record ofPleistocene coastal settlement is likely tobe severely under-represented.

Some idea of the amount of landexposed by sea level lowering in theRed Sea can be gained from the –100m bathymetric contour (Figure 4). Theshallowest part of the continental shelfis clearly in the southern part of thebasin, where the coastline on either sideextended seawards by 60 to 100 km,incorporating the present day DahlakArchipelago and Farasan Islands into theadjacent mainland. To the north theoffshore topography steepens and in thefar northeastern corner in the Gulf ofAqaba, the offshore gradient is so steepthat there is almost no additional incre-

ment of land at lowered sea level. Thecross-sections emphasize this variation,with the deep axial trough reaching to adepth of 2000 m and more in the centralbasin, and becoming dramatically shal-lower and narrower towards the HanishSill and the Bab al-Mandab. With theHanish Sill at a depth of 137 m, closeto the maximum eustatic sea-level re-gression, the accuracy of paleoshorelinereconstructions becomes critical, and inthis respect two further effects need tobe examined.

Isostatic Effects

Loading and unloading of ice andseawater results in isostatic adjustmentof the Earth’s crust. Far from the conti-nental ice sheets, the principal changesare due to seawater loading on the seafloor, and these hydro-isostatic effectsmay account for up to 30% of theeustatic signal during lowstands. Thisis the dominant isostatic factor in theRed Sea. As an example, the loading ofseawater resulting from a sea-level rise of100 m on a linear coastline would causesubsidence of the sea floor by as muchas 30 m, although there would be a timelag in the response of the mantle andlithosphere. The land at the coast wouldbe dragged down to a lesser extent,while inland there would be no changeor perhaps a slight uplift as material inthe underlying mantle was redistributedto compensate for the loading offshore.With lowering of sea level of course, thereverse effects would take place, namelya slow rebound of the sea floor and thecoast margin and slight subsidence ofthe hinterland. The crust also respondsto glacio-isostatic effects, that is loadingand unloading of ice, and while this isthe dominant signal in areas of glacia-tion, it may also have some effect atgreater distance from the ice margin. Icemodels based on the configuration of the

138 VOLUME 2 • ISSUE 2 • 2007

Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


Figure 4. The Red Sea showing area of land exposed at the –100 m bathymetric contour (afterHead [1987, figure 1.2]; drawn by G. Bailey).

ice sheets, and earth models based onthe thickness of the lithosphere and theviscosity of the upper mantle, togetherwith information on the local geometryof the continental shelf, can be usedto estimate the position of shorelinesunder varying isostatic conditions (e.g.,Lambeck 1995, 1996b, 2004). Detailsof the parameters and models used inthe calculation of shoreline positions inthe Red Sea are described elsewhere,together with calibration of the resultsagainst the available evidence of datedshoreline features and elevated terraces(Bailey et al. in prep). Here we presentsome maps illustrating the results and

consider some of their paleogeographi-cal implications.

Our first set of maps illustratessuccessive shoreline positions in thevicinity of the Hanish Sill and the Babal-Mandab during the last glacial cycleand their approximate dates (Figure 5).The narrowest sea crossing today is29 km, divided into two channels byPerim Island, the western channel 26km wide, and the eastern 3 km. At theLast Glacial Maximum 20,000 years ago,the channel would have been so shallowand narrow that it is not possible to becertain within the available margins oferror whether the seabed was above or


Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


Figure 5. Shoreline positions in the region of the Bab al-Mandab Straits and the Hanish Sill atdifferent dates in the most recent sea-level cycle (data compiled by K. Lambeck; drawn byG. Bailey).

below sea level (Figure 5d). In any casethe land elevation at this crossing couldhave been at most only a few metersand would not have formed an effectivebarrier isolating the waters of the RedSea from the Indian Ocean. The deep-seacore record from the Red Sea reinforcesthis conclusion, demonstrating that wa-ter circulation with the Indian Oceanwas never completely cut off at themaximum regression of sea level, or notfor long enough to register in the isotoperecord. Fernandes et al. (2006) refer to awater depth of 15 ± 6 m over the HanishSill and a width of channel of 4 km atthe glacial maximum. But it is unclear

how accurately the depth and width ofthe channel can be inferred from theisotope data, or whether it was a singlechannel or a series of braided channels,and Siddall et al. (2003) include anerror term of ± 12 m in their estimatesof sea-level variation. The isotopic andisostatic records, which are derived fromindependent sources of information, arein very close general agreement, whichgives confidence in the overall recon-struction. Neither method appears suffi-ciently sensitive to capture a situation ofalternating wet and dry conditions fluc-tuating over short time scales across thesouthern end of the Red Sea, however,

140 VOLUME 2 • ISSUE 2 • 2007

Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


or to reject decisively the existence ofan intermittently dry pathway betweenthe African and Arabian shores at lowestsea-level stands. Whether wet or dry, thecrossing at the glacial maximum posedno significant barrier to the dispersal ofhumans or many other mammals. Evenat sea levels up to about −50 m, thechannel would have remained long andnarrow, with intervening islands to actas stepping-stones, and water crossingsno more than a few kilometers widefor the period between about 90,000and 12,000 years ago (Figure 5a). Theopposite shore would have offered avisible target extending for tens of kilo-meters. The probability of successfullandfall in such circumstances is veryhigh, regardless of whether the crossingoccurred intentionally or accidentally,or by swimming, rafting, or in boats,because there is little danger of failurewhen being carried parallel to the shoreby water currents (cf. Birdsell 1977).

Estimates of water currents at low-ered sea level can provide some controlon this assessment. Using a two-layermodel to compute the flow in and outof the Red Sea, and assuming a waterdepth of 32 m above the Hanish Sill, theannual average velocity of water flow inthe upper layer is 0.4396 m/s, equivalentto 1.58 km/hr (Biton, pers. comm., July2007). Given a channel width of 4 kmat that depth, and assuming that prehis-toric peoples could swim at 1 km/hr, thecrossing time is 4 hours, during whichthe swimmer would be swept sidewaysa distance of 6 km. At a water depth of 17m above the Hanish Sill, the water flowis 0.7098 m/s, and the extent of side-ways movement nearly 10 km assumingthe same channel width and swimmingspeeding. Neither situation would poseany danger of being swept far awayfrom land, given that the channel wasmore than 100 km long (see Figure5). If the water above the Hanish Sill

was even shallower, current velocitiesmight increase accordingly, though theireffects on lateral displacement wouldhave been offset to some degree byfurther narrowing of crossing distances.But even in this case the dangers wouldbe slight given the length of the chan-nel. Moreover, there were probably sea-sonal fluctuations in current flow, withslacker conditions at certain times ofyear.

If lower sea levels enhanced theattractions of the southern corridor, weshould also look at sea-level effects onalternative pathways, across the north-ern end of the Red Sea, and to the eastin the Gulf region between Arabia andIran. In both cases there are shallow con-ditions that would have been exposedas dry land quite early on in the glacialcycle, and the maps illustrated show suc-cessive shorelines corrected for hydro-isostatic effects at selected stages of thecycle. The Gulf of Suez would haveremained an exposed basin until about14,000–15,000 years ago when sea lev-els rose above about –50 m (Figure6), broadening the potential pathwayvia the Sinai. On the eastern side ofthe Peninsula, the Tigris-Euphrates riversystem would have extended out into anow submerged low-lying region witha narrow sea inlet throughout a largepart of the glacial period, offering apotentially attractive pathway for dis-persal on a northwest-southeast axis(Figure 7). The continental shelf aroundthe Gulf of Oman, though relativelynarrow in places, would also have of-fered a significant addition of low-lyingcoastal land that could have facilitatedmovement from the Arabian Peninsulato the Indian subcontinent, whetheroriginating from the north via the Tigris-Euphrates catchment or from the southvia southern Arabia. In other words, thepaleogeographical effects of a loweredsea level would have been advantageous


Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


Figure 6. Shoreline positions in the Gulf ofSuez at different dates in the mostrecent sea-level cycle (data compiledby K. Lambeck; drawn by G. Bailey).

to ‘northern’ as well as ‘southern’ routesof dispersal.

On the time scale of the last glacialcycle, the dominant temporal and spatialvariation in relative sea level within theRed Sea is likely to have been the com-bined eustatic-isostatic effect. Extrapo-lation further back in time becomesmore debatable because global glacialhistory becomes increasingly uncertainand tectonic contributions, tending to

be secular or cumulative, may becomemore significant. Thus the situation wehave described so far applies with someconfidence to the last glacial cycle,which covers the span of time thatencompasses the postulated date ofdispersal out of Africa of anatomicallymodern humans, and with decreasingconfidence to earlier glacial cycles. Whatof earlier periods and the possibilities formovement of the earliest populations toleave Africa (Homo ergaster or Homoerectus), according to current opinionat about 1.8 million years ago (but seeDennell and Roebroeks 2005)? If weconsider only the global isotope recordsavailable from the early Pleistocene,conditions like those at the last glacialmaximum would have been absent anda situation closer to that shown inFigure 5b or perhaps Figure 5c wouldhave been the best available by way ofeasing transit. Even under these condi-tions, island hopping, sea crossings overrelatively short distances, and a longand relatively narrow channel, wouldseem to offer possibilities of dispersalacross the straits for archaic humanswith quite modest capacities for floating,swimming, or improvised rafting. How-ever, an additional variable that becomessignificant in this earlier period is theeffect of tectonic movement associatedwith rifting. This raises the possibilitythat the straits may have been narrowerthan today at 1.8 Myr, and perhapsshallower, because of ongoing seafloorspreading since then that has widenedthe gap between Africa and Arabia.

Tectonic Effects

Spreading rates for the opening ofthe Red Sea have been derived from theseafloor paleomagnetic record. Girdlerand Styles (1974) proposed an averagerate of separation between the Africanand Arabian margins of 13 mm yr−1

142 VOLUME 2 • ISSUE 2 • 2007

Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


Figure 7. Shoreline positions in the Persian Gulf at different dates in the most recent sea-level cycle(data compiled by K. Lambeck; drawn by G. Bailey).

during the earlier phase of rifting over30 Myr ago, with a slower rate of about 9mm yr−1 following resumption of riftingafter 5 Myr. If we take the lower figure,and assume that it can be applied tothe southern opening of the Red Sea ata uniform rate over the past 2 millionyears, that implies a separation of 18 km,clearly enough to substantially alter thepossibilities of transit, even with a loweramplitude of sea-level variation duringglacial periods. There are, however, con-siderable uncertainties about these ratesand the appropriateness of applyingthem to the Bab al-Mandab Straits.

Some recent measurements haveproduced rates for the last 3.2 Myr thatrange from ∼16 mm yr−1 near 18◦ Nto 10 mm yr−1 at 25◦ 30′ N (Chu andGordon 1998). Separation between Sinaiand Africa during the Late Pleistocenehas been 0.8–1.2 mm yr−1 (Bosworthand Taviani 1996). GPS data for the Red

Sea are still of limited value as they arebased on only one continuous GPS sta-tion and three survey-mode GPS stationson the Arabian Peninsula (McClusky etal. 2003), but they yield a Nubia-ArabiaEuler vector in statistical agreement withthat derived from magnetic anomalies,with an associated angular displacementat the central Red Sea of ∼0.4◦ Myr−1,amounting to extension at about 7 mmyr−1 on a roughly NE azimuth.

These rate variations may be ex-plained by the great disparity in thelength of the periods in question, butmight suggest a real slowing in the rateof extension. Spreading history can alsobe assessed by reference to the isotopiccomposition of foraminifers from corestaken from the Red Sea floor (Vita-Finziand Spiro 2006), on the assumptionthat spreading is primarily coseismicand thus associated with pulses of hy-drothermal activity. This approach has


Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


yielded evidence of a major tectonicevent prior to 40,000 years ago but nonesubsequently, and suggests that tectonicwidening of the Red Sea has been moremodest for the last 40,000 years andpossibly longer. This is consistent withthe evidence from many locations ofelevated marine beach deposits datedto the Last Interglacial at heights abovethe present sea level of c. 5 m (e.g.,Gvirtzman 1994; Plaziat et al. 1998) closeto the global norm, which strengthensthe inference that for much of the RedSea there has been little overall verti-cal movement of the rift flanks duringthe past 125,000 years. In contrast,the on-land portion of the Dead SeaRift indicates a recurrence interval of1,500 years for earthquakes of M = 7.3and 42 events of M = 6.7–8.3 in thepast 2,500 years (Ben-Menahem 1991;Marco et al. 1996). Episodic extensionof the Red Sea, according to Drury etal. (2006), is well established on bothits flanks on the basis of fault polari-ties and cooling histories. Evidently asingle recurrence interval is unlikely toapply to the entire structure, but ifone concludes that only large eventsare recorded geochemically in Red Seacores, the available measurements mayunderestimate the rate of spreading. Inother words, simple extrapolation onthe basis of long-term extension ratessuggests that the Bab al-Mandab couldhave opened as recently as 2 Myr ago.These arguments assume that the Bab al-Mandab Straits have always been directlyconnected to the main stretches of theRed Sea and the Gulf of Aden throughoutthe past 2 Myr. Deformation in the Afarshould throw some light on this assump-tion, but here the tectonic history ismore complex. The current depressionthat forms the Bab al-Mandab Straitsexhibits no significant current activity,and motion has jumped to the highlyactive Danakil depression (e.g., Ayele et

al 2007; Manighetti et al. 1997). Whenthis occurred is poorly established, butis unlikely to have been much morerecent than 2 Myr ago, suggesting thatany tectonic activity in the region overthat time span is minor with at mostmodest alterations to the coastlines dueto tectonic activity. Absence of evaporiteformation during the past 5 millionyears also indicates that some form ofsea connection has been maintainedover that period, although it does notrule out short-lived episodes when dry-land transit from Africa to Arabia waspossible.

In conclusion, for the period ofapproximately the last 200,000 years,which encompasses the proposed dis-persal of anatomically modern humansout of Africa, we are confident thattectonic effects would have had nomajor impact on the reconstructionsof paleoshorelines shown in Figure 5.For earlier periods of human dispersal,and especially for a postulated dispersalevent at about 1.8 Myr, the positionis much less clear. Further evaluationof the implications for ease of humantransit in this period would requiremore detailed geological and seismicinvestigation of the straits region thanare currently available, or extension ofthe Red Sea isotope record into earlierperiods. In the absence of such detail,we can only point to possibilities inneed of further exploration. Sea levelsmay not have been lowered to the sameextent as in the late Pleistocene, butconversely tectonic effects may havecreated a channel geometry that wasnarrower or shallower than today. Atmost, the southern crossing of the RedSea during low sea-level stands wouldprobably have been comparable to theconditions that obtained during the LastGlacial (Figure 5b), while dry-land transitis unlikely to have been available exceptfor very short-lived periods.

144 VOLUME 2 • ISSUE 2 • 2007

Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


COASTAL HABITATS AND MARINERESOURCES

Getting across the Red Sea is one issue,and the above discussion suggests thatfor much of the Pleistocene the Babal-Mandab Straits posed little obstruc-tion to at least intermittent movementbetween Africa and southern Arabiaduring long intervals of time. That raisesthe further question of whether theenvironments available in Arabia weresufficiently attractive to make the cross-ing worthwhile and to sustain humansettlement and dispersal further east-wards. To answer that question we needto examine three closely related issues:(1) the nature of climatic conditions incoastal regions at the southern end ofthe Red Sea at different stages of theglacial-interglacial cycle and their likelyimpact on the availability of animal andplant foods on land; (2) the evidencefor the exploitation of marine foods andthe potential impact of such resourceson the possibilities for settlement anddispersal; (3) local conditions of climateand resource availability on the emergedcontinental shelf during periods of lowersea level.

Climate Change

With regard to climate, the preferredview is that wetter intervals in the aridzones of Arabia are correlated with inter-glacials, and hence that conditions weremore arid than today during glacial peri-ods (Derricourt 2005). Some support forthis model comes from investigations inthe Rub’ al-Khali desert and elsewhere,which demonstrate the strengtheningof the Indian Ocean Monsoon and theformation of lakes in now-dry inlandareas between about 9,000 and 7,000years ago (Parker et al. 2004, 2006).Deep-sea isotope records also indicategenerally increased aridity on a global

scale during glacial periods. However,the earlier conclusion of Deuser et al.(1976) that the deep-sea oxygen isotoperecord of the Red Sea indicates morearid conditions during glacial periodsshould be discounted, since they in-ferred increased aridity from high ratesof evaporation associated with highsalinities, whereas Siddall et al. (2003)show that high salinities are primarilydue to reduced inflow from the Gulf ofAden (see also Assaf 1977). In addition,it has been shown that Mediterraneancyclones brought increased winter rain-fall to northern Arabia and the north-ern Red Sea region during the earlyHolocene. On the other hand wetterinterludes have been identified duringpart of the last glacial, creating largelakes in the Nafud region in north-ern Arabia between about 34,000 and24,000 years ago (Arz et al. 2003; Schultzand Whitney 1986). Sanlaville (1992),in summarizing evidence in the formof alluvial sediments, travertines, tufas,calcretes, and lake deposits indicatinghigher levels of precipitation than today,also identified evidence for four wetphases since the Last Interglacial. Thefirst, well-dated and corresponding toMIS Stage 5e, suggests higher levels ofprecipitation than during the wettestphase of the Holocene, under the in-fluence of a northward movement ofthe Indian Ocean Monsoon, and themechanisms underlying such fluctua-tions seem to be of general applicationduring interglacial periods (Rohling etal. 2002). A second and wetter phaserepresents a renewal of extended mon-soon conditions corresponding to Stage5a, followed by renewed aridity inStage 4 after about 80,000 years ago. Athird and prolonged phase of increasedprecipitation extended from at least30,000 years ago, and possibly from asearly as 45,000 years, to about 20,000years ago. These conditions were less


Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


wet than in the previous two phases,but were widespread throughout thePeninsula, and included the formationof extensive lakes in the Rub’ al-Khali(McClure 1976). Sanlaville attributes thisepisode to the southward movement ofMediterranean cyclones bringing winterrainfall. There then followed more aridconditions, with widespread dune build-ing, but with brief intervening phasesof increased humidity, until the onset ofthe early Holocene wet period referredto earlier.

To sum up the above climatic se-quence, there have been wet-dry fluctua-tions during both glacial and interglacialperiods, though of varying amplitudeand periodicity. The wettest and pre-sumably most favorable conditions forplant and animal life on land occurredduring periods of relatively high sealevel, typically at sea levels slightly be-low the present on the evidence of theearly Holocene and MIS 5a phases, butan extended and relatively humid periodalso coincided with a large part of MISstage 3 when sea levels were betweenabout –40 m and –60 m. Only duringthe very maximum of the glacial period,when sea levels reached their maximumregression and crossing the southernend of the Red Sea was easiest, doesit seem that arid conditions might haverestricted further movement.

There is, however, one very impor-tant qualification to this assessment, andthat is the nature of the environmentalconditions in the enlarged coastal regionexposed by a drop in sea level. Faureet al. (2002) have presented cogentarguments to suggest that as sea leveldropped, the flow of groundwater fromthe enormous underground aquifers ofthe Arabian Peninsula, as in many othercoastal regions, would have increased.There are many examples of submarinefreshwater springs in the Mediterranean,the Red Sea, and the Persian Gulf, some

of which are well known to local fisher-men and used to replenish drinking sup-plies. Faure et al. suggest that this flowof freshwater would have been greatlyincreased at lowered sea level becauseof the increased hydrostatic head, andthe removal of the downward pressureinhibiting stream flow imposed by theoverlying column of sea water duringhigh sea levels, and that this effect wouldhave transformed the emerged shelvesinto extensive areas of well-watered landwith springs and wetlands. Thus as sealevels dropped and conditions becamemore arid inland, so refugia for plantand animal life became available in thenewly emerged coastal zone. If thisargument is correct, then the periodwhen local conditions of water supplyand plant and animal life were at theirmost favorable in coastal regions wouldalso have coincided with periods whendispersal across the southern end of theRed Sea and around the coastal marginsof the Arabian Peninsula would havebeen easiest.

Marine Resources

In considering the additional im-pact of marine resources on humandispersal, our first task is to considerthe available evidence that they wereactually exploited for food during thePleistocene, or that they were availablefor exploitation. We begin with therecently reported material from Abdur inEritrea (Figure 1, Bruggeman et al. 2004;Walter et al. 2000). Here artifacts havebeen found stratified within a series ofelevated marine coral terraces, and welldated to c. 125,000 years (MIS 5e). Findsinclude Acheulean bifacials, vertebratebone fragments, including hippopota-mus, crocodile, elephant, rhinocerosand bovid, and oyster shells. Takenat face value the material suggests amixed diet that included a substantial

146 VOLUME 2 • ISSUE 2 • 2007

Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


dependence on land mammals. Walteret al. (2000) originally suggested that theoyster shells were very early evidenceof a significant use of marine resourcesthat helped to facilitate the dispersalof anatomically modern humans out ofAfrica, and cited in support the evidenceof marine resources from deposits inSouth African caves of similar date.These sites, which include caves suchas Die Kelders, Klasies River Mouth andBlombos Cave, and open air sites suchas Sea Harvest and Hoedjies Punt (Averyet al. 1997; Henshilwood and Marean2003; Henshilwood et al. 2001; Klein etal. 2004) date between about 130,000and 75,000 BP, and contain variablequantities of marine shells, mostly rockyshore species of limpets and mussels,often forming layers of quite dense shellmidden, and bones of seals, penguins,and fish.

However, there are unresolved ques-tions about the taphonomy of the Ab-dur material. Bruggeman et al. (2004)indicate that the oysters are a naturaldeath assemblage and are not thereforecertainly food remains. The artifacts as-sociated with the shells could have beenused for activities near the shorelinequite unrelated to shellfish gathering andconsumption, and then discarded in theshallow water nearby, or on the exposedreef flat during a period of sea-level re-gression. In any case, as in South Africa,the Abdur evidence is at best the earliestvisible expression of shoreline activityin this region, made visible by proximityto the coastline during a period of highsea level, rather than the earliest actualevidence. In short, the Abdur evidenceis an important indication of visits tothe Last Interglacial shoreline, and rein-forces the significance of similar findsin the Afar (Faure and Roubet 1968).However, there is currently no clearevidence for shellfish consumption, letalone a significant dependence on it,

and certainly no evidence that it formedthe basis for a new adaptation to theexploitation of marine resources or thekey to dispersal of anatomically modernhumans.

Similar sites have been recordedalong the Saudi Red Sea coastline inassociation with raised marine terracesand lava fields, with finds of Lower andMiddle Paleolithic stone tools, particu-larly near Al Qahmah and Al Birk (Figure1, Zarins et al. 1981). All the sites andartifacts we have examined in the fieldare surface finds. None of them arestratified within the underlying deposits,and extensive inspection of exposedsections in marine beach deposits andalluvial sediments has so far failed toreveal any artifacts in a stratigraphiccontext that might provide some meansof dating and preservation of associatedfood remains. Two basalt samples fromthe lava cone at site 216–208 near AlQahmah (Zarins et al. 1981:18, fig. 5/A)gave K/Ar ages of 1.37 ± 0.02 millionyears (KSA04/AR1) and 1.25 ± 0.02 mil-lion years (KSA04/017), which are max-imum ages for the Acheulean artifactsin the vicinity, and Middle Paleolithicartifacts occur on the surface of a nearbycoral beach terrace some 3–4 m abovepresent sea level, presumed to be of LastInterglacial date (Bailey et al. 2007). Sofar, we have found no convincing insitu deposits or shell middens clearlyassociated with beach deposits datedearlier than 6,000 years ago.

From this evidence, the scale andtime depth of exploitation of marineresources in the Red Sea area remainsunclear. There is an added problem insome parts of the Red Sea that duringextreme low sea-level stands salinitieswould have been high enough to pre-vent plankton growth (Fenton et al.2000; Hemleben et al. 1996) and ma-rine resources dependent on the plank-ton food chain. Even if some marine


Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


resources, such as intertidal molluscs,were able to persist during these aplank-tonic episodes because of adaptability tohigh salinities and an ability to subsist ondetrital material, the opportunities forhuman subsistence on marine resourcesduring these periods would seem to bevery limited, at least within the mainbody of the Red Sea. No such problemswould have affected the narrow channelsouth of the Hanish Sill because thesalinity of the inflowing water wouldhave been identical with that in theadjacent Gulf of Aden and Indian Ocean,and that moderating effect may haveextended some distance into the mainbasin, sufficient to permit biologicalproduction in the region of the Farasanand Dahlak archipelagos.

In any case it is unclear whethermarine resources by themselves andespecially molluscs could be relied onas a major source of subsistence inthe absence of complementary foodsupplies derived from terrestrial plantsor animals. All the marine taxa recordedas food remains in sites such as theMiddle Stone Age caves of South Africareferred to earlier could have beenobtained on the seashore—molluscs inthe intertidal zone, fish trapped in rockpools by the receding tide, and seamammals and marine birds scavenged ascarcasses beached by storms or trappedon land in the case of seals. None ofthem presuppose the existence of aspecialized technology or the use ofboats. Molluscs are valuable in providinga predictable and relatively accessiblesupply of food (cf. Meehan 1982), buteven in the largest shell mounds ofthe mid to late Holocene, the molluscsdemonstrably supply much less foodthan the bulk of their remains mightotherwise suggest, are almost alwaysassociated with exploitation of terres-trial resources as well as marine, andrarely show the sort of evidence of catas-

trophic overexploitation that might becaused by over-dependence on molluscsor overwhelming human impact (Baileyand Milner 2002, in press).

Our analysis suggests that resourceson the coastal shelf may have beenmore abundant and more easily availableduring periods of lowered sea level thantoday, particularly on the landward sideof the shoreline, although the precisecombination of circumstances evidentlyfluctuated with changes in climate andsea level and from region to region.There is certainly no basis for arguingfor a persistent barrier, whether physicalor climatic, to dispersal via the southernroute on Pleistocene time scales. Atthe extreme lowering of sea level atthe glacial maximum it may have beenwell-watered conditions on the emergedcoastal shelf, together with ease of tran-sit in regions such as the Bab al-Mandaband Hormuz that was the key attractionto human settlement and dispersal. Whatboth climatic and archaeological consid-erations emphasize is the importance ofthe emerged continental shelf during pe-riods of lowered sea level, and the needto undertake explorations both on landand under water in search of relevantarchaeological and paleoenvironmentalevidence.

THE FARASAN ISLANDS

The Farasan Islands today are some 40km from the mainland and are composedof coral platforms uplifted and deformedby salt tectonics, resulting in a com-plex onshore and offshore topography(Dabbagh et al. 1984; Macfadyen 1930).Indigenous resources include the Ara-bian gazelle, a rich inshore and intertidalmarine environment with fish, molluscs,turtles and sea mammals, and migratorybirds. The sea area around the FarasanIslands supports the largest fishery in

148 VOLUME 2 • ISSUE 2 • 2007

Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


Figure 8. The Farasan Islands showing offshore bathymetry (bathymetric data from HydrographicOffice [2000]: British Admiralty Chart No. 15; drawn by G. Momber).

Saudi Arabia. The Islands offer a numberof advantages for underwater survey.The seabed between the Islands andthe mainland is relatively shallow andat sea levels between about 20 and 50m they would have formed a singleland mass accessible to human settle-ment at intervals throughout the Pleis-tocene without requiring sea crossings(Figure 8).

In the other direction, the shelfslopes more steeply towards the axialtrough of the Red Sea, offering the pos-sibility of discovering shorelines associ-ated with different stages of the sea-levelcycle. At all but the lowest sea levels,the complex configuration of coastlines,islands, and archipelagos would havecreated shore areas protected from thefull destructive force of wave actionduring sea-level rise, with a greaterlikelihood that landscape features andarchaeological material have survivedinundation. The offshore location alsomeans that the seabed is less likely to

have been covered with thick sedimentswashed into the sea from the mainland.Many inshore areas are also unsuited tocoral growth, which would otherwiseobscure the underlying surface. Thecomplex topography of the seabed ap-parent from the bathymetric contours,and in particular the deep depressionsscattered across the underwater land-scape, also suggest the possibility ofnumerous locations that would havetrapped freshwater when the landscapewas uncovered during periods of lowersea level. Here we report on our strategyof investigation and the results of ourfirst season’s work in 2006 (see alsoBailey et al. 2007).

Farasan Archaeology

Almost nothing has been reportedpreviously on the archaeology of theFarasan Islands apart from the results ofbrief surveys published in Zarins et al.(1981) and Zarins and Zahrani (1985),


Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


Figure 9. General distribution of shell mounds on the Farasan Islands (drawn by G. Bailey).

two Latin inscriptions, some Islamic ma-terial, and an unreadable but supposedlypre-Islamic inscription on a traditionalbuilding. Zarins et al. noted the presenceof shell middens in Janaba Bay (Figure9), the remains of structures of probablepre-Islamic age made from blocks ofcoral, and the presence of ceramicstypical of the Islamic period, the pre-Islamic South Arabian Civilization, datedon the mainland between about 1200BC and the sixth century AD, and the‘Neolithic’ period. No excavations wereundertaken as part of these early surveysapart from a small test excavation ofone of the shell middens to providesome radiocarbon dates. We thus haveonly the outlines of an archaeologicalsequence and a basic chronology ex-tending back to about 5400 radiocarbonyears BP.

We are pursuing a joint strategy ofarchaeological investigation above andbelow modern sea level. On land orbelow the water, our interest is inthe prehistoric landscape as a seamlesswhole and the human activities andarchaeological materials associated withit.

Survey Results

In our first season of survey, our ob-jectives were to survey more widely forarchaeological materials on land, to ob-tain more radiocarbon dates, to examinethe relationship between archaeologicalsites on land and local geomorphologicalfeatures as clues to where to searchunder water for earlier sites, and to un-dertake preliminary underwater survey.

Our initial surveys on landhave revealed an immensely rich

150 VOLUME 2 • ISSUE 2 • 2007

Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


archaeological landscape comprisingsome 1,000 archaeological sites. Mostare shell mounds of substantial size, upto 4 m thick, sometimes forming a vir-tually continuous line of deposits alongstretches of the shoreline, and many oth-ers have marine shells associated withthem (Figure 9). The sites often formclusters, with thicker mounds on thebeachfront and shallow shell deposits orshell scatters situated further back fromthe shoreline. Some are associated withpotsherds of Islamic and pre-Islamic age.Others appear to be older prehistoricdeposits without ceramics. Small testtrenches through the larger moundsshow that they have the characteristicfeatures of a midden deposit, with strat-ified layers of broken marine shells, ashlenses, occasional imports of stone mate-rial, and fish bone. The taxonomic com-position of the molluscs varies from areato area, and includes reef, sand, and rockdwelling species, with conch (Strombusfasciatus) and pearl oyster (Pinctadanigra) especially well represented. Thenumber of artifacts recovered fromthese text excavations is currently toosmall to provide any useful clues tocultural affiliation or chronology.

Further radiocarbon dates areawaited from our own investigations.In the meantime it is clear that thereis an abundant archaeological recordextending back to at least the midHolocene. Since the shorelines haveundergone localized uplift as well asdownwarping, we may also find shellmiddens of significantly earlier dateabove sea level. Our survey strategyis also directed to the discovery ofmaterial of earlier date deposited onland when the Islands formed part ofan extensive hinterland on the emergedcontinental shelf, as well as to thediscovery of sites associated with thenow submerged hinterland. There isalso a possibility of discovering coastal

archaeological middens from the LastInterglacial, since there are traces ofan extensive fossil coral terrace thatformed an ancient shoreline at the inneredge of the modern shore platform.

Underwater archaeological sitescould be shoreline sites like the shellmiddens on the present-day coast, or‘inland’ sites associated with freshwatersprings or lakes. We suspect thatshoreline sites are likely to be easier tolocate given their probable associationwith shell deposits. Hence, our initialstrategy for underwater exploration isto use the location of the shell moundsand other archaeological features onthe present-day shoreline, and theirrelationship to local geological andtopographic features, to predict thetypes of surface features that weshould be targeting underwater. Acharacteristic feature is the associationof shell mounds with a wave-undercutcoral platform resulting primarily fromthe chemical solution of the coralmaterial at the shore edge (Figure 10).In some cases these undercut coralplatforms on land are now separatedfrom the modern shoreline by ashallow sand-filled embayment, whichwas formerly an extensive intertidalzone with productive conditions forthe molluscs whose shells form themain constituent of the archaeologicaldeposits. This infilling is the result ofcumulative deposition of sediments invery shallow conditions, accentuatedin some cases by localized tectonicwarping resulting from salt movement,and highlights the fact that theecological conditions for abundantintertidal shell beds may have beenrelatively short lived and quite variableon different shorelines.

Preliminary underwater survey hasbeen directed to locating similar shore-line features associated with earlier pe-riods of sea level, using a combination


Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


Figure 10. Shell mound and undercut coral platform on Qumah Island (photo G. Bailey, 2006).

of diving, video, and remote sensing.We have placed particular importanceon the diving work, because video andremote sensing by themselves are insuffi-cient to evaluate underwater features inthe early stages of exploring unfamiliarterrain, and cannot make drawings orcollect dating samples from cementedcoral. An important part of the initial sur-vey was therefore to establish the feasi-bility of prolonged and deep dives usingmixed gas technology and to solve thelogistics of doing so in a remote location.The diving program was facilitated bythe use of a large vessel (the MV Midyan)as a mobile offshore base for a six-strongdiving team, diving equipment, and twosmall boats to provide access to divesites. For short and shallow exploratorydives we used normal scuba tanks withcompressed air and for prolonged diving

at greater depth, nitrox (oxygen andnitrogen) and trimix (oxygen, nitrogen,and helium), which enable divers towork with greater mental acuity, reachdeeper locations that would otherwisebe inaccessible, to work there safely forextended periods, and to explore under-water features to a maximum depth ofabout 90 m.

Diving was concentrated in near-shore locations with relatively steepdrop-offs and good potential for ex-posure of wave-cut coral platformsrepresenting shorelines formed duringperiods of lowered sea level. Pre-liminary acoustic (single-beam echo-sounder) transects were conducted toidentify breaks of slope elsewhere thatmight indicate similar features, and tohelp focus the selection of diving loca-tions. Work was concentrated on the

152 VOLUME 2 • ISSUE 2 • 2007

Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


Figure 11. Mapping of the underwater paleoshoreline at Slick Point on Qumah Island, showingunderwater features. The cross-sections show the profile of the terrace viewed from southto north. The wave-undercut notches at the base of the terrace indicate the position of thepaleoshoreline and both the longitudinal profile and the cross-sections show the extentof warping resulting from salt tectonics (compiled and drawn by G. Momber).

north side of Zufaf Island and on thesouthern side of Qumah Island in thevicinity of Slick Point.

At Zufaf, well-defined coral terraceswere identified at depths of 60 m and6 m, but smothering of sand obscureddetail, and more detailed work was con-centrated at Slick Point. Here there waslittle overburden of obscuring sedimentor recent coral growth, and it was pos-sible to identify coral reefs representingold shoreline terraces at approximately6 m and 20 m depth and to map theirfeatures over a considerable distance(Figure 11). These features are typical

of those associated with the present-dayshoreline and represent a beach lineat a sea level about 20 m below thepresent, which could have formed atabout 80,000–90,000 years ago and beenre-occupied when sea level reached thesame level about 12,000 years ago (seeFigure 2). This former shoreline featureis now tilted in a north-south directionat an angle of about 1 in 20 as aresult of salt tectonics, and rises to alevel some 9 m below the present sealevel within Qumah Bay. Within theBay itself, there are shell mounds abovethe modern sea level that are located


Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


directly over wave undercut platforms,and other mounds located further inlandon an undercut coral platform frontedby a dry, shallow sandy bay. Similartopography has been identified in asso-ciation with the underwater shoreline,including notches and undercuts (Fig-ure 11). Such undercuts occur abovemodern sea level and provide shelteredand shady locations at the present-day,often used as modern campsites witha typical archaeological signature ofbottles, cans, and remains of fireplaces.We would expect to find prehistoricartifacts associated with the equivalentfeatures that are now submerged, as wellas shell midden accumulations abovethe wave cuts in ecologically favorableconditions. Slick Point provides an ex-ample of a submerged shoreline wherecoastal geomorphological processes canresult in clearly defined platforms andunderlying notches, some of which ex-tend up to 4 m into the cliff (Figure11). Unfortunately, the exposed natureof the headland reduces the chance ofdiscovering archaeological material insitu at this location. However, recon-naissance further into the bay led to thediscovery of another series of wave cutnotches in 10 m of water. These possibleoverhanging ‘shelters’ are fronted bya gentle sandy slope into Qumah Bayand would have been protected fromthe full force of the sea to the south.These tranquil conditions increase thechances for survival of archaeologicalmaterial, which, even if disturbed, maynot have been dispersed far from itsoriginal position. Further exploration onthis occasion was curtailed by logis-tical constraints, but we shall extendthis work in future with more detailedinvestigation of archaeological sites onland, and more extensive underwatersurvey using multi-beam bathymetry,sub-bottom profiling, shallow coring,and diving.

CONCLUSION

In reviewing the evidence of paleo-geographical and paleoclimatic change,we conclude that human transit acrossthe Bab al-Mandab Straits would havebeen relatively easy during the lower sealevels that prevailed throughout mostof the Pleistocene, and that neither thephysical barrier of a water channel norclimatic disincentives of arid climate canbe invoked to reject the likelihood ofhuman settlement and dispersal in thesouthern corridor over the time span ofthe Pleistocene as a whole. Occasion-ally, unfavorable paleogeographical andclimate conditions may have combinedto reduce the attractions of this region,but probably only for relatively shortperiods. Extreme high sea levels wouldalways have posed a more significantbarrier to crossing at the Bab al-Mandabbefore the invention of seaworthy boats.

If coastlines were especially attrac-tive places for human settlement, eitherbecause they represented primary corri-dors for initial population dispersal, orbecause they offered additional attrac-tions in the form of marine resourcesand better-watered and more fertile con-ditions on land, it is clear that settlementwas not confined to coastal zones. Dur-ing climatically favorable intervals, lakebasins developed in inland depressionsand attracted human settlement from anearly period in the Pleistocene.

Nor should we overstate the relativeattractions of the southern route incomparison with others. Lowered sealevels would probably have enhancedthe attractions of the northern routevia Suez, Sinai, and the Tigris-Euphratescatchment, as well as in the south, whilehigh sea levels would have had ratherless impact in the north. The loss of acoastal plain at high sea levels along theeastern coastline of the Mediterraneanwould have posed no serious obstacle to

154 VOLUME 2 • ISSUE 2 • 2007

Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


movement. It may even have heightenedthe attractiveness of caves and shelters,such as those of Mt. Carmel, by provid-ing easier access to marine resources aswell as the added security of complexand protective topography in a coastalsetting (Vita-Finzi and Stringer 2007).Conversely the steep topography andaridity of the Makran Coast may haveposed a serious barrier to movementduring periods of high sea level. Thisregion today is mountainous and aridwith annual rainfall of 150 mm, and isfamous as the location where the army ofAlexander the Great came to grief fromlack of food and water on their returnfrom India in 325 BC. It is, however, aregion that would have become moreattractive and easily traversable with arelatively narrow but significant addi-tion of low-lying and potentially well-watered coastal plain when sea levelswere lower, and Middle Stone Age sur-face finds from the Konarak Peninsulashow a human presence there at somepoint in the later Pleistocene (Vita-Finziand Copeland 1980).

The case for the influence of marineresources in facilitating dispersal alongPleistocene coastlines remains uncertainin the absence of clear evidence fortheir exploitation. On the other hand,the exposure of relatively well-wateredcoastal lowlands around the coastlinesof the Arabian Peninsula and the Makranmay well have been a more decisivefactor in opening up easier pathwaysfor movement to the east. One of thebiggest uncertainties in assessing thepaleogeographical possibilities in moredetail is our ignorance of conditions onthe now submerged coastal plain. Allthe indications are that when sea levelswere lower this was a key refugium forplants, animals, and humans during themore arid episodes of the glacial peri-ods. Survey of this submerged landscapefor relevant paleoenvironmental and ar-

chaeological evidence is now essential ifsuch hypotheses are to be tested.

ACKNOWLEDGEMENTS

We acknowledge funding from theNatural Environment Research Coun-cil (NERC), UK, through its EFCHEDprogram (Environmental Factors inHuman Evolution and Dispersal), theBritish Academy, and the LeverhulmeTrust, and additional funding andassistance in kind for the underwaterwork from Saudi ARAMCO, the SaudiBritish Bank (SABB), and Shell Compa-nies Overseas. For the issue of permitswe are grateful to the following Saudigovernmental organizations and indi-viduals: HRH Crown Prince Sultan BinAbdul Aziz Al Saud, Minister of Defenceand Aviation; HRH Prince Sultan BinSalman Bin Abdul Aziz, Secretary Gen-eral to the Supreme Commission forTourism; the Deputy Ministry of Antiq-uities and Museums; the Military Sur-vey Department of the Ministry of De-fence and Aviation; the Saudi BorderGuard; Prof. Saad Al-Rashid and Dr. AliS. Al-Moghanam, former Deputy Minis-ters of Antiquities and Museums; Dr.Ali Al-Ghabban, Supreme Commissionfor Tourism; Major General MurayaAl-Sharani and Admiral AbdulrahmanAl-Shihiri, Military Survey Department,Ministry of Defence and Aviation; andDr. Dhaifallah AlTalhi, Deputy Min-istry of Museums and Antiquities. Wealso gratefully acknowledge the help ofHM British Ambassador to Saudi Ara-bia, Sir Sherard Cowper-Coles; Dr. AliAl-Muhana, Public Relations Advisor tothe Minister of Petroleum and MineralResources; the Governor of Farasan,Abdulrahman Mohammed Abdulhak;the Captains of Saudi ARAMCO’s oilvessel, the Midyan, Yusuf Dukak, Al-Amin Gizani, and Salem Enazi, and


Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


their crew; Captain Ahmed Mirdad andMohammed Saber of Saudi ARAMCO,Jeddah; and our representatives in thefield, Faisal Tamaihi, Deputy Ministryof Antiquities and Museums, Sabiya,and Lt. Cdr. Abdulla M. Ahmari, Min-istry of Defence and Aviation, GeneralStaff Headquarters, Military SurveyDepartment, Riyadh. We also thankthe Hampshire and Wight Trust forMaritime Archaeology for provisionof survey equipment and additionalsupport; Eva Laurie, Department ofArchaeology, University of York, forthe identification of molluscan taxa;Eliyahu Biton, Department of Environ-mental Sciences, Weizmann Institute,for information on current flows in theHanish Channel; and two anonymousassessors for their comments.

REFERENCES

Al-Ma’mary, A. R. 2002. Archaeological findsin geological layers from Sana’a University.Chroniques Yemenites. Sana’a: French Insti-tute for Archaeological and Social Studies (inArabic).

Al-Sharekh, A. M. 2006. The Archaeology ofCentral Saudi Arabia: Investigations of LithicArtefacts and Stone Structures in NortheastRiyadh. Riyadh: Deputy Ministry for Antiqui-ties and Museums.

Allen, J. 1989. When did humans first coloniseAustralia? Search 20:149–154.

Allen, J. and S. Holdaway. 1995. The contamina-tion of Pleistocene radiocarbon determinationsin Australia. Antiquity 69:101–112.

Amirkhanov, K. A. 1991. The Palaeolithic ofSouthern Arabia. Moscow: Nauka (in Russian).

Arz, H. W., F. Lamy, J. Patzold, P. J. Muller,and M. Prins. 2003. Mediterranean moisturesource for an Early-Holocene humid periodin the northern Red Sea. Science 300:118–121.

Assaf, G. 1977. Sea straits and glacial periods inthe Red Sea. Science 195(4273):90.

Avery, G., K. Cruz-Uribe, P. Goldberg, F. E. Grine,R. G. Klein, M. J. Lenardi, and C. W. Marean,et al. 1997. Excavations at the Die Kelders Mid-dle and Late Stone Age cave site, South Africa.Journal of Field Archaeology 24(4):263–291.

Ayele, A., E. Jacques, M. Kassim, T. Kidane, A.Omar, S. Tait, A. Nercessian, J. B. de Chabalier,and G. C. P. King. 2007. The volcano-seismiccrisis in Afar, Ethiopia, starting September2005. Earth and Planetary Science Letters255:177–187.

Bailey, G. 2004a. World prehistory from the mar-gins: The role of coastlines in human evolution.Journal of Interdisciplinary Studies in Historyand Archaeology 1(1):39–50.

Bailey, G. 2004b. The wider significance of sub-merged archaeological sites and their relevanceto world prehistory. In Submarine PrehistoricArchaeology of the North Sea: Research Prior-ities and Collaboration with Industry (N. C.Flemming, ed.): 3–10. London: CBA ResearchReport 141.

Bailey, G., A. Al-Sharekh, N. Flemming, K.Lambeck, G. Momber, A. Sinclair, and C. Vita-Finzi. 2007. Coastal prehistory in the southernRed Sea Basin, underwater archaeology, andthe Farasan Islands. Proceedings of the Semi-nar for Arabian Studies: 37.

Bailey, G., K. Lambeck, C. Vita-Finzi, and A.Al-Sharekh. In prep. Sea-level change and thearchaeology of human dispersal in the Red Searegion (to be submitted for publication).

Bailey, G. N. and N. J. Milner. 2002. Coastalhunters and gatherers and social evolution:Marginal or central? Before Farming. The Ar-chaeology of Old World Hunter-Gatherers3–4(1):1–15.

Bednarik, R. G. 2003. Seafaring in the Pleistocene.Cambridge Archaeological Journal 13(1):41–66.

Beech, M. J. 2004. In the Lands of theIchthyophagi: Modelling Fish Exploitation inthe Arabian Gulf and Gulf of Oman from the5th Millennium BC to the Late Islamic Period.Oxford: Archaeopress. BAR International Se-ries. 1217.

Ben-Menahem, A. 1991. Four thousand years ofseismicity along the Dead Sea Rift. Journal ofGeophysical Research 96:20195–20216.

Beyin, A. 2006. The Bab al Mandab vs the Nile-Levant: An appraisal of the two dispersal routesfor early modern humans out of Africa. AfricanArchaeological Review 23:5–30.

Biagi, P. and R. Nisbet. 2006. The prehistoricfisher-gatherers of the western coast of theArabian Sea: A case of seasonal sedentarization?World Archaeology 38(2):220–238.

Birdsell, J. B. 1977. The recalibration of aparadigm for the first peopling of GreaterAustralia. In Sunda and Sahul: PrehistoricStudies of Southeast Asia, Melanesia andAustralia (J. Allen, J. Golson, and R. Jones,eds.):113–167. London: Academic Press.

156 VOLUME 2 • ISSUE 2 • 2007

Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


Bonatti, E. 1985. Punctiform initiation of seafloor-spreading in the Red Sea. Nature 316:33–37.

Bosence, D. W. J., M. H. Al-Aawah, I. Davison, B. R.Rosen, C. Vita-Finzi, and E. Whitaker. 1998.Salt domes and their control on basin marginsedimentation: A case study from the TihamaPlain, Yemen. In Sedimentation and Tectonicsof Rift Basins: Red Sea—Gulf of Aden (B. H.Purser and D. W. J. Bosence, eds.):448–478.London: Chapman and Hall.

Bosworth, W. and M. Taviani. 1996. Late Quater-nary reorientation of stress field and extensiondirection in the southern Gulf of Suez, Egypt:Evidence from uplifted coral terraces, meso-scopic fault arrays, and borehole breakouts.Tectonics 15:791–802.

Bowdler, S. 1977. The coastal colonisation ofAustralia. In Sunda and Sahul: PrehistoricStudies in Southeast Asia, Melanesia andAustralia (J. Allen, J. Golson, and R. Jones,eds.):205–246. London and New York: Aca-demic Press.

Bowdler, S. 1990. Peopling Australasia: The‘coastal colonization’ hypothesis re-examined.In The Emergence of Modern Humans: An Ar-chaeological Perspective (P. Mellars, ed.):327–343. Ithaca, NY: Cornell and Edinburgh Univer-sity Press.

Braithwaite, C. J. 1987. Geology and palaeo-geography of the Red Sea region. In KeyEnvironments: The Red Sea (A. J. Edwards andS. M. Head, eds.):22–44. Oxford: Pergamon.

Bruggemann, J. H., R. T. Buffler, M. M. Guillaume,R. C. Walter, R. von Cosel, B. N. Ghebretensae,and S. M. Berhe. 2004. Stratigraphy, palaeoen-vironments and model for the deposition ofthe Abdur Reef Limestone: Context for animportant archaeological site from the lastinterglacial on the Red Sea Coast of Eritrea.Paleogeography, Palaeoclimatology, Palaeoe-cology 20:179–206.

Caton-Thompson, G. 1953. Some Palaeoliths fromSouth Arabia. Proceedings of the PrehistoricSociety 19:189–218.

Chappell, J. and N. J. Shackleton. 1986. Oxygenisotopes and sea level. Nature 324:137–140.

Cherry, J. F. 1990. The first colonization ofthe Mediterranean Islands: A review of recentresearch. Journal of Mediterranean Archaeol-ogy 3:145–221.

Chu, D. and R. G. Gordon. 1998. Current platemotions across the Red Sea. Geophysical Jour-nal International 135:313–328.

Dabbagh, A., H. Hotzl, and H. Schnier. 1984.Farasan Islands. General considerations andgeological structure. In Quaternary Period inSaudi Arabia, Vol. 2 (A. R. Jado and J. G.

Zotl, eds.):212–220. Vienna and New York:Springer-Verlag.

Dennell, R. and W. Roebroeks. 2005. An Asianperspective on early human dispersal fromAfrica. Nature 438:1099–1104.

Derricourt, R. 2005. Getting “out of Africa”:Sea crossings, land crossing and culture inthe hominin migrations. Journal of WorldPrehistory 19(2):119–132.

Deuser, W. G., E. H. Ross, and L. S. Waterman.1976. Glacial and pluvial periods: Their re-lationship revealed by Pleistocene sedimentsof the Red Sea and Gulf of Aden. Science191(4232):1168–1170.

Drury, S. A., W. Ghebreab, M. E. AndrewsDeller, C. J. Talbot, and S. M. Berhe. 2006.A comment on “Geomorphic development ofthe Eritrean margin, southern Red Sea fromcombined apatite fission-track and (U-Th)/Hethermochronometry” by Balestrieri, M. L. et al.[Earth Planet. Sci. Lett. 231 (2005) 97–110].Earth and Planetary Science Letters 242:428–432.

Erlandson, J. M. 2001. The archaeology of aquaticadaptations: Paradigms for a new millennium.Journal of Archaeological Research 9:287–350.

Erlandson, J. 2002. Anatomically modern humans,maritime voyaging, and the Pleistocene colo-nization of the Americas. In The First Ameri-cans: The Pleistocene Colonization of the NewWorld (N. G. Jablonski, ed.):1–19. Memoirs ofthe California Academy of Sciences 27. SanFrancisco: University of California Press.

Erlandson, J. M. and S. M. Fitzpatrick. 2006.Oceans, islands, and coasts: Current perspec-tives on the role of the sea in human prehistory.Journal of Island and Coastal Archaeology1:5–32.

Faure, H. and C. Roubet. 1968. Decouverte d’unbiface Acheuleen dans les calcaires marinsdu golfe Pleistocene de l’Afar (Mer Rouge,Ethiope). Comptes Rendus Hebdomadairesdes Seances de l’Academie des Sciences Paris267:18–21.

Faure, H., R. C. Walter, and D. R. Grant. 2002.The coastal oasis: Ice age springs on emergedcontinental shelves. Global and PlanetaryChange 33:47–56.

Felis, T., J. Patzold, Y. Loya, M. Fine, A. H. Nawar,and G. Wefer. 2000. A coral oxygen isotoperecord from the northern Red Sea documentingNAO, ENSO, and North Pacific teleconnectionson Middle East climate variability since the year1750. Paleoceanography 15:679–694.

Fenton, M., S. Geiselhart, E. J. Rohling, and C.Hemleben. 2000. Aplanktonic zones in the


Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


Red Sea. Marine Micropaleontology 40:277–294.

Fernandes, C. A., E. J. Rohling, and M. Siddal.2006. Absence of post-Miocene Red Sea landbridges: Biogeographic implications. Journalof Biogeography 33:961–966.

Field, J. S. and M. M. Lahr. 2005. Assessmentof the southern dispersal: GIS based analysesof potential routes at oxygen isotope stage 4.Journal of World Prehistory 19:1–45.

Field, J. S., M. D. Petraglia, and M. M. Lahr.2007. The southern dispersal hypothesis andthe South Asian archaeological record: Exam-ination of dispersal routes through GIS analy-sis. Journal of Anthropological Archaeology26(1):88–108.

Flemming, N., G. Bailey, V. Courtillot, G. King,K. Lambeck, F. Ryerson, and C. Vita-Finzi.2003. Coastal and marine palaeo-environmentsand human dispersal points across the Africa-Eurasia boundary. In The Maritime and Under-water Heritage (C. A. Brebbia and T. Gambin,eds.): 61–74. Southampton: Wessex Institute ofTechnology Press.

Forster, P. and S. Matsumura. 2005. Did earlyhumans go north or south? Science 308:965–966.

Gamble, C. 1993. Timewalkers: The Prehistoryof Global Colonization. London: Sutton.

Garfunkel, Z. and Z. Ben-Avraham. 1996. Thestructure of the Dead Sea Basin. Tectonophysics266:155–76.

Garfunkel, Z., I. Zak, and R. Freund. 1981. Activefaulting in the Dead Sea Rift. Tectonophysics80(1–4):1–26.

Girdler, R. W. and P. Styles. 1974. Two stage RedSea floor spreading. Nature 247:7–11.

Girdler, R. W. and R. B. Whitmarsh. 1974.Miocene evaporites in Red Sea cores, theirrelevance to the problem of the width and ageof oceanic crust beneath the Red Sea. InitialReport of the Deep Sea Drilling Program23:913–921.

Gvirtzman, G. 1994. Fluctuations of sea levelduring the past 400,000 years: The record ofSinai, Egypt (northern Red Sea). Coral Reefs13:203–214.

Hardy, Sir Alister. 1960. Was man more aquatic inthe past? New Scientist 7:642–645.

Head, S. M. 1987. Introduction. In Key Environ-ments: The Red Sea (A. J. Edwards and S. M.Head, eds.):1–21. Oxford: Pergamon.

Hemleben, C., D. Meischner, R. Zahn, A.Almogi-Labin, H. Erlenkeuser, and B. Hiller.1996. Three hundred eighty thousand year longstable isotope and faunal records from the RedSea: Influence of global sea level change on

hydrography. Paleoceanography 11(2):147–156.

Henshilwood, C. S. and C. W. Marean. 2003.The origin of modern behavior: Critique ofthe models and their text implications. CurrentAnthropology 44(5):627–651.

Henshilwood, C. S., J. C. Sealy, R. Yates, K. Cruz-Uribe, P. Goldberg, F. E. Grine, R. G. Klein,C. Poggenpoel, K. van Niekerk, and I. Watts.2001. Blombos Cave, southern Cape, SouthAfrica: Preliminary report on the 1992–1999excavations of the Middle Stone Age levels.Journal of Archaeological Science 28(4):421–448.

Hubert-Ferrari, A., R. Armijo, G. C. P. King,B. Meyer, and A. Barka. 2002. Morphology,displacement and slip rates along the NorthAnatolian Fault, Turkey. Journal of Geophys-ical Research 107 (0,10.1029/2001JB000393,2002):ETG:X1–X32.

Hubert-Ferrari, A., G. C. P. King, I. Manighetti,R. Armijo, B. Meyer, and P. Tapponnier.2003. Long-term elasticity in the continentallithosphere: Modelling the Aden Ridge prop-agation and the Anatolian extrusion process.Geophysical Journal International 153:111–132.

King, G. C. P. and G. N. Bailey. 2006. Tectonicsand Human Evolution. Antiquity 80(308):265–286.

Klein, R. G. 1999. The Human Career: HumanBiological and Cultural Origins, 2nd ed.Chicago: University of Chicago Press.

Klein, R. G., G. Avery, K. Cruz-Uribe, D. Halkett,J. E. Parkington, T. Steele, T. P. Volman, andR. Yates. 2004. The Ysterfontein 1 MiddleStone Age site, South Africa, and early humanexploitation of coastal resources. Proceed-ings of the National Academy of Sciences101(16):5708–5715.

Lahr, M. and R. Foley. 1994. Multiple dispersalsand modern human origins. Evolutionary An-thropology 3(2):48–60.

Lambeck, K. 1995. Late Devensian and Holoceneshorelines of the British Isles and North Seafrom models of glacio-hydro-isostatic rebound.Journal of the Geological Society, London152:437–448.

Lambeck, K. 1996a. Sea-level change and shore-line evolution in Aegean Greece since UpperPalaeolithic time. Antiquity 70:588–611.

Lambeck, K. 1996b. Shoreline reconstructions forthe Persian Gulf since the last glacial maximum.Earth and Planetary Science Letters 142:43–57.

Lambeck, K. 2004. Sea-level change through thelast glacial cycle: Geophysical, glaciological

158 VOLUME 2 • ISSUE 2 • 2007

Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


and palaeogeographic consequences. ComptesRendus Geoscience 336:677–689.

Lambeck, K. and J. Chappell. 2001. Sea levelchange through the last glacial cycle. Science292:679–686.

Lourandos, H. 1997. Continent of Hunter-gatherers. Cambridge: Cambridge UniversityPress.

Macaulay, V., C. Hill, A. Achilli, C. Rengo, D.Clarke, W. Meehan, J. Blackburn, et al. 2005.Single, rapid coastal settlement of Asia re-vealed by analysis of complete mitochondrialgenomes. Science 308:1034–1036.

Macfadyen, W. A. 1930. The geology of theFarasan Islands, Gizan and Kamaran Island, RedSea. Geological Magazine:310–332.

Mandryk, C. A., H. Josenhans, D. W. Fedje,and R. W. Mathewes. 2001. Late Quaternarypaleoenvironments of Northwestern NorthAmerica: Implications for inland versus coastalmigration routes. Quaternary Science Reviews20:301–314.

Manighetti, I., P. Tapponnier, V. Courtillot, S.Gruszow, and P. Gillot. 1997. Propagation ofrifting along the Arabia-Somalia plate boundary:The Gulfs of Aden and Tadjoura. Journal ofGeophysical Research 102:2681–2710.

Mannino, M. A. and K. D. Thomas. 2002. Deple-tion of a resource? The impact of prehistorichuman foraging on intertidal mollusc commu-nities and its significance for human settlement,mobility and dispersal. World Archaeology33:452–474.

Marco, S., M. Stein, A. Agnon, and H. Ron. 1996.Long-term earthquake clustering: A 50,000-yearpaleoseismic record in the Dead Sea Graben.Journal of Geophysical Research 101:6179–6191.

McClure, H. 1976. Radiocarbon chronology oflate Quaternary lakes in the Arabian Desert.Nature 263:755–756.

McClusky, S., R. Reilinger, S. Mahmoud, D. BenSari, and A. Tealeb. 2003. GPS constraintson Africa (Nubia) and Arabia plate motions.Geophysical Journal International 155:126–138.

Meehan, B. 1982. Shell Bed to Shell Midden.Canberra: Australian Institute of AboriginalStudies.

Mithen, S. J. 2002. Stepping out: A com-puter simulation of hominid dispersal fromAfrica. Journal of Human Evolution 43:433–462.

Morgan, E. 1982. The Aquatic Ape. London:Souvenir Press.

Morgan, L. H. 1877. Ancient Society, or Re-searches in the Lines of Human Progress from

Savagery through Barbarism to Civilization.New York: Holt.

Morwood, M. J., F. Aziz, P. O’Sullivan, Nasruddin,D. R. Hobbs, and A. Raza. 1999. Archaeologicaland palaeontological research in central Flores,east Indonesia: Results of fieldwork 1997–98.Antiquity 73:273–286.

Omar, G. I. and M. S. Steckler. 1995. Fissiontrack evidence on the initial rifting of theRed Sea: Two pulses, no propagation. Science270:1341–1344.

Parker, A. G., L. Eckersely, M. M. Smith, A. S.Goudie, S. Stokes, S. Ward, K. White, and M. J.Hodson. 2004. Holocene vegetation dynamicsin the northeastern Rub’ al-Khali desert, Ara-bian Peninsula: A phytolith, pollen and carbonisotope study. Journal of Quaternary Science19(7):665–676.

Parker, A. G., A. S. Goudie, S. Stokes, K. White,M. J. Hodson, M. Manning, and D. Kennet.2006. A record of Holocene climate changefrom lake geochemical analyses in southeasternArabia. Quaternary Research 66(3):465–476.

Petraglia, M. 2003. The Lower Palaeolithic of theArabian Peninsula: Occupations, adaptations,and dispersals. Journal of World Prehistory17:141–179.

Petraglia, M. and A. Alsharekh. 2003. The MiddlePalaeolithic of Arabia: Implications for modernhuman origins, behaviour and dispersals. An-tiquity 77(298):671–684.

Plaziat, J.-C., F. Baltzer, A. Choukroi, O. Conchon,P. Freytet, F. Orszag-Sperber, A. Raguideau, andJ.-L. Reyss. 1998. Quaternary marine and conti-nental sedimentation in the northern Red Seaand Gulf of Suez (Egyptian coast): Influencesof rift tectonics, climatic changes and sea-levelfluctuations. In Sedimentation and Tectonicsof Rift Basins: Red Sea—Gulf of Aden (B. H.Purser and D. W. Bosence, eds.):537–573.London: Chapman and Hall.

Roberts, R. G., R. Jones, and M. A. Smith. 1994.Beyond the radiocarbon barrier in Australianprehistory. Antiquity 68(260):611–616.

Rohling, E. J., T. R. Cane, S. Cooke, M. Sprovieri,I. Bouloubassi, K. C. Emeis, and R. Schiebel,et al. 2002. African Monsoon variability duringthe previous interglacial maximum. Earth andPlanetary Science Letters 6304:1–15.

Rose, J. I. 2004. The question of UpperPleistocene connections between East Africaand South Arabia. Current Anthropology45(4):551–555.

Sanlaville, P. 1992. Changements climatiquesdans la Peninsule Arabique durant lePleistocene Superieur et l’Holocene.Paleorient 18(1): 5–26.


Dow

nloa

ded

By:

[Aus

tralia

n N

atio

nal U

nive

rsity

Lib

rary

] At:

05:3

5 8

Nov

embe

r 200

7


Sauer, C. O. 1962. Seashore—Primitive home ofman? Proceedings of the American Philosoph-ical Society 106:41–47.

Schultz, E. and J. W. Whitney. 1986. UpperPleistocene and Holocene lakes in the AnNafud, Saudi Arabia. Hydrobiologia 143:175–190.

Shackleton, N. J. 1987. Oxygen isotopes, icevolume and sea level. Quaternary ScienceReviews 6:183–90.

Siddall, M., E. J. Rohling, A. Almogi-Labin, C.Hemleben, D. Meischner, I. Schmelzer, andD. A. Smeed. 2003. Sea-level fluctuations duringthe last glacial cycle. Nature 423:853–858.

Siddall, M., D. A. Smeed, S. Matthiesem, and E. J.Rohling. 2002. Modelling the seasonal cycle ofthe exchange flow in Bab el Mandab (Red Sea).Deep-Sea Research I:1551–1569.

Stanford, D. and B. Bradley. 2002. Ocean trailsand prairie paths? Thoughts about Clovis ori-gins. In The First Americans: The PleistoceneColonization of the New World (N. G. Jablon-ski, ed.):255–271. Memoirs of the CaliforniaAcademy of Sciences 27. San Francisco: Uni-versity of California Press.

Straus, L. G., D. J. Meltzer, and T. Goebel. 2005.Ice Age Atlantis? Exploring the Solutrean-Clovis‘connection’. World Archaeology 37:507–532.

Stringer, C. 2000. Coasting out of Africa. Nature405:53–55.

Templeton, A. 2002. Out of Africa again and again.Nature 415:45–51.

Thangaraj, K., G. Chaubey, T. Kivisild, A. G.Reddy, V. K. Singh, A. A. Rasalkar, and L. Singh.2005. Reconstructing the origin of AndamanIslanders. Science 308:996.

Turner, A. and H. O’Regan. 2007. Afro-Eurasianmammalian fauna and early hominin dispersals.In The Evolution and History of HumanPopulations in South Asia (M. Petraglia andB. Allchin, eds.):23–39. New York: Springer.

Van Andel, T. 1989a. Late Quaternary sea-levelchanges and archaeology. Antiquity 63(241):733–745.

Van Andel, T. 1989b. Late Pleistocene sea levelsand the human exploitation of the shore andshelf of southern South Africa. Journal of FieldArchaeology 16:133–155.

Vita-Finzi, C. 1973. Palaeolithic finds fromCyprus? Proceedings of the Prehistoric Society39:453–454.

Vita-Finzi, C. and L. Copeland. 1980. Surface findsfrom Iranian Makran. Iran 18:149–155.

Vita-Finzi, C. and B. Spiro. 2006. Isotopic gaugesof deformation in the Red Sea. Journal ofStructural Geology 28:1114–1122.

Vita-Finzi, C. and C. Stringer. 2007. The setting ofthe Mt. Carmel caves reassessed. QuaternaryScience Reviews 26(3–4):436–440.

Walter, R. C., R. T. Buffler, J. J. Bruggemann,M. M. M. Guillaume, S. M. Berhe, B. Negassi,Y. Libsekal, et al. 2000. Early human occupa-tion of the Red Sea Coast of Eritrea during theLast Interglacial. Nature 405:65–69.

Warren, J. K. 1999. Evaporites, Their Evolutionand Economics. Oxford: Blackwell.

Whalen, N. M., J. S. Siraj-Ali, and W. Davis.1984. Excavation of Acheulean sites near Saf-faqah, Saudi Arabia, 1403 AH 1983. Atlal, theJournal of Saudi Arabian Archaeology 8:9–24.

Zarins, J., A. Al-Jawad Murad, and K. S. Al-Yish. 1981. The Comprehensive Archaeologi-cal Survey Program, a. The second preliminaryreport on the southwestern province. Atlal,the Journal of Saudi Arabian Archaeology5:9–42.

Zarins, J. and A. Zahrani. 1985. Recent archaeo-logical investigations in the Southern TihamaPlain (The sites of Athar, and Sihi, 1404/1984).Atlal, the Journal of Saudi Arabian Archaeol-ogy 9:65–107.

160 VOLUME 2 • ISSUE 2 • 2007

The Journal of Island and Coastal Archaeologypeople.rses.anu.edu.au/lambeck_k/pdf/264.pdf · The...

Documents

Transcript of The Journal of Island and Coastal Archaeologypeople.rses.anu.edu.au/lambeck_k/pdf/264.pdf · The...