Where River Meets Sea.pdf

315

Where River Meets Sea

The Southern-Dispersal Hypothesis

A Parsimonious Model for Homo sapiens

Colonization of the Indian Ocean Rim and Sahul

David Bulbeck

For approximately the past decade, the Out-of-Africa or re-

placement theory has enjoyed the status of the new orthodoxy

on the origins of H. sapiens. It is certainly compatible with

the fossil record (Stringer and McKie 1996; Cameron and

Groves 2004) and is strongly supported by the mitochondrial

DNA and Y-chromosome evidence (Oppenheimer 2003; Har-

ding and Liu 2005). As emphasized by Oppenheimer (2004,

175–84), Macaulay et al. (2005 a ), and Thangarej et al. (2005),

the rakelike structure of deeply rooted mtDNA lineages (hap-

lotypes M and N) right along South Asia, Southeast Asia, and

Australia–New Guinea indicates a Late Pleistocene dispersal

event that was rapid on an archaeological time scale.

The most creditable opposition comes from geneticists

working with nuclear DNA, who reconstruct time depths for

many non-African polymorphisms on the order of hundreds

of thousands of years. Commenting on the three Science ar-

ticles referenced above, Harpending and Eswaran (2005)

brieﬂy summarized the misgivings of human nuclear genet-

icists about the Out-of-Africa theory. In response, Macaulay

et al. ( 2005 b ) pointed to methodological problems with the

time depths estimated by nuclear geneticists and noted that

our article (Macaulay et al. 2005 a ) had not touched on

whether H. sapiens might have incorporated relics of other

populations during its Out-of-Africa expansion. I leave ad-

judication on the genetic evidence to my qualiﬁed colleagues

(Richards et al. 2006) and address this article to readers not

entirely opposed to the Out-of-Africa theory.

The rapid-southern-dispersal hypothesis could possibly be

critiqued in terms of the absence of a rationale for the col-

onists’ having kept moving ever eastward. Spencer Wells (re-

ported in Shreeve 2006, 68) suggests that “it was less of a

journey and more like walking a little further down the beach

to get away from the crowd.” This idea, however, raises the

questions why individual colonists had wished to escape their

social milieu and how they had survived away from their

group. The colonization model presented here will resolve

such questions. It is inspired more by the mtDNA evidence

than by other considerations, though its relevance for ar-

chaeology will be addressed.

School of Archaeology and Anthropology, The Australian

National University, ACT 0200, Australia (david.bulbeck@

anu.edu.au). 31 X 06

Recent genetic research suggests an expansion along the trop-

ical coastline of the Indian Ocean, between 75,000 and 60,000

years ago, of the population which included the ancestors of

all of the non-African human mitochondrial DNA lineages

known today. In view of the arid sections along this coastal

stretch, irregularly punctuated by resource-rich estuaries, and

the crossings over open sea during the last leg to Australia/

New Guinea, this expansion would necessarily have involved

the features of watercraft, portage of potable water, and ad-

aptation to estuaries. These features could well have been the

cultural basis for the rapid tropical dispersal of Homo sapiens

out of Africa to Australia/New Guinea.

The Out-of-Africa theory proposes a Late Pleistocene expan-

sion of Homo sapiens from its African cradle into Eurasia.

Proponents of the theory have disagreed on whether there

had been a southern dispersal along the coastal belt of Arabia

and southern Asia, a more inland dispersal through the Mid-

dle East, or both (Kingdon 1993; Lahr and Foley 1994; Klein

1999; Oppenheimer 2003, 2004; Mellars 2006). The southern-

dispersal hypothesis is strengthened by mtDNA research

which suggests a rapid eastward migration of Homo sapiens

along the northern rim of the Indian Ocean, starting from

East Africa at around 75,000 years ago (Macaulay et al. 2005 a ;

Thangarej et al. 2005). As Forster and Matsumura (2005)

remark, this hypothesis has profound implications for un-

derstanding when early colonists to Wallacea and Sahul staged

the necessary water crossings from Sunda (ﬁg. 1). In this

contribution I propose a model, based on an estuarine ad-

aptation, to account for the average rate of 0.7 to 4 km per

year calculated by Macaulay et al. (2005 a ) for the eastward

dispersal of H. sapiens. It is neither a “push” nor a “pull”

model and does not presume of the colonists any grand vision

of occupying and populating unknown terrain. Instead, it

proposes a tradition of seeking out suitable estuarine habitat

as the main impetus for the migratory movement.

Colonization along the Northern Rim of

the Indian Ocean

Well before the Out-of-Africa theory took root, Sauer (1963,

311) had proposed that “the dispersal of early man took place

most readily by following along the seashore.” Sauer pointed

out that coastally adapted populations would have had access

to a continuous string of familiar habitats little affected by

seasonal changes to the climate and biota. With regard to arid

landmasses, Sauer argued that desert shores are richer in food

resources than the adjacent hinterland and that even here

potable water could be found in seeps at low tides. “Coastwise

2007 by The Wenner-Gren Foundation for Anthropological Research.

316

Current Anthropology Volume 48, Number 2, April 2007

Figure 1. Proposed Out-of-Africa dispersal routes and sites mentioned

in the text.

there was scarcely a barrier to the spread of early man through

tropical and subtropical latitudes” (p. 312).

Sauer’s line of thought was further developed by Kingdon’s

(1993) attempt to relate the Out-of-Africa theory to mam-

malian dispersal routes. Kingdon suggested that Homo sapiens

strandloopers had essentially followed the marine shoreline

(or dugong–ﬂying fox) route from the Bay of Bengal to Sahul

in one direction and along South Asia to East Africa in the

other (ﬁg. 1). He noted the similarity in resources between

tropical rivers and tropical seas, especially at river mouths, in

mangrove swamps, and on the foreshore. He also remarked

that watercraft would have opened up reefs and marine shal-

lows as niches for human foragers. As for the original mi-

gration of Homo sapiens from Africa to the Bay of Bengal,

Kingdon hypothesized that these emigrants had taken the

tropical Indian or monkey-porcupine route. He dated this

event to 80,000–115,000 years ago on the basis of the dates

then available for anatomically modern skeletons from Skhul

and Qafzeh in the Levant (Kingdon 1993, 77–80, 120–22,

135–36).

Lahr and Foley (1994) foreshadowed the southern-dispersal

hypothesis in their model of two movements by Homo sapiens

out of Africa: a lesser migration along the northern rim of

the Indian Ocean and a subsequent main migration through

Asia Minor. Oppenheimer (2003) reversed the order and im-

portance of these dispersals. He associated the Skhul/Qafzeh

people with a failed early exodus from Africa, as their age of

90,000–120,000 years antedates the estimated age of

70,000–85,000 years for the common ancestor of all non-

African mtDNA. He also noted that at around 80,000 years

ago the overland route out of Africa would have been con-

fronted by continuous desert. In contrast, the southerly route

across the Hormuz Strait would have been assisted by patches

of coastal scrub along East Africa and the southern Arabian

Peninsula. Observing that the Eritrean site of Abdul docu-

mented the human exploitation of reef resources by 125,000

years ago, Oppenheimer (2003) proposed a dispersal of

“beachcombers” from East Africa to Sahul along the northern

rim of the Indian Ocean (ﬁg. 1). Macaulay et al. (2005 a )

subsequently proposed a slightly later chronology, beginning

in the Horn of Africa at around 75,000 years ago and reaching

Sahul at around 60,000 years ago.

The ﬁnal leg to Sahul would clearly have involved watercraft

capable of voyages of 50–100 kilometers (Mulvaney and Kam-

minga 1999, 109–10). While the colonization of Australia by

at least

years ago is clear (Gillespie 2002), the Wal-

50,000

lacea archaeological record is more complex. Flores, with its

highly distinctive record of hinterland habitation from the

Early Pleistocene till 12,000 years ago (Morwood et al. 2005;

Brumm et al. 2006), stands apart from the other islands of

Wallacea, whose well-documented archaeological time depth

is consistently 35,000 years or less (Bellwood 2001; Bulbeck

∼

317

2004). Irrespective of the enigma surrounding the precocious

settlement of Flores, there is minimal evidence for an Aus-

tralasian maritime capacity that could have reached Australia

before 60,000 years ago (Gillespie 2002; O’Connor and Chap-

pell 2003) or a widespread colonization of Wallacea before

35,000 years ago.

The arrival of Homo sapiens in Australasia at around 60,000

years ago, as indicated by the genetic evidence (see also van

Holst Pellekaan et al. 2006), is the likely cause for the new-

found capacity to colonize Australia. Given the dating im-

precisions, this could have occurred as much as 10,000 years

after Homo sapiens had reached Sunda—a long period of time

for the local development of a maritime capacity. By the same

token, the crossing to Sahul could have occurred instanta-

neously (in archaeological terms), as asserted by the rapid-

southern-dispersal hypothesis. In that case it would be likely

the beachcombers already had boats when they arrived rather

than fortuitously inventing them on their arrival. This is en-

tirely reasonable in view of the archaeological evidence from

East Africa of fully modern capacities in artifact manufacture

and exploitation of littoral and reef resources by 125,000 years

ago (Walter et al. 2000). Thus, acceptance of the rapid-south-

ern-dispersal hypothesis would strongly suggest watercraft

assistance.

enclosed coastal embayment where fresh water and seawater

meet and mix,” as the point of origin for successful lengthy

forays.

All of the advantages mentioned by Sauer (1963) and King-

don (1993) for coastal foragers would have been optimized

in the Indian Ocean’s estuaries. Timber (for boat building)

and potable water would have been available, even if requiring

some upstream travel. The occurrence of coral reefs, sea-grass

beds, and/or mangrove belts along the northern rim of the

Indian Ocean (Sheppard 2000) would have supported a sim-

ilar range and abundance of marine and littoral resources

along the entire route. This continuity of familiar resources

would also have prevailed at river mouths and stream outlets,

with many of the same or closely related organisms as are

found at other coastal locations (Nybakken 2001, 335–38).

Finally, watercraft would have provided protection against

saltwater crocodiles and other potential dangers.

Given their rich resources, estuaries would have supported

healthy population growth, conveying an impression of

crowded conditions even if ﬂanked by stretches of vacant land.

This would have stimulated the search for new estuarine hab-

itats, which, to the experienced, advertise their presence out

to sea. Colonists towards the front of the wave would have

known how to travel back to their parent community, when-

ever conditions were ripe, producing a chain of interlinked

estuarine settlements (cf. Moore 2001). In summary, an es-

tuarine specialization would have been the ideal adaptation

to carry colonizing groups out of Africa past the Arabian

Peninsula—a demanding section of the route in terms of

drinking water (Oppenheimer 2004, 68; Petraglia and Al-

sharekh 2003, 681)—to South Asia and onward to Sunda.

Virtually the entire route is tropical (ﬁg. 1), minimizing bar-

riers against the passage of individuals maladapted to cooler

conditions.

The questions raised earlier with respect to Wells’s suggestion

of “beachcomber creep” are readily resolved by the model of

estuarine-focused paddlers. The model speciﬁcally invokes a

chain of linked estuarine nodes with regular trafﬁc between

neighbors and a “front” of the chain at any point in time.

Onward colonization at the front would have been achieved

by individuals or groups seeking out new habitat and/or being

blown along the coast. The capacity would always have been

present for community members to explore unfamiliar coast-

line until they could return with news of a suitable estuary or

until their water dwindled to a bare sufﬁciency for the return

journey—a version writ small of Irwin’s (1992) model for the

colonization of the Paciﬁc. Thus, expansion of the social milieu

at its front would have been a function of maintaining the

social milieu, equipped with technology that both held society

together and allowed it to expand. By assuming serviceable

watercraft, portage of potable water, and an estuarine focus, no

further assumptions would be necessary to underwrite the

rapid-southern-dispersal hypothesis.

Watercraft and Estuaries in the Southern Dispersal

We may assume that early watercraft was paddle-propelled,

limiting its role in rapid transport to rare or inadvertent oc-

casions. Experienced canoe paddlers today can achieve merely

6–18 miles (10–30 kilometers) per day, depending on wind

conditions (Search and Rescue Society of British Columbia

2006). This suggests a maximum allowance of 10 kilometers

on a usual day for Late Pleistocene paddlers, depending on

the quality of their craft—in any case, a shorter distance than

could be achieved by walking. However, strong winds and

especially gales enforce downwind travel on any paddle craft

caught in the open. Regular users of coastal watercraft would

have needed to be prepared for occasions when all they could

do was hold tight during a rapid maritime journey.

One major advantage boat users have over pedestrians is

the much lower friction of water compared with land. A boat

can double as a mobile shack without incurring prohibitive

transport costs. The vagaries of the wind would have exerted

selection pressures on boat users to carry containers for fresh

water along with gear for foraging and tool-kit maintenance.

These safeguards would additionally have allowed those near

the dispersal front to skirt shores bereft of even the low-tide

water seeps mentioned by Sauer (1963). Stretches of 100 ki-

lometers or more of the most forbidding coast could have

been traversed by single colonists or small cooperative groups,

intentionally or otherwise. The most critical factor for the

harshest coasts would have been to start the journey with

decent water supplies. This point suggests that we should look

at the estuary, deﬁned by Nybakken (2001, 328) as “a partially

318

Current Anthropology Volume 48, Number 2, April 2007

Usefulness of a Darwinian Perspective

coastal exploitation with simple watercraft would have been

greatest at times of rising sea levels, when reefs, mangrove

belts, and estuaries would have been richest and in closest

proximity to each other. They accordingly indicate

59,000–62,000 years ago as the optimum time for colonization

of Australia across the Timor Sea. This dating is in complete

accord with the rapid-southern-dispersal hypothesis, and the

newcomers’ tradition of watercraft and an estuarine focus, as

suggested here, would have preadapted them for the rigors

of the southern route. 1

The advantages of the northern route include intervisibility

between islands all the way to New Guinea, superior fresh

water provisioning, and an eastward direction of the pre-

vailing winds (Mulvaney and Kamminga 1999, 108). However,

the profusion of debouching streams along the coasts of

Sunda and the wetter Wallacea islands would have blurred

the distinction between estuarine and other habitats. When

human groups with a narrow niche breadth (as posited here)

came upon a wider range of options, how would they have

responded? The most likely answer is variably. Some groups

may have used the safety net of ready drinking water to un-

derwrite even more ambitious exploration of the surrounding

coasts and seas. These would have been the ﬁrst colonists of

the well-watered coasts, explaining the “maritime nomad”

adaptation (Gosden and Robertson 1991; see also Torrence

et al. 2004) that emerges in the earliest archaeology of the

Bismarck Archipelago. Other groups, however, would have

settled for the estuaries at hand, ﬁlling in the coastline with

directionally irregular movements. Of course, ﬁlling in coast-

lines, river fronts, and the landscape generally would have

been the major solution to population growth anywhere be-

hind the colonizing front, albeit involving the slower process

of individual learning.

Pardoe (2006) suggests that Sahul could have been colo-

nized through one or more organized expeditions, perhaps

numbering thousands of transmigrants. This is possible but

less parsimonious than a scenario of disorganized coloniza-

tion. The considerable number of Wallacea islands within

striking distance of Sahul (ﬁg. 1) suggests multiple sources

and destinations. Particularly with the likelihood of periodic

genetic replenishment, even a small band of arrivals may have

thrived much better than would be suggested by Moore’s

(2001) population genetic models. This is because the incest

prohibitions assumed by Moore relate speciﬁcally to ﬁlled

landscapes, not to isolated bands short of unrelated marriage

partners but very well endowed with land. There would have

been a dynamic balance between the enhanced ability of rel-

atively small parties to make early landfall and the greater

viability of larger populations of colonists (Webb 2006,

98–101). The original contribution which the present model

A Darwinian perspective would remove any need to invoke

human teleological intentionality either to exploit the synergy

between watercraft and estuarine habitation or to plan their

combination for purposeful colonization. A Darwinian par-

adigm characterizes cultures by their stream of experimental

innovation within the social and technological limits of pre-

vailing social practices (Rindos 1986). As is the case with

genetic mutations, random processes strongly affect the in-

ception and establishment (“ﬁxation”) of innovations, but if

they survive the initial trials their frequency will increase or

decrease under the natural selection mechanisms that operate

on cultural traits (Boyd and Richerson 1985, 2005). In par-

ticular, it is not necessary for an estuarine focus to have stim-

ulated innovation in watercraft or vice versa. As long as a

tradition combining watercraft and estuarine habitation hap-

pened to develop in some population, the adherents would

have reaped the beneﬁts. The incubation period may have

been selectively neutral or even maladaptive from a func-

tionalist point of view; Boyd and Richerson’s (1985, chap. 8)

model of “runaway cultural evolution” explains how cultural

evolution can, within limits, take on its own life independently

of the reproductive success of the culture bearers.

East Africa, with its record of littoral and reef exploitation

by 125,000 years ago, was a likely nursery or welcome recipient

of the watercraft-estuarine tradition at some point during the

succeeding 50,000 or so years. A “rapid” dispersal of this

tradition along the northern rim of the Indian Ocean would

have been aided by the similarity of coastal resources along

the entire route. Hinterland colonists, in contrast, would have

been severely slowed down by the need for “individual learn-

ing,” the slow process whereby Darwinian selection leads to

new modes of successful behavior in a novel environment

(Boyd and Richerson 2005, pt. 1).

Colonization of Sahul

The colonization of Sahul could have occurred along a north-

ern route through Sulawesi or a southern route via Nusa

Tenggara (ﬁg. 1). Whichever route was followed, colonization

had presumably occurred in the search for habitable islands

rather than as a heroic thrust from Sunda to Sahul. Travel

between islands could have been accelerated by paddling to-

wards parts of the sea with tell-tale signs of land nearby or

by being swept away by tides and storms (Mulvaney and

Kamminga 1999, 107–12). With regard to the model of es-

tuarine seeking developed here, preparedness for short voy-

ages over open sea would need to be added to the coastal

travel that had brought Homo sapiens to the edge of Wallacea.

An estuarine focus would have been particularly beneﬁcial,

if not obligatory, for movement along the typically dry shores

of Nusa Tenggara (Monk, de Fretes, and Reksodiharjo-Lilley

1997, 76–77) and northwestern Australia. O ’ Connor and

Chappell (2003, 21) point out that the opportunities for

1. The previous period of higher sea levels, according to O’Connor

and Chappell (2003, ﬁg. 2), would have prevailed from

70,000

skirted Arabia according to the rapid-southern-dispersal hypothesis.

∼ 75,000 ∼

years ago, just when Homo sapiens departed East Africa and

319

may make to understanding the colonization of Sahul is to

suggest a possible motive for exploration—seeking out new

estuaries—and to assign colonists a habitat search image that

would have optimized their chances of survival upon landfall.

model that did not speciﬁcally rule out use of coastal re-

sources. Conceivably, after recovering a satisfactory sample of

uplifted Indian Ocean coastal sites dated to 60,000–75,000

years ago, archaeologists may be able to determine a pattern

of increasing or decreasing site density towards ancient es-

tuaries, either consistent with or falsifying the model. This

would be a tall order but perhaps not impossible given on-

going advances in geoarchaeology and dating techniques.

Three current anomalies in Australasia’s archaeological rec-

ord can perhaps be explained by an estuarine focus. First,

O’Connor and Chappell (2003) point out that the terminal

Pleistocene in Australia seems to have witnessed colonization

from savanna to coast rather than the reverse. They do not

doubt that the initial colonists had a maritime capacity, but

they see no evidence that this had conﬁned early dispersal to

littoral belts. My model would provide an archaeologically

undetected means for colonists to have discovered and settled

estuaries, moved upriver, and adapted to savanna resources

(cf. Bowdler 1990). Secondly, early H. sapiens colonists may

have had limited advantages over earlier established Eurasian

populations except at estuaries (or perhaps the coastline gen-

erally) which they commandeered. Modeling early H. sapiens

more as supertramp (cf. Diamond 1977) than as superman

could explain the occupation of the Flores interior by H.

ﬂoresiensis till the end of the Pleistocene (Morwood et al.

2005; Brumm et al. 2006). Thirdly, despite passage through

Wallacea before 50,000–60,000 years ago, a true maritime

culture evidently appeared there only after 30,000–35,000

years ago (Bellwood 1998). The explanation provided by the

present model is that the colonists had more an estuarine

than a maritime orientation. The possibility of 50,000-year-

old sites in Sulawesi, along one of its main rivers (Keates and

Bartstra 1994), is particularly interesting in this regard.

Effective testing of the rapid-southern-dispersal hypothesis

can be expected from ongoing genetic research and its cor-

relation with growing archaeological knowledge. Merriwether

et al. (2005) note that their evidence on M haplogroups may

be consistent with the rapid-dispersal hypothesis of Macaulay

et al. (2005 a ), but it could also be explained by the existence

of ancestral hapolotype M for around 20,000 years. Should

further research bear out the latter explanation, the rapid-

southern-dispersal hypothesis would require revision. In ad-

dition, as previously implied, expert opinion is currently split

between

Testing of the Model and Its Implications

for Archaeology

The present model posits a process of colonization of estu-

arine habitats between approximately 75,000 and 60,000 years

ago. Any sites dating to the last 40,000 or 50,000 years ago,

when radiocarbon dating comes in, would refer to adaptations

postdating the colonization thrust by 10,000 to 35,000 years.

The relevance of these sites to the testing of the model would

be debatable. It may draw support from radiocarbon dates

of around 40,000 years ago at Kupona na Dari in New Britain

(Torrence et al. 2004) and Buang Merabak in New Ireland

(Leavesley et al. 2002), but alternative colonist models would

be equally comfortable with the same archaeological evidence.

Absence of evidence of 75,000–60,000-year-old sites along

the northern rim of the Indian Ocean would hardly amount

to evidence of their absence. Late Pleistocene coastal sites are

rarely available for archaeological study except where the land

has risen faster than the sea, as with the Bismarck sites, or

where a steep continental shelf translates into a short distance

from the Pleistocene coast (e.g., Mandu-Mandu in Western

Australia). Other coastal sites would have been destroyed

through wave action or, at best, buried beneath marine sed-

iment. While we may not have a single site in Arabia or

southern Asia that documents use of the coast between 60,000

and 75,000 years ago, South Asia has possible candidates,

including Badalpur (Marathe 1981, 47, 111–12; see James and

Petraglia 2005, S7), Patirajewalla Site 50 (Deraniyagala 1992,

685; Abeyratne 1996, 101), and Ramayogi Agraharam (Rath,

Reddy, and Prakash 1997). This small number should be seen

in the context of the general rarity of South and Southeast

Asian archaeological sites dated to around 60,000 years ago

(Bellwood 2001; James and Petraglia 2005; Mellars 2006). To

falsify the southern-dispersal hypothesis, we would need pos-

itive evidence of avoidance of the relevant stretch of coast.

This could come from archaeologically sterile deposits, se-

curely dated to the period concerned, in a location where we

could conﬁdently expect traces of any coastal habitation. I

am unaware of any documented case to this effect.

The prospects for archaeological testing are further dimmed

when we consider the estuarine focus of the model. During

times of falling sea levels such as the Late Glacial Maximum,

rivers would have cut down to lower levels, eroding sites along

their banks or near their mouths. Moreover, this destruction

of estuarine sites would have been most pronounced where

tectonic uplift would otherwise have lent these sites some

chance of survival. To be sure, an estuarine focus hardly im-

plies avoidance of the adjacent coast, so the model would

predict sites on uplifted land adjacent to ancient estuaries.

However, so would any other habitation or colonization

50,000

Australia. Should the former estimate become the consensus

and the departure date from East Africa of

and 60,000 years ago for initial entry to

75,000

ago still hold up, this would imply a less than rapid dispersal

to Sunda or a stasis at Sunda, either of which would be in-

consistent with the model developed here.

∼

years

Conclusion

This contribution develops an estuarine colonist model to

address and explain the rakelike structure currently observed

for the most deeply rooted mtDNA lineages outside of Africa.

∼

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika: