From Sunda to Sahul

The first crossings and early settlement of the Pacific

Lian Bua Cave in Flores, Indonesia, where the remains of Homo floresiensis were discovered in 2003


Excerpted from Voyagers: The Settlement of the Pacific by Nicholas Thomas, copyright © 2021 by Nicholas Thomas. Originally published in 2021 by Head of Zeus in the UK. First US Edition: June 2021, published by Basic Books, an imprint of Perseus Books, LLC, a subsidiary of Hachette Book Group, Inc. All rights reserved. Reprinted with permission.

It is almost impossible to imagine the worlds we now know as Maritime Southeast Asia and Australia as they were 50,000 years ago. The last great Ice Age began around 2.6 million years ago and lasted until some 11,700 years ago. The periodic advances and retreats of Antarctic ice resulted in sea levels as much as 430 feet below those of the present. Throughout the world, islands were connected for extended periods with adjacent continental masses. Java and Sumatra, together with Borneo and the now slender Malay Peninsula, were part of a vast extension of the Asian continent reaching south and east, forming a broad and massive hook around the great gulf of the South China Sea. New Guinea and Australia were connected by expansive land bridges extending north of the present Australian state of Queensland and the Northern Territory. Although the Torres Strait—now between the northern tip of Queensland and New Guinea—is nearly four times wider than the English Channel, it is also much shallower, in general no more than fifty feet deep.


    Yet there were deep water trenches that always separated what are referred to by biogeographers as Sunda (old continental Southeast Asia) and Sahul (greater Australia). Just east of Bali, which formed the far southeastern tip of Asia at its greatest extent, and west of Lombok in the archipelago of the Lesser Sunda Islands, running north to south between Borneo and the Philippines, is Wallace’s Line. This line, discovered by British zoologist and traveler Alfred Russel Wallace (1823–1913), acknowledged that the fauna of Southeast Asia was basically different from that of New Guinea and Australia, reflecting an ancient history of separation between the great regions [see “Alfred Russel Wallace: An Appreciation” by David Attenborough, Natural History, September 2015]. Wallacea, a sea  of some seventeen thousand islands, constitutes a zone between Sunda and Sahul, embracing Sulawesi and Maluku. For as long as forty million years, life in Sahul had evolved separately, resulting in the distinctive varieties of marsupial mammals that so astonished Europeans when they first encountered them. Those explorers would have been yet more astounded had they ever seen the giant marsupials that had earlier occupied the continent. Of a whole cast of great kangaroos, wombats, and lion– and rhino-like marsupials, Diprotodon optatum was the star. Weighing over 6,000 pounds and with a length of about thirteen feet, it was a herbivore and the largest marsupial known to ever exist.

Diprotodon optatum, the largest known marsupial


    Sahul’s environments included tropical rain forests, tropical seasonal forests, tropical deciduous forests, savanna woodlands and grasslands, montane forests, and subalpine and alpine regions—the highest mountains in New Guinea approach 16,500 feet above sea level—as well as swamp and mangrove forests across lowland, estuarine, and coastal regions. Fifty thousand years ago, all that these environments had in common was that humanity had yet to make anything more than the most limited impact on them.
    There is hard data about the movement of people from Sunda into Sahul, but it primarily relates to “when,” not “how” and still less “why.” Even the most basic aspects of “when” and “where” are subject to dispute and revision. The “who” question is also complex. The mainstream view used to be that the peoples who first engaged in sea travel were all Homo sapiens, and that this was true worldwide. It has been assumed that movement over water involved social coordination and therefore communication, symbolism, and other aspects of modern human identity, which have not generally been considered attributes of earlier species. The issue is now being reconsidered, alongside evidence for symbolic behavior among pre–Homo sapiens species and other discoveries over the last fifteen years. In 2004, a find was made on the eastern Indonesian island of Flores of remains attributed to a new hominid species, Homo floresiensis. The remains recovered included those of a female just one meter high, though aged about thirty at the time of her death more than 60,000 years ago.
    More recently, a further species, Homo luzonensis, has been proposed on the basis of bones found in deep layers in the sediments in Callao Cave, in northern Luzon in the Philippines. Homo luzonensis, dated 50,000–70,000 years ago, coexisted in the region with modern Homo sapiens. These finds are arresting in the sense that they call for an understanding of human evolution that is diverse and multistranded rather than marked by steady progress toward our own state. But they are also striking in that neither Flores nor Luzon were ever joined to the great Sunda landmass. Homo floresiensis and Homo luzonensis could only have reached the islands upon which they lived by crossing water. Although the gaps between the Balinese coast and the intervening islands of Lombok, Sumbawa, and Komodo are narrow and would not have involved passages of more thirty miles, it appears highly unlikely that (at a minimum) a couple able to breed would have swum (or even attempted to swim) such distances. A range of evidence has been adduced in support of the possibility that, following a tsunami, hominids clinging to tree debris might have been accidentally carried across such straits. The gap between mainland Sunda and the ancient, greater Philippines was of a similar order, but a water crossing still had to be made. The most recent modeling concludes that “the chances of randomly making the voyage to Sahul is low except when unrealistically high numbers of adults are washed off an island at unrealistically high frequencies.” Hence, while argument around the issue will surely continue, it appears that the crossings were planned, attempted, and successfully undertaken.

    Remarkable as their stories are—and further extraordinary details may be added as archaeological research advances—neither Homo floresienses nor Homo luzonensis had the migratory, expansionist tendencies of Homo sapiens. So far as can now be established, neither species ventured significantly beyond the regions in which their remains were first identified. The hominids who would have a greater and more enduring impact were Homo sapiens. The species began to move out of Africa some 120,000 years ago, was present in the Middle East at least 90,000 years ago, and was in Southeast Asia 70,000 years ago. As people migrated, they adapted to changing environments, and at times early sapiens encountered and interbred with older, non-sapiens populations who had previously left Africa. The earliest migrants into the Australasian region were bearers of a hybrid biological and cultural identity, which diverged from Eurasian populations and acquired its distinctiveness in the Sunda region.
    Dates for Aboriginal sites in Australia began to be published around 1960, but it was at the end of that decade that scientists made discoveries pointing to a far more ancient history in Australia than had been suspected. In 1969, the remains of a cremated woman were found in the sand dunes on the edge of Lake Mungo in western New South Wales. Five years later, a male skeleton was discovered in the nearby deposits. The woman’s remains were dated to 19,000–24,000 years earlier. Evidence that her body had been burned, as well as the presence of ocher from a considerable distance away, implied a ceremonial interment. The man’s remains were far older, from around 42,000 years ago. These findings were not only scientifically significant, but they pointed to new imaginings: Australia, conceived for so long as a vast, arid wilderness, was revealed as a continent with a deep inhabited history that had seen ritual for millennia. Yet its ancient civilization was one of nomads, and its deep histories were quite unlike those of other parts of the world.
    Some areas, such as the islands of Wallacea, through which humans must have initially traveled, are still to be extensively investigated. But archaeological research across the wider region has advanced dramatically, and numerous early dates for sites situated in what are now the separate landmasses of New Guinea and Australia—as well as the Bismarck Archipelago, which extends northeast of Papua— have been reliably established. The earliest dates for northern Australia suggest that human settlement may have taken place around or even before 60,000 years ago, but the accuracy of the dates has been extensively debated. There is a greater density of archaeological sites for the period 50,000–45,000 years before the present; if humans did arrive 10,000 years or so earlier, they most likely did so in very small numbers.
    The “how” is a realm of enigma in two senses. First, what route might the first voyagers have taken? The most obvious journey from the Sunda landmass to Sahul must have brought Homo sapiens into contact with Homo floresiensis. The island chain runs precisely west-east from Bali through Flores to Timor, and the distance separating each island is not great. Fifty thousand years ago, the crossing from Timor to the expanded north Australian coast was far shorter than the passage today. But, in the absence of more fine-grain archaeological evidence (for instance, a pattern of early sapiens sites through one group of islands and their absence in the other), alternative routes farther north—through Sulawesi and Ceram toward the Bird’s Head (the great peninsula of northwestern New Guinea), or some variation—cannot be ruled out.

This print from the atlas of Dumont d’Urville’s second voyage of 1837–1840 shows the scale of Tongan double canoes.

Second, what kinds of vessels enabled sea crossings to be made? As there are no boats of this age extant from any part of the world, this is entirely a matter for conjecture. It is worth observing that there is, for the likely period of first crossing, likewise no evidence for any kind of deep sea fishing. Coastal subsistence was a matter primarily of living off shellfish or other species that could be readily gathered from the shore. In other words, people did not make or use boats for fishing that could be enlarged or adapted for longer passages. The vessels most probably used were rafts or simple canoes made from materials that were not challenging or time-consuming to handle, such as a variety of bamboo. Bamboo has the great advantage of being inherently buoy-ant, as the hollow voids between the nodes are naturally air-tight. A species such as Dendrocalamus giganteus (dragon or giant bamboo), which was common throughout the region, grew swiftly and could reach over one hundred feet in height. Individual tubes could be as much as fourteen inches in diameter, meaning that comparatively few, or a mix of very thick tubes and thinner ones, could be tied together with vines to form a long but relatively narrow boat, susceptible to basic steering. Even a raft of squarer design could have been paddled and perhaps maneuvered to a limited degree; it is likely that vessels of either sort would have enabled people to cross open water to another land with the assistance of favorable winds and currents.
    Assuming a deliberate intention on the part of a group to resettle in a territory across water, a party would presumably have consisted of more than one male-female couple. Demographic modeling of the necessary size for a viable population points toward greater numbers: some hundreds of thousands, presumably arriving over decades or hundreds of years. We might assume that boats needed to carry not only people, but also their utensils and belongings, such as hunting implements, baskets, bags, garments, and items of personal adornment. If the accidental drift of refugees from some volcanic eruption or similar catastrophe cannot be absolutely excluded, the establishment of a population that survived—or, indeed, thrived and spread comparatively swiftly over new lands once they were reached—is suggestive of a venture that was both deliberately conceived and sustained over a period.
    Those who inhabited Wallacea did not become seafarers at the time of these early crossings. The evidence from the distribution of animals and from genetics does not point toward back and forth voyaging, or to an extended series of migrations by successive groups. The relative homogeneity of the population of Sahul rather suggests settlement of this vast and diverse region by Homo sapiens through a single episode or several closely linked passages, presumably by related groups.
    Relatively soon after—though it is difficult to be specific about what “soon” means in the context—people covered vast distances to occupy the northern and eastern extremities of New Guinea and southeastern Australia. The distance from an assumed landfall south of Timor to an archaeological site such as Keilor, close to Melbourne’s main airport, is just under 2,000 miles. The peoples of mainland Australia became separated from those of New Guinea and, again, biological anthropology suggests that subsequent interaction among the groups did not take place. The history was thus one of relocation—that is, travel that led to the establishment of a mode of life that would be adapted to new environments, some of them utterly different from the coastal or small island setting. This new mode of living was then localized. It was not apparently part of a sustained or regionally extensive social network, nor were ancestral connections with other places maintained through exchange or any form of intermarriage. From one perspective, this is exactly what one would expect: while the archaeological record is not rich, it implies comparatively simple modes of localized hunting and gathering. But in fact the populations that moved through the diverse and challenging environments of New Guinea and Australia developed locally specific technologies and new ways of living. The large blades found in the raised sediments on the Huon Peninsula have been interpreted as forest-clearing tools, which would have opened up areas in which edible plants such as sago, bananas, and yam vines might have spread. These people were not horticulturalists in the strict sense, but they appear to have been intervening in the environment to enable edible plants to flourish and spread to a greater extent than they would have if left unattended. If these really were experiments toward cultivation, they took place at a remarkably early date, far earlier than conventional histories of agriculture have acknowledged.

A 44,000-year-old, cave painting on the island of Sulawesi, Indonesia, possibly a hunting scene depicting an anoa, or miniature buffalo, facing several human–animal figures


    Relative to the peoples who arrived subsequently, these ancestors of Papuans and indigenous Australians appear to have been less specialized and, inevitably, less dynamic. While it is difficult to paint any finely detailed picture, given the relatively sparse archaeological record for these early periods, there is abundant evidence for distinctive creativity, innovation, and adaptation. Recent discoveries from the Leang Bulu’ Sipong cave site in southern Sulawesi have revealed what appears to be the earliest narrative scene in cave art from any part of the world. In particular, a fifteen-foot panel constitutes a scene in which human-animal hybrids—with human bodies but indeterminate, animal heads—are engaged in hunting pigs and a bovid, such as a wild ox of some kind. No specific belief, nor a spiritual practice such as shamanism, can be identified from this scene, other than speculatively. However, it has generally been assumed that similar representations from later periods were not just images of hunting, but were part of some magical or ritual effort to enhance the success of hunters, who may have danced or chanted in the presence of the painting. The Leang Bulu’  Sipong cave art is at least 43,900 years old. It provides a tantalizing glimpse of the beliefs of the people of the place at that time, and it is presumably an indication of the importance of hunting in their culture.
    Yet, at least some sapiens groups in the Wallacea-Papua region in this early period were oriented as much to sea as to land. People preferred to live along coastlines, and the earliest foragers no doubt collected and consumed shellfish, crabs, and other species that were readily accessible on reefs and in shallow waters. However, evidence for fishing is sparse from any part of the world before around 12,000 years ago—no doubt in part because older coastal sites were eroded or submerged as Pleistocene sea levels rose. It does not help that fishing materials such as spears, nets, and lines are generally perishable. Some early Sahul sites that are close to ancient shores and which have been preserved, in some cases because of geological uplift, have yielded shell, fish, and shark remains. These do not, however, indicate the presence of a maritime population, as the fish species represented are associated with inshore habitats, and smaller sharks were found on rocky reefs and could have been speared by people lacking even basic fishing-line technology. However, more abundant fish remains from lower levels at the Kilu rock-shelter on the island of Buka, at the northern extremity of the Solomon Islands, include some deepwater species (“pelagic,” in the scientific literature) such as tuna, which could only be caught offshore.

A complete fish hook from the Pleistocene levels of a cave site at the east end of Timor

Jerimalai Cave, in a raised terrace at the eastern extremity of Timor, is especially significant for a substantial assemblage of fish bones, including a high proportion—in the earliest layers of the site—from tuna and a variety of other deepwater species. The remains in question date back to 42,000–38,000 years ago. The finds imply, first, a local population skilled in offshore fishing techniques, which require strong lines and hooks, as well as an understanding of suitable bait or lures. Second, they suggest that the people routinely constructed, over a long period, vessels appropriate to fishing offshore. These were probably dugout canoes that might not have been large (line fishing can be undertaken by just one or two individuals) but would have been well balanced and seaworthy, given the challenges of bringing in any larger tuna or similar fish. Third, the finds indicate a broader pattern of social life and subsistence that complements this expertise. As the people were not horticulturalists, any use of stable canoes, as opposed to more improvised temporary watercraft, might imply a semi-settled residence pattern rather than the nomadism usually associated with hunter-gatherer lifestyles. People would surely not have invested the time and energy involved in boatbuilding only to leave vessels behind, unless they were maritime nomads who took their canoes wherever they went.
    The Papuan settlement of the Bismarck Islands and the Solomons can hardly be imagined without adaptation to life on water and a developed capacity to produce seaworthy vessels. Indeed, the necessary boats would have been bigger than those that could have taken a couple of men fishing (fishing was, historically, a male rather than a female or mixed activity across Oceania). The passage from the New Guinea mainland across to New Britain could be undertaken while maintaining two-way visibility—that is, both land behind and land ahead were visible. The crossing from New Ireland to Buka (achieved by 32,000 years ago) was either 87 or 109 miles, and the destination would not become visible until the boat was 25 to 34 miles away from the departure point, depending on the specific route. But the settlement of Manus, at the northeastern extremity of the Bismarck Archipelago, was a challenge of an entirely different order. The shortest sea crossing, successfully made by 25,000 years ago, was some 140 miles. For around a third of the voyage, both the land behind and the land ahead would have been out of view. So this was a speculative venture, toward islands that might not have existed, undertaken by people presumably confident in their capacity to explore and return home in the event that they encountered nothing but open water. No comparable or longer open-ocean voyage has been evidenced archaeologically up to this time in history. The seas to the north and east of New Guinea appear to have been realms of human experimentation of an extraordinary and unprecedented kind.
    Notable evidence of innovation in subsistent societies was found in the New Guinea Highlands. Investigations in the Wahgi Valley from the 1960s revealed drains and related features in swamps dating back to 9,000 years ago. These appear to reflect some basic form of horticulture, which was overlain by extensive gray clay, associated with widespread clearing. Even the early levels provide evidence for the cultivation of bananas, yam, and taro, plants that have loomed large ever since in the lives of Pacific com-munities. Pollen samples suggest a relatively sudden loss of forest in the surrounding area, which is strongly indicative of human clearing, presumably for gardening. The nature of early horticulture across Papua and in adjacent islands is yet to be more fully understood, but there is no doubt that diverse forms of tree cropping and the cultivation of root crops were practiced across the region, even if the people did not always constitute fully settled agricultural communities.
    The subsequent few thousand years would transform Papuan cultures in the interior of New Guinea and the neighboring islands. Change would be brought about by the dynamism of these communities and—in island Melanesia—by interaction with new arrivals. Archaeological discoveries have brought into view aspects of human history across Sahul during the millennia when, we now know, the region was occupied. But the signs of human presence regarding ways of life and belief are like candles in a living, fertile, and vibrant forest at night. We are lucky to glimpse scenes and moments but have little sense of the social landscapes and stories around them. Notwithstanding the primeval associations of tropical rain forests, we do know that there was nothing static or unvarying in the lives of the Papuan ancestors. To the contrary, the region was distinguished not only by a plethora of local adaptations and the invention of new ways of life on water, but also, it appears, by humanity’s first experiments in the daunting business of maritime exploration. —NT

Recent Stories

The way they live, the food they eat, and the effect on us

A true but unlikely tale

Story and Photographs by William Rowan

Increasing day length on the early Earth boosted oxygen released by photosynthetic cyanobacteria.

Genomic evidence shows that Denisovans and modern humans may have overlapped in Wallacea.