What Were the Early Seafaring Cultures, and How Did They Do It?/info


Introduction

The colonisation of islands and continents separated from mainland landmasses by open water is among the most significant behavioural markers in human evolution. Maritime crossings require not only watercraft capable of surviving ocean conditions, but also navigational knowledge, provisioning for multi-day voyages, and — in many cases — the social organisation necessary to sustain a founding population on arrival. The archaeological record demonstrates that water crossings were accomplished far earlier and more frequently than twentieth-century scholarship assumed, with evidence now extending into the Lower Palaeolithic for at least short-distance strait crossings and deep into the Upper Palaeolithic for open-ocean voyages.

The colonisation of Sahul (the Pleistocene landmass comprising Australia and New Guinea) by approximately 65,000 years BP required crossing at least 90 kilometres of open water through the Wallacean archipelago — a feat that presupposes functional watercraft and deliberate navigation. The settlement of the Ryukyu Islands, the Mediterranean islands, and ultimately Remote Oceania by Austronesian-speaking populations demonstrates an escalating trajectory of maritime capability spanning tens of thousands of years. Most remarkably, the presence of Homo erectus and Homo floresiensis on islands east of Wallace's Line raises the possibility that non-sapiens hominins accomplished water crossings over one million years ago, though whether these crossings were intentional or accidental remains vigorously debated.

This brief reviews the evidence for early seafaring across multiple regions and time periods, evaluates competing models for the mechanisms of maritime dispersal, and identifies the principal unresolved questions in the field.

What we know

The earliest indirect evidence for water crossings comes from the Indonesian archipelago, where stone tools dating to over one million years BP have been recovered from islands that were never connected to the Asian mainland by land bridges, even during glacial maxima when sea levels were at their lowest. Sulawesi, which lies east of Wallace's Line, has yielded stone artefacts dated to approximately 194,000 years BP (van den Bergh et al. 2016), and more recently to approximately 1.04 million years BP (Brumm et al. 2025). Flores, the island on which Homo floresiensis was discovered, has stone tool evidence extending back approximately 1 million years. These finds imply that Homo erectus or a related species crossed at least one significant water gap, though the minimum crossing distance during glacial low-stands may have been as short as 10–20 kilometres at certain straits.

The debate over whether these early crossings were intentional or accidental centres on the question of watercraft. Some researchers, notably Bednarik (1999, 2003), have argued that Homo erectus possessed bamboo rafts and the navigational awareness to use them. Others, including Shipton et al. (2021), propose that early hominin colonisation of western Wallacea was incidental — individuals swept out to sea while swimming, clinging to natural flotsam, or caught in storm surges. The absence of any direct evidence for pre-sapiens watercraft leaves this question unresolved.

The colonisation of Sahul represents the earliest unambiguous evidence for planned, deliberate maritime migration. Clarkson et al. (2017) dated the Madjedbebe rock shelter in northern Australia to approximately 65,000 years BP (with some debate; Allen and O'Connell 2020 argue for a more conservative estimate of approximately 50,000 years BP). Regardless of the precise date, reaching Sahul required multiple water crossings through the Wallacean island chain, with the longest single voyage spanning approximately 90 kilometres of open sea. The founding population must have included enough individuals — likely including both sexes — to establish a viable breeding community, implying organised group migration rather than accidental drift.

Pelagic fishing at Jerimalai Cave in East Timor, dated to approximately 42,000 years BP, provides direct evidence of advanced maritime subsistence. O'Connor et al. (2011) reported over 38,000 fish bones from 23 taxa, including tuna and other deep-water species that cannot be caught from shore. The assemblage also included the world's earliest known fishhooks. This evidence demonstrates that by the Late Pleistocene, populations in Island Southeast Asia had developed specialised maritime technologies — boats capable of reaching deep water, fishing tackle for pelagic species, and the knowledge to locate and exploit oceanic fish stocks.

In the Mediterranean, evidence for pre-Neolithic island colonisation has accumulated steadily since the early 2000s. Stone tools attributed to Lower Palaeolithic traditions have been reported from Crete (Strasser et al. 2010), though these claims remain contested. More securely dated evidence places Homo sapiens on Sardinia by approximately 20,000 years BP, on Sicily by approximately 30,000 years BP, and on Cyprus by approximately 12,000 years BP — all during periods when these islands were separated from the mainland by significant water gaps.

The experimental archaeology programme led by Yosuke Kaifu of the National Museum of Nature and Science, Tokyo, has contributed substantially to understanding Palaeolithic seafaring in East Asia. Kaifu et al. (2019) tested the bamboo raft hypothesis by constructing and sailing replicas, concluding that bamboo rafts were inadequate for the 200-kilometre crossing from Taiwan to the Ryukyu Islands against the Kuroshio Current. A subsequent experimental crossing in 2019 using a dugout canoe succeeded, suggesting that more advanced watercraft were required for this route.

The Lapita cultural complex, emerging approximately 3,500 years BP in the Bismarck Archipelago, represents the most archaeologically visible phase of Pacific seafaring. Lapita peoples — ancestors of the Polynesians — colonised a vast oceanic territory from Island Melanesia to Fiji and Tonga within a few centuries, using outrigger canoes and sophisticated non-instrument navigation based on star paths, ocean swell patterns, and biological indicators (bird flight, cloud formations over land). By approximately 1200 CE, Polynesian voyagers had reached every habitable island in the Pacific, including Hawai'i, Rapa Nui (Easter Island), and Aotearoa (New Zealand).

The navigational systems of Polynesian wayfinders, documented ethnographically and preserved in oral traditions, represent the most complex pre-modern maritime knowledge systems known. Navigators memorised star compasses, swell signatures for individual islands, and seasonal current patterns — a body of knowledge requiring years of apprenticeship and demonstrating that maritime expertise constituted a formal intellectual tradition transmitted across generations.

Classification of theories

A. Plausible explanations (supported by evidence, not refuted):

  1. Homo sapiens accomplished deliberate, planned maritime crossings to colonise Sahul by at least 50,000–65,000 years BP, using watercraft of unknown type through the Wallacean island chain.
  2. Pelagic fishing technologies, including offshore-capable boats and tackle, were established in Island Southeast Asia by approximately 42,000 years BP.
  3. Austronesian-speaking populations developed outrigger canoe technology and non-instrument navigation systems that enabled the colonisation of Remote Oceania beginning approximately 3,500 years BP.
  4. Pre-Neolithic Homo sapiens reached Mediterranean islands separated from the mainland during the Upper Palaeolithic and Mesolithic.

B. Possible explanations (raised by researchers, less well supported):

  1. Homo erectus or related species intentionally constructed and used watercraft (bamboo rafts or similar) to colonise Wallacean islands over one million years ago (Bednarik 2003). Direct evidence for pre-sapiens watercraft is absent.
  2. Deliberate seafaring, rather than accidental drift, explains all Lower Palaeolithic water crossings in Southeast Asia (Barham and Everett 2021).
  3. Lower Palaeolithic stone tools on Crete indicate Homo erectus maritime capability in the Mediterranean (Strasser et al. 2010). Dating and attribution remain contested.

C. Highly unlikely but argued by some:

  1. All pre-Neolithic island colonisations can be explained by accidental drift on natural flotsam without any watercraft. The demographic requirements for founding viable populations make purely accidental colonisation implausible for most documented cases.
  2. Polynesian navigation was unsystematic and relied primarily on luck rather than formalised knowledge systems. Ethnographic evidence and experimental voyaging have thoroughly refuted this position.

---

Research Papers

Landmark Studies

1. Clarkson, C., Jacobs, Z., Marwick, B., Fullagar, R., Wallis, L., Smith, M., Roberts, R.G., Hayes, E., Lowe, K., Carah, X., Florin, S.A., McNeil, J., Cox, D., Arnold, L.J., Hua, Q., Huntley, J., Brand, H.E.A., Manne, T., Fairbairn, A., Shulmeister, J., Lyle, L., Salinas, M., Page, M., Connell, K., Park, G., Norman, K., Murphy, T., and Pardoe, C. (2017). "Human occupation of northern Australia by 65,000 years ago." Nature, 547(7663), 306–310. Full text: https://www.nature.com/articles/nature22968 [publisher paywall] [Publisher abstract] The time of arrival of people in Australia is an unresolved question. We report the results of new excavations at Madjedbebe, a rock shelter in northern Australia. Optically stimulated luminescence dating places human occupation at 65,000 years ago. Full-text notes: Clarkson et al. applied single-grain optically stimulated luminescence (OSL) dating to sediments at Madjedbebe, producing a robust chronological framework that pushed back the earliest occupation of Australia by approximately 15,000 years relative to previous estimates. The assemblage includes ground-edge axes, ochre processing, and plant food processing — a full behavioural repertoire from the earliest occupation layers. The maritime implications are profound: reaching northern Australia from the nearest Wallacean island required crossing at least 90 km of open water, likely in a single voyage. The study has been challenged by Allen and O'Connell (2020), who argue that sediment disturbance may have displaced artefacts to older levels, but the OSL methodology has been defended in subsequent publications. The finding remains one of the strongest pieces of evidence for early deliberate seafaring.

2. O'Connor, S., Ono, R., and Clarkson, C. (2011). "Pelagic Fishing at 42,000 Years Before the Present and the Maritime Skills of Modern Humans." Science, 334(6059), 1117–1121. Full text: https://www.science.org/doi/10.1126/science.1207703 [publisher paywall] [Publisher abstract] Maritime migration by anatomically modern humans into island Southeast Asia represents one of the great achievements of our species. We report remains of a variety of pelagic and other fish species dating to 42,000 years before the present from Jerimalai shelter in East Timor, as well as the earliest definite evidence for fishhook manufacture in the world. Full-text notes: The Jerimalai assemblage is transformative because pelagic fish — particularly tuna — cannot be caught from shore; their capture requires boats, deepwater tackle, and knowledge of fish behaviour and seasonal migration patterns. The presence of tuna bones from the earliest occupation levels demonstrates that maritime subsistence was not a late addition to the behavioural repertoire of Sahul colonists but was present from the outset. The fishhooks, made from shell, represent the oldest known examples of this technology class. O'Connor et al. argued that the sophistication of maritime subsistence at Jerimalai implies a long prior history of maritime adaptation, likely developed during the colonisation of Wallacea itself.

3. Kaifu, Y., Fujita, M., Yoneda, M., and Yamasaki, S. (2019). "Palaeolithic seafaring in East Asia: testing the bamboo raft hypothesis." Antiquity, 93(372), 1424–1441. Full text: https://doi.org/10.15184/aqy.2019.90 [publisher paywall] [Publisher abstract] The colonisation of the Ryukyu Islands in southwestern Japan during the Late Pleistocene required multiple sea crossings. The authors test the hypothesis that bamboo rafts were used for these crossings through construction and experimental navigation. Full-text notes: Kaifu et al. built bamboo rafts using only materials and techniques plausible for the Late Pleistocene and attempted sea crossings in the waters between Taiwan and the Ryukyu Islands. The rafts proved inadequate against the Kuroshio Current — the strong western boundary current that dominates this region — suggesting that dugout canoes or similarly advanced watercraft were required. This negative experimental result is methodologically significant because it narrows the range of plausible watercraft types and implies a higher level of technological sophistication than simple raft construction. The team's subsequent successful crossing by dugout canoe in 2019 demonstrated that the route was viable with more advanced technology.

4. Fujita, M., Yamasaki, S., Katagiri, C., Oshiro, I., Sano, K., Kurozumi, T., Sugawara, H., Kurita, D., Yamamoto, Y., Takeshita, Y., Kimura, H., Kaifu, Y., et al. (2016). "Advanced maritime adaptation in the western Pacific coastal region extends back to 35,000–30,000 years before present." Proceedings of the National Academy of Sciences, 113(40), 11184–11189. Full text: https://www.pnas.org/doi/10.1073/pnas.1607857113 [open access] [Publisher abstract] Archaeological evidence from the Sakitari Cave on Okinawa Island demonstrates advanced maritime adaptation, including shell fishhook manufacture and pelagic resource exploitation, dating to 35,000–30,000 years before present. Full-text notes: The Sakitari Cave assemblage extends the evidence for advanced maritime subsistence in the Ryukyu Islands back to at least 35,000 years BP. Shell fishhooks, seasonal shellfish harvesting, and consumption of freshwater crabs and eels demonstrate a diversified aquatic subsistence base. The paper's significance lies in its demonstration that maritime adaptation in the western Pacific was not a late or marginal phenomenon but a core subsistence strategy from the earliest colonisation of the island chain. Combined with the Jerimalai evidence, the Sakitari data support a model in which maritime expertise was a prerequisite for, rather than a consequence of, island colonisation.

5. Leppard, T.P. (2015). "Passive dispersal versus strategic dispersal in island colonization by hominins." Current Anthropology, 56(4), 590–595. Full text: https://www.journals.uchicago.edu/doi/10.1086/682325 [publisher paywall] [Agent-generated summary] Leppard examines the theoretical distinction between passive (accidental) and strategic (intentional) dispersal models for hominin island colonisation, arguing that the demographic requirements for founding viable populations on islands — particularly the need for propagule groups containing both sexes and sufficient genetic diversity — make passive dispersal an inadequate explanation for most documented cases. The paper proposes a framework for distinguishing between the two mechanisms based on island remoteness, founding population size estimates, and the speed of post-colonisation population growth. Full-text notes: Leppard's framework is useful because it transforms the intentional-versus-accidental debate from a binary question into a graduated assessment based on specific variables. For short-distance crossings (under 10 km) with visible target islands, accidental arrival of small groups is demographically plausible. For crossings exceeding 50 km to islands beyond the horizon, the probability of multiple accidental arrivals contributing to a viable founding population drops sharply. This analysis strongly supports intentional seafaring for the colonisation of Sahul, the Ryukyu Islands, and the more remote Mediterranean islands.

6. Kealy, S., Louys, J., and O'Connor, S. (2016). "Islands Under the Sea: A Review of Early Modern Human Dispersal Routes and Migration Hypotheses Through Wallacea." The Journal of Island and Coastal Archaeology, 11(3), 364–384. Full text: https://doi.org/10.1080/15564894.2015.1119218 [publisher paywall] [Agent-generated summary] Kealy et al. reviewed the evidence for Homo sapiens dispersal through Wallacea en route to Sahul, evaluating multiple proposed sea routes through the archipelago using bathymetric reconstructions of Pleistocene shorelines. The study identified a northern route through Sulawesi as the most likely path based on minimum water crossing distances during glacial low-stands. Full-text notes: This comprehensive review synthesised bathymetric, archaeological, and genetic data to evaluate which of several proposed routes through Wallacea was most feasible for Pleistocene seafarers. The northern route via Sulawesi offered shorter individual water crossings and more inter-visible islands than southern routes through Flores and Timor. However, the authors noted that different routes may have been used at different times, and that the existence of pelagic fishing at Jerimalai suggests that at least some colonising populations possessed maritime capabilities sufficient for the longer southern crossings.

Recent Studies

7. Shipton, C., et al. (2021). "Pleistocene Water Crossings and Adaptive Flexibility Within the Homo Genus." Journal of Archaeological Research, 29, 235–326. Full text: https://link.springer.com/article/10.1007/s10814-020-09149-7 [open access] [Publisher abstract] This review examines evidence for Pleistocene water crossings by multiple Homo species, including H. erectus, H. floresiensis, H. luzonensis, and H. sapiens. The authors propose that colonisation of western Wallacea by pre-sapiens hominins was incidental rather than intentional, while H. sapiens colonisation of eastern Wallacea and Sahul required deliberate seafaring. Full-text notes: Shipton et al. drew a distinction between western Wallacea (where Homo erectus and other archaic species are found) and eastern Wallacea/Sahul (reached only by Homo sapiens). The argument is that short water crossings in western Wallacea could be accomplished through swimming, clinging to flotsam, or storm-driven drift — consistent with the colonisation patterns of other large mammals including elephants and stegodons. Eastern Wallacea and Sahul, however, required crossings of 50–90 km beyond the horizon, which the authors consider achievable only through intentional seafaring with constructed watercraft.

8. Kaifu, Y. (2022). "A synthetic model of Palaeolithic seafaring in the Ryukyu Islands, southwestern Japan." World Archaeology, 54(4), 573–595. Full text: https://www.tandfonline.com/doi/full/10.1080/00438243.2022.2121317 [publisher paywall] [Agent-generated summary] Kaifu synthesises a decade of archaeological, experimental, and oceanographic research on Palaeolithic seafaring in the Ryukyu Islands, proposing a model in which colonisers used dugout canoes and possessed sufficient navigational knowledge to plan crossings of 100+ km against the Kuroshio Current. Full-text notes: This synthetic paper integrates Kaifu's experimental voyage data with archaeological evidence from multiple Ryukyu Island sites and oceanographic modelling of Late Pleistocene currents. The model proposes that colonisers departed from the coast of Taiwan (then connected to the Asian mainland) and navigated to islands that were, in several cases, not visible from the point of departure. Kaifu argues that the colonisation was planned, deliberate, and required both advanced watercraft and substantial navigational knowledge — making the Ryukyu colonisation one of the most cognitively demanding maritime events in the Palaeolithic record.

9. Brumm, A., et al. (2025). "Stone tools on Sulawesi date to over 1 million years ago." Nature (in press/2025). Full text: https://www.nature.com/articles/s41586-025-09348-6 [open access] [Agent-generated summary] Stone artefacts recovered from Sulawesi, Indonesia, have been dated to approximately 1.04 million years BP, extending the evidence for hominin presence on this Wallacean island deep into the Early Pleistocene. The toolmakers were likely Homo erectus, implying water crossings at this early date. Full-text notes: The Sulawesi findings push back the evidence for hominin island colonisation by hundreds of thousands of years and raise fundamental questions about the maritime capabilities of pre-sapiens species. While the minimum crossing distance to Sulawesi during glacial maxima was relatively short, any crossing required surviving open water. Whether this was accomplished through intentional use of watercraft or accidental drift remains unresolved, but the presence of stone tools implies that enough individuals survived the crossing to establish a population capable of lithic production.

---

Current Discussions

  1. "Ancient humans were seafaring far earlier than we realised." New Scientist, January 2026. https://www.newscientist.com/article/2511681-ancient-humans-were-seafaring-far-earlier-than-we-realised/ [publisher paywall] — Review article surveying the accumulated evidence for pre-Neolithic maritime capability, including recent finds from Wallacea and the Mediterranean.
  1. "They Crossed Oceans a Million Years Ago and Left Clues in Stone." Science News Today, August 2025. https://www.sciencenewstoday.org/they-crossed-oceans-a-million-years-ago-and-left-clues-in-stone [open access] — Coverage of the Brumm et al. 2025 Sulawesi stone tool discovery and its implications for Homo erectus water crossing capabilities.
  1. "An ancient human species was island hopping in Asia over a million years ago." Natural History Museum, London, August 2025. https://www.nhm.ac.uk/discover/news/2025/august/ancient-human-species-island-hopping-asia-over-a-million-years-ago.html [open access] — Institutional summary discussing multiple hominin species in Wallacea and the ocean current modelling that informs dispersal hypotheses.
  1. "A Stone Age ocean voyage re-enacted by daring scientists." EarthSky, July 2025. https://earthsky.org/human-world/stone-age-ocean-voyage-re-enacted/ [open access] — Report on the ongoing analysis of Kaifu's experimental voyage programme, including discussion of the limitations of single-voyage experiments.

---

Future Research Directions

The most significant gap in the evidence base is the complete absence of preserved watercraft from any period before the Mesolithic. The oldest known boats — the Pesse canoe (Netherlands, approximately 8,000 BCE) and dugouts from several Mesolithic European sites — postdate the earliest maritime crossings by tens of thousands of years. Targeted investigation of waterlogged contexts in coastal Southeast Asia, northern Australia, and the Mediterranean could potentially recover earlier specimens, though the probability of wood preservation in tropical marine environments is low.

Oceanographic modelling of Pleistocene current systems offers a productive approach to constraining dispersal hypotheses without relying on direct vessel evidence. By reconstructing glacial-period ocean currents and simulating drift trajectories, researchers can estimate the probability of accidental arrival on specific islands and identify which crossings required active navigation against prevailing currents. Such modelling has been applied to the Ryukyu Islands and parts of Wallacea but remains underdeveloped for the Mediterranean and East African coast.

Ancient DNA from founding populations on colonised islands could resolve questions about propagule size, sex ratios, and source populations. The recovery of aDNA from early Sahul settlers would be particularly informative, as it could reveal whether the founding group was small (consistent with accidental dispersal) or genetically diverse (consistent with planned migration). Current aDNA recovery from tropical and subtropical contexts remains technically challenging, but advances in sedimentary eDNA extraction are expanding the geographic range of viable sampling.

The relationship between seafaring capability and cognitive/linguistic evolution remains largely theoretical. Barham and Everett (2021) have argued that the Homo erectus water crossings in Wallacea imply language capacity, on the grounds that coordinated group migration requires symbolic communication. This hypothesis is currently untestable with archaeological methods alone, but comparative analyses of the cognitive demands of maritime versus terrestrial dispersal — drawing on experimental navigation, ethnographic analogy, and computational modelling — could refine the theoretical framework.

Experimental archaeology programmes, following the model established by Kaifu in Japan, should be extended to other regions where Palaeolithic maritime crossings are documented. The Mediterranean, the East African coast, and the Wallacean straits all present distinctive oceanographic challenges that would benefit from systematic experimental testing of watercraft types, navigation methods, and provisioning requirements.

---

Summary of Existing Research and Public Opinion

The scholarly consensus has shifted substantially over the past two decades from a model in which deliberate seafaring was a relatively recent human innovation (Neolithic or later) to one in which maritime capability extends deep into the Palaeolithic and possibly into the Lower Palaeolithic for short-distance crossings. The colonisation of Sahul is now universally accepted as requiring deliberate maritime migration, and the evidence for pelagic fishing at 42,000 years BP and shell fishhook manufacture at 35,000 years BP demonstrates that maritime subsistence was a core adaptation in Island Southeast Asia from the earliest occupation.

The most contentious area of current debate concerns the Lower Palaeolithic water crossings in Wallacea. The discovery of million-year-old stone tools on Sulawesi has intensified the question of whether Homo erectus possessed watercraft and navigational intent, or whether colonisation occurred through natural dispersal mechanisms. The field is divided between researchers who see these crossings as evidence for pre-sapiens cognitive sophistication (Bednarik, Barham and Everett) and those who model them as passive processes analogous to the dispersal of other large mammals (Shipton et al.).

Public awareness of early seafaring has grown significantly through high-profile media coverage of Kaifu's experimental voyages and the Sulawesi stone tool discoveries. The Polynesian navigational tradition, in particular, has achieved broad cultural recognition through documentary films, museum exhibitions, and the revival of traditional wayfinding practices by Pacific Islander communities. The experimental voyage of the Hōkūleʻa and subsequent Pacific voyaging canoes has been instrumental in demonstrating the viability of non-instrument navigation to both scientific and general audiences.

---

Where Do I Come In?

For maritime archaeologists and geoscientists, the priority is expanding the geographic and chronological scope of seafaring evidence through targeted fieldwork in understudied regions. The East African coast, the Arabian Peninsula, and the western Indian Ocean represent major gaps in the current evidence base despite being critical zones for the initial Out-of-Africa dispersal. Submerged landscape archaeology — surveying continental shelves exposed during glacial maxima — offers the highest potential for transformative discovery.

For oceanographers and climate modellers, collaboration with archaeologists on Pleistocene current reconstruction and drift simulation provides a quantitative framework for evaluating dispersal hypotheses that has been applied to only a handful of case studies so far. Extending these models to cover the full range of documented water crossings would significantly constrain the space of plausible explanations.

For geneticists, the recovery and analysis of ancient DNA from founding island populations remains a high-value target. Advances in sedimentary eDNA extraction and low-coverage whole-genome sequencing from degraded tropical samples are beginning to make this feasible, though substantial methodological challenges remain.

For cognitive scientists and linguists, early seafaring provides a natural experiment in the relationship between technological complexity and cognitive capacity. The debate over whether Homo erectus water crossings imply language is fundamentally a question about the minimum cognitive prerequisites for specific behavioural outcomes — a question that connects maritime archaeology to broader programmes in cognitive evolution.

For the public and for educators, the history of human seafaring offers a powerful counter-narrative to technological determinism. The Polynesian colonisation of the Pacific — accomplished without metal, writing, or any form of instrument — demonstrates that the most remarkable navigational achievements in human history were products of accumulated oral knowledge, social organisation, and environmental observation rather than material technology.