Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2025 May;641(8061):137-143.
doi: 10.1038/s41586-025-08780-y. Epub 2025 Apr 9.

Hunter-gatherer sea voyages extended to remotest Mediterranean islands

Eleanor M L Scerri  1   2   3 , James Blinkhorn  4   5 , Huw S Groucutt  6   7 , Mathew Stewart  8 , Ian Candy  9 , Ethel Allué  10   11 , Aitor Burguet-Coca  10   11   12 , Andrés Currás  13 , W Christopher Carleton  14 , Susanne Lindauer  15 , Robert Spengler  16 , Kseniia Boxleitner  16 , Gillian Asciak  17 , Margherita Colucci  13   18 , Ritienne Gauci  19 , Amy Hatton  14   20   21 , Johanna Kutowsky  13 , Andreas Maier  7 , Mario Mata-González  13   6 , Nicolette Mifsud  6 , Khady Niang  13   22 , Patrick Roberts  14   23 , Joshua de Giorgio  24 , Rochelle Xerri  25 , Nicholas C Vella  26
Affiliations

Hunter-gatherer sea voyages extended to remotest Mediterranean islands

Eleanor M L Scerri et al. Nature. 2025 May.

Erratum in

  • Author Correction: Hunter-gatherer sea voyages extended to remotest Mediterranean islands.
    Scerri EML, Blinkhorn J, Groucutt HS, Stewart M, Candy I, Allué E, Burguet-Coca A, Currás A, Carleton WC, Lindauer S, Spengler R, Boxleitner K, Asciak G, Colucci M, Gauci R, Hatton A, Kutowsky J, Maier A, Mata-González M, Mifsud N, Niang K, Roberts P, de Giorgio J, Xerri R, Vella NC. Scerri EML, et al. Nature. 2026 Feb;650(8102):E10. doi: 10.1038/s41586-025-10024-y. Nature. 2026. PMID: 41611888 Free PMC article. No abstract available.

Abstract

The Maltese archipelago is a small island chain that is among the most remote in the Mediterranean. Humans were not thought to have reached and inhabited such small and isolated islands until the regional shift to Neolithic lifeways, around 7.5 thousand years ago (ka)1. In the standard view, the limited resources and ecological vulnerabilities of small islands, coupled with the technological challenges of long-distance seafaring, meant that hunter-gatherers were either unable or unwilling to make these journeys2-4. Here we describe chronological, archaeological, faunal and botanical data that support the presence of Holocene hunter-gatherers on the Maltese islands. At this time, Malta's geographical configuration and sea levels approximated those of the present day, necessitating seafaring distances of around 100 km from Sicily, the closest landmass. Occupations began at around 8.5 ka and are likely to have lasted until around 7.5 ka. These hunter-gatherers exploited land animals, but were also able to take advantage of marine resources and avifauna, helping to sustain these groups on a small island. Our discoveries document the longest yet-known hunter-gatherer sea crossings in the Mediterranean, raising the possibility of unknown, precocious connections across the wider region.

PubMed Disclaimer

Conflict of interest statement

Competing interests: The authors declare no competing interests.

Figures

Fig. 1
Fig. 1. Maps and image of Latnija.
Top, the position of Malta in the Mediterranean. Bottom left, digital elevation model of Latnija, showing the current dripline in dashed lines. Bottom right, the site, showing the sea channel and Gozo in the background, with past sea levels based on a previous study. LGM, Last Glacial Maximum; MASL, metres above sea level. The edge of Trench 4 is denoted by the hessian sacks. Data from refs. , and created using ArcMap 10.5.
Fig. 2
Fig. 2. Stratigraphic section of the northwest wall.
Top, illustration of the key stratigraphic sequence (numbered Beds are described in Supplementary Information 2) highlighting a thick bed of ash (A; bottom left), and a hearth deposit or combustion structure (B; bottom right), with combustion residue (ash on top), thermal impact zone and a natural substrate (Supplementary Information 3), at the base of the Mesolithic Horizon. Note also the Phorcus turbinatus tip line, starting in the mid-right of box A.
Fig. 3
Fig. 3. Chronological model.
Model (OxCal 4.4; IntCal20) shows the phase boundaries in the Mesolithic Horizon, Phases III–V. The model indicates that the mean start date of the Mesolithic Horizon is 8.5 ka. Laboratory codes are included in the left box.
Fig. 4
Fig. 4. Fauna and lithics from the Mesolithic Horizon.
a–q, Selected fauna and lithics from the Mesolithic Horizon. All the fauna are wild. Limestone flakes (a–d,f,g), red deer left mandible (e), and metatarsal (q), Phorcus turbinatus (h), Patella sp. (i), crab claw (k), turtle or tortoise carapace (l,p), fish vertebra (m), seal proximal phalanx (n), bird humerus (j) and coracoid (o). Scale bar is 50 mm and applies to all. r, Percentage of the number of reported specimens (NRSP) of piece-plotted bone. This includes terrestrial animals and marine mammals. s, Percentage of the number of identified specimens (NISP) of fish and marine invertebrates from squares L2 and N2 recovered during wet sieving and flotation.
Extended Data Fig. 1
Extended Data Fig. 1. Regional OxCal phase model of the Mesolithic-to-Neolithic transition.
OxCal phase modelling of the estimated start and end dates for the Mesolithic and Neolithic phases in and around Malta125 using the IntCal20 terrestrial calibration curve. Results indicate a general geographical cline in the spread of the Neolithic in mainland Italy from north to south, to Sicily, Sardinia, and Corsica, and then finally to Malta. See the OxCal script in https://github.com/wccarleton/mesoneomalta for specifics. As the legend indicates, dashed lines represent the ‘sigma’ type boundary whereas filled areas represent the ‘uniform’ type.
Extended Data Fig. 2
Extended Data Fig. 2. Bayesian evaluation of the earliest Maltese Neolithic.
a, Plotted radiocarbon dates available for the sequence from the Salina Deep record in Malta, along with the pollen sequence reported by Farrell and colleagues, using the Bchron R package, one of the methods used by Hunt and colleagues. Results show many potentially intrusive samples used to date the sediments and very few sequences of dates in strict stratigraphic order. b, Our Bayesian regression model to relate depth to age in the Salina Deep core using the IntCal20 calibration curve to calibrate the dates. The results suggest that the first Neolithic evidence in the Salina Deep record has a date with a broad error range of around two thousand years. Depth shown = cm.
Extended Data Fig. 3
Extended Data Fig. 3. Harris matrix of the Latnija excavation contexts organized by phase.
Harris matrix of the Latnija excavation organized by phase illustrating the stratigraphic relationship between excavation contexts, with numbers shown in red indicating deposits containg dated material.
Extended Data Fig. 4
Extended Data Fig. 4. Results of archaeobotanical analyses.
a, Pollen diagram from Latnija archaeological site, with percentage values of pollen remains identified from contexts (034) and (048). b, The deep Mesolithic hearth from Phase V, square N2, with sample locations for phytolith and FTIR studies. The internal structure of the hearth can be observed, from top to bottom, combustion residue, thermal impact and natural substrate (control). c, ESEM images of Pistacia cf. lentiscus and Juniperus sp. charcoal remains showing wood anatomical characters: (i) Juniperus sp. charcoal fragment tangential section; (ii) Juniperus sp. charcoal fragment tangential section showing rays and tracheids; (iii) Juniperus sp. charcoal fragment tangential section showing a detail of tracheid pits; (iv) Pistacia cf. lentiscus charcoal fragment transverse section showing ring porous distribution and vessel clusters; (v) Pistacia cf. lentiscus charcoal fragment tangential section showing spiral thickenings and biseriated rays; (vi) Pistacia sp. charcoal fragment transverse section showing cracks and vitrification altering the wood cell structure.
Extended Data Fig. 5
Extended Data Fig. 5. Chronological sample locations.
Oblique view illustrating the location of dated contexts (red) with respect to sediment phase boundaries spanning grid squares J–N, showing the distribution of sediments at the base of Phase II; the location of dated contexts from Phase III, with the upper boundary of Phase III deposits shown as a wireframe; the location of dated contexts from Phase IV, with the upper boundary of Phase IV deposits shown as a wireframe; the location of dated contexts from Phase V, with the upper boundary of Phase V deposits shown as a wireframe; and the location of dated contexts from Phase VI, with the upper boundary of Phase VI deposits shown as a wireframe.
Extended Data Fig. 6
Extended Data Fig. 6. Section drawing.
a, Section drawing of the Latnija exposure with the detailed section shown in Fig. 3 highlighted (recorded September 2022). b, Detailed record of the contact between the upper part of Unit 4 and the lower contact of Bed 10 (for location of the recorded section see Fig. 1). The section was recorded in September 2024, note that some of the large clasts that were recorded in the September 2022 section (a) had been removed from the section by September 2024.
Extended Data Fig. 7
Extended Data Fig. 7. Photomicrographs of the Latnija sediments.
a, Overview of the sediments of Unit 1. The sediments are dominated by limestone derived clasts and fine-grained material. The clear grains are fine-sand/silt sized grains of quartz. b, Biological material within Unit 1 (Sh – shell, Bo – Bone, BF – Burrow fill). c, Overview of the sediments of Unit 2 (Bed 10). The high birefringence colours show the dominance of limestone material in both the coarse and fine component. d, Limestone clasts within Bed 10 showing unweathered limestone fragments (UWLF) and blackened, burnt limestone fragments (BLF). e, Burnt limestone fragment (BLF) next to an iron oxide enriched intraclast (IC) of reworked sediment. f, Charcoal Fragment (ChF) and Burnt limestone fragment (BLF) in Bed 10. g, Iron enriched sediments of Unit 1 from directly below the contact with Bed 12. h, Pelleted microfabric of lens b below Bed 10.
Extended Data Fig. 8
Extended Data Fig. 8. High-resolution scans of micromorphological thin sections from the Latnija sequence.
See Extended Data Fig. 6 and sediment section (Supplementary Information 2) for the location of thin-section samples. a, MM2 is taken from the unaltered sediments at the base of the sequence. This sample presents the key characteristics of patterns of sedimentation prior to the Mesolithic occupation. These can be summarized as consisting of: 1) rare limestones clasts (frequently showing evidence for in situ decay), 2) intraclasts of reworked sediment and 3) fragments of terrestrial mollusc shell. Limited materials that are indicative of burning or burning products are present. b, MM3 is taken from the ash-rich sediments of the sequence, a factor that can be seen in the colour difference between the brown matrix of MM2 and the grey matrix of MM3. MM3 is rich in >1 mm sized charcoal fragments and limestone clasts, some of which show evidence for a strong degree of burning. c, MM5 is taken from the contact between the ash-rich sediment (MM3) and the underlying cave floor sediment (MM2). The sediment of MM5 is, consequently, characterized by a mixture of both sediment types. Charcoal fragments are abundant (but only rarely >1 mm) but the matrix overall is more typical of MM2. Bones fragments are present as are circular features that are characteristics of deformation which could be either biological or physical in origin. The sediments have formed in association with cutting into pre-existing sediments prior to the deposition of the ash-rich sediments. d, MM6 is taken from below darkened sediment believed to be in situ burning. The sediments of MM6 are more reddened than any other sampled sediments and occur directly below sediments the colour of which are more typical of the unaltered cave sediments of MM2. This unit is interpreted as being thermally altered as a direct result of in situ burning directly above these deposits.
Extended Data Fig. 9
Extended Data Fig. 9. Phytolith study.
a, Phytoliths image identified in the Phase V deep hearth. a,b, Elongate entire from grass leaf; c,d, Elongate dentate from grass inflorescence; e, Bulliform flabellate from grass leaf; f, Acute bulbosus from grasses leaf; g,h, GSSCP Bilobate from grasses-C4; i–j, GSSCP Rondel from grasses-C3; k,l, Spheroid echinate from Palms. b, Different dynamics of phytoliths identified in the layers that make up the hearth in relation to the number of phytoliths identified as spheroid echinate in each sample. Key: C, Control; TI, Thermal Impact; CR, Combustion Residue. c, Infrared spectra of sediments from some representative samples. Key: C, Control; TI, Thermal Impact; CR, Combustion Residue; Ca, Calcite; Cl, Clay; Qz, quartz; b, thermally altered clay; nb, not thermally altered clay.
Extended Data Fig. 10
Extended Data Fig. 10. Lithics from the Maltese Mesolithic.
a, Photographs of lithics from Phase III (left) and Phase V (right) showing terrestrial and coastal raw material forms. b–d, Illustrations of lithics from Phases V (b), IV (c) and III (d). All are flakes except b3 (core) and b4 (retouched flake). All limestone.
Extended Data Fig. 11
Extended Data Fig. 11. Taxonomic and isotopic analyses of faunal remains.
Red deer remains, taphonomic modifications, and stable isotope results. a, Remains of red deer including a proximal radius (1), distal radius (b), proximal metatarsal (3), proximal metacarpal (4), scapula (5), and a distal metatarsal (6). Examples of taphonomic modifications including a midshaft fragment with a green fracture and a double notch with corresponding negative flake scars (7), a midshaft fragment with a green fracture (8), examples of charred bone (9–11), and examples of bone covered in adhering matrix (12–14). b, Results of the stable carbon δ13C and δ18O analysis by taxa. Roman numerals refer to ref. . Left scale bar applies to A1–6 and right scale bar to A7–14.
See this image and copyright information in PMC

References

    1. Broodbank, C. The Making of the Middle Sea: A History of the Mediterranean from the Beginning to the Emergence of the Classical World (Thames & Hudson, 2013).
    1. Broodbank, C. The origins and early development of Mediterranean maritime activity. J. Mediterr. Archaeol.19, 199–230 (2007). - DOI
    1. Kelly, R. L. The Lifeways of Hunter-Gatherers: The Foraging Spectrum (Cambridge Univ. Press, 2013).
    1. Cherry, J. F. & Leppard, T. P. Patterning and its causation in the Pre-Neolithic colonization of the Mediterranean islands (Late Pleistocene to Early Holocene). J. Isl. Coast. Archaeol.13, 191–205 (2018). - DOI
    1. Clarkson, C. et al. Human occupation of northern Australia by 65,000 years ago. Nature547, 306–310 (2017). - DOI - PubMed

Publication types

Grants and funding

LinkOut - more resources