East meets West: the Middle Pleistocene site of Rodafnidia on Lesvos, Greece

Nena Galanidou, James Cole, Giorgos Iliopoulos & John McNabb

Figure 1
Figure 1. The island of Lesvos and the site of Rodafnidia (© ArcGIS).
Click to enlarge.

This paper introduces a new inter-disciplinary and international research project focused on the Palaeolithic site of Rodafnidia on the Greek island of Lesvos, located in the north-eastern Aegean Sea (Figure 1). Rodafnidia, near the village of Lisvori, is less than a kilometre away from the south-western shore of the large Kalloni Gulf. It is a prolific open-air site spanning the Lower and Middle Palaeolithic. What makes this site unique is the richness of its Acheulean lithic assemblage which, so far, makes it unparalleled in Greece (Galanidou 2004).

The island of Lesvos is separated from the Anatolian coast by two sea straits: Mouselim and Mytilene. A glacial sea-level drop of only 50m would be enough to expose the eastern strait and connect the island with the Asian mainland, allowing hominin and terrestrial animal migration. Several fossiliferous sites have been found in Early Pleistocene deposits on the south coast of the island with over 15 mammal taxa including the giant macaque (Paradolichopithecus arvernensis), which is characterised as continental (Lyras & van der Geer 2007) and reflects a proximity to Anatolia. The site of Rodafnidia introduces the human perspective into the rich palaeontological record of Lesvos and adds a new point to the Palaeolithic map of the north-eastern Mediterranean.

The site: surface survey and excavation methodology
Figure 2
Figure 2. Looking north towards Rodafnidia and the Kalloni Gulf, showing the topography of the gorge and thermal spring flood plain to the south.
Click to enlarge.

Rodafnidia extends over four hectares to the north and north-west of the Lisvori thermal springs (Figure 2). The main concentration of finds is on a spur of a low hill, defined by a small gorge with running water coming from the hot springs. The southern and western sides of the hill are rather steep, rocky and composed of ignimbrites, whereas the northern side presents a smooth relief with low inclination, covered with olive groves.

A brief report by a group of amateur archaeologists (Harisis et al. 2000) who identified an extensive surface concentration of Palaeolithic artefacts instigated an initial visit (by NG) to the site in 2009, and a subsequent surface survey in 2010 (also by NG) to establish the site's boundaries and content. It produced a number of Large Cutting Tools (as described by Kleindienst 1962; McNabb et al. 2004) and highlighted the potential richness and antiquity of the site.

Systematic investigations in the field began in 2012 and included surface survey, as well as excavation to establish the site's stratigraphy. A series of 14 1m x 1m test pits (A to Ξ) were laid out at 20m intervals along a south–north transect across the top of the spur (Figure 3). In the end, 10 test pits and 3 larger L-shaped trenches, extensions A (11 x 1–3 x 1m), B (7.5 x 1–3 x 1m) and H (7 x 1–3 x 1m), were excavated to a maximum depth of 2.8m.

Figure 3
Figure 3. The trench layout, geology and site location of Rodafnidia in relation to local topography.
Click to enlarge.

A summary of the stratigraphy of the site is exemplified in the sections from extensions B and H (Figure 3). Units 0 and 1 contained archaeological finds whilst units 2 and 3 were devoid of finds. The lithology of the stratigraphic units suggests a relatively small alluvial plain as the depositional environment for the artefacts. Two types of deposit can be distinguished: floodplain (red to red-brown muds) and fluvial (conglomerate deposits). The fluvial deposits of Unit 1 characterise a river or stream network that shifted course over time, eroding and forming new river beds cut through the floodplain sediments (Unit 2) deposited during earlier flood events.

This situation can be identified in most of the trenches except the northernmost excavated trench, M, which is closer to the shore of the Kalloni Gulf (Figure 3) where green clays indicate the presence of a pond, marsh or small lake.

Sediment samples were collected for IR-OSL and TCN dating, for microfaunal preparation and, from trench M, for palynological preparation (results forthcoming).


The major technogroups identified in the Rodafnidida lithics are summarised in Table 1 and Figures 4–6. They are knapped on fossiliferous flint and andesite, raw materials associated with the island's volcanic history. They comprise:

  • Large Cutting Tools (LCTs)
  • Prepared Core Technology (PCT)
  • Ordinary flake cores
  • Flakes and detached pieces
  • Retouched flakes
Artefact provenance
ExcavationSurface finds (2010)Surface finds (2012)Total
Table 1. Summary of the main artefact categories for the Rodafnidia lithic assemblage.
Artefact typeBlade2--2
Core (tool)-123
Discoidal core--11
Flaked flake-11112
PCT (core)2-24
PCT (flake) (Figure 4)--22
Retouched flake-178
Simple prepared core--22
LCT typeCleaver (Figure 5)--11
Cleaver flake (Figure 5)1--1
Cleaver? (Figure 5)-112
Handaxe (Figure 6)231217
Handaxe tip--11

Figure 4
Figure 4. Examples of two PCT flakes from Rodafnidia.
Click to enlarge.

Figure 5
Figure 5. Examples of the LCT cleaver category from Rodafnidia.
Click to enlarge.
Figure 6
Figure 6. Examples of LCT handaxes from Rodafnidia.
Click to enlarge.

In sum, the artefacts recovered from controlled excavations and from field-walking clearly demonstrate the presence of Early Stone Age Acheulean hunters in this locality and possibly Middle Stone Age hominins associated with PCT (Levallois) technology. Within trench extension B it was observed that diagnostic Early Stone Age artefacts were present within the channel fills and a major research question for subsequent seasons will be to clarify the spatial relationship of the various stages of occupation indicated by the Lower Palaeolithic LCTs and the Middle Palaeolithic prepared cores.


Rodafnidia stands out as an exceptional and exciting island site, a target for enriching the Palaeolithic record of the north-eastern Mediterranean (e.g. Kuhn 2010) and for obtaining dates of Acheulean activity. The importance of this site, and our project, lies in:

  • The time and length of hominin presence. This covers the Lower and the Middle Palaeolithic; its onset should be placed provisionally in the Middle Pleistocene.
  • The size of the site. The Acheulean site may be extensive and the preliminary results presented here suggest that, in comparison to other Lower Palaeolithic sites of Greece and the broader Anatolian region, where large Lower Palaeolithic sites are relatively rare, future research at Rodafnidia has much more to tell us about the Middle Pleistocene settlement of the north-eastern Mediterranean.
  • The geography of the site. Rodafnidia's position locally (in a fluvio-lacustrine environment of the Kalloni basin in a landscape dominated by volcanic rocks) and regionally (a prominent position on the border of two continents) makes it a unique target for research addressing the chronology, timing, direction and pattern of Middle Pleistocene hominin dispersals out of Africa and out of Asia into Europe, and vice versa.


The authors would like to thank Kostas Athanassas, Laboratory of Archaeometry, OSL Unit, Demokritos; Andrew Chamberlain, Professor of Biological Anthropology, University of Manchester; Mimi Hill and Neil Suttie, Palaeomagnetism Unit, University of Liverpool; Thanos Katerinopoulos, Professor of Geology, University of Athens; Geoffrey King, Équipe de Tectonique - Institut de Physique du Globe, Paris; Andreas Magkanas, Professor of Geology, University of Athens; Chronis Tzedakis, Professor of Geography, University College London; Katerina Vasileiadou, Palaeontology Curator, Sigri Natural History Museum and Nikos Zouros, Professor of Physical Geography, University of the Aegean. Their help and support for the project have been invaluable.


  • GALANIDOU, N. 2004. Human presence in the Balkans during the Palaeolithic, in H.J. Griffiths, J. Reed & B. Kristufek (ed.) Balkan biodiversity: pattern and process in the European hotspot: 147–65. Dordrecht: Kluwer Academic.
  • HARISIS, H.B., P. DURAND, M. AXIOTIS & T.B. HARISIS. 2000. Remains of Palaeolithic occupation on Lesvos. Archaeology and Arts 76: 83–87.
  • KLEINDIENST, M.R. 1962. Components of the East African Acheulean assemblage, in G. Mortelmans & J.A.E. Nenquin (ed.) Actes du IVe Congrès Panafricain de Préhistoire et de l'Étude du Quaternaire 81–103. Tervuren: Musée royal de l'Afrique centrale.
  • KUHN, S. 2010. Was Anatolia a bridge or a barrier to early hominin dispersals? Quaternary International 223–224: 434–35.
  • LYRAS, G.A. & A.A.E. VAN DER GEER. 2007. The Late Pliocene vertebrate fauna of Vatera (Lesvos Island, Greece). Cranium 24: 11–24.
  • MCNABB, J., F. BINYON & L. HAZELWOOD. 2004. The large cutting tools from the South African Acheulean and the questions of social traditions. Current Anthropology 45: 653–77.


*Author for correspondence

  • Nena Galanidou
    Department of History & Archaeology, University of Crete, University Campus Gallou, Rethymno, 74100, Greece (Email: ngalanidou@phl.uoc.gr)
  • James Cole*
    Institute of Archaeology, University of Oxford, 36 Beaumont Street, Oxford OX1 2PG, United Kingdom (Email: james.cole@arch.ox.ac.uk)
  • Giorgos Iliopoulos
    Department of Geology, University of Patras, 26500 Rio, Patras, Greece (Email: iliopoulosg@upatras.gr)
  • John McNabb
    Department of Archaeology, University of Southampton, Avenue Campus, Highfield, Southampton SO17 1BF, United Kingdom (Email: J.McNabb@soton.ac.uk)