Cambodia – 69,000 years ago

Laang Spean Cave

The Laang Spean Cave (Cave of the Bridges) is one of the major prehistoric sites in mainland South-East Asia, not only because of its size but also because of its stratigraphic sequence, which is nearly 13 metres deep.

Laang Spean Cave during excavations in the western wall area, Unit 1, Chamber 2

The excavations carried out by the Franco-Cambodian Prehistoric Mission (Mission Préhistorique Franco-Cambodgienne [MPFC]) form part of an international training programme and aim to trace the chronological and cultural evolution of prehistoric settlement in Cambodia, from the last Palaeolithic hunter-gatherers to the first Neolithic settled communities on the Battambang Plain.

The Franco-Cambodian Prehistoric Mission (Mission Préhistorique Franco-Cambodgienne [MPFC])

Following a fifty-year hiatus in prehistoric research due to the civil war, the MPFC resumed work at Laang Spean Cave in 2009 thanks to the support of the MEAE, the MNHN, the CNRS, the SCAC at the French Embassy in Phnom Penh and, of course, our partners: the Cambodian Ministry of Culture and the Royal University of Fine Arts in Phnom Penh.

Nestled at the summit of a limestone massif (Phnom Teak Trang), the site covers an area of more than 1,000 m2, with three chambers that were discovered and excavated between 1964 and 1971 by Roland and Cécile Mourer, French archaeologists stationed in Cambodia at that time. With its exceptionally thick sedimentary layer, this site has, over the years, become a laboratory in which we have tested a wide range of modern methods for archaeological science (AMS C-14 dating, C-14 dating by mineral fractions, as well as OSL, ESR, sedimentary DNA, DNA, isotopes, geochemistry, etc.).

The main aim of the MPFC is to present a new, clear and well-dated chrono-cultural and chrono-stratigraphic framework, which is now based on four previously unpublished archaeological levels that have been discovered in recent years:

  • the Neolithic period (6 graves containing pottery, polished stones, jewellery, ornaments, etc.): dating from around 1,500 years BC, these burials confirm that the cave was used as a necropolis by sedentary groups living on the Battambang plain,
  • the Upper Palaeolithic, known as the “Hoabinhien” (a chipped pebble tool industry of the last hunters of the forests of South-East Asia), has been dated to between 11,000 and 5,000 years BP (OSL and AMS C-14 dating). The reference profile was published in 2015 in Journal of Archaeological Science,
  • the pre-Hoabinhian: This is an intermediate layer containing flint artefacts dating from between 11,000 and 26,000 years ago (OSL dating),
  • the very ancient level containing tools on “ancient” boulder-like pebbles (3 artefacts) discovered in 2017, and corresponding to a deep stratum dated to -71,000 years BP (OSL dating, at –5 m). Following on from this, an ash layer one level below likely corresponds to one of the eruptions of the Toba supervolcano (Sumatra). This is currently undergoing geochemical and geochronological analysis (tephrochronology, Ar/Ar). Identifying the signature of such a global climatic event in the continental fill would be a major finding in the context of scenarios concerning the settlement of modern humans in the Far East.

The longest sequence of records to date from Laang Spean provides sufficient archaeo-stratigraphic markers to trace, both synchronically and diachronically, the chronological and cultural evolution of prehistoric settlement in Cambodia from its very origins: history before Angkor.

Laang Spean Cave

The Franco-Cambodian project at Laang Spean, which began twelve years ago (in 2009), has made it possible to resume excavations in an exceptional cave measuring 1,000 m2, at ground level, which had been discovered fifty years ago by two French archaeologists, Roland and Cécile Mourer (CNRS, Lyon), who were based in Cambodia at that time. The resumption of excavations by the Franco-Cambodian mission, under the auspices of the MEAE, the CNRS, the MNHN, the Cambodian Ministry of Culture and the SCAC at the French Embassy in Phnom Penh, has uncovered an exceptional chronological sequence extending to a depth of thirteen metres and comprising several archaeological layers. That is to say, different human occupations over time.

After twelve years of excavations carried out across an area of nearly 80 m2, the bedrock of the cave was reached at a depth of 13 metres (12.77 metres), which now makes it possible to launch an intensive programme of dating (OSL, ESR, U-Th, etc.), palaeo-environmental and climatic studies (geochemistry) and, of course, archaeological research involving the analysis of archaeological (lithic and faunal) and anthropological remains. As for the human remains, they date from two periods: the Hoabinhian (hunter-gatherer) and the Neolithic.

The four archaeological strata discovered span the period from the Upper Palaeolithic to the Neolithic, which remains a relatively little-known period in Cambodia compared with subsequent periods such as the Metal Age (protohistory), the Pre-Angkorian period and the incomparable Angkorian period.

The aim of the excavations carried out by the Franco-Cambodian Prehistoric Mission is to provide information, from both a diachronic and synchronic perspective, on all the recorded prehistoric periods. The extensive archaeological and sedimentary sequence demonstrates the full scientific value of reopening the excavation site at Laang Spean, both in terms of its depth and its surface area. The archaeological discoveries at this site, which is exceptional for peninsular South-East Asia, provide us with the necessary evidence to give an account of the long “history before Angkor”.

The oldest site in South-East Asia

The Lang Spean cave site is a vast cavity measuring 1,000 m2 and comprising three adjoining chambers where the Franco-Cambodian Prehistoric Mission undertook excavations in Chamber 2 between 2009 and 2020. The work currently underway is marked by the conclusion of the excavation (2020 campaign) upon discovery of the bedrock, and thus the uncovering of the full stratigraphy. The team is now focusing on launching a programme involving the analysis of archaeological material and cross-referenced scientific analyses, with a view to studying this exceptional 13-metre-deep archaeological-sedimentary sequence from an environmental and natural history perspective.

Continuous monitoring of the site is extremely helpful in identifying markers of time and space, in particular human presence in the cave 71,000 years BP, as evidenced by the discovery of three stone tools. This makes Laang Spean the oldest cave site in mainland South-East Asia. It can therefore be said that (modern) humans were already present at the site during the IOS4 isotopic stage, whilst the sedimentary fill dates back to between 200,000 and 300,000 years BP, which means that the cave began to fill during the Late Middle Pleistocene period (IOS 7–8).

The discovery of a Neolithic burial site in a cave in Cambodia

Chamber 2 yielded a long stratigraphic sequence, the most recent layer of which consists of Neolithic burial remains in the form of six grave pits (five men and one woman) dated to a few decades after around 3,300 BP. Of the six burials in Chamber 2 at Laang Spean, the one that we have chosen to discuss is the “Chief” burial or Burial D26-29, which is undoubtedly the most spectacular from an archaeological point of view and the best preserved. It was dated directly to 3,370±30 BP/Cal using a bone sample. BC (Oxcal 4.3) -1 745-1566 Cal BC using mineral fraction by A. Zazzo (MNHN-CNRS). The excavation of this pit involved around ten excavators under the direction of palaeoanthropologist V. Zeitoun (CNRS) and lasted one month.

This is the grave containing the greatest wealth of archaeological (funerary) material, along with a rather robust skeleton. The site contains the remains of an adult whose large body was buried in an extended position, alongside numerous archaeological artefacts such as pots, polished stone blades (adzes) and a variety of ornaments including bracelets and ochre beads.

The body is lying on its back, oriented along the SE–NW axis of the cave, with its head lying to the SE. Five vases were placed around the body of the deceased. Two pots were placed between his knees. There is a vase to the right of the right knee and a pot with a ring-shaped foot on the outside of the right shin. Another pot is located at the upper end of the right femur. Orange-coloured, biconical white limestone beads were found around the neck, strung in a row (see necklace) beneath the lower jaw and along the pelvic girdle from the right wrist to the left coxal bone. White calcite bangles, carved from a large seashell (Murex type), encircle the upper limbs: one at the right elbow, two at the left wrist.

The skull, which is visible with the face facing towards the feet, shows a slight rotation to the right. Anatomically, the mandible is joined to the maxilla. Ante-mortem dental avulsions affect all the incisors (upper and lower) and the two upper canines. The skull is wedged in place at the occipital region by a small limestone block. The upper limbs are held out parallel to the body. Their positions appear to be restricted by the weight of the bangles, which have locked out the joints. Consequently, the right humerus has rotated inwards, as evident from its lateral aspect, whilst the right clavicle has been straightened. This movement appears to be caused by the ring pressing against the elbow. The supinated forearm is positioned beneath the closed right coxal and beneath a pot. The hand is extended. The verticalisation of the left clavicle is accompanied by the outward displacement of the humeral head. The humerus is visible from its anterior surface. The forearm is extended and pronated. The left wrist is located beneath the coxal bone, which is vertical, and two calcite rings cover the hand.

The spine is connected. The rib cage has dropped. The pelvis remains intact, with no collapse of the pubic or hip bones. Numerous biconical beads are strung in a line from the right elbow to the left hip, passing over the sacrum. The lower limbs are extended in anatomical alignment with the pelvis. Outward rotation reveals the medial (inner) surface of the right femur. The lower right leg is enclosed by the row of pots. A decimetre-sized block of limestone restricts the space outside the left femur. The right foot is extended. The inside of the left foot can be seen. The collapsed position of the metatarsals indicates that the feet were in an upright position. The left leg rotated outwards. The anthropobiological data indicate that this individual is male with a probability of 99.99 per cent, according to the DSP method (Murail et al. 2005). His height was 174.15 ± 2.9 cm, according to Pureepatpong et al. (2012). The artefacts associated with this individual, together with his imposing stature, confer upon him a special status amongst all the burials that have been identified to date.

Unlike other burials, such as P31-32, which has been dated to 2,960±30 BP, and I31, which has been dated to 3,310±29 BP, this one shows no evidence of the knees being raised during the burial of the body, nor of the grave being covered with limestone blocks when it was sealed. Several common features in the arrangement of the graves (the inclusion of pottery as offerings and/or the positioning of the body at the time of burial) as well as ante-mortem practices (tooth extraction) or physical characteristics (relatively tall stature of 1.70 m for men) link the burials at Laang Spean. Based on our current research, the question of the social, chronological or cultural significance of differences in funeral practices cannot be answered due to an insufficient number of individuals being recovered from the graves and/or a lack of regional archaeo-anthropological points of reference. The development of a programme focusing on ancient genetics could help to establish the degree of kinship between these few individuals.

Conservation, restoration, public engagement and scientific training activities

To preserve the imprint of ground dating back 11,000 years (Hoabinhien = the last hunter-gatherers of South-East Asia), strewn with burnt bones of hunted animals, shells and carved stones, the MPFC organised a moulding operation which served as a scientific, museological/conservation and training event for young Khmer archaeologists in the techniques of casting both archaeological ground and artefacts. It was also the first time that an operation of this kind had taken place at a prehistoric site in Cambodia. This project was made possible by the visit of a specialist in prehistoric casting, Mr Cyril Planchand (CERP Tautavel), who led the field and laboratory work in Phnom Penh as part of the Franco-Cambodian Prehistoric Mission.

The area of which a mould was to be made, measuring approximately 4 m2, corresponds to the entire surface of square Aa26, together with part of square Ab26, as well as the stratigraphic section bordering the square on its left-hand side and a portion of the (western) wall against which the ground rests.

We will now set out the stages of the moulding process step by step: a 7-step moulding process.

  • Preparing the surface of which a mould is to be made. The moulding process begins with cleaning the ground to remove stones, bones, shells (Unio) and stone tools, in order to preserve them. They were secured to the ground using a mixture of sediment and Paraloid B72. We then proceeded to harden the floor. Three layers of Paraloid B72, diluted to 3% in acetone, were applied to the entire surface of which a mould was to be made (ground, cut-out, wall). Three coats of release agent were also applied evenly to the surface with a brush, using standard soap diluted in water.
  • Taking a mould. Three layers of silicone were applied in succession to the surface of which a mould was to be made. The silicone used (RTV 838) has a pot life of approximately three minutes, which leaves very little time to apply it. This type of operation is tricky because time is of the essence, which meant we had to prepare small quantities (500 g) to be applied in succession, requiring perfect coordination amongst the team members. Once the three layers of silicone had been applied, it was time to treat the undercuts. Numerous and substantial, these were dealt with using flexible “screed layers” incorporated into the silicone membrane. Cotton was placed over the undercuts and a final layer of silicone was applied to cover them.
  • Screed layers. It was necessary to make two of these from laminated polyester resin in order to remove the mould of the stratigraphic section without any risk of it tearing away. First screed layer: The point of separation between the two screeds is marked by two battens fixed together and covered with aluminium foil. Release wax was applied to the aluminium foil to prevent the screed from sticking to the formwork. Unfortunately, we couldn't find any Vaseline in Cambodia, so we used wax as a release agent. The silicone membrane was also coated with release wax. As we were not certain that the wax would be sufficient to prevent the resin from sticking to the silicone, the membrane was also covered with aluminium foil.
    • The first layer was made from two layers of 100 g/m² fibreglass impregnated with polyester resin and reinforced by a third 300 g/m² fibreglass layer, which was also impregnated with polyester resin. The fibreglass was laid so as to follow the contours of the ground as closely as possible and was extended up onto the battens forming the separating layer to create a new support surface on which the second screed layer would rest. Battens were then fixed to the screed to ensure its rigidity.
    • Second screed layer: Once the first screed had dried completely, the separation layer was removed. The second part of the untreated membrane was covered with release wax and aluminium foil. The surface formed by the first layer was carefully waxed and covered with aluminium foil to prevent the two layers from sticking together. The second screed layer was made using the same method as the first (two layers of 100 g/m² fibreglass and one layer of 300 g/m² fibreglass). As with the first screed layer, battens were fixed to it to ensure its rigidity. To join the two screeds together in preparation for making the positive cast, holes were drilled in the separating layer between the two screeds. Screws would be inserted there when the screed layers were reassembled. Two battens were also attached to the battens that reinforce the screed layers using wood screws. These two battens were removed immediately afterwards and would be put back in place when the positive cast was made.
  • Removing the mould. The screed layers were removed one by one and the silicone membrane was peeled away from the floor.
  • The positive cast was made using a gel coat and laminated polyester resin. The cast was reassembled (the cast halves were joined and the membrane was repositioned over them). A layer of gel coat was applied over the entire cast. Three layers of fibreglass (two 100 g/m2 layers and one 300 g/m2 layer) impregnated with polyester resin were then applied to reinforce the gel coat layer. Battens were fitted all the way round the mould. They served to stiffen the structure as a whole and would also act as a starting point for the battens used to reinforce the casing.
  • Casing. Casing refers to the wooden frame that surrounds the cast. First, a framework made of battens is constructed all the way round the cast. Sheets of plywood are fixed to the battens. The joints between the different sheets of plywood are filled with dental plaster (the only plaster that is hard enough in Cambodia).
  • Patina and finish. The patina was created using shellac, natural pigments, talc and sediment. The colours were prepared by mixing natural pigments and/or talc with shellac. They were applied to the cast using brushes. Sediment from the cave was then glued onto the cast between the objects to ensure that the finish, realism and aesthetic quality match those of the original ground, thereby enhancing archaeological objectivity and credibility, as well as serving museum display purposes.

The oldest tools from mainland South-East Asia

The deep “Roland Mourer” (R.M) test pit revealed a depth of 13 metres (Z5=79) below the cave's current ground level, which made it possible both to work vertically and to extend the excavation into the upper, more recent and richer layers (Hoabinhian and Neolithic). The impressive stratigraphic extent of this R.M. test pit had been the subject of a preliminary study by MPFC geologist S. Puaud, who had identified sedimentary and, probably, geochemical and climatic changes. This long chronological sequence has been partially published and is, to date, the longest cave sequence in South-East Asia; its lower section (-5 m) has been dated to 71,000 years using OSL dating.

The excavated area of this test pit was small (4 m²) but was sufficient to uncover three “ancient” lithic artefacts in a sandy-clay layer dating back to 71,000 years BP, containing rare bone fragments and flint hand axes (chert, slightly silicified limestone).

These three spectacular discoveries serve as key evidence of human activity pointing to the early settlement of Cambodia. These are previously unpublished lithic tools made from boulders and pebbles, which differ from the Hoabinhian (Holocene) material — also made from pebbles — found in the upper layers.

These are previously unpublished, exceptional specimens, eagerly awaited by the team, as they enable the characterisation of thetechnological behaviours of groups of the late Upper Pleistocene — which are still poorly understood in this region of Asia. These three pieces are very important because they shed light on a new chronological and cultural milestone within the stratigraphic context of the cave:

  • 1st tool: a chopper found in square E32, no. 6 (Z3 = 514 cm), i.e. almost at the level of the 71,000 years BP date obtained by OSL. It is a parallelepiped-shaped block of sandstone-limestone with a sinuous transverse edge, carved on one side.
  • 2nd tool, in F31, No.  7 (Z3= 546): a massive, multiple-chopper-type tool (distal and proximal sections) on an ovoid-shaped limestone pebble.
  • 3rd tool, in F31, No.  8, (Z3= 556): a chopping tool i.e. a tool made from a limestone-sandstone pebble, cut on two sides.

These three tools are fundamental and previously unknown indicators that shed light on the ancient migrations of modern humans from southern China to the territories of South-East Asia, where Cambodia was an important stopover, as evidenced by the Laang Spean cave refuge.

Useful links