Andavakoera gorge, Montagne des Français, Madagascar
Lakaton'i Anja
Tools in sand dated about 4,380 to 3,470 years ago, or much later (optically stimulated luminescence against charcoal and termite burrows)
Lakaton'i Anja is a limestone shelter in the Andavakoera gorge, looking northeast, 18 metres long and 4.5 metres wide inside the drip line. A 2013 study reports small stone tools in the lower layers and optically stimulated luminescence ages of 4,380 plus or minus 400 and 3,470 plus or minus 370 years for layer 5. Charcoal from those layers is much younger. So is a bead that fell down a termite burrow.
Why this is extraordinary
South of Antsiranana, the Montagne des Français breaks into limestone gorges. Lakaton'i Anja sits in the Andavakoera gorge, a shelter 18 metres from east to west and 4.5 metres deep to the drip line, opening northeast over a meadow. You climb about 6 metres of fallen blocks to the floor. The first digs, in 1986–1988, found an upper occupation with local pottery and trade goods of the 11th to 14th centuries, and a poorer lower unit with bone and charcoal. In 2011 the team came back looking for small stone tools, because a sieve that is too coarse throws them away.
They found them. Coarse chert flakes, fine reddish chert, crystal quartz, small blades, crescentic pieces, a trapezoid. Many are under 2 centimetres. The 2013 paper reads them as a forager’s kit, closer to microliths from Africa and southern Arabia than to the stone tools of Southeast Asia, and unlike the iron of the later villages. Tortoise, birds, tenrecs, lemurs, fish, and coastal shell are in the lower layers. The coast is a couple of kilometres off. Optically stimulated luminescence on the quartz of layer 5 gave 4,380 plus or minus 400 years and 3,470 plus or minus 370 years. The paper’s plain-language span is roughly 1460 to 2370 BC. An optical age higher in the section matches the imported pottery, which is a reason to trust the method where the layer is obvious.
The other clock does not agree. Charcoal in layers 4 and 5 has been put at about 1,460 to 930 years before present. The original paper says termites carried young charcoal down into older sand, because termites want moisture, and that the majority of the quartz grains still date the layer. A reconsideration looks at the same burrows and refuses the exception. In five of six optical samples, 28 to 40 percent of the grains belong to a younger population. Pottery moved down. A glass bead from layer 1 was found in layer 5. Flakes smaller than 2 centimetres are exactly the size of thing a burrow moves. If the tools rode with the bead, the shelter’s stone age is the medieval occupation, and the old ages belong to sand.
Christmas River, far to the south, is a different argument: cut marks on an elephant-bird bone dated to more than 10,500 years, with no stone tools in that bone bed. This room does not borrow that bone, and that room does not borrow these flakes. No paper used here aims the northeast mouth at a solstice. A meadow and a cliff are a place to live, not yet a measured sky. The photograph is the massif’s skyline, dry woodland in front, not the gorge and not the sieve. What this changes: if the host sand dates the tools, foragers with microliths were in the far north more than a millennium before the villages. If the burrows date them, Madagascar’s dug archaeology still starts with those villages, and the old quartz is a sediment story.
Why it matters
If the tools are as old as the sand’s luminescence, foragers were in northern Madagascar more than a thousand years before iron-using villages. If termites moved the tools down into old sand, the island’s archaeological clock stays with the villages of the first millennium CE. The cut-marked elephant bird at Christmas River is a different site and a different kind of evidence.
The shelter was dug in 1986–1988 and again in 2011. Upper layers hold local pottery and imported sherds, including Persian Gulf sgraffiato of the 11th to 13th centuries and a fragment of a southern Chinese bowl of the same broad centuries. Those layers are the 11th to 14th centuries by the pots and by luminescence that agrees with the pots. Below them, flakes of chert, some of it a fine red stone, small blades, crescents, and a trapezoid sit in redder sand. The 2013 paper reads them as a forager kit, unlike later Malagasy iron tools, and more like microliths of Africa and southern Arabia than like Southeast Asian stone tools. Bones of tortoise, birds, tenrecs, lemurs, fish, and coastal shell are in the lower layers. The objection, set out in a reconsideration and in a 2018 review of the colonisation, is bioturbation. Termite burrows from above penetrate layers 4 and 5. In five of six luminescence samples, 28 to 40 percent of the quartz grains were a younger population mixed in. A glass bead from the upper occupation was found in layer 5. Pottery moved down. The flakes are smaller than 2 centimetres. Anything that small can follow a bead. No star alignment of the shelter is published. The opening faces northeast over a meadow. A direction of a mouth is not a solstice unless someone has measured the horizon, and nobody in these papers has.
Photo tour
1 photograph, unedited. On a phone, swipe the row. Choose one to look closely.
- The shelter mouth, the section, and the tools are not in the photograph. The frame is the limestone massif.
What is still unexplained
These sit beside the consensus account, with the same visual weight. None of them is marked solved.
Location and dating
- Location
- Lakaton'i Anja, in the gorge of Andavakoera, Montagne des Français, south of Antsiranana. The 2013 paper prints S 12° 20′, E 49° 21′, the pin at 12.333 south, 49.35 east. The photograph is the massif, not the shelter floor.
- Period label
- Current consensus — open to revision Upper layers, 11th to 14th centuries CE, by imported pottery and by luminescence that matches it. Lower layers, either about 4,380 to 3,470 years ago on the host sand, or within the last 1,500 years if the tools moved with the charcoal and the bead.
- How this dating was done
- Optically stimulated luminescence on single quartz grains, and radiocarbon on charcoal. Pottery from the Persian Gulf and southern China dates the top. Termite burrows are the stratigraphic objection.
- What could change it
- A tool-bearing patch with no burrows, dated by both methods on samples that contain one grain population, would end the split. A microscopic study that finds termite traces on the flakes themselves would end it the other way.
What is actually there
- The shelter floor is 18 metres east–west and 4.5 metres wide inside the drip line.
- Layer 5 optical ages are 4,380±400 and 3,470±370 years.
- Charcoal in the lower layers is reported around 1,460 to 930 years before present.
- A glass bead from the upper layers was found in layer 5. Pottery also moved down.
- In five of six optical samples, 28 to 40 percent of grains were a younger population.
- The tools include small blades, crescents, and a trapezoid, many under 2 centimetres.
Open questions
| Question | Status |
|---|---|
| Are the stone tools as old as the host sand? | Open. The luminescence majority says yes. The bead, the pots, and the charcoal say objects in that sand are younger. |
| Do these tools show a forager visit before the villages? | Open. They are a different kind of object from the cut-marked bird bones at Christmas River, and they are not yet a village. |
The current narrative
Where publication stands today. It is the best-supported reading, and it is open to revision. It is not a verdict.
The current narrative — open to revision
Medieval tools in older, burrowed sand
People used the shelter in the centuries of the imported pottery. The old optical ages date sand, and termites mixed younger artefacts down into it.
What supports this
- Charcoal in layers 4 and 5 is about 1,460 to 930 years before present, in the same broad era as the upper occupation.
- A bead and sherds moved down. The paper records significant sedimentary contamination.
- Flakes under 2 centimetres are the size of object burrows move. The criticism says all the cultural material was susceptible, not only charcoal.
What pushes back
- If only charcoal moved, the tools can be old. The original paper makes that distinction on purpose.
- An optical age at the top that matches Song-dynasty and sgraffiato pottery shows the method can hit the right century when the layer is richer.
- A forager kit unlike village iron is a real difference in the objects, even if the year is wrong.
What would settle it
It is the centre because a bead in layer 5 is an object that travelled, and the tools are the same size class. It moves if a burrow-free patch of tools still dates to the fourth millennium before the present.
Emerging narratives
Other published readings. Each may be right. None is adopted here unless the evidence carries it.
An emerging narrative
Foragers here before 2,000 BC
The lower tools belong with the host sand, older than 3,470 years, and record intermittent foraging long before iron-using villages.
What supports this
- Two optical ages in layer 5, in stratigraphic order with the ages above them, and an upper optical age that matches the imported pottery.
- The paper argues that termites moved charcoal down but that the main grain population still dates the layer the tools sit in.
- Red chert tools are concentrated below, which is odd if every flake fell from the medieval floor.
What pushes back
- A bead and pottery of the last thousand years are in layer 5. Small flakes can take the same ride.
- 28 to 40 percent introduced grains means the samples are mixed. A majority age is not a sealed age.
- The same industry at Ambohiposa is associated with the last millennium. One unchanged kit across 3,000 years needs a reason.
What would settle it
It is not the centre, because the burrows are described by the original paper and then used by the critics as the simpler account. It stays because the optical ages are in order and the upper one matches the pots, which is a real check.
Should the current view be reevaluated?
Open to revision
Two clocks in one shelter disagree, and the shelter’s own section shows the burrows that could cause that. The tools are real. Their year is not. Open keeps both readings, and keeps this shelter distinct from the cut marks at Christmas River.
What drives this
- Optical ages of 4,380±400 and 3,470±370 years in layer 5.
- Charcoal, pottery, and a bead that are much younger in the same sands.
- Termite mixing, with 28 to 40 percent younger grains in most optical samples.
This is the museum’s own weighing of the evidence. It is not a visitor vote, and it is separate from the slider.
Weigh it yourself
Anonymous. Nothing here is stored except on this browser. Your lean does not change the assessment beside it.
Visitor lean so far — not evidence. On this browser only; a shared tally is not collected.
If the contested reading is right
What would follow if a reading that is not the consensus turned out to be true, and what could settle it. This explains the stakes. It does not adopt the reading.
The tools are older than 3,500 years
Northern Madagascar had foragers with a microlithic kit long before the iron-using villages, and the island’s environmental history has to include them.
Why it is still worth a test
A patch without burrows, one grain population, and a radiocarbon sample that agrees with it.
Sources
- Stone tools and foraging in northern Madagascar PNAS, 2013, the tools, the shelter size, and the optical ages
- New evidence of megafaunal bone damage indicates late colonization of Madagascar PLOS ONE, 2018, the bioturbation reading of the same layers
Image credits
- The Montagne des Français, near Antsiranana. Lakaton'i Anja is a shelter in the Andavakoera gorge on this massif. The gorge mouth is not this skyline. The photograph is the hill, not the dig. . Hiroki Ogawa, CC BY 3.0. Shown unedited.