Kalambo River, near Mbala, Zambia
Kalambo Falls
At least 476,000 ± 23,000 years (the date is the sand that buried the logs)
Two logs cross in the sand below Kalambo Falls. One has a notch cut across it so the other can pass through. The sand that buried them is at least 476,000 years old, plus or minus 23,000, before the oldest fossils assigned to our own species.
Why this is extraordinary
On the bank of the Kalambo River, just above the falls and the border with Tanzania, waterlogged sand kept wood that almost never lasts. In 2019 a trench below the present river level found two logs locked together. The upper log, of the tree Combretum zeyheri, is 141.3 centimetres long and 25.6 centimetres across. Where it crosses the trunk beneath, at about 75 degrees, a U-shaped notch 13.2 centimetres long and 11.4 centimetres wide has been cut across the grain, and the lower trunk passes through that notch. Both pieces have shaped ends. Chopping and scraping marks sit on the wood, and stone tools, including handaxes and cleavers, lay in the same sands.
The clock is the sand, not the timber. Sixteen samples were dated by luminescence, the light minerals release after they are buried and no longer see the sun. Younger sands were dated on quartz by optically stimulated luminescence, a light clock in sand grains, reset by sunlight. The older sands, which quartz cannot reach, were dated on potassium feldspar by a post-infrared method. The ages fall in order, with one reversal, and cluster. The deepest cluster, the one that holds this join, has a mean of 476,000 years, plus or minus 23,000. The catch is that the paper prints that figure as a minimum, because it dates the burial of the wood. A middle cluster, 390,000 plus or minus 25,000, holds one wooden tool. An upper cluster, 324,000 plus or minus 15,000, holds two more. No hominin bone was found with any of it, so the maker has no species name.
Higher in the same bank the 2019 dig recovered a wedge, a digging stick, a cut log and a tapered piece with a chop mark. An experiment with a hafted cleaver reproduced the stepped chops on the cut log, though the experimental cuts were a little shallower than the archaeological ones, which run 13 to 25 millimetres deep. Infrared spectroscopy did not show that fire had been used to shape the notch, and the evidence on that point is indeterminate. An earlier excavation at the falls had already described a grooved log of similar length and read it as part of a structure. Its age was not secure. The 2019 join is the one with a luminescence cluster around it.
Shaped wood is older elsewhere. A polished plank fragment from Gesher Benot Ya'aqov, in Israel, is more than 780,000 years old. What Kalambo adds is a carved interlock, two large pieces made to hold each other. The report's reading is a construction: a walkway, a platform, or a raised foundation above seasonally wet ground. Two locked logs are not a house, and they are a joint with no published parallel in the African or Eurasian Palaeolithic.
The locked logs are a joint in wood, the earliest of that kind in the sources this room uses. No azimuth of the riverbank, and no star, is published for the site. A notch that locks two timbers is carpentry. It is not a tally of months unless someone prints a count, and none is on these pieces. The waterfall and the river are why people camped. They are water, not a sighted sunrise.
What this changed
The shift, in short. The weighing further down is the longer argument.
- Older textbook view
- Textbooks kept structural carpentry late, and wood was allowed as fuel, as a spear, or as a rare preserved tool. A building, even a modest platform, was a thing of modern humans or of the last few tens of thousands of years. The oldest famous wooden weapons were the Schöningen spears, then placed near 300,000 years and in Europe.
- New evidence
- The 2019 trench recovered two interlocking logs and, in younger sands, a wedge, a digging stick, a cut log and a notched branch. The join is a transverse notch, and stone tools lie in the same deposits. An earlier, less securely dated grooved log from the same falls had already been read as structural.
- How it is dated
- Luminescence on the sand, not on the wood. Quartz optically stimulated luminescence covers the younger samples, and post-infrared infrared stimulated luminescence on potassium feldspar covers the older ones, which quartz cannot date. The deepest cluster is 476,000 ± 23,000 years, the middle is 390,000 ± 25,000, and the upper is 324,000 ± 15,000. One pair of samples is out of order, and the deepest figure is a minimum age for the burial.
- What moved
- The oldest known structural use of wood moved to Zambia, to at least 476,000 years ago, and to a hominin who is not named and is not Homo sapiens on any current clock. Wooden tools at the same site now sit between about 390,000 and 324,000 years. The European spears are no longer the start of the wooden record.
- What is still disputed
- A notch is not yet a building. A platform, a walkway and a foundation are readings of two locked pieces. No bone names the maker. The date is the sand. Fire-shaping was tested and came back indeterminate. Absence of wood at other sites of this age is mostly decay, so one waterlogged bank cannot show how common the skill was.
Why it matters
If the notch is a joint, people were fitting large timbers together hundreds of thousands of years before Homo sapiens. If it is only a shaped log, the date still leaves worked wood in Africa far older than the famous European spears. The open word is "structure."
The figures are the paper's own, shown unedited: the trench and its plan, the tools, a model of the notch, close views of the marks, and the section with the luminescence ages. The waterfall is the landscape beside the bank, not the logs.
Photo tour
6 photographs, unedited. On a phone, swipe the row. Choose one to look closely.
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
- Kalambo Falls, on the Kalambo River near Mbala, northern Zambia, at the border with Tanzania. The coordinate is the falls. The 2019 trench is on the riverbank there, in deposits that go below the present water line.
- Period label
- Current consensus — open to revision Deepest wood in sands dated to at least 476,000 ± 23,000 years. Wooden tools in clusters at 390,000 ± 25,000 and 324,000 ± 15,000 years. Current consensus on the notch and the burial age — open on what was built, and on who built it.
- How this dating was done
- Luminescence on sand that brackets the wood. Younger samples use single-grain quartz optically stimulated luminescence. Older samples use post-infrared infrared stimulated luminescence on potassium feldspar, because quartz saturates. The ages are mostly in stratigraphic order, the order of layers, which says what came before what. The join's cluster is the mean of the deepest samples, printed as a minimum age of burial.
- What could change it
- A second laboratory on the same feldspar samples, or a reversal that will not go away, would move the 476,000-year cluster. A hominin fossil in the same sand would name a maker. It would not, by itself, decide whether the notch was a platform.
What is actually there
- Upper log 141.3 cm long and 25.6 cm across, of Combretum zeyheri.
- Notch 13.2 cm by 11.4 cm, the lower trunk passing through it at about 75 degrees.
- Shaped ends, and chop and scrape marks, with stone tools in the same sand.
- Deepest luminescence cluster 476,000 ± 23,000 years, a minimum for the burial.
- Wedge, digging stick, cut log and tapered piece between about 390,000 and 324,000 years.
- No hominin bone with the wood. Fire-shaping of the notch was indeterminate.
- No month-count is carved on the locked logs.
Open questions
| Question | Status |
|---|---|
| Was the join a platform, a walkway, or a simpler fastening? | Open. The lock is observed. The rest of any structure is not. |
| Who cut it? | Open. The age is before Homo sapiens on the current fossil clock. The species is not in the trench. |
| How common was this, away from a waterlogged bank? | Open. Wood of this age almost never survives, so absence elsewhere is weak evidence. |
| Do the notches tally a moon? | Not in the sources used here. They lock one timber to another. |
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
An intentional joint, older than our species
Hominins at Kalambo cut a notch so that two large timbers would lock, and the sand that buried that join is at least 476,000 years old, plus or minus 23,000. Younger layers at the same bank hold wooden tools. The maker is unnamed. The purpose of the joint is not demonstrated beyond the lock itself.
What supports this
- The notch is transverse, shaped, and the lower trunk passes through it.
- Ends are tapered, and marks match chopping and scraping. A hafted-cleaver trial reproduced the stepped chops on the cut log.
- Luminescence samples around the join cluster at 476,000 ± 23,000 years, in stratigraphic order with the clusters above.
- Stone tools, including handaxes, lie in the same sands.
What pushes back
- Two pieces are not a building. Platform and walkway are inferences.
- The date is the burial. The cutting could be older by an unknown gap, and one pair of samples higher in the bank is reversed.
- No fossil identifies the species. Fire was not shown to have shaped the notch.
What would settle it
Best-supported for a cut joint and for a Middle Pleistocene age. The word structure is earned as far as an interlock. It is not earned as a house.
Emerging narratives
Other published readings. Each may be right. None is adopted here unless the evidence carries it.
An emerging narrative
A wooden architecture we mostly cannot see
The join is the surviving scrap of a real structural habit. People this early invested in a place, raising a walkway or a platform above wet ground, and the rest of that habit rotted away everywhere the ground was dry.
What supports this
- The report finds no parallel for this construction in the African or Eurasian Palaeolithic, which is what a vanished habit predicts when only a waterlogged site preserves it.
- An earlier grooved log at the same falls was already read as structural.
- Wooden tools in the layers above show that shaping timber here was not a single accident.
What pushes back
- One joint cannot prove a tradition. A clever fastening of two logs is also what the trench contains.
- Decay explains missing wood. It does not, by itself, restore the missing buildings.
- The experiment reproduced chops. It did not rebuild a platform and show that this notch was its footing.
What would settle it
A fair reading of why the find feels larger than two logs. It asks the trench to stand for a technology the trench does not fully contain.
Should the current view be reevaluated?
Open to revision
The old picture, in which structural wood begins with our species, does not survive this trench. The new picture is solid on the notch and on the burial age, and it is open on what the joint belonged to. Calling it the oldest building would say more than two locked logs can say.
What drives this
- A measured notch with the lower trunk still through it
- A luminescence minimum of 476,000 ± 23,000 years
- Wooden tools in two younger clusters at the same bank
- No plan of a larger structure, and no hominin bone
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 joint is the footing of a raised structure
Technical planning with large timber would sit hundreds of thousands of years before Homo sapiens, and dry-land sites of that age would be missing their organic half.
Why it is still worth a test
More waterlogged trenches of this age, and a refit that shows further pieces of the same construction.
The notch is a fastening and not a building
Worked structural timber would still be this old. Labels that say "the first architecture" would shrink to "the first known joint."
Why it is still worth a test
A published account of how else a notch of this size gets made, tested on wood of this species, beside the cleaver trial already reported.
Sources
Image credits
- The 2019 trench at Kalambo. A plan with a one-metre scale sits beside the logs themselves. The upper piece crosses the lower trunk. A ranging rod stands at the right of the photograph. The overlap is a cut notch, and the sand around the logs is at least 476,000 years old, plus or minus 23,000.. Barham, Duller, Candy and colleagues, CC BY 4.0. Shown unedited.
- The river and the 2019 sections. Individual luminescence ages are printed on the layers, from a few tens of thousands of years near the surface to samples around 450,000 to 490,000 years beside the deepest wood. The 476,000-year figure is the mean of that deep cluster, not a single label on this drawing.. Barham, Duller, Candy and colleagues, CC BY 4.0. Shown unedited.
- Worked wood from the 2019 dig. The paper identifies a as the notched structural log, b as a wedge, c as a digging stick, d as a cut log, and e as a tapered piece with a chop mark. The white bars are 10 centimetres.. Barham, Duller, Candy and colleagues, CC BY 4.0. Shown unedited.
- A model of the notched log, with a 30-centimetre scale. The waist marked Notch is the join. Letters and arrows on the model belong to the published figure.. Barham, Duller, Candy and colleagues, CC BY 4.0. Shown unedited.
- Close views of the same logs. Boxes and arrows are part of the published figure. They point at shaping marks on the dark wood, including chops and a cluster of small hewing marks.. Barham, Duller, Candy and colleagues, CC BY 4.0. Shown unedited.
- Kalambo Falls, the drop beside the archaeological bank. The wood was in the waterlogged sands, not in the falling water.. Kelvin Chidoti, CC BY-SA 4.0. Shown unedited.