In the 1950s, the archaeologist J. Desmond Clark was excavating below Kalambo Falls, a 235-metre waterfall on the Kalambo River near the border of Zambia and Tanzania, when he came across something that didn’t sit easily with what he expected to find. Among the stone tools and animal bone typical of a Palaeolithic site, there was a log with tapered ends and a distinct notch cut into it. Clark was convinced the shaping was deliberate, not the work of a river or a beaver. He had no way to prove it.
Radiocarbon dating, the standard tool of the day, only reaches back around 50,000 years before its signal runs out entirely. Applied to Clark’s wood, it simply returned an answer of infinite, telling him only that the sample was older than the method could measure. An attempted uranium-series date came back at roughly 180,000 years, a figure nobody, including Clark, fully trusted. He died in 2002 without the dating technology that could have settled the question.
A team led by Larry Barham at the University of Liverpool, working on the Deep Roots of Humanity project returned to the same stretch of riverbed and found overlapping logs with matching notches, essentially the same feature Clark had identified sixty years earlier. This time, the tools existed to date it properly.
What luminescence dating actually measures
The dating method used, published by Barham, Geoff Duller of Aberystwyth University, and colleagues in Nature in September 2023, is luminescence dating, applied not to the wood itself but to sand grains buried around it. The technique measures how long it has been since those mineral grains were last exposed to sunlight; once buried, they accumulate a faint radiation signal over time that can be read back as an age. Applied to the sediment surrounding the wooden structure, it returned a date of at least 476,000 years, with a margin of roughly 23,000 years either way.
That is an order of magnitude older than Homo sapiens as a species. The paper itself doesn’t name which hominin species is responsible, and is notably restrained about the point. What it does establish is that whoever it was had the capacity to select suitable timber, shape it deliberately with stone tools, and join two pieces together in a way that served some functional purpose, roughly half a million years before anyone in that lineage had a written language, agriculture, or, as far as the evidence shows, permanent architecture of the kind usually associated with the emergence of complex culture.
Not a hut, and the authors are careful to say so
Coverage of the find has sometimes reached for the word hut, and it’s worth being precise about why the researchers themselves avoid it. Analysis of the wood’s surface found scraping and adze-like tool marks on the notch, and V-shaped cut marks on the log it fits against, evidence consistent with intentional shaping rather than natural wear. But a joined pair of logs is a long way from a roofed structure, and the paper’s authors are explicit that they can’t claim more than the wood itself shows.
Barham’s own account of the find, as reported in coverage of the paper, is more modest: the structure was probably part of a wooden platform, perhaps a walkway, a base for keeping food or firewood dry, or the foundation for a dwelling built from other, less durable materials that haven’t survived. The description offers three possible functions rather than committing to one. The same excavation also recovered four separate wooden tools, dated between 324,000 and 390,000 years old, which at least establishes that working wood at this site wasn’t a one-off event but a sustained practice across a long span of time.
Why wood almost never survives this long
Part of what makes the find unusual has nothing to do with the notch and everything to do with the material. Wood rots. Stone tools dominate the archaeological record from this period not because early hominins used more stone than wood, but because stone survives and wood generally doesn’t. Kalambo Falls preserved this piece because the site has stayed permanently waterlogged, an oxygen-poor environment that slows the decay processes that would otherwise have destroyed it long ago.
That preservation bias means the wooden structure isn’t necessarily as anomalous a behaviour as it looks. It may simply be one of the only surviving examples of something that was, at the time, unremarkable and common. Barham told the Associated Press that it took him a while to appreciate what the team was looking at: “It didn’t look very nice, to be honest. But it is much more complex than I thought.” Clark saw the same thing decades earlier and reached the same intuition, without a way to prove it. What changed between his excavation and Barham’s wasn’t the wood sitting in the riverbed. It was a dating method that had finally caught up.