Kalambo Falls has been known to archaeologists since the 1950s. Excavations run by J. Desmond Clark recovered wooden fragments from the waterlogged sand beside the river - extraordinary finds, because wood almost never survives from the Stone Age. Soil fungi, insects and weather reduce timber to nothing within a few centuries. But Kalambo's sediments held the water table permanently high, starving oxygen from the buried layers and freezing decay in its tracks. Clark's team could not quite prove the fragments were human-made. Wear and taphonomic damage muddied the picture. The wood stayed in museum drawers for decades, waiting. When Larry Barham of the University of Liverpool reopened the site in the 2010s, better dating methods and high-resolution analysis were ready for it.

What Barham's team found was specific in a way that matters. The upper of the two interlocking logs has a deliberate notch. Infrared spectroscopy suggests fire was used during its working. Under the microscope the notch shows the short, regular scoring of scraping and adzing. The lower log, the one the upper piece rests against, carries V-shaped cut marks at its midpoint and along a narrowed end that fits inside the notch. These are not accidents. Separately, across the site, archaeologists recovered a wedge and a digging stick dating to between 390,000 and 324,000 years ago - younger than the structure itself but belonging to the same tradition of worked wood. Somebody, for a long, long time, was building things here.

The discovery predates the earliest known appearance of our species by more than 100,000 years. No hominin remains have been found at Kalambo Falls, so nobody can say for certain whose hands did the work. A 300,000-year-old Homo heidelbergensis skull has been found elsewhere in Zambia, and heidelbergensis is a leading candidate - a tall, heavy-browed relative whose toolkits we already know included careful woodwork. The timing of the Kalambo structure, the team notes, coincides with a period when the river basin was forested, the water table was high, and elevated platforms above the floodplain would have made life easier during the rainy months. In other words: a good reason to build. And someone, archaic human but recognisably thoughtful, did.

Barham's team argues, carefully, that wood may have been at least as important as stone in the Palaeolithic toolkit - we simply cannot see it. Stone tools survive. Wooden tools rot. For every handaxe in a museum drawer there were probably dozens of spears, hafts, digging sticks, platforms and traps that vanished completely. The Kalambo structure suggests that Lower Palaeolithic hominins were already doing something conceptually enormous: linking multiple shaped pieces to make a single object. That idea - combining parts into a whole - is the intellectual seed of hafting, architecture, mechanism, furniture, everything that followed. Standing on the riverbank at Kalambo, looking at wet sediment the colour of old tea, it is hard to keep the scale of the find in your head. The people who made this are so far back in time that we do not even have a name for them. And yet they already knew how to join two pieces of wood so they would hold.