Copper from Everywhere, Tin from One Valley: What Metal Fingerprints Reveal About Iron Age Trade

For the past twenty years, researchers studying the Iron Age of southern Africa have been arguing about what drove societies to become more complex: was it competition to attract Swahili traders, stockpiling masses of glittering gold and gleaming ivory in exchange for imported glass and cloth? Or was it an accumulation of wealth in cattle?

The problem is that all these arguments largely focus on goods imported from outside Africa, with hardly anyone looking at the trade that existed within the region and its peripheries. Since everything had to be moved on foot or by canoe, only lightweight, hard-to-get, valuable items were worth transporting far, things like cotton, salt, dried fish, skins, ivory, graphite, specularite, copper, tin, bronze, iron, and gold. Of these, many simply rot away with time, leaving only specularite, ivory, and metal behind.

Ivory could, in principle, be traced using isotopes and DNA, but there isn't enough of it, and no database to reference its origin exists yet. Gold is rarer than other metals, but so valuable that huge amounts of effort were put into mining it, even in tiny amounts, with pre-colonial Zimbabwe and Botswana having at least 1119 gold mines, too many to trace reliably. It's also easily melted, recycled, and reused, while iron is so common in the Earth's crust it's nigh impossible to pin down a source.

Copper, though, is a Goldilocks metal; it's rare in the Earth's crust, often turns up at sites, and there are well-established techniques to trace its source. Tin is scarcer still, with only a handful of known deposits, and can be used to pinpoint tin and bronze objects' origins, which is exactly what archaeologists Jay Stephens and his colleagues did for three ancient Botswanan sites to figure out where their copper and tin came from, and what those journeys reveal about who was trading with whom across Africa.

Three Sites at the Edge of the Kalahari

Botswana's early Iron Age began with the arrival of Bantu-speaking farmer-herders in the 6th and 7th centuries AD. In the centuries that followed, people kept goats, cattle, and sheep, grew millet, sorghum, and cowpeas, and occasionally hunted wild animals.

Among the many Iron Age sites, only three have thus far produced Lose-type pottery: Bosutswe, Khubu la Dintša, and Lose. Lose-type ceramics were a locally made elite pottery style, decorated with triangular incisions similar to those found at Mapungubwe.

Bosutswe, sitting at the head of the Motloutse River, became one of the region's main Iron Age trading hubs. Glass beads and cowrie shells indicate Indian Ocean trade networks, while chickens and black rat bones show that trade carried both deliberate and accidental passengers. Local and regional goods link the site to the Makgadikgadi Pans and Okavango Delta.

The site had humble beginnings, first established as a cattle post in the 600s before becoming a trade center by the 1100s, trading in salt, glass, specularite, and iron. In the 1300s, a large fire swept through the site, coinciding with changes in how the houses were built.

Its Lose pottery appears mainly in the elite central precinct, within the distinctive red-floored, double-walled houses, alongside other exotic goods such as cowrie shells, glass, and metal. Some researchers suggest this indicates it controlled long-distance trade, while others argue that power was more decentralized and that trade depended more on cooperation among different communities, including Khubu la Dintša and Lose.

Khubu la Dintša, 14 km away, is only a tenth the size of Bosutswe yet has yielded a higher concentration of glass beads than Bosutswe's own elite precinct, suggesting Bosutswe used its wealth and access to exotic goods to secure good grazing land for their cattle.

Each sits near rivers that drain east to the Limpopo and the Indian Ocean, making them perfectly positioned for east-to-west trade.

To try and figure out where the metal from these three sites came from, the team drew on two decades' worth of lab work. The first steps began in 2005, when Duncan Miller examined 69 copper and bronze fragments, plus one odd lump of lead and tin, from Botswana under an electron microscope, and then took 27 of those pieces s plus the lead-tin lump to the University of Arizona in 2007, where their lead isotopes were measured in 2009.

Then there was an over-10-year pause, with trace chemistry conducted only in 2021. During this time, in 2019, Stephens and co-author Dave Killick collected 27 more metal samples from the Botswana National Museum in Gaborone; this time, the metals included those sourced from Lose and Khubu la Dintša. From here, nine bronzes, a piece of 'dirty copper' and that same odd lump of lead and tin made their way to labs at Juniata and Brooklyn College for tin isotope analysis, while brand new radiocarbon dates anchored the metals in time.

So what did they find out?

Copper from Everywhere, Tin from One Valley

Earlier research had shown that copper and tin traveled vast distances in southern Africa. Some sites always had preferred copper suppliers. For example, the Tsodilo Hills (Botswana) always sourced theirs from the Central African Copperbelt, while Mapungubwe and K2 drew their copper almost exclusively from the Phalaborwa Igneous Complex.

Tin, on the other hand, always came from the same place regardless of site: the Rooiberg mines in South Africa.

But the Lose-pottery-style sites break that pattern: Bosutswe’s metal in particular came from an extraordinary array of sources hundreds of kilometers away. Some copper came from the Copperbelt and Kipushi more than 950 km away; others came from Musina (380 km), Phalaborwa (515 km), and Rooiberg (220 km) in South Africa. Lose’s copper came from the Copper Queen mine in Zimbabwe, over 500 km away, while Khubu la Dintša's copper came from an unknown deposit somewhere in South Africa.

The tin pattern is rather consistent; every bronze whose tin could be traced pointed at the same spot, the Rooiberg valley, the only tin deposit in all of southern Africa with evidence of pre-colonial mining.

That mismatch is the key. If tin arrives from one place and copper from everywhere else, someone at the receiving end is doing the mixing. In this case, someone at or near Bosutswe.

Almost all of the bronze took the form of personal adornment: bangles, beads, helices and clips. Bronze is often known for its shiny gold-like appearance, yet it doesn't seem like the people here were trying to mimic gold. The tin content is too variable, ranging from 2 to 10 percent, resulting in bronze colors that span from near copper-red to salmon pink and warm yellow, with almost none reaching the golden yellow that would pass for gold.

The authors suggest these different colors weren't the result of someone botching their attempt to make 'fake' gold, but rather extended the color palette of jewelry, allowing the elite of Botswana to mix and match.

But another mystery arose when the chemical fingerprints began pointing seemingly everywhere except Botswana.

The Mines Outside the Doorstep

Botswana had many pre-colonial copper mines, several of them practically on the doorsteps of Lose, Khubu, and Bosutswe. Yet of the items the team tested, only a handful could be traced to two Botswana copper mines, Selebi Phikwe and Dukwe, 150 km and 140 km from Bosutswe, respectively.

If Botswana had so many pre-colonial copper mines, why did it seem like they were being ignored in favor of those over 200, 500, or even 950 km away?

That's where those radiocarbon dates come in handy. Before 1300, Botswanan copper is almost entirely absent from the sites, with Selebi Phikwe the only Botswanan source represented at all, while Dukwe copper only begins to appear after 1300. Meanwhile, mines like Thakadu and Dihudi, which are just as close as Selebi Phikwe, don't show up at all.

The likeliest explanation isn't that traders or miners were avoiding these mines but rather that they simply weren't open yet. For Dukwe, this was definitely the case, as independent excavations have revealed it only opened after 1300, matching the metal evidence.

You simply can't source metals from mines that don't yet exist. If that is the explanation, it also makes this study an indirect means of dating, by working out when objects from them first began to appear.

For a long time, Bosutswe has been treated as a smaller site orbiting the much bigger and frankly more well-known polity to the south, Mapungubwe. Yet Mapungubwe and K2 sourced their copper almost exclusively from a single deposit, Phalaborwa, of which there is only one certain piece at Bosutswe. Both Mapungubwe and Bosutswe sourced their tin from Rooiberg. Yet, their copper networks don't overlap, suggesting each traded with Rooiberg independently, neither truly controlling it, and each making their own bronze as a result. And because Bosutswe was already receiving Rooiberg copper before anyone mined tin there, its existing connections would have carried tin north almost as soon as production began.

Also, while Mapungubwe flourished for less than a century, Bosutswe flourished for centuries, perhaps due to its decentralized nature and ability to draw on multiple copper sources.

While for 20 years people have been focusing on what's been coming in from the outside, they have overlooked an enormous internal trade in copper and tin that ran across the southern African interior for centuries.

The three sites in eastern Botswana weren't end-of-the-line backwaters but trading hubs pulling in metal from across southern Africa. Copper came from deposits hundreds of kilometers away, while tin, regardless of where one looked, always came from Rooiberg, and both seem to have been mixed into an array of differently colored bronzes upon arrival, adorning the elite of Bosutswe.

Whatever pushed these Iron Age societies towards complexity, we now know it wasn’t only what arrived from outside Africa or from the coast that mattered.

What do you think was traveling north in exchange for all that copper? Salt and ivory are the best guesses, but could it have been something else?

Let me know what you think, and if you'd like to support this blog and the person writing it, you can chip in once, monthly, or yearly. Only if you'd like to.

Reference

Stephens, J., Klehm, C., Segadika, P., & Killick, D. (2026). Isotopic and chemical evidence for long-distance trade of copper and tin into east-central Botswana, c. cal. AD 800–1650. Azania: Archaeological Research in Africa, 1–32. https://doi.org/10.1080/0067270X.2026.2715248