Mapungubwe

Mapungubwe is the name of a sandstone hill, of the kingdom whose rulers lived on it, and of the wider landscape around it. The hill rises about 30 metres above the surrounding plain and is roughly 300 metres long, with sheer sides and a flat summit. It stands a short distance south of the place where the Shashe river flows into the Limpopo, in what is now the far north of South Africa, close to the borders of Botswana and Zimbabwe.

Between about AD 900 and 1300 this valley held a series of large settlements. Archaeologists usually describe three capitals in turn: Schroda, then a large town called K2 (also known as Bambandyanalo), and finally the settlement around Mapungubwe Hill itself. In the last phase, from about 1220, the ruling family lived on top of the hill while most people lived below. Because of this, Mapungubwe is often described as the earliest known class-based society and kingdom in southern Africa, although, as later sections show, both claims are debated.

The name is usually explained as meaning “place of jackals” or “hill of the jackal”, from a word for jackal found in several Bantu languages of the region. We do not know what the people of the kingdom called themselves or their capital, because they left no written records. Everything we know comes from archaeology and from comparison with later societies of the region.

In July 2003 UNESCO added the Mapungubwe Cultural Landscape to the World Heritage List. Its value lies in the whole sequence: the hilltop capital, the villages around it and the earlier centres of Schroda and K2 survive with little modern disturbance, so the growth and decline of a kingdom can be followed over some four centuries.

The best way to understand Mapungubwe is to imagine walking across it. Start at K2, about a kilometre south-west of the hill. Today it looks like a low mound, but it is made up of an enormous midden, a dump of ash, broken pottery, animal bones and household rubbish that built up over two centuries at the heart of the town. Archaeologists read these layers like the pages of a book to see how life changed.

Then look at Mapungubwe Hill. The ways up are steep and narrow, and in one cleft the first modern visitors found steps that had been deliberately cut into the rock. On the summit, the hill’s residents carried up large amounts of soil to create level ground for houses. This was where the royal family lived, and the cemetery where gold-rich graves were found lies on the summit too.

At the foot of the hill lay the homes of ordinary people. Stone walls marked off important areas and helped separate the ruler from everyone else. Further out, smaller villages dotted the valley. Some hilltops in the region also contain deposits linked to rainmaking ceremonies, a reminder that rain was the most precious resource in this hot, dry land.

Finally, look at the setting itself. The river floodplains could be farmed when rains were good, and the area was rich in elephants, whose tusks were valuable trade goods. UNESCO treats the whole valley, ruins and land together, as a single cultural landscape.

Most people at Mapungubwe were farmers. They grew grain crops such as sorghum and millet on the river floodplains and stored the harvest in raised granaries. They also kept cattle, sheep and goats. Cattle were more than food: they were a form of wealth, and herds were sometimes grazed on the land of other communities, which helped to build ties between them.

The people were skilled metalworkers. They smelted and forged iron for hoes, axes and spears, and worked copper and gold into wire, beads and bangles. Scientific dating of fibres wound inside gold anklets from the hilltop graves shows that these objects were made while the hill was occupied, by local craftspeople with a mature metalworking tradition, not by visitors from elsewhere.

Before farmers arrived, hunter-gatherers had lived in the valley for thousands of years and left rock paintings in shelters nearby. Older accounts, including Carruthers’s, describe these foragers as gradually pushed aside as farming villages multiplied. More recent research by Tim Forssman paints a more varied picture: at some rock shelters foragers kept their Stone Age way of life while trading with farmers, producing crafts such as beads and even gaining wealth and local status. Foragers were present throughout the rise of the kingdom, though their place in it changed over time.

Around AD 900 people whose pottery archaeologists call Zhizo settled in the Shashe–Limpopo valley, apparently drawn by elephant ivory for trade. Their main settlement, at Schroda, may have been home to a few hundred people. In the early eleventh century a different group, associated with the Leopard’s Kopje pottery tradition and with roots further north on the Zimbabwe plateau, founded K2. Most researchers see K2 as a new community rather than a continuation of Schroda, and date it to about 1000 or 1030 until 1220.

K2 grew into a large town of perhaps a thousand or more people. Its leaders became wealthy through cattle, through crafts such as ivory working and through their link to long-distance trade. Yet ruler and subjects still lived together in the same settlement, with the chief’s household simply the largest.

Around 1220 this changed. The elite moved to the top of Mapungubwe Hill, while commoners settled below and in the surrounding villages. Stone walls separated important areas, and the ruler was secluded from ordinary people. The archaeologist Thomas Huffman, whose model has long shaped the field, argued that this move marked the moment when class distinction and “sacred leadership” crystallised: the ruler was no longer simply the richest member of the community but a figure of ritual power, associated with rainmaking, who lived above and apart from his subjects.

For Huffman and many others, this physical separation of rulers and ruled was the key sign of an early state, and the pattern was later repeated, on a larger scale, at Great Zimbabwe and Khami. A study that dated the hilltop gold found that its use also changed at this time: gold that had once been mainly an export to the coast became, on the hill, a symbol of power and status. Not all archaeologists accept that Mapungubwe was the first or only place where these changes happened, a debate discussed below.

Mapungubwe’s location gave it an advantage. Paths linking the Zimbabwe plateau with the lands south of the Limpopo crossed the valley, and the river corridor led eastwards towards the coast. Two goods made the kingdom valuable to the outside world: elephant ivory, which had first attracted settlers to the valley, and gold. Neither was easy to obtain elsewhere along the coast, and both were prized in Arabia, India and China. They were carried, probably through a chain of intermediaries, to ports on the coast of what is now Mozambique: an early link was with Chibuene, and later the Sofala region became the main outlet.

From there the trade joined the network of the Swahili coast. Swahili merchants and visiting traders from Arabia, the Persian Gulf and India shipped African gold, ivory and other goods north and east with the monsoon winds. In the late twelfth and thirteenth centuries the island city of Kilwa won control of the Sofala trade, and its sultans grew rich by taxing gold passing through their port. How much of Mapungubwe’s gold reached Kilwa itself cannot be measured, but the kingdom was part of the same chain that linked the southern African interior to the wider ocean.

What came back is visible in the ground. Many thousands of imported glass beads have been excavated at K2 and Mapungubwe; one hilltop grave alone contained tens of thousands. Archaeologists group them into series of different shape and composition, made at workshops around the Indian Ocean, with South and Southeast Asia among the likely sources. Their changing styles help archaeologists date layers and trace shifts in trade partners. Beads were not mere trinkets: they were valued for display, exchange and, in the case of the burials, for honouring the dead.

A handful of fragments of Chinese celadon, a green-glazed stoneware, were also found on the hilltop in 1934, and another piece of the same vessel turned up in 1991. They were first identified as Longquan ware of the late Southern Song dynasty (1127–1279), but a 2005 analysis of the glaze suggested they might be somewhat later, from the Yuan or even early Ming period. Either way, a vessel made in south-eastern China ended up on a hill above the Limpopo, a vivid sign of how far the kingdom’s connections reached, even if it travelled through many hands.

On the summit of the hill was a small cemetery. The excavations of the 1930s recorded more than twenty burials, of which three were exceptionally rich in gold. In one of them, now called the Original Gold Burial, lay the famous gold rhinoceros together with other gold objects: fragments of a bowl, beads, bands of gold wire and many pieces of decorated gold sheet. Another rich grave held a person, probably a woman, buried with more than a hundred gold bangles and thousands of gold and glass beads.

The rhinoceros was made by shaping thin gold foil over a carved wooden core and fixing it in place with tiny gold tacks; the head and horn were made from separate pieces. The wood has long since decayed. When it was first dug up, the foil was crushed and broken, and the pieces were divided among the members of the party who found them. The University of Pretoria later bought them, and early attempts were made to piece the figure together. In 1999 a conservator from the British Museum carried out the restoration that is displayed today.

What the rhinoceros meant is uncertain. The rhinoceros was a powerful animal, and among later Shona and Venda societies certain animals served as emblems of clans and lineages. It is often presented as a symbol of royal power. The archaeologist Alex Duffey noticed that a gold rhinoceros, a bovine and a feline from the same grave are of similar size, with tack holes in their feet, and argued that they were once attached to an elaborately carved wooden divining bowl, comparable to later Venda divining bowls.

Since the end of apartheid the gold rhinoceros has become a national symbol. It appears at the centre of the badge of the Order of Mapungubwe, a South African national honour created in 2002, and in 2016–17 it was shown at the British Museum in London. The objects themselves remain in the University of Pretoria’s Mapungubwe Collection.

In the conventional chronology, the capital on Mapungubwe Hill was abandoned around 1290 and its people dispersed. A set of new radiocarbon dates published by Huffman and Woodborne in 2021 suggests that people using Mapungubwe-style pottery remained in the valley until about 1320, and Chirikure has also argued for occupation after 1290. Over the following decades Great Zimbabwe, more than 250 kilometres to the north-east on the Zimbabwe plateau, became the leading centre of the region, with even larger stone-walled enclosures and a strongly marked separation between elite and commoners. Later still, power passed to Khami in the west.

Why Mapungubwe was abandoned is one of the central questions about the site. UNESCO’s statement of the site’s value names climate as the cause: in its account, farming in this marginal valley depended on favourable rainfall, and a drier period around 1300 forced the population to scatter. Other suggested factors include soil exhaustion and overgrazing, shifts in the coastal trade towards ports and routes that favoured the north, and a loss of faith in rulers who could no longer ensure rain.

Huffman and the environmental scientist Stephan Woodborne, combining archaeological evidence of rainmaking and burnt grain bins with a 1,000-year rainfall record from baobab trees, concluded that a drought did occur in the early fourteenth century, but that the abandonment “owed as much to cultural factors as to environmental pressures”. Their 2021 dates place the end of the Mapungubwe phase close to a severe drought dated by the baobab record to about 1310. Other researchers question whether climate played a decisive role at all, as the section on different perspectives explains.

Mapungubwe’s people did not vanish. Their descendants mixed with neighbouring communities, and the cultural pattern of the kingdom lived on in the states of the Zimbabwe plateau. Several present-day groups, including the Venda and the Kalanga (a people related to the Shona), have claimed ancestral links to Mapungubwe, though historians caution that the evidence for direct descent of any single group is limited.

Mapungubwe was not isolated. Its gold and ivory entered a trading system that stretched across the Indian Ocean, and the best way to place it in world history is to follow those goods. On the Swahili coast, towns such as Kilwa were growing rich as middlemen. Kilwa’s seizure of Sofala and its gold trade, which the Kilwa Chronicle places in the late twelfth century, happened while K2 was at its height; the city’s fourteenth-century heyday, when Ibn Battuta visited in 1331, came after Mapungubwe had faded and Great Zimbabwe had taken its place as the main inland source of gold.

At the other end of the ocean, China under the Southern Song dynasty (1127–1279) depended heavily on maritime trade, and Quanzhou in Fujian was one of its great ports. Around 1225 its superintendent of maritime trade, Zhao Rugua, compiled the Zhufan zhi, a handbook on foreign lands based on interviews with merchants. In his entry on Cengba, a name for part of the East African coast, he lists among its products elephant tusks, “native gold”, ambergris and yellow sandalwood, and says that ships from Gujarat and Arab ports came every year with white cotton cloth, porcelain and copper. Nothing in his text refers to Mapungubwe, but it shows that by the early thirteenth century Chinese officials knew East Africa as a source of exactly the goods Mapungubwe exported.

Zhao’s information came second-hand, from foreign merchants, and goods crossed the ocean in stages; the exact routes by which particular objects reached the Limpopo are not known. The glass beads at K2 and Mapungubwe, many of them probably made in South Asia, and the Chinese celadon on the hilltop are physical traces of these exchanges, although the celadon may date from the Yuan period or later rather than the Song. Seen this way, the kingdom on the Limpopo, the Swahili coast, the ports of western India and China were connected parts of a single Indian Ocean world in the thirteenth century.

Local communities had long known of the hill and regarded it as a sacred, even forbidden, place. In December 1932 a farmer’s son, J. C. O. van Graan, a student at the University of Pretoria, heard about the “sacred hill”. Around New Year 1932–33 he returned with his father and three other men. They pressured a local man, recorded only as Mowena, into showing them the hidden path to the summit, which he did reluctantly and in fear. At the top they found gold beads, bangles and pieces of gold sheet, and on 2 January 1933 they dug into the grave that held the rhinoceros.

Van Graan reported the find to his history professor, Leo Fouché. Fouché arranged for the finders to be paid for the gold, which went to the University of Pretoria, and for the government to buy the farm, Greefswald. The discovery was announced internationally in the Illustrated London News on 8 April 1933, and in the same month the first archaeological team arrived. The excavations of 1933–1940, later led by Neville Jones and Guy Gardner, recovered vast amounts of material, but by modern standards they were poorly recorded, and UNESCO notes that valuable evidence was lost. Fouché’s report of 1937 was titled Mapungubwe: Ancient Bantu Civilization on the Limpopo.

The political setting soon became hostile. In the 1930s the University of Pretoria came under the influence of Afrikaner nationalism; Fouché, regarded as too liberal, left for another university. Settler history taught that Bantu-speaking Africans had arrived in the region at about the same time as Europeans, the so-called “myth of the empty land”. A wealthy African kingdom on the Limpopo several centuries earlier contradicted this. Some physical anthropologists also produced racial interpretations of the skeletons that played down their link to living African populations; later studies, from the 1970s onwards, rejected these conclusions.

In 1947 the government of Jan Smuts created the Dongola Wild Life Sanctuary, a national park that included Mapungubwe and was meant to become part of a cross-border reserve. After the National Party won the 1948 election, the park was abolished. For the following decades the site received little public attention, although research continued intermittently. In the 1980s K2 and Mapungubwe Hill were declared national monuments, but the story of the kingdom entered school textbooks and national symbols only after the end of apartheid in 1994.

After 1994, Mapungubwe was given a new role as a symbol of African achievement and of the “African Renaissance” promoted by the post-apartheid government. In 2003 UNESCO inscribed the Mapungubwe Cultural Landscape on the World Heritage List, the first cultural landscape in South Africa to be listed. The property covers about 28,000 hectares, with a much larger buffer zone, and is managed by South African National Parks (SANParks).

Human remains excavated at Mapungubwe Hill, K2 and Schroda over several decades had been kept in museum and university collections, above all at the University of Pretoria. After claims by communities who regard the people of Mapungubwe as their ancestors, a steering committee of institutions, claimants and government bodies arranged their return. Most remains were first documented with standard anthropological methods, but no samples for destructive analysis such as DNA were allowed. The remains of 143 individuals, each sealed in its own container, were reburied at the principal sites; on Mapungubwe Hill this took place on 20 November 2007, accompanied by rites performed by claimants and traditional healers. According to the University of Pretoria team, a planned official ceremony was cancelled amid political disputes. The gold and other grave goods remain in the university’s museum.

The surrounding protected area, managed by SANParks since the late 1990s, was first known as Vhembe-Dongola National Park; it was renamed Mapungubwe National Park in 2004 and officially opened on Heritage Day, 24 September of that year. It was the first South African national park created mainly for its cultural heritage. Elephants and other large animals have returned to the landscape alongside the archaeological sites.

Because the kingdom’s territory once stretched across what are now three countries, South Africa, Botswana and Zimbabwe signed a memorandum of understanding in 2006 to create a transfrontier conservation area around the confluence; in 2009 it was renamed the Greater Mapungubwe Transfrontier Conservation Area. Pressures remain, including irrigated farmland and mining nearby; UNESCO has expressed concern about a coal mining licence granted just outside the property.

Was climate to blame? The view that a drier climate around 1300 made the valley unable to support a large farming population is written into UNESCO’s statement of the site’s value and was long the standard explanation, associated especially with Huffman’s work on the onset of the Little Ice Age. Huffman and Woodborne’s baobab studies and new radiocarbon dates link the final abandonment to an early fourteenth-century drought, while acknowledging cultural factors as well. Critics, including Shadreck Chirikure and colleagues, argue that this reasoning is too environmentally deterministic: communities upstream and downstream, in similar settings, persisted through the same climatic conditions, so drought alone cannot explain why Mapungubwe’s capital was abandoned. The debate is unresolved.

Was Mapungubwe a “state”, and was it the first? For Huffman, the move of the elite to the hilltop around 1220 marked the crystallisation of class distinction and sacred leadership, the hallmarks of what archaeologists call the Zimbabwe culture, and thus the birth of the first state in southern Africa, from which Great Zimbabwe descended. In 2014 Chirikure and colleagues published dates from Mapela Hill in south-western Zimbabwe, about 90 kilometres to the north-west, showing massive stone-walled terraces and elite floors from the eleventh century, earlier than Mapungubwe’s hilltop phase. They concluded that Mapungubwe can no longer be regarded as the sole cradle of the Zimbabwe culture and proposed a model in which several centres developed in parallel. Others continue to defend Mapungubwe’s central role. The word “kingdom” is itself a comparison with European institutions; the historian Jane Carruthers places it in quotation marks for this reason.

How have heritage narratives changed? The story told about Mapungubwe has shifted with politics. Early excavators accepted that it was an African site, but racialised theories about its people, the Afrikaner nationalist “empty land” narrative and official neglect kept it out of the public story for decades. In 1999 Mamphela Ramphele claimed that the gold rhinoceros had been “carefully hidden away”; the University of Pretoria rejected this, pointing to decades of press coverage. Carruthers, for her part, concludes that the university did little to bring the site into national consciousness. On this reading the finds were not secret, but they were marginalised. After 1994 Mapungubwe became a national emblem, and some scholars, such as Carruthers, warn that this new story can also be simplified, for example by celebrating a society built on inequality and on the displacement of foragers, or by presenting a single group as its only heirs.

Because Mapungubwe left no written records, its history rests on archaeology and science. The basic sequence of Schroda, K2 and Mapungubwe Hill comes from pottery styles, stratified middens and radiocarbon dates, refined over decades. Radiocarbon results must be calibrated, and wiggles in the calibration curve can give several possible calendar ranges for a single sample, so archaeologists rely on stratigraphy and regional models to choose between them. This is why the commonly cited dates (K2 about 1000/1030–1220; the hilltop capital about 1220–1290) differ from the revised sequence Huffman and Woodborne proposed in 2021, which places the end of the Mapungubwe phase at about 1320.

Imported objects provide cross-checks. Glass beads can be grouped by shape, colour and chemistry, and the appearance of new bead series in the deposits is used to date layers and to trace shifts in trading partners. But imports can also complicate the picture: the 2005 analysis suggested the celadon might be Yuan (1279–1368) or even early Ming. The Yuan period overlaps the conventional occupation dates, but if the vessel were shown to be a fourteenth-century product, it would imply that the hill was still in use later than usually assumed. The researchers who proposed the re-dating called for more radiocarbon dates from the hill.

The gold presents its own problem. Metal cannot be radiocarbon dated, but in 2009 Woodborne, Pienaar and Tiley-Nel dated organic fibres preserved inside coiled gold anklets and showed that they were contemporary with the hilltop occupation. For the rhinoceros and the other foil objects, which were dug out hastily in 1933 without proper records, no such date is possible.

Finally, interpretation depends on models drawn from later ethnography. Huffman’s reading of sacred leadership and rainmaking hills relies heavily on comparisons with later Shona, Venda and other southern African societies; critics warn against projecting recent practices back eight centuries without independent evidence. The reburial of 2007 also limits future work: the remains were sealed in containers intended to preserve them, but destructive analyses were not permitted beforehand, and any new study would require fresh agreement.