A garbage site close to K2, where commoners lived, indicates that rich and poor ate very different foods. Funeral traditions were also different. The rich had a graveyard at the top of the hill with a beautiful view of the region. They were also buried with gold and copper ornaments and glass beads, showing the people of Mapungubwe were skilled in working with gold. Ivory was traded with Arab merchants and contributed greatly to the wealth of the kingdom.
It is difficult to find a single explanation for the desertion of Mapungubwe. Some archaeologists feel that the kingdom began to decline in the 's because the climate changed.
The weather became colder and drier and reduced the grazing land making cattle farming difficult. Others think there was a change in trade routes. Mapungubwe relied on trade and any blow to this activity would have forced people to move away. The people of Mapungubwe were wealthy and farmed with cattle, sheep and goats, and also kept dogs. They produced large harvests that allowed them to trade and store extra food. Archaeologists found traces of millet, sorghum and cotton in the remains of storage huts.
Riches also came from ivory, gold and the rich farmland caused by the flooding of the area. From about to Mapungubwe was an advanced trading centre and its inhabitants traded with Arabia, China and India through the East African harbours.
Farm animals supplied meat and hides, but they also hunted, snared and gathered other food. The city could trade because it was so close to the Limpopo River, which connected it with the coast. They exchanged salt, cattle, fish, gold and iron, ivory, wood, freshwater snail and mussel shells, chert and ostrich eggshell beads were used for glass beads and cloth.
The first settlers of Mapungubwe were early iron age settlers. They lived there from about ad tonad, and around iron age subsistence farming also settled there. Their existence is confirmed by the discovery by archeologist of a few potsherd identified as early iron age pottery. Mapungubwe was strucutured along social classes. The elite lived at the top of mapungbwe and their followers stayed at the bottom of the hill and in the surrounding area.
A garbage site close to k2, where commoners lived, indicates that rich and poor at very different food. They were buried with gold and copper ornaments and glass beads, showing the people of mapungubwe were skilled in working with golf. Unit 2 focus on Mapungubwe as the first state in Southern Africa, this is after Mapungubwe had discovered itself as a kingdom. The civilization that was taking place in the area, the opportunities, the rule of law, the bureaucracy was visible in the area.
There was royalty with kings and queens ruling over the masses. There was also the working class and the upper class. The upper class will be the individuals that were allowed to live on top of the hills because of their social and economic status. The working class generally occupied the lower parts of the hills.
The king and his advisors were the decision making body in the state. This unit deals with the royal power and other objects that symbolized power and political leadership. Now Mapungubwe had discovered itself as a function state that has a population, economy and law.
The rise of the area from a small kingdom into a greater power in Southern Africa was visible. Referencing a combination of radiocarbon dating, Raman spectrometric studies of cultural artefacts and Bayesian modelling, he concludes that Mapungubwe might well have survived for centuries after AD Further, he suggests that it and its cultural trademarks bears a more striking cultural resemblance to other ancient communities in the adjacent regions of Botswana and south-western Zimbabwe.
New dating evidence indicates that a number of related communities in southern Africa - which are contemporary with the foundation of Mapungubwe and Great Zimbabwe - were trading with the Indian Ocean coast, were experimenting with stone wall construction, and kept large herds of cattle, he adds. Our interpretations are conditioned by the fact that some of sites were looted, making comparisons difficult.
But there is hope for the future, a hope that is created by ever improving techniques of study. Please view the republishing articles page for more information. By embedding this news article on your site you are agreeing to the University of Cape Town's terms of use.
Further research by archaeologists such as Robinson [13] , [14] , [15] and Garlake [16] , [17] among others, led to the development of a credible culture history of farming communities in southern Zambezia.
The Leopard's Kopje culture CE— was named after the site of Leopard's Kopje approximately two kilometres north-east of the Khami World Heritage Site, and about 24 kilometres west of the modern city of Bulawayo. This culture was widely distributed from west-central Zimbabwe to north-eastern Botswana, and adjacent areas of northern South Africa. It was evident in the sequence of Great Zimbabwe and appeared at many other sites, such as Taba Zika Mambo [14]. Owing to a lack of imposing monumental architecture and a 15 th century CE date, Mapungubwe was considered an extension of Great Zimbabwe's influence [6].
Garlake [17] carried out perfunctory excavations on top of the heavily terraced Leopard's Kopje site of Mapela Hill, about 90 kilometres north-west of Mapungubwe in south-western Zimbabwe. On the basis of solid dhaka houses, abundant glass beads and evidence of social distinction, Garlake concluded that it was a capital of an independent Leopard's Kopje state. From the s until the late s, attempts to understand the evolution of socio-political complexity were, however, Great Zimbabwe-centric in nature.
Based on the available chronological and spatial data, it was assumed and widely accepted that Great Zimbabwe was the cradle of the Zimbabwe culture [6].
Although archaeologists at the time recognised the broad similarities between glass beads and ceramics from Leopard's Kopje sites and those of Periods II and III at Great Zimbabwe, little attention was paid to the possibility that the Zimbabwe culture evolved out of the Leopards' Kopje and was therefore more important than was acknowledged at the time.
Great Zimbabwe was considered to be the capital of a very extensive empire stretching from the Indian Ocean to the Kalahari. The hundreds of differently-sized dry stone-walled settlements were accorded the status of provincial, district and ward centres in the super state. As a result, an opportunity to unravel the contribution of the Leopard's Kopje in the development of the Zimbabwe culture was missed. This was achieved by Huffman [7] who coupled cognitive structuralist approaches with archaeological data and radiocarbon dates to argue that spatially, Mapungubwe CE— exhibited the Zimbabwe culture pattern earlier than Great Zimbabwe CE— Based on cultural precedence, Mapungubwe became the origin of the Zimbabwe culture.
Because Mapungubwe was thought to be the only Leopard's Kopje site with evidence of class distinction, it was seen as the capital of southern Africa's first state which flourished until CE [7] , [21]. Southern African archaeologists, like their international counterparts, widely accept this framework [22] , [8] , [23]. This entrenched position is, however, problematic because its linearity has caused more important sites to be assumed to be the capitals of the major phases of Leopard's Kopje Mapungubwe , Great Zimbabwe Zimbabwe culture and Khami Zimbabwe culture.
As such, archaeologically well-explored places became the theatres of innovation, thereby editing out of historical significance the many sites that chronologically overlap with the so-called capitals. In fact, sites such as Mapela Figure 1 were, without much research, granted the status of provincial centres under Mapungubwe see [24] , [8]. To establish the basis for a polycentric model, Chirikure et al.
This wider expression of the Zimbabwe culture challenged the assumption that in an approximately one million square kilometre-large southern Zambezia, K2 and Mapungubwe were sole propagators of socio-political complexity.
However, although Chirikure et al. As such, supporters and sceptics demanded the substantiation of the conclusions with the aid of carefully excavated sites. This provided the motivation for this work — it presents the outcome of detailed research carried out at Mapela in the Shashe region of south-western Zimbabwe. Mapela was naturally attractive because it contains substantial terrace walls, abundant local pottery, solid dhaka floors, and glass and shell beads on the surface — attributes which collectively define the Zimbabwe culture.
Geographically, Mapela possesses all the advantages which K2 and Mapungubwe have, and perhaps a lot more. It is situated in rich elephant hunting country close to perennial water sources and is within 20 kilometres of the strategic Gwanda-West-Nicholson gold belt. The fieldwork mapped and excavated the site to establish its spatial extent and the density of stone walling, as well as collecting samples for dating.
The results indicated that Mapela is easily the largest known Leopard's Kopje site in southern Africa. The Bayesian chronology, when combined with local pottery and glass bead typologies, suggests that Mapela flourished between CE and CE and therefore pre- and post-dates the generally accepted dates for the flourishing of Mapungubwe CE— Furthermore, it contained prestige stone walls and evidence of class distinction, which suggest that the Zimbabwe culture was expressed earlier at Mapela than at K2 and Mapungubwe.
Finally, a polycentric model informed by Actor Network theory, historical data, material culture patterning and modelled Bayesian dates was developed.
Mapela Figure 2 is the name of a prominent hill which lies two kilometres due east of the confluence of the Shashe and Shashani rivers in the south-western Zimbabwe lowveld [17]. As the crow flies, it is approximately two kilometres due north of the Zimbabwe-Botswana border. Mapela is, however, notoriously difficult to access because it lies beyond the major road networks in the area.
Because of this inaccessibility, it is not surprising that Garlake [17] is the only known archaeologist to have studied the site before our visit. All interpretations of the site, past and present, are rooted in Garlake's observations. Mapela Hill is characterised by a flat top, with steep cliffs terminating in ledges at different elevations Figure 2. The resulting flat areas were terraced following the contours of the hill from the bottom to the top, to create spaces for homesteads, some of which were associated with clearly-defined kraals.
Amazingly, Mapela is heavily terraced on all sides with revetment walls which often reach up to two and half metres in height. To establish the cultural affiliation of the site, Garlake [17] test excavated the hilltop area and a midden abutting the north-facing cliff floor.
The excavations uncovered typical Leopard's Kopje ceramics, glass beads, fauna, iron work and few copper-based objects. Other finds include spindle whorls and pellets of slag. The hilltop stratigraphic sequence consisted of alternating layers of dhaka , or earthen floors, with curved kerbs typical of Great Zimbabwe and related sites. Garlake [17] : 24 concluded that the glass beads from Mapela were more abundant than any other Leopard's Kopje site in south-western Zimbabwe.
Garlake further observed that although stone walling is a characteristic feature of Leopard's Kopje sites, the size and extent of the terraces of Mapela was not approached elsewhere in the region.
Therefore, the stone walls of Mapela, according to Garlake, were only dwarfed by those of later sites such as Great Zimbabwe, Khami and others, typical of the flowering of the Zimbabwe culture. Two radiocarbon dates, SR and SR, were obtained from charcoal. However, because of the high error term, the dates are uncertain and when calibrated at 2 sigma, yield a date of between CE and Nevertheless, they indicate that Mapela was occupied during the Leopard's Kopje period. It is this conclusion which persuaded Africanist archaeologists, despite their never having visited the site, to believe that Mapela was smaller than Mapungubwe.
Unfortunately, Garlake did not include the massive terraces on different contours of the hill below the summit. Despite this omission, Garlake concluded that Mapela was a centre of political and economic influence over some considerable area. Furthermore, he noted the presence of class distinction and surmised that further excavation of Mapela, on a more substantial scale, had the potential to fundamentally change perspectives on the development of socio-political complexity in the region.
This research was carried out in Zimbabwe. It is an offence to alter, disturb or destroy an archaeological site without written permission from the Executive Director of National Museums and Monuments of Zimbabwe, in terms of the National Museums and Monuments Act No. This research was carried out under Permit no. The permit allowed excavation and the export of charcoal samples for radiocarbon dating. No destructive analyses were performed on any other objects which are archived in Harare. Detailed fieldwork was carried out between May and October in a stepped approach that began with pedestrian surveys, followed by a combination of field and desktop mapping, and ending with stratigraphic excavations.
A dedicated pedestrian survey was conducted across the entire site. In the process, extensive terraces Figure 4 , vitrified dung and middens with a mix of fauna, pottery, metal objects and occasional glass and shell beads were recorded on the flats, terraces and hilltop.
On both the terraces and hilltop, erosion often exposed a succession of fired dhaka or Zimbabwe cement floors Figure 5. Surprisingly, for a stone-walled Leopard's Kopje site, vitrified dung was recorded on numerous terraces and flats pointing to the presence of kraals on and below the hill. It has been argued that with the advent of the Zimbabwe culture, cattle kraals were sited away from residential areas 7 but Mapela clearly contradicts such thinking.
Surface pottery comprised a mixture of K2 Figure 6 and Mapungubwe ceramics, with occasional Zhizo sherds. On the western side, the main hill narrows into a smaller kopje which is conjoined to the site of Little Mapela.
When these observations are reconciled with Garlake's 17 report, it becomes immediately obvious that only the summit of the main hill was mapped, thereby excluding approximately three quarters of the site.
Unfortunately, this underestimation of Mapela's size has given it a subordinate role to Mapungubwe, when the opposite is true if settlement size is the only variable that determines political importance. These far-reaching observations demonstrated that it was impossible to understand the complexity of the site without mapping it in full. Owing to advances in satellite imaging and GIS software, it was possible to use a combination of desk and field based mapping techniques.
Sadr and Rodier [29] demonstrated the utility of this approach when they mapped stone-walled sites around Gauteng in South Africa.
To begin with, Garlake's map of the summit was superimposed on a Google Earth image of Mapela Hill to establish control points. Once the best-fitting overlay was established, the terraces and other prominent features were screen digitised using GIS software.
In cases where opacity was poor, field walking and geocoded GPS recording enabled the capturing of accurate details. Such an endeavour, for the first time, finally produced a complete map of Mapela Hill Figure 7. As is clear from the difference between the ground and the summit elevation, Mapela Hill is just over 90 metres high.
Apparently, most contours of the hill were heavily terraced indicating that the site was intensively occupied. Furthermore, if the map of the summit is considered in relation to the whole site, it becomes decisively clear that Garlake's descriptions — and all subsequent interpretations based on it — severely underestimated the size of the site.
In order to understand the chronology and material culture of Mapela, stratigraphic excavations were conducted at three areas: 1 Excavation Area 1 on the flats marked I on Figure 7 ; 2 Terrace Excavation Trench 1 on a substantial north-facing terrace marked H on Figure 7 ; and 3 Lower Summit Excavation Trench 1 on the lower summit marked G on Figure 7.
The intention was to develop an impression of the chronology and activities taking place in different areas of the site — the flats, the terraces and the summit. Such a strategy allowed us to explore the issue of class distinction at the site.
Terrace Excavation Trench 1 was located on the eastern edge of a sloping terrace covered by vitrified dung Figure 4. The lower summit trenches were sited on a midden purposely selected to provide a comparison with Garlake's stratigraphy on the upper platform. The excavations proceeded in 10 centimetre spits. As Figure 8 shows, the stratigraphy of Excavation Area 1 was not more than 50 cm deep and was not clearly defined except in a few cases with a lens of ash.
The finds included slag, K2 and Zhizo ceramics, together with a small number of shell beads. Terrace Excavation Trench 1 yielded an authoritatively informative stratigraphy Figure 9. A very thin layer of vitrified dung constituted the topmost layer and was followed by two successive middens, one brownish and another greyish in colour. Underneath this was a layer of vitrified dung, followed by another dhaka floor which rested on top of a brown fill with midden debris.
The fired nature of the floor ruled out other alternatives, such as caps which may result from repeated cattle or animal hoof stamping see Huffman, [24]. Below the fill was a thin layer of burnt grass, carbonised sorghum seeds, charcoal and dhaka in level A very thick dhaka floor followed underneath level It was succeeded by yet another event associated with burning. A mixture of wood charcoal, grass and sorghum seeds was recovered in this level The recovery of sorghum seeds suggests that the circular stone feature that continued from the layer above, on the northern side of the trench, was a grain bin foundation.
Below this was another floor underlaid by a layer comprising of broken dhaka fragments, pottery and charcoal. This was followed by a midden which accumulated on top of a floor, and covered one side of the trench.
The southern corner of the trench contained charcoal, ash and burnt dhaka in level It seems that this material collected on top of a floor levels 17 and 18 which sealed a very thin midden that contained K2 pottery, charcoal and three glass beads levels 18 and Because of the depth and integrity of this stratigraphy, nine samples of carbonaceous materials from levels 7, 9, and 13 to 19 were submitted for radiocarbon dating.
Lower Summit Trench 1 was just over half a metre deep. However, the stratigraphy was complex, comprising alternating layers of midden and floors. A decision was made to extend the trench by one metre Extension A and to excavate that to bedrock Figure Only four more layers were encountered before reaching bedrock. A floor lay directly on top of this bedrock, where a decorated K2 sherd was found.
When combined, the excavations produced a number of significant finds ranging from domestic pottery, spindle whorls, ferrous metal objects, copper-based ornaments, glass beads, fauna, fish bone, carbonised seeds to fragments of dhaka.
Only domestic pottery and glass beads were typologically and stratigraphically analysed to determine the sequence of occupation and the identity of the inhabitants of Mapela. Since the ceramic and bead sequences in southern Africa are reasonably well dated and are reproducible, it is easy to cross-date new sites by comparing their ceramics and beads to established types [21] ; [30].
With high levels of stratigraphic integrity, Terrace Excavation Trench 1 provided useful insights for building ceramic and bead sequences on the terrace.
The ceramics were analysed using the standard typological technique of considering vessel shape, decoration position, technique and motif see [21]. The main indication from the typological analysis was that typical K2 pottery Early Leopard's Kopje dominates the bottom of the sequence with transitional K2 ceramics in the intermediate layers.
The same pattern was also mirrored in Lower Summit Excavation Trench 1 where the bottom layers were dominated by K2 pottery, followed by transitional and Mapungubwe pottery. Occasional Zhizo ceramics were recovered in all excavation areas but more work is required to interpret the implications of this association. Therefore, the ceramic sequence at Mapela consists of all the major pottery phases associated with socio-political complexity in the Middle Limpopo Valley — Zhizo, K2, Transitional K2 and Mapungubwe.
The glass beads from Mapela were placed within a chronological and typological framework established by Robertshaw et al. This classification scheme is based on a meticulous combination of visual and metric attributes of the glass beads, but was also independently verified using geochemical techniques Table 1.
Geochemically, K2 and Mapungubwe beads are distinct which further confirms the macroscopic evidence. The earliest bead series from the terrace excavation belonged to the K2 series. These were followed by Mapungubwe series glass beads. Apart from being chronological markers, glass beads are also seen as status markers [32].
Over a thousand beads were recovered from the terrace, while far fewer came from Excavation Area 1 and the lower summit. However, a significant amount of glass beads was eroding from several areas. On the northern side, thousands of Mapungubwe glass beads were eroding out of a context which also contained Mapungubwe beakers labelled glass bead cache on Figure 7. A decision was made to salvage these beads through scraping the surface and sieving the soil Figure Overall, the glut of glass beads indicates that Mapela was a major player in trading and exchange relationships with the Indian Ocean.
The pottery and glass bead typology indicated that Mapela contains both Leopard's Kopje Phases I and II, or early Zimbabwe culture traits on the terraces, flats and hilltop. However, samples of carbonised seeds and charcoal were radiocarbon dated using AMS and conventional radiocarbon dating techniques to develop an absolute and independent chronology. Samples for radiocarbon dating were obtained from charcoal and short-lived samples, such as carbonised seeds and twigs. Table 2 presents the context, laboratory numbers and the dated material.
The dates were modelled following Bayesian techniques in the software OxCal version 4. Bayesian models are conditional probabilities which allow for pre-existing information to be incorporated into the current data, to permit the development of an integrated interpretation process [33]. The prior is a formal statement of what is known before the process of data collection, while the posterior is the desired outcome. Bayes' theorem relates posterior likelihood X to the prior.
Based on the stratigraphy, and the observation that K2 ceramics and glass beads were at the bottom, followed by transitional pottery and Mapungubwe material culture, a sequence model was run in OxCal version 4.
The recommended Southern Hemisphere Calibration Curve SHCA13 was used as it was developed using dendrochronologically dated wood from the corresponding hemisphere [35] , [36]. Because the dates were from a single stratified sequence, all the dates were combined into a single model. The results are shown in Table 3 and Figure The presence of Mapungubwe pottery and glass beads in levels 1 to 6 suggests that the sequence extends to the 14 th century.
The lack of dates from the top section of the trench was considered unimportant in view of the need to provide a tight sequence relating to the K2, Transitional K2 and Mapungubwe intervals. The dates from Garlake [17] were modelled separately as they are not from the same stratigraphic sequence.
0コメント