Ndebele Raids Effect on Shona Power

Ndebele Raiders and Shona Power

Author: D. N. Beach

Source: The Journal of African History, Vol. 15, No. 4 (1974), pp. 633-651

Published by: Cambridge University Press

FROM the twelfth to the nineteenth centuries the southern Zambezian plateau was dominated by the Shona-speaking peoples. Although the Portuguese influenced chiefs in the north during the seventeenth century, they did not destroy the basic Shona economic, social and political structure of the country. But the end of the eighteenth and the beginning of the nineteenth centuries saw the permanent loss of Shona territory in the south-west and south-east to the Tswana and Tsonga-Hlengwe, as well asthe invasions of the mfecane period, when Nguni- and Sotho-speaking peoples crossed the plateau. The middle of the century saw two Nguni speaking dynasties established, the Ndebele under Mzilikazi in the southwest and the Gaza under Soshangane in the eastern highlands. Finally, in the 1890s, the whole country was claimed by the Portuguese and the British, the latter making a large-scale settlement under the control of Cecil Rhodes's British South Africa Company.

The precise extent and nature of Gaza influence upon the Shona has not been given much attention by historians, largely because the Gaza state moved bodily to the south-east in 1889. Although the Ndebele have been examined more closely, several factors have affected both the available evidence on the Ndebele relations with the Shona and the viewpoints of historians. Firstly, most European observers of Shona-Ndebele contacts were influenced by the fact that they approached the subject both literally and figuratively from the angle of the Ndebele state itself, which added to the preconceptions they already possessed.' Secondly, exaggerated estimates of the number, scope and brutality of Ndebele raids on the Shona were later used to justify the conquest of the Ndebele by the British South Africa Company in 1893.² Thirdly, Rhodes's claims to the Shona country were based upon an exaggeration of the extent of Ndebele power.³ Fourthly, most accounts of Ndebele history after 1840 have tended to interpret the Ndebele in terms of their Nguni ancestry, ignoring the fact that the Ndebele state was a successor to the Shona-speaking Changamire Rozvi state, which influenced it in many ways. This article aims to place the impact of the Ndebele upon the Shona in perspective by relating the foundation of the Ndebele state to the Rozvi state it succeeded, and by showing how the balance of power between the Ndebele and the independent Shona changed during the nineteenth century.

The basic population of the area occupied by the Ndebele in the 1840s spoke the Kalanga dialect of Shona and was descended from the people of the Leopards Kopje culture, who occupied the area from about 1000.4 In the fifteenth century, people of the Zimbabwe culture moved west into the Kalanga country, and some time after 1450 the Torwa dynasty ruled over a state that was a successor to Zimbabwe, based on Khami. Some time between 1644 and 1683 the Torwa dynasty was succeeded by that of the Changamire Rozvi, whose main centres were the stone buildings of Danangombe (Dhlodhlo) and Manyanga (Tabazikamnambo).5 But although the Khami culture was a continuation of that of Zimbabwe,6 the Kalanga remained the basic population, and by the nineteenth century the Rozvi of the Changamire dynasty were speaking a variant of Kalanga.7 The dry environment of the southwestern plateau favoured cattle breeding, and from the earliest days of Kalanga settlement the economy of the area was strongly committed to the build-up of herds. In addition, until the early nineteenth century the goldfields of the area were in production on a reduced scale and the Torwa and Changamire dynasties exported gold and ivory to the Portuguese centres on the Zambezi and the coast in return for cloth and beads, as well as other articles.8

When, in the early nineteenth century, the migrations of Mpanga, Ngwana Maseko, Zwangendaba and Nyamazana, set off by the mfecane, struck the Changamire state, it was poorly prepared to stand the shock. Droughts, wars with the Karanga advancing from the north-east, Tswana raids, strife between the royal dynasty and the Mwari cult and civil wars within the dynasty had seriously affected the strength of a state that had earlier been one of the foremost military powers of southern Africa.9 By the early part of the nineteenth century the most important Rozvi families were grouped in a ring around the Changamire capitals of Danangombe and Manyanga east of the Bembesi river. West of the Bembesi, the Khami area was under a Kalanga sub-ruler, Ndumba.'0 The most serious split within the Changamire dynasty was between the houses of Miutinhima, a son of the Changamire Gumboremvura, and of Chirisamhuru, who was Changamire in the early 1830s.11 The Mutinhima faction occupied the Mulungwane hills and influenced a wide area east of the upper Lundi. Other Rozvi sub-rulers of the main dynasty were Lukuluba of the Ghoko hills, Rozani of the Vungu river, Swabasvi of the Somabula forest, and a ruler praise-named Dlembeu on the Mpopoti range.12 Nevertheless, weakened as it was, the Changamire state withstood the attacks of Mpanga, Ngwana Maseko and Zwangendaba, who were driven off, although they did a great deal of damage, taking grain and cattle. Even the death of Changamire Chirisamhuru at the hands of Nyamazana's group did not destroy the Changamire state.13 It seems likely that no Mambo was installed immediately after Chirisamhuru's death, but at some point well before 1852 his son Tohwechipi succeeded him, and as he received the support of the Mutinhima house he can fairly be said to have been the next Mambo.14

Ndebele State and its Tributaries

The Ndebele invasion differed from those of the Ngoni in several ways. Firstly, Mzilikazi's people made no attempt to attack the central part of the Changamire state, but settled in the western province of Ndumba, westof the Bembesi river. Secondly, they consolidated their power by exploiting the splits between the Rozvi, and by entering into an economic relationship with them. The main body of Ndebele under Gundwane arrived from the Umzingwani valley in I838-9.15 Ndumba's dynasty vanished from the scene relatively early,16 and the main resistance in the immediate area was led by Mutinhima from the Mulungwane hills. Mutinhima, nicknamed mafuta, was at first successful in his defence, and may not have been pushed out of the hills until after Mzilikazi arrived.17 The Ndebele succession crisis undoubtedly delayed the impact of the Ndebele upon the Shona, but even so it seems to have been surprisingly mild. There are reports of some raids made upon the local people in the first year,18 but tradition from the Kalanga pointed out that although 'they killed a lot of people . . . none of my family were killed. We did not regard the Matabele as bad people. The only thing they fought over was grain.... There was no trouble when the Matabele came'. This was contrasted with the rapacity of the Ngoni during previous invasions.19 On the Manzamnyama river the local Kalanga fled briefly and then returned as tributaries of the new overlords. 20 In short, west of the Bembesi the place of the Ndumba dynasty was simply taken by Mzilikazi and his followers, who settled down among the Kalanga as the Rozvi had done before them. It was thus logical that the Ndebele should not remain on hostile terms with the local Shona because they needed supplies of grain, which would not be forthcoming if raiding was continued for a long time. In I854 Moffat noted the Ndebele prosperity in grain, and in 1858 he confirmed that the Shona were continuing to live inside the Ndebele-settled area in their own villages.21

The Ndebele had thus become rivals of the Rozvi as rulers of the Kalanga and other Shona peoples, and in the period when the Changamire dynasty was weakened by the death of Chirisamhuru, Mzilikazi actually took his place as overlord of certain Rozvi families of the main dynasty. These, including Swabasvi, Lukuluba and Rozani, may have been motivated by internal political jealousies among the Rozvi such as their exclusion from the centre of power and the succession, but they also had an economic motive. The Ngoni invasions had resulted in the loss of a great number of the prized Rozvi cattle, so that there was a considerable shortage.22 The Ndebele, on the other hand, had plenty of cattle but desperately needed more people. The result was that an exchange took place, in which Mzilikazi distributed cattle to the Rozvi mentioned above in return for young people, who were incorporated into the Ndebele state and society.23 This state of affairs extended over the eastern half of the Changamire state and even as far as Tsunga, the land between the upper Umniati river and the Mwanesi range. There, the Nyandoro dynasty had acted as intermediaries between the main Rozvi dynasty and the peoples north-east of them.24 At some time before the great campaigns in that direction in the 1850s and 1860s Ndebele cattle were distributed there as well. 25

It was not likely that this situation would endure for long without serious trouble. Firstly, there was the problem of the main Rozvi dynasty, which had withdrawn into the hills that fringed the Changamire state to the east.26 It was hardly likely to accept the loss of its position without making some attempt to regain it. Secondly, the economic exchange created by the Ndebele was, in the long run, extremely disadvantageous to the Rozvi and other Shona who took part in it. The Ndebele appear to have retained ultimate ownership of the cattle that they distributed, although the milk and limited slaughter rights would presumably have been accorded to the herders, but the young people who were levied by the Ndebele were not allowed to return to their own societies. Moffat noted in 1854 that 'there is nothing they deplore so much as their children being taken from them just at a time when they become useful to their parents', 27 and this practice, taken to excess, could ruin a Shona society. Later the Ndebele did not need to recruit so many from their tributaries, and so caused less damage and created less resentment. But the combination of a serious grievance and the existing organization of the Rozvi state led to the first serious Shona resistance to Ndebele rule.

It is important to stress that the Ndebele did not believe in total war any more than the Shona believed in total peace. During the warfare between the Chirumanzu dynasty and the Ndebele in the 1850s Moffat was able to note that between the fighting in 1854-5 and the surrender of Chirumanzu in I857 there had been no further fighting.28 In August 1866, the Ndebele attempted to trade with Mashayamombe's people, despite their attack on the latter earlier in the year.29 As for the Shona, even the Njanja, whose exploitation of the Wedza ironfield and wide-ranging hoeselling network was one of the great economic success stories of the late eighteenth and nineteenth centuries, and who depended for their sales on good relations with the surrounding peoples, did not hesitate to rob 'Portuguese' zungu traders passing through their territory, even though this hardly endeared them to the peoples who expected to receive the trade goods that the zungu were importing.30

Even when Ndebele raids were major ones, directed against specific targets on the orders of the king, there was a tendency for other people in the area to suffer as well. This was because the Ndebele, whose famed military discipline existed far more in the minds of European writers than it ever did in reality, were prone to scatter across a wide area in search of cattle and women. This emerges clearly from all detailed accounts from

Europeans living among the Shona. Thus, quite unlike Montagu Kerr's fictitious stereotype of a surprise attack on a surrounded village, Mauch, in 1872, gives a picture of a series of raids which continued over a period of about three weeks over a very wide front, from the western Duma on the Mtilikwe-Pokoteke confluence to the upper Pokoteke, an area of some forty miles. The Shona, who had at least three days' warning, suffered various losses but were rarely taken completely by surprise.31 The same picture emerges from accounts of the 1892 raids on the country from Chivi to Gutu and from those of I893 on Zimuto.32 Ndebele raiders also tended to follow up their targets if they fled, as when they pursued people from the Chaminuka medium's base near the Umfuli to the northern Shawasha country beyond the Umwindsi in I883.33 Even Ndebele on a peaceful mission, such as the delivery of a message, would sometimes cover

If the main stimulus behind the great Ndebele campaigns of the 1850s and 1860s was the political threat of the Rozvi dynasty, the economic stimulus of the Shona trade system was also extremely important. Indeed, it appears to have provoked the first important expansion of Ndebele power. In inheriting the Changamire state, the Ndebele had inherited its basic economic framework, which, in spite of a regional emphasis on cattle, was also aligned to the traditional exchange of gold and ivory for cloth and beads. Thus, even during the fighting of 1854-5, the trade system linking the Ndebele with the Zambezi and the coast through the north-eastern Shona country continued to function. Nevertheless, by the 1850s it had become clear to Mzilikazi that the Shona were difficult to dislodge from their mountain strongholds.35 At this point neither side had guns in quantity, although the Shona had been importing a certain number of guns for a very long time.36 The Ndebele had learned from their experiences south of the Limpopo that guns were useful, and in the 1850s and 1860s they did their best to acquire them.37 The basic Kalanga population of the Changamire state had been accustomed to import cloth, and their needs also had to be supplied. Although supplies of both guns and cloth were available through the variously friendly, neutral or hostile Shona dominions to the east, it was obviously desirable for the Ndebele to control the trade routes to a greater extent.

The first expansion of the Ndebele was to the north-west, however. One reason for this was probably that the Ndebele state was too weak in the 1840s to attempt the more hazardous-because more heavily populated and thus better defended-route to the north-east. The badly-watered sand country to the north-west was almost uninhabited and thus offered no resistance to raiders who could easily cross it to strike at the Shona under Hwange, Pashu and Saba on the Deka, Gwai and Zambezi rivers. These people were not only vulnerable but also offered access to one of the trade routes to the sea. This route was along the Zambezi through the Tonga country to Zumbo, Tete, Sena and the sea, and was economically viable in spite of the distance involved because nearly all the distance could be covered by some sort of water transport. Water transport, as Selous noted, made goods from these Portuguese ports much more competitive than those hauled by wagon from the South African ports,38 and in the 1860s 'Portuguese' traders were operating near the Victoria Falls.39 By the early I850s the Ndebele appear to have established their authority over the Zambezian polities,40 especially after the death of the Hwange in 1853,41 although intermittent raids on the area occurred for various reasons as long as the Ndebele state survived.

A Rozvi tradition from the Insiza area suggests that Mzilikazi extended his policy of co-operation with the Rozvi to the point of requesting Chirisamhuru's son Tohwechipi to return from his exile in the direction of the eastern Highlands and settle down in his own country, and that it actually worked for a few years before Tohwechipi broke away.42 It seems certain that the Ndebele tried to get the Mutinhima house to join them, but that they refused. The 1850s saw a rapid revival of the Rozvi power, and as mentioned above they seem to have sunk their differences sufficiently for the Mutinhima group to recognize the paramountcy of Tohwechipi. Even the Swabasvi house broke away from the Ndebele and joined the Mambo's Rozvi.43 But it does not seem that there was a single Rozvi command over the rest of the Shona. Even some of the Rozvi under Lukuluba and Rozani remained under the Ndebele,44 and accounts of the period are full of stories of quarrels and warfare between the Rozvi and rulers such as Hwata, Gutu and the people of the upper Sabi valley.45 Other Shona rulers did attack the Ndebele at the same time as the Rozvi, but it appears to have been on their own initiative.

If the Ndebele exactions of young people provided the basic motive for the Shona resistance to Ndebele rule in the early 1850s, and the revival of the Changamire dynasty gave an example to be followed, the resistance

took a thoroughly traditional form. Shona raiders penetrated deep into the country of the Ndebele, stealing cattle and-according to the Ndebele committing atrocities on women. The most prominent of the raiders were

the Mambo Tohwechipi, his relative Mutinhima, and Chizema the son of the Govera ruler Chirumanzu on the Shashe.46 But these raids provoked an Ndebele response that proved too strong for the Shona. Battles were

fought in the mountains to the east of the Ndebele state: at the Mipopoti range against the Rozvi ruler 'Dlembeu Kupengobuta',47 at Umgulugulu (Guruguru) mountain,48 and against the Mhari ruler Zingwe, who was killed for refusing to supply young people as tribute.49 Tohwechipi was forced to retreat through Chivi past Nyaningwe hill in the direction of Zimbabwe,50 and it may be near there that he won the defensive battle of Chikato. By employing zvitunya-strong people51-who came from the Zambezi to trade and who possessed guns, he was able to defeat the Ndebele some time before 1852, winning himself the name of Chibambamu in the process.52 The fighting continued into 1854, and then there was a lull. But by 1857 the situation had resolved itself in Mzilikazi's favour. The Chirumanzu dynasty surrendered early in 1857,53 and from then until 1889 became a strong ally of the Ndebele.54 Indeed, Chizema, who had been so prominent in raids on the Ndebele, was aided by them in his unsuccessful attempt to win a new land for himself in southern Buhera in the years that followed.55

As for the Rozvi, they suffered from the lack of unity among the Shona peoples. In July 1857 it was noted that 'the rulers holding these lands [goldfields in the central Shona country] were tributary to the emperor Changamire, but today, by a betrayal, the "Uata Mezircase" has taken possession of them from the said Changamire, who lives as a refugee in Njanja, land of the ruler Gambiza, in the district of the Hera, on the edges of his vast domains.' This 'Uata' could have been Mzilikazi, but 'Uata' was also the Portuguese rendering of 'Hwata'. Hwata was the ruler of a comparatively small Hera polity at the head of the Mazoe valley, but if his territory was small his economic influence was considerable. He controlled the goldfields in the northern Shawasha country to the east of him,56 and the locality of the old Portuguese feira of Dambarare.57 This strategic position-probably of importance far back into early Shona history enabled Hwata to dominate much of the trade of the central Shona country, buying ivory and reselling it to the 'Portuguese' traders,58 whose houses were to be found in the upper Mazoe valley.59 Hwata guarded this economic advantage jealously, and when the people near the old feira of Maramuca60-probably the Devera group that owned the Shurushuru goldfield61-attempted to re-open it to 'Portuguese' trade in c. 1830-50, he attacked both them and the 'Portuguese'. He lost the battle but won the war, for the feira was not re-opened.62 Mzilikazi devoted as much effort to the defeat of Hwata as he did to Tohwechipi, and the continued subjection of Hwata to Ndebele rule until I889 suggests that Mzilikazi was fully aware of the economic importance of Hwata's area and intended to profit by it.

There seems to have been a period of peace from 1854 to 1860, but from the latter date to 1873 the Ndebele made what was probably their greatest concerted effort to dominate the Shona. They raided over a wide front from Chivi in the east to Mangwende in the north-east and Hwata in the north, and in the northern areas in particular the relatively few raids mentioned in traditions most probably occurred during this period. Even so, it does not seem likely that the Ndebele were numerous enough to affect all these areas at once, and in one year, I863, when the main strength of the kingdom was turned against the Ngwato to the south-west, the only noted effort to the north-west was a raid by associates of the Ndebele on the Deka river area,63 while another force raided Hwata's associate, Chiweshe, in the upper Mazoe valley.64

The Ndebele effort of 1860 was confined to a small raid to the north-east and another to the south-east. After this, it is possible to make some estimate of the sequence of events in each area affected by the surge of Ndebele activity. In the east, the peoples of Chivi,65 Bere, Zimuto, and the Njanja were attacked in 1861. Bere's Mhari bore the brunt of the attack,66 and were severely weakened in consequence,67 while Chivi's Mhari appear to have succumbed to the power of the newly-imported guns68 and became tributary to the Ndebele. A combination of ambition on the part of Chivi Matsweru's son, Makonese, and the Ndebele expansion led to the deaths of both Chivi and Bere at about this time.69

The attack on the Njanja mentioned above brought the Ndebele back into contact with the Changamire Rozvi, who had arrived in the Hera country-dominated by the two Hera rulers Mutekedza and Nyashanu and

the rapidly-expanding Njanja confederacy under Gambiza-by 1857.70 The Rozvi, led by the Mambo Tohwechipi Chibambamu and his cousins of the Mutinhima house, occupied hills on the frontier between Nyashanu and Gambiza such as Bedza and the Mavangwe range.71 It is stated that the Ndebele made three major attacks in order to rid themselves of the menace of the remnant of Rozvi imperial power, until in 1866 a prolonged siege forced Tohwechipi to surrender.72 He was brought to Mzilikazi but later allowed to leave. Tradition is emphatic that he left Mavangwe and went to Gutu, where he died,73 but in view of the fact that in 1873 Mtikana Mafu led a major force against the Rozvi in Gutu it seems possible that even the defeat of 1866 did not crush Rozvi resistance to the Ndebele.74 In view of the tendency of Ndebele raiders to spread across country it seems likely that most of the damage suffered by the Njanja and the Hera of Mutekedza and Nyashanu occurred at this time. Certainly by 1870 the Ndebele had raided Mutekedza, since they had mutilated his sub-ruler Nyoka.75 This may also be the period of Chizema's attempt to conquer southern Buhera from Nerutanga, which was repulsed by that ruler and the Njanja in spite of his Ndebele backing. The Njanja recall having aided Gutu after this, which may coincide with the I873 raid there.76 The Nhowe of Mangwende remembered in I898 that 'it was in following up the Abarosis that the Matabele first came to know our country, with the result that they commenced killing and raiding through the different districts'.77 The Rozvi were not the only ones responsible, however, because the Ndebele tendency to follow up their enemies applied to Nyandoro of Tsunga as well. Nyandoro had been herding cattle for the Ndebele, but at about this time the murder of an Ndebele nduna led to hostilities,78 and the Ndebele attacked Tsunga, which was flat and nearly indefensible. This was thus one of the very few areas occupied by the Shona that was depopulated by Ndebele action, as the Nyandoro people left en masse in the general direction of their seventeenth-century home in Fungwe. They moved to the nearby Nyoka river, and after a year or two raids pushed them further north-east to the Chirume. A few years later more raids drove them to the Matswitswi caves in Samuriwo's land, from which they fled after Nyandoro's death and further raids as refugees to Mangwende.79 Like the Rozvi, they brought the attention of the Ndebele to those peoples who lived nearby, so that Samuriwo, Chihota, Svosve and Mangwende all suffered.80 But, as the Mangwende people pointed out in 1898, 'The first time they entered the country very few of Mangwende's tribe were killed, and very few were taken prisoners, but they took away with them large numbers of cattle and goats.... The Matabeli never came back into this district again but every year they were raiding the districts on the Sabi river.'80 This marked the farthest point of Ndebele raiding to the north-east.

The pursuit of the Rozvi Mambo and his associate Nyandoro led the Ndebele straight to the north-east up the watershed of the whole country, over open, grassy plains. These were of little significance to their economy except as sources of cattle to replenish their herds, after the lungsickness of 1861 had so reduced them that the kingdom actually contracted in size at this time.82 The route to the north, however, not only led to the trade routes of the heart of the old Mutapa state, but was also of considerable economic importance in itself, running as it did through some of the biggest goldfields still being worked in the early part of the century83 and

across river valleys running west from the watershed, that were full of elephants.84 The great Ndebele efforts of 1860-8 in this area hit the inhabitants very hard. The Ngezi dynasty of Rimuka partly broke up,85 and the Mashayamombe and Chivero people of the Umfuli valley also suffered, so that at one point their rulers were forced to flee to the north.86 However, Mashayamombe at least appears to have returned to his land by 1866, in time to be raided once more. Mzilikazi's attempt to trade for ivory later that year suggests economic motives.87 Economic motives almost certainly also lay behind the very determined efforts made to subject the Hwata trading centre. For four years from 1860-1 the Ndebele attacked, even sending Lotshe to raid Hwata's associate and neighbour Chiweshe at a time when the greatest need for men lay on the Ngwato front to the south-west in 1863. Finally Hwata surrendered in 1864, and was captured to be returned to his home as a tributary ruler.88 However, Hwata, like the Mambo Tohwechipi, appears to have attempted to break away from this relationship, because a major campaign was required in 1868 to subject him again. Even in 1870 his allegiance to the Ndebele was thought to be superficial,89 but he remained at least nominally tributary until 1889.90

After these major campaigns the Ndebele succeeded in making a number of Shona political units tributary to them. This relationship involved the payment of articles of partly symbolic value, such as skins, feathers, hoes, tobacco or spears, or the provision of services, such as labour for hutbuilding.91 Nevertheless, such tribute represented a considerable diversion of valuable man-hours among a people living in a largely subsistence economy. Consequently there was a tendency to break away from Ndebele domination that ran counter to the strong attraction of Ndebele society to many Shona. It was noticeable that in many of the tributary lands farthest from the main Ndebele state the tributary relationship took the form of an alliance between a Shona faction, that relied upon Ndebele support against its rivals within the dynasty, and the Ndebele, who themselves could rely upon this faction. This seems to have been the case with Hwata, Mutekedza and Chirumanzu. The main body of Ndebele tributaries was bounded by a line that ran from the KweKwe river east to Chirumanzu, south to Chivi and then south across the lowveld to Matibi. This boundary varied from year to year, depending upon the political situation in the border areas. Hwange, the Shangwe of the Mafungabusi plateau, Nemakonde, Hwata and Mutekedza were all outlying tributaries, separated from the main body by almost uninhabited land or independent Shona lands.92

The wars of the 1850s established the dominance of the Ndebele in the vicinity of the old Changamire state. The campaigns of the 1860s wiped out the last power of the Rozvi and gave the Ndebele strong economic advantages in the north. In spite of the dissensions of the succession crisis of 1868-72, in terms of relations between the Shona and the Ndebele the latter were, by I873, at the zenith of their power. It is thus ironic, in view of the myths of Ndebele supremacy, to note that their first serious defeat, and the first sign of a change in the balance of power that was to lead in the end to the revolt of many of the Shona tributaries, occurred only six years later, in I879. Ironic, but not surprising. The Shona after all were descendants of the creators of the most impressive Iron Age material culture in southern Africa, the Zimbabwe-Khami culture. They worked what was left of considerable goldfields, and had access to many elephants. Their political institutions and territories were small only by comparison with the few super-states of southern, central and east Africa. By comparison with most polities of that area many Shona rulers held quite big territories. Most of them owned superb defensive sites. Moreover, developments to the south were beginning to aid the Shona. The opening of the Kimberley diamond fields in 1867, the increased availability of guns as Europeans adopted rifles, the expansionist ambitions of Britain and the Afrikaners-and Portuguese counter-moves-all tended to aid the Shona in the short run, though not in the long. Under the circumstances it is surprising that the Ndebele accomplished as much as they did. The sale of gold and ivory and the labour opportunities of Kimberley and the Rand made it possible for the Shona, who in the 1860s had been fatally short of guns by comparison with the Ndebele,93 to re-arm. Guns, which entered the Shona country in the hands of 'Portuguese' traders, Venda mercenaries and gun-runners from the Ndebele kingdom, as well as through Shona long-distance traders and migrant labourers, made the hill strongholds of the Shona almost impregnable even against gun-using Ndebele and Europeans, as the 1896-7 risings were to show.94

In 1879 the missionary Cockin wrote that 'latterly some of the kraals attacked have shewn fight and being many days away and the towns denser, the Amandebele are becoming afraid to go there so much. Cattle and sheep and slaves (are) not coming in so freely now from these distant raids . . .'95 and it is probable that he was referring in particular to the war with Chivi. In the 1860s Chivi was evidently tributary to the Ndebele, but in the reign of Mazorodze, who ruled from 1870 at the latest, the Mhari began to acquire guns from the Venda and to build up a considerable herd of cattle, guarded by a group of men. This represented a threat to Ndebele power in the area, and in October 1879 a major force under Lotshe and Manyewu attacked the Mhari capital of Nyaningwe. Although the Ndebele force consisted of the Mbizo ibuto and probably outnumbered the defenders of Nyaningwe, they were repulsed with the loss of twenty men, their only success being the capture of the Chivi himself on an outlying hill. The loss of twenty men was not significant in itself, but the defeat was, and even the execution of Chivi did not hide the fact that although the Ndebele could operate over the open ground, they could not take the hill-strongholds of the Mhari, who were henceforth independent.96

The year 1880 saw the defeat of the Gaza by Gutu in the similar battle of Rasa mountain,97 and the beginning of a rift between the Ndebele and their ally the Chaminuka medium of the upper Umfuli. Up to then, Lobengula had paid the medium tribute in return for religious services,98 and in that year they combined to raid the Shona north of the Hunyani. But at the same time the Chaminuka medium claimed that he, and not Lobengula, had the power to grant hunting-rights to Europeans in the area, and his son Jugu 'had said that his father would now show Lobengula that the country beyond the Umniati river belonged to him, Chameluga' and that if necessary he could drive away the Europeans by his magic. It is therefore not surprising to find that in 1883 Lobengula had the Chaminuka medium killed. His men raided as far as the Shawasha country of Chinamhora, 99 whose people had taken Chaminuka's cattle,100 which in all probability had been taken from them in 1880.

In 1882 Selous noted that Ndebele had reached the Mukwadzi river west of the Umvukwe range,100 and in 1887 there was a major raid on the Umvukwe area,102 probably the one led by Gwasagwasa against the Shona ruler Chipuriro, far to the north.103 This may have represented a revival of the policy of the 1860s of gaining control of the trade routes to the Zambezi, for beyond Chipuriro lay the prazos of Matakenya, Jose de Araujo Lobo, who had earlier been in contact with the Ndebele, buying their ivory. In 1888 a major raid struck at the Mashayamombe and Rwizi people of the Umfuli valley.104 The reason for this is not known. The attack on Rwizi may have been to prevent a renewal of the Chaminuka cult, while the fact that a very large number of people were removed from Mashayamombe's may mean that this was the raid, recalled in tradition, that resulted from a civil war among Mashayamombe's people in which one side called in the Ndebele.105 But it may also have resulted from the fact that the Shona were undeniably growing stronger. Isolated Ndebele were liable to be killed if they were discovered. In 1887 a whole party of Ndebele was killed, and indeed the 1888 raid 'suffered so severely that Lobengula was very angry and another one was sent out in another direction.' Montagu Kerr, with his preconceived ideas, was amazed to hear the Shona at the head of the Mazoe valley in 1884 coolly discussing their chances of success, with some hope of victory, but it seems that in the 1880s the Shona were indeed beginning to turn the tide of Ndebele power.106

This trend was greatly advanced by the Portuguese, who, seeking to counter British expansion by limiting the power of the Ndebele upon which Rhodes's claims were based, sent two expeditions into the Shona country in 1889. Nominally led by Vitor Cordon and Paiva de Andrada, but actually relying upon the forces of such prazo-holders as 'Kuvheya', Manoel Antonio de Sousa, and 'Chimbango', Vicente Jose Ribeiro, they reached Nemakonde and Mangwende respectively, and made treaties with Shona rulers in a wide belt across the country from the Mafungabusi plateau to the upper Sabi. Portuguese flags were distributed in acts of vassalage that were not taken seriously by the Shona, and large numbers of guns were handed over by the Portuguese.107 The effect of the treaties was to give nearly every Shona polity north of the Umniati river a considerable increase in the size of its armoury. The 'Portuguese' zungu expedition to the whole Charumbira-Mapanzure-Bere group of peoples in 1872 had only had forty-eight guns for sale,108 but the 1889 expeditions gave this many to a single ruler. Even the small polities received ten guns, and powder and ammunition were supplied as well.109 This was a huge increase in Shona fighting strength, and from both oral traditions and some of Cordon's treaties110 there is no doubt that the whole tenor of this major political development was anti-Ndebele. The implications of this in Shona history proper cannot be discussed here, but the effect on the balance of power between the Ndebele and the Shona was immense. Hwata, Nemakonde, Mutekedza and some of the Shangwe all abandoned their allegiance to Lobengula and accepted the Portuguese guns and flags which were to be found as far south as the Njanja country and beyond.111 No major raiding forces of Ndebele ever entered the central Shona country again. There are strong suggestions that the revolt against Ndebele power even extended as far south as Gutu, where from 1889 the rulers no longer had the Gaza state to balance against the Ndebele, and to Chirumanzu, where the death of Bangure allowed his brother Chatikobo, aided by some Rozvi, to lead the people into their first revolt since 1857.112 But the arrival of the British in 1890 altered the situation.

Although there can be no doubt that Lobengula was thoroughly opposed to the arrival of Rhodes's men on his eastern frontier, he remained functionally neutral to the extent that he did not attack the Pioneer column, and once the British had driven away the Portuguese and captured the formidable Sousa, he took advantage of the British presence to regain control over Nemakonde and Chirumanzu, although Hwata and Mutekedza remained lost to him. Even so, this was only possible with the co-operation of Shona interest groups. In a coup d'etat in 189I, Chinyama, son of Bangure, drove out his uncle Chatikobo and became the new Chirumanzu with Ndebele aid.113 At the end of the year an Ndebele force visited the Nemakonde area and, after consultation with the most important spirit medium, killed Nemakonde Hodza and four others in an action that has all the marks of a coup d'etat by an internal group.114 During 1892 a similar split in the Gutu dynasty and an appeal for Ndebele help by Makuvaza led to a joint Chirumanzu-Ndebele force installing him as ruler;115 at this time a small party of Ndebele even reached the highlands across the Sabi, perhaps the furthest point ever attained in that direction, in this last rather feeble demonstration of Ndebele power.116

It is beyond the scope of this paper to discuss the events and negotiations that led to the war in 1893 that broke the Ndebele state, except insofar as they concern the Shona. The lowveld area had been subjected to raiding for years, partly at least by unofficial raiding parties, causing Matibi to move away from his northern lands into the remote lowveld in the late 1880s.117 But the decisive area was around Chivi and Zimuto. Chivi was raided in late 1891, to the delight of Rhodes, who was trying to prove Lobengula's dominance of the area,118 but in July and August 1892 a major raid on the recalcitrant Chivi and Zimuto led to an appeal by Chivi to Rhodes's deputy Jameson.119 This in turn led to a demand that Ndebele raiders stay away from the town of Victoria and the main road,120 a demand that was fully complied with, as far as the Ndebele ruler was concerned, until the crucial raid of July 1893.121 Even this raid came about partly as a result of Shona actions. Early in June 1893 a joint party of raiders from Bere and the Makamure house of Zimuto stole cattle from Mpakame, a Shona tributary of the Ndebele at Guruguru hill. Mpakame complained to his overlord the Ndebele-ized Rozvi Lukuluba at the Ghoko range. Lukuluba raided Bere in retaliation, but on being turned back by Company police, reported in turn to his superior, Mgandane of Nxa.122 This led directly to the famous raid in July on Bere and Zimuto near Victoria, to the fight of 18 July and to Rhodes's decision to overthrow the Ndebele kingdom.

Even before the British columns set out, however, Shona raiders were moving in to take Ndebele cattle.123 As the Victoria column began to move towards its rendezvous in the north with the Salisbury column, it was joined by large forces of Shona. Zimuto sent 120, Madziviri 50, and Gutu abandoned the Ndebele who had put him in power the previous year and sent eighty men. As the force approached Chirumanzu its ruler Chinyama followed Gutu's example and offered 300 men.l24 A few days later Chivi's men marched through Victoria to catch up with the advancing columns.125 These Shona, acting in concert for the first time in their particular histories, fought at the Shangani battle with some success, considering that they were left outside the defensive laagers.126 Meanwhile in the south, Matibi, whose relations with the Ndebele had been deteriorating to the point of outright war as he was repeatedly raided, experienced a further raid in late 1893 and retaliated in force, together with the police. They penetrated deep into Godhlwayo.127 These Anglo-Shona alliances of 1893 had a profound effect on the subsequent history of the southern Shona, especially in 1896. As the Ndebele state fell, Shona raiders from all over the southern Shona country and from as far as the upper Sabi valley began to move towards the Ndebele herds, and the end of Ndebele power in the summer of 1893-4, saw Shona raiders striking deep into the Ndebele kingdom, as they had done in the early 1850s.128

SUMMARY

For a variety of reasons the extent, number and severity of Ndebele raids upon the Shona-speaking peoples have been greatly exaggerated in the past. Moreover most studies of the Ndebele have failed to take into account the fact that the Ndebele conquered an already well-established Shona state with an economy linked with the Indian Ocean trade. This article seeks to show that the bulk of Ndebele raiding before 1873 was a response to both the political and military threat of the Changamire Rozvi dynasty and to the economic needs of the state that had been taken over. Even so, Ndebele raids were limited in extent and duration, and the two decades after 1873 saw a steady revival of Shona strength. In spite of Ndebele raids aimed at preventing this process, the independent Shona strengthened themselves by re-arming and 'alliances', and were able to take part in the eventual overthrow of the Ndebele kingdom.

This article is a condensed version of a paper presented at the History Workshop,

Gaborone, September 1973.

Notes

1.  Sources with biases, conscious or unconscious, include: The Matabele Journals of Robert Moffat, 1829-I860, ed. J. P. R. Wallis (London, 1945); The Matabele Mission ...of John and Emily Moffat, I858-I878, ed. J. P. R. Wallis (London, I945); W. Montagu Kerr, The Far Interior (London, I887); Gold and the Gospel in Mashonaland 1888, eds. C. E. Fripp and V. W. Hiller (London, 1949); The Northern Goldfields Diaries of Thomas Baines, ed. J. P. R. Wallis (London, 1946); J. Mackenzie, Ten Years North of the Orange River (Edinburgh, I871); F. Coillard, On the Threshold of Central Africa (London, I897) A. A. Anderson, Twenty-five Years in a Waggon (London, I888); The Southern African Diaries of Thomas Leask I865-I870, ed. J. P. R. Wallis (London, 1954).
2. 'I hope they do raid the Barotses', wrote the Company's secretary in I892. 'All these raids and deaths and murders ought to be entered into a book, so that we may always be able to prove justification and their being a cruel damnable race.' Hist. MSS CO 4/I/I F. Rutherfoord Harris to J. W. Colenbrander, 9 Feb. I892. Unless otherwise stated all reference codes relate to the National Archives of Rhodesia, Salisbury.
3. T. O. Ranger, 'The rewriting of African History during the Scramble: the Matabele dominance in Mashonaland', Afr. Soc. Res. iv (I967); D. N. Beach, 'The Adendorff Trek in Shona History', S.A.H.J. III (197I).
4. T. N. Huffman, The Leopards Kopje Tradition (in press).
5. This interpretation, which differs radically from that generally made previously, was made independently by Dr Mudenge (S. I. Mudenge, 'The Rozvi Empire and the Feira of Zumbo', unpubl. Ph.D. thesis (London, 1972), 35-43, 54-70) and myself (D. N. Beach, 'Historians and the Shona Empires, II, 3, The Changamire Problem', Univ. of Rhodesia, Hist. Dept., Henderson Seminar Paper, 20, 1972), but to Dr Mudenge belongs the credit for convincingly identifying the Khami culture in its earlier phase with the Torwa dynasty.
6. T. N. Huffman, 'The Rise and Fall of Zimbabwe', J. Afr. Hist. XIII, 3 (I972), 356-7.
7. G. Fortune, 'A Rozvi text with translation and notes', NADA, 33 (1956).
8. Mudenge, 'The Rozvi Empire', 52, 70, 159-60; N. Sutherland-Harris, 'Trade and the Rozvi Mambo', in J. R. Gray and D. Birmingham, eds., Precolonial African Trade (London, 1970).
9. F.W.T. . Posselt, Mambo and his Court (Salisbury, 1923), 2; N. 3/33/8, J. W. Posselt, N. C. Charter to Acting C. N. C. Salisbury, c. I Jan. 1904; Mudenge, 'The Rozvi Empire', 187-91. Dr Mudenge's use of Portuguese documents to date Rozvi reigns in the late eighteenth century is one of the most important breakthroughs in the study of Rozvi history. His explanation of the origins of the Jiri-Gumunyu dispute is another, and his argument has influenced my use of the evidence cited below. 
10. K. R. Robinson, Khami Ruins (Cambridge, 1959), Appendix 6, Tradition 2. This is a reflection of a tradition recorded in one of a set of very valuable traditions collected in 1898-1906 (NB 6/i/i, N 3/33/8 and A 3/I8/I8): A 3/18/28 C. T. Stuart, N. C. Gwelo to C. N. C. Bulawayo, 16 Feb. 1906. This is not the Ndumba after whom the Ndumba hills at the Bubi source are named, A 3/18/28 R. Lanning. N. C. Inyati to C. N. C. Bulawayo, 21 Apr. 1906.
11. The main traditions of Rozvi politics are found in J. W. Posselt in N. 3/33/8; F. W.T. Posselt, Mambo; M. V. J. Rukara, in Hist. MSS Msc. RU 4/I/I; A. Marwodzi, 'The Barozwi', NADA, 2 (1924); E. M. Lloyd and S. Muhlanga, 'Mbava', NADA, 3 (I925) and 'Mbava and the others', NADA, 4 (1926); Fortune, 'A Rozvi text'; K. R. Robinson, 'A History of the Bikita District', NADA, 34 (1957); S. Rhodesia Ministry of Internal Affairs Delineation Reports, Bikita and Buhera, I964-5. The simplest and most convincing explanation of the dispute is that the modern Jiri group are descended from Mutinhima, son of Gumboremvura, and that the Gumunyu group are descended from Chirisamhuru. The co-operation of the Mutinhima house with the main house in the 1850s is described below. The death of Mambo Tohwechipi Chibambamu in c. 1873 after his defeat in i866 led to the political eclipse of his house and the rise of Mutinhima's. The earliest known reference to Tohwechipi's son or brother Chikore makes it clear that he was not regarded as Mambo-although the reference must be regarded with caution on this point because it comes from a traditionally pro-Mutinhima area (A 3/I8/28, F. G. Elliott, N. C. Selukwe to C. N. C. Bulawayo, I9 May I906). By I890 the people near Zimbabwe stated that 'Tihina' (Mutinhima) was ruling the Bikita Rozvi (Argief van die N. G. Kerk, Cape Town, Report of S. P. Helm, I89I). In I896 the Gumunyu-descended Chiduku group joined the ritual-position holder Mavudzi and the Mbava group in the installation of a Mutinhima-house member, Chikohore Chingombe, as Rozvi Mambo in the Mavangwe hills, Buhera (N i/i/8, N. C. Ndanga to C. N. C. Salisbury, 2 Mar. I897). Subsequently the Mutinhima house dominated the Buhera and Bikita Rozvi until the revival of the Chirisamhuru-Gumunyu house in Bikita in the I950s.
12. F. W. T. Posselt, Mambo, 6; Sr Mary Aquina, O.P., 'The tribes in Chilimanzi Reserve and their relation to the Rozvi', NADA, IX, 2 (1965), 4I; Fortune, 'A Rozvi text', 72; A 3/18/28 T. M. Thomas, Acting N. C. Insiza to C. N. C. Bulawayo, 16 May 1906; MIA Delineation Report, Que Que, I963; Stuart in A 3/1i8/28; NB 6/i/i S. N. G. Jackson, N. C. Belingwe to C. N. C. Bulawayo, 31 Mar. 1898.
13. G. J. Liesegang, 'Nguni migrations between Delagoa Bay and the Zambezi, I821-I839', Afr. Hist. Studies, III, 2 (1970); W. A. Elliott, 'The Ma-Shuna' in D. Carnegie, Among the Matabele (London, I894), II3; F. G. Elliott in A 3/I8/28; S. N. G. Jackson in NB 6/i/i; Hist. MSS WE 3/2/4; Hist. MSS Wi 8/i/i Nkani I passim.
14. Robert Moffat, J7ournals, I, 369; F. G. Elliott A 2.I8/28. In this case the Mutinhima faction admitted their subordination to Chirisamhuru's house.
15. Hist. MSS Wi 8/I/2 Ntabeni I-2. 
16. Stuart in A 3/I8/28.
17. Hist. MSS Wi 8/i/i Ngungu 2; T. M. Thomas, Eleven Years in Central South Africa (London, i873), I65; S. N. G. Jackson in NB 6/i/i; 'Mziki' (pseudonym of A. A. Campbell), 'Mlimo (Pietermaritzburg, I926), 85. This latter account, which existed in draft form in I905 and in final form in I9II, has been described as a fabrication. But the presence of information about Lukuluba and Chizema in particular, and about the affairs of the eastern Ndebele in general, suggests that it is indeed a body of genuine tradition collected in Insiza I897-I905 and added to a fictional romantic story.
18. Hist. MSS Wi 8/I/2 Ntabeni 4.
19. Hist. MSS Wi 8/i/iI Nkani I-2.
20. G. A. Taylor, 'The Matabele Headring', NADA, 3 (1925), 4I.
21. Robert Moffat, Yournals, I, 2I4-I5, 224, 265; II, I58.
22. F. G. Elliott in A 3/I8/28.
23. F. W. T. Posselt, Mambo, 6; Stuart in A 3/I8/28; Lloyd and Muhlanga, 'Mbava and others', 92.
24. N 3/33/8 E. Morris, N. C. Marandellas to Acting C. N. C. Salisbury, I Jan. 1904.
25. W. Edwards, 'The Wanoe', NADA, 4 (I926), I8. Edwards gives Nyandoro's Fungwe origins correctly (Morris, in N 3/33/8), but omits the period in which he ruled Tsunga under the Rozvi.
26. Robert Moffat, Journals, I, 233, 240, 369. Moffat originally used the term 'Shona' to mean, principally, 'Rozvi', and 'Bamakalaka' to mean 'Kalanga'.
27. Ibid., I, 319. 
28. Ibid., II, 59.
29. Leask, Diaries, 81, 88.
30. The Journal of Carl Matuch I869-I872, ed. E. E. Burke (Salisbury, 1969), 222.
31. Ibid., 209-214.
32. E.g. A i/o/i, H. Paulet to L. S. Jameson, 28 July I892.
33. F. C. Selous, Travel and Adventure in South East Africa (London, I893), II6.
34. Baines, Diaries, I, I63.
35. Robert Moffat, Journals, I, 224, 234, 241, 250, 36I; II ,79-80, 104.
36. Mudenge, 'The Rozvi Empire', 117-I8. Zumbo sent two guns every three years.
37. Robert Moffat, Yournals, I, 234; II, 79, 104; John Moffat, Matabele Mission, I 52, i6 i.
38. Selous, Adventure, 474. 
39. Baines, Diaries, II, 469.
40. Robert Moffat, Yournals, I 243.
41. E. C. Tabler, The Far Interior (Cape Town, 1955), 212.
42. T. M. Thomas in A 3/i8/28.
43. F. G. Elliott and Stuart in A 3/i8/28; F. W. T. Posselt, Mambo, 6.
44. Stuart and T. M. Thomas in A 3/i8/28; M\IA Delineation Reports, Que Que and Gatooma, 1963-5.
45. D. J. van der Merwe, 'Some history of the Vakaranga in the Gutu Reserve', NADA, I4 (1936-7), 73. The episode described immediately preceded the Jiri Rozvi move to Bikita in the late nineteenth century; Marwodzi, 'The Barozwi', go. By 'this country of the Batonga' he meant the upper Sabi valley; for Hwata, see below.
46. Robert Moffat, Journals, 1, 233, 24I, 250, 368-9; Campbell, 'Mlimo, 103.
47. S. N. G. Jackson in NB 6/i/i.
48. Leask, Diaries, 114.
49. NB 6/i/i W. T. T. Driver, N. C. Selukwe to C. N. C. Bulawayo, 3I Mar. I898; I am indebted to Mr R. G. Mwanza for his tradition confirming and explaining Zingwe's death. 
50. S. A. Dornan, 'Rhodesian Ruins and Native Tradition', S.A.J. Science, xii, 2 (I9I6), 508.
51. I am indebted to Mr A. Samasuwo for his translation of this Rozvi dialect word.
52. Rukara in Hist. MSS Misc/RU 4/I/I; MIA Delineation Report, Bikita, I964; Robert Moffat, Journals, I, 369; J. Chapman, Travels in the Interior of South Africa 1849-1863 (Cape Town, i868), I, 63. I am indebted to Mr J. D. Cobbing for this latter reference.
53. Robert Moffat, Journals, II, 59.
54. D. N. Beach, 'The Rising in South-western Mashonaland, 1896-7', unpubl. Ph.D. thesis (London, I971), 205.
55. Univ. of Rhodesia, Hist. Dept., Text 78 Bha; MIA Delineation Report, Buhera, 1965.
56. Izidoro Correia Pereira, 'Mappa das minas conhecidas na distrito de Senna', 31 July I857, in Memoria e Documentos acerca dos Direitos de Portugal aos territorios de Machona e Nyassa I890 (Lisbon, I890), 296. 'Mezircase' could be a variant of the Hwata praisename mufakose.
57. P. S. Garlake, 'Excavations at the Seventeenth-century Portuguese site of Dambarare, Rhodesia', Proc. Trans. Rhod. Scientific Assn. LIV, I (I969).
58. Mauch, Journal, 220; P(ublic) R(ecord) O(ffice), London, CO 417(I4) South Africa 1887, II, Mandy to Jones, 8 Feb. I887, 46. I am indebted to Dr N. M. B. Bhebe for this reference, and for having drawn my attention to the importance of Hwata's trade system.
59. Hist. MSS 6/z/1 29 Dec. I890.
60. P. S. Garlake, 'Seventeenth-century Portuguese earthworks in Rhodesia', S.A. Archael. Bull. xxi, 84 (I966).
61. Pereira, 'Mappa das minas', 296.
62. Baines, Journals, II, 472-5.
63. Chapman, Travels ii, 161.
64. Campbell, 'Mlimo, 139.
65. John Moffat, Matabele Mission, 136-7, 152; T. M. Thomas, Eleven Years, 3I4. I am indebted to Mr J. D. White for his help on this point.
66. Campbell, 'Mlimo, 137.
67. Mauch, Journal, 176-8.
68. John Moffat, Matabele Mission, 152.
69. Beach, 'Adendorff', 38-9; Driver in NB 6/i/i.
70. Pereira, 'Mappa das minas', 296.
71. MIA Delineation Report, Bikita, I964; Fortune, 'A Rozvi text', 73.
72. Leask, Diaries, IO2; T. M. Thomas, Eleven Years, 339-40.
73. MIA Delineation Report, Bikita, I964.
74. Driver in NB 6/i/i; C. G. Oates, Matabeleland and the Victoria Falls (London, 1881), 59; Campbell, 'Mlimo, I5I; Hist. MSS Wi 8/i/2 Ntabeni 53.
75. Baines, Diaries, II, 413.
76. See 57.
77. N 3/33/8, W. Edwards, N. C. Mrewa to Acting C. N. C. Salisbury, ii Dec. I903, encl. 'A Short History of Mangwendi's People, I3 Apr. I898'.
78. Edwards, 'Wanoe', i8. 
79. Morris, in N 3/33/8.
80. Morris and Edwards in N 3/33/8. 
81. Edwards in N 3/33/8.
82. Leask, 74, 221.
83. Pereira, 'Mappa das minas', 296.
84. Large-scale ivory shooting by European hunters in these valleys began in 1865 (Tabler, Far Interior, 27I), and by I884 African gunmen had decimated the game there (Kerr, The Far Interior, I, 44).
85. N 3/33/8 S. N. G. Jackson, N. C. Hartley to Acting C. N. C. Salisbury, c. 1 Jan. 1904; PRO, FO I79/279 No. 57, 'Memorandum on the Rights of Portugal in the territories to the south of the Zambezi, communicated by M. de Freitas,' I2 July 1890; Baines, Diaries, 1I, 498. 
86. S. N. G. Jackson in N 3/33/8. 
87. Leask, Diaries, 86.
88. Baines, Diaries, II, 498. T. M. Thomas (Eleven Years, 325) incorrectly made the date 1863, but Baines was closer to the event in time and place, as well as being backed up by John Moffat. 
89. Baines, Diaries, I, XXXVII, 55, 498; Selous, Adventure, 47, 295.
90. Mandy, in PRO, CO 417(14) South Africa I887, II.
91. On tribute, see Beach, 'Rising', 126, and N. M. B. Bhebe, 'Christian Missions in Matabeleland, 1859-1923', unpubl. Ph.D. thesis (London, 1972).
92. Beach, 'Rising', 124-35.
93. John Moffat, Matabele Mission, 152; Leask, Diaries, 69, I90.
94. Beach, 'Rising', 143-9.
95. Hist. MSS LO 6/I/4, J. Cockin to Mullen, May I879.
96. Beach, 'Adendorff', 39-40; the I877 date for the war given in my earlier articles has been revised by Hist. MSS TH 2/I/I, for which reference I am indebted to Mr J. D. Cobbing.
97. N 9/I/6, Statistical Report for year ending 31 Mar. 1900, Gutu.
98. Beach, 'Rising', 155-7.
99. Selous, Adventure, 113-I6, 465-6.
100. J. Chidziva, 'History of the vaShawasha', NADA, ix, i (I964), 29.
101. Selous, Adventure, 50-3.
102. Fripp and Hiller, Gold and the Gospel, 32, 34.
103. Driver in NB 6/i/i; Hist. MSS, Wi 8/i/2, Ntabeni 53-5. Gwasagwasa also visited Chipuriro in I892 (A i/9/i, Lendy to Jameson, i Apr. I893), probably to investigate the removal of Chipuriro by the Portuguese (L. Decle, Three Years in Savage Africa [London, I898], 2I7).
104. Fripp and Hiller, Gold and the Gospel, 27-30, 51, 69; Decle, Three Years, 141.
105. D. N. Beach, 'Kaguvi and Fort Mhondoro', Rhodesiana, 27 (1972), 38 n.
106. Fripp and Hiller, Gold and the Gospel, 27-8, 3I, 65, 67; Kerr, Far Interior, i, 64, 151.
107. Memoria e Documentos, 27-8, 268-76, 333; Boletim Oficial... de Mofambique, 51, 2I Dec. I889, 725; B.O.M. I2, Mar. I890, I52; B.O.M. 20, 17 May I890, Caetano de Carvalho Montez, 'Documentos para a historia de Mozambique', Mojambique, documentatio trimestral, 25, I94I, 86, 98-iii; J. C. Paiva de Andrada to Minister of Marine, 18 Aug. I889, Arquivo do Ministerio dos Negocios Estrangeiros, Lisbon, Soberania de Portugal na Zambezia, Caixa 2, I889; J. C. Paiva de Andrada to Francisco Costa, 15 and 27 Dec. I889, Arquivo Hist6rico Ultramarino, Lisbon, Mo4ambique, 2a. Repartiio, 6a. Caixa I889; Freitas, op. cit.; Beach, 'Rising', I74-97.
108. Mauch, Journal, 213.
109. CT i/i16, C. Lendy to A. R. Colquhoun, 18 Dec. 1890.
110. Memoria e Documentos, 27i-6; University of Rhodesia, Hist. Dept. Texts 53-4, 6o-i Hty.
111. CT I/I2/8, A. R. Colquhoun to F. Rutherfoord Harris, io Dec. I890; CT I/5/2, R. G. Nicholson to F. Rutherfoord Harris, 7 Nov. I890.
112. LO 5/2/3, Lt.-Col. Pennefather to F. Rutherfoord Harris, 4 Sept. 1890; URHD Text 26 Czi; N 9/4/I3, N. C. Chilimanzi to C. N. C., I Nov. 1902.
113. URHD Text 26 Czi; Hist. MSS WI 5/I/I, 7 Apr. I897.
114. A 2/8/I, L. S. Jameson to Acting Secy. BSAC Cape Town, 2 Dec. I89I.
115. DV 5/I/I, Chaplin to L. S. Jameson, 25 Mar. I892; CT I/I5/2, L. S. Jameson to Acting Secy. BSAC Cape Town, I5 July I892; A I/9/I, telegraphic conversation A. 19 Aug.-Sept. I892; NB I/I/9, A. N. C. Gutu to C. N. C., 9 Sept. 1900.
116. Houghton Library, Harvard, Diary of W. L. Thompson, 29-3I Aug. I892. I am indebted to Dr J. K. Rennie for this reference.
117. Hist. MSS BE 2/I/I, Diary of Tshakoma mission, 1st quarter, 1887.
118. Beach, 'Adendorff', 40, 43.
119. LO 5/2/2I, F. R. Harris to London Board of BSAC, 27 July I892.
120. A 2/1/4, H. M. Hole to J. W. Colenbrander, 5 Sept. I892.
121. Beach, 'Rising', 230-5.
122. E. G. Howman and J. H. Stanley Adam, 'The hand of glory', NADA, 6 (1928), 33-4; Campbell, 'Mlimo, -I; CT 1/14/I, C. Lendy to L. S. Jameson, 18 June 1893.
123. Hist. MSS CO 4/I/I, B. Dawson to J. W. Colenbrander, 24 Sept. I893.
124. Hist. MSS MO 14/2/I, 8 and io Oct. I893; Hist. MSS WE 3/2/3; Hist. MSS WI 9/2/4, 16 Oct. 1893.
125. CT 1/14/2, F. R. Harris to H. Loch, 28 Oct. I893.
126. W. A. Wills and L. T. Collingridge, The Downfall of Lobengula (London, 1894), 107.
127. DV i/i/i, Hay and Crowne to F. R. Harris, 7 Mar. I894; DV 7/2/2, trial of H. D. Hay, 14 Dec. I893; L 2/3/8, Vigers to Duncan, 7 Dec. 1893.
128. N I/1/3, N. C. Hartley to C. N. C., I9 Feb. and I5 July I895; N I/I/2, N. C. Charter to C. N. C., 30 Jan. and I9 Feb. I895; N I/I/I2, N. C. Victoria to C. N. C., 2 and 16 Oct. I894; A IS/I/I, N. C. Victoria's report to 10 Dec. I894.

Read more...

Languages of Southern Rhodesia in 1917

THE LANGUAGE OF SOUTHERN RHODESIA

By ALICE WERNER, Lecturer in Swahili and Bantu Languages

Abstract of a public lecture given by Miss Werner at the School on February 21st, 1917.

THOUGH the development of the mining industry consequent on European occupation has brought natives of several different tribes into Southern Rhodesia, there appears to be virtually but one indigenous language in this territory. The Zulu spoken by the Matebele and the Sesuto of Sebituane's people, who settled on the Zambezi about the middle of last century, are comparatively recent intrusions.

This language is sometimes called "Mashona", "Chino" , "Chiswina", apd "Chizwina" (sw and zw appear to be attempts at the peculiar labial sibilant also found in Ronga and the other Delagoa Bay dialects); but none of these names are to be recommended, The people called "Mashona" or "Maswina" strongly object to these appellations, which seem to be opprobrious terms applied to them by their Matebele conquerors. Moreover,

the name Chiswina seems to be confined to the Salisbury District, while elsewhere there are countless local designations: Chimanyika (Umtali), Chiungwe (Rusapi), Chikaranga (Victoria), Chirozwi, etc. (Chi- is the prefix indicating "language ", corresponding to Se- in "Sesuto ", Lu- in " Luganda", etc.)

The most satisfactory name would seem to be " Karanga". Mrs. C. S. Louw (Manual of the Chikaranga Language, Bulawayo, 1915), says: "In the district of Victoria "-in the south-east, on the head-waters of the Sabi-" the natives call it Chikaranga ". But there seems reason to follow Father Torrend in taking this name to include all the sub-tribes. The Bechwana, when they came in contact with these people in the north-west, called them " Makalaka" (= Makalanga or Makaranga), and the fact that the Great Zimbabwe, which was the King's Kraal, or chief tribal centre, lies in the Victoria district seems to indicate that their headquarters were formerly here. It was the Paramount Chief of the Makaranga who was vaguely known to sixteenth century geographers as the " Emperor of Monomotapa"

Karanga is so nearly akin to Nyanja (which covers a large area north of the Zambezi) that the differences may be set down as merely dialectical. It includes Chindau (spoken in Gazaland, just east of the Portuguese border), of which Mr. Daniel Jones has made a careful phonetic study.1 Its affinities appear to be rather with the languages north of the Zambezi than with the more southern ones; it has not, like Zulu, adopted the suffixed locative, or the Hottentot clicks, and the "laterals" (usually written, in Zulu and Sesuto, hl, dhl, ti, tlh) occur, if at all, only to a limited extent. But this point cannot be determined without a full examination of the language by an exact phonetician. The three books before me, that of the Rev. H. Buck (Penhalonga), of Father Biehler, S.J. (Chishawasha, Salisbury District), and Mrs. Louw (Victoria) not only deal with different parts of the country but use different systems of orthography, so that it is impossible to gather from them what variations of sound are due to dialectical divergence, or, in some cases, exactly what sounds are intended. Mrs. Louw's book is by far the most satisfactory from a phonetic point of view.

East of the area covered by Karanga comes the Delagoa Bay group of languages included by M. Junod under the name Thonga (so written to mark the aspirated t and also to distinguish them from the Tonga ( = Chopi) of Inhambane, the Tonga ( = Toka) of the Middle Zambezi, and the Tonga of West Nyasa). They are more nearly allied to Zulu than is any other language, except the Xosa of Cape Colony; but they also have links with Karanga (notably the labial sibilant already mentioned) and with the languages beyond them to the north.

One interesting point about Karanga, which deserves further investigation, is the occurrence of words identical or cognate with Swahili forms, which seem to be absent from intervening languages. Thus we have sona= "sew" (Swahili shona, but Nyanja tsoka, soca, Yao tota); tswimbo, "stick"; dikiti, " melon"; ganda, " skin," "husk," etc.; and others. This is not the place to attempt giving a, complete list, still less to draw any inferences

from this phenomenon.

A different interest attaches to a few stray Arabic words. Ndarama, "gold," "money," is evidently (as pointed out long ago by Meinhof) from dirham, and must have spread in both directions from the Arab settlements on the Lower Zambezi, as it is also found in Nyanja. (Here it means either with the addition of "white", "silver ", with the addition of " red ", "gold ", or simply "money ". Curiously enough, it is not used in Swahili.) 

Mari, "money," " property"= (arabic symbol used) , is also found in Zulu as i-mali and in Swahili as mali. It may be a question whether the Zulus borrowed it from the Makaranga or vice versa; it might have reached the former through the medium of the coast tribes, though M. Junod says the Thonga have taken it from Zulu. An interesting point in the latter language is the occurrence of the forms isi-kati, "time, "um-kati, "point of time." The Swahili wa-kati is so evidently derived from "(arabic symbols used)', the initial (arabic symbol) being treated as a prefix, that one is tempted to think of the Zulu words as having the same origin (prefixes being variable); but it might be possible to establish a connexion with the Bantu root KATI, "middle," "interval," or "space between ". It seems to be accepted that, in primitive languages, conceptions of space precede conceptions of time.

Notes

1. The Pronunciation and Orthography of the Chindau Language. University of 

London Press, 1911.

Read more...

Matabele Toys collected in 1917

Below are some Matabele toys collected in 1917 from the land of the Ndebele in the then Rhodesia by Matthew Park. They were on exhibition at Glasgow Art Gallery in Scotland, UK


Matabele Toys, Zimbabwe, (rhino; woman; giraffe; leopard).



Read more...

Ndebele Amabutho or Ndebele Military Structure

The Evolution of Ndebele Amabutho

Author: Julian Cobbing

Source: The Journal of African History, Vol. 15, No. 4 (1974), pp. 607-631

Published by: Cambridge University Press

IT has become a well-established cliche to say that the Zulu state and its Nguni off-shoots, such as the Ndebele of Mzilikazi and his son, Lobengula, were 'military states', and that at the root of their social and political organizations lay the 'regimental system'. It comes therefore as a surprise to discover that rarely, if at all, has the 'regimental system' of an Nguni state been examined in detail. Barnes, for example, in his study of the Fort Jameson Ngoni, describes 'regiments' as military age-sets which are peripheral to the main political structure of the Ngoni state. As with the Ndebele, regimental age-sets cut across the 'segmentary' structure of Ngoni society; yet one is not told what happened to the 'regiments' when they grew collectively older, nor whether they made any lasting impact upon the political and residential structure of the state.2 Similarly in Zulu historiography there is no adequate manual of how the Zulu 'regimental system' worked. Judging from many existing accounts, indeed, Zulu society was unworkable.3

The same is true for the Rhodesian Ndebele who crossed the Drakensberg into the country north of the Vaal River in about 1822 and passed on northwards into modern Matabeleland (western Rhodesia) between 1837 and 184I. The student fresh to the subject of the Ndebele finds himself informed that the Ndebele state in the Matopos region was highly centralized and consisted of a royal capital surrounded by a relatively small

number of 'regimental' towns, each with a population of at least one thousand people and often much more.4 They were 'colossal strongholds' with a purely military raison d'etre.5 Each was 'commanded' by a chief(induna) who was 'appointed' by the king and who was an 'official' rather than the representative of an important local lineage.6 'Regiments' were grouped into four 'divisions', each with its superior commander.7 At this and every other point in the state hierarchy, the political function was overlaid by the military one, so that state and army were inseparable. The kingdom expanded via the creation of new regiments formed every so often to absorb a new generation of youths, to which were added captives taken in the annual raids upon adjacent Shona tribes. Settlement was thus everywhere uniform, the whole one 'warlike aggregation [rather] than a cultural unit'.8 Early descriptions, especially, were invariably accompanied by generalizations about the sanguinary nature of the state and of its militaristic objectives.9

This picture at first seems plausible. But one soon discovers that it departs in essentials from the historical evidence available. For example, the majority of Ndebele settlements in the pre-colonial era were smallscale, probably averaging from about fifty to two hundred people. Villages were collected together into clusters10 and the outlying political unit of the state was not the 'regimental' town, but the partially decentralized chieftaincy or isigaba (lit: sub-division), which contained several villages. One family supplied a succession of chiefs within an isigaba according to strict laws of patrilineality, which even the king was rarely able to disturb.11 The emergence of the isigaba was inseparably linked with the formation of 'regiments'. It is hoped that the following account illustrates the simplicity of this development whilst accommodating the main historical facts.



Ndebele Settlement 19th Century



The Evolution of the Typical Ibutho

Ndebele settlements not only moved physically from time to time, but were simultaneously in a state of evolution. This can best be understood by comparing the concepts of the umusi (pl. imisi: kraal, village) with that of the ibutho (pl. amabutho: 'regiment'). An umusi was a residential unit containing families of men, women and children, and was essentially concerned with non-military activities, whereas an ibutho was a concept with a specifically military and masculine sense. Generically, the term ibutho meant men gathered together for martial functions; but it was used to designate three different, though related ideas. In the first place it was a theoretical concept defining men from a given group of imisi, who, if called out for military duty, would be called out together. This reflected a relationship between men from imisi which usually had a common origin or geographical location, or both. An ibutho was, secondly, an actual squad of men assembled from the adult male inhabitants of imisi, to which they returned when the specific operation was completed. The Ndebele army was thus essentially temporary and melted away after combat. Thirdly, an ibutho was young men (amajaha) grouped together (ababebuthwa) in separate and temporary settlements of varying sizes for both offensive and defensive fighting purposes. Only in this last sense did ibutho approximate

to the British idea of regiment, but the differences in concept are so enormous as to make the translation misleading. The thing to note, however, is that imisi-the normal residential units of the Ndebele kingdom- were not purely military formations as in the third sense of the word ibutho. An umusi only had a military function in so far as its menfolk were potential soldiers. But in this respect the Ndebele did not differ from their neighbours such as the Shona and the Tswana.

Hughes and Omer-Cooper have already suggested that an ibutho in the third sense could transform itself over a number of years into an umusi, or, more probably, a cluster of imisi.12 In doing so the character of the settlement underwent a metamorphosis, and its purely military function evanesced. The military potential of the kingdom would come to reside in newly-created amabutho, into which the children of existing imisi might be drafted. This process explains, firstly, the perpetuation of settlements over several generations (age-set units disappear); and, secondly, the internal settlement proliferation of the state. Only in the most general sense was it a form of segmentation.13 It is better described as a process of unplanned internal self-regeneration, at once both politically and militarily expedient, which reversed any tendency towards in-breeding of close kin: a tendency already emphasized by the ban on marriage between people of the same isibongo (praise-name). The process, finally, led to the social advancement of brave warriors (abaqawe) and the permanent elevation of new aristocratic families.14

This process of settlement evolution is pivotal to an understanding of the structure of the Ndebele state, and so warrants a more detailed analysis. The Ndebele youth was brought up to be cattle-herder, hunter and soldier functions which were obviously linked. Depending on his promise, but usually at about the age of ten, the Ndebele child would begin to help in the herding of his father's cattle. He would follow the cattle into the winter veldt, mix with his elders and learn the rigours of life.15 At puberty, an initiation ceremony, the tomba, was performed: a medicine called izembe was burnt, blown into the boy's mouth through a reed, and official adulthood was reached.16 The youth would not however make the next transition, that of entry into an ibutho, until he had had full sexual experience, for otherwise the protective medicines (intelezi) of the izanusi(doctors) would not work.17

The raising of amabutho in the Ndebele state was an irregular occurrence, depending firstly on the particular political or fighting requirements of the moment, and secondly on the availability of sufficient numbers of adult males. The examples which follow are taken from the raising of Imbizo, Insuga and Umcijo. Imbizo was raised during the civil war crisis between i870 and I872, the first two years of Lobengula's reign, and only two years after Mzilikazi had raised Inqobo. The raising of Insuga which followed in c. I883, a gap of about twelve years, was more clearly connected with the availability of recruits, but was also probably linked with the wars against the Tawana of I883-5. And the drafting of Umcijo and another ibutho, the Ihlati, in I890 followed by only approximately two years the raising of Isiziba. This outburst of ibutho-raising was almost solely the result of the European invasion of Mashonaland.

Ntabeni Khumalo, giving us our first example, was born at Dukadeni in the western Malungwane and in I87I was put into Imbizo.18 The recruits trained at the Nkantor Mountain in the central Matopos for two years before establishing a settlement on the Koce River.

We first went by ourselves and built the kraal, and then our mothers came. We cooked our own food whilst we built the kraal, and when our mothers came they cooked for us. Our sisters also came. Our fathers stayed behind. Some younger women stayed behind to look after our fathers.19

And a contemporary observer, Thomas Baines, wrote in August 1871,

giving a different emphasis:

The king is forming a new regiment called La Besa or 'the called'. They are lads of 10 to 18 years of age . . . all removed from their mammas and other softening influences, that they may grow up as hardy and daring as possible. He sent them out twice to hunt.20

Ginyalitshe Hlabangana told a similar story of being drafted into Insuga in the mid-1880s.21 Ginyalitshe was born at Amangubeni (Inqobo), which was near Inyati, and for many years he herded cattle there for his father, Sigozwana. After Sigozwana's death he went to stay at the umusi zamathanga (private village) of his grandfather, Dlamlomo, which was not far from Inqobo. One day Lobengula's messenger came to Inqobo, and the induna sent round the imisi under his control and collected a group of youths amongst whom was Ginyalitshe. They went to Bulawayo, were, with others, collectively named Insuga ('those who depart in the day') and sent to the upper Nata River. 'Today', Lobengula informed them, underlining their rise to independence and maturity, 'you leave your fathers and you have your own kraal'. They stayed in the west of the kingdom for about six years before being transferred to the upper Gwelo in 1890.

Ndansi Khumalo, lastly, tells of being drafted into Umcijo.22

We spent about a month singing and dancing near the King's kraal; he used to give us lots of food and we were well looked after. We brought our own outfit, our skins and shields. Our parents provided our spears.

During their stay at Bulawayo the recruits bathed in a specially treated pool in the Umgusa River to counteract any evil which might level itself against them. They were then sent to the Shangani River to guard the king's cattle. Each pair of amajaha (unmarried soldiers) built cattle shelters and looked after from five to ten head, having the use of the milk. They learnt to fight, and were posted to the Gwelo River in March i892.23

A newly-formed ibutho, consisting of young soldiers-to-be under the leadership of a mature warrior, would be 'butha'd' in a particular isigaba (induna's area), sometimes moving from isigaba to isigaba for particular purposes. Dr Smith noted that 'they then only take their war-dresses and implements. The karosses they leave behind them saying that the people of the kraal whither they go must find them in karosses'.24 They had an internal law-and-order function; and the local induna could call upon the amajaha to make themselves useful around the king's villages, cutting poles and grass for huts or being utilized for building and repair works.25 Moving around continually, the amajaha would erect temporary grass 'loobas' or shelters at night, with special lodging for any attendant women.26 An ibutho would usually divide into a number of encampments or groups termed isigatshana, each training and hunting under a separate leader.27 As time went by, a young ibutho would be mixed with maturer forces for raiding experience. After a few years it would become an elite group of soldiers, who enshrined in themselves the major military potential of the state.

Although it was not uncommon for men who were already married to be drafted into an ibutho,28 the decisive stage in its evolution into a number of settled imisi occurred when the amajaha were officially encouraged to marry, between five and ten years after formation. During this period the ibutho remained closely connected with the imisi which had supplied its recruits, irrespective of where it was posted. Ginyalitshe even emphasized: 'I continued to live at Dlamlomo's kraal after I became a member of the Insuga regiment. The king did not separate the young men from their homes'; and, 'it did not mean that we could not go back and live with our fathers if we wanted to. It was not a prohibition.'29 Ginyalitshe seems here to be referring to periodical visits home, which most frequently took place during the summer months, the time of planting and of the rains. Thomas Morgan Thomas mentions on one occasion how 'the impiso [Imbizo] are scattered here and there visiting among their friends',30 a form of regular 'home-leave' called umusi wamaxhegu.31 During these constant contacts, a young man would perhaps take a sister, or even one of his father's wives, back to the ibutho. Love-making between the amajaha and the attendant females was a normal and condoned practice, and if pregnancies resulted, marriages were arranged. The rate of marriage would accelerate when the warriors were in their late twenties or early thirties, at this stage with the specific approval of the king.

Although Maund wrote in i885 that Lobengula 'has just permitted a military kraal to marry, because several hundred girls in it were reported to be in an interesting condition',32 it need not be thought that some huge marriage ceremony or a large number of simultaneous marriages took place.33 Marriages occurred over an extended period of time; and each was a matter for the families of the couple concerned. A man would leave his ibutho and return to his father's umusi for the ceremonies (bavulelwe). Ntabeni Khumalo at Imbizo recalled how 'when my wife came to me I paid lobolo: three beasts, two cows and an ox, to her father. We killed an ox for the occasion'.34 The bride would come with her attendants to her future father-in-law's umusi in a procession called the umtimba, after which dancing and feasting lasted several days. Shortly after his marriage Ntabeni returned to Imbizo and his wife remained at his father's umusi of Dukada. Later it is probable that his wife moved to Imbizo; and here they would settle as a family where the children would grow up. Women might even go to an ibutho in search of marriage. The induna Geletu remembered how a girl called Nhlombi had left his umusi to go and marry a man in Imbizo, the girl's relations moving with her.35 On another occasion the daughter of Jozana of Mbambanjeni ran away to her sweetheart, Simila, at Imbizo, after she had first been betrothed against her will to Labalala of Nxa.36

Although a percentage of wives would remain at the imisi of their fathers-in-law, the presence of favourite wives and their children would make the main centre of the ibutho static, and it would gradually transform itself into a main umusi with a number of smaller linked imisi. At an interim stage of development the settlement would still have some of the characteristics of an ibutho. The transition towards a greater demilitarization could be slowed, for example, by the drafting of new recruits, and even by the military training of the children of the first warriors. The first generation of children to grow up in the Imbizo imisi were collectively nick-named Litshelentaba ('stones off a mountain'); and by the early 1890s Imbizo-Litshelentaba, strengthened by new drafts from other imisi, was itself an ibutho. The children of later recruits were known as Imbizo-Nembe ('porridge for children'), but they were still too young for military service at the time of the collapse of the Ndebele state in the mid-1890s.37 Nevertheless, in spite of this generation of new military talent, Imbizweni became more static. Some of the older men returned to their original imisi, whilst others who had acquired cattle in warfare or otherwise distinguished themselves, established their imisi zamathanga in the vicinity of the main umusi.38 The original ibutho had become a new isigaba or chieftaincy.

How long it took for a full transition of loyalties to the new imisi to be made is difficult to judge. Imbizo for example was not sufficiently detached from its root imisi to be able to perpetuate itself into the period of the European conquest. After the burning down of the Imbizo main kraal in 1894 its people preferred to return home, and Imbizo became merely the umusi of its chief, Mjaan Khumalo.39 Amabutho formed later, such as Insuga, Isiziba and Ihlati, also disappeared in the period after I894. But Induba and Mhlahlanhlela (formed in the early i86os) and Inqobo (late i86os) did perpetuate themselves into the twentieth century as groupings with which Ndebele identify themselves today, even if far removed from the nineteenth-century sites. Examples of fully matured imisi which exist as concepts providing group identity today are: Ndinana (descendants living in the Lundi Tribal Trustland under chief Jobe Matshazi, of the main nineteenth-century line); Intunta (descendants either living near the old site of Intunta at Bushtick in the Essexvale district, or along the Lundi River in the Belingwe Tribal Trustland under chief Ngungumbane Mkwananzi of the main nineteenth-century line); and Nqama (descendants living under the traditional line of Mathema chiefs in Wenlock Tribal Trustland). All these, and others such as Ujinga, Matshetshe and Godhlwayo, are imisi which had become established by the I85os and sometimes before the migration of Mzilikazi northwards across the Limpopo in the late I830s. A minimum period for the establishment of group-identity into the era after the conquest is thus c. twenty-five years of independent existence before I893, although under less violent circumstances it would probably have been less.

How then did pre-existing imisi survive intact over several generations in the face of the periodic creation of new amabutho? Why did settlements not mature, age and die? The answer is that the perpetuation of settlements was ensured since only a percentage of young men in a given age group would go to form a new ibutho and many less fortunate youngsters would remain behind.40 Some would not be given the chance to enter, especially if there was a lengthy time-lag between two drafts. There were also exemptions, one informant for example mentioning that an only son would not be 'butha'd'. Other informants emphasized the permanent return of men from the ibutho to their fathers' villages, and their replacement by young men in a fresh draft. This exit from an ibutho occurred especially with men who had a political function to play in the imisi of origin, such as sons of chiefs. Thus Hliso Mkwananzi of Intunta was drafted into Induba in the i86os, but by the i88os he was back at Intunta, having assumed the chieftaincy in succession to his father, Mahubohubo.41 Also, according to Nqagwana Khumalo: 'When an old man died his eldest son would be withdrawn from his regiment so that he could look after his father's village and property.'42 Yet the reverse move could take place. When parents of a 'butha'd' son grew old, they could ask their induna to be allowed to move their home to where the son was living.43 But irrespective of whether an eldest son remained in his 'ibutho-umusi' or returned to his father's umusi following his father's death, he would inherit his father's property such as cattle.44 In this way the rise and maturation of amabutho did not break across the hereditary aspects of family life, and the functioning of existing settlements was only partially disrupted. If major imisi disappeared, it was not usually a result of a failure to regenerate, but because of deliberate dispersal for political reasons, as with Gibixegu in c. I84I-2 and Zwangendaba in I870.45

Since it is possible to date the sequence of creation of amabutho approximately, it is also possible to estimate the age or generation of informants or their ancestors if one knows the ibutho into which they were drafted. Mapunda Msiza, for example, was drafted into Iliba probably in the early I840s. His son, Sonjobo, was born at Iliba, but subsequently in the I870s was put into Imbizo; and Sonjobo's sons, Mpuqa, Matete and Mguqu, were born at Imbizo in the 1880s.46 Another example is that of Sikwebu Ncube, who was born at Iliba and drafted into Induba in the i86os. His daughter, Magoli, was born at Induba, and in the I870S went to marry Mukontshwana Tjili, a young Sotho who had gone into Mbuyaswe in about I869.41

In general, someone born into Iliba (formed probably in the I840s), is older than someone born into Inqobo or Imbizo (I867-77I).48 One can detect the relative age of brothers in the same way. For example, Bozongwana Khumalo's (presumably) eldest son, Gabeni, went into Imbizo, whilst his three younger sons, Mukenke, Mqhewa and Ntabeni went into Insuga.49 Finally, men of older-established imisi such as Godhlwayo would, over the generations, be present in virtually all newly-created amabutho.50 Theoretically, then, settlements evolved, and at any given moment in the last three-quarters of the nineteenth century settlements of different composition and 'stage' existed side by side. This assumption is largely derived from informants who, whilst themselves not able to explain the evolution, were nevertheless agreed that there was a generic difference between imisi (established since time immemorial) and amabutho (time of establishment remembered, if only vaguely). Tito Matshazi and Ntolwana Dhlodhlo, two particularly knowledgeable female informants, drew up the following lists51:

I. Imisi                                 2. Amabutho                                Approx. dateof formation

Godhlwayo                             Mahlokohloko                                   1849-50

Ndinana                                   Induba                                              1862-4

Mzinyatini                                Mhlahlanhlela                                     1863

Matshetshe                              Inqobo                                               1867-9

Nqama (etc.)                           Imbizo                                                1871

(all in existence by                    Insuga                                                1883-4

the early I840s)                        Isiziba                                                1887-8

                                               Umcijo                                                1890

                                                Ihlati                                                   1890

Finally, with the help especially of Mpuqa Msiza and Zimu Moyo, the following diagrammatical representation of settlement evolution was established in the sand of several villages52:

Diagrammatical Presentation of Ndebele settlement evolution

The Historiography of the 'Division' or 'Province'

Besides imisi and amabutho, several key writers on the Ndebele have described the 'division' or 'province'. There is, however, no Ndebele word for such a concept, and modern informants deny that they existed.53 Because of this conflict, a re-examination of the historiography of the European idea of the 'division-province' is necessary. The relevance to the argument will become clear in the next section.

The first hint of something higher than the induna in the Ndebele political hierarchy came from Andrew Smith, who divided Mzilikazi's chiefs into two grades: the numzan (umnumzana) and the tuna (induna), the former, according to Smith, being the senior. Smith went on to isolate certain 'big' chiefs (Kalipi, Gwabalanda, Mpondo and Mncumbata) who, he implied, held authority over large portions of Mzilikazi's territory.54 The missionary, Thomas, was more explicit:

The king divided the land into several provinces, these again were divided into smaller districts, and the districts into towns. Over each province, a very loyal subject was appointed to represent the sovereign.... This officer was called the induna enkulu (great under chief), and was next in rank to the king.55

Most writers have cited Thomas to show firstly, that the province existed as a formal administrative sub-division of the kingdom, and secondly that there was a hierarchy of appointment with an induna enkulu being an official senior in rank to an induna, the former ruling the 'province', and the latter the 'district'. But both Smith's and Thomas's terms were applied inexactly. Smith's umnumzana was simply the head of an umusi, even though it was a title of respect. If anything, the umnumzana was inferior to the induna, although one man could be both.56 Similarly the induna enkulu was just an important chief. This carried implications of power and prestige, but not of officially appointed rank.

In 1885 Maund (who had read Thomas) added a military meaning to both province, which became 'division', and induna enkulu, who became primarily a military figure. It is a military framework from which Ndebele historiography has never escaped.57 'The country', wrote Maund, 'is divided into four divisions [Amabuto, Amagapa, Imhlope and Likanda], which really constitute the four great territorial divisions of the army'. Each division embraced a number of kraals, some of which were regimental headquarters headed by 'indunas, or officers of regiment in the field'. The system was not only purely military but highly centralized, as a glance at Maund's illustrative diagram shows:



Ndebele Hierarchy



And to underline the organizational interlocking of army and state, Maund added:

When an impi goes out it is in two divisions under two Indunas acting under a selected general. They march in columns. One from each territorial division and another consisting of the king's men, i.e. Embezu [Imbizo] and his own town [Bulawayo].

Apart from the general meaninglessness of the diagram, and the improbability of such a precise army organization, the important things to note are that Maund was a soldier who at this point had known the Ndebele for a fortnight, who came from the England of the generation of the Cardwell army reforms, and believed that militarily the Ndebele were 'similar to the Zulus', whose own degree of militarization was being vastly exaggerated during the Zulu war of I879. There is in fact no evidence to support Maund's theories as to Ndebele military divisions. Yet his analysis of the Ndebele became gospel, and has done more to influence Ndebele studies than any other piece of writing. For example, in 1898 the Matabeleland

Administrator, Captain Lawley, informed the assembled Ndebele izinduna about Maund rather than about themselves. 'As you know', he said, 'in the days of Lobengula there were five divisions which were each governed by five [sic] chiefs'.58 By 1919 Posselt had 'confirmed' the idea of the province as a division. Each head of division was now a 'divisional commander', and each division was divided into 'regiments' just like the British army.59

The army analogy developed alarmingly after the Second World War. In a celebrated article the archaeologist Roger Summers elaborated on Maund's diagram, which now looked like this60:



Ndebele Hierarchy



Maund's influence is plain, although the degree of militarization has increased. The king is now commander-in-chief; there is a commander at 'H.Q. Kraal' (presumably Bulawayo); and now there is a 'Great Council' (the isikulu) to advise the king. In the context it can only be a military council. Otherwise Summers's diagram is as unworkable and as detached from any evidence as is Maund's. In their publication, The Warriors, in 1970, Summers and Pagden threw all reservations aside and described the Ndebele kingdom in purely military terms. It consisted of regimental components of an army inseparable from the state.62

The anthropologist Hughes, although a collaborator of Summers, was more cautious. He renounced the term 'division' in favour of 'province' again; and whilst believing that the provinces were territorial entities with clearly demarcated boundaries, recognized that 'in one sense provinces were primarily divisions of the people not of territory'.63 And, the first writer to hint that provinces were not administrative sub-divisions of the state, but rather evolutionary antecedents of the later nineteenth-century imisi, Hughes noted that 'the four provinces formed administrative segments of the nation, and . . . the regiments formed by Lobengula comprised another segment of the same order of segmentation'.64 Nevertheless, the imisi were for Hughes still essentially military units; and although he thought that because of the administrative aspects of their function 'there might seem to be a good case for adopting some English term other than "regiment" to describe them [amabutho]', he did not do so on the ground that 'it would create unnecessary confusion to adopt another translation of this Ndebele word [ibutho] here'.65 And so the influence of the nineteenthcentury commentators still holds the field.

The original Amabutho and the development of settlement under Mzilikazi

This persistent historiographical 'militarization' of the structure of Ndebele society has obscured the true origin and structure of the 'province'. Present-day informants deny the existence of military or administrative divisions larger than the isigaba, or single chieftaincy, and register incomprehension when the English ideas of 'division', as used by Maund, or of 'province', as used by Thomas, are explained to them. Yet they are well-acquainted with the concepts of Amakanda, Amhlope, Amnyama (for Maund's Amabuto) and Igapha.66 A man from Nxa will, for example, use Nxa and Amakanda interchangeably, even though Amakanda refers hazily to a general area rather than to a specific point on the map. Another man will identify Nyamandhlovu with Igapha in the same way. What is the separation of the umusi Nxa from the larger but less precise concept of Amakanda? The answer is, as Ntabeni Khumalo pointed out to a surprised Windram (he had been looking for military divisions), that:

There were no groupings of the regiments under main divisions.... All the later regiments came from the original regiments, the Amhlope, Amakanda, the Amnyama and Igapha.... Although they had their own regiments, the later people used to regard themselves as the descendants of one or other of the older regiments. 67

The slender historical evidence that exists confirms that the 'divisions' were indeed early amabutho which later 'spawned' fresh amabutho, and in the process themselves disappeared. The initial close links between amabutho spawned in this way were increased by their future geographical proximity in the Matopos region. By the mid-nineteenth century however, the original proto-amabutho were becoming memories, and there had been a fusing of the various descent groups. The attempt to place all the later amabutho into neat lists according to 'division' is a reflection of the historian's mania for tidiness, and of the influence of Maund.68

Literature and tradition combine to produce the following chronology. Mzilikazi left Tshaka's country in the early 1825 with a small body of men divided into two major groups, Amakanda and Amhlope.69 By the late I820s Amakanda had sub-divided into smaller groups, among which were Nxa and Ndinana, and Amhlope had given rise for example to Isizinda, Magolosa and Dibinhlangu. These became the nuclei for the assimilation of Sotho peoples as the Ndebele passed through what is now the Transvaal. By the I830S other break-offs, new accretions or family groupings had emerged, among which were Insinda, Insinga, Nzwananzi

and Dhlodhlo, identifying themselves with Amakanda; and others such as Mfakuceba and Ntembusa associated with Amhlope.0

The third of the proto-amabutho was Amnyama (the black-haired ones, i.e. young men), who were regarded as the 'sons' of Amakanda-Amhlope.7 Mzilikazi was reputed to have formed Amnyama just before his quarrel with Tshaka; and we know of leaders of later Amnyama-identified imisi, such as Mahubohubo of Intunta and Dambisamahubo of Godhlwayo, who earned for themselves reputations as abaqawe in the stormy period of Mzilikazi's break-away.72 In the early I83os Amnyama gave rise probably first to Mzinyati, and a little later to Eyengo, Dukada, Matshetshe, Godhlwayo, Nkenenkene and Isiphezi.73 Several of the Amnyama and Amakanda settlements broke away from Mzilikazi after the Boer attack on Egabeni towards the end of I837, and, led by Gundwane Ndiweni and Majijili Gwebu, migrated across the Limpopo to the Malungwane Hills

APPROXIMATE CHRONOLOGICAL TABLE OF THE MAJOR NDEBELE AMABUTHO

Chronology of major Ndebele Amabutho

in what is now Rhodesia. Amnyama-connected imisi settled to the west of the Malungwane range along the valleys of the Ncema and upper Umzingwani, and Amakanda further to the east along the upper Insiza. Gundwane established a new capital at Gibixegu on the Umzingwani; but this was destroyed by Mzilikazi when he arrived from the Zambezi in c. 1841.

The fourth proto-ibutho was Igapha, the name coming from Mzilikazi's I835-7 capital of Egabeni, and its recruits from the Amakanda-Amhlope- Amnyama group already in existence.75 Igapha and several imi.si connected with Amhlope, such as Isizinda and Magolosa, followed Mzilikazi to the upper Nata and Gwaai valleys, after a migration taking them through Ngwato country to the lower Chobe and south again.76 North of the Matopos Igapha disintegrated into the off-shoot settlements of Amagogo, Nqama and Nyamandhlovu,77 groupings which probably bore the brunt of the fighting against Gundwane's Amakanda-Amnyama in the civil war of c. I842. Amhlope imisi tended to site themselves immediately to the north of the Matopos, whereas imisi descended from Igapha subsequently came to function as a defensive cordon reaching from the northern Matopos through the upper Gwaai, Khami and Umgusa Rivers to the Bembesi and Shangani in an umbrella-handle shaped crescent. Later the term Igapha (or Amagapha) referred to the collective idea of men from Igapha-descended imisi. It developed south-western and north-eastern facets, so that Ndebele spoke of the imisi of 'Igapha towards the Matopos' as Likamagqegenya or Linkgcweli, and of the imisi of 'Igapha towards Que Que' as Lamasandhla. Lamasandhla was in particular used to describe the artificial extension of Ndebele settlement to the Gwelo and Ingwenya Rivers at the time of the European occupation.78

Settlement now evolved in the Matopos region along the lines already suggested, but without precise reference to the proto-amabutho. Zwangendaba, Mbambanjeni, Ujinga, Iliba, Ilanga, Babamba and Hlambabaloi probably emerged in the I840s along the upper Umgusa and upper Bembesi valleys. In about I849-50 Mzilikazi established another ibutho on the watershed between the Bembesi and Umgusa called Mahlokohloko. This was both royal capital and the major ibutho of the 1850s, much as Igapha had been during the mid-i83os. By the end of the 1850s, however, Mzilikazi had moved again, this time about twenty miles to the north-east, where Inyati was formed on the Inkwegwesi.79 Induba was formed at some time during the early i86os, and Mhlahlanhlela-successor to Inyati as royal capital-in I863, possibly as a result of the Ndebele-Ngwato war of that year. Either in the 1850S or early 1860s Intemba was drafted under the leadership of Xukutwayo Mlotshwa, and sent to the extreme northeast of the kingdom between Inyati and the Mambo Hills.80

Lobengula's Amabutho

Lobengula raised, or began to raise, about a dozen new amabutho, concerning which there is happily a little more evidence. Inqobo ('the blanket') was raised at about the time of Mzilikazi's death in i868,81 and during the early part of I87I was moved to the Inkwegwesi River near Emhlangeni, replacing the more mature amabutho of Inyati and Intemba as the kingdom's most northerly outpost.82 Inqobo became a number of imisi under the leadership of Mazwi Gumede and Mtini Mpoko Ndhlovu. In I89I or early I892 a contingent from Inqobo was positioned on the Gwelo River to help defend the Ndebele kingdom from anticipated European incursions. Mazwi assumed authority over Rozvi groups on the watershed between the Gwelo and Lundi Rivers such as Lozani and Gambiza, both of whom went to him for land.83 Generally, Inqobo usurped

the position of Mbambanjeni, which, during the time of Mncumbata Khumalo, had for a time been sited along the upper Gwelo. In 1866 Mbambanjeni had moved south-west to the headwaters of the Shangani, but in 1892, nfow under Mncumbata's son, Mhlabi, was destroyed for political reasons.84

By far the most prestigious of Lobengula's amabutho was Imbizo, something of whose social evolution we have already examined. Its formation in I87I was a result of Lobengula's fears of an invasion by those Ndebele supporting 'Nkulumane' and Mangwane, his rivals for the kingship.85 Its first leader, according to Mahlangu, was Habimane Mthimkulu, but later Lobengula's second-cousin Mjaan Khumalo assumed control.86 In about June I87I Imbizo was probably involved in an attack upon a force of Macheng's Ngwato near the Tati River; and later that year participated in raids against the Karanga of Zimundu and Sibanda in the present Fort Victoria region. The cattle captured during these raids were given to Imbizo, which soon afterwards moved to the Koce River for better grazing, and perhaps to stand between Bulawayo, the new royal capital, and the still disaffected north-east.87

A whole generation of scions of important families was drafted into Imbizo during the first years of its formation.88 Imbizo took part in most of the major campaigns of the I870s, including the prolonged siege of Chivi's stronghold in October i879.89 It had the reputation of being particularly fearsome, and came in the end to challenge even Lobengula's authority. In about I874 men from Imbizo were attacked by a band from Godhlwayo in the storm created by the execution of the Godhlwayo chief, Mtikana Mafu.90 A decade later, in July I885, there was a symbolic battle of the generations between Imbizo and Mhlahlanhlela over the ownership of a captive (ifuyiweyo), in which about twenty men were killed or wounded.91 In October I890 J. S. Moffat wrote: 'the induna of Imbizo [Mjaan] is in trouble with his men, they want to get rid of him, he does not suit them.'92 Mjaan supported Lobengula's militarily unaggressive policy towards the British South Africa Company, whose forces had occupied Mashonaland without the king's permission. His men however

were more extreme, and made the lives of local whites a misery in an endeavour to bring about a military collision.93

Lobengula's next major ibutho was Insuga, which was formed in about 1883 during the time of the Tawana wars under the leadership of Manondwaan Tshabalala, and which had close connexions with the Inqobo- Ujinga complex of villages along the Inkwegwesi valley. Predictably, men of younger average age went into Insuga than had gone into Imbizo, and these included Nyamanda, Lobengula's second son and heir. One of Insuga's abalisa (headmen) was Sepalo Ndiweni; and the leader of one of its izigatshana, of which there were at least three, was Siatcha, the son of Matilingwane.4 In May I890 Insuga was transferred from the Nata River to the upper Gwelo as a result of the crisis caused by the entry of the pioneer column into Mashonaland.95 In I89I or I892 a body of men was detached from the main ibutho forming a sub-ibutho called Insizwa which had a separate umusi.96 Like Imbizo, Insuga was impatient to anticipate what was regarded as the inevitable attack from the Europeans in Mashonaland. Men from Insuga sang treasonable songs at the nxwala of 1892,97 and later that year they quarrelled over cattle with the Nxa people (upper Shangani), producing tensions which, White suggests, helped lead to their precipitate actions at Fort Victoria in July i893.98 Although many of warriors had married by I893, most of the families stayed behind in their home imisi in the troubled period preceding the war of I893. Having no roots, Insuga therefore disintegrated in the period after I894, and Manondwaan led a somewhat lone fight in the Gwelo region in the Rising of 1896. 99

Isiziba ('still waters in a river') was formed in about 1887,100 and its sub-chiefs were Nhlapo, and an unmarried son of Qaqa Ndiweni, leader of the Kwesincinyane isigatshana of Mhlahlanhlela.101 In I890 it was moved east to the Shangani River, after which it came under the overall command of the Insuga chief, Manondwaan Tshabalala. Isiziba participated in the raid on Nemakonde in September I89I, the July I892 raid on Chivi and the Shangani battle of October I893, following which it dissolved.102

Ihlati ('a forest of short bush'), referred to by Mahlangu as an isigatshana of Mhlahlanhlela, was raised during the crisis of May-June i89o, an event to which Colenbrander was probably alluding when he wrote:

'Lobengula is raising a new regiment consisting of two hundred to three hundred young men just returned from the diamond fields armed with Martini- Henry rifles which they have earned there.'103 Initially Ihlati appears to have trained under Somfulu Fuyane in the Nyamandhlovu area; but later it was mixed with Insuga and sent to the Ingwenya River under Matshe, the eldest son of Gambo Sithole.104 Matshe's brother, and Gambo's eventual successor, Dlomo, was in Ihlati, as also was Nyamanda's younger brother, Tshakalisa, who was about seventeen years old in i892.105 At about the same time, Umcijo ('a sharp end' as of a spear) was being raised in the Emhlangeni vicinity under Mlumbi.106 In March I892 Lobengula moved Umcijo to the Gwelo River, about twenty miles to the north-west of Ihlati, where it formed the extreme left-wing of the Ndebele defensive cordon.107

There are, lastly, occasional references to other of Lobengula's amabutho, though whether they were ever military settlements proper, or were simply villages of mainly non-Ndebele assimilants, is uncertain. One was

Amazibalongwe, also known as Edutjwa, which contained Kalanga and was led by Manyeu Ndiweni of Mhlahlanhlela. It was positioned somewhere on the Nata River, and, according to Posselt, took part in the attack on the Tawana in 1885.108 Others were Ngwangwa under Ngogode Thebe; Nkanyezi, on the Umgusa River not far from Imbizo, under Rubane Ngwenya (whom Posselt described as 'a Bechuana nobleman, who had tendered his allegiance to the king, and was thus rewarded with the royal confidence'); and Makupekupe, a predominantly Rozvi group in the eastern Malungwane which the king was reputed to have been organizing just before the war of i893.109 There was also a personal bodyguard of the king known as the Imbovane, 'the king's black ants', as Vaughan-Williams called them. These 'slave youngsters slept outside the king's stockade with the dogs' and attended the king wherever he travelled.110

Some conclusions

The critical reader might conclude from the foregoing: there are some interesting new facts here, and perhaps it is true that Maund overstated his case regarding the military structure of the Ndebele state. But the Ndebele state did evolve from military origins; the underlying structure of the state was still therefore military, and existing accounts have only to be modified in detail. There are several reasons, however, for going further than this. In the first place, it is clear from the above analysis that even if the word 'regiment' is retained as a translation for ibutho, only a very small proportion of Ndebele settlements at any given time were 'mobilized' amabutho. Even though men from the reproductive imisi continued to be called out to fight from time to time, the 'army', that is the fighting amabutho, was an organ (if that is not an over-exact term) within the state and subordinate to it. To continue to describe the Ndebele state as an army would be tantamount to replacing the word state in its various contexts by the word army. At almost every point in the Ndebele state structure a word redolent of military meaning has been misleadingly applied to what were essentially political phenomena. The umusi was a settlement concerned with the whole gamut of human activity rather than with military affairs alone; the chief was a political, judicial and even spiritual leader, rather than a military 'officer', a hereditary 'baron' as opposed to a royally appointed 'official'; the isikulu was an informal and ever-changing group of friends, relatives and advisers of the king with a political role, rather than a military 'council'; the 'division' was not something akin to a Prussian canton but a social grouping antecedent to later nineteenth-century settlement; the king was not simply a commander-in chief (Lobengula never led an army into battle) but an administrator and a ruler in the full sense of the words.

Secondly, a correct appreciation of amabutho evolution provides an understanding of how the state operated at other levels. For example, the manner in which political interests emerged away from the centre of the state can only be explained by reference to the evolution of amabutho. The majority of the great Ndebele families, such as the Sitholes of Amagogo, the Masukus of Matshetshe, the Matshazis of Ndinana or the Mafus of Godhlwayo, owed their rise to their identification with the creation and maturation of an ibutho. The process set up centrifugal tensions which, along with splits within the royal family, were at the root of the major crises the state faced in its short career (such as the civil wars of I841-2 and 1868-72). All Ndebele owed allegiance to the king; but their most proximate allegiance came to be to the chiefly family which held a stranglehold over power in the isigaba. Rivalries at high levels of the state were paralleled by similar struggles for power and position within the isigaba. The isigaba was the state writ small, a mini-state almost. As an ibutho matured, divided into imisi, occupied a definable geographical area for a generation or so, and experienced a succession of chiefs from the same family, it emerged as a rival power centre to the larger state. Paradoxically, the initial process of ibutho formation, partly a defence against the danger of state tensions (cf. the raising of Imbizo), led eventually to their magnification. Over-powerful chiefs such as Mtikana Mafu of Godhlwayo or Gambo Sithole of Amagogo were either killed or exiled.1"' Sooner or later one of them would have reacted as Mzilikazi did to Tshaka by seceding.

In the third place, the evolved ibutho was the nucleus around which the state grew. It was reproductive and therefore expanded in population, even though some of its male offspring went into new amabutho, and many of its young females married into other izigaba. The mature warrior (iqawe) with several wives, a village of his own (umusi zamathanga), and many children, occupied a central position in Ndebele society. His family grew

with the addition of captives (abafuyiweyo) taken in war, distributed to him by the king or by the chief and distributed by him amongst his wives. In an analogous way a brave man could become wealthy in cattle. His status within the isigaba increased if he married the daughter or other female relative of the chief. If he caught the king's eye he might be given the command of an ibutho (as Mhlumbi was in I890) and raise himself to chiefly rank. Only at certain points along this tortuous path was a man a purely military figure: his means may have been the exploitation of his martial abilities,112 but his aims were control of people, cattle and territory, the major standards by which others judged him.

Many of these processes are either incompatible with or unaccounted for by the caricature of the Ndebele state as a conglomeration of uniformly structured 'regimental' towns. The descriptions of authors like E. A. Maund, F. W. T. Posselt and R. Summers are not so much unworkable (which is true) but invalid because looking at only a part, they generalized for the whole. The lessons are threefold. The historical analysis of the major Nguni states has for far too long been derived from tendentious or incomplete nineteenth-century accounts; the anthropological study of the nineteenth-century Nguni is all too frequently vitiated by poor use of what anyway is inadequate historical source material; lastly, it is possible that a fresh historical examination of other Nguni states, more particularly that of the Zulu, will lead eventually to substantial revisions in how we

view their organizations and dynamic logic.

SUMMARY

The assumptions that Ndebele settlement was purely militarily orientated and that it was composed of a hierarchy of 'regiments' and 'divisions' or 'provinces' are false. The parallels between the Ndebele ibutho and the English regiment are so tenuous that the translation is best dispensed with. On the other hand, an Ndebele ibutho was, at least in one sense, a military organization which evolved over a number of years into a cluster of imisi (villages). These imisi were residential, essentially non-military, and composed by far the largest proportion of Ndebele settlement. The 'division' did not exist. Amhlope, Amakanda, Amnyama and Igapha were collective group concepts comprising imisi descended from four original or proto-amabutho created in the period before the migration of the Ndebele to the Matopos region. The sequence of creation of amabutho in the post-1850 period can roughly be determined, outbursts of ibutho formation usually coinciding with crises in the history of the kingdom. Amabutho created after the late 1860s such as Imbizo and Insuga disintegrated after the European conquest during the 1890s, whereas older, established imisi survive as concepts of allegiance to this day. A knowledge of how the Ndebele ibutho evolved is essential for a more complete understanding of the dynamics of Ndebele society at other levels, for example the way the chieftaincies (izigaba) and the great local lineages emerged, as well as the social and political context within which individuals sought to realize their ambitions. The unexpected way in which a closer look at the ibutho reveals much of the early literature on the Ndebele to be at best incomplete, at worst, caricature, but either way misleading, suggests that other nineteenth-century Nguni states may benefit from renewed historical examination.

Notes

1. The following is based partly upon documents of the British South Africa Company (BSAC), and a number of Historical Manuscripts contained in the National Archives of Rhodesia, Salisbury. A particularly rich source for the Ndebele is material relating to the Chief Native Commissioner for Matabeleland, I897-19I4, the most important files of which are numbered NBi (In Letters), NB3 (Correspondence) and NB6 (Reports). Other file series of use are those designated CT (Correspondence of the Kimberley and later Cape Town offices of the BSAC), and LO (Correspondence of the London Board of the BSAC). Most of the material used represents oral traditions collected by Company officials and their correspondents for a variety of immediate purposes. Historical Manuscript references are distinguished in the footnotes by the prefix Hist. MSS. The most valuable single source is filed under the code Hist. MSS W18, and contains the reminiscences of R. Foster Windram, in fact a series of traditions collected from Ndebele between I937 and I94I. I have supplemented this material considerably by fieldwork carried out in the Belingwe, Lundi, Godhlwayo, Matopos, Ntabazinduna, Khumalo and Nkai Tribal Trustlands (TTLs) of western Rhodesia. I am particularly indebted to my research assistants E. M. Mhlanga, J. B. Dube and S. H. Khumalo, to M. Mahlangu for translating the work Umthwakazi by P. S. Mahlangu from Sindebele into English, to my interpreters C. Mkwananzi and M. Senda, to A. R. Simmens for drawing the map, and to the Research Board and the Department of History of the University of Rhodesia for helping to finance this field-work.
2. J. A. Barnes, Politics in a Changing Society: A Political History of the Fort Jameson Ngoni (London, 1954), 36-41.
3. For example, E. J. Krige, The Social System of the Zulus (Pietermaritzburg, I936), 261-5; M. Gluckman, 'The Kingdom of the Zulu of South Africa', in M. Fortes and E. E. Evans-Pritchard, African Political Systems (London, I940), 3I-2.
4. A. J. B. Hughes and J. van Velsen, 'The Ndebele', in The Shona and Ndebele of Southern Rhodesia, Ethnographic Survey of Africa, Southern Africa, Part iv (London,1955), 68.
5. P. Becker, Path of Blood: The Rise and Conquests of Mzilikazi, Founder of the Ndebele (London, Panther Books, 1972 reprint), 78.
6. J. D. Omer-Cooper, The Zulu Aftermath: A Nineteenth Century Revolution in Bantu Africa (London, I966), 135.
7. R. Summers and C. W. Pagden, The Warriors (Cape Town, 1970), 41-2.
8. E. C. Tabler, The Far Interior: Chronicles of Pioneering in The Matabele and Mashona Countries, 1847-79 (Cape Town, I955), 13.
9. F. C. Selous, Travel and Adventure in South-East Africa (London, I893), 8I, for example.
10. The villages were not in any way physically linked but separated by tracts of countryside. The habit of referring to Ndebele settlements as 'towns' may have reflected the familiarity of most Ndebele travellers with Ngwato settlements which were much larger. The map on p. 609 is misleading because it shows only a small percentage of Ndebele villages. The majority were those of private individuals (imisi zamathanga) which are now extremely difficult to locate. The main conclusion from this is that the settlement "revolution' which was meant to have taken place in 1894-5 never occurred.
11. For a detailed discussion of the major Ndebele chiefly families see my Ph.D. thesis on the Ndebele under the Khumalos, now in preparation.
12. Hughes and Van Velsen, 'The Ndebele', 67; Omer-Cooper, Zulu Aftermath, 148-9.
13. A. J. B. Hughes, Kin, Caste and Nation Among the Rhodesian Ndebele, Rhodes-Livingstone Papers No. 25 (Manchester, I956), 5. The terms 'segment' and 'segmentation' are of little practical use in describing Ndebele settlement or society.
14. It is not true, as W. F. Lye claims in 'The Ndebele Kingdom South of the Limpopo', Journal of African History, x, i (I969), 99, that 'this regimental organization of the male population broke down any clan or local affiliations'. Indeed, the amabutho and the imisi which evolved from them became local affiliations; and the hereditary families which established themselves in power parallel to the evolution of settlement provided clan affiliations. Thus the Mafus of Godhlwayo or the Mkwananzis of Intunta were families with a tremendous local and personal following, as Mzilikazi found to his cost, for example, in the civil war of c. 1841-2, when both Godhlwayo and Intunta opposed him.
15. P. R. Kirby (ed.), The Diary of Dr Andrew Smith, I834-6, 2 vols. (Cape Town, 1939-40), I, 293-
16. G. G. B. Woods, 'Extracts from Customs and History; Amandebele', Native Affairs Department Annual (henceforth NADA), Ix (1931), 18.
17. D. G. Lewis, 'Lobengula's Regiments: Recruiting and Lobolo', NADA, xxxiii (1956), 5.
18. Hist. MSS WI8/I/2, Reminiscences of R. Foster Windram, Statement of Ntabeni Khumalo, 4 Feb. 1940.
19. 19 Ibid. Some informants say that it was the grandmothers who did the menial tasks for the amajaha; others that at the beginning of an ibutho's life there were no women at all.
20. J. P. R. Wallis (ed.), The Northern Goldfields Diaries of Thomas Baines, 3 vols. (London, I946), in, 658.
21. Hist. MSS WI8/i/i, Reminiscences of R. Foster Windram, Statement of Ginyalitshe Hlabangana, 6 April I940.
22. J. W. Posselt and M. Perham, 'The Story of Ndansi Khumalo of The Matabele Tribe', in M. Perham (ed.), Ten Africans (London, I936), 66-7.
23. For other accounts of ibutho-raising see Wallis, Baines Diaries, I, 261; J. Mackenzie, Ten Years North of the Orange River (Edinburgh, I871), 327; and T. M. Thomas, Eleven Years in Central South Africa (London, I873), 183
24. Kirby, Smith Diary, II, 246.
25. Interview with Nyamabani Khumalo, Shabani, 6 Dec. 1973.
26. J. P. R. Wallis (ed.), The Matabele Journals of Robert Moffat, I829-60, 2 vols. (London, 1945), I, 28I.
27. Interview with Mabhena by E. M. Mhlanga, Avoca, Godhlwayo TTL, 24 Aug. 1972.
28. For example, Mukontshwana Tjili had married at Ndinana before being drafted into Mbuyaswe, where he married a second wife; interview with Tjavu Tjili, Godhlwayo TTL, 24 Jan. 1974.
29. Hist. MSS W18/i/i, Statement of Ginyalitshe Hlabangana, 6 April I940.
30. Hist. MSS TH2/i/i, T. M. Thomas Journal, I874-83, entry for 25 April I877.
31. Interview with Mkhwezeli Bozongwana by J. B. Dube, Ntabazinduna TTL, 24 June 1972.
32. C 4643, 'Further Correspondence Respecting the Affairs of The Transvaal and Adjacent Territories', Feb. i886, Enclosure 8 in No. 34, Matabeleland, by Lieut. Maund, B.F.F., 115.
33. As implied in Summers and Pagden, Warriors, 148.
34. Hist. MSS W18/i/2, Statement of Ntabeni Khumalo, 4 Feb. 1940. Lye, 'Ndebele Kingdom', 99, is incorrect when he writes: 'the wives were either a gift of the king or were paid for with royal cattle'. A man's choice of wife was his own. Lobola was never paid with the king's cattle (izinkomo enkosi), but, at least for the first marriage, usually out of the private cattle (izinkomo zamathanga) of the father.
35. NBI/I/2, Lanning to C. N. C., ii Dec. I897.
36. NB3/i/6, Campbell to Jackson, 20 May I909.
37. Maund, C 4643, iI6; interview with Mpuqa Msiza, Lundi TTL, 4 Dec. 1973. Summers and Pagden, Warriors, 47, refer to a 'junior company' of Imbizo called Encinyane. The view of H. Child, The amaNdebele (Salisbury, I968), 20, that 'regiments did not expand within themselves, losses in war being made up from recruitment' is thus incorrect.
38. Mjaan Khumalo, chief of Imbizo, had at least two imisi zamathanga where he kept some of his wives; interview with Nqagwana Khumalo by J. B. Dube, Ntabazinduna TTL, 5 June 1972.
39. NB6/i/i, C. N. C. Annual Report for year ending 31 Mar. I898, and Annual Report, Bulawayo, same year. Imbizo's dispersal may have been largely a result of the excessive attention it received from the Europeans, as well as its position on the main Bulawayo- Inyati wagon road. Other settlements probably broke up in the migrations prior to I840.
40. Hughes, Kin, Caste and Nation, 13.
41. Interview with Chief Ngungumbane Mkwananzi, Ngungumbane, Belingwe TTL, 5 Dec. 1973.
42. Interview with Nqagwana Khumalo by J. B. Dube, 5 June 1972.
43. Interview with Makwakwa Ngwenya by J. B. Dube, Nkai TTL, IO-I5 Nov. 1972.
44. Interview with Sihlangu Tshabalala by J. B. Dube, Bulawayo, 16-24 Sept. 1972.
45. Haynes was clearly referring to Zwangendaba when he claimed in I885, 'old regiments die out or are broken up and distributed among others'; see C 4643, Enclosure 9 in No. 34, 'Report on Matabililand', by Lieut. C. E. Haynes, I20-x.
46. Interview with Mpuqa Msiza, 4 Dec. 1973.
47. Interview with Tjavu Tjili, 24 Jan. 1974.
48. This is only a rough guide since settlements such as Iliba and Induba did not break up but continued to reproduce until the end of the kingdom. Thus Manqikila, the eldest son and heir of Mehlomakulu, head of the Dhlodhlo family, was drafted into Inqobo some time after i869 and was still attached to Inqobo in I893, where and in which year his son, Ndabayenduku, was born.
49. Interview with Mkhwezeli Bozongwana by J. B. Dube, 24 June 1972.
50. On the other hand, informants often say that recruits to an ibutho were taken from a particular group of imisi. For example, Insuga had special links with the imisi of Inqobo and Ujinga in the Inkwegwesi valley. The identification of post-I840s imisi in the Matopos area with one or other of four original or proto-amabutho (discussed below) also suggests the linkage of an ibutho with a specific hinterland. Nevertheless, it seems that the amabutho of Lobengula's reign contained men from most parts of the kingdom.
51. Interview with Ntolwane Dhlodhlo and Tito Matshazi, Lundi TTL, 4 Dec. 1973. This is in no way a complete list of Ndebele settlement.
52. Interviews with Mpuqa Msiza, 3 Dec. I973; and with Zimu Moyo, Ngungumbane, Belingwe TTL, 5 Dec. 1973. Informants often arrived at this process only after much thought, and that such an evolution took place had seldom occurred to them.
53. Hughes, Kin, Caste and Nation, II, uses the word ixhiba for 'province', but ixhiba simply means a hut (for unmarried boys or girls) or a group of warriors.
54. Kirby, Smith Diary, II, 76, 79.
55. Thomas, Eleven Years, 224-5.
56. The exact status of both the umnumzana and the induna in nineteenth-century Ndebele society awaits clarification. Amongst the Zulu umnumzana means 'villageheadman', whereas Barnes, Politics, 9, describes him as a 'regional governor' of the Fort Jameson Ngoni. Barnes seems, however, to be referring to no more than the head of a lineage. Amongst the Ndebele the umnumzana was, roughly, a 'headman', and is today, according to Pelling, the equivalent of 'Mr'; see Wallis, Baines Diaries, I, 79; and J. N. Pelling, A Practical Ndebele Dictionary (Bulawayo, I966), 52.
57. For Maund's interpretation, see C 4643, Enclosure 8 in No. 34, Matabeleland, 113-i6.
58. L05/6/8, Lawley's speech at the Meeting of Headmen and Indunas at Bulawayo, 5 Jan. 1897.
59. F. W. T. Posselt, 'The Rise of the Amandebele', Proceedings of the Rhodesian Science Association, xviii (I919-20); this is reprinted with minor changes in F. W. T. Posselt, Fact and Fiction (Bulawayo, 1I935), 1161-93.
60. R. Summers, 'The Military Doctrine of the Matabele', NADA, xxxii (I955), 7-15.
61. Summers, 'Military Doctrine', 8, admits that there is no Ndebele word for 'divisional commander'; he suggests induna yomuzi, which means, however, induna of a village.
62. The military bias of Summers and Pagden is especially implicit in Appendix 2, 'Estimates of Regimental Strength at various times in 1893', Warriors, 148-51. Isikulu, for example, means 'important men,' and was never a formal council.
63. A. J. B. Hughes, 'The Restructuring of Ndebele Society under European Control', unpublished Ph.D. typescript, no date (copy held by the National Free Library, Bulawayo), 132; Hughes, Kin, Caste and Nation, ii, 69.
64. Hughes, 'Restructuring', 134. See also footnote I3.
65. Hughes, Kin, Caste and Nation, 15-I6. It is rather the use of the word regiment which has created unnecessary confusion in Ndebele studies.
66. They are usually referred to as amabutho.
67. Hist. MSS W18/i/2, Statement on the Organization of the Matabele Nation by Ntabeni Khumalo.
68. See especially P. S. Mahlangu, Umthwakazi (Cape Town, I957), 23-7. For Maund's lists see C 4643, 115-I6.
69. Hist. MSS W18/i/i, Statement of Mvumi, ii Nov. I937. Lye, 'Ndebele Kingdom', 9I, refers to Amhlope as Mzilikazi's older married regiment.
70. 'Mziki' (A. A. Campbell), 'Mlimo. The Rise and Fall of the Matabele (Pietermaritzburg, I926), 53, 59, 67; A. T. Bryant, Olden Times in Zululand and Natal (London, I929), 426, 438-9; Mahlangu, Umthwakazi, 23-7.
71. Hist. MSS WI8/i/i, Second Statement of Ginyalitshe, 6 April 1940.
72. Interview with Ngungumbane Mkwananzi, 5 Dec. 1973; interview with Mabhena by E. M. Mhlanga, 24 Aug. I972; Hist. MSS WI8/i/i, Statement of Ginyalitshe, 23 Nov. I937.
73. 'Mziki', 'Mlimo, 59; Hist. MSS WI8/i/i, Statement of Mvumi, ii Nov. I937, who said: 'when Amnyama became very few they were put with Inzimazana [Mzinyati]'. Mzinyati, led by the Gwebu and Khumalo families, was always regarded as the most important of the Amnyama-descended imisi. Godhlwayo was formed at Mkwahla in the 1830s, some say as an offshoot of Mzinyati.
74. 'Mziki', 'Mlimo, ch. I2. Gundwane's Amakanda-Amnyama group made Mzilikazi's son Nkulumane king. It is likely that this was at first regarded as a permanent solution, and symbolic of the emergence of a new but small Nguni state north of the Limpopo.
75. Hist. MSS WI8/i/i, Second Statement of Ginyalitshe.
76. 'Mziki', 'Mlimo, ch. ii.
77. Hist. MSS WI8/i/i, Second Statement of Ginyalitshe; Mahlangu, Umthwakazi, 25.
78. Interview with Zimu Moyo, 5 Dec. 1973. A glance at the map is enough to dispose of the idea of four large administrative divisions. The umbrella-handle is made by drawing a line connecting the imisi of Nqama, Amagogo, Nyamandlovu, Ujinga and Umcijo, all of which identified themselves with Igapha. Note that Nqama is separated from Amagogo by Amhlope-connected imisi such as Isizinda. Que Que is a modern town just to the north of the map.
79. For Zwangendaba, Mncwazi and Ujinga see H. M. G. Jackson, 'Boer Invasion of Rhodesia', NADA, ii (1924), 58-60; Mhlagazanhlansi (N. Jones), My Friend Khumalo (Bulawayo, I944), ii. For Mahlokohloko and Inyati see Wallis, Moffat Journals, I, 229; II, 71-2, 249.
80. For Induba see 'Mziki', 'Mlimo, IIo; for Mhlahlanhlela see J. P. R. Wallis (ed.), The Matabele Mission, a Selection from The Correspondence of John and Emily Moffat (London, 1945), 21 I; for Intemba see Wallis, Baines Diaries, ii, 535.
81. Summers and Pagden, Warriors, say that Inqobo was formed out of the Mbuyaswe ibutho which Mzilikazi formed at the close of his life. Thomas, Eleven Years, I83 (writing in about 1871), describes umbuyazwi as 'the preparatory course of the [in the context, any] ibuto', which may or may not support this.
82. London Missionary Society Archives (School of Oriental and African Studies), Matabeleland, Box i, Folder 3, Jacket B (henceforth LMS MLI/3/B), Sykes to Mullens, 2I June I878. The reasons for the change are unclear. Perhaps Intemba and Inyati had 'matured' and were no longer militarily effective; perhaps Lobengula was doubtful of their loyalty, or perhaps they were moved south to defend against the anticipated invasion by Mangwane, Lobengula's half-brother, who was supported by the Ngwato king, Macheng. The attack did not take place until January I872, however.
83. Rhodesian Government Delineation Report, Gwelo District, The Gambiza Chieftainship, 1963.
84. Thomas, Eleven Years, 190-2; CTi/8/4, Colenbrander to Harris, Telegram, 5 July 1892.
85. 85 Interview wvith Sihlangu Tshabalala by J. B. Dube, 16-24 Nov. I972; Hist. MSS W18/i/i, Second Statement of Ginyalitshe, 6 April I940, who observed that Lobengula 'only wanted young men [for Imbizo] because they would have no knowledge of Nkulumane'. Nkulumane was almost certainly dead by 187I, but was being impersonated by a man called Kanda. The real threat to Lobengula was Mangwane.
86. Mahlangu, Umthwakazi, 6o.
87. LMS MLi/2/D, Thompson to Mullens, I4 Aug. i87I; Wallis, Baines Diaries, iIl, 657-8; 'Mziki', 'Mlimo, II9.
88. See, for example, NBi/i/i, Robinson to C. N. C., 6 Aug. I900; NBE7/I/I, Stuart to C. N. C., 8 June 1897; NBI/I/5, Statement of Tala Ndiweni, 19 Nov. I902; Hist. MSS MAi/2/2, Diary of Major Thomas Maxwell I889-9I, entry for 27 Dec. 1890.
89. Hist. MSS TH2/I/I, 'Thomas Diary', entry for October I879.
90. Mbizo (F. V. Johnstone), 'Mtikana ka Mafu', NADA, IV (I926), 54-5.
91. M. Gelfand (ed.), Gubulwayo and Beyond: Letters and Journals of the early Jesuit Missionaries to Zambesia, I879-87 (London, 1968), 463.
92. CTi/I3/IO, J. S. Moffat to Harris, 6 Oct. I890.
93. Hist. MSS MOI/I/5/6, J. S. Moffat Correspondence-Official and Political, Moffat to Loch, 24 May 1892. For Jameson's misrepresentation of Lobengula's permission for the pioneer column to enter Mashonaland, see J. Cobbing, 'Lobengula, Jameson and the Occupation of Mashonaland', to be published in Rhodesian History, IV (0974).
94. Hist. MSS WI8/i/i, Joint Statement of Mvutu and Posela; and Statement of Siatcha, zo Nov. 1937; J. A. Pitout, 'Lobengula's Flight and the Shangani Battle', NADA, XL (i963), 7I; C 8547, 'Report by Sir Richard Martin on the Native Administration of the British South Africa Company', I897, 32-3, Carnegie to Martin, no date. The distinction between the umlisa and the umnumzana has not yet been cleared up.
95. CTi/I3/4, Doyle to Harris, 25 July I890.
96. Hist. MSS W18/i/z, Statement of Ntabeni Khumalo.
97. D. N. Beach, 'The Rising in South-Western Mashonaland, I896-7', unpublished Ph.D. thesis, London University (1971), 221.
98. J. White, 'History of Shabani' (forthcoming), draft chapter: 'The Ndebele Influence'.
99. L05/6/7, W. I. S. Driver to C. N. C., 2o Dec. I896.
100. It was present in the Bulawayo area by I889; see MOI/3/I1/, J. S. Moffat Diary I887-92, entry for 10 Dec. 1889.
101. Hist. MSS W18/i/i, Statement of Mvumi; Mahlangu, Umthwakazi, 23.
102. Beach, 'Rising', 2z2; Hist. MSS WIi/i/i, G. T. Wilkerson, 'The Matabele Nation', 19; Summers and Pagden, Warriors, I15.
103. Hist. MSS C04/3/I, J. W. Colenbrander Diary, June-July I890, entry for 8 June I890.
104. F. Gambo, 'The Royal House of the "Gambos"', NADA, XXXIX (I962), 49; Rhodesian Government Delineation Report, Nyamandhlovu District, The Matapula Chieftainship, I965.
105. C. L. Norris Newman, Matabeleland and How We Got It (London, I895), I6i.
106. Hist. MSS W18/i/z, Statement of Ntabeni Khumalo; Perham, Ten Africans, 67-8.
107. For this defensive cordon see CTi/i3/3, Colenbrander to Curry, I June I892. A major link in it was Amaveni, a Rozvi settlement which Ndansi Khumalo referred to as 'baboons; we called them this because they were a mixed lot, not pure Matabele'. Amaveni was in the upper Gwelo region at least by the mid-i88os; see Perham, Ten Africans, 68; CTi/6/8, Selous to Selous Exploration Syndicate, z and IZ Oct. I889.
108. Posselt, Fact and Fiction, 76; Hist. MSS WIi/i/i, Wilkerson, 'Matabele Nation', I9. Other Kalanga settlements under Ndebele control on the highveld to the west of Bulawayo were Usaba, Zinyama, Mpande and Lulwane (see map). Although 'governed' by Ndebele izinduna, they do not seem ever to have been amabutho in the formal sense.
109. Hist. MSS WI8/i/z, Statement of Ntabeni Khumalo; Posselt, Fact and Fiction, 76; WIi /i /i, Wilkerson, 'Matabele Nation', 19.
110. Interview with Nqagwana Khumalo by J. B. Dube, 5 June I972; Mhlagazanhlansi, My Friend Khumalo, z3; H. Vaughan-Williams, A Visit to Lobengula in I889 (Pietermaritzburg, I94I), 103; Hist. MSS FRz/z/i, The Reminiscences of Ivon Fry, 1938, 92. According to Nqagwana the Imbovane lived in a part of Bulawayo called Engoqeni.
111. The careers of Mtikana and Gambo, as well as other chiefs, are discussed in my forthcoming Ph.D. thesis.
112. Men could exploit other talents such as those of the isanusi ('doctor') or umkhanda (blacksmith), or even a personal hold over the king.

Read more...