Origins of the Tswana L. Ngcongco The Tswana chiefdoms form part of the larger group of Sotho peoples, while the Sotho group itself is one of the three great sub-divisions of the Bantu- speaking peoples situated north of the Nguni. 1 In addition to the Batswana or 'Western Sotho', the Sotho group includes the Basotho of Lesotho and the Orange Free State, to whom the term 'Sotho' has come to be more speci- fically and almost eXClusively applied. This group some- times also is referred to as the 'Southern Sotho'. A third group comprises the Bapedi who have been generally referred to as the 'Northern Sotho,.2 These different Sotho groups that together may be more conveniently described as 'Sotho-Tswana' at the very earliest stage of their history, shared a number of linguistic and cultural characteristics that distin- guished them from other Bantu-speakers of southern Africa. These are features such as totemism, a pre-emptive right of men to marry their maternal cousins, and an architec- tural style characterised by a round hut with a conical thatch roof supported by wooden pillars on the outside.3 Other minor distinguiShing features included their dress of skin cloaks or dikobo and breech-cloths,4 a variety of Buispoort-type pottery closely related to Schofield's NC2 variety,5 and a predilection for dense and close 6 settlements, as well as a tradition of large-scale building in stone.? 21 While the Sotho-Tswana developed these dis- tinguishing characteristics, they did, of course, also share a number of characteristics with other southern Bantu-speaking peoples. These included physical features which generally speaking, make it hard to distinguish Sotho or Tswana from Xhosa, Zulu or Swazi; although many Tswana, especially those living south of the Molopo tend to be of a lighter complexion than others, as well as being slightly lankier with prominent cheek-bones - fea- tures which clearly point to considerable inter-marriage and other forms of interaction with such groups as the Khoi, Koranna and Griqua.8 The languages of the Sotho- Tswana and other Bantu-speakers have a number of common features - they are agglutinative in construction, nearly all the words ending in vowels or with a nasal consonant, nouns do not indicate masculine or feminine gender, and these nouns are highly alliterative in character owing to an elaborate system of noun classes functioning in much the same way that gender does in European languages. Also, there are similarities in idiom which are not easy to express in a precise manner.9 Among the cultural affinities shared by the southern Bantu-speakers are their lineage descent systems. All these groups are patrilineal - a factor which is of tremendous importance to the pattern of succession and therefore to both dynastic and filial generations. They all practised polygamy, observed the levirate or sororate forms of marriage, gave bridal cattle on marrying their wives, and in varying degrees, observed the age-set orga- nisation. There are also indications that at one time all these groups practised circumcision.lO The traditions of the Sotho-Tswana people point to a northward origin, and indicate that their southward movement was part of the great migrations of the Bantu-speaking iron-age peoples. Early Sotho people have been associated with the B.l culture that is thought to have flourished between the Zambezi and the Limpopo, and were also thought to have developed the gold trade with Sofala. According to L.Fouch~ this is also attested by evidence from pottery remains, and the Sotho period ter- minated when the early Shena invasion entered the Zambezi-Limpopo area about A.D. 1200.11 Although the direction from which the Sotho and other Bantu-speakers came is readily accepted by all writers, there must be considerable reservation about locating the place of origin of these groups in either Egypt or Ethiopia.12 Other indications favouring the theory of northward origins of the ancestors of the Sotho-Tswana peoples are linguistic features, pottery styles and their architecture. Malcolm Guthrie has pointed out that there are indications that such languages as Sotho, Venda and Nguni have developed from Zezuru, which is a Shona language. If this view was tenable, it would imply a considerable period of close settlement, or at any event, a very intimate association over a long time among the speakers of those languages.l3 However, a close reading of the writings of Christopher Ehret gives the impression that save for Venda, he questions the gist of the Guthrie thesis with respect to the development of languages like Sotho and Nguni which formed elements of a proto- Southeast Bantu network.l4 His general conclusion is that the sub-Zambezi languages, which were part of the 'Pembela complex' could be divided into two groups - namely Shona and Southeast Bantu, and that what these two groups share in language they also share with other Bantu languages north of the Zambezi. Consequently 'any common period in their linguistic histories would have to be attributed to historical events outside southeastern Africa.15 D. P. Abraham who conducted extensive oral research among societies of Rhodesia added to the evidence confirming the north~d origins~ Sotho-Tswana peoples by referring to a period of close interaction between such early Sotho groups as the Bafokeng and the Barolong and Rhodesian peoples. This is thought to have taken place in the Guruuswa district of Rhodesia.l6 Again John Schofield has drawn attention to typological analogues of the ceramic wares of the Iron Age B.l culture with that of the Sotho-Tswana. It is striking, however, that Summers' discussion of the Rhodesian Iron Age B.l culture carefully refrains from any specific correlation of Sotho-Tswana and Rhodesian societies on the basis of pottery styles, while Inskeep warns against the dangers of attempting to identify ethnic groupings on the basis of pottery assemblages, especially when the samples of pottery examined have been so few. This is, of cours, a typical example of the frustration that the student of history encounters so frequently in his quest for the historical conspectus amidst the plethora of esoteric and predominantly typological writings of archeologists. The archeologist is quite prepared to describe material culture but reluctant to ascribe it to the ancestors of any particular ethnic or linguistic group. A problem flowing directly from that of the origins of the Sotho-Tswana concerns the chronology of their arrival in southern Africa. For a long time written traditions have repeated the theory that the Sotho-Tswana or at least the Tswana, arrived in South Africa in a succession of migration waves, and speculations about the time of their arrival have been inextricably bound up with this 'wave theory' of immigration. Usually the theory asserts that the Sotho-Tswana separated from other Bantu-speaking peoples in the vicinity of the Great Lakes of East Africa, and that they proceeded downwards along the western part of present-day Rhodesia in three series of migrations.l? The first wave is accordingly thought to have comprised such groups as the Dighoja, the Bathammaga, Batsatsing and other early groups simply known as Bakgalagadi, who settled in parts of the Transvaal High- veld, the eastern portion of Botswana and of the northern Cape Colony, where they intermingled freely with the pre-existing Khoi and San communities. The second wave is said to have brought the ancestors of the Bafokeng, Barolong and Batlhaping societies who settled along the upper reaches of the Molopo spreading south and west from the neighbourhood. The third and largest migration is alleged to have comprised the other major Sotho-Tswana groups whose descendants have survived as the present-day societies of Botswana, Lesotho, the Transvaal, Orange Free State and the northern Cape in South Africa. Accord- ing to this 'wave theory' each group of immigrants into South Africa subdued or conquered groups that preceded it in the area of settlement. Implied in this 'wave theory' is the notion that by the time the Sotho-Tswana crossed the Limpopo they were already ethnically well-defined and fully differentiated into the modern chiefdoms we know today; and that many Sotho-Tswana chiefdoms can be traced back to a far distant past ante-dating their migration into South Africa. Another misleading impression conveyed is that of large-scale migrations that originated hundreds or even thousands of miles from the vicinity of the equatorial lakes keeping together as ethnically homoge- neous groups that invaded the sub-Limpopo area conquering pre-existing, smaller and probably non-iron using peoples. This is, of course, hardly tenable. Several scholars have questioned, and through their researches, discredited this popular myth of mass migrations, and have demons- trated that where these have occured at all, they have been a rare phenomenon.18 With reference to the Sotho- Tswana as well, it would be far more realistic to think of their earliest migrations in terms of small-scale scattered movements of several segmentary lineage groups occuring slowly and gradually in many directions over a wide area. While during the later stages of the ten centuries preceding 1500 A.D., there may have been a few ancient but well-recognisable Sotho-Tswana groups such as the Bafokeng and, perhaps, the Barolong, it seems fairly plausible to think of the appearance of most of the Sotho- Tswana peoples we know today as the result of a slow but steady process of mingling of several segmentary groups sharing a number of cultural features. Some groups probably increased their numerical size through being joined by other small migrant groups. Many of the groups vere probably mixed and Sotho-Tswana culture would accordingly be a blend of many cultural traits that developed over a long time in some cradle-land. The increase in the size of the Sotho-Tswana population like- vise ought to be regarded as having taken place in situ through the absorption of other groups in the sub-Limpopo region. Monica Wilson has argued convincingly that patrilineal and polygynous lineages with traditions of exchanging cattle during marriage tend quickly to increase both their numerical strengths as groups and their wealth in stock at the expense of those without cattle.19 In this connection it is interesting that both the Sotho-Tsvana and the Nguni emphasized patri- lineage in their marriages. This, together with the Sotho-Tswana practice of preferential marriages - a practice encapsulated in the proverb, Ngwana rrangwane, nnyale. dikgomo di boele sakeng: 'Child of my father's younger brother, marry me, so that the (bogadi) cattle may return to our kraal,20 - would have the effect of keeping wealth in the same lineages, thereby perpetuating their preferred positions. The suggestion made here is that proto-Sotho- Tswana lineages moved very slowly into the sub-Limpopo region - a process that took several centuries and during vhich they slowly difussed a 'Sotho-Tswana' culture over groups they found in that region. It also seems fair to conclude that whatever the linguistic and cultural foun- dation they brought with them, the developments that have giyen rise to the distinctive language and culture of the Sotho-Tswana probably occured in the cradle area or 'homeland' lying immediately south of the Limpopo area.2l Such a hypothesis - of scattered proto-Sotho- Tswana lineages dispersed over a wide area between the Limpopo and the Vaal rivers and the eastern limits of the Kgalagadi desert - would also help to account for the fact that many Batswana believe firmly in the creation myth of 26 the 'cave of Lowe' at Matsieng not far from Motshodi.22 There must have been Sotho-Tswana or a proto-Sotho-Tswana population that dwelt in the vicinity long enough for the legend to gain currency and be widely diffused that they and other humans originated there. When it is also remembered that the earliest Iron-Age dates for the Transvaal go back to the fifth century A.D., the presence of iron-using, cattle-keeping and mining peoples in that region becomes highly significant. This will be so even if we do not assume, in a facile manner, that the first iron-workers in South Africa were Bantu-speakers.23 Turning more specifically to the question of chronology we may begin by considering the archeological testimony, insofar as it is intelligible to the non- archeologically trained student of history. The area between the Limpopo and the Vaal was penetrated by iron- using peoples from about the middle of the fifth century A.D. By the eighth century, some of these people were mining copper at Phalaborwa and by the eleventh, smelting metal at Melville Koppies. Two Iron-Age cultures have been identified by the Transvaal's leading Iron-Age archeologist, Revil Mason. These cultures have been named Uitkomst and Buispoort from the type sites representing those cultures. Uitkomst sites have been found to be concentrated in the central part of the southern Transvaal, around the source of the Odi (Crocodile) river near Tshwane or modern Pretoria. Radio-carbon dating has yielded two dates, A.D. 1060 for Melville Koppies and A.D. 1650 for the Uitkomst cave occurrence. The Uitkomst culture was an extension of the Rhodesian Leopard's Kopje culture which dates from the eighth to the fifteenth centuries. Buispoort culture sites which have a much higher density, occur mainly in the vicinity of the Rustenburg and Zeerust districts, that is, in the area to which the traditions of the Sotho- Tswana point as their centre of dispersion. The first date for the Buispoort culture was about A.D. 13'0 but recent work by the archeologists in Johannesburg has resulted in a fifth century A.D. date for the 'Buispoort' site o.fBroederstroom 24/73.24 27 In a recent re-evaluation of recorded tradi- tions of the Sotho-Tswana, Martin Legassick has suggested a geographical identification of the Buispoort culture complex with the Kwena-Hurutshe cluster of lineages, and the Uitkomst complex with the Kgatla or what will here be termed Kgatla-Pedi cluster.25 Mason is confident that the site Broederstroom 24/73 represents the earliest relatively complete settlement known to date in Southern Africa. The date of A.D. c.460 for that site brings the radio carbon dates for the cradle-land of the Tswana into line with that of Castle Peak in Swaziland, for which a 26 fourth to fifth century date was recently announced. Mason has further reported a date for the earliest Iron-Age penetration of South Africa. This was from a site at Silver Leaves in the districts of Tzaneen, in the northern Transvaal, where charcoal samples yielded 27 dates such as A.D. 270, A.D. 330 and A.D. 1100. These dates have the significance of bringing the chronology of the Transvaal into line with that of Rhodesian and Zambian societies. The possible links with societies north of the Limpopo are not indicated by chronology alone, but also ceramW typological analogues, which suggest influences or associations with first millenium Iron-Age communities in Malawi, Zambia and possibly Kenya.28 Finally, Iron-Age research in the predominantly Tswana populated region of Magaliesberg-Witwatersrand appears to have made great strides since 1971 and to have yielded such an abundance of data and radio-carbon dates, that the sequences are now being schematically presented in three stages. The Early Iron-Age dates from A.D. 460 to A.D. 1000, the middle Iron- Age from A.D. 100 to A.D. 1500/1600 and the Late Iron- Age dating from about A.D. 1550 to A.D. 1820.29 As far as the traditions of the Sotho-Tswana are concerned, we find we have to deal with several clusters of chiefdoms. The most important are the Fokeng-Dighoya, Rolong-Tlhaping, Kwena-Hurutshe and Kgatla-Pedi clusters. These may be briefly considered sequentially. The antiquity of the Bafokeng cluster is widely attested. Their traditions are, however, confused and incoherent, and render it extremely difficult to reconstruct their royal genealogy and thereby find a sound basis for working out their chronology. Consequen- tly, the student of Bafokeng history is bound to lean heavily on the works of such writers as Stow, Theal and Ellenberger whose pioneering and sometimes laudable syntheses of the traditions of the Sotho-Tswana tend to be vitiated by their proneness to a priori assumptions.30 Ellenberger states that the Bafokeng crossed the Zambezi during the eleventh or twelfth century. For an indeterminate period they dwelt together with the Bahurutshe and the Barolong, for says Ellenberger, there is a tradition that tells of the separation of the Bafokeng and the Barolong at the same time and place as the separation of Bafokeng and Bahurutshe.31 When they left their neighbours - the Barolong and the Bahurutshe - at a place that Ellenberger described as 'Bechuanaland', the Bafokeng migrated eastwards to the vicinity of the Magaliesberg range, named after Mogale. This is stated to have occurred before the start of the sixteenth century. The sanguinary conflict that is supposed to have given rise to this migration is alleged to have been over some young bulls that the 'Bahurutshe' wished to castrate contrary to the wishes of the Bafokeng.32 In the Magaliesberg area, the Bafokeng are said to have split into two sections. One of these remained in the Magaliesberg area, and in the nineteenth century suffered much from Mandebele raids. The o~er section further subdivided into a number of clans which migrated southwards across the Vaal, thereby becoming the first Sotho-Tswana to cross that river.33 Some of these settled near the hill Ntsuanatsatsi between modern Free State towns of Frankfort and Vrede, well before the year 1530.34 At Ntsuanatsatsi the Bafokeng intermingled and intermarried with various San and half-caste groups found in the vicinity. Tradition states that it was such a marriage by the Bafokeng chief, Napo, at Ntsuanatsatsi which led to a serious civil strife. When the chief died the sons of his San wife were denied recognition as legi- timate heirs, a situation that resulted in the disgruntled San-Fokeng sons of the late chief Napo hiving off. They migrated with their followers across the Drakensberg mountains and down along the Natal coast. Their migra- tion route is marked by the type of pottery classified by John Schofield as Natal Coast pottery,35 HC.2, which bears a strong resemblance to early Bafokeng pottery found in the Orange Free State and Lesotho.36 These Bafokeng- ba-'Hutla, literally 'the Bafokeng of the hare', whose totem was the hare, lived for a short period among the Hpondo people along the Transkei coast before moving down further south and settling for some time, among the Thembu society as the Ama-Yundle clan.3? It is thought that Hpondo and Thembu pottery styles bearing very close affinities to Fokeng pottery were a result of this interaction. This migration of the Bafokeng-ba-'Mutla from Ntsuanatsatsi has been dated to about A.D. 1600 by Ellenberger, but Schofield dated their settlement in Natal towards the middle of the sixteenth century. Walton is of the opinion that Schofield's chronology for this group is nearer the mark because the Ama-Yundle traditions confirm that they lived in the Transkei for about eight to nine generations.38 Fagan, however, casts doubt on the dependability of conclusions based on such slender or tenuous evidence, as well as on the grounds that the NC pottery had not been dated.39 Then there are the corbelled stone hut settle- ments found widely distributed over the Sotho-Tswana area. As had once been the case with the Zimbabwe ruins, there has been much discussion about whether these were indeed built by ancestors of the people who lived in that area _ in this case the Sotho-Tswana. Walton demonstrated from the pattern of the cave dwellings at Ntlokholo in Lesotho that the stone villages were the work of the Bafokeng, or people influenced by the way of living of the Bafokeng.40 Since the radio-carbon date for these settlements is about 1445-1495, it tends somewhat to support Ellenberger's estimate that the Bafokeng settlement at Ntsuanatsatsi and other high-veld sites predates A.D. 1500, but does not support his estimate of the Bafokeng-ba- 'Mutla sojourn in Natal, for which A.D. 1600 appears too late a date. A pre-1500 date for Bafokeng occupation of Ntsuanatsatsi is in harmony with the tradition that they were found there by the Lesotho line or branch of Kwena clans. As will be seen later, the latter clans are supposed to have migrated from the Kwena dispersion centre in the Limpopo-Odi-Madikwe watershed about the middle of the fifteenth century. When we turn to the Barolong-Batlhaping cluster, we find that the chiefdoms belonging to this cluster comprise the various Barolong chiefdoms found by the sons of Tau - namely the Barolong-Ratlou, Barolong- TShidi, Barolong-Seleka, Barolong-Rapulana and Barolong- Mariba - as well as the Batlhaping-Phuduhutswana and Batlhaping-Maidi sections and the Bakaa chiefdom. With the exception of the Bakaa, and small sections of the Barolong-Seleka and Barolong-Rapulana in the southwestern Transvaal, all the states belonging to this cluster are situated in the northern Cape, south of the Molopo River. Like the Bafokeng, the Barolong were among the earliest Sotho-Tswana kingdoms to establish themselves in South Africa, they appear to have been, without doubt, earlier than those chiefdoms claiming descent from Masilo (C.1415-c.1445).4l Ellenberger and Macgregor dated their first ruler Morolong, from whom the people took their 42 name, to about 1270. It has been suggested that the name Morolong is derived from an old Sotho verb rola, 'to forge', suggesting one who was versed in or a practi- tioner in the craft of a blacksmith. The association with iron or metal implied in this explanation, is carried further in the name of the son of Morolong as second ruler of the Barolong called, Noto or 'hammer' as well as in the totem of these people which was tshipe or iron.43 To 31 arrive at the date 1270, Ellenberger and Macgregor were calculating on the basis of thirty-year generations. Using thirty years we arrive at a date like c.1325-c.1355 for Morolong. Thought by Abraham to have been among the Sotho-Tswana clans that had been interacting with the Shona clans in the Guruuewa district of Rhodesia between the twelfth and fifteenth centuries,44 the Barolong were already spread widely between the headwaters of the Molopo and the Modder Rivers by the time they were ruled by their eighth king MADIBOYA, c.1535-c.1565. The rule of the ninth Barolong king Tshesebe (c.1565-c.1595) witnessed the emigration of a group of clans under the sub-king (kgosana), Phuduhutswana, and their southward trek to establish themselves at Dikgatlong45 near the confluence of the Vaal (Noka-e- Tshehla or Lekwa) and the Harts (Kolong) rivers. Tradi- tions are silent on the cause of the exodus. But these emigrant Barolong retained their links with the capital. Famine compelled this group, now living among the Korana Khoi and Griqua, to break with tradition and eat fish. 46 Since then they were known as Batlhaping. During the reign of the fourteenth Rolong king Tau the Batlhaping refused to continue paying sehuba (tribute) to the Rolong monarch. They were thereby declaring themselves independent of the Rolong state. Tau died, early in the eighteenth century fighting a war against a joint Rolong-Korana army.47 Within the same generation another group of Barolong seceded under a leader called Maidi. They joined the Batlhaping and were 48 called Batlhaping-ba-Maidi. Another off-shoot from the Rolong kingdom was that of the people later called Bakaa. Their secession was led by Tseme, a grandson of Maleka under whom friction with the main group started. After migrating to several places in what is now southern Botswana the Bakaa even- tually settled near Shoshong hills, where they overthrew the Khurutshe state they found there.49 The Kaa state was ultimately destroyed by the Ngwato. Fragments of the 32 Kaa joined the Kwena state or fled towards the Kalanga peoples in the north, where they were called Chwizina or Sebina. This brings us to the Kwena-Hurutshe cluster. The traditions of the Bahurutshe and all Bakwena chief- doms indicate that at some time in the past they were all under the same ruling line of kings. The chiefdoms that claim descent from a common ancestor, Masilo, are the Bahuruthse chiefdoms in the western Transvaal with the Tlharo and Khurutshe off-shoots, the Bakwena chiefdoms of the Transvaal - the Bakwena-Mogopa, Bakwena-Modimosana with its four sections: Ramanela, Mmatau, Matlhaku and Haake; Bakwena-Moletswane, Bakwena-Moletse, the Baphalane and Bakwena cluster of Botswana which comprise the Bakwena of Molepolole, the Bamangwato, the Bangwakets~ and the Batawana. The Bakwena clans of Lesotho also belong to this same group of chiefdoms.50 The places that occur as the earliest remembered settlement sites are Rathatheng, said to have been near the confluence of the Odi and Hadikwe Riversj and Habjanamatswana, also known as Swart Koppies, near the modern town of Brits. Both the Transvaal and Botswana cluster share with the Bahurutshe the same kings until a break-up (c.1475-c.l~05) that resulted in the existence of separate Bahurutshe and Bakwena chiefdoms. There are conflicting traditions accounting for this historic split, the consequences of which were the wide dispersal of Sotho-Tswana chiefdoms over the South African highveld up to the limits of the Kgalagadi desert on the west and almost as far as the Orange River on the south; as well as the diffusion of Sotho-Tswana language and culture. According to one tradition, the cleavage that resulted in Kwena-venerating peoples becoming Bahurutshe and Bakwena, had to do with the first fruits ceremony.51 Another tradition attributed to Kwena-Hogopa informants is that the first born child in Malope's senior house was a daughter, Mohurutshe, while the first born child in the second house was a son, Kw~na. According to this yersion, 33 the dispute was about whether the chiefdom should be in the hands of the eldest child in the senior house regard- less of whether it was female, or whether the leadership should be kept male by electing the senior son of the second house. Whichever version is preferred, it was clear there was a leadership crisis that followed upon the death of Malope. From this it will be seen that while the followers of Mohurutshe were, in consequence, forced to leave Mabjanamatshwana and move south as a separate group, with a separate totem, and while their seniority in rank as well as in ritual matters was generally recognised, it is incorrect to speak of the Bahurutshe as the parent group of others. They were as much a splinter as the group that followed Kwena and became known as Bakwena. The only difference is that they were the senior splinter. A close look at developments in the Sotho- Tswana world suggests that events of great moment were unfolding in the sub-Limpopo area - events which call for a causation of different character in addition to the usual stereotypes used to explain dynastic rivalry and fission. The break-up that was occurring in the Sotho- Twana societies was on such a scale as to render feasible the conjecture that the cleavage which produced separate Hurutshe and Bakwena kingdoms, was only the beginning of what grew to be a huge phenomenon. Breutz dates the 52 Hurutshe-Kwena split to about A.D. 1400_1480. When we glance at events occurring among the neighbouring Shona societies to the north, we are pro- vided with clues that might help to interpret develop- ments in the Sotho-Tswana world south and east of the Limpopo. Abraham records an infiltration by proto-SothO- Tswana clans from the north of the western limits of the Guruuswa district of modern Rhodesia. This infiltration, which he dates c.A.D.1250-1400, was according to him, due to progressive dessication and increasing aridity of the Kgalagadi desert.53 During the period c.1450-1480, a critical drought supervened in the Dande area and the Makorekore (so named by the Tavara clans, 'to indicate they are as numerous in their occupation of the land as the clouds of locusts that periodically descend upon the Dande'>,54 who had moved there two generations earlier, were obliged to migrate northwards across the Zambezi into the kingdom of Maravi.55 It is possible that the same drought that drove the Makorekore into the Dande, also led to the break-up of the Hurutshe-Kwena. Further, the fact that the tradi- tion of the dispute between Mohurutshe and Kwena had to do with agricultural, as well as with religious matters, may be an oblique indicator of an economic crisis. In other words, is the tradition about go loma ngwaga, 'biting the year' to be taken at face value? Is it not likely that it masks a deeper and more pervasive factor which was that of drought and its menacing effects on the entire Hurutshe-Kwena community?56 It is possible that centuries of population build-up in the area below the Limpopo, plus a gradual dessication of the Kgalagadi area, a failure of rains from the direction of the Indian Ocean during the middle decades of the fifteenth century resulted in an ecological pressure that called for a re- distribution of the Sotho-Tswana peoples over a wider area. It is likely that the split of the Hurutshe and Kwena was the first of that distribution. Between five and seven generations after the separation of the Bahurutshe and the Bakwena, while Hogopa was still ruling the latter, a terrible famine occurred, 'tlala ee boitshegang', which scattered and dispersed the Kwena clans far and wide. Calculating by generations, gives a date in the bracket c.1625-l655. It will be noted that this date correlates well with dates cited for periodic droughts in the Indian Ocean locality during the seventeenth century.57 As a result of this famine, many Kwena clans - the Modibedi, Mogorosi, Bahlakwana, Bamonaheng, Ba- Mokotedi Makhoakhoa - migrated south of the Lekwa or Vaal River into the modern Free State. One or more of these Kwena clans went to settle at Ntsuanatsatsi near the Bafokeng settlement. These Kwena clans that migrated ~ southwards beyond the Lekwa were the ones that were later organised into the ruling lineage of Lesotho. Other Kwena lineages such as the Bamo1etse. Ba-Phogo1e. Pha1ane and others, migrated eastwards where they set themselves up as separate chiefdoms. Mogopa and the remaining Kwena groups. which still included the Modimcsana cluster and 58 migrated those that later formed the Botswana branch. to Mabjanamatshwana along the Odi River to its confluence with the Madikwe and there built a settlement named Rathatheng. After a period of very strenuous or difficult existence. owing to scarcity of food and water,59 Mogopa migrated back to Mbjanamatshwana. in the modern Brits district of the Transvaal. Mogopa's return migration to Mbjanamatshwana was not joined by his brother Kgabo II who. together with his followers. remained at Rathatheng. Kgabo II's follo- wers included all the wards and divisions that were later to separate as the Bamangwato and Bangwaketse. While recorded tradition is silent on the specific reasons for Kgabo II and his followers declining to follow Mogopa to Mbjanamatshwana. it may be surmised that reservations about pressure of too dense settlement in one area in a 60 situation of droughts could hardly help matters. Personal ambition and lust for power on the part of Kgabo cannot, of course. be ruled out. Thus, partly as a result of the droughts and famines that occurred during the generation c.1625-c.1655. there emerged two Kwena kingdoms in the western Transvaal. These were the Bakwena-Mogopa based on Mabjanamatshwana. also known as Swart Koppies. and the Bakwena-Kgabo at Rathatheng. Segmentation caused by droughts and famines also resulted in the Bahurutshe state splitting into the Many.n and Gopane chiefdoms. the Bakaa and the a Phuduhutswana-T1haping hiving off from the Baro10ng-Tshidi. and the Bakwena-Modimosana splitting up with the four ~ chiefdoms known as Ramane1a. Maaka. Mmatau and Matlhaku. It was probably at Rathatheng that Kgabo II was succeeded by his son Motshodi. although according to some traditions. he (Kgabo II) led the migration of his 36 followers across the Madikwe into present-day Botswana. Another version attributes the leadership of that migra- tion to his son Motshodi. Whichever we finally settle on, it will be found that their generations occurred nearly a century earlier than the chronology suggested by Schapera and Sillery for these kings. The Kwena-Kgabo went to occupy Dithejwane hills in the present Kweneng district. There they intermingled with groups such as Bakgwatleng, Banakedi, Baphaleng and others now commonly referred to as Bakgalagadi. Towards the end of the long reign of Motshodi the huge Kwena-Kgabo kingdom broke up. Consequently two new independent states came into being - the Ngwato and Ngwaketse kingdoms. These developments 62 probably occurred late in the seventeenth century. Towards the end of the eighteenth century a section of the Ngwato kingdom seceded under the leadership of a junior son of Mathiba called Tawana. They founded a new state on Lake Nghabe - the Tawana kinr,dom named after the founder. We may now turn to the Kgatla-Pedi cluster. The Kgatla complex comprise the Bakgatla at Motshodi, and in the Pilanesberg district, the Bakgatla-Mmakau, Bakgat1a-Mosetlha, Bakgatla-Hotsha in Hammanskraal and the Bakgatla-Hmanaana in the Ngwaketse and Kweneng districts of Botswana. The Bapedi are also regarded as part of the Kgatla cluster. The Kgatla cluster claims descent from the Bahurutshe.63 If they did, and there seems to be no evidence to challenge this tradition, then they may have seceded from the 'United Phofu Confederacy' or Hurutshe-Kwena complex since the Bahurutshe state did not come into separate existence until about 1475-1505, while the founder of the Bakgat1a, Halekeleke was, according to the Bakgat1a regnal list, of the generation 64 C.1385-c.1415. Furthermore, Bamalete traditions report a very severe famine in that same generation (c.1385-c.1415) which could also account for the Bakgatla secession from the 'United Phofu Confederacy'. The problem here may be that the royal genea- logies are faulty in that collateral successions have been 37 represented as father-son successions, or the traditions or origins are incorrect.65 On the other hand, the Hurutshe-Kwena may have been telescoped or there are very likely omissions and/or extensions in both lists. Thus, the Malekeleke tradition could only tie in chronologically with early Hurutshe royal genealogies if the supposition of Breutz is accepted that a separate Bahurutshe kingdom became a reality between c.1400 and c.1480. But it is not unusual for a tradition to say that a particular group hived off from a certain chiefdom, e.g., the Bahurutshe when they in fact mean that they broke from the chiefdom of which the Bahurutshe were also a part. In this case, the evidence is even stronger when one considers that traditions generally have seen the Bahurutshe state as the senior and legitimate inheritor of the 'United Phofu Confederacy'. Further, the near similarity of the totem of the Bahurutshe and the Bakgatla groups (that is the baboon and the monkey), could also lead to a superficial conclusion of identical or common origins. But until that is proved on other grounds, the basis for such a conclusion would be tenuous and mis- leading. The tshwene, or baboon of the Bahurutshe was not their original totem, but one which they adopted much later after the split, not between Mohurutshe and Kwena, but between Mohurutshe's sons, Motebele and Motebejana, nearly a century after Malekeleke is alleged to have led the Bakgatla secession from the Bahurutshe.66 The Balete and the Batlokwa are comparative recent arrivals to Botswana. While they have interesting, rich and complicated traditions of origins, these cannot for problema of space be discussed in this essay. The traditions of the Bakalanga of Botswana also require serious study before a clearer picture of their past can emerge. Finally, we need to embark on concerted collec- tion and analysis of the traditions of societies of northwestern Botswana, the Chobe district and Kgalagadi areas. By way of conclusion it may be observed that there appears little or no value in searching for Tswana origins outside of southern Africa. Fanciful notions pointing to origins in the Sudan, Ethiopia or even Egypt may now be completely discarded. Migrations of iron- working and iron-using peoples into southern Africa probably occurred as early as the beginning of the Christian era. Among such communities were the ancestors of people later called the Tswana. These were small-scale migrations of families, clans and groups of clans migra- ting in a variety of directions. Some groups became the ~ for the emergence of larger states which we call kingdoms. Indications are that these large states developed during the second millenium of the Christian era and that the plethora of modern chiefdoms were due to fissions caused as much by droughts, famine, pestilence search for better conditions as by dynastic ambitions and rivalries. 39 F 0 0 T NOT E S lWilson, H. 'The Sotho, Venda and Tsonga', in M. Wilson and L. Thompson, eds., The Oxford History of South ~, I, p. 131 2Legassick, 'The Sotho-Tswana Peoples b~fore 1800', in L. Thompson, ed., African Societies in Southern Africa, London 1969, p. 97; E. Westphal, 'The Linguistic Pre- history of Southern Africa', Africa, XXXIII, July 1963 3Legassick, 'Sotho-Tswana Peoples before 1800', p. 86; N. J. van Warme10, 'Grouping and Ethnic History', in I. Schapera, ed., The Bantu-SpeakinK Tribes of South Africa, London 1937, p. 57. Christopher Ehret draws attention to indirect archeological evidence of Bantu borrowing from Central Sudanic communities 'in the shift of Bantu-speaking communities from their older rectangular house style - an ancient Bantu feature - reconstructible on comparative ethnographic grounds and still present in some early Iron-Age Bantu sites in Zambia - to the round, conic1e-roofed house prevalent in more recent periods among east-central and southern African Bantu'. (C. Ehret, 'Patterns of Bantu and Central Sudanic Settlement in Central and Southern Africa', Transafrican Journal of History, III, nos. 1 and 2, 1973, pp. 25, 27 4E11enberger, History of the Basuto: Ancient and Modern, translated by J. C. Macgregor, London, 1912; reprinted New York, 1969, p. 34 5Inskeep, R. R., 'Archeological Background', in Wilson and Thompson, eds., Oxford History, I, p. 36; J. Walton, 'Early Bafokeng Settlement in South Africa', African Studies, XV, 1956, pp. 37-38 6Wilson, 'The Sotho, Venda and Tsonga', p. 139 7Mason, 'Prehistory of the Transvaal, pp. 378-381; Wilson, 'The Sotho, Venda and Tsonga', p. 139 9Doke• C. M •• 'Language'. in Schapera. ed •• Bantu- speakin~ Tribes •••• pp. 309-312; S~ M. Mole ... The Bantu - Past and Present: An Ethnographical and Historical Study of the Native Races of South Africa, Edinburg. 1920. pp. 153-157 10Krige. E. J •• 'Individual Development'. in Schapera. ed •• Bantu-speakin~ Tribes. pp. 100-107. 113 llQuoted in A. J. Wills. An Introduction tQ History of Central Africa. London. 1967, p. 19; J. D. Clark. The Prehistory of Southern Africa, Harmondsworth,1959 l2Ellenberger, History of the Basutho. pp. xviii. xix; Molema, Bantu - Past and Present, p. 40 l3Guthrie, H •• 'Some Developments in the Prehistory of Bantu Languages'. Journal of African History. III. 2. 1962. pp. 277-278. But see Ehret, 'Patterns of Bantu and Central Sudanic Settlement', pp. 10-13. where he shows that all the Bantu languages today covering most of Malawi, Mozambique and eastern Zambia and southeastern Africa, and including Bemba, Cewa. Makua. Shona and Nguni belong to a grouping known as Pembela bear traits of interaction with Central Sudanic societies. The Pembela grouping later sub- divided into seven sub-groups which became distributed relatively evenly across a vast span of country stretching from far north Zambia to South Africa. Ehret further hypothesizes that: 'Because Shona is only one language. though with many diverse dialects it is possible to argue •••that ancient Shona-Central Sudanic interactions might have taken place before or as ancestral Shona-speakers began to settle across the Zambezi perhaps in an area as far south as Zambia'. l41n addition to Ehret's views referred to in preceding footnote, see C. Ehret and others. 'Outlining Southern African History ARe_evaluation. A.D. 100-1500'. Efahamu. III. 1. 1972. pp. 9-72 l5Ehret• 'Outlining Southern African History'. pp. 11-15 l6Abraham. D. P., '1he Early Political History of the Kingdom of Hwene Mutapa 850-1589'. Leverhulme Inter- Collegiate History Conference. Hi~torians in Tropical Africa. Salisbury. 1960, pp. 62. 77; Idem •• ~ramuca : An Exercise in the combined use of Portuguese Records and Oral Tradition'. Journal of African History. II. 2, 1961. p. 211 41 l7Stow, G., The Native Races of Southern Africa, London 1905, facsimile reprint Cape Town, 1964, chapters 21-26; Schapera, Ethnic Composition, pp. 5-6; Idem., The Tswana, p. 14; Breutz, Mafeking District, pp. 23-24; Transvaal Native Affairs Department, Short History of the Native Tribes, Pretoria 1905 pp. 8-13; Ellenberger and Macgregor, History of the Basuto, pp. 15-20 18Vansina, J., Kingdoms of the Savanna, Madison, 1966; Ehret, 'Patterns of Bantu and Central Sudanic Settlement ••• ', pp. 1-2, 9; Idem., 'Outlining Southern African History', pp. 13-20; Legassick, 'The Sotho- Tswana ••• ', PP. 92-93; Parsons, 'Origins of the BaaaNgwato', pp. 82, 91; D. W. Phillipson, 'Early iron-using peoples of Southern Africa', in Thompson, ed., African Societies, p. 45 19Vilson, M., 'Changes in Social Structure', in Thompson, ed., African Societies, pp. 76-83 20Schapera, I., 'Kingship and Marriages Among the Tswana', in A. R. Radcliffe-Brown and Daryll Forde, ed., African Systems of Kingship and Marriage, London 1950 pp. 150-151 21Cf• Ehret, 'Outlining Southern African History', p. 13 where he makes a similar suggestion regarding lingui- stic developments. See also Legassick, 'The Sotho- Tswana ••• ', p. 93 220n the 'Cave of Lowe' at Matsieng, see Wookey, Dico, p. 43; P. L. Breutz, The Tribes of Marico DistrICt, Pretoria, 1953-1954, pp. 67-68; Schapera, The Tswana p. 59 23Summers, R., 'The Southern Rhodesian Iron-Age', Journal of African History, II, 1, 1962, p. 7; A. Galloway, The Skeletal Remains of Bambandyanalo, Johannesburg, 1959, p. 11; Brian Fagan, 'The Later Iron Age in South Africa', in Thompson, ed., African Societies, indicates that evidence points to Mapungubwe being an industry of Bantu Speakers. This view is endorsed by Inskeep. See R. R. Inskeep, 'The Archeological Background', in Wilson and Thompson, eds., Oxford History, I, p. 39; ~., 'South Africa~ in Shinnie, ed., The African Iron Age, p. 261 42 24 Mason, R., 'First Early Iron Age in South Africa: Broederstroom 24/73, Brits District, Transvaal', S.A.J.S., LXIX, November 1973. In a subsequent article, Mason released new dates for Buispoort sites ranging from A.D. 1240-18501 see R. Mason, 'Background to the Transvaal Iron Age - new discoveries at Olifantspoort and Broederstroom', S.A.J.S., LXXIV, 6, January 1974 25Legassick, 'The Sotho-Tswana ••• ', pp. 100 et seq.; Parsons, 'Origins of the BamaNgwato', pp. 91-92 26 Fagan, B., Southern Africa, London 1965, p. 157 27Klapwijk, M., 'An Early Iron Age site near Tzaneen, N.E. Transvaal', S.A.J.S., LXIX, November, 1973, pp. 324-326 28 Mason, R., 'Background to the Transvaal Iron Age _ new discoveries at Olifantspoort and Broederstroom', J.S.A.I.M.M., LXXIV, 6, January 1974. 30For examples of such tendencies, see Legassick, 'The Sotho-Tswana ••• ', pp. 92-93 31Ellenberger and Macgregor, History of the Basutho, pp. 15, 333; J. Walton, 'Early Bafokeng Settlement in South Africa', African Studies, XV, 1, 1956, pp.37-38 32Ellenberger and Macgregor, History of the Basutho, pp. 15, 16; P. L. Breutz, The Tribes of the Rustenburg and Pilanesberg districts, Pretoria 1952-53, pp. 59-61 33Ellenberger and Macgregor, History of the Basuto, p. 68; Walton, 'Early Bafokeng Settlement ••• ', p. 38. It will be noticed that Ellenberger and Macgregor con- tradict themselves here. On page 53 they stated, more accurately, that the Dighoja were the first of the Sotho-Tswana peoples to migrate south of the Vaal, and that they were also responsible for the stone ruins of Dithakong near Kudumane. 34Walton, 'Early Bafokeng Settlement ....', p. 38 35Schofield, J. F., Primitive Pottery, Cape Town, 1948 pp. 151-152 43 36Ibid.; Inskeep, 'South Africa', in Shinnie, ed., ~e African Iron A~e, pp. 255-256 37Walton, 'Early Bafokeng Settlement •••', p. 38 39Fagan, 'The Later Iron Age', in Thompson, ed., African Societies, pp. 64-65 40Mason, Prehistory of the Transvaal, pp. 379-380; ~., 'The Origin of South African Society', S.A.J.S., LXI, 7, 1965, pp. 255-267 41Breutz, Mafekin~ District, p. 25, states that the first two kings of Kwena-Hurutshe ('United Phofu' confederacy) were contemporaneous with the Rolong kings, Morare and Mabe. 42 Ellenberger and Macgregor, History of the Basuto; p.393 43Ibid., pp. xviii-xix; Wilson, 'The Sotho, Venda and ~onga', in Oxford History, I, p. 145; Breutz, Mafeking District, p. 25. Breutz differs with Ellenberger and Macgregor and states that Morolong means 'kookoo-man' in the same way that Mokwena means 'crocodile-man'. In the Interlacustrine region, along the shores of the Nyanza (Victoria) an early ethnic group of iron workers were known as Barongo or Balongo. Their earliest chi~' dates to about A.D. 1250. Newcomers who were not iron workers mixed with the Balongo and new dynasties emerged in the fifteenth century. Iron working ritual was prominent among these new kings. 44Abraham, 'The Early Political History of the Kingdom of Mwene Mutapa (850-1589)', pp. 62, 77 45Dikgatl0 means confluence or meeting place 46Language, F. J., 'Herkoms en Geskiedenis van die Tlhaping', African Studies, I, 2, 1943, pp. 120, 133; Stow, Native Races •••, pp. 434-439; Brown, Bantu ~, p. 206; Wookey, Dico, p. 12 --- 47Matthews, Z. K., 'A Short History of the Tshidi Barolong', Fort Hare Papers, I, 1, June 1945, p. 12; ~., 'Barolong', in I. Schapera, ed., Ditirafalo, p. 9; Breutz, Marico District, pp. 30, 31 48Language, 'Herkoms en Geskiedenis •••', p. 121 49 Ngcongco, L. D., 'Aspects of the History of the Bangwaketse to 1910', pp. 38-40 50Ellenberger and Macgregor, History of the Basuto, chapter 7, passim; TNAD, Short History of Native Tribes; R. Massie, The Native Tribes of the Transvaal, London 1905; Breutz, The Tribes of Rustenburg and Pilanesber~ Districts, pp. 19-20; Sillery, Bechuanaland Protectorate, pp. 104-105; Schapera, Ethnic Composition, p. 9 51According to this tradition, when the time for biting the curcubitae ('go loma thotse') came, it was found that baboons had already eaten the melons and Kwena accordingly refused to eat the saliva-stained melons, while Mohurutshe and his people did so. In this way, Mohurutshe acquired seniority or rank while Kwena appeared to have gained political control. (See Wookey, Dico, p. 44; also footnote 56 below) 52Breutz, Marico District, p. 25; Idem., Mafekin~ District, p. 28 53Abraham, 'Early History of Mwene Mutapa', pp. 62, 77. According to Abraham, this entry of militarily strong Sotho-Tswana clans stimulates a military organization of Shona peoples under Rozvi leadership, in order to consolidate Rozvi suzerainty. The result is that by the start of the fifteenth century there was what Abraham calls 'an established political hierarchy'. This is followed by a shift of the Karanga kingdoms from the Guruuswa to the Dande area just below the Zambezi, around A.D. 1400 54Ibid., p. 63 55Ibid., p. 64 4S 56The baboons and the crisis associated with the biting of the melon may in fact be an allegory which masks an embarassing historical event. It is possible to look upon the baboons which ate the melons as famine- stricken refugees of the tshwene or baboon totem who joined the Hurutshe-Kwena complex, and who broke the ritual taboo of eating the new harvests before the mo1omo or first fruits ceremony. This would make the dISPUte between Mohurutshe and Kwena one over whether to accept these ill-mannered refugees as citizens of the chiefdom or not. Thus the phofu peoples formed the Bahurutshe state while Kwena set up his own chief- dom of Kwena-venerating people. Upon the death of Mohurutshe, his son Motebe1e ruled, and again conflict with the tshwene or baboon group arose and was again told in allegory. Motebele gave a young baboon he had caught to his younger 'brother' Motebejana who lost it and was chastised for it by the king and seceded in consequence. It seems possible that Motebejana may have been of the tshwene totem or had been made a sub-chief over a tshwene-venerating clan. In any case, when Motebe1e died, the group was re- united under Motebejana of the tshwene totem, the ~hofu having disappeared upon the death of Motebele. The story strongly suggests a change in the dynasty rather than a simple change of totem of the old dynasty as the tradition purports. 57Schapera, 'Bakwena', Ditirafa10, p. 35; Abraham, 'Early History of Mwene-Mutapa', pp. 70 and 87, where Abraham cites de Souza, Oriente Conquistado, I-V-II 58Which is being referred to in this study as the Kwena- Kgabo in order to distinguish them from the Kwena- Mogopa. 59Schapera, 'Bakwena', Ditirafalo, pp. 35-36 60Conditions of drought and famine did not disappear for long periods. For instance, during the years 1600- 1700, droughts have been recorded for the areas near the Indian Ocean in 1630, 1645 and 1660 61Ngcongco, 'History of the Bangwaketse', pp. 53-57 62van Warme10, N. J., The Bakgat1a of Moset1ha, p. 4 63See Chart VII in the Appendix 64For an illuminating discussion on the dangers of lengthy king-lists supposed to be based on regular father-son successions, see Henige, The Chronology of Oral Tradition, Chapter 2, passim. 65Wookey, Dico, pp. 38-39. During the generations c.1565-1595 and c.1595-1625, the Kgat1a threw off segments that went to form the independent Kgat1a- Mmakau, Kgatla-Motsha, Kgat1a-Mosetlha as well as the Pedi chiefdoms. It is possible that the 1561 famine cited by Abraham may have been felt here too