1. Stayt, H.A., The Bavenda (London, 1931), p. 16.Google Scholar The Venda usually make the point, however, that there were some Venda in Rhodesia even before chief Mphephu's flight in 1898.

2. Ibid., p. 1, has “Vesha” but a nineteenth-century praise poem from Moçambique (which I hope to publish soon) has “Vetsha.” Portuguese usage in the nineteenth century changed this into Feza or Beja. Elsewhere Stayt had “Dzweda.” In the nineteenth century transformations or cognates of the term ‘Venda’ were occasionally applied to a larger group. In the 1850s and 1860s all of northeastern Transvaal seems to have been designated as “Beja” by the people near Delagoa Bay. See, for example, das Neves, D.F., Itinerário de uma viagem à caça doe elephantes (Lisbon, 1878), pp. 98–108et passim.Google Scholar In 1882 the Swiss missionary Creux described the inhabitants of “Mphalaora” (Phalaborwa) as “Bassueta de Madjadji soumis à Mozila.” Bulletin publié par le conseil de la Mission des Eglises libres de la Suisse Romande, 4(1882/1883), p. 39.Google Scholar A different, although slightly related, problem is discussed by Bothma, C.V., “Pedi Origins” in Ethnological and Linguistic Studies in Honour of N.J. van Warmelo (Pretoria, 1969), pp. 167–97.Google Scholar

3. Trigardt, Louis, ed. Le Roux, T.H., Die Dagboek van Louis Trigardt (Pretoria, 1966), pp. 3–5.Google Scholar

4. Gründler, W., Geschichte der Bawenda-Mission in Nord-Transvaal (Berlin, [1897]), pp. 13–14.Google Scholar A date of ca. 1700 for the migration of the royal Khwinde group appeared in several later publications. Alpers, E.A., “Dynasties of the Mutapa-Rozwi Complex,” JAH, 11(1970), p. 216CrossRefGoogle Scholar assigned it to “the end of the seventeenth century.” A more recent description of Dzata is in Stayt, , Bavenda, p. 6.Google Scholar

5. van Warmelo, N.J., ed., Contributions toward Venda History, Religion, and Tribal Ritual (Pretoria, 1950 [orig. 1932]), pp. 6–12.Google Scholar For variant versions see idem, ed., The Copper Miners of Musina and the Early History of the Zoutpansberg (Pretoria, 1940). One contribution in this work even referred to conquests of an ancestor of Thoho-ya-Ndou.

6. Van Warmelo, , Contributions, p. 6.Google Scholar

7. Ibid., pp. 6, 9. It was supposed that Bele-la-Mambo had lived north of the Limpopo. It is quite possible of course that all of the ancestors attributed to Thoho-ya-Ndou are not historical. One also appears in the genealogies of the Rozwi.

8. Blacking, J., “Initiation and the Balance of Power: the Tshikanda Girls' Initiation School of the Northern Transvaal” in Ethnological and Linguistic Studies, pp. 26–31.Google Scholar

9. Bothma, , “Pedi Origins,” pp. 193–94Google Scholar; N. Ralushai, personal communication. It is possible that at a fairly early time it was also a geographical term. Krige, J.D., “Traditional Origins and Tribal Relationships of the Sotho of Northern Transvaal,” Bantu Studies, 11(1937), p. 330CrossRefGoogle Scholar, referred to a country called Thobela in the traditions of the Mmamaila, which was situated to their north and may have been in the western part of the present Venda country, probably identical with the area the Dutch called Inthowelle.

Mahumane's account and the Dutch records suggest that it may have been a title-name in the 1720s as well.

10. Ibid., pp. 326, 329-30, 353; Krige, E.J. and Krige, J.D., The Realm of the Rain Queen (London, 1943), p. 305.Google Scholar Some of the origin traditions of the ruling lineages (particularly those of the wild pig clan) may have been inventions like those of the Rozwi who fabricated a new migration story after their subjection by the Ndebele.

11. Fouché, L., ed., Mapungubwe: Ancient Bantu Civilization on the Limpopo (Cambridge, 1937), p. 24 and Pls. XII-XIVGoogle Scholar; Frobenius, Leo, Evythräa (Berlin, 1931), Pl. 20.Google Scholar The ruin at Solvent is situated just west of the Sand River and is close enough to Venda country to qualify as a possible residence site of a Venda chief or even king.

12. Coetzee, C.G., “Die Kompanjie se besetting van Delagoabaai,” Archives Yearbook for South African History, 11/2(1948), pp. 167–276Google Scholar; Smith, , “Struggle,” pp. 49ff.Google Scholar In 1724 the Dutch started buying slaves, though this had not been envisaged originally, ibid., pp. 46-47, 71. Gold dust with many impurities was first obtained in 1726, K.A. 12205, f. 391v.

13. Theal, G. McCall, ed., Records of South-Eastern Africa, (7 vols.: London, 1898-1899), 1: pp. 407–506Google Scholar; Molsbergen, E.C. Godée, ed., Reizen in Zuid-Afrika in de Hollandse Tijd, (4 vols.: The Hague, 1916-1932), 3, passim.Google Scholar

14. In K.A. 12205, ff. 569-574 there is one copy of this report but two copies of the covering letter dd 2 May 1730. Another copy of the letter is in the Cape Town archives but I do not know whether a copy of the report is there as well.

15. In his letter of 2 May 1730 Van de Capelle referred repeatedly to “Lacroix.” K.A. 12205, ff. 392v, 398v, 402, 404. This is de la Croix, A.P., author of Algemeene Waareldbeschriving, (3 vols.: Amsterdam, 1705).Google Scholar In an earlier letter Van de Capelle quoted an earlier work of La Croix published in Utrecht in 1683.

16. Smith, “Struggle.” Cattle also played an important role in these exchanges. Cf. Theal, , Records, 1: p. 417, referring to 1723.Google Scholar

17. Smith, , “Struggle,” p. 62.Google Scholar

18. Theal, , Records, 1: p. 420Google Scholar; K.A. 12205, “Memortje getroken uijt de dagregister …, 9 April 1730, f. 583v-584. The term “Deessa” sounds rather like the “Beja” of nineteenth-century Portuguese sources and refers to the same general geographical area.

19. Junod, H.A., Moeurs et coutumes des Bantous (Paris, 1936), p. 29Google Scholar; Krige, , “Traditional origins,” pp. 337-38, 340.Google Scholar

20. K.A. 12205, f. 384. Van de Capelle mentions this chief Unbaaij in describing the rivers debouching into Delagoa Bay.

21. Theal, , Records, 1: p. 414.Google Scholar In K.A. 12205. f. 393 it was stated that copper and tin came from “Chiremandelle, Machicosje, and Paraotte” (or Phalaborwa). The alluvial tin deposits at Mbabane in Swaziland were probably not those referred to since this area could hardly have been under Venda control. It is possible that “Machicosje” was situated on the Mogalakwena to the southwest of Venda territory since as late as 1868 Africans spoke of a hard white metal there. British Parliamentary Papers, Africa, 36: p. 399Google Scholar, statement of Ntekwane, June 1868. The evidence is too meager to base an identification of the Langa Ndebele with “Machicosje.” Paver, F.R., “Trade and Mining in pre-European Transvaal,” South African Journal of Science, 30(1933), p. 605Google Scholar, identified “Machicosje” as Mashakatsi, chief of the Maluleke Tsonga. There are grounds, however, for believing that Mashakatsi was the personal name of a chief in the second half of the nineteenth century. This and the absence of tin in the Maluleke area militate against Paver's identification. For “Savuko” see note 47 below.

22. Caton-Thompson, G., The Zimbabwe Culture: Ruins and Reactions, (Oxford, 1931), p. 174, Pls. XL/1, XLIV/2.Google Scholar

23. K.A. 12205, f. 28. Van de Capelle to Lafontaine, 28 October 1729. Cf. K.A. 12205, f. 585, which spoke of “neeger van Maphumbe [Mpfumo] genaamt homan of Mahomane.”

24. Theal, , Records, 1: p. 420Google Scholar; Smith, “Struggle,” passim.

25. This report in K.A. 12205 is a copy. Most of the names are uncapitalized and I have rectified this, as well as changing some orthographies to forms more familiar to English readers. Evidently in the Dutch transcription -oe- and -ou- both stood for -u-. The long -a- was written either as -ae- or -aa-.

26. The information on the location of Matoeme tallies with the information gathered in 1728 near modern Xinavane on the Nkomati. According to this report Matoeme was the northernmost of two chiefs in the area called Sanguano. Theal, , Records, 1: p. 437.Google Scholar

27. This may be translated as “the area inhabited by the Laute” who were probably the people of the chiefdom called Chivuri or Chiburi (Xivuri) later. One of their two related chiefly lineages stills retains the term “Laute” in its praises while the other group regards itself as having come from the Venda area (“Bvexa”). Both claim to have come from “the wild fig tree [nkuwa] of Mbingadzi,” a claim also made by other groups in southern Moçambique. Peixe, J. Santos, “Os Va-Chivuri,” Notícias, [Lourenço Marques], Christmas number, 1970, p. 56.Google Scholar See also Jacques, A.A., Swivongo swa Maahangana (Cleveland, Transvaal, 1958) sub Xivuri.Google Scholar

28. The existence of these hut ‘hostels’ indicates that there was an established trade route through the low veld. The fact that this area was not densely populated may have been of some advantage to the traveler since it freed him from the extortions often practiced by the populations of more heavily inhabited areas.

29. Loucotte may have been further east than indicated on the map since during the nineteenth century the Olifants was often crossed closer to its junction with the Letaba.

30. van Warmelo, N.J., “The Classification of Cultural Groups” in Hammond-Tooke, W.D., ed., The Bantu-Speaking Tribes of South Africa, (London, 1974), p. 69Google Scholar, seems to have been the first to identify Paraotte with Phalaborwa. It is possible that the correct transcription should have been “Paravotwa” and that the change from Tsonga -t- to NE-Sotho -r- is a regular one (cf. Vatwa -Barwa =Bushmen, Nguni). The geographical identity of both groups is sufficient, however, to identify them. “Phalaborwa” has been translated as “excelling the south,” Bothma, , “Pedi Origins,” p. 192.Google Scholar

31. I assume here that the river was crossed below the junction of the Klein and Groot Letaba.

32. On the accompanying map two possible locations are given -- one south of the Luvuvhu, which would tally with the absence of any reference to this river in Mahumane's account and with the fact that he reported that mining operations were conducted not far from the Venda capital. The other location is in the Nzhelele valley near or at Dzata. This could probably have been reached without increasing the distances of the daily marches during the final four days. It is possible that the Luvuvhu was crossed at a place and in a time when it does not look very impressive. A third possibility -- that Mahumane did not cross the Luvuvhu at all but took a more westerly route -- would increase the distances to be covered in this period considerably.

33. The -te in brackets is mere speculation. “Woulou” is written on the very edge of the leaf and the -te may or may not have been there once. Access to another copy in better preservation might resolve the matter.

34. In modern Venda “capital” is “musanda,” B.M. Tshivase III-3, f. 409, letter of Endemann dd 26 May 1926, with another mention in 1877. The -tt- in “insatti” would put it closer to the Sotho cognate “moshate” but a certain allowance must be made for the double translation from Venda to Tsonga to Dutch.

35. It was not specified who these neighbors were but one might think of the inhabitants of the Tati area in modern Botswana and parts of southwestern Rhodesia.

36. Not only the term “Kalanga” but also the name “Gole” appears in the Portuguese records, albeit in a slightly modified form. Randies, W.G.L., L'empire du Uonomotapa du XVe au XIXe siècle (Paris, 1975), p. 32CrossRefGoogle Scholar, cites a document from ca. 1794 in which “Gole” appears in the composite form “Goromucuro,” i.e., “Great Goro.” “Goromucoro” was supposed to be another name for “Urobze,” the area of the Rozwi and was known to be situated west of Manyika. Gole is also mentioned in Venda songs, Van Warmelo, , Contributions, pp. 64–7.Google Scholar Sangamene or Tsangamene may be a version of Changamire, although -mene probably does not really correspond to -mire.

37. Portuguese correspondence from 1760 to 1769 mentioned the Loyi or Baloyi (i.e., people from “Woloij”) as “Maloyo” and the Venda as “Mafeces,” Arquivo Histórico Ultramarino, Lisbon: Moçambique, Caixa 8, Caixa 13. As the text below indicates, the Loyi lived near the Limpopo as they still do. This is consistent with the Portuguese sources noted.

38. Sena and Manica were well known in the geographical literature of the period and it seems likely that neither name was mentioned spontaneously by African traders or Mahumane, although there were some Lemba traders in contact with the Portuguese before 1726.

39. Much of the Rozwi-Portuguese trade went via Zumbo since the Portuguese used the Zambezi to transport most of their trade goods. But some of this trade may have gone via Manyika and Sena, although the Sena feira had been closed for nearly two decades by the 1720s. Nor can it be excluded that the Venda may have had trade contacts with Sofala at or before this time. After 1730, though, Inhambane became more important and then Lourenco Marques.

40. “Basarette” seems to be a distorted version of “Bazaruto” which the Dutch were unable to locate correctly at this time.

41. This passage reflects certain ethnocentric attitudes of the Ronga near Delagoa Bay. Smith, , “The Peoples of Southern Mozambique: a Historical Survey,” JAH, 14(1973), p. 569CrossRefGoogle Scholar, interpreted it as referring to the Chopi but the ancestors of the modern Chopi were living northeast rather than southwest of the Limpopo; that is, in the area between the river Nkomati and Limpopo referred to by Mahumane.

42. K.A. 12205, f. 576.

43. N. Ralushai advises me that Lawellan would probably be “Lavhelani” in modern Venda orthography. He adds that the name of a southeastern Venda outpost, Tshivulane, which I had tried to connect with “Matchiwalaane,” is not a personal name but a place name. If it had been the title of a chief one might have tried to look for the former overlord of the Venda among the Laudzi who formerly inhabited that area. Personal communication, January 1976.

44. One might speculate whether some of the old Laute or Laudzi south of the Olifants might not have been inte-trated into the Venda kingdom and whether a Tsonga chiefly lineage ruling the old Mabila chiefdom at the confluence of the Sable and Nkomati might not have been related to this postulated southerly extension of the Venda kingdom. The name of this lineage is Tovele and their neighbors say that they were originally “Suto.” Information received in a group interview in 1969 at the homestead of chief Júlio Pambe Cossa, who is their northern neighbor. Silva, J.M. da in Ferrão, F., Ciroumscripcões de Lourenço Marques: resposta aos questões (Lourenço Marques, 1909), p. 83Google Scholar, related a somewhat different account of the origins of the Mabila ruling lineage.

45. Human or Humane seems to have been a relatively common name in the area and apparently is still in use. The informant for the report was in Venda country when Masinke Human Magayi gave his information.

46. The original has “t Chiendanine” and one might translate it in this context as “then to Tchiendanine …”

47. In March of 1728 a Dutch ensign ascended the Nkomati as far as Ntimane (near modern Xinavane) and heard of a country called “Savuko” where gold was to be found and which was “still far away.” Theal, , Records, 1: p. 435.Google Scholar This might be identical with “Saluweke.”

48. K.A. 12202, f. 292v, I. Monna, “Dagregister gehouden op de gedestineerde Landtogt naar 't Landschap Sanguano …, 22 July 1727 to 26 August 1727.” The Sanguano area which Monna visited is not the same as the Sanguano or Hlanganu area further northwest.

49. One should exempt Krige's work from these strictures and acknowledge that N.J. van Warmelo's work as ethnographer and editor have made available many sources which otherwise might have been lost.

50. There are Sotho groups who formerly lived in Transvaal but are now in Botswana and who claim some connection with the Venda. They may have formed part of the Venda state in the eighteenth century. Schapera, I., The Ethnic Composition of Tswana Tribes (London, 1952), pp. 74–75.Google Scholar