¹ See, for example, Thornton, John, ‘The slave trade in eighteenth century Angola: effects on demographic structures’, Canadian J. Afr. Studies (1980), 417–27.Google Scholar

² Thornton, John, ‘Demography and history in the Kingdom of Kongo, 1550–1750’, J. Afr. Hist. XVIII (1977), 507–30CrossRefGoogle Scholar and ‘An eighteenth century baptismal register and the demographic history of Manguenzo’in Fyfe, Christopher and McMaster, David (eds.), African Historical Demography, I (Edinburgh, 1977), 405–15.Google Scholar

³ Magyar, Lázló, Reisen in Süd Afrika (Pest, 1859, reprinted New York, 1973), 242–5, 362, 392–400.Google Scholar

⁴ For Magyar's background, see de Kun, Nicholas, ‘La vie et le voyage de Ladislaus Magyar dans l'intérieur du Congo en 1850–1852’, Bulletin des Séances, Académie Royale des Sciences d'Outremer, n.s. VI (1960), 605–36.Google Scholar

⁵ On the Ovimbundu kingdoms, see Heywood, Linda, ‘Production, trade and power: the political economy of central Angola, 1850–1930’ (Ph.D. thesis, Columbia University, 1984).Google Scholar

⁶ Magyar, Reisen, 432–3.Google Scholar

⁷ Ibid. 242–44. Cf. Kanyungo district, with 10,000 people in 60 villages and 2,000 in the Kombala (central town), averages 133 per village; Dyindyoya with 8,000 people and 70 villages (adding 2,000 for the central district, though Magyar gave no estimate) averages 875 per village.

⁸ Ibid. 392–6 (Caconda: 100,000 people, a total of 400 villages and four kombalas with a total population of 8,00) and 399–400 (Ngalangi: 250,000 people with 2,050 villages and kombalas totalling 20,500).

⁹ Our conception of a ‘normal’ age distribution derives largely from John Thornton's earlier work on eighteenth-century Kongo populations, in ‘Demography and history’ and ‘Baptismal register’, which is partly based on documentary research and suggests that certain mathematically generated age distribution tables can be used to estimate agepyramids for seventeenth- and eighteenth-century central Africa. The degree to which these can be distorted by such processes as the slave trade can be seen in ‘Slave trade in eighteenth century Angola’.Google Scholar

¹⁰ Heywood, Linda and Thornton, John, ‘Demography, production and labor in Central Angola, 3890–3950’, in Cordell, Dennis and Gregory, Joel (eds.), African Population and Capitalism: Historical Perspectives (Boulder, Colorado, 1987).Google Scholar

¹¹ Ibid. 236–7; 260 n.26 on Magyar's wife and her background; 445 on the cooperation her marriage elicited.

¹² Ibid. 387. Magyar illustrates the principle by citing the case of the district Onduma in Mbailundu.

¹³ Ibid. 245. This princess was named Ina-Kullu-Sake.

¹⁴ de Miranda, José and Brochado, Antonio (eds.), Viagens e apontamentos de urn portuenese em Africa: Excerptos do ‘Diario’ de Antdnio Francisco da Silva Pôrto (Lisbon, 1942), 82–3.Google Scholar The original version of the diary in the Biblioteca da Sociedade de Geografia de Lisboa, Reservados (henceforward: SGL, Res.) 146–C–6, 1 (1846–1853), contains material not in the published version, but in this passage there is nothing of consequence omitted.

¹⁵ Ibid. 84–7.

¹⁶ Magyar, Reisen, 278, 291, 308.Google Scholar

¹⁷ Ibid. 308.

¹⁸ Pôrto, Silva, in Viagens, Miranda and Brochado (eds.), 192.Google Scholar

¹⁹ SGL, Res. 146–C–6, I(1846–1953), and ibid. vii (1882), 139.

²⁰ Ibid. (both citations) and Magyar, Reisen, 200.Google Scholar

²¹ Chart 1 is drawn from material taken from Magyar, Reisen, 73, 123, 18, 242–4, 362–400 and map. Also see map notes. We have followed modern orthography for kingdoms, and Magyar's German orthography for the lesser units.Google Scholar

²² The Map is drawn from the same sources as Chart 1. We have drawn the base map onto squared paper using the (U.S.) Army Map Service Series at 1:2,000,000 scale as a base for hydrography and some physical features (this map also places contours at 100-meter intervals) and have matched these against the map illustrating Magyar's book. Magyar's map contains several errors: although he calculated the position of a number of places by astronomical methods, his latitude and longitude grid is off by nearly a degree in each dimension. Secondly, he made numerous minor errors in the hydrography for the north of the map, and several serious errors in hydrography for the southern areas (the rivers that flow north and south are shown flowing northwest to southeast). Thus we have had to redraw Magyar's map, making use of his hydrography and physical features and narrative description in the text. In view of the importance of the surface area calculations, however, we have tried to harmonize Magyar's estimates of area with the areas in the redrawn map (not a difficult task, in many cases there is a very close correlation). Magyar attempted to draw kingdom boundaries on the map, and we have respected these as much as possible, although his boundaries are much changed in our map in the south and southeast. We have supplied district boundaries, using mountain ranges and rivers as boundaries where apparently indicated, and have been more arbitrary where such obvious features are lacking.Google Scholar

²³ Magyar, Reisen, passim.Google Scholar

²⁴ Ibid.passim; Graça, Joaquim Rodrigues, ‘Derrota em direitura de Loanda…(1843) and ‘Derrota da viagen da provincia de Bihé…’ (1846) in ‘Expedicão ao Muatayanvua’, Boletim da Sociedade de Geografia e da História de Lisboa, IX (1890), 371–85 and 402–408.Google Scholar

²⁵ Heywood and Thornton, ‘Demography’.Google Scholar

²⁶ Centro de Documentação e Investigao Histórica, Luanda Estante 69, Caixa i, Huambo, documento 42, ‘Recenseamento de Posto de Polícia Civil do Lepi’, 1919–1920.Google Scholar

²⁷ de Cerqueira, Ivo, Relatório, 1931–1932, Direcção dos Serviços e Negócios Indígenas (Luanda, 1932), 67.Google Scholar

²⁸ Ibid. 67–9. Cerqueira also gives densities for Bié (Viye), but the administration's decision to add substantial lands of low population lying east of the old kingdom reduces the average density sufliciently that a direct comparison is not possible. Cerqueira's density, for example, is v86, Magyar's is 42 (although one should note that the eastern parts of the kingdom had densities in the 1·0–1·5 range).

²⁹ On the overall history of this period, see Heywood, ‘Production’. For some suggestions on the demographic effects of this see Heywood and Thornton, ‘Demography’.Google Scholar

³⁰ For a discussion of the nature of the age-sex distribution in the 1920s and subsequent population change, see Heywood and Thornton, ‘Demography’.Google Scholar

³¹ de Vasconcellos, Alexandre José Botelho, ‘Descripcáo de Captania de Benguella, suas Provincias, Portos, Rios mais Caudelosos…’ (August 1799) in Annaes do Conseiho Ultramarino, Parte não oficial, series 4, no. 4 (1844), 161 (‘Mappa de Capitania de Benguella’).Google Scholar

³² Ibid. 161.

³³ Heywood, ‘Production’, 103–5.Google Scholar

³⁴ Pôrto, Silva, in Miranda and Brochado (eds), Viagens, 82–3.Google Scholar

³⁵ Heywood, ‘Production’, 113.Google Scholar

³⁶ Ibid., 105–10.

³⁷ For a useful introduction to Magyar's unpublished work, see de Kun, ‘Vie et voyage’.Google Scholar

³⁸ Petermann published the letter, dated Lueira, 16 November 1858, as ‘Ladislaus Magyar's Erforschung von Inner-Afrika. Nachrichten über die von ihm in den Jahren 1850, 1851 and 1855 bereisten Länder Moluwa, Moropu und Lobal’, Petermann's Geographische Mitheilungen, v (1860) 227–37. A Hungarian version was published from a lecture Magyar gave to the Hungarian Academy of Sciences on 10 Oct. 1859 which is practically identical, published in Académai Ertesitö, XIX (1859) and translated into English in L´sló Krizsán, ‘“Homo Regius” in Africa (in commemoration of the centenary of David Livingstone's death)’, Studies on Developing Studies (Institute for World Economics of the Hungarian Academy of Sciences, no. 78, Budapest, 1975), Appendix I, L´zló Magyar, ‘A brief account of the Moluva or Moropuu and Lobal countries’, 1730.Google Scholar

³⁹ Ibid. 24 (German version, 231).

⁴⁰ Graça, Joaquim Rodrigues, ‘Expedição ao Muatayanvo. Diario de Joaquim Rodrigues Graça’, Boletim da Sociedade de Geografia de História de Lisboa, IX (1890), 468.Google Scholar

⁴¹ Ibid. 468, states tribute was taken in ‘ivory, slaves and money (fazenda)’. See also Magyar, ‘Brief account’ (tr. Krizán), 24 (German version, 231).

⁴² We have accepted the boundaries of Lunda c. 1850 as they appear on the map in Vansina, Jan, Kingdoms of the Savanna (Madison, 1966), 217. Total area estimated by redrawing the map on squared paper and counting squares.Google Scholar

⁴³ Magyar, ‘Brief account’ (tr. Kriz´n), 24, (German version, 231).Google Scholar

⁴⁴ The 1960 census of Angola gives the population of Moxico province as 1·3 per square kilometre, and shrinking, since the 1970 census showed only 1·1 per square kilometre. Lunda Province, on the other hand, had about 1·5 people per square kilometre and was rising to 1·8 by 1970. See Kaplan, Irvin (ed.), Angola: A Country Study (Washington, 1979), 66.Google Scholar Population enumeration in this part of Angola was very lax in the preceding period and the numbers should probably not be taken too seriously. In Zaire, where the heart of the Lunda empire lay, densities were higher, running generally around 3–4 in 1930, though lower in places; see the map in Jewsiewicki, Bogumil, ‘Rural society and the Belgian colonial economy’, in Birmingham, David and Martin, Phyllis (eds.), History of Central Africa (London and New York, 1983), 11, 102. Here again, further research will clarify the evolution of the population in the colonial period.Google Scholar

⁴⁵ Magyar, ‘Brief account’ (tr. Krizán), 24 (German version, 231).Google Scholar

⁴⁶ Heywood, ‘Production’, 102, 109.Google Scholar

⁴⁷ Thornton, ‘Effect of slave trade’, 424–5.Google Scholar

⁴⁸ Nachtigal, Gustav (tr. Allan, B. and Fisher, Humphrey J.), Sahara and Sudan, iv (London, 1971), 358.Google Scholar

⁴⁹ Earlier versions of this paper were presented at the International Conference on the Demography of Colonial Central Africa, Milwaukee, Wisconsin, August 1986 and at the conference “Peuplement du Monde” in Paris in June 1987. The authors would like to thank the conference participants, and especially Dr. Bruce Fetter, for their comments on this paper. We would also like to thank Dr. Joseph C. Miller and Dr. Andrew Roberts for their criticism and comments. Finally, we wish to thank Dr. Charles Geiger of the Geography Department of Millersville University (Pennsylvania) for drawing the map.Google Scholar