¹ Curtin, P. D., Economic Change in Precolonial Africa: Senegambia in the Era of the Slave Trade (Madison, 1975)Google Scholar; Klein, H., The Middle Passage (Princeton, 1978)CrossRefGoogle Scholar; Kopytoff, I. and Miers, S. (eds), Slavery in Africa; Historical and Anthropological Per-spectives (Madison, 1977)Google Scholar; Inikori, J. E. (ed.), Forced Migration: the Impact of the Slave Trade on African Societies (New York, 1982)Google Scholar; Manning, P., Slavery, Colonialism and Economic Growth in Dahomey (Cambridge, 1982)CrossRefGoogle Scholar; Robertson, C. and Klein, M., (eds), Women and Slavery in Africa (Madison, 1983)Google Scholar; Lovejoy, P., Transformations in Slavery (Cambridge, 1983)Google Scholar; Galenson, D., Traders, Planters and Slaves (Cambridge, 1986)CrossRefGoogle Scholar; Eltis, D., Economic Growth and the Ending of the Transatlantic Slave Trade (New York, 1987).Google Scholar

² However, Lovejoy, , Transformations, 62Google Scholar, incorporates a small sample from Mettas's, JeanRépertoire, vol. 1.Google Scholar

³ Mettas, J., Répertoire des expéditions négrières françaises au xviii^e siècle, edited by Daget, Serge, 2 vols. (Paris, 1978, 1984).Google Scholar

⁴ As the surviving data somewhat overrepresent the trade of the smaller French colonies, especially in the 1780s, it is useful to assign weights to the trade of each colony for each of the three periods used in Table 1, in order to reflect their true share of total French Caribbean imports. For this purpose the data in Curtin, P. D., The Atlantic Slave Trade: a Census (Madison, 1968), 166, 180Google Scholar were used. The resultant changes are small. The overall sex ratio rises to 184, and the periodic ratios to 183, 172, and 200.

⁵ Robertson, and Klein, , Women, 4, 32, 39.Google Scholar In fact, in three of the four data-sets that Herbert Klein presents in this study males made up less than two-thirds.

⁶ Inikori, , Forced Migration, 24.Google Scholar

⁷ The Pará/Maranhão statistics cited in Table 1 point to a fairly low sex ratio, but there is every reason to believe that this cotton-growing region supplied primarily from Cacheu and Bissau, and where the male/female price differential was negligible, was quite atypical. Census data indicate that slave sex ratios were much higher in other Brazilian provinces, but themselves are only crude guides to the composition of slave imports. Conversely, some sources mention large numbers of women and youths among Brazilian slave imports. See Carreira, A., As companhias pombalinas de navegaçao (Porto, 1969), 92, 161–8Google Scholar; Conrad, R., World of Sorrow (Baton Rouge, 1986), 9–12Google Scholar; Schwartz, S., Sugar Plantations in the Making of Brazil (Cambridge, 1985), 348–53Google Scholar; Miller, J., ‘Slave Prices’, in Lovejoy, P. (ed.) Africans in Bondage (Madison, 1986), 57, 61, 72Google Scholar; Miller, J., Way of Death: Merchant Capitalism and the Angolan Slave Trade, forthcoming, 130, 163–4.Google Scholar In the 1983 edition of his study, Carreira states that the data are too unreliable for analysis.

⁸ Klein, H., ‘Cuban slave trade’, in La traite des noirs par l'Atlantique (Paris, 1978).Google Scholar

⁹ Geggus, D., ‘The demographic composition of the French Caribbean slave trade’, Boucher, P. (ed.), Proceedings of the 14th Annual Conference of the French Colonial Historical Society, Natchez, 1988 (Washington, 1989)Google Scholar, forthcoming.

¹⁰ On the seventeenth century French slave trade see, Gautier, A., Les soeurs de Solitude: la condition féminine dans l'esclavage aux Antilles du xvii^e au xix^e siècles (Paris, 1985), 80Google Scholar, which states that around 1660 French slavers carried an equal number of men, women, and children; also Petitjean-Roget, J., La société des habitations à la Martinique (Lille, 1980), vol. 2, 1448.Google Scholar

¹¹ Males amounted to 71.5 per cent of total imports into Havana during the period 1790–1820 only by virtue of local importations of Creole slaves from other islands: Robertson, and Klein, , Women, 32–3.Google Scholar A slightly lower proportion is given in Klein, , Middle Passage, 222.Google Scholar

¹² These were the two later Dutch trades, and the British trades to Barbados, 1663–67, to the British Caribbean, 1673–1723, and to Spanish America, 1715–38.

¹³ See above, note 4.

¹⁴ French planters generally classed as adults those seemingly aged fifteen and over, as did British traders of the late seventeenth and early eighteenth centuries: Davies, K. G., The Royal African Company (London, 1957), 300.Google Scholar In Dutch practice, the adult classification began at age sixteen: Gemery, H. and Hogendorn, J. (eds), The Uncommon Market: Essays in the Economic History of the Atlantic Slave Trade (New York, 1979), 256.Google Scholar In late eighteenth century trade to Jamaica, a height of 4 4 was the criterion used. This suggests an age range for children of about 0–11 or –12 years. Cf. Friedman, G., ‘The heights of slaves in Trinidad’, Social Science History VI, 4 (1982).Google Scholar For Latin American practice, see Klein, , Middle Passage, 223Google Scholar; de Queiros Mattoso, K. M., Etre esclave au Brésil, xvie–xixe Steele (Paris, 1979), 97.Google Scholar

¹⁵ Unfortunately, the relative proportions of young children and teenagers were not at all constant. In the Cuban trade children under eleven were much more numerous than those aged eleven to seventeen. However, among imports into Parã/Maranhão the opposite was true, and in the British trade to Spanish America, 1715–35, children under ten were greatly outnumbered by those aged ten to fourteen. This also seems to have been the case in Saint Domingue.

¹⁶ Klein, in La traite des noirs, 85.Google Scholar The extremely low percentages of children reported in Portuguese records prior to 1811 seem scarcely credible in view of the very high percentages obtained for the post-1811 period from more trustworthy sources. Note also the uncertainty surrounding the British trade of the 1790s: above, table 1, n. 13. Cf. above, note 7, and Miller's, J. comments in Actes du Colloque International sur la Traite des Noirs, Nantes, 1985, 2 vols., forthcoming, vol. I, 1–33.Google Scholar

¹⁷ Klein, , Middle Passage, 150Google Scholar; Robertson, and Klein, , Women, 31, 33.Google Scholar

¹⁸ Geggus, D., ‘Slaves of British-occupied Saint Domingue: an analysis of the work-forces of 197 absentee plantations’, Caribbean Studies, XVIII, 1 (1978), 24n.Google Scholar; Inikori, , Forced Migration, 23.Google Scholar

¹⁹ Robertson, and Klein, , Women, 31, 33.Google Scholar

²⁰ Inikori, , Forced Migration, 23.Google Scholar

²¹ It is perhaps relevant that, while Africans described as Congos and Igbos had approximately the same average height in the British as in the French colonies, Man-dingoes in Jamaica were three to four inches shorter than in Saint Domingue and Trinidad: Geggus, D., Saint Domingue Slave Revolt, forthcoming, ch. 2.Google Scholar

²² de Saint-Mery, M. L. E. Moreau, Description topographique…, de l'isle Saint-Domingue [1797] (Paris, 1958), vol. I, 49–50.Google Scholar

²³ Sex ratio: 150; 46 per cent men, 31 per cent women, 24 per cent children; (7,419 slaves).

²⁴ See Curtin, , Census, 150Google Scholar; Curtin, P. D., ‘Measuring the Atlantic slave trade’, in Engerman, S. and Genovese, E. (eds), Race and Slavery in the Western Hemisphere: Quantitative Studies (Princeton, 1975), 111–26.Google Scholar

²⁵ See Curtin, , Census, 200.Google Scholar

²⁶ Marrero, L., Cuba: Economia y Sociedad (Madrid, 1972), vol. 9, 38–56.Google Scholar

²⁷ Curtin, , Census, 247.Google Scholar

²⁸ Fraginals, Manuel Moreno, ‘Africa in Cuba’, in Rubin, V. and Tuden, A. (eds), Comparative Perspectives on Slavery in New World Plantation Societies (New York, 1976), 190.Google Scholar

²⁹ However, the rise in sex ratio may also have been due to increased numbers of Hausa slaves being sent down the Niger to Brass and New Calabar.

³⁰ Sample sizes were as follows: 1716–53: 10,897; 1764–78: 47,317; 1784–92: 4,851.

³¹ Children, apparently aged under seven or eight, dropped from 127 per cent of exported slaves in 1733–5, to 7.7 per cent in 1749–52, and for the period 1734–59 made up 72 per cent. After 1759, children formed under 1 per cent of those recorded. See Goulart, M., Escravidão africana no Brasil (São Paulo, 1950), 203–9.Google Scholar This latter reduction may derive partly from the introduction of a new system of record-keeping, which henceforth took no account of babies. They, however, had represented less than 2 per cent of the 1733–5 and 1749–52 samples.

³² See Peukert, W., Der atlantische Sklavenhandel von Dahomey, 1740–1797 (Wies-baden, 1978), 64–65Google Scholar; Law, R., The Oyo Empire, c. 1600–c. 1836 (Oxford, 1977), 222–3.Google Scholar

³³ Sample sizes were Porto Novo and Badagri: 4,028; Bight of Benin: 41,121. Slaves exported from Epe in the middle of the coast had a demographic profile midway between those of the eastern and western ports.

³⁴ Prior to Eltis's recent Economic Growth, which finds a comparable pattern for the nineteenth century, only Inikori, Forced Migration, 23 recorded a low regional sex ratio – among 4,813 slaves shipped from Whydah to Jamaica in the period 1764–88. Data showing a relatively high sex ratio, but deriving from more restricted samples, appear in Robertson and Klein, , Women, 33Google Scholar, and Lovejoy, , Transformations, 62.Google Scholar

³⁵ Law, Robin, ‘Dahomey and the slave trade: reflections on the historiography of the rise of Dahomey, J. Afr. Hist. XXVII, 2 (1986), 237–67.CrossRefGoogle Scholar

³⁶ Sample sizes for the three periods were 8,583, 7,554, and 13,933. No data were available for the years 1727–34.

³⁷ Inikori, , Forced Migration, 22–3.Google Scholar

³⁸ The number of slaves in each group was 6,395, 3,983, 2,167.

³⁹ Curtin, , Senegambia, vol. I, 180.Google Scholar

⁴⁰ Curtin, , Senegambia, vol. I, 176Google Scholar; Geggus, , ‘Slaves’, 23–4.Google Scholar However, the extremely high regional sex ratios Curtin found in the 1680s and 1720s would appear quite unusual.

⁴¹ See Curtin, P., ‘Abolition’, in Walvin, J. and Eltis, D. (eds), The Abolition of the Atlantic Slave Trade (Madison, 1981), 86.Google Scholar

⁴² Eltis, , Economic Growth, 60.Google Scholar On export volume see Lovejoy, , Transformations, 50Google Scholar; Manning, , Dahomey, 35.Google Scholar

⁴³ Because of the uneven importance of a foreign contraband slave trade in different parts of Saint Domingue, and the contrasting preferences displayed by planters of different crop-types for slaves of certain ethnic groups, the ethnic make-up of the slave population was subject to substantial local variations. Future efforts to reconstruct the ethnic composition of the French slave trade from these sources need to take this into account: Geggus, , ‘Slaves’, 14–23Google Scholar; Geggus, , ‘Les esclaves de la plaine du Nord à la veille de la Révolution française, IV,’ Revue de la société haitienne d'histoire, 149 (1985), 23Google Scholar; above, note 9.

⁴⁴ Littlefield, D., Rice and Slaves: Ethnicity and the Slave Trade in Colonial South Carolina (Baton Rouge, 1981), 21–34Google Scholar; Geggus, Slave Revolt, ch. 2; Geggus, , ‘Slaves’, 14–33.Google Scholar

⁴⁵ Both ‘Miserables’ and ‘Mesurades’ were as yet unidentified peoples from the area of Cape Mesurado. The latter were also known in the South and West Provinces of Saint Domingue as ‘Canga’, an Akan term meaning ‘barbarous outsider’. It was seemingly bestowed on them by their easterly neighbours, who were more numerous in those parts of the colony, being introduced by British contraband slavers.

⁴⁶ See the data in Higman, ‘Trinidad’, table 2, in Crahan and Knight, Africa. Unfortunately, Higman gives only regional but not ethnic sex ratios.

⁴⁷ See Geggus, , ‘Slaves’, 17–20, 24, 30.Google Scholar

⁴⁸ Akan and Ga speakers were known as ‘ Coromantees’ in the British West Indies, where they were quite numerous and highly prized. ‘Caramenty’ were rarely found outside those parts of Saint Domingue where the contraband slave trade flourished: Geggus, D., ‘On the eve of the Haitian revolution’, in Heuman, G. (ed.), Out of the House of Bondage: Runaways, Resistance and Marronage in Africa and the New World (London, 1986), 122.Google Scholar

⁴⁹ Although the French distinguished ‘Aoussa’ from ‘Gambary’, I follow Law, , Oyo Empire, 227nGoogle Scholar in assuming that the terms were synonyms, ‘Gambary’ being the Yoruba word for Hausa.

⁵⁰ The sole exception would seem a footnote in Gautier, Soeurs, 80.

⁵¹ Robertson, and Klein, , Women, 61.Google Scholar

⁵² Although numerous peoples were no doubt enveloped in the generic term ‘Congo’, a majority were probably Kikongo speakers, either native or ‘naturalized’. This is suggested by linguistic evidence from colonial and modern Haiti. The chief sub-groups identified in colonial records, apart from Kongo, were Nsundi, Yombe, Mbala and Yaka. While regional trade routes extended far into the interior, there is little evidence even in the nineteenth century that the most distant sources supplied many captives to the coast: Curtin, , Census, 256, 295–6Google Scholar; Degrandpre, L., Voyage a la cote occidentale d'Afrique fait dans les années 1786 et 1787 (Paris, 1801), vol. I, xiv–xxv, vol. 2, 37, 48.Google Scholar For a contrary view see Miller, Way of Death. Among the ‘ Mozambiques’ the only peoples identified were Makonde and Makwa.

⁵³ Igbo constituted 91 per cent of the region's slaves in the French plantation sample; but only 52 per cent in the 1813 Trinidad sample (excluding Hausa); and 68 per cent in a sample of 1,008 slaves sold in 1821–2, 26 per cent being Ibibio: Northrup, D., Trade Without Rulers (Oxford, 1978), 231.Google Scholar Although the latter sample, drawn from only two ports, is not strictly representative, it is broadly corroborated by the 1848 Sierra Leone census of freed slaves, which similarly shows a much increased ratio of N. W. Bantu to Igbo: Northrup, , Trade, 60–1.Google ScholarEltis, , Economic Growth, 358–9Google Scholar, however, found that in this period Igbo and Ibibio captives exhibited similar sex ratios.

⁵⁴ In West Africa with the exception of Senegambia the coastal price of adult males in the eighteenth century usually exceeded that of adult females by 25 to 40 per cent: Atkins, J., Voyage to Guinea, Brasil and the West Indies (London, 1735), 163–6Google Scholar; Barbot, J., Description of the Coasts of North and South Guinea (London, 1732), 326Google Scholar; Peukert, , Sklavenhandel, 112Google Scholar; Curtin, , Senegambia, vol. I, 173, 176Google Scholar; below, note 83.

⁵⁵ Miers, and Kopytoff, , Slavery, 22, 31Google Scholar; Lovejoy, , Transformations, 118.Google Scholar

⁵⁶ Descriptions of the sexual division of labour were taken from Murdock, G. P., Ethnographic Atlas (Pittsburgh, 1967)Google Scholar; Vansina, J., Ethnographie du Congo (Kinshasa, nd), 14Google Scholar; Manning, , Dahomey, 70–71.Google Scholar American planters frequently found the women among their Igbo and Congo slaves better fieldworkers than their male counterparts, whereas among Bight of Benin slaves males and females were equally prized: Edwards, B., History Civil and Commercial of the British West Indies (Dublin, 1793), vol. 2, 68, 70Google Scholar; Littlefield, , Rice, 18, 151Google Scholar; de Saint-Méry, Moreau, Description, vol. I, 51, 53Google Scholar; Malenfant, C., Des colonies modernes (Paris, 1814), 210–11Google Scholar; Long, E., History of Jamaica (London, 1774), vol. 2, 403–4Google Scholar; Ducoeurjoly, S., Manuel des habitants de Saint Domingue (Paris, 1802), vol. I, 24.Google Scholar Some also made the same observation of the Akan-speaking ‘Coromantee’: Long, , Jamaica, vol. 2, 446, 471–3.Google Scholar

⁵⁷ See Goody, J., ‘Slavery in time and space’, in Watson, J. L. (ed.), Asian and African Systems of Slavery (Berkeley 1980), 38–42Google Scholar; Thornton, J., ‘The slave trade in eighteenth century Angola’, Canadian Journal of African Studies, XIV, 3 (1980), 424–5.Google Scholar

⁵⁸ Lovejoy, , Transformations, 114, 116, 119–20Google Scholar; Manning, , Dahomey, 10, 39Google Scholar; Meil-lassoux, , ‘Female slavery’, in Robertson, and Klein, , Women, 61.Google Scholar

⁵⁹ Northrup, , Trade, 75–77Google Scholar; Brown, S. D., ‘From the tongues of Africa: a partial translation of Oldendorp's interviews’, Plantation Society, II, 1 (1983), 49.Google Scholar

⁶⁰ Of course, the two categories were not necessarily mutually exclusive. See Edwards, , History, 56–59Google Scholar; Long, , Jamaica, vol. 2, 473Google Scholar; de Saint-Mery, Moreau, Description, vol. I, 48, 50.Google Scholar

⁶¹ Originally advanced by J. D. Fage, this view is espoused in Manning, , Dahomey, 32, 42Google Scholar, and Northrup, , Trade, 80–4.Google Scholar

⁶² This is argued in Thornton, , ‘Angola’, 426–7Google Scholar; A. van Dantzig, ‘Effects of the Atlantic slave trade on some West African societies’, in Inikori, , Forced Migration, 200Google Scholar; Lovejoy, , Transformations, 64–5.Google Scholar

⁶³ Inikori, , Forced Migration, 25.Google Scholar

⁶⁴ Manning, , Dahomey, 371.Google Scholar

⁶⁵ Cf. Edwards, B., Historical Survey of the Island of Saint Domingo (London, 1801), 268–71Google Scholar, which contrasts slave shipments from Bonny and Bassam consisting of one-half children with a shipment from the Gold Coast with only one-tenth children.

⁶⁶ Galenson, , Traders, 104–10.Google Scholar

⁶⁷ Curtin, , Senegambia, vol. I, 175.Google Scholar However, this may have been true only in the early eighteenth century, when the Portuguese were seeking children for the Iberian market. Cf. Atkins, , Voyage, 177.Google Scholar

⁶⁸ On ethnic preferences and price levels see Curtin, , Census, 155–62, 181–3, 189–90, 208–9Google Scholar; Littlefield, , Rice, 8–21, 45–7, 54Google Scholar; Eltis, , Economic Growth, 264Google Scholar; Craton, M., Walvin, J. and Wright, D., Slavery, Abolition and Emancipation (London, 1976), 51–3Google Scholar; Miller, ‘Slave Prices’, in Lovejoy, (ed.), Africans in Bondage, 66–8Google Scholar; Bowser, F., African Slavery in Peru (Stanford, 1974), 80Google Scholar; Donnan, , Documents, vol. 2, ‘Report on the Trade to Africa, 1709’, 56Google Scholar; Palmer, , Cargoes, 15, 62Google Scholar; Geggus, , ‘Slaves’, 14–18, 19nGoogle Scholar; Geggus, , ‘Toussaint Louverture and the Slaves of the Bréda plantations’, Journal of Caribbean History (1985–1986), 46Google Scholar; LeVeen, E. P., British Slave Trade Suppression Policies (New York, 1977), 114–15, 146.Google Scholar

⁶⁹ See above, note 56.

⁷⁰ Davies, , Royal African Company, 78.Google Scholar

⁷¹ Klein, in Traite des Noirs, 83Google Scholar; Klein, , Middle Passage, 151, 241–2Google Scholar; Robertson, and Klein, , Women, 34–6Google Scholar; Patterson, O., ‘Recent Studies on Caribbean Slavery and the Atlantic Slave Trade’, Latin American Research Review XVII (1982), 272–3.Google Scholar

⁷² See Robertson, and Klein, , Women, 4–5, 10–11, 67–89Google Scholar; Curtin, , Senegambia, vol. I, 175–7Google Scholar; Miers, and Kopytoff, , Slavery, 21–2, 53, 62, 125, 297, 320Google Scholar; Harries, P., ‘Slavery, social incorporation and surplus extraction’, J. Afr. Hist, XXII (1981), 322–6Google Scholar; Tambo, D., ‘The Sokoto Caliphate slave trade in the nineteenth century’, International Journal of African Historical Studies IX (1976), 192–202Google Scholar; Van Wing, J., Etudes Bakongo: Histoire et Sociologie (Bruxelles, 1921), vol. I, 136.Google Scholar

⁷³ See MacCormack, , ‘Slaves, slave owners, and slave dealers’, in Robertson, and Klein, , Women, 288Google Scholar; Miers, and Kopytoff, , Slavery, 108, 161–2.Google Scholar

⁷⁴ See above, note 54.

⁷⁵ Bean, R., British Transatlantic Slave Trade, 1650–1775 (New York, 1975), 134.Google Scholar Galenson's sophisticated analysis of slave sales in Barbados in the period 1673–1723 also reveals a male:female price ratio of 100:81: Traders, 62.

⁷⁶ See Table 1; above, note 10; Dunn, R., Sugar and Slaves: the Rise of the Planter Class in the English West Indies (London, 1973), 316Google Scholar; French Caribbean censuses in Arch. Nat., Aix en Provence, GI/509.

⁷⁷ Littlefield, , Rice, 56–66, 72.Google Scholar

⁷⁸ See above, note 9.

⁷⁹ Despite a large urban market for female slaves, Spanish American demand was usually considered by contemporaries strongly pro-male: Littlefield, , Rice, 59–60.Google Scholar Cf. Table 1. The two French slave cargoes listed by Mettas that were sold in the Spanish Caribbean significantly had a sex ratio of 204:100.

⁸⁰ See Galenson, , Traders, 94–6.Google Scholar

⁸¹ See, Klein, , Middle Passage, 222–5Google Scholar; Fraginals, Moreno, ‘Africa’, 192–3Google Scholar; Eltis, , Economic Growth, 257.Google Scholar

⁸² Klein, , Middle Passage, 32–4.Google Scholar

⁸³ Isolated data suggest that on the Gold Coast the male:female price ratio rose between 1720 and 1770 from about 100:67 to 100:80, then fell back to around 100:77 in the mid-1770s as slave prices declined: see Littlefield, , Rice, 62nGoogle Scholar; Metcalf, G., ‘Gold, Assortments and the Trade Ounce: Fante Merchants and the Problem of Supply and Demand in the 1770s’, J. Afr. Hist, XXVIII (1987), 37.Google Scholar

⁸⁴ LeVeen, , Suppression, 117–21Google Scholar; Manning, , Dahomey, 37.Google Scholar Eltis, Economic Growth, passim. The widening gap between African and American slave prices may additionally account for the high percentages of children that also were a feature of the nineteenth century trade.