¹ See, for example, Fogel, Robert William, “Nutrition and the Decline in Mortality since 1700: Some Preliminary Findings,” in Engerman, Stanley L. and Gallman, Robert E., eds., Long-Term Factors in American Economic Growth (Chicago, 1986), pp. 439–527.Google Scholar The basic reference work on the physiology of this issue is Eveleth, P. B. and Tanner, J. M., Worldwide Variation in Human Growth (Cambridge, 1976).Google Scholar

² Steckel, Richard H., “Height and Per Capita Income,” Historical Methods, 16 (Winter 1983), pp. 1–7.CrossRefGoogle Scholar

³ Fogel, Robert W. et al. , “Secular Changes in American and British Stature and Nutrition,” in Rotberg, Robert I. and Rabb, Theodore K., eds., Hunger and History: The lmpact of Changing Food Production and Consumption Patterns on Society (Cambridge, 1983), pp. 247–83;Google Scholar and the articles in the special issue of Social Science History, 6 (Fall 1982). edited by Robert W. Fogel and Stanley L. Engerman;Google ScholarHigman, B. W., “Growth in Afro-Caribbean Slave Populations,” American Journal of Physical Anthropology, 50 (03 1979). pp. 373–85;CrossRefGoogle ScholarEltis, David, “Nutritional Trends in Africa and the Americas: Heights of Africans, 1819–1839.” Journal of Interdisciplinary History, 12 (Winter 1982), pp. 453–75;CrossRefGoogle ScholarPubMedJohnson, T. O., “Height and Weight Patterns of an Urban AfricanPopulation Sample in Nigeria,” Tropical and Geographical Medicine, 22 (03. 1970), pp. 65–76Google ScholarPubMed; Eveleth and Tanner, Worldwide Variation in Human Growth, p. 340.Google Scholar

⁴ Tobias, Phillip V., “The Negative Secular Trend,” Journal of Human Evolution, 14 (05 1985), PP. 347–56. It might be noted that very few West African and no Yoruba peoples are included in the Tobias survey.CrossRefGoogle Scholar

⁵ Hiernaux, Jean, La diversité humaine en Afrique subsaharienne: Recherchés biologiques (Brussels, 1968). See further development of this point below.Google Scholar

⁶ See Eltis, David, Economic Growth and the Ending of the Transatlantic Slave Trade (New York, 1987), p. 400Google Scholar; and Eltis, “Nutritional Trends in Africa and the Americas,” pp. 453–57. The copies of the registers are scattered in the Foreign Slave Trade series (FO 84) of the British Public RecordOffice, volumes 4, 9, 15, 21, 38, 63, 64, 76, 86, 87, 98, 100, 101, 102, 110, 116, 127, 134, 135, 147, 166, 189, 212, 231, 269, 273, 308, 310, 345, 346, 393, 450, 507, 559.Google Scholar

⁷ Children's growth curves calculated from the Sierra Leone data are improbably steep as well as bimodal in distribution. The reporting bias that almost certainly lies at the root of these features likely has two sources. First, the British paid their naval crews a bounty for each slave landed alive. This bounty was greater for adults than for children (see Lloyd, Christopher, The Navy and the Slave Trade [London, 1949, pp. 79–84). and it is possible that to reduce costs, court officials underestimated the ages of young adults. Second, European unfamiliarity with the growth patterns of Africans, and the fact that officials had in any event to estimate African ages, could also have contributed to reporting bias.Google Scholar

⁸ Heat dissipation occurs through the skin. A low mass-to-surface ratio, which facilitates cooling, is a characteristic of both shorter body types and tall, slender physiques. According to Hiernaux, which of these adaptive routes is taken depends on humidity, because in hot dry climates sweating is a more efficient method of cooling than in humid environments. For a full review, see the work of Hiei-naux, Jean, in particular La diversité humaine and the summary in The People of Africa (New York, 1974). pp. 65–90.Google Scholar

⁹ For twentieth-century data on stature, see Eveleth and Tanner, Worldwide Variation in Human Growth, pp. 338–39;Google ScholarTalbot, P. A. and MuIhall, H., The Physical Anthropology of Southern Nigeria (Cambridge, 1962), p. 82Google Scholar; Hiernaux, La diversité humaine, pp. 167–76.Google Scholar

¹⁰ See citations in fn. 9. These data are conveniently drawn together in Eltis, “Nutritional Trends in Africa and the Americas,” p. 462.Google Scholar

¹¹ Over half of the Sierra Leone recaptives spoke Yoruba dialects. Data on countries of origin in the Sierra Leone registers indicate that all Yoruba left from the six major embarkation points for slaves in the Bight of Benin. The registers also show that the most easterly of the six, 13enin, shipped some Igbo peoples as well as Nupe and Hausa, though most of these left from points farther east. We can infer that the two westernmost, Little and Grand Popoe, must have been the source of the significant Ewe, Gun, and Adja populations at Sierra Leone. Some Adja may have left Whydah and this embarkation area also is ignored. The Yoruba must have made up the overwhelming majority of those leaving Badagry and Lagos because few ports in this region other than the ones discussed embarked slaves, and the large numbers of Yoruba could not have set sail from anywhere else. There were relatively few Nupe, Hausa, and other peoples living to the north of the Yoruba at Sierra Leone, and some of these embarked in the Bight of Biafra rather than points to the east. See Koelle, Sigmund W. in Poivglotta Africana, Hair, Philip E. H. and Dalby, David, eds. (Sierra Leone, 1963)Google Scholar; Hair, P. E. H., “The Enslavement of Koelle's Informants,” Journal of African History, 6 (No. 2, 1965), pp. 193–203CrossRefGoogle Scholar; Curtin, Philip D. and Vansina, Jan, “Sources of the Nineteenth Century Atlantic Slave Trade,” Journal of African History, 5 (No. 2. 1964), pp. 185–208CrossRefGoogle Scholar; Eltis, Economic Growth, pp. 165–80. For the purposes of this study, then, “adult Yoruba” are defined as Africans embarking at Badagry and Lagos (even though only Lagos was actually a Yoruba town) in the period from 1821 to 1845, with a recorded age of between 23 and 40.Google Scholar

¹² See Eltis, “Nutritional Trends in Africa and the Americas,” for a fuller discussion of this.Google Scholar

¹³ These data are from Trinidad, which received most of its slaves between the British takeover of the island in 1797 and 1805, when an order-in-council prohibited direct shipments of slaves from Africa. For a few years thereafter slaves arrived from other British territories, though many of these slaves had been born in Africa.Google Scholar

¹⁴ The latest information on this is in Richardson, David, “Slave Exports from West and West-Central Africa, 1700–1810: New Estimates of Volume and Distribution,” Journal of African History, 30 (No. 1, 1989), pp. 1–22.CrossRefGoogle Scholar For the U.S. distribution. Richardson uses Donnan, Elizabeth, Documents Illustrative of tile History of the Slave Trade to America, 4 vols. (Washington, 1935), vol. 4, pp. 474, 475, 477, 490, 504, 508, 513–15, 521–22, 633.Google Scholar The African distribution of the British slave trade is explained more fully in Richardson, David, “The Eighteenth Century British Slave Trade: Estimates of its Volume and Coastal Distribution in Africa,” Research in Economic History, 12 (1989), pp. 151–95.Google Scholar As pointed out in the notes to Table 1, the U.S. and British weights are based on the share of each country's slave trade rather than arrivals in U.S. and British territory. The British Caribbean was overwhelmingly supplied by British slavers at this time, however. In the U.S. case the British and Americans were together responsible for most of the slaves arriving in the early nineteenth century, but the distribution of the two trades is rather similar. For the African origins of the Cuban and Brazilian traffic in slaves after 1820, see Eltis, David, “The Export of Slaves from Africa, 1821–1843,” this JOURNAL, 37 (06 1977), p. 423.Google Scholar

¹⁵ In the eighteenth century St. Domingue drew heavily from the Bight of Benin, and slaves from this area were not noted for their tall stature. See the discussion in Fouchard, Jean, The Haitian Maroons: Liberty or Death, trans. by Watts, A. Faulkner (New York, 1981), pp. 177–81.Google Scholar

¹⁶ Higman, Barry, Slave Populations of the British Caribbean, 1807–1833 (Baltimore, 1984), pp. 282–83, shows that slaves born in non-sugar colonies were taller than those born in newly developing sugar-producing regions.Google Scholar The evidence is not conclusive, however, as slaves on Trinidad sugar estates were taller than those on non-sugar-producing units. See Friedman, Gerald C., “The Heights of Slaves in Trinidad,” Social Science History, 6 (Fall 1982), pp. 482–515CrossRefGoogle ScholarPubMed; and John, A. Meredith, Plantation Slaves in Trinidad, 1783–1816: A Mathematical Inquiry (Cambridge, 1988), pp. 117–18.Google Scholar

¹⁷ While it is not my main concern here, it is worth noting that about a quarter of the difference in height between slaves in the United States and slaves in the British Caribbean is explained by the shorter initial stature of new arrivals in the British Caribbean compared to those arriving in the United States. Indeed, after allowing for the unrepresentativeness of the Cuban data (see text), it is clear that slaves arriving in the United States were generally taller than slaves arriving in any other major region in the Americas apart from Bahia. The tall stature of U.S. and Bahian arrivals is explained by the fact that the United States drew heavily from the Senegambia and Sierra Leone areas, where the taller Susu and Mandingo peoples embarked, and Bahia drew almost exclusively from the Bight of Benin-which in the nineteenth century meant Yoruba-regions.Google Scholar

¹⁸ Margo, Robert A. and Steckel, Richard H., “The Heights of American Slaves: New Evidence on Slave Nutrition and Health,” Social Science History, 6 (Fall 1982), pp. 522–23.CrossRefGoogle ScholarPubMed

¹⁹ Time profiles for both non-Yoruba and Yoruba data were fitted with higher-order polynomials, but no improvement in significance levels resulted.Google Scholar

²⁰ One indeterminate, but probably small, influence would result in some further understatement of the difference. Mortality rates in Trinidad were greater for slaves of shorter stature; see Friedman, “The Heights of Slaves in Trinidad,” p. 488. There is the possibility that the substantial mortality associated with the slave trade yielded a survival group in the Americas slightly taller than the African population from which these survivors were drawn. Most of the Sierra Leone sample had already undergone this selection process because the ocean voyage to Sierra Leone from most parts of Africa was as severe in terms of duration and loss of life as the transatlantic voyage;Google Scholar see Eltis, David, “Mortality and Voyage Length in the Middle Passage: New Evidence from the Nineteenth Century,” this JOURNAL, 44 (06 1984), pp. 301–8. None of the samples of the African-born in Table I were collected in the homelands before the slave trade took its effect. Thus the height difference between the Sierra Leone sample and samples from the Americas could understate the true difference between Africa and the Americas. It should also be noted that very few Africans arrived in the United States after 1807 and even fewer after the Piracy Act of 1821—the period covered by the Sierra Leone data. Thus the non-Yoruba series as used in this context is a projection for comparative purposes based on pre-1807 trends of U.S. imports.Google ScholarPubMed

²¹ A second-degree polynomial significant at the I percent level suggests a 2.3-centimeter decline between 1801 and 1821.Google Scholar

²² See fn. 10.Google Scholar

²³ Morton-Williams, Peter, “The Oyo Yoruba and the Atlantic Trade, 1670–1830,” Journal of the Historical Society of Nigeria, 3 (12 1964), pp. 25–45Google Scholar, reprinted in Inkori, J. E., ed., Forced Migrations: The Impact of the Export Slave Trade on African Societies (New York, 1982), pp. 167–86.Google Scholar

²⁴ Steckel, Richard H., “A Peculiar Population: The Nutrition, Health, and Mortality of American Slaves from Childhood to Maturity,” this JOURNAL, 46 (09 1986), pp. 721–41.Google Scholar

²⁵ Curtin, Philip D., “Epidemiology and the Slave Trade,” Political Science Quarterly, 83 (06 1968), pp. 190–216.CrossRefGoogle ScholarPubMed

²⁶ Morgan, W. B. and Pugh, J. C., West Africa (London, 1969), pp. 112–13. For height data on Gold Coast slaves, see Higman, “Growth in Afro-Caribbean Slave Populations,” p. 364.Google Scholar

²⁷ See fn. 10.Google Scholar

²⁸ “In stature the Egboes differ very considerably according to the region whence they are taken. In Isuama and the north-central districts the inhabitants are tall and majestic looking, some well formed and shaped… In other parts they are of middling size, averaging five feet two. In the delta of the river the inhabitants are short and stout…scarcely above four and a half feet.” West African Countries and Peoples (Edinburgh, 1969, reprint of 1868 edition), p. 157. It might be added that both the Sierra Leone and modern data suggest that Horton somewhat exaggerated the intra-Igbo differentials.Google Scholar

²⁹ Johnston, Bruce F., The Staple Food Economies of Western Tropical Africa (Stanford, 1958), pp. 55–90.Google Scholar

³⁰ There is also the point that the Oyo absorbed several non-Yoruba groups into the Yoruba socioeconomic “system” during its period of hegemony. On the basis of the above evidence, these non-Yoruba were very probably of shorter stature than the dominant group. Once more the relatively low standard deviations in the Yoruba Sierra Leone data indicate that such non-Yoruba groups quickly came to share in the Yoruba advantage over neighboring peoples.Google Scholar

³¹ For 1911 and 1921 data, see Talbot, P. Amaury, The Peoples of Southern Nigeria, 4 vols. (London, 1926), vol. 4, pp. 1–15.Google Scholar

³² The fullest assessment of the historical Yoruba economy is in Law, Robin, The Oyo Empire, c. 1600–c. 1836 (Oxford, 1977), pp. 202–36Google Scholar, but the orientation of both this and the same author's “Slaves, Trade and Taxes: The Material Basis of Political Power in Precolonial West Africa,” Research in Economic Anthropology, 1 (1978), pp. 37–52, is toward explaining the sources of state economic power.Google Scholar

³³ Clarke, W. H., in Travels and Exploration in Yorubaland (1854–1858), Atanda, J. A., ed. (Ibadan, 1972), p. xxxviii.Google Scholar

³⁴ Law, , The Ovo Empire, pp. 8–10Google Scholar; Clapperton, H., Journal of a Second Expedition into the Interior of Africa (London, 1829), p. 13Google Scholar; Bascom, William, “Urbanism as a Traditional African Pattern,” The Sociological Review, 7 (07 1959), pp. 29–43.CrossRefGoogle Scholar

³⁵ See the missionary map of Yorubaland containing population estimates for urban centers from the early 1850s reproducedGoogle Scholar in Clarke, Travels and Explorations, p. xxxviii;Google Scholar also Bowen, T. J., Central Africa: Adventures and Missionary Labours (Charleston, 1857), p. 218.Google Scholar

³⁶ Skinner, E. P.,“West African Economic Systems,” in Herskovitz, M. J. and Harwitz, M.. eds., Economic Transition in Africa (London, 1964), p. 97Google Scholar; Mabogunje, Akin and Omer-Cooper, J. D., Owu in Yoruba History (Ibadan, 1971), pp. 42–43.Google Scholar

³⁷ Urbanization might increase the incidence of diseases spread by congestion, such as intestinal complaints, though presumably nothing like the conditions in the slave trade that produced enormous mortality rates from dysentery existed in cities or indeed anywhere else in Africa under normal conditions. The major diseases in West Africa tend, in any event, to be endemic rather than epidemic, and it is hard to see how urbanization would reduce their impact. Johnston, Staple Food Economies, pp. 55–90, 200–1;Google ScholarNicol, B. M., “Nutrition of Nigerian Peasant Farmers, with Special Reference to the Effects of Vitamin A and Riboflavin Deficiency,” British Journal of Nutrition, 3 (1949), pp. 25–43CrossRefGoogle ScholarPubMed; idem, “The Nutrition of Nigerian Peasants, with Special Reference to the Effects of the vitamin B complex, Vitamin A, and Annual Protein,” British Journal of Nutrition, 6 (1952), pp. 34–55; U.S. Department of Health, Education, and Welfare, Republic of Nigeria: Nutrition Survey: February-April 1965 (Washington, 1967).Google Scholar Regarding specialized food production, see Talbot, , The Peoples of Southern Nigeria, vol. 3, p. 699.Google Scholar

³⁸ Iliffe, John, The African Poor: A History (Cambridge, 1987), PP. 85–86.CrossRefGoogle Scholar

³⁹ Steckel, “Height and Per Capita Income”: Margo, Robert A. and Steckel, Richard H.,“Heights of Native-Born Whites During the Antebellum Period,” this JOURNAL, 43 (03 1983), pp. 167–74.Google ScholarPubMed

⁴⁰ Nearly 12 percent of a sample of 13,334 slaves on eighteenth-century St. Domingue plantations have been identified as Yoruba; see Geggus, David, “Sex Ratio, Age and Ethnicity in the Atlantic Slave Trade: Data from French Shipping and Plantation Records,” Journal of African History, 30 (No. 1, 1989), p. 32.CrossRefGoogle Scholar

⁴¹ A similar case could be made for a nutritional component in the height advantage of West Africans as a whole compared to those from west-central Africa. Specialization in the production of both food and nonfood items seems to have been greater in most West African societies compared to most west-central African groups. Genetic variation was probably greater in the latter region than the former, but it would be difficult to argue that the humidity and heat combinations discussed by Hiernaux explain stature differences between these two broad nineteenth-century slave provenance zones.Google Scholar

⁴² The following is based on Smith, Robert S., Kingdoms of the Yoruba (3rd edn., Madison, 1988), pp. 109–24;Google ScholarMabogunje and Omer-Cooper, Owu in Yoruba History, pp. 19–82; and especially Law, The Oyo Empire, pp. 245–77.Google Scholar

⁴³ Fogel, Robert William, “Nutrition and the Decline in Mortality since 1800: Some Preliminary Findings,” pp. 497–502; idem et al., “Secular Changes in American and British Stature and Nutrition.”Google Scholar

⁴⁴ The only other active slave traders at this time were the British and the Americans. In the 1790s and early 1800s neither group traded heavily in the Bight of Benin. See Richardson, “Slave Exports from West and West-Central Africa.” For slave departures from the Bight of Benin after 1810, see Eltis, Economic Growth, p. 250.Google Scholar

⁴⁵ Eltis, Economic Growth, p. 243.Google Scholar

⁴⁶ The omission of the dummy variable from the equation does nothing to impair the systematic relationship between final heights and slave departures. A simple regression of mean annual stature on slave exports yielded a reduced R² of 0.34, but a slave export coefficient significant at the 0.001 level.Google Scholar

⁴⁷ Law, The Oyo Empire, pp. 305–6; Eltis, Economic Growth, p. 74.Google Scholar The traditional view is represented by Akinjogbin, I. A., “The Prelude to the Yoruba Civil Wars of the Nineteenth Century,” Odu, 2nd series, 1 (1964), pp. 24–46.Google Scholar

⁴⁸ A third possible interpretation is that as the volume of slave exports increased, slave traders were forced to take shorter slaves. Given the fact that the final heights here are grouped into year of birth cohorts, this is very unlikely. If this tendency was in fact a feature of the trade, we would expect final heights grouped by year of export (rather than year of birth) to vary inversely with annual slave departures. To test for this, an ordinary least squares regression equation was estimated with mean heights in year of departure as the dependent variable, and annual slave departures from the Bight of Benin as the independent variable. The sign of the coefficient was negative, but the result was nonsignificant and the R² less than 0.02. Rerunning the regression with lagged values did not improve the result, It should also be noted that the distribution of slaves from all regions was normal, whereas a predisposition for taller slaves would have made the distribution of recaptives skewed. See Eltis, “Nutritional Trends in Africa and the Americas,” for a discussion of this last point.Google Scholar