COMPARED to the information now available on the shipping part of the slave trade, and, indeed, on the experiences of slaves after they arrived in the Americas, we still know little about the people who were forced on board slave vessels on the African coast and only slightly more on how they came to be captives. Apart from Sigmund Koelle's Sierra Leone survey from the 1840s and the fragmentary accounts of European travelers, merchants, missionaries and those who carried out the slaving activities stretching back to the sixteenth century, the sources for the African end of the slave trade are sparse indeed. And of African sources on the subject (unmediated by Europeans), there are few, if any. This situation is partly responsible for the strongly divided historiography that exists on issues such as the role of slavery in Africa, the impact of the slave trade and which people became captives. With scholars so far away from consensus on important issues, it is hardly surprising that the image in the public mind of the African operations of the slave trade seems further removed from reality than public perceptions of almost any other historical issue.
One small step toward adding detail to the story of the African end of the traffic is made possible by the use of a source long-known to historians, but severely under-exploited -- the Registers of Liberated Africans made in Sierra Leone, Havana and Rio de Janeiro between 1819 and 1845. (1) This source, in effect, allows us to identify for the first time the ethnic origins of those put on board slave ships from evidence provided by the captives themselves. The Rio registers do not appear to have survived but the other two courts processed several hundred vessels (many captured prior to embarking slaves) and over 67,000 individuals. Each liberated African falling within the jurisdiction of the court was asked to provide name, age and place of habitation. The court added the height and sex of the person and a description of the most obvious cicatrization. (2) In over 99 per cent of the cases, the name that was recorded was clearly African. All information appears to have been mediated by an African interpreter, usually an earlier recaptive from the same part of the coast. In the Havana registers, the identity of the interpreter is recorded. Despite the fact that the practice of recording place of habitation was quickly discontinued, the new data provide a basis for identifying region of origin of each recaptive, without traversing the minefield of European identifications of ethnicity that have plagued attempts to pin down the homelands of Africans in the Americas. In contrast to many plantation records in the Americas, the ethnic or regional basis of many of the names is recognizable, and makes it possible to identify broad groupings of peoples and in some cases sub-groupings on which the slave trade probably drew. With fieldwork and considerable help from African-based scholars in several disciplines, we are able to draw inferences about the geography of the trade. The social and cultural implications of the names are not explored here. (3)
It should be stressed that the linking of names and ethnicity carries no necessary implications for the definition of ethnicity itself. Of all the elements that go to make up an ethnic identity, only name recognition and geographic location are essential to our analysis. We have a set of names which people who today identify themselves, for example, as Igbo, Efik or Banyangi can recognize. Whether those populations on which the transatlantic slave trade drew in the 1820s and 1830s saw their group identities in these terms is largely irrelevant to our quest to establish a profile of who was pulled into the slave trade. Nonetheless, because geography and language are fairly central to the way people see themselves, we would expect our findings to make a contribution to the broader debate on ethnicity and the extent to which it was transferred from the Old World to the New, and vice versa. It is at least clear that Europeans played a very small role in the identification of peoples that this source makes possibl e. (4) This situation is not very different from the peasants of Brittany and Normandy who went to Quebec in the seventeenth century and the German migrants who traveled from various parts of the Rhine to Pennsylvania in the eighteenth century. If a vessel carrying either of these groups had been stopped mid-Atlantic and the names and descriptions of everyone on board had been entered into registers without any indication of the nationality of the passengers or the political circumstances under which they traveled, would historians today have any hesitation in drawing inferences on the geographic origin and language of the passengers? As this European analogy implies, ethnolinguistic continuity in West Africa has been considerable over the centuries. (5) Notwithstanding such continuity, ethnic groups have certainly absorbed new members and varied in size over time, and awareness of changes in the range and meaning of ethnicity is a central concern in our analysis.
The potential benefits of the source are mostly easily demonstrated in smaller areas that were less prominent in the slave trade. The hinterland of the Cameroons River is one of these. The region was never a major player in the transatlantic slave trade and became a source of captives for a relatively short time. Generally, the traffic in slaves focused on a much narrower range of African regions before 1750 than after. For reasons not yet fully understood, the time a vessel spent acquiring a full consignment of captives increased strongly after the mid-eighteenth century, in response to which merchants greatly widened the African geographic range of their operations. The third quarter of the eighteenth century saw Sierra Leone and the Windward Coast rise to prominence for the first time as suppliers of captives to the Americas. Within the Bight of Biafra, the hitherto largely ignored Cameroons and the Gabon areas were pulled into the trade as part of this process and the Cameroons remained involved until the late 1830s. Before 1808, the merchants were mainly British; afterwards they were mainly Portuguese and Spanish. It now appears that from 1751 to 1840, about 62,000 captives left from the Cameroons River and Bimbia for the Americas, with possibly a few hundred more taken to the islands in the Bight. This was between 5 and 6 per cent of the total carried off from the Bight of Biafra in this period and, of course, a much smaller share again of the total traffic. (6) Departures averaged about 1,000 a year in the forty years after 1766, and again between 1821 and 1837. Outside these years, transatlantic trade was sporadic. Overall, about 14,000 captives are estimated to have left Cameroons ports between 1822 and 1837, so that the present sample constitutes about 7 per cent of the whole. (7) Given that captures occurred off-shore and that our sample of recaptives has a time profile similar to known total departures from the Cameroons, bias is not likely to be a major problem. Before 1808, the captives were sold ma inly in the British Caribbean; after 1807, they were taken in the ratio 40:35:25, to Rio de Janeiro, Cuba and Bahia respectively, though the British navy diverted many to Sierra Leone (as well as Fernando Po). The region under study was ethnically diverse. In its distribution of peoples around the Atlantic, it came close to replicating the remarkable diaspora from Africa as a whole in the years under study.
The region has always had a wide range of environments, peoples and languages -- perhaps the most diverse in Africa after the littoral of what is now south-east Nigeria to the immediate west, which it adjoins. Excluding vehicular languages and dialects, the Atlas linguistique lists 234 separate languages. (8) The south-western boundary is part of the Cross River region, while to the east its peoples have more in common with the Bantu peoples of West-Central Africa. The one third of the Cameroon population who live in the north have affinities with neither of these groups. The south-east of the country also contains Mbuti and related peoples. Ethnic affiliations are diverse. Twentieth-century colonial administrators did not inaugurate the idea of ethnicity in Africa, but neither did they recognize the complex reality of ethnic categories when setting up administrative units. For instance, 'Widekum', 'Keyaka' and 'Tikari' refer to administrative groupings as well as weakly related ethnolinguistic formations. (9) This complication in itself is less a problem than a lack of awareness of it would be. Whatever the nature of these groupings, our consultants were able to identify names from the Sierra Leone registers in common use today.
In 1913, some 75 to 90 years after the present data were collected, the estimated population of the German colony, roughly the present-day Cameroon Republic, was 2,650,000, yielding a density of about 7 per square kilometer. (10) According to German administrators three-quarters of the territory had a density of less than 5 per square kilometer, leaving two widely …