SPECIAL ISSUE: The Trans-Atlantic Slave Trade Database and African Economic History

CONTENTS

Preface

The Trans-Atlantic Slave Voyage Database and the 1–28 History of the Upper Guinea Coast Paul E. Lovejoy

The Slave Trade from the Windward Coast: The Case 29–52 of the Dutch, 1740–1805 Jelmer Vos

The Supply of Slaves from Luanda, 1768–1806: 53–76 Records of Anselmo da Fonseca Coutinho Daniel B. Domingues da Silva

Reexamining the Geography and Merchants of the 77–106 West Central African Slave Trade: Looking Behind the Numbers Filipa Ribeiro da Silva and Stacey Sommerdyk

The Shipping Registries of the Havana Slave Trade 107–136 Commission: Transcription Methodology and Statistical Analysis Henry B. Lovejoy

Extending the African Names Database: New Evidence 137–162 from Sierra Leone Suzanne Schwarz

Contributors 163 PREFACE

Under the leadership of David Eltis, and including Herbert Klein, Stephen Behrendt, David Richardson, Manolo Florentino, Paul Lachance, and many other scholars, the records of over 35,000 voyages have been assembled into a user friendly, open source, on-line database, The Trans-Atlantic Slave Trade Database (www.slavevoyages.org). The database, which is hosted at Emory University, builds on the pioneering work of Philip D. Curtin (The Atlantic Slave Trade: A Census, Madison, WI: University of Wisconsin Press, 1969). Subsequent demographic analysis of the forced migration of African peoples under slavery resulted in a more elaborate database, viz., David Eltis, David Richardson, Stephen Behrendt, and Herbert S. Klein, The Atlantic Slave Trade: A Database on CD-ROM (New York: Cambridge University Press, 1999). The current version greatly expands the database as a result of extensive collaboration among scholars whose research has been incorporated to generate the on-line database. The great debt that scholars owe to the editors and compilers of this database is enormous. Their work has challenged scholars to reconsider the impact of the slave trade on the Atlantic world. In May 2010, the Harriet Tubman Institute for Research on the Global Migrations of African Peoples, York University, provided a forum for discussion of the relevance of the database in understanding African economic history. The workshop specifically challenged some of the achievements of this monumental collaboration and offered new insights into how the impact of slavery on Africa can be assessed. It should be noted that there is no standardized way to refer to the on-line database. In the essays that follow, the database is sometimes referred to under the names of the principal compilers, David Eltis, David Richardson, Stephen Behrendt and Manolo Florentino; sometimes simply as Slave Voyages Database or the Voyages database, and sometimes among specialists as the Transatlantic Slave Trade Database, and to distinguish the on-line version from 1999 CD-ROM publication, as TSTD2. The Web site is not entirely clear on this matter, although generally Web sites are not cited with their author(s). In addition to the support of the Harriet Tubman Institute, the Workshop was funded by the Canada Research Chair in African Diaspora History and the Social Sciences and Humanities Research Council of Canada through its Major Collaborative Research Initiatives Programme.

Paul E. Lovejoy José C. Curto Editors

THE UPPER GUINEA COAST AND THE TRANS-ATLANTIC SLAVE TRADE DATABASE

Paul E. Lovejoy York University

e are now better able to understand the impact of the slave trade on the evolution of the Atlantic world to a degree that Wcould not have been imagined a generation ago, in a manner that might well fulfill the visions of Thomas Clarkson, W.E.B. Du Bois and other commentators.1 Thus there is a long tradition of assessing patterns of the slave trade. Clarkson was using muster lists to determine voyage information, and at the same time abolitionists were gathering testimony on pathways of enslavement in Sierra Leone and elsewhere. Du Bois searched for comparable information a century later, but all this evidence is patchy. The Trans-Atlantic Slave Trade Database represents the culmination of inputting data into a common format that was collated by a team of scholars under the leadership of David Eltis, David Richardson and Paul Lachance. The first version of the database accounted for approximately 27,233 voyages and was published as a CD- ROM by Cambridge University Press in 1999, thirty years after the pioneering census by Philip D. Curtin that launched the modern study of the slave trade in 1969.2 The current version accounts for more than 35,000 slaving voyages and has been available online (www.slavevoyages.org) since 2009.3 This version accounts for the overwhelming majority of voyages that carried enslaved Africans across the Atlantic. Its construction is an example of collaborative research at its best, drawing on the collective research of more than a dozen scholars over the past several decades. The attempt to identify all voyages carrying Africans on ships across the Atlantic world is essential in determining where people came from in Africa and where they went, although tracking the ships is only one part of these much bigger questions of the types of research approaches that are necessary. In the case of the Trans-Atlantic Slave Trade Database, unprecedented recognition has almost given this African Economic History v.38(2010):1–27

2 PAUL E. LOVEJOY database an identity of its own as “TSTD2,” that is the “transatlantic slave trade database 2,” although the title of the project significantly now emphasizes “voyages” and not “trade.” Nonetheless, TSTD2 has emerged as if it is a primary document(s) or a new persona of scholarship representing the collective input of many scholars, rather than a series of imputed variables. Inevitably, scholars will be examining the primary documentation behind TSTD2 and not rely on the imputed variables that underlay its construction. The Database has stimulated debate on a whole range of research issues, but in many ways the organisation of the data reflects the original priorities of the authors and the nature of the source material. A database is a tool that is used to organize data for purposes of analysis and is only as good as the imputations that underlay, which requires careful examination. This raises the issue of how to re-organise material, given that people with different interests will raise questions which are different to those originally raised by the compilers. In many ways, the existence of the Database has helped to frame new questions and identify new priorities. This critique assesses some of the claims, methods, contributions, and also lapses of the Slave Trade Database from the perspective of African economic and social history and therefore addresses one of the issues that the Slave Trade Database addresses – the origins of people in Africa. There are alternate ways of organizing data, and inevitably in the digital age alternate approaches to identifying, searching and structuring data are being explored, partly in response to the collaborative effort behind the TSTD2, which has become the reference. My critique arises from a workshop that was held at the Harriet Tubman Institute, York University in 2010, and whose essays are the basis of this special edition of African Economic History. This essay is intended to explore and critique the interface between the Slave Trade Database and African economic history, and by implication, African social history. What can and what cannot be said in relation to the trans-Atlantic migration under slavery is crucial to the reconstruction of African history and the history of the African diaspora in the Americas and elsewhere. Following on the intention of the Workshop, this essay raises issues that have to be THE UPPER EAST GUINEA COAST AND TDSD2 3 considered in the extrapolations that invariably will be attempted on the basis of the Slave Trade Database, which my own work on the history of slavery in Africa and the diaspora demonstrates. The numbers and the corresponding demographic implications in studying both Africa and the Americas are major historical issues. Most especially, I used much of the material that went into the first version of database, and subsequently used the CD-ROM version, in my overview history of slavery in Africa, Transformations in Slavery, originally published in 1983, revised in 2000 and again in 2011.4 In each version, I attempted to correlate the projected numbers of departures of slaves from different regions and places along the Atlantic coast and southeast Africa, and from these correlations projections were made to identify historical circumstances that could explain the scale of departures. The inadequacies of a European “regional” approach to the division of Africa for purposes of analysis became increasingly apparent in the revisions of the history of slavery in Africa. While the possibility of correlating slave departures with the contours of African history appear to be considerable, the way in which the database is constructed raises many problems. Elemental questions about how to interpret inputted data might well change any assessment of the materials that underpin the database, and it is difficult even for specialists to understand the implications of questioning the imputations. Any attempt to correlate migration with African history depends on some assessment of ethnicity, religion, and other factors in historical context. Transformations in Slavery presented a model of how historical change occurred in the context of slavery and should be treated as a first approximation only. The forced-migration data have to be connected with what is known about the history of specific places and peoples, and the attempts to correlate information have to be continually updated as demographic estimates are improved.5 There is now an extensive literature that attempts to understand this correlation, including publications in Portuguese and Spanish.6 Moreover, the considerable knowledge that has been achieved, some of which is reflected in the Slave Trade Database affords about the 4 PAUL E. LOVEJOY migration of Africans will surely increase even more because of the documentation still to come. The database has become an innovative institution in its own right. Having spawned an editorial review board to permit new data to be entered on a periodic basis, it promises to affect research and the dissemination of information on slavery and the slave trade for a long time. From the perspective of African history, however, certain questions and problems have to be addressed. TSTD2 is a database about slave traders and where they came from in Europe and the Americas and not about the enslaved, except as they figure as items of commerce, and hence TSTD2 does not take into consideration an African historical perspective on the slave trade. The advances in the reconstruction of African history, particularly arising from research at universities in Africa, suggest that it is no longer possible to rely on European categories of designation. And the ones in the Slave Trade Database are no exception.

The Problem of Choosing National Carriers to Organize the Database

The TSTD2 data are almost exclusively derived from the “national” carriers responsible for the movement of enslaved Africans throughout the Atlantic world. The legitimacy of the database depends upon the input of voyage records and the effort to assess the significance of the combined data. As the compilers clearly state, the database cannot demonstrate where Africans originally lived in Africa or where they eventually found themselves in the Americas. To determine the origins of enslaved Africans requires extrapolation and synthesis from the known details of African history, and the determination of their ultimate destinations depends on the ability to trace their subsequent movement after ships arrived in the Americas. Although we now know more about such topics as the intra-Caribbean enslaved migration than ever before, we have barely begun to understand the actual origins of the enslaved in Africa. Thus it can be seen that 42.4 percent of all slaves transported in the period 1700 to British abolition in 1807 went on British and North American ships, while Brazil and Portugal appear to have accounted for THE UPPER EAST GUINEA COAST AND TDSD2 5

34.6 percent and ships from France accounted for approximately 16 percent of all slave departures (see Table 1). The problems arising from the organization of TSTD2 data by national carrier relate to false assumptions about the majority of slaving voyages originated in Europe, which is clearly true in some periods but not in others. In fact references to Portugal could well mean Brazil and Angola, as well as Portugal, and in the nineteenth century really referred only Brazil, which is complicated by the fact the Portuguese monarchy was there for part of the period. Geography and politics become confused, and the impression acquired by students and non specialists is that “Portugal” means Lisbon, and not the whole kingdom let alone the empire. Rio de Janeiro and Bahia were major cities of empire but not of the kingdom. Similarly, “Spain” was not a factor in the trade until the nineteenth century, although in fact Spanish merchants were heavily involved in slaving ventures throughout the period of the trade, as revealed in archives across Spain. Moreover, the “Spain” of the nineteenth century was really Cuba, in terms of where ships operated from. In fact, Rio de Janeiro and Havana were more important as the home ports from where ships involved in the trade operated than Liverpool, Lisbon or Nantes. The trade was heavily concentrated in the Americas, not just Europe, especially by the nineteenth century. The assumption that the major ports of origin for slave ships were in Europe—London, Bristol, and Liverpool in Britain; Amsterdam and Middleburg in the Netherlands; Nantes and Rochelle in France; Lisbon in Portugal; and even Cadiz in Spain—is complicated by the fact that Rio de Janeiro, Salvador (Bahia) in Brazil, and Havana accounted for half of the slave traffic.7 Likewise, Newport and Charleston in North America were significant points of origin for slave voyages but not the only ones in the Americas. Adam Jones and Marion Johnson questioned the way in which French shipping data was being interpreted, arguing that French records of shipping in West Africa mostly targeted the Bight of Benin and to a lesser extent Senegambia.8 In 1997, Hernaes offered a critique of the

Table 1: Slave Departures by Nationality of Slave Ships, 1700–1807 North Spain Portugal/ Britain America Netherlands France Denmark/ Totals Brazil Baltic 1700– 16,000 2,497,000 2,845,000 288,000 334,000 1,152,000 85,000 7,217,000 1807 Percent 0.2 34.6 39.4 4.0 4.6 16.0 1.2

Source : Eltis et al., Transatlantic Slave Trade Database

Table 2: Region of Origin of Africans Destined for the Americas, 1500–1867

Bight of Bight of West South Sierra Windward Senegambia Gold Coast Benin Biafra Central East Total Leone Coast Africa Africa

755,500 388,700 336,900 1,209,300 1,999,100 1,594,600 5,694,600 542,600 12,521,300

6.0% 3.1% 2.9% 9.7% 16.0% 12.7% 45.5% 4.3%

Source: Eltis et al., Transatlantic Slave Trade Database.

THE UPPER EAST GUINEA COAST AND TDSD2 7

European regional approach that also questioned Curtin’s schema. He doubted the accuracy of projections based on the complicated assumptions behind the imputed variables that are often employed to “fill in” missing data, such as when the number of people on board a ship is not known but the size of the ship is known.9 Patrick Manning has also demonstrated the difficulties of following the regional designations of the Slave Trade Database in terms of how French shipping records were to be understood for the Bight of Benin. Manning has tried to assess the impact of trans-Atlantic slavery through demographic simulation, and his analysis is invariably tied to the Slave Trade Database and hence his projections are as robust as the basic assumptions of how data have been assigned to regional origins that are based on shifting European nomenclature.10 As David Wheat has demonstrated, there are similar problems with respect to missing data on the upper Guinea coast.11 Another problem in using Europe as the central parameter of analysis is they do not indicated that some slave traders who owned ships were based in Africa. And many crucial merchants in the trade were based in African ports of embarkation, although the links between these merchants and the ships is not a factor of analysis. The focus is on the owners and captains of ships from Europe and the Americas, not their partners on the coast of Africa. Moreover, people of Igbo, Kongo or Yoruba background became more numerous in the Americas than English, French, Spanish, or Portuguese, and hence neither the Africans involved in the trade nor the victims of the trade are part of the database. The on-line Slave Voyage Database does have a section on the names of those taken off slave ships, but this preliminary effort to understand the origins of individuals through the identification of names is seriously flawed, as is discussed in essays in this special issue.12 Although efforts are underway to improve this feature of the database, what has been available since 2009 lacks an adequate perspective that focuses on the people who were enslaved.

8 PAUL E. LOVEJOY

The Problem of the Regional Organization of TSTD2 Data for Embarkations

Which “regions” are relevant to the study of trans-Atlantic slavery, both in the Americas and in Africa? The Slave Trade Database provides a list of six regions in which the African coast is divided, including Senegambia and off-shore Atlantic (i.e., Cape Verde Islands); Sierra Leone; Windward Coast, Gold Coast, Bight of Benin, Bight of Biafra, West Central Africa and St. Helena; Southeast Africa and Indian Ocean Islands. The “regions” actually refer to places of embarkation or groups of ports along the coast and not geographical regions of the interior. Despite the difficulty of interpreting the historical significance of these regional categories, it is clear about half of all Africans entering the Atlantic world came from Bantu-speaking backgrounds and largely matrilineal social structures, sometimes influenced by the presence of Christianity. The importance of this pattern is evident in the study of Brazil, a pattern that Goulart and Curtin already recognized and confirming what Heywood and Thornton have maintained for some time.13 The database also confirms that more than 45 percent of Africans went to Brazil and that about 50 percent of all Africans came from the Bantu speaking regions of Angola and other parts of west central Africa, as well as Mozambique (Table 2). The South Atlantic route between Africa and Brazil emerges as singularly important avenue of migration, distinct from the North Atlantic routes above the Equator to the Caribbean and its hinterland on the Spanish mainland and North America. This finding alters our understanding of the cultural transfer from Africa to the Americas. The relationship between Brazil and west central Africa, especially Luanda, Benguela, Cabinda, Loango and their hinterlands, highlights the extent to which cultural factors mattered in the migration; the overwhelming majority of people from these areas brought a shared heritage of language and custom to the Americas, particularly noticeable in Brazil. The problem of determining the significance of cultural differences, similarity in language and such THE UPPER EAST GUINEA COAST AND TDSD2 9 commonly held cultural features as matrilineality is beyond the scope of this paper.14 The false sense of precision in defining the coastal regions from where people came from results in considerable confusion in understanding how to interpret TSTD2. What is called a region is actually the coast line, from one port to another, ignoring the African routes from the hinterland to the coast. While “Windward Coast” was used by ship captains to designate the area to the windward of the Gold Coast, but the term has little if any meaning in Africa. Similarly, mainland Hispanic America is combined into an amorphous area called “Central America” that no Latin American historian would recognize. The reference to questionable regional designations raises methodological issues. How can we be sure that these artificial, European-determined, concepts reflect what actually happened from an African perspective? Much of the way in which the database is constructed is based on basic assumptions of the regional designations of coastal origins that are not sustainable upon close study. Another issue to be considered is the ways that differing levels of knowledge of those who compiled the data might have affected how the data were organized. Unfortunately, the artificial “regions” that Curtin initially used in 1969 to contrast departure and arrival data predominate in the expanded database without addressing criticisms that have been in the literature since the 1980s. The designations “Senegambia,” “Sierra Leone,” “Windward Coast,” “Gold Coast,” etc. are fundamental to the structure of the database, but these “pseudo-Atlantic” terms are meaningful only from an “European” (and “American”) point of view. The problems with Curtin’s categories have been well known for thirty years. There was no “windward” coast from the perspective of the western Sudan; nor was “Sierra Leone” a clearly defined region and often only referred to a single river. Although the “Gold Coast” and the Bights of Benin and Biafra might comply with regional logic, there are overlapping connections that cannot easily be delimited. The designation “West Central Africa” conceals major differences between one side of the Congo River and the other. The presence of a kingdom of Kongo and 10 PAUL E. LOVEJOY a Portuguese colony of Angola at Luanda certainly were different than the trade at Loango and Cabinda, and distinctly different from the trade of all parts of West Africa. Muslim areas in the interior of West Africa and extending to Senegambia were relatively self-defined as well, further identifying regional designations from the perspective of African history, not European naval guides and western scholarly homogenization. “European” and “American” designations matter from an African perspective. It is important to understand how slave traders on the Atlantic delimited the African coast and for what purposes. The problem is that African regions that Europeans may have defined for purposes of the trans-Atlantic slave trade do not correspond with the designations of TSTD2. In determining coastal designations, consequently, the Slave Trade database has to be used with extreme care. The attempts to apply TSTD2 to historical contexts in Africa require projections that involve imputations on top of the imputations of the database. If it were possible to reorganize the Slave Trade Database to dispense with the European derived regional designations and instead use historical African designations of regions, it is likely that the structure of the database about the migration of Africans across the Atlantic would be completely altered. There have been various attempts to use the database for purposes of estimating the demographic impact of the migration in the Americas, and to a lesser extent in trying to figure out where people came from in Africa, but these attempts are limited by the structure of the database. St. Helena and the Cape Verde Islands (the offshore islands of Senegambia) figure prominently in all tables in the Slave Trade database for reasons that are curious. In both cases, students and non-specialists will see west central African and St. Helena as part of the same category, but this is a TSTD2 design mistake. St. Helena was only involved in the trade as a receiving point for slaves taken off ships by the British navy only in the mid nineteenth century. Slaves came from southeast Africa as well as west central Africa, and hence the St. Helena data reflects trade in the South Atlantic but only between c. 1840 and 1860.15 Similarly, Cape Verde was closely connected to the upper Guinea coast for the THE UPPER EAST GUINEA COAST AND TDSD2 11

Portuguese and Brazilian trades, and was really a reflection of trade mostly with Bissau and Cacheo and other areas on the upper Guinea coast, which otherwise in the database is identified as “Senegambia,” even though located to the south the Gambia River, and really more connected with the region that is designated as “Sierra Leone,” which included the rivers and islands immediately south of Bissau and Cacheo as far as Cape Mount. Yet, the offshore islands of Cape Verde are prominent as part of “Senegambia,” although the islands were tied to the rivers south of Senegambia. Uruguay, St. Helena, and other obscure “slaving” areas also enter prominently, and misleadingly, into the online database, but the fact that some ship owners were resident in Africa virtually slips by unnoticed. Francisco Felix de Souza, the leading slave trader in Ouidah in the early nineteenth century, invested in many slave ships during the 1820s, and is mentioned thirteen times, although identified with his place of residence in West Africa.16 The online TSTD2 database does not attempt to explore the subject of African-based merchants who were slave shippers, despite what is known about Lagos and Ouidah in the nineteenth century, or indeed about Luanda even earlier. Given that the great strength of the volume and the database is supposed to be its focus on voyage data, this is a curious oversight or omission. Could merchants based in Africa own ships? Could they be partners in Atlantic contracts? In the interpretation of the Slave Trade Database, what is often referred to as place of origin generally refers to regions of departure from the African coast, and often the search ends there, despite the continuous expansion of the frontiers of knowledge about the African past. This can be explained by the fact that it is a slave trading voyage database and not a slave migration database, even though most users read it like this. For more than thirty years, there has been criticism about the regional organization of the data, but the compliers of the database have preferred to retain the inaccuracies of a Eurocentric approach, which can lead to naïve and ahistorical assumptions about the African background. From the perspective of specialists of African history, these regional designations are not rooted in African historical reality, with the result 12 PAUL E. LOVEJOY that there is limited explanatory power in the use of these designations. Somewhere along the way, the database project became more concerned with the ships and the language of ship captains rather than with focusing on the people and what we need to know to decipher historical context. TSTD2 is based on imputed variables and otherwise requires sophisticated interaction and knowledge of SPSS programming to decipher its codes. Nonetheless, despite the complex layers of storage and imputation, to the credit of the compilers, the database has taken on an identity. I am suggesting here, however, that the reliance on shipping data has shaped the organization of data in a way that raises problems for those interested in the reconstruction of African history and the influence of Africans in the Americas and Europe, as well as in Africa. In considering the links to Africa, the Slave Trade Database is implicitly Eurocentric and therefore raises serious problems with respect to how the database can be applied in the reconstruction of African history. Because the data are derived from ship registers, commercial documentation, and slave traders, the sources are dangerously biased. The problem with the construction of the database is that this factor has not always been recognized. As a result, the Slave Trade Database has an implicit weakness that prompts discussion of some of the uses to which it can be put in the reconstruction of trans-Atlantic migration. Inevitably, despite the overwhelming amount of shipping data that is available, it is often difficult to trace where individuals who were enslaved actually came from and how they had become enslaved, but the database makes it even more difficult to trace origins. The distillation of information into a form that can be plugged into a database complicates the effort to get from ship to shore, and from shore inland, and thereby understand the historical context of enslavement. The European perspective is in the construction of the database and also in the documentation behind the database, and compilers have not made it easy to discern differences between them or attempt to confront this bias with an “African” perspective that was changing over time.

THE UPPER EAST GUINEA COAST AND TDSD2 13

Why Look at a Database About the Enslavement of Africans from an “African” Perspective?

The simple answer is, “because Africans were the ones who were enslaved.” The organization of data from the perspective of the slave ship has certain advantages since the nature of commercial records has led to the preservation of considerable amounts of documentation. The maritime destinations of passengers have considerable value, but the ports of disembarkation were not where most people ended up. We also have to understand the data in light of where the enslaved originally lived, not only their port of departure, defined by regions that have little historical meaning. This can only be done through careful analysis of primary data in the context of known African history. We have to assess available documentation with reference to local African conditions with a degree of specificity that is not always discernible from an uninformed reading of documentation derived from the slave ship. The danger is that historical clues are reduced to categories that obfuscate rather than clarify. The reconstruction of how the development of the African diaspora had an impact on slavery and society in Africa are intimately related to the path-breaking contribution of Curtin’s Census and the subsequent efforts of Eltis, Richardson and many others in generating a mechanism of historical analysis through the construction of databases that can be revealing of trends and can raise new questions of enquiry. Nonetheless, certain problems arise in the use of the Slave Trade Database that have to be examined or there is danger that trends and patterns will be misunderstood, or that some things that might seem clear and decisive are in fact more apparent than real. The Slave Trade Database needs to be placed in context of the historiography of African history and the slave trade. Unfortunately, the TSTD2 contains only a superficial introduction that provides no context for the database and its construction. Anyone wanting to know the history of the project and what the aims of the project might be has to go elsewhere. Clearly, the online database would benefit from a fuller explanation of categories and some of the decision-making processes 14 PAUL E. LOVEJOY underlying the identification of categories. The presentation of the background to the project ignores a long trajectory of efforts to quantify the slave trade, dating at least to the late eighteenth century. Thomas Clarkson, and perhaps Bryan Edwards of Jamaica, made the first serious quantitative attempts to analyze the scale of the traffic in slavery. Some appreciation of the magnitude of the trade underlay the anti–slave trade movement of the 1780s and 1790s, which culminated in British abolition (1807), although Clarkson could only estimate a portion of trade before British abolition.17 There were various subsequent efforts to address the scale of the trade in the 1850s which had little basis in statistical data. As Philip D. Curtin discovered, a little known publicist for the Mexican government, the American, Edward E. Dunbar, published an overview of the slave trade in The Mexican Papers in 1861, which nonetheless was only based on estimates of the migration.18 In 1896, W.E.B. Du Bois completed his doctoral dissertation at Harvard on the slave trade, in which he attempted to quantify the migration, although as Curtin has argued, Du Bois probably relied on Dunbar for his own estimates.19 Other efforts to document the migration included the work of Mauricio Goulart, who in 1950 attempted to summarize the research that had been done on the number of Africans arriving in Brazil.20 Curtin provided an overview of this historiography, and his preliminary census of 1969 launched what has become known as the “numbers game,” which has raised scholarship on slavery to a new plateau. In comparing demographic data on the estimated size of slave populations in the Americas with available shipping data for departures in different places in Africa and arrivals at specific places in the Americas, Curtin tried to determine the relative scale and direction of the migration and how the migration changed over time. The present Slave Trade Database is the latest version of this scholarship. Curtin identified the key problem in any demographic analysis of the enforced migration of Africans to the Americas—the apparent discrepancy between the known demographic profile of the enslaved population in the Americas and data derived from the shipping records for the numbers of Africans leaving Africa and arriving in the Americas. THE UPPER EAST GUINEA COAST AND TDSD2 15

Curtin’s analysis revealed that records from some parts of Africa yielded a reasonable estimation of departures in some periods, and that data on arrivals in the Americas could be used to fill in some gaps in data. He divided Africa into several “regions” to aggregate data that sometimes depended upon fuzzy assumptions because of gaps in data, and then he compared his results with what was known about the size of the enslaved population in the various colonies of the Americas. The accumulation of extensive new data has reduced the discrepancy in what Eltis and Richardson have labelled the “Curtin dilemma”—the aforementioned problem of reconciling African shipping records with demographic data from the Americas. There is now much more shipping data available, so that voyages are now the focus of analysing demographic movements. Hence the discourse has changed and gives the appearance of greater accuracy in determining the origins of people. However, any attempt to assess population flows relies on assumptions that are not directly observable in the database. How data have been entered into the database requires explanations that are not easily discernible. The designations in the sources are sometimes very specific and other times are often confusing. For example, a ship leaving the coast of the Galinhas, often considered part of the “Windward” coast but which was clearly tied to a specific commercial network on the upper Guinea coast in what is now southern Sierra Leone. Similarly, the term Mina, which derived from the Portuguese trading fort at Elmina on the Gold Coast, had different meanings for Portuguese, Dutch, and French traders and frequently included slaves leaving from Ouidah and other ports in the Bight of Benin. These problems are not adequately addressed in the Slave Trade Database or in the publications that so far derive from the database.21 This weakness in regional descriptions raises questions as to the extent to which the database can be used to assess the social, economic, political and cultural impact of the slave trade on Africa. In its categorization, TSTD2 does not allow for change over time. The use of Guinea for the Atlantic coast of Africa is acceptable; its use dates to the fifteenth century, but the analysis has to relate to specific places on the African coast that were important rather than larger coastal categories 16 PAUL E. LOVEJOY that have little correspondence to African history. The different meanings of the Mina coast demonstrate the risks of relying on ahistorical categories.

The Case of the Upper Guinea Coast

The earliest European designations of the Atlantic coast referred to the upper Guinea coast and, correspondingly, to the lower Guinea coast, “Guinea” being a term applied to the Atlantic coast of Africa at least as far south as the Bight of Biafra. The TSTD2 does not use the term Guinea, but rather identifies Senegambia as a region that incorporates modern Senegal, Gambia and Guinea-Bissau. In fact the region of Senegambia was historically part of the upper Guinea coast, which the Slave Trade Database divides between Senegambia, Sierra Leone and parts of an uncertain region that is referred to as the Windward Coast, which seems to lie between Cape Mount and the Gold Coast. The regions in fact overlap, and the distinctions are questionable. Far more useful for analysis is the identification of specific places on the upper Guinea coast, especially the Senegal River, the Gambia River, and the various rivers and islands to the south, including, Bissau, Cacheo, Rio Pongo, Isles des Los, the Sierra Leone River, Sherbro Island, the Galinhas, and similar points of departure. As the terms have been used by historians of West Africa, the interior includes the region of the western Sudan and the basins of the Senegal, Gambia and upper Niger. The coastal zone is in fact the frontier of the western Sudan.22 The calculation of the number of Africans departing from the database regions of Senegambia, Sierra Leone and the Windward Coast has increased the figure from 689,000 in the 1999 database to 854,800 in the on-line version of 2009, an increase of 24 percent (See Appendix). Yet the relative proportions of the designated sub-regions did not increase in the same proportion; the number of departures registered for Senegambia came mostly from St. Louis, the Gambia and Bissau. The estimated departures increased 15.3 percent from 314,900 to 363,200, while the number of those from “Sierra Leone,” i.e., the region of modern Guinée and Sierra Leone, actually decreased from 230,900 to THE UPPER EAST GUINEA COAST AND TDSD2 17

202,000, a drop of 12.5 percent. Astonishingly, the region that has been most criticized as a concept, the so-called Windward Coast registered a remarkable 102.2 percent increase, rising from 143,200 in the 1999 calculations to 289,000 in 2009. One asks, is this an increase in the uncertainty of the categories of ascription, since the Windward Coast is a dubious concept? Why in the one sub-region where there is most evidence of slave trading is there a decline in perceivable departures? More recently, Eltis and Richardson have calculated that 69.6 percent of Africans left upper Guinea coast after 1740, and 20.9 percent came from Iles de Los, Sierra Leone estuary and the Galinhas, largely in the nineteenth century (Table 3).23

Table 3: Departures from Upper Guinea Coast

Departures Percent Upper Guinea 1501–1641 185,000 12.5 Upper Guinea 1642–1740 266,000 18.0 Upper Guinea 1741–1807 804,000 54.3 Upper Guinea 1808–1856 226,000 15.3 TOTAL 1501–1856 1,481,000

St. Louis 1668–1829 145,000 9.8 Gambia 1644–1816 258,000 17.4 Bissau 1686–1843 126,000 8.5 Iles de Los 1759–1820 64,000 4.3 Sierra Leone River 1563–1808 148,000 10.0 Galinhas 1731–1856 98,000 6.6 SUB-TOTAL 839,000 56.7 Source: Eltis and Richardson, Atlas of the Transatlantic Slave Trade, 96-108.

The problem arises from the TSTD2 confusion of coastal categories with regions that make sense in terms of African history. The dominant regional classification in western Africa during the era of the trans- Atlantic slave trade was the western Sudan, whose relationship to the 18 PAUL E. LOVEJOY voyage database is missing. Why the architecture of the Slave Trade Database could not be developed to conform to the outlines of this scholarship is not clear. Historians of West Africa continue to trace the continuities and disjuncture for the vast region that stretched southward from the Sahara to the Atlantic shores from the Senegal River to the Bights of Benin and Biafra, both of which are located in the Gulf of Guinea. The Slave Trade Database, for purposes that have more to do with European shipping patterns than to African history, reduces this region basically into six zones, Senegambia and Cape Verde Islands, Sierra Leone, the Windward Coast, the Gold Coast, and the Bights of Benin and Biafra. The boundaries between these zones may give the impression of being precise. They have the dubious advantage of corresponding to the modern political boundaries of Senegal/Guinée- Bissau, and from there to the boundary between Sierra Leone and Liberia. Liberia and Cote d’Ivoire conform exactly to the Windward Coast, which conveniently extends to the border with Ghana, the “Gold Coast.” Togo, Bénin and southwestern Nigeria form the Bight of Benin, and finally the Bight of Biafra essentially incorporates the ports in the central and eastern Niger delta, as well as the Cross River. This was a schema imposed by Philip Curtin, whose pioneering census relied on far less documentation than has been assembled in the Slave Trade Database. Without adequate explanation, the Slave Trade Database retains the preliminary delineations of the coast and ignores the criticisms of scholars of this regionalization profile. A comparison of figures from the earlier version with the latest version also presents problems that should be discussed. Overall figures for the number of Africans leaving what I would call the upper Guinea coast, vary considerably between the 1999 version of the Slave Trade Database and the current on-line 2009 version. Without question, the overwhelming number of African departures left from the region stretching from the River Gambia to Sherbro Island, at least in the eighteenth century. The region “Senegambia” is euphemism for the stretch of coast from the Senegal River to the Gambia River, but in terms of departures, the overwhelming majority of captives leaving from the THE UPPER EAST GUINEA COAST AND TDSD2 19

“region” left from the Gambia River or immediately to the south, and indeed connecting with the many rivers as far south as Sherbro Island. This is the region of upper Guinea. The logical geographic organization would include this region as one, with the Senegal a distant and northern tributary, as was the region to the leeward of Sherbro Island. In this configuration, the revisions in the database raise more questions than provide answers.24 It is well known that “Portuguese” slavers who were very active on the upper Guinea coast in the eighteenth century, especially at Cacheo and Bissau, came largely from Brazil. Research on the “creole” culture of the islands, estuaries and lagoons of this region is now extensive. The names of the commercial households that are known attest to a close connection with the slave trade. Moreover, the routes to the interior were extensive, following the Senegal River in the far north, and then the heavily travelled Gambia, before following the many rivers that extended from Casamance to Sherbro, including the Sierra Leone River and Bunce Island, which have the ignominy of giving their name to a regional category in the database, a designation that has no historical context. How is it that Brazilian traders only account for less than 10 percent of slave departures from this broad region, which was less than the Dutch as being recorded to have transported from this broad region in the same period. The Dutch maintained no shore facilities on the coast, while the Brazilians did establish partnerships in both Cacheo and Bissau. Where is the Cape Verde connection, which lends its name to a category of the database? Britain is credited with transporting the overwhelming majority of Africans from the whole of the upper Guinea coast, 503,400 of a total of 854,800, or almost 60 percent. North American ships accounted for 65,800 people, or another 10 percent, which suggests that the regional categories disguise changes over time. Most obviously, why are there such significant variations between the two editions of the Slave Trade Database, the CD-ROM of 1999 and the online version of 2009? The new findings do not seem to correspond to events in Africa, or at least raise questions of earlier interpretations. What does this mean in terms of African history? The current version 20 PAUL E. LOVEJOY compounds the problem of correlation with African history. This problem alone demonstrates that the model of trans-Atlantic history proposed by Eltis and Richardson has to be challenged because it is based on questionable data imputations. The Gambia trade had links to the western Sudan that were comparable to the trade inland from most parts of the region that is designated as Sierra Leone. Perhaps it is possible that the Gold Coast can be divided from the Bight of Benin, at least with the justification of being separated by the Volta River, but much of what is identified with the “Windward Coast” was contiguous to the Gold Coast—there was no geographical feature in the west that demarcated these regions, not even for ship captains. Moreover, Portuguese sources often referred to the coast east of Elmina as the Mina coast, which was an equally “foreign” nomenclature, just as Onim was the Portuguese name for Lagos, while Yoruba and Hausa called the town Eko. For some reason not discussed, the Slave Trade Database tries to reorganize the geography, thereby missing crucial historical and environmental contexts. The geography distinguishes the Gold Coast from the area the east of the Volta River because of the lagoons behind the Bight of Benin, which run parallel to the sea and extend all the way to the Niger River delta as far as the Cross River estuary in the Bight of Biafra. For our purposes here, we need not venture further south to critique the category, West Central Africa. The point is that the regional designations make sense in some cases but not at all for the upper Guinea coast as used as categories in the Slave Trade Database. And by extension, neither do the other regional designations make sense, but for different reasons. The context for this critique is the debate over the impact of the slave trade on African societies and economies, represented in its extremes by David Eltis, Philip Morgan and David Richardson, who minimize the impact and dismiss the evidential discourse of Africanist specialists, on the one hand, and Joseph Inikori, whose attempt at relabeling historic events, processes, and terminology has led to an exaggeration of the impact of slavery on Europe, on the other hand.25 According to the transformation thesis that I have proposed for THE UPPER EAST GUINEA COAST AND TDSD2 21 understanding the African past attempts to understand the process of change, slavery was an integral part of African history, so the institution of slavery and its localized implementation changed over time. The incidence of slavery seems to have increased dramatically over time, especially from the sixteenth century, and its demise only occurred under colonial transformations, and then has not entirely disappeared. In conclusion, the TSTD2 as it is presented allows virtually no critique of the problems that have been raised here, despite the establishment of a scholarly review board. The “Old World” for Eltis and Richardson, and apparently conceptually implicit in the construction of the database, really means where ships came from in Europe, and by extension Brazil, Cuba and North America. The focus is not on where people came from. There is disjuncture between the generators of the database and historians who are attempting to analyze the database in the context of specific research projects and questions that relate to African provenance. Despite these weaknesses, the data that are beneath the surface of the population “estimates” that are most easily accessible in the online database offer the promise of a truly innovative analysis of African ports and places of origin that does not take the Eurocentric tradition at face value. The raw data should allow researchers to compensate for any unrecognized or accidental biases of the compilers, particularly when caused by the need to stay abreast of the ever increasing knowledge about African history. 22 PAUL E. LOVEJOY

Appendix: Comparison of 1999 and 2009 versions of Slave Trade Database

Table 1: Regional Origins of Slaves in the Eighteenth-Century Atlantic Trade: 1999 Senegambia Sierra Leone Windward Coast Total 1701–10 7,600 700 2,100 10,400 1711–20 17,100 3,800 300 21,200 1721–30 26,900 2,200 3,500 32,600 1731–40 41,500 3,800 4,600 49,900 1741–50 16,700 9,500 8,000 34,200 1751–60 47,600 28,400 30,900 106,900 1761–70 51,200 63,600 49,300 164,100 1771–80 50,600 20,400 30,000 101,000 1781–90 30,100 57,500 7,300 94,900 1791–1800 25,600 41,000 7,200 73,800 Total 314,900 230,900 143,200 689,000 Source: Eltis et al., Slave Trade Database (1999)

THE UPPER EAST GUINEA COAST AND TDSD2 23

Table 2: Regional Origins of Slaves in the Eighteenth-Century Atlantic Trade: 2009 Senegambia Sierra Leone Windward Coast Total

1701–10 16,300 1,200 3,100 20,600

1711–20 22,700 3,100 4,400 23,800

1721–30 34,900 9,400 4,500 48,800

1731–40 44,800 1,500 9,400 55,700 1741–50 24,200 8,000 25,200 57,400 1751–60 50,600 17,400 44,100 112,100 1761–70 52,400 42,300 76,500 171,200 1771–80 51,300 36,600 65,200 153,100 1781–90 37,900 31,400 36,100 105,400 1791–1800 28,000 51,100 21,200 100,300 Total 363,200 202,000 289,600 854,800 Source : Eltis et al., Slave Trade Database (2009)

Summary of Discrepancies Between Slave Trade Database 1999 and 2009 Senegambia increase: 48,300 + 15.3 percent Sierra Leone decrease: 28,900 – 12.5 percent Windward Coast increase: 146,400 + 102.2 percent Total increase: 165,800 + 24.0 percent

24 PAUL E. LOVEJOY

1 The research for this paper was funded by the Social Sciences and Humanities Research Council of Canada and the Canada Research Chair in African Diaspora History. I wish to thank Mariza Soares, José C. Curto and the anonymous readers for their comments, although I accept full responsibility for the interpretation presented here. 2 Philip D. Curtin, The Atlantic Slave Trade: A Census (Madison, WI: University of Wisconsin Press, 1969); and David Eltis, David Richardson, Stephen Behrendt, and Herbert S. Klein, The Atlantic Slave Trade: A Database on CD-ROM (New York: Cambridge University Press, 1999). 3 David Eltis and David Richardson, eds., Extending the Frontiers: Essays on the New Transatlantic Slave Trade Database (New Haven, CN: Yale University Press, 2008) and Eltis and Richardson, Atlas of the Transatlantic Slave Trade (New Haven, CN: Yale University Press, 2010). To the best of my knowledge, there are no reviews of the online database, although there are reviews of publications arising from the database; see Lovejoy, “Extending the Frontiers of Transatlantic Slavery, Partially,” Journal of Interdisciplinary Studies 11:1 (2009), 57-70, which is a review of Eltis and Richardson, Extending the Frontiers; and Joseph C. Miller’s review of Atlas of the Transatlantic Slave Trade, in Slavery and Abolition 32:4 (2011), 589-92, neither of which is a review of the dababase itself. 4 Paul E. Lovejoy, Transformations in Slavery: A History of Slavery in Africa (Cambridge: Cambridge University Press, 3rd ed., 2011). 5 For a discussion of the methodological issues in analyzing ethnicity, ethnic transfer and transformations, see, for example, Paul E. Lovejoy, “Methodology through the Ethnic Lens: The Study of Atlantic Africa,” in Toyin Falola and Christian Jennings (eds.), African Historical Research: Sources and Methods (Rochester, 2002). Also see Robin Law and Kristin Mann, “West Africa in the Atlantic Community: The Case of the Slave Coast,” William and Mary Quarterly 56 (1999), 307-34; Lovejoy, Identity in the Shadow of Slavery (London: Continuum, 2000); Kristin Mann and Edna G. Bay, eds., Rethinking the African Diaspora: The Making of a Black Atlantic World in the Bight of Benin and Brazil (London: Frank Cass, 2001); Lovejoy and David V. Trotman (eds.), Trans-Atlantic Dimensions of Ethnicity in the African Diaspora (London: Continuum, 2003); José C. Curto and Lovejoy (eds.), Enslaving Connections: Changing Cultures of Africa and Brazil during the Era of Slavery (Amherst, N.Y.: Humanities Press, 2004); Lovejoy, ed., Slavery on the Frontiers of Islam (Princeton, NJ: Markus Wiener, 2004); Matt Childs and Falola (eds.), The Yoruba Diaspora in the Atlantic World (Bloomington, IN: University of Indiana Press, 2005); Curto and Renée Soulodre-La France (eds.), Africa and the Americas: Interconnections during the Slave Trade (Trenton, NJ: Africa World Press, 2005); Gwendolyn Midlo Hall, Slavery and African Ethnicities in the Americas: Restoring the Links (Chapel Hill, NC: University of North Carolina THE UPPER EAST GUINEA COAST AND TDSD2 25

Press, 2005). The publications since 2005 are extensive, and their influence is not adequately reflected in modification and elaboration of the database. 6 Mariza de Carvalho Soares, Devotos da cor: Identidade étnica, religiosidade e escravidão no Rio de Janeiro, século XVIII (Rio de Janeiro, 2000), now available in English, People of Faith: Slavery and African Catholics in Eighteenth-Century Rio de Janeiro (Durham, NC: Duke University Press, 2011); Rina Cáceres (ed.), Rutas de la Esclavitud en África y América Latina (San José: Editorial de la Universidad de Costa Rica, 2001); Claudia Mosquera, Mauricio Pardo, and Odile Hoffmann (eds.), Afrodescendientes en las América: Trayectorias socials e identitarias (Bogota: Universidad de Colombia, 2002); Juliana Barreto Farias et al., No labirinto das nações: africanos e identidades no Rio de Janeiro, do século XIX (Rio de Janeiro: Arquivo Nacional, 2005); Manolo Florentino, Alexandre Vieira Ribeiro adn Daniel Domingues da Silva, “Aspectos Comparativos do Tráfico de Africanos Para o Brasil (Séculos XVIII e XIX),” Afro-Ásia 31 (2004), 83-126; and João José Reis, Rebelião escrava no Brasil (2a edição revista e ampliada. São Paulo: Companhia das Letras, 2003). 7 Luís Henrique Dias Tavares, Comércio proibido de escravos (São Paulo: Ática, 1988); Tavares, Desembarque da pontinha (Salvador: CEB, 1971); Jaime Rodrigues, O infame comércio. Propostas e experiências no final do tráfico de africanos para o Brasil (1808-1850) (Campinas, SP: Unicamp/Cecult, 2000). 8 Adam Jones and Marion Johnson, “Slaves from the Windward Coast,” Journal of African History 21:1 (1980), 17-34. 9 Per O. Hernaes, Slaves, Danes, and African Coast and Society: The Danish Slave Trade from West Africa and Afro-Danish Relations in the Eighteenth-Century Gold Coast (Trondheim: University of Trondheim Press, 1997). See also Lovejoy, “The Impact of the Atlantic Slave Trade on Africa: A Review of the Literature,” Journal of African History 30 (1989), 365-94. 10 Patrick Manning, Slavery and African Life: Occidental, Oriental, and African Slave Trades (New York: Cambridge University Press, 1990). 11 For the most critical evaluation of estimates of departures from the upper Guinea coast, see David Wheat, “The First Great Waves: African Provenance Zones for the Transatlantic Slave Trade to Cartagena de Indias, 1570-1640, Journal of African History 52 (2011), 1-22, whose assessment largely replaces the earlier work of Enriqueta Vila Vilar, Hispanoamerica y el comerico de esclavos: Los Asientos Portugueses (Seville, 1977). Also see Walter Hawthorne, From Africa to Brazil. Culture, Identity, and an Atlantic Slave Trade, 1600-1830 (Cambridge: Cambridge University Press, 2010). 12 The names of Africans taken off slaving vessels are in a separate database; see G. Ugo Nwokeji and Eltis, “The Roots of the African Diaspora: Methodological Considerations in the Analysis of Names in the Liberated African Registers of 26 PAUL E. LOVEJOY

Sierra Leone and Havana,” History in Africa 29, (2002), 365-79. Also see The African Origins Project, www.african-origins.org/about. 13 Linda M. Heywood (ed.), Central Africans and Cultural Transformations in the American Diaspora (New York: Cambridge University Press, 2002); Heywood and John K. Thornton, Central Africans, Atlantic Creoles, and the Foundation of the Americas, 1585–1600 (New York: Cambridge University Press, 2007); Thornton, Africa and Africans in the Making of the Atlantic World, 1400–1800 (New York: Cambridge University Press, 1998; orig. pub. 1992); Livio Sansone, Eliseé Soumonni and Boubacar Barry, eds., Africa, Brazil and the Construction of Trans Atlantic Black Identities (Trenton, NJ: Africa World Press, 2008); and José C. Curto and Paul E. Lovejoy, eds., Enslaving Connections: Changing Cultures of Africa and Brazil during the Era of Slavery (New York: Humanity Books, 2003). Also see Mauricio Goulart, Escravidão Africana no Brasil (São Paulo, 1950); and Curtin, Atlantic Slave Trade, 47-49. 14 The meaning of ethnicity in this region differed from other parts of Africa, as reflected in cultural practices such as scarification and tattooing; see Paul E. Lovejoy, “Scarification and the Loss of History in the African Diaspora,” in Andrew Apter and Lauren Derry, eds., Activating the Past Historical Memory in the Black Atlantic (Newcastle: Cambridge Scholarly Publishing, 2009). 15 Andrew Pearson, Ben Jeffs, Annsofie Witkin and Helen MacQuarrie, Infernal Traffic: Excavation of a Liberated African Graveyard in Rupert’s Valley, St Helena (York: Council for British Archaeology, 2011). 16 Robin Law, “Francisco Felix de Souza in West Africa, 1820-1849,” in José C. Curto and Paul E. Lovejoy, eds., Enslaving Connections: Changing Cultures of Africa and Brazil During the Era of Slavery (New York: Humanity Books, 2004); Law, Ouidah: The Social History of a West African Slaving ‘Port’ 1727- 1892 (Athens, OH: Ohio University Press, 2004), 165-71; and Alberto da Costa e Silva, Francisco Felix de Souza, Mercador de Escravos (Rio de Janeiro: Eduerj/Editora Nova Fronteira, 2004). For da Souza’s ships, see Jamie Bruce Lockhart and Paul E. Lovejoy, eds., Hugh Clapperton into the Interior of Africa: Records of the Second Expedition 1825-1827 (Leiden: Brill, 2005), 35, 90. 17 Thomas Clarkson, The Substance of Evidence of Sundry Persons on the Slave- Trade Collected in the Course of a Tour made in the Autumn of the Year 1788 (London: James Phillips, 1779); and Clarkson, History of the Rise, Progress, and Accomplishment of the Abolition of the African Slave Trade by the British Parliament (London: John W. Parker, 1839). In addition to collecting extensive data on the British slave trade, Clarkson also tried to synthesize the data. See, for example, Thomas Clarkson Manuscript CN 33, c. 1823, Huntington Library. 18 Curtin, Atlantic Slave Trade, 6-8, citing Edward E. Dunbar, “Commercial Slavery,” in The Mexican Papers 1:5 (April, 1861), 269-70. 19 W.E.B. DuBois, The Suppression of the African Slave-Trade to teh United States of America, 1638-1870 (Ph.D. thesis, Harvard University, 1896); also see THE UPPER EAST GUINEA COAST AND TDSD2 27

Curtin, Atlantic Slave Trade, 6, citing DuBois, “The Negro Race in the United States of America,” in Papers on Inter-Racial Problems Communicated to the First Universal Races Congress (London, 1911), 349. 20 Goulart, Escravidão Africana no Brasil. Also see Curtin, Atlantic Slave Trade, 47-49. 21 See Eltis and Richardson, Atlas of the Transatlantic Slave Trade, and Eltis and Richardson, Extending the Frontiers. 22 Virginia Rau, Walter Rodney, P.E.H. Hair, Yves Person, and other historians clearly established the historical context for this broad region. See especially Virginia Rau, Uma tentative de colonização da Serra Leoa no Século XVII (Madrid: C. Bermejo, 1946); P.E.H. Hair, “The Spelling and Connotation of the Toponym ‘Sierra Leone’ since 1461,” Sierra Leone Studies 18 (1966), 43-58; Walter Rodney, “Slavery and Other Forms of Social Oppression on the Upper Guinea Coast in the Context of the Atlantic Slave Trade,” Journal of African History 7:4 (1966), 431-43; Rodney, A History of the Upper Guinea Coast 1545 to 1800 (Oxford: Clarendon, 1970), P.E.H. Hair, “Ethnolinguistic Continuity on the Upper Guinea Coast,” Journal of African History 8 (1967), 247-68; and Yves Person, Samori: une revolution Dyula (Dakar: IFAN, 1968-1975), 3 vols. Also see P.E.H. Hair, Africa Encountered: European Contacts and Evidence 1450- 1700 (Aldershot, 1997). More recently, see Toby Green, The Rise of the Trans- Atlantic Trade in Western Africa, 1300-1589 (Cambridge: Cambridge University Press, 2012). 23 Eltis and Richardson, Atlas of the Transatlantic Slave Trade, 96-108. 24 For recent studies of the slave trade along the upper Guinea coast, see Suzanne Schwarz and Paul E. Lovejoy, eds., Slavery, Abolition and the Transition to Colonialism in Sierra Leone (Trenton, NJ: Africa World Press, 2012). 25 David Eltis, Philip Morgan and David Richardson, “Agency and Diaspora in Atlantic History: Reassessing the African Contribution to Rice Cultivation in the Americas,” American Historical Review 112 (2007), 1329-58; and the corresponding exchange with Walter Hawthorne and Gwendolyn Midlo Hall. For Inikori’s perspective, as stated at the conference, “Ending the International Slave Trade: A Bicentenary Inquiry,” College of Charleston, March 25, 2008.

THE SLAVE TRADE FROM THE WINDWARD COAST: 1 THE CASE OF THE DUTCH, 1740–1805

Jelmer Vos Old Dominion University

n the rapidly advancing scholarship on African coastal societies in the Atlantic slave trade, the Windward Coast remains a much Ineglected area.2 The only study fully devoted to the slave trade from this region is an article published in 1980 by Adam Jones and Marion Johnson, which critiqued the census data produced by Philip Curtin.3 Jones’s and Johnson’s commentary ended with a warning that, for lack of historical knowledge of the trade from this part of West Africa, it was still far too early for scholars to produce a meaningful census. “What is needed now is not a slanging match over the totals which Curtin proposed, but detailed analysis of individual regions and indeed of individual ships.”4 Using the ship records of the Middelburgse Commercie Compagnie (MCC), this paper examines regional patterns of trade between Dutch buyers and African suppliers on the Windward Coast in the second half of the eighteenth century. After losing access to West Africa’s more productive slaving zones, around 1740 the Dutch found new markets for slaves on the Windward Coast, where conditions for trade were far less favorable. The archives of the MCC, the largest private slave trading firm in the eighteenth-century Netherlands, document unusual methods of trading in this part of Africa. In most African regions slave trading was concentrated in a limited number of ports, but trade on the Windward Coast was spread out over numerous small embarkation points. Through systematic analysis of the MCC records, this paper highlights the most significant features of this trade and doing so addresses a number of enduring misunderstandings about the economic history of the Windward Coast. First, Kru-speakers formed a considerable part of the slaves the Dutch carried from Africa after 1740, despite the absence of major embarkation points on the Kru Coast. Second, among the Kru slaves were disproportionally high numbers of African Economic History v.38(2010):29–51

30 JELMER VOS women and children, although the overall ratio of children on board Dutch slave ships was drastically reduced by selective purchasing on the Ivory and Gold Coasts. Third, Cape Lahou, on the eastern Ivory Coast, was by far the most important slaving port for Dutch free traders in West Africa and was, in this period, rivaled only by Malembo on the northern Angolan coast.

I

The significance of the Windward Coast for eighteenth-century slavers from the Netherlands was first acknowledged by Johannes Postma in his pioneering work on the Dutch Atlantic slave trade.5 Postma calculated that after 1740, when the West India Company (WIC) had lost its monopoly on the African trade and disengaged from Atlantic slaving, Dutch free traders on the coast of West Africa obtained roughly 70 percent of their slave cargoes on the Windward Coast, while the remaining purchases were completed on the Gold Coast. He further estimated that, overall, the Dutch colonies in the Americas drew roughly 40 percent of their slave labor force from the Windward Coast. Recent calculations based on Voyages: The Trans-Atlantic Slave Trade Database (TSTD2) as well as the data presented in this paper indicate that Postma’s findings were largely correct.6 However few scholars have thus far recognized the implications of these results for the history of slavery in the Atlantic world, in particular the ascendancy of Kru- speaking groups and Cape Lahou among the African suppliers of slaves to the Dutch Americas. Postma highlighted the prime position of Lahou on the Windward Coast, an issue taken up by Philip Curtin in his 1975 re-assessment of the volume of the Atlantic slave trade. According to Curtin, the pre- eminence of Lahou was due, first, to its location at the mouth of the partly navigable Bandama River, which gave local trade brokers access to Juula merchant networks in the interior; and, second, to the eighteenth century migration of the Baule (an Akan sub-group), towards the lower Bandama, which by provoking warfare with established Senufo and Guro THE SLAVE TRADE FROM THE WINDWARD COAST 31 communities created a supply of captives for export.7 In later years, Curtin linked the Akan expansion into modern Ivory Coast to the creation around 1750 of a new trade route connecting Lahou to Jenne in the Sudan.8 Adam Jones and Marion Johnson criticized these explanations. Although they primarily questioned the foundations of Curtin’s measurements of eighteenth-century slave exports from the region, they also hinted at flaws in Curtin’s historical account of the prominent role of Lahou in the Atlantic trade. Jones and Johnson doubted that a trade route connecting the lower Bandama with the Sudan still existed in the middle of the eighteenth century. By contrast, they reasoned that this route had probably been closed due to the disruptive impact of the Akan migrations. More fundamentally, in their view, there was too little evidence available at the time to sustain the theory that increasing slave exports from the Ivory Coast were the result of state expansion and warfare. “The real origins of slaves were far more complex and varied,” they suggested, cheekily referring to the slave narratives collected in Curtin’s own Africa Remembered.9 It will be clear from this paper that the growth of Cape Lahou as a major port in the Atlantic slave trade is a subject in need of further study.10 The origins of the slaves exported from Lahou may as yet be unknown, but very likely local merchants were able to tap into slave supply networks that were non-existent elsewhere on the Windward Coast. Paradoxically, Curtin, as well as Jones and Johnson, borrowed their views on Juula commerce to the lower Bandama River from the seminal work of Yves Person on the southern savanna region of West Africa.11 Supporting the position taken by Jones and Johnson, Person postulated that the Akan migrations cut off Lahou from access to Sudanese markets deeper inland. However, as pointed out in the final section of this paper, other studies suggest that even if the Baule excluded the Juula merchants from their realm, the former continued the north-south trade connecting the Sudan with the Atlantic coast.

32 JELMER VOS

II

The Windward Coast stretched from Cape Mount in the northern corner of modern-day Liberia to Assinie on the eastern border of modern Côte d’Ivoire. This definition of the Windward Coast is the one used in TSTD2 and was first employed by Curtin in his 1969 Census.12 The name refers to the coast’s location west, or windward, of the Gold Coast and thus originated from a European outlook on the African coast; for the British and the Dutch on the Gold Coast, windward started where the coastal forts ended.13 Paul Lovejoy has recently argued that such a definition has little meaning from an African perspective – and there is no point in denying that – although the Atlantic slave trade was of course created by Europeans and Africans alike.14 From the perspective of both African and European merchants it would make sense to consider the area around Cape Lahou, the so-called Quaqua Coast, as part of the Gold Coast. The MCC records and other primary sources indicate that, besides slaving, Lahou also participated in the gold trade and Lahou was probably connected to the same commercial networks that supplied the Gold Coast with commodities for export.15 As this paper will point out, however, trade on the coast west of Lahou was in a number of ways significantly different from other slaving areas in Africa. To facilitate comparison with Voyages: The Trans-Atlantic Slave Trade Database, the regional demarcations originally used by Philip Curtin are maintained. Much like Sierra Leone to the north, the Windward Coast was an area for which slave traders showed little interest before the eighteenth century. European visitors to West Africa around 1700 observed that exports from the region centered on ivory and did not include slaves. Along the coast from Cape Mount to Cape Palmas, Kru-speaking communities also still exported pepper grains (Aframomum melegueta). This region, hence known to Europeans as the Malaguetta or Grain Coast, was furthermore frequented by Atlantic traders in transit to lower Guinea to buy provisions, notably rice, water and firewood. Avoiding the controlled monopoly zones of Britain and the Netherlands on the Gold Coast, interlopers of different nationalities targeted the coast between THE SLAVE TRADE FROM THE WINDWARD COAST 33

Cape Lahou and Assinie to purchase ivory and small amounts of gold dust.16 These interlopers may have purchased small numbers of slaves as well.17 Throughout the history of the Atlantic slave trade, no region in Africa supplied fewer slaves to the Americas than the Windward Coast. According to the TSTD2, an estimated total of 337,000 slaves were carried from the Windward Coast, of which 290,000 embarked in the eighteenth century alone. These estimates are much lower than those previously calculated by Curtin, whose latest figure was 530,000 for the period 1711-1800,18 but much higher than the 143,000 suggested for the eighteenth century by the 1999 slave trade database on CD-ROM.19 Furthermore, the Windward Coast was the last region in Africa which European merchants opened for transatlantic slaving, although it already had exported pepper to European markets since the late fifteenth century. In the sixteenth century, a small number of slaves were sold to the Portuguese, who also used to carry slaves from the Windward Coast to Elmina.20 During the seventeenth century, just a few European ships appear to have loaded slaves there.21 The Windward Coast only started to supply slaves on a significant scale for the Atlantic market after 1700. Slave exports increased in the second quarter of the eighteenth century, while a peak was attained in the third quarter, which was largely due to British activities. Liverpool merchants in particular traded heavily on the northern part of the coast (especially at Cape Mount), often to complement a trade that started in Sierra Leone. As the British withdrew from the trade after 1807, the Windward Coast’s function in Atlantic slavery was severely cut. Adding to declining British demand for slaves, another nation with a major stake in the region, the Netherlands, also abandoned the slave trade shortly after 1800. The Netherlands were, after Britain, the second largest national carrier of slaves from the Windward Coast. In the eighteenth century, Dutch vessels shipped roughly 90,000 slaves from the region, almost all of them destined for the Guyanas. But it is remarkable that the Dutch only began trading slaves on the Windward Coast after 1740, that is, after the WIC had largely withdrawn from the Atlantic trade.22 Whereas

34 JELMER VOS the WIC never had a strong commercial interest in the region, interlopers from Zeeland frequented the Windward Coast when the WIC held a monopoly over the West African trade (which effectively ended in 1734), thereby setting a pattern for the later free traders.23 The directors of the Zeeland-based MCC were well informed about the interloping activities of the Middelburg and Vlissingen merchants and the African markets in which they operated; in the 1720s they even participated in an illegal trade expedition to the Loango Coast.24 Although Dutch purchases fell far behind those of Britain in terms of volume, no other slave trading nation relied more on the Windward Coast than the Netherlands. Whereas British vessels loaded 6 percent of all slaves they carried across the Atlantic on the Windward Coast, Dutch traders obtained 16 percent of their slave cargoes from this region. In the eighteenth century alone, the Windward Coast accounted for nearly a quarter of all Dutch slave exports from Africa. For Dutch free traders the Windward Coast gained particular importance as by the middle of the eighteenth century British, Portuguese and French merchants had squeezed the Dutch out of the Bight of Biafra and then the Bight of Benin, where some of Africa’s most productive slaving ports were located. In places like Old Calabar and Ouidah, large numbers of slaves could be obtained in short spans of time. As business in these regions proved more cost-effective for European traders, it also became competitive, and the Dutch were eventually boxed out.25 Dutch exports from the Bight of Biafra dropped from six thousand slaves in the decade prior to 1670 to two thousand slaves in the decade after; in subsequent years these numbers dwindled even further. For some time the Dutch still maintained a foothold in the Bight of Benin, but after 1740 their slave purchases in this region also collapsed. Seeking new alternatives for the growing plantation colonies in the Guyanas, Dutch traders migrated to the Windward Coast while they also increased their dependence on the Loango Coast and the Gold Coast. In each of these areas, the Dutch slave trade always centered on a single port. Thus, as the eighteenth century progressed, on the Loango Coast Dutch slavers increasingly concentrated in Malembo. On the Gold THE SLAVE TRADE FROM THE WINDWARD COAST 35

Coast, Elmina emerged as the major center of trade, while on the Windward Coast Cape Lahou was the only major supplier of slaves to the Dutch. Together these three ports accounted for roughly two-thirds of the Dutch slave trade out of Africa between 1730 and 1800.26 It should be emphasized, however, that Lahou and the Gold Coast were part of the same trade route, commonly described as the Dutch “Guinea trade.” On this route, ship captains (or their supercargoes) typically started purchasing slaves on the upper Windward Coast, most often near Cape Mesurado. Then they sailed down along the coast, all the while trading slaves, until they reached Lahou on the Ivory Coast, where most of the purchases were done. Ultimately, the ship’s hold was “filled up” in one or more of the Gold Coast ports. In the 1740s, Axim was often the final port of slave embarkation, followed in the 1750s by Anamabo and Apam, and by Elmina from the 1760s onward. By contrast, Malembo was part of the so-called “Angola trade,” which also included Loango and Cabinda, although in the latter ports the Dutch had far less purchasing power as the century wore on.

III

This study is based on a database of 72 slaving voyages undertaken by the MCC between 1740 and 1802. For each voyage, the database provides the number of slaves purchased at every single embarkation point, with breakdowns for the age and gender characteristics of the slaves. The database also indicates the time vessels spent on the African coast purchasing slaves, as well as slave mortality rates on the coast, during the Middle Passage and after arrival in the Americas. As the largest private slave trading firm in the Netherlands, the MCC organized a total of 114 slaving voyages between 1732 and 1802, all but two of which were carried out between 1740 and 1794.27 Eighty voyages were destined to the Guinea Coast, three to Gabon, and thirty-one to the Loango Coast. The database builds on information taken from the surviving log and/or trade books of 69 Guinea voyages and the three

36 JELMER VOS voyages to Gabon, all located in the provincial archives of Zeeland in Middelburg.28 The database lists 88 different places on the West African coast where MCC vessels purchased slaves. For the purpose of analysis, these ports have been divided into five groups.

Table 1: Regional Distribution of Slave Purchases in the MCC Guinea Trade, 1740–1803 (72 voyages) Region Ports Slaves Percent Total TSTD2 Dutch estimates Guinea trade, 1741–1803 Senegambia & off-shore – – – – 426 Atlantic Sierra Leone 5 97 0.5 679 829 Windward Coast west of Lahou 44 4,520 23.9 31,638 (Kru Coast) Cape Lahou 1 7,179 38.0 50,249 Windward Coast 77,867 east of Lahou 9 1,039 5.5 7,272 (Quaqua Coast) Gold Coast 21 5,339 28.2 37,370 48,942 East of Gold 8 735 3.9 5,145 4,715 Coast All regions 88 18,909 100 132,353 132,353 Source: MCC database and David Eltis, David Richardson, Manolo Florentino, and Stephen D. Behrendt, Voyages: The Trans-Atlantic Slave Trade Database (www.slavevoyages.org).

The first group includes five ports in the Sierra Leone region, north of Cape Mount, which accounted for less than one percent of the MCC’s THE SLAVE TRADE FROM THE WINDWARD COAST 37 slave trade in West Africa. The second group comprises 44 ports on the Windward Coast from Cape Mount to Little Lahou, a long stretch of coast inhabited by Kru-speaking groups and also known as the Kru coast. Cape Lahou (aka Grand Lahou), which alone accounted for nearly 40 percent of the MCC’s slave trade in West Africa, constitutes the third group. The fourth group includes nine ports on the Windward Coast east of Lahou. European merchants named this region the Quaqua coast, which also included Lahou and was famed for its exports of striped cloth.29 The Gold Coast, with 21 ports from Newton to Accra, makes up the fifth group, while eight ports located east of the Volta River form the final group (better known among historians as the Bight of Benin and the Bight of Biafra). None of the MCC vessels in the database purchased slaves in Senegambia. Table 1 shows the regional distribution of the total number of slaves purchased by the MCC in West Africa. As is clear from this table, the Windward Coast accounted for two-thirds of the MCC’s slave exports from Guinea. This figure is representative of the total Dutch slave trade in the free trade era. Extensive consultation of MCC ship journals proves that vessels from Amsterdam, Rotterdam, and Vlissingen followed the same trading pattern as the MCC ships. We can, therefore, confidently project the regional percentages calculated from the MCC database onto the TSTD2 estimate for the total Dutch slave trade in West Africa between 1741 and 1805. The result, as shown in Table 1, is that the Windward Coast supplied approximately 89,000 slaves to the Dutch in this period, which means 11,000 more slaves than the Transatlantic Slave Trade Database suggests. The difference between the MCC database and TSTD2 stems from the fact that the former has not yet been fully integrated in the latter. The Transatlantic Slave Trade Database consequently underrates the size of the slave trade from Cape Lahou, for which it suggests an estimated total of nearly 15,000 slaves embarked by all national carriers between 1709 and 1820. At the same time, TSTD2 is biased towards Elmina, which in the absence of contrary evidence has been viewed as the main location of slave embarkation in case it was the last port of call on the African coast,

38 JELMER VOS which it often was for Dutch vessels. As a result, the Transatlantic Slave Trade Database inflates the number of slaves carried from the Gold Coast at the cost of the Windward Coast, where in most cases the majority of slaves was really purchased. The method of slave procurement proves as important as the numbers. Compared to other regions in Africa, trade on the Windward Coast was following an unusual pattern. Ship captains generally called at different places along the coast (of the 54 identified places of purchase on the Windward Coast, some were almost standard ports of call, whereas others were visited on a less frequent basis) but at most they purchased only small numbers of slaves. Table 2 shows that, except for Cape Lahou, not a single place on the Windward Coast supplied more than seven slaves per voyage. One reason for this seemingly haphazard nature of the trade can be found in the region’s geography. This part of the West African coast is characterized by steep beaches, heavy surf, few decent river inlets and a shortage of good anchorages throughout. “For lack of natural harbors,” Curtin explains, “trade along this coast was fragmented, divided among many small ports or shipping points.”30 Due to the absence of good landing spots, maritime trade had to be ship- based. The MCC logbooks describe in detail how captains used to stop at several points along the coast, often sailing back and forth, waiting for African traders to come out by canoe or for a smoke signal from land telling there was an opportunity to barter. Alternatively, captains sent boats ashore to collect information about the possibilities of trade. In this “troque au vol” style of trade, Dutch captains sometimes bought a handful of slaves at a single place, but quite often they bought no slaves at all. In exploiting practically all the markets for slaves that existed on the Windward Coast, the Dutch distinguished themselves from British and French buyers, who generally focused on the northern part of the coast (especially Cape Mount and Bassa). The fact that, according to TSTD2, only twenty-four British and eighteen French vessels traded slaves at Cape Lahou throughout the eighteenth and early nineteenth THE SLAVE TRADE FROM THE WINDWARD COAST 39 centuries is astounding, especially since this port had the capacity to supply the Dutch with an average of 800 slaves per year.

Table 2: Ten Most Important Places of Purchase in the MCC Guinea Trade, 1740–1803 (72 voyages) Place Region Slaves Slaves/voyage % of total trade Cape Lahou Quaqua coast 7,179 100 38.0 Elmina Gold Coast 1,996 28 10.6 Axim Gold Coast 1,111 15 5.9 Anamabo Gold Coast 554 8 2.9 Kru Settra Kru coast 501 7 2.6 Apam Gold Coast 458 6 2.4 Sassandra Kru coast 451 6 2.4 Cape Mount Kru coast 416 6 2.2 Grand Bassam Quaqua coast 408 6 2.2 Rio Cavaly Kru coast 360 5 1.9 Total 13,434 71.0 Source: MCC database

Equally striking are the irregular age and gender distributions in the different port groups. As Table 3 demonstrates, the share of males among the slaves purchased on the coast west of Lahou was a mere 52 percent; meanwhile children formed an extraordinarily large fraction of the slaves taken from the Kru coast (42 percent). Cape Lahou exported relatively few children, but the female share was still exceptionally high (45 percent). East of Lahou, slave purchases focused increasingly on males and included very few children. Compared to the Atlantic slave trade in general, females were overrepresented in the Guinea trade and children underrepresented, despite the large numerical presence of the latter among the slaves carried from the coast west of Lahou. On average, the MCC vessels spent 222 days on the African coast buying slaves. The average number of slaves carried per voyage was

40 JELMER VOS

269, which translates to a purchase rate of 1.2 slaves per day (which was similar to daily loading rates of Liverpool ships trading on the Windward Coast). One would expect that the long trading time in Africa raised not only the costs of outfitting a slaving voyage, but also the mortality rates among crew and slaves.31 The MCC data on slave mortality, however, run contrary to expectations. First, the overall death rate of 12.9 percent was low by Dutch standards. TSTD2 indicates that of all slaves shipped by the WIC from West Africa, 16.8 percent had died before sale in the Americas. Furthermore, mortality rates were higher during the Middle Passage, which on average lasted about seventy days, than during the seven or eight months that MCC vessels spent loading slaves on the coast. The number of slaves who died on board ship in Africa was 4.3 percent of the total purchased, while the death rate during the Atlantic crossing amounted to 7.7 percent. The reason for this comes from the fact that the vast majority of slaves embarked during the latter part of the loading period, whereas the impact of disease factors rose as the voyage progressed.

Table 3: Age and Gender Distribution of Slaves in the MCC Guinea Trade, 1740–1803 (72 voyages) Region Percent male Percent children Windward Coast west of 51.9 42.1 Lahou Cape Lahou 54.4 11.3 Windward Coast east of Lahou 58.2 7.4 Gold Coast 65.8 8.7 Total Guinea Coast 57.1 18.5 Total Africa (TSTD2) 64.6 20.9 Source: MCC database and TSTD2

Although the study of slave mortality falls largely outside the boundaries of this paper, it touches directly upon questions about regional differences in slave supplies. Using a sample of 39 MCC slaving voyages, Simon Hogerzeil and David Richardson have recently linked THE SLAVE TRADE FROM THE WINDWARD COAST 41 shipboard mortality data to differing “purchasing strategies” in the Dutch Guinea trade.32 They demonstrate that the number of slaves obtained per week by the MCC vessels rose as time progressed. “During most of the purchase period, [the] pattern of acquisitions was fairly homogeneous or constant... with slaves being acquired at the rate of up to five a week up to 20 weeks before departure from Africa, and rising to about eight a week... between 20 and six weeks before departure.”33 During the last month acquisition patterns changed radically, as the vessels used to complete their purchases at one of the Gold Coast forts, often by placing an order in advance. Indeed, 22 percent of the slaves were bought in the last of the seven or eight months spent on the African coast, and 15 percent during the final week before sailing. Unfortunately, Hogerzeil and Richardson fail to recognize the importance of Lahou as a supplier to the MCC vessels. “Slaves were acquired at the Windward Coast at various places, with no single location predominating,” they state.34 However, the small increment in the purchase rate between twenty and six weeks before departure came from the vessels’ stay at Lahou in this period. Nevertheless, the implication of these findings for the analysis of shipboard mortality is that, although MCC vessels spent an exceptionally long time on the African coast, the time during which slaves were exposed to risk factors on board ship was on average significantly shorter. Hogerzeil and Richardson also underline that “the last minute surge in purchases… centered disproportionately on adults, especially men.”35 Mortality normally increased for all slaves throughout the loading phase, as well the Middle Passage, but death rates of children and women usually remained lower than those of men. The fact that most men embarked during the last part of the loading phase helped to keep mortality rates down. Examining the peculiar age and gender ratios of the MCC slave acquisitions, Hogerzeil and Richardson introduce the concept of “purchasing strategies.” In their view, the disproportionately high number of women and children loaded on the Windward Coast was the result of a deliberate choice of the MCC captains. “As time spent on

42 JELMER VOS board ship was widely considered a health hazard for slaves, a deferral in purchasing high-value adult males was perhaps to be expected. In other words, it reflected a policy adjustment by MCC shippers to coastal price differentials among categories of slaves largely dictated, it seems, by planter preferences for adult males in the Americas.”36 In addition, the comparatively high propensity of slaves from the Windward Coast to rebel on board ship might also have influenced regional patterns of slave exports.37 “The relative lateness in finalizing the loading of adult males may… have been just as much an adjustment to fears of slave rebellion as to the price premium on adult male slaves.”38 This study argues for a more Afrocentric explanation of the regional differences in the MCC’s Guinea trade. Hogerzeil and Richardson do not exclude the possibility that African factors influenced Dutch “purchasing strategies” on the Windward Coast. The rising demand for African workers in Dutch Guyana after 1750, combined with a strong competition for slaves in other regions of Atlantic Africa, “brought MCC shippers… into contact with emergent slave supply systems that were perhaps less efficient than the well-established systems of the Gold Coast in satisfying shippers’ preferences for slaves.”39 Indeed, the picture that emerges from the MCC ship journals is that, on the Windward Coast west of Lahou, Dutch captains had little choice in their purchases. The slaves offered for sale were small in number and were mainly women and children.

IV

We now return to the origins of the slaves exported from the Windward Coast, analytically divided in three port groups. Cape Lahou was by far the largest slaving port on the Windward Coast, and it was predominantly trading with vessels coming from the Netherlands. Between 1740 and 1805, when Lahou exported an estimated 50,000 slaves via Dutch vessels (Table 1), it stood out as the most important supplier of African labor to the Dutch Americas after Malembo, the center of the Dutch slave trade on the Loango Coast. Local brokers and THE SLAVE TRADE FROM THE WINDWARD COAST 43 small independent traders sold on average 100 slaves per voyage to the MCC (Table 3). Although prices were much higher here than on the coast further windward, Lahou was attractive because it was able to supply slaves in much larger numbers and more swiftly. With average annual sales of more than 800 slaves after 1740, there can be little doubt that Lahou merchants tapped into hinterland markets that were unavailable to most other ports on the Windward Coast. The historical literature provides a few, though contradictory, clues about the commercial networks that connected Lahou with the interior and, therefore, about the possible origins of the exported slaves. The work of Yves Person suggests that by the time the Dutch intensified their commercial relations with Lahou, the latter was already cut off from a direct north-south connection with the Sudan. This route, which linked the lower Bandama River via Kong and Old Boron to the Sudanese centre of Jenne, had been established by Malinke and Juula groups in the early 1600s. When the Baule migrated into the Bandama region in the early eighteenth century, however, the Juula trade networks were dismantled. Long-distance trade was replaced by a “system of relays carrying goods from place to place under the control of local chiefs. Thus from the first quarter of the eighteenth century the Bandama route was closed to Sudanese trade… On the coast, Lahou fell into a decline and ships stopped calling there.”40 Excluded from the trade networks in the Baule savanna, Juula merchants concentrated on alternative channels to the Atlantic coast, notably Kumase in the Asante kingdom, but also the Comoe River that gave access to outlets at Bassam and Assinie.41 The evidence presented in this paper indicates, however, that Lahou did not fall into decline after 1700 but instead flourished through its commercial relations with the Dutch. In a number of publications on the pre-colonial history of the Baule and the hinterland of the Quaqua Coast, Timothy Weiskel has also failed to recognize the economic vitality of Lahou in the eighteenth century. “During most of the eighteenth century,” Weiskel has argued, “the trade arriving at Grand Lahou, the natural outlet for the Baule southern trade, was never significant enough to attract sustained European interest.”42 In his view the commercial

44 JELMER VOS networks of the Baule that replaced the long-distance trade routes of the Juula were, because of the former’s Akan connection, directed more towards the east than to the south. It is worth pointing out, in this regard, that except for three voyages to Gabon, every single MCC vessel in our database purchased slaves at Lahou, which was moreover the principal port of embarkation in 48 out of a total of 69 cases. In contrast to Person and Weiskel, Jean-Piere Chauveau has argued that the historical north-south trade between Lahou and the Sudan was not interrupted by the Baule migrations. In his analysis, the Baule monopolized the transportation of goods within their country while they traded with foreign merchants at so-called “transit markets” at the frontiers of their territory. Thus they maintained commercial relations with the Juula in the north, with the Guro in the west (via Toumodi and Kokumbo), and with merchants from Lahou in the southern town of Tiassalé, located at the banks of the lower Bandama (only from here to the coast was the river navigable). In this the Baule largely followed the old north-south axis of trade, through which Atlantic imports were exchanged for products like ivory, gold and slaves.43 As for the origins of the slaves traded by the Baule, Chauveau suggests that the majority was purchased from Guro and Juula merchants and probably came from the north.44 The strong ties between the Baule and the Guro have been confirmed by the oral traditions Claude Meillassoux collected in Guro country. According to his informants, European imports reached the Guro via the port of Bassam and the Baule town of Toumodi. In return, the Guro supplied the Baule with ivory, war captives as well as slaves purchased from the northern Malinke.45 Although Lahou is not mentioned in this context, it seems clear that a substantial part of the slave imports of the Baule, on whom Lahou merchants relied for their supplies, stemmed from Sudanese sources. Exports through ports on the Quaqua Coast east of Lahou, accounting for a mere 5 percent of the Dutch slave trade in West Africa, seem to have relied on a north-south trade with the Sudan as well. The work of Claude Hélène Perrot shows that the two most important ports on this part of the coast, Grand Bassam (or Bassam) and Assinie, were THE SLAVE TRADE FROM THE WINDWARD COAST 45 outlets for the Mbrasue market in the land of the Anyi of Ndenye. Founded in the late eighteenth century on the banks of the Comoe River, Mbrasue obtained its slaves predominantly from Baule suppliers in the northwest, Kumasi in the east, and Bonduku in the north; from Bonduku there were connections through Kong and Buna with the Sudanese commercial hub of Jenne.46 In short, all ports on the Quaqua Coast, including Lahou, were dependent on Sudanese slave supplies. Even slaves imported through the Guro and the Asante probably originated in large part from the northern savannas. Warfare in the Baule and Guro regions, or at the fringes of the Asante kingdom, might also have produced captives for the Atlantic trade. A current trend in the scholarship on African slavery and the slave trade is to emphasize the production of slaves by means more subtle than inter-societal warfare or the tapping into far away markets by long- distance trade caravans.47 It seems that Africans sold into the Atlantic slave trade from the coast west of Lahou were mostly victims of such local or internal means of enslavement. Again geography was one of the underlying factors. The long stretch of coast between Sierra Leone and the Bandama River was covered by dense, often impenetrable forests. Overall the region was sparsely populated; although rivers crossed it, waterside settlements were usually found only at the navigable sections near the river mouths. According to Person, “this coast, hedged in by the great forest, made commerce unattractive up to the colonial era.”48 The expansion of Malinke, or Juula, trade networks that connected the Atlantic coast of nearby regions with the Sudanese interior, in this region halted at the forest edges. Thus “the impenetrable wall of the great forests” cut off local Kru and Mande populations from commercial connections with the coastal hinterland.49 In the absence of professional long-distance traders, commodities were carried from place to place through networks of local big men. The lack of extensive trade connections with the interior has, in the historical literature on the Windward Coast, supported the idea that Kru- speakers were insignificant participants in the Atlantic slave trade. “These coastal peoples engaged in a good deal of fishing and trading,”

46 JELMER VOS

Person argued, “but the slave trade itself often bypassed these desolate areas, which rarely saw slaves from the interior.”50 Another scholar observed that “A few isolated reports of slaving can be found for the eighteenth century, but by 1750 or so it appears that the Malaguetta Coast no longer competed seriously with other areas as a source of slaves.”51 This paper argues the opposite: in the latter half of the eighteenth century Kru-speaking groups became important suppliers of slaves especially to traders from Liverpool and the Netherlands. Indeed, as Table 1 shows, in this period they accounted for nearly a quarter of all slaves the Dutch carried from West Africa.52 The main features of this trade may be summed up as follows. First, numerous communities on the coast between Cape Mount and Lahou consistently provided slaves to Dutch vessels, but they all did so on a very small scale. Second, more adult women than men were sold into the Atlantic slave trade, while children formed an exceptionally high portion of the exported slaves (among the children boys predominated, which accounts for the overall male-to-female ratio of 52:48). Curtin suggested that “most of these ports drew on their immediate hinterland within the forest.”53 The coastal geography prohibited the development of large- scale embarkation points and of connections with potential supply networks in the interior. But Kru-speaking groups disposed of criminals in society by selling them in exchange for European imports, while other slaves were the victims of violent conflicts between communities.54 Kru villages also raided their neighbors for slaves. According to the supercargo of a Dutch slaving vessel, near the Cestos River night-time marauding was a common means of obtaining slaves.55 Women and children were of course more vulnerable to village raids than adult males. The fact that these two groups made up a large portion of the slaves exported from the region, might indicate that kidnapping was a prime method of slave production on the Windward Coast.

V

THE SLAVE TRADE FROM THE WINDWARD COAST 47

Around 1740, because of a pressing demand for slaves in the Americas and sturdy competition in the dominant supply regions of West Africa, Dutch slave traders had a strong incentive to explore new markets on the Windward Coast. The African response was affirmative, sending many thousands into slavery, but with the exception of Cape Lahou, local supply systems were rather inefficient. The time slave ships spent on this part of the coast was exceptionally long as purchases were spread over numerous small embarkation points, each of which produced slaves in only very small numbers. Among the enslaved, moreover, the share of women and children, who were probably obtained from villages in the immediate hinterland, was disproportionately large. But as long as prices and mortality rates in Africa were kept low, while prices in the Americas remained high—and the composition of the slave cargoes could be adjusted in Lahou and on the Gold Coast—Dutch slave trading firms saw the Windward Coast as a creditable market. This paper has shown that, as a result of these market decisions, in the second half of the eighteenth century the plantation colonies of Surinam, Demerara and Essequibo began to receive a fair number of Kru-speaking slaves and an even larger group of slaves hailing from the hinterland of Cape Lahou.

1 Earlier versions of this paper were presented at the Social Science History Association Conference in Chicago (2007), the European Social Science History Conference in Lisbon (2008), the Atlantic History seminar at the University of Turku (2008), the African Economic History Workshop at the London School of Economics (2008), and the Centre International de Recherches Esclavages in Paris (2009). I thank Ann Carlos, Frank Lewis and my colleagues from the history department at Old Dominion University for their useful comments on the paper. Of course I assume all responsibility for its contents. 2 Recent monographs include George E. Brooks, Eurafricans in Western Africa. Commerce, Social Status, Gender, and Religious Observance from the Sixteenth to the Eighteenth Century (Athens: Ohio University Press, 2003);Walter Hawthorne, From Africa to Brazil. Culture, Identity, and an Atlantic Slave Trade (Cambridge: Cambridge University Press, 2010); Robin Law, Ouidah. The Social History of a West African Slaving ‘Port’, 1727-1892 (Athens: Ohio University Press, 2004); Kristin Mann, Slavery and the Birth of an African City. Lagos, 1760-1900 (Bloomington: Indiana University Press, 2007); G. Ugo

48 JELMER VOS

Nwokeji, The Slave Trade and Culture in the Bight of Biafra. An African Society in the Atlantic World (New York: Cambridge University Press, 2010); Mariana Candido, An African Slaving Port on the Atlantic World. Benguela and its Hinterland (New York: Cambridge University Press, forthcoming); Roquinaldo Ferreira, Atlantic Microhistory. Slaving, Transatlantic Networks and Cultural Exchange in Angola (New York: Cambridge University Press, forthcoming). 3 Philip D. Curtin, The Atlantic Slave Trade. A Census (Madison: The University of Wisconsin Press, 1969); “Measuring the Atlantic Slave Trade,” in Stanley L. Engerman and Eugene D. Genovese, eds., Race and Slavery in the Western Hemisphere. Quantitative Studies (Princeton: Princeton University Press, 1975). 4Adam Jones and Marion Johnson, “Slaves from the Windward Coast,” Journal of African History 21 (1980), 34. 5 Johannes Menne Postma, The Dutch in the Atlantic Slave Trade, 1600-1815 (Cambridge: Cambridge University Press, 1990), 122-23. 6 Jelmer Vos, David Eltis, and David Richardson, “The Dutch in the Atlantic World: New Perspectives from the Slave Trade with Particular Reference to the Origins of the Traffic,” in David Eltis and David Richardson, eds., Extending the Frontiers. Essays on the New Transatlantic Slave Trade Database (New Haven: Yale University Press, 2008), Table 8.2. 7 Curtin, “Measuring the Atlantic Slave Trade,” 120-21. Curtin used Postma’s 1970 Ph.D. thesis, which was published as a book in 1990. 8 Phillip D. Curtin et al., African History. From Earliest Times to Independence. Second Edition (London: Longman, 1995), 199. 9 Jones and Johnson, “Slaves from the Windward Coast,” 30-31. Philip D. Curtin, ed., Africa Remembered. Narratives by West Africans from the Era of the Slave Trade (Madison: The University of Wisconsin Press, 1967), 145-69, documents two cases of individuals of Sudanese origin who were sold into slavery and shipped from ports on the Gold Coast. 10 One recent study of slavery in the forest societies of Côte d’Ivoire ignores Cape Lahou as a slaving port and instead focuses on places that probably only became important export outlets in the produce trade of the nineteenth century. Harris Memel-Fotê, L’esclavage dans les sociétés lignagères de la forêt ivoirienne, XVIIe-XXe siècle (Paris: les éditions CERAP, IRD, 2007). 11 Yves Person, “Le Soudan nigérien et la Guinée occidentale,” in Hubert Deschamps, ed., Histoire générale de l’Afrique noire, de Madagascar, et des archipels, vol. 1 (Paris: Presses Universitaires de France, 1970); “The Atlantic Coast and the Southern Savannas, 1800-1880,” in J. F. H. Ajayi and Michael Crowder, eds., History of West Africa. Volume 2, Second Edition (London: Longman, 1987). 12 Curtin, Atlantic Slave Trade, 128. 13 Only the French attempted settlement on the Windward Coast. They constructed a fort at Assinie in 1698, which they abandoned in 1704. See Jean- THE SLAVE TRADE FROM THE WINDWARD COAST 49

Michel Deveau, L’Or et les esclaves. Histoire des forts du Ghana du XVIe au XVIIIe siècle (Paris: UNESCO, 2005), 301. 14 Paul Lovejoy, “Extending the Frontiers of Transatlantic Slavery, Partially,” Journal of Interdisciplinary History 40, 1 (2009), 63. 15 In 1762, the supercargo of a Dutch merchant ship pointed out that the ounce was the standard unit of exchange from Lahou eastward, as it was on the Gold Coast. See Jean Pierre Plasse, Journal de bord d’un négrier (Marseille: Éditions le mot et le reste, 2005), 79. See also Willem Bosman, A New and Accurate Description of the Coast of Guinea. Divided into the Gold, the Slave, and the Ivory Coasts (New York: Barnes & Noble, 1967), 492; P.E.H. Hair et al., eds., Barbot on Guinea. The Writings of Jean Barbot on West Africa 1678-1712, 2 vols. (London: Hakluyt Society, 1992), 300-301. 16 Hair, Barbot on Guinea, 234-316, 331-37; Bosman, New and Accurate Description, 469-93. 17 For instance, a Vlissingen interloper carried slaves from Lahou in 1709 (TSTD2, voyage id 33652). 18 Curtin, “Measuring the Atlantic Slave Trade,” 112. 19 David Eltis et al., The Trans-Atlantic Slave Trade. A Database on CD-ROM (Cambridge: Cambridge University Press, 1999). 20 George E. Brooks, Landlords and Strangers. Ecology, Society, and Trade in Western Africa, 1000-1630 (Boulder, CO: Westview Press), 283. 21 The Transatlantic Slave Trade Database contains only three records of vessels carrying slaves from the Windward Coast before 1700 (voyage ids 24358, 33752, and 33831). 22 The WIC never withdrew completely from the Atlantic slave trade. According to the Transatlantic Slave Trade Database, at least fourteen WIC vessels carried African slaves to the Americas between 1741 and 1793. 23 On the Dutch interloper trade, see Ruud Paesie, Lorrendrayen op Africa. De illegal goederen- en slavenhandel op West-Afrika tijdens het achttiende-eeuwse handelsmonopolie van de West-Indische Compagnie, 1700-1734 (Amsterdam: De Bataafsche Leeuw, 2008). 24 C. Reinders Folmer-van Prooijen, Van goederenhandel naar slavenhandel. De Middelburgse Commercie Compagnie, 1720-1755 (Middelburg: Koninklijk Zeeuwsch Genootschap der Wetenschappen, 2000), 76, 89. See also Postma, Dutch in the Atlantic Slave Trade, 123. 25 David Eltis and David Richardson, “Productivity in the Atlantic Slave Trade,” Explorations in Economic History 32 (1995), 465-84; Vos et al., “Dutch in the Atlantic World,” 239. 26 Vos et al., “Dutch in the Atlantic World,” 240-242. 27 For the early MCC slave trade, see Reinders Folmer-van Prooijen, Van goederenhandel naar slavenhandel. 28 I thank Simon Hogerzeil for sharing his dataset of 39 MCC voyages with me.

50 JELMER VOS

29 Brooks, Landlords and Strangers, 300, 302; Colleen Kriger, “‘Guinea Cloth’. Production and Consumption of Cotton Textiles in West Africa before and during the Atlantic Slave Trade,” in Giorgio Riello and Prasannan Parthasarathi, eds., The Spinning World. A Global History of Cotton Textiles, 1200-1850 (Oxford: Oxford University Press, 2009), 114, and the literature there cited. 30 Curtin, African History, 198-99; also “Measuring the Atlantic Slave Trade,” 118. 31 For a comparison of crew wages between the commodity trades of the WIC and the MCC in West Africa, see Henk den Heijer, “The West African Trade of the Dutch West India Company, 1674-1740,” in Victor Enthoven and Johannes Postma, eds., Riches from Atlantic Commerce. Dutch Transatlantic Trade and Shipping, 1585-1817 (Leiden: Brill, 2003), 163. 32 Simon Hogerzeil and David Richardson, “Slave Purchasing Strategies and Shipboard Mortality: Day-to-Day Evidence from the Dutch African Trade, 1751-1797,” Journal of Economic History 67 (2007), 166-67. 33 Hogerzeil and Richardson, “Slave Purchasing Strategies,” 168. 34 Hogerzeil and Richardson, “Slave Purchasing Strategies,” 167. 35 Hogerzeil and Richardson, “Slave Purchasing Strategies,” 170. 36 Hogerzeil and Richardson, “Slave Purchasing Strategies,” 172. 37 David Richardson, “Shipboard Revolts, African Authority, and the Atlantic Slave Trade,” William and Mary Quarterly 58 (2001), 69-92. 38 Hogerzeil and Richardson, “Slave Purchasing Strategies,” 173. 39 Hogerzeil and Richardson, “Slave Purchasing Strategies,” 172. Note that, because of low productivity, prices for all slave categories were much lower on the Kru Coast than in Lahou or on the Gold Coast. From this angle, it would have made sense to buy as many slaves as possible on this part of the coast, whoever they were. 40 Person, “Atlantic Coast and Southern Savannas,” 265. 41 Person, “Soudan nigérien,” 284-5, 301. 42 Timothy Weiskel, “The Precolonial Baule: A Reconstruction,” Cahiers d’Études africaines 18 (1978), 511-12; also Weiskel, French Colonial Rule and the Baule Peoples. Resistance and Collaboration, 1889-1911 (Oxford: Clarendon Press, 1980), 8; Weiskel, “L’histoire socio-économique des peoples baule: problems et perspectives de recherche,” Cahiers d’Études africaines, 16 (1976), 380. Weiskel recognized the special relationship Lahou maintained with the Dutch, describing it nonetheless as a “feeble participation in the slave trade.” 43 Jean-Pierre Chauveau, “Notes sur les échanges dans le Baule précolonial,” Cahiers d’Études africaines 16 (1976), 578-84. 44 In the nineteenth century, especially under the rule of Samory, the Baule also imported large numbers of slaves from the Anyi and Asante in the east; see Chauveau, “Échanges dans le Baule précolonial,” 578, 591, 596. A fundamental problem in the reconstruction of Baule trade networks, and hence the origins of THE SLAVE TRADE FROM THE WINDWARD COAST 51

the slaves exported from Lahou, is that the most informative data on trade routes stems from European sources from the late nineteenth-century. 45 Claude Meillassoux, Anthropologie économique des Gouro de Côte d’Ivoire. De l’économie de subsistance à l’agriculture commerciale (Paris : Mouton, 1964), 265, 270-71, 274. 46 Claude Hélène Perrot, Les Anyi-Ndenye et le pouvoir aux 18e et 19e siècles (Paris: Publications CEDA, 1982). 47 See especially the set of articles on the slave trade in decentralized societies in the Journal of African History 42 (2001). 48 Person, “Soudan nigérien,” 292. 49 Person, “Atlantic Coast and Southern Savannas,” 257. For more details and a slightly different view, see Brooks, Landlords and Strangers, 72-73, 96, 282- 319. 50 Person, “Atlantic Coast and Southern Savannas,” 258. 51 Ronald W. Davis, Ethnohistorical Studies of the Kru Coast (Newark: Liberian Studies, 1976), 32. 52 Based on Postma’s figures of slave exports from the African coast, Alex van Stipriaan has postulated that Mandingos made up about a third of all slaves arriving in Surinam after 1730, with Kormantins (from the Gold Coast) and Loangos forming roughly equal shares. His assumption was that all slaves traded on the Windward Coast were Mandingo, hence of Sudanese origin, and thus ignored the possibility that many hailed from Kru communities along the coast itself. See ““Een verre verwijderd trommelen…” Ontwikkeling van Afro- Surinaamse muziek en dans in de slavernij,” in Ton Bevers et al, eds. De Kunstwereld.Produktie, distributie en receptive in de wereld van kunst en cultuur (Hilversum: Verloren, 1993), 144-45. 53 Curtin, African History, 199. 54 Jean-Baptiste Labat, Voyage du chevalier Des Marchais en Guinée, isles voisines, et à Cayenne, fait en 1725, 1726 & 1727 (Paris: Saugrain, 1730), 118- 19. 55 Plasse, Journal de bord, 47. See also Bosman, New and Accurate Description, 480-81.

THE SUPPLY OF SLAVES FROM LUANDA, 1768–1806: RECORDS OF ANSELMO DA FONSECA COUTINHO

Daniel B. Domingues da Silva University of Missouri at Columbia

Luanda was not a town in which the unwary amateur could hope to turn a profitable trade.1

he supply of slaves from Luanda, in Angola, was an activity that required significant experience in the transatlantic slave Ttrade. In the late sixteenth century, the Portuguese established a colony at Luanda, which became one of the principal ports of slave embarkation on the coast of West Central Africa. It supplied slaves mostly to the plantations and mines of Portuguese and Spanish Americas. As the demand for slaves tended to increase over time, the economy of Luanda expanded based largely on the trade of human beings. However, in the eighteenth century, at the peak of the transatlantic trade, foreign competitors challenged Luanda’s position in this activity by purchasing slaves in ports located north of Luanda.2 Additionally, Benguela, a Portuguese port situated south of Luanda, emerged at the beginning of the eighteenth century as an important source of slaves, providing further competition for merchants based in Luanda.3 As a consequence, Luanda merchants had to adjust to the new circumstances and devise strategies to face the increasing competition in the slave trade from West Central Africa, as the records of Anselmo da Fonseca Coutinho will show. The lives and commercial strategies of merchants based in Luanda have been explored comparatively recently. In 1972, Herbert Klein accessed the account books for the years between 1750 and 1760 of a Portuguese merchant resident in Luanda, Captain João Xavier da Proença e Sylva.4 In 1984, Joseph Miller analyzed the account books of a Portuguese royal officer, António Coelho Guerreiro, who resided temporarily in Luanda and had participated in the slave trade between 1684 and 1692.5 In 1985, Clarence-Smith explored the commercial strategies of slave traders operating in

African Economic History v.38(2010):53-76 54 DANIEL B. DOMINGUES DA SILVA

Brazil and Portuguese Angola during the period of suppression of the slave trade, in the second quarter of the nineteenth century.6 More recently, Roquinaldo Ferreira and Mariana Cândido have highlighted the complexity of merchant operations in Luanda and Benguela, exploring their commercial strategies in the slave trade as a group.7 Despite these contributions, no study has ever traced the individual career of a slave merchant based in Luanda from the beginning through to its completion. In 1806, Governor and Captain General of Angola, D. Fernando António de Noronha, sent the service records of Anselmo da Fonseca Coutinho to the Portuguese Regent Prince, emphasizing Coutinho’s contribution to the royal revenue.8 These records included two lists of slaves shipped by Coutinho, which run for a time period of almost 40 years, from 1768 until 1806.9 Together, these lists provide the most extensive record available of slaves shipped by a single merchant based in Luanda at the peak of the transatlantic slave trade. More importantly, they provide a rare opportunity to trace the formation of a successful slave merchant from the beginning of his career to maturity. Anselmo da Fonseca Coutinho, son of António da Fonseca Coutinho, Knight of the Order of Christ, was born in Luanda.10 Little is known about Coutinho’s early years at this point, but as an adult he was clearly an ambitious man who sought to climb to the top of Luanda’s social ladder by accumulating titles and highranking within the military. In 1784, the Portuguese Queen D. Maria confirmed Anselmo da Fonseca Coutinho in the rank of Colonel of the Auxiliary Troops of Massangano, in the interior of Angola.11 Two years after that, the queen made him knight of her own house and granted him with a symbolic stipend of 600 réis per month.12 Coutinho was then promoted to Colonel of the Militia of Luanda. In 1799, he followed in his father’s footsteps and became Knight of the Order of Christ; the most distinctive title of status that one could aspire to in the Portuguese Empire.13 He was able to apply for this title thanks to his sister, D. Ana Maria, whom he had supported into adulthood in exchange for the remuneration and recognition of all services performed by their father for the Crown.14 Finally, between 1807 and 1810, Governor and Captain General of Angola, D. António Saldanha da Gama, promoted Coutinho to commander of the Militia of Luanda.15 THE SUPPLY OF SLAVES FROM LUANDA, 1768–1806 55

The governors of Angola had a favorable opinion of Anselmo da Fonseca Coutinho. The Baron of Moçâmedes (1784-1790), for instance, referred to him as “the most trustworthy merchant in Luanda.”16 D. Manuel de Almeida e Vasconcelos (1790-1797) called Coutinho “one of the most condign vassals of Her Majesty,”17 and D. Fernando António de Noronha (1802-1806) said that Coutinho was a “credit worthy and useful inhabitant of Angola.”18 All these comments derived from the commercial power that Coutinho held. In fact, between 1768 and 1806, Coutinho alone embarked about 5 percent of all slaves shipped from Luanda.19 As the slave trade was the principal economic activity of this port, the share of slaves shipped by Coutinho impressed the highest authorities in Angola. Coutinho operated in a critical moment of the transatlantic slave trade from West Central Africa. Economic growth and industrialization in Europe, particularly in Britain, increased the demand for primary products from the Americas, thereby expanding the volume of slaves carried from Africa.20 Figure 1 compares the annual estimated number of slaves shipped from Luanda with the annual estimated number of slaves embarked from West Central Africa. Since the mid-eighteenth century between 25 and 30 thousand slaves were shipped annually from West Central Africa. Luanda had a consistent share of the total number of slaves departing from this region. However, in the late 1770s, the slave trade from West Central Africa declined sharply. War in Europe and the Americas mobilized part of the slaving fleet of several nations, notably the French and the English. In the 1780s, the slave trade from West Central Africa recovered but at higher rates than before. Luanda experienced a slight increase in the number of slaves shipped, but was unable to follow the general trend. The slave trade from Luanda recovered in part only in the beginning of the nineteenth century.

56 DANIEL B. DOMINGUES DA SILVA

Figure 1: Luanda in the Transatlantic Slave Trade from West Central Africa, 1768–1806

50

45

40

35

30

25

20

15

10 Number of slaves exportedthousands)slaves(in of Number 5

0 1768 1770 1772 1774 1776 1778 1780 1782 1784 1786 1788 1790 1792 1794 1796 1798 1800 1802 1804 1806 Years

WC Africa Portuguese vessels Luanda Source: Eltis et al., Voyages, www.slavevoyages.org (accessed on June 2009) and José C. Curto, “Quantitative Reassessment,” 20-25.

The competition for slaves on the coast undermined the activities of Luanda merchants. In the eighteenth century, British and French traders increased their participation in the slave trade from West Central Africa. In general, they offered to Africans commodities such as gunpowder and textiles at cheaper prices than their Portuguese competitors. British and French vessels exchanged most of these commodities for slaves in ports situated north of Luanda, outside Portuguese rule, such as Cabinda, Loango and Molembo.21 Figure 2 locates these ports on the coast of West Central Africa. Africans tended to divert the supply of slaves from the interior to the northern ports, threatening the commercial activities of merchants based in Luanda. Portugal prohibited foreign vessels from trading with any of its colonial ports. As a consequence, Luanda merchants had to purchase slaves in the interior with commodities imported mostly from India, Portugal and Brazil.22 In addition to the competition from the northern ports, Luanda also faced competition THE SUPPLY OF SLAVES FROM LUANDA, 1768–1806 57 from the Portuguese port of Benguela, south of West Central Africa. Benguela had emerged as a major port in the transatlantic slave trade at the beginning of the eighteenth century, attracting part of the investments previously made in Luanda from Brazil.23 Hence, the competition for slaves on the coast of West Central Africa made the slave trade a difficult business for merchants based in Luanda.

Figure 2: Ports of Slave Embarkation on the Coast of West Central Africa, c.1750–c.1800

© Copyright Daniel B. Domingues da Silva

58 DANIEL B. DOMINGUES DA SILVA

Regional competition reduced the ability of new individuals in Luanda to break into the slave trade. The coastal supply of slaves required capital as well as access to the sources of slaves and the commercial routes connecting the interior to the coast of Africa. As competition grew, access to these resources became increasingly difficult for the merchants of Luanda. The records of Anselmo da Fonseca Coutinho suggest that the career of a slave merchant in eighteenth century Luanda required many years of experience to build capital and access to the internal sources of slaves. Table 1 shows the number of slaves shipped by Coutinho over a period of approximately 40 years, drawn from the two lists of slave shipments copied from the customs books of Luanda and attached to his records. It indicates that he took almost half of this period to become a major slave merchant, shipping slaves only occasionally during the first 12 years of activity. Coutinho became a major slave merchant after 1785, when he was able to load a couple of vessels per year with slaves, considering 350 slaves as the average carrying capacity of each vessel.24 The first list of shipments lacks the dates of embarkation, but both of them followed the books’ order chronologically, allowing distribution of Coutinho’s shipments according to the opening and closing dates of each book. Although the first book started on 9 May 1767, the first list begins in the following year, presumably when Coutinho shipped his first slaves, and ends on 31 May 1796. The second list continues from this date until 10 March 1806. Together, the lists confirm the years of experience required to participate in the slave trade during an intense period of competition. The records of Anselmo da Fonseca Coutinho also shed light on the commercial strategies that Luanda merchants employed in the transatlantic slave trade. The shipment of slaves across the Atlantic was generally conducted by merchants situated in Europe or the Americas. In the Portuguese slave trade, the majority of the vessels began their voyages from the ports of Brazil, although a few of them also departed from Portugal.25 In any case, lack of control over the shipment of slaves across the Atlantic deeply affected the interests of Luanda merchants, because they held ownership over the slaves until the final point of sale in the Americas.26 As a result, Luanda merchants were more vulnerable to slave casualties during the THE SUPPLY OF SLAVES FROM LUANDA, 1768–1806 59

Middle Passage. However, they devised ways to reduce casualties at sea. Merchants new to the business tended to trade slaves only occasionally. Additionally, they tended to ship the majority of their slaves in fast sails, such as curvetas, sumacas and bergantins, in the expectation that a fast trip across the ocean would result in fewer casualties. Indeed, studies of mortality at sea demonstrated that a successful slaving venture depended largely on the length of the voyage.27 Table 2 provides the percentages of slaves shipped by Anselmo da Fonseca Coutinho distributed by rig of vessels and according to the periods covered by the customs books of Luanda. It shows that in the years before 1785, Coutinho shipped the majority of his slaves in curvetas and sumacas. The remaining slaves he shipped in other fast sails, such as bergantins and iates, or in larger vessels like the galeras. As the merchants succeeded in the trade, they tended to distribute their shipment of slaves across different rigs. Table 2 shows that between 1785 and 1796 Coutinho shipped slaves in a variety of sails, including larger and slower vessels, such as naus and galeras. In the following years, he increased the share of slaves shipped in larger vessels. Since these vessels came mostly from Portugal, the shipment of slaves in naus and galeras suggests that, as Coutinho gained experience in the trade, he attempted to negotiate with merchants based in Portugal. Although any of the documents attached to Coutinho’s records provide information on his financial activities, they also indicate that he conducted business with merchants in Brazil as well as Portugal. All government authorities of Luanda emphasized Coutinho’s access to credit and resources in both places. In 1797, Governor and Captain General of Angola D. Manuel de Almeida Vasconcelos reported that Coutinho held the largest credit in Luanda, trading with ports in Portugal as well as Brazil.28 In 1799, the Crown Justice Official Dr. João Álvares de Melo expressed a similar opinion, noting that Coutinho held ownership over his own vessels, which often sailed to Lisbon and the Americas.29

60 DANIEL B. DOMINGUES DA SILVA

Table 1: Number of Slaves Shipped by Coutinho, 1768–1806

Number Number of Number of Lists Periods of years slaves slaves per year

1 January 1768 to 12.0 478 40 11 January 1780

12 January 1 1780 to 5.1 733 145 4 February 1785

5 February 1785 to 11.3 7,933 701 31 May 1796

31 May 1796 to 2.4 1,865 787 16 October 1798 2 17 October 1798 to 7.4 5,833 789 10 March 1806

Total 38.2 16,842 441 Source: Certidão de António José Manzoni de Castro, 31 May 1796, enclosed in the Petição de Anselmo da Fonseca Coutinho, s.d., AHU, CU, Angola, box 115 doc. 45 and Certidão de António Martiniano José da Silva e Sousa, 10 March 1806, enclosed in the Petição de Anselmo da Fonseca Coutinho, s.d., AHU, CU, Angola, box 115 doc. 45. THE SUPPLY OF SLAVES FROM LUANDA, 1768–1806 61

Table 2: Coutinho’s Shipment of Slaves by Rig of Vessels (Percent), 1768–1806

Approximate years of shipment Rig of 1768– 1780– 1785– 1796– 1798– Vessels 1780 1785 1796 1798 1806 Bergantim 1.8 – 11.0 20.0 41.8 Curveta 75.0 84.1 63.00 53.3 31.3 Fragata – – – 6.7 1.5 Galera 1.8 4.5 9.6 20.0 22.4 Iate 3.6 – – – – Nau – – 1.4 – – Paquete – – 1.4 – – Sumaca 17.8 11.4 13.6 – 3.0 Total 100 100 100 100 100 Source: Certidão de António José Manzoni de Castro, 31 May 1796, enclosed in the Petição de Anselmo da Fonseca Coutinho, s.d., AHU, CU, Angola, box 115 doc. 45 and Certidão de António Martiniano José da Silva e Sousa, 10 March 1806, enclosed in the Petição de Anselmo da Fonseca Coutinho, s.d., AHU, CU, Angola, box 115 doc. 45.

Indeed, vessel ownership was one of the first steps for Luanda merchants to succeed in the slave trade. This not only provided them with more control over the trade, but also allowed them to expand their activities into the shipping business. Additionally, it provided merchants with the opportunity to participate in governmental activities.30 Coutinho’s records provide insufficient information on the vessels that he owned. In 1797, D. Manuel de Almeida Vasconcelos said that Coutinho owned two vessels,31 but in the previous year Dr. João Álvares de Melo said that Coutinho owned “several vessels.”32 The records, however, indicate the names of three vessels only, all fast sails, the Curveta Rainha dos Anjos, the Bergantim Flor do Mar and the Sumaca Santo António e Almas.33 Coutinho used to employ the first two in the slave trade, though in 1799 he lent the second one for an expedition to Benguela organized by Governor and Captain General of Angola D. Miguel António de Melo.34 Coutinho always seemed to have reserved the Sumaca Santo

62 DANIEL B. DOMINGUES DA SILVA

António e Almas for these occasions but, at some point in the government of D. Miguel António de Melo (1797-1802), French pirates captured the Curveta Rainha dos Anjos en route to Rio de Janeiro, so Coutinho asked the governor to release his sumaca from the royal service so he could use it in the slave trade.35 All in all, ownership of vessels gave slave merchants in Luanda more access to both power and the slave trade. Another way of reducing the costs of loss in transit was to ship slaves in smaller numbers per vessel. Luanda was a busy port in the slave trade, with a large number of vessels arriving and leaving annually. The frequency of vessels calling at the port allowed merchants to make several small shipments of captives in different vessels. This strategy reduced the risks of losing an entire cargo of slaves at sea, in the event that a vessel was captured, destroyed or sank. It also allowed merchants in Luanda to dispose quickly of a highly vulnerable “commodity.” Slaves often arrived at the coast from the march from the interior exhausted, undernourished and, as a consequence, susceptible to diseases. The slaves’ condition tended to deteriorate as merchants accumulated too many captives on their properties, increasing the risks of death among slaves as a result of famine and contagious diseases.36 Hence, merchants sought to sell their slaves as soon as possible. Table 3 shows the size of Coutinho’s shipments between 1768 and 1806. It indicates that merchants shipped their slaves typically in small numbers per vessel. Coutinho rarely shipped a full cargo at any one time. Table 3 shows that the majority of his shipments consisted of no more than 50 slaves per vessel, but in fact almost half of all his shipments, 48 percent, ranged between 1 and 10 slaves only. Coutinho’s records, therefore, confirm the merchant practices of reducing the costs of loss in transit by shipping small numbers of slaves per vessel. A final way of reducing risks in a slave venture was to ship slaves in partnership with other merchants. This was a practice common among merchants in Europe and the Americas, but it was also practiced in Africa.37 Partnerships allowed merchants to invest less capital in a single voyage than if they were shipping slaves on their own account. Table 4 shows the structure of ownership in Coutinho’s shipment of slaves between 1768 and 1806. It indicates that almost 60 percent of Coutinho’s shipments were made in THE SUPPLY OF SLAVES FROM LUANDA, 1768–1806 63 partnership with other merchants. Additionally, he shipped about 30 percent of the slaves by himself and almost 10 percent on behalf of others. Merchants in Luanda rarely shipped slaves on their own account, let alone on behalf of others. Thus, the confidence that other merchants had in Coutinho’s ability to attend to their demands clearly indicates that he enjoyed a prominent status among the merchant community of Luanda.

Table 3: Size of Coutinho’s Slave Shipments, 1768–1806

Size of shipments Number of vessels

1–50 207 51–100 10 101–150 2 151–200 7 201–250 5 251–300 4 301–350 3 351 over 17 Total 255 Source: Certidão de António José Manzoni de Castro, 31 May 1796, enclosed in the Petição de Anselmo da Fonseca Coutinho, s.d., AHU, CU, Angola, box 115 doc. 45 and Certidão de António Martiniano José da Silva e Sousa, 10 March 1806, enclosed in the Petição de Anselmo da Fonseca Coutinho, s.d., AHU, CU, Angola, box 115 doc. 45.

However, it required time for a slave merchant to achieve prominence in Luanda. The slave trade was an activity that required large investments. These investments came from markets situated overseas in the form of commodities used to barter for slaves. African slave traders demanded these commodities in advance, so merchants in Luanda acted as brokers, sending commodities to the interior in exchange for future payments from the sale of slaves in the Americas.38 As in many other ports of the Portuguese Empire, Luanda suffered from a scarcity of metal currency circulating in the

64 DANIEL B. DOMINGUES DA SILVA market.39 The fact that a few individuals were able to ship slaves on behalf of others indicates that they had a reputation of being successful merchants so others would feel confident enough to trust them with commodities in advance to purchase slaves. Table 4 shows that, at the beginning of his career, between 1768 and continuing through 1785, Coutinho owned the majority of slaves he embarked. This long apprenticeship in the slave trade surely provided him with sufficient experience and access to resources to attract the partnership of other merchants. As he succeeded in his career, the investment of partners in his commercial activities increased to include shipments on behalf of others. Since slaves were regarded as vulnerable “commodities,” thus involving high risks, well established merchants in Luanda preferred to control the logistics of the business instead of holding ownership over the slaves embarked. Mastery of the business logistics provided merchants with a privileged commercial position, which allowed them to rise in prominence within the merchant community of Luanda.

Table 4: Structure of Ownership in Coutinho’s Shipment of Slaves (Percent), 1768–1806

On Approximate On In behalf List years of Coutinho’s partnership of shipment account others

1 1768–1780 90.4 9.6 – 1780–1785 78.2 21.8 – 1785–1796 33.8 66.2 –

2 1796–1798 16.5 43.1 40.4 1798–1806 24.0 61.3 14.8

Total 32.2 58.4 9.4 Source: Certidão de António José Manzoni de Castro, 31 May 1796, enclosed in the Petição de Anselmo da Fonseca Coutinho, s.d., AHU, CU, Angola, box 115 doc. 45 and Certidão de António Martiniano José da Silva e Sousa, 10 March 1806, enclosed in the Petição de Anselmo da Fonseca Coutinho, s.d., AHU, CU, Angola, box 115 doc. 45. THE SUPPLY OF SLAVES FROM LUANDA, 1768–1806 65

Unfortunately, Coutinho’s records provide little information on his trading partners, with exception of the names of some of them, which appear in his lists of shipments. In the early period of his career, Coutinho was the partner of two other merchants, both presumably residents in Luanda, Manoel António Magalhães and José António Pereira. In the later period of his career, Coutinho built partnerships with several individuals. Some of them were crew members of his Curveta Rainha dos Anjos or of vessels belonging to other merchants, like the Galera NS da Conceição S Bento Bela Africana; Galera Bom Jesus; Curveta NS da Conceição e Almas and the Bergantim Africano. Some other individuals were in all likelihood residents in Luanda, including Álvaro de Carvalho Matoso or Sergeant Major Álvaro Matoso de Carvalho; D. Ana Maria de São Miguel; António da Fonseca Coutinho; António de Azevedo Galiano; António José Rodrigues; António Manuel de Melo; Bento António Moreira; Father Boaventura José de Melo; Caetano Soares; Domingos Antunes Guimarães; Estevão da Fonseca Negrão; Francisco de Benício Carvalho; Jacinto Lopes; João de Araújo de Barros; Father João Luís Fortunato; João Teixeira; Joaquim da Fonseca Negrão; Joaquim de Matos; Joaquim Vieira de Abreu; José da Costa Lisboa; José Joaquim Jorge; José Pinheiro Salgado; José Teodósio Vilaça; Colonel José Vaz Salgado; Manuel da Rosa; Manuel Luís Gonçalves; Silvestre José de Seixas; Tomás Corrêa Porto. Finally, in one of Coutinho’s shipments, he appeared as partner of the famous institution present in any major city of the Portuguese Empire: the Santa Casa da Misericórdia.40 Although Coutinho’s records lack information on his partners, they provide more complete information on his slave markets in the Americas. The first list of shipments attached to his records clearly indicates the intended port of slave disembarkation for each shipment that he made. This information is lacking for the second list, but the destination of the slaves can be retrieved with the help of Voyages: The Trans-Atlantic Slave Trade Database.41 The second list provides the dates of each shipment as well as the captain and vessel names. This information was crosschecked with the shipping records

66 DANIEL B. DOMINGUES DA SILVA available in the Voyages Database to determine the destination of the slaves contained in the second list. Fortunately, the Voyages Database provides information on the ports of disembarkation for all the vessels available in the second list. Table 5 shows the destination of slaves according to each list for the years between 1768 and 1806. It shows that Coutinho shipped slaves to Luanda’s principal slave markets in the Americas, shown in Figure 3. In the second half of the eighteenth century, the decline in gold exports from Brazil resulted in a revival of agricultural production. This revival was further stimulated by economic growth and industrialization in Europe, which increased the demand for products from the Americas. The slave markets in the Brazils responded positively to this impulse, increasing the demand for slaves from Africa. Although Luanda experienced a difficult moment in the slave trade, it is no surprise that Coutinho shipped most of his slaves to Rio de Janeiro. Rio de Janeiro had long been the principal slave market of Luanda.42 In the second list, the number of slaves shipped from Luanda to this port declined significantly. Two reasons explain this decline. First, the revival of the sugar industry in Bahia increased the demand for African slaves, offering new commercial opportunities for merchants based in Luanda.43 Second, Montevideo began trading slaves directly with Africa. Montevideo was a port dominated by the Spanish, where contraband slaves were traded in Rio de Janeiro, in Brazil. However, in the last decade of the eighteenth century, the demand for slaves in Montevideo increased, leading local merchants to break into the transatlantic slave trade and explore connections their partners in Rio de Janeiro had with Luanda. This activity was no less illegal, but merchants in Luanda were willing to take the risks because traders from Montevideo paid in money made of silver extracted from the mines of the Spanish Americas.44 The second list of shipments available in Coutinho’s records does not indicate the intended destination of the slaves, but all the ships listed are also included in the Voyages Database, which shows their port of disembarkation in the Americas. Since the list was copied directly from the official customs records, the information available in the list and the database allows measuring the extent of the illegal activities of merchants based in Luanda. Between 1796 and 1806 Coutinho was at the peak of his career and, although Montevideo traders THE SUPPLY OF SLAVES FROM LUANDA, 1768–1806 67 offered ready access to metal currency, Coutinho shipped only 7 percent of his slaves to this destination. Thus, at the peak of the transatlantic slave trade, well established merchants in Luanda avoided shipping slaves illegally to the Americas. In contrast to Rio de Janeiro, Coutinho seems to have reserved some of his slaves to sell exclusively in the markets of Pernambuco and Maranhão. Both held the same percentage of slaves shipped in the two lists. In the second half of the eighteenth century, two trading companies based in Portugal held the monopoly over the slave trade to Pernambuco, Maranhão and Pará. The first, founded in 1755, was the Companhia Geral do Comércio do Grão Pará e Maranhão. The second, founded in 1759, was the Companhia Geral do Comércio de Pernambuco e Paraíba. These two companies were part of a project formulated in Lisbon to stimulate the production of cash crops such as rice, cotton and sugar in remote areas of Brazil.45 They were also responsible for providing labor to these areas through the transatlantic slave trade. The companies thus had a slaving fleet, which enjoyed reduced freight charges to ship slaves from Africa. Additionally, slaves carried in the vessels of the Companhia Geral do Comércio do Grão Pará e Maranhão to the captaincies of Maranhão and Pará were disembarked tax free. Hence, the establishment of these companies provided Luanda merchants with regular access to major slave markets, such as Pernambuco, as well as an opportunity to ship slaves to markets previously outside of their domain, such as Maranhão and Pará. In the second half of the 1770s the companies’ monopoly ended, but they continued to trade slaves to Pernambuco, Maranhão and Pará until the end of the 1780s, competing against other slave traders.46 Finally, as these companies were financially based in Lisbon, they provided slave merchants in Luanda with an opportunity to do business with merchants based in Portugal. As previously noted, government authorities in Luanda often mentioned Coutinho’s access to credit and business in Portugal, which may have been built with his shipment of slaves to Pernambuco, Maranhão and Pará.

68 DANIEL B. DOMINGUES DA SILVA

Table 5: Destination of the Slaves Shipped by Coutinho, 1768– 1806

List Number Percentage Destination of slaves (Period) of slaves of slaves

Intended port of

disembarkation (Coutinho’s records) Rio de Janeiro 4,762 52 Pernambuco 2,682 29 1 Maranhão 1,321 14 (1768– Bahia 375 4 1796) Pará 4 - Total 9,144 100

Principal port of disembarkation (Voyages Database) Rio de Janeiro 2,536 33 Pernambuco 2,252 29 2 Bahia 1,240 16 (1796– Maranhão 1,046 14 1806) Montevideo 562 7 Pará 41 1 Total 7,677 100 Source: Certidão de António José Manzoni de Castro, 31 May 1796, enclosed in the Petição de Anselmo da Fonseca Coutinho, s.d., AHU, CU, Angola, box 115 doc. 45 and Certidão de António Martiniano José da Silva e Sousa, 10 March 1806, enclosed in the Petição de Anselmo da Fonseca Coutinho, s.d., AHU, CU, Angola, box 115 doc. 45. THE SUPPLY OF SLAVES FROM LUANDA, 1768–1806 69

Figure 3: Ports of Slave Disembarkation in Brazil, c.1750–c.1800

© Copyright Daniel B. Domingues da Silva

The records of Anselmo da Fonseca Coutinho also reflect the age profile of the slaves shipped from Luanda. In the second half of the eighteenth century, the Portuguese taxed slaves shipped from Luanda according to three age categories. Adult slaves, termed cabeça or “head,” were taxed regardless of their sex. Standing children, crias de pé, were taxed half the adult levy, and children at the breast, crias de peito, were shipped tax free together with their

70 DANIEL B. DOMINGUES DA SILVA mothers. Although these categories were defined as age categories, criteria separating children from adults were based on the height of the slaves. Slaves measuring over four palmos tall, approximately 88 centimeters or 3 feet tall, were considered adults.47 Slaves measuring this height or less were considered children, and were further divided into children who could stand and children who were still suckling. The Portuguese created these categories through the law of 25 January 1758, which regulated the tax collection over the slaves shipped from the Portuguese ports in West Central Africa. The law became in effect 5 January 1760 and because it focused on tax collection, only late in the eighteenth century customs officials began reporting the number of crias de peito shipped from Luanda.48 Furthermore, it should be noted that in the eighteenth century the Portuguese, as did most Europeans, accepted that a person achieved adulthood at the age of puberty, when they were eligible for marriage.49 Therefore, the total number of slaves shipped from Luanda tended to be lower than the number of slaves recorded and it is likely that the total number of children shipped was higher than the reported. Coutinho’s records allow an assessment of the age profile of slaves shipped from Luanda. Table 6 shows the number of slaves that he shipped was distributed across the three principal age categories. The first list of shipments contains only the number of adults and crias de pé, but the second list provides the number of slaves shipped for all three categories. Table 6 shows that Coutinho tried to address the preferences for slaves of slave markets in the Americas, shipping more adult slaves than children. Slaveholders in the Americas demanded slaves who were ready to work, hoping to obtain a fast return from their investments in labor power. They did not wish to shoulder the costs of raising slave children until their productive age. As a consequence, slaveholders in the Americas demanded more adult slaves than children from Africa. Thus, as slave merchants had ownership over the slaves until the final point of sale in the Americas, they tended to avoid selling children into the transatlantic slave trade.

THE SUPPLY OF SLAVES FROM LUANDA, 1768–1806 71

Table 6: Age Profile of the Slaves Shipped by Coutinho, 1768– 1806

Approximate Number of Number of Number of List years of adults crias de crias de shipment pé peito 1768–1780 477 1 – 1 1780–1785 731 2 – 1785–1796 7,925 8 –

1796–1798 1,805 4 56 2 1798–1806 5,644 7 182

Total 16,582 22 238

Source: Certidão de António José Manzoni de Castro, 31 May 1796, enclosed in the Petição de Anselmo da Fonseca Coutinho, s.d., AHU, CU, Angola, box 115 doc. 45 and Certidão de António Martiniano José da Silva e Sousa, 10 March 1806, enclosed in the Petição de Anselmo da Fonseca Coutinho, s.d., AHU, CU, Angola, box 115 doc. 45.

Although children made up a small fraction of the total number of slaves shipped, there were remarkable differences between the number of crias de pé and crias de peito embarked. Coutinho’s records indicate that he shipped over ten times crias de peito in relation to crias de pé, but this difference only confirms the preferences of slave markets in the Americas. According to the law, crias de peito were shipped tax free along with their mothers, who were taxed at the same level as any other adult slave. Unfortunately, the records do not distinguish the sex of the slaves. However, research on the transatlantic slave trade suggests that, in the eighteenth century, two male slaves were shipped for every female.50 If this ratio is correct then, at the peak period of his career, between 1796 and 1806, Coutinho shipped at least 2,483 slave women, and if all children at the breast were indeed shipped with their mothers, this suggests that about 9 percent of these slave women were also

72 DANIEL B. DOMINGUES DA SILVA mothers. Crias de peito entered the slave trade not because slaveholders in the Americas demanded young slaves, but because they contribute to a percentage of the number of slave women sold into the transatlantic slave trade. All in all, the records of Anselmo da Fonseca Coutinho show that the merchants of Luanda strove to participate in the transatlantic slave trade. The competition for slaves on the coast of West Central Africa increased during the eighteenth century, making it difficult for merchants based in Luanda to secure access to the sources of slaves in the interior of Africa and address the demand for labor in the Americas. The Portuguese restrictions on foreign imports forced the merchants of Luanda to compete on unequal terms with merchants based in other ports of West Central Africa. As a consequence, merchants had to accumulate a great deal of experience in the slave trade before rising to prominence among the merchant community of Luanda. The slave trade was a commercial activity demanding huge investments and involving high risks. Merchants had to master the logistics of the business, build commercial strategies and please the local authorities if they wanted to prosper in the trade. All these strategies required time to develop and they did not guarantee a successful career, as merchants remained largely dependent on the market fluctuations of the transatlantic slave trade.

1 Joseph C. Miller, “Some Aspects of the Commercial Organization of Slaving at Luanda, Angola - 1760-1830,” in Henry Gemery and Jan Hogendorn, eds., The Uncommon Market: Essays in the Economic History of the Atlantic Slave Trade (New York: Academic Press, 1979), 92. 2 David Birmingham, Trade and Conflict in Angola: The Mbundu and their Neighbours under the Influence of the Portuguese, 1483-1790 (Oxford: Clarendon Press, 1966), 137-141; Joseph C. Miller, “The Slave Trade in Congo and Angola,” in The African Diaspora: Interpretative Essays (Cambridge, MA: Harvard University Press, 1976), 96-97; Miller, Way of Death: Merchant Capitalism and the Angolan Slave Trade, 1730-1830 (Madison: University of Wisconsin Press, 1988), 226-27; Roquinaldo Amaral Ferreira, “Transforming Atlantic Slaving: Trade, Warfare and Territorial Control in Angola, 1650-1800” (Ph.D. thesis, University of California at Los Angeles,, 2003), 17, 69, 83-88. 3 Birmingham, Trade and Conflict, 137-41; Miller, “Slave Trade,” 97-98; Miller, Way of Death, 226-27; Ferreira, “Transforming Atlantic Slaving,” 70-71; Mariana P. Candido, “Enslaving Frontiers: Slavery, Trade and Identity in Benguela, 1780-1850” (Ph.D. thesis, York University, 2006), 22- 25. THE SUPPLY OF SLAVES FROM LUANDA, 1768–1806 73

4 Herbert S. Klein, “African Women in the Atlantic Slave Trade,” in Claire C. Robertson and Martin A. Klein, eds., Women and Slavery in Africa (Madison: University of Wisconsin Press, 1983), 29-38; Klein, The Middle Passage: Comparative Studies in the Atlantic Slave Trade (Princeton: Princeton University Press, 1978), 38-41. 5 Joseph C. Miller, “Capitalism and Slaving: the Financial and Commercial Organization of the Angolan Slave Trade, According to the Accounts of António Coelho Guerreiro (1684-1692),” International Journal of African Historical Studies 17, no. 1 (1984), 4-10. 6 W.G. Clarence-Smith, The Third Portuguese Empire, 1825-1975: A Study in Economic Imperialism (Manchester: Manchester University Press, 1985), 30-34, 38-56. 7 Roquinaldo Amaral Ferreira, “Dos Sertões ao Atlântico: Tráfico Ilegal de Escravos e Comércio Lícito em Angola, 1830-1860” (M.A. thesis, Universidade Federal do Rio de Janeiro, 1996), 118-49; Ferreira, “Transforming Atlantic Slaving,” 126-43; Candido, “Enslaving Frontiers,” 101-18; Cândido, “Merchants and the Business of the Slave Trade in Benguela c. 1750-1850,” African Economic History 35 (2007), 1-30 8 D. Fernando António de Noronha to the Regent Prince, 31 March 1806, Arquivo Histórico Ultramarino, Conselho Ultramarino (hereafter, AHU, CU), Angola, box 115 document 45. 9 Certidão de António José Manzoni de Castro, 31 May 1796, enclosed in the Petição de Anselmo da Fonseca Coutinho, s.d., AHU, CU, Angola, box 115 doc. 45 and Certidão de António Martiniano José da Silva e Sousa, 10 March 1806, enclosed in the Petição de Anselmo da Fonseca Coutinho, s.d., AHU, CU, Angola, box 115 doc. 45. 10 Atestação de D. Manuel de Almeida e Vasconcelos, 7 February 1793, AHU, CU, Angola, box 115 doc. 45. Instrumento em pública forma sobrescrito por Felipe Benício e Rosa Mascarenhas, 08 June 1795, AHU, CU, Angola, box 115 doc. 45. 11 Arquivo Nacional Torre do Tombo (hereafter, ANTT), Registo Geral de Mercês, D. Maria I, Book 16, f. 126. 12 Idem. 13 ANTT, Registo Geral de Mercês, D. Maria I, Book 29, ff. 224v and 243. 14 Instrumento em pública forma sobrescrito por Felipe Benício e Rosa Mascarenhas, 08 June 1795, AHU, CU, Angola, box 115 doc. 45. 15 João de Oliveira Barbosa to Conde das Galvêas, 2 December 1810, AHU, CU, Angola, box 121 A doc. 31. 16 Atestação do Barão de Moçâmedes, 6 October 1790, AHU, CU, Angola, box 115 doc. 45. 17 Atestação de D. Manuel de Almeida Vasconcelos, 2 January 1796, AHU, CU, Angola, box 115 doc. 45. 18 Atestação de D. Fernando António de Noronha, 27 February 1806, AHU, CU, Angola, box 115 doc. 45.

74 DANIEL B. DOMINGUES DA SILVA

19 Calculated with data available in the lists of shipments and in José C. Curto, “A Quantitative Reassessment of the Legal Portuguese Slave Trade from Luanda, Angola, 1710-1830,” African Economic History 20 (1992), 20-25. Note that Curto’s figures may include untaxed slaves, while part of Coutinho’s lists of shipment does not, so this percentage is likely to be higher. See the text below for further discussion on the shipment of slaves from Luanda. 20 David Eltis, Economic Growth and the Ending of the Transatlantic Slave Trade (New York: Oxford University Press, 1987), 31-33. 21 Birmingham, Trade and Conflict, 137-38; Miller, Way of Death, 73-78; Joseph C.Miller, “Imports at Luanda, Angola: 1785-1823,” in Figuring African Trade: Proceedings of the Symposium on the Quantification and Structure of the Import and Export and Long-Distance Trade of Africa in the Nineteenth Century, c.1800-1913 (St. Augustin, 3-6 January 1983) (Berlin: Dietrich Reimer Verlag, 1986), 165-66. 22 Miller, “Imports at Luanda,” 192-200 and Table VI.3. 23 Benguela participated in the transatlantic slave trade already in the seventeenth century, but until the beginning of the eighteenth century, most of the slaves shipped from Benguela were actually delivered in Luanda for sale into the Atlantic. David Birmingham states that Governor and Captain General of Angola, Rodrigo César de Menezes (1733-1738), reported that the bulk of the slaves coming to Luanda originated from the Benguelan hinterland. Roquinaldo Ferreira notes that Lisbon authorities reported in 1688 that about one third of the slaves shipped from Luanda came in fact from Benguela. Mariana Candido claims that until 1716 slaves shipped from Benguela had to be delivered in Luanda for sale into the Atlantic, because Benguela had no customs house to collect export duties. Cf. Birmingham, Trade and Conflict, 141; Ferreira, “Transforming Atlantic Slaving,” 71-80 and 112-121; Candido, “Enslaving Frontiers,” 21-22. 24 Herbert Klein has observed that the legal carrying capacity of ships loading slaves in Luanda between 1762 and 1765 averaged about 420 slaves, but they actually carried an average of 394 slaves per vessel. Klein noted later that this pattern changed little in the late eighteenth century. Cf. Herbert Klein, “The Portuguese Slave Trade from Angola in the Eighteenth Century,” Journal of Economic History 32, 4 (1972), 902-03, and Table 6; Miller, Way of Death, 30-32. 25 Daniel B. Domingues da Silva, “Africans in Transit: From the Angolan to the Brazilian Hinterland,” in Allen Morris Conference on Florida and the Atlantic World (Tallahassee: Florida State University, 2006), 5-9 and Table 2. 26 Klein, “Portuguese Slave Trade,” 906-907; Klein, Middle Passage, 38-39; Miller, Way of Death, 252-53. 27 Klein has suggested that slave mortality at sea correlated with length of voyage, but has observed that other factors could also affect the mortality rates among the slave population on board during the Middle Passage. Miller notes that slave casualties at sea tended to decline from relatively THE SUPPLY OF SLAVES FROM LUANDA, 1768–1806 75

high levels on ships making fast passages to a minimum on voyages of medium and longer duration prior to rising again toward the end of voyages of exceptionally slow progress. Cf. Klein, Middle Passage, 65, 87-89; Joseph D. Miller, “Mortality in the Atlantic Slave Trade: Statistical Evidence on Causality,” Journal of Interdisciplinary History 11, 3 (1981), 395. Further information on the different types of rigs in the Portuguese slave trade is available in Miller, Way of Death, 366-74. 28 Atestação de D. Manuel de Almeida e Vasconcelos, 28 August 1797, AHU, CU, Angola, box 115 doc. 45. 29 Atestação do Dr. Ouvidor Geral João Álvares de Melo, 10 September 1799, AHU, CU, Angola, box 115 doc. 45. 30 Miller has argued that only by 1820 a group of traders in Luanda acquired sufficient capital to purchase sailing vessels and ship slaves on their own account. Cf. Miller, “Some Aspects,” 87. 31 Atestação de D. Manuel de Almeida e Vasconcelos, 28 August 1797. 32 Atestação do Dr. Ouvidor Geral João Álvares de Melo, 10 September 1799. 33 Carta de Ofício de D. Miguel António de Melo, 9 May 1799, AHU, CU, Angola, box 115 doc. 45. Petição de Anselmo da Fonseca Coutinho a D. Miguel António de Melo, s.d., AHU, CU, Angola, box 115 doc. 45. The Voyages Database provides information on only two records of slaving voyages with Coutinho’s name as owner; one for an unnamed ship, which set sail in 1799, and the other for the Bergantim Flor do Mar, which sailed between 1810 and 1811. Cf. David Eltis, David Richardson, Manolo Florentino, and Stephen D. Behrendt, Voyages: The Trans-Atlantic Slave Trade Database (www.slavevoyages.org), voyage id 46,313 and 49,898 (Accessed in June, 2009). 34 Carta de Ofício de D. Miguel António de Melo, 9 May 1799. 35 Petição de Anselmo da Fonseca Coutinho a D. Miguel António de Melo, s.d. It appears that the unnamed ship listed in the Voyages Database with Coutinho’s name was the Bergantim Flor do Mar, because the voyage outcome in the database shows that the vessel was pressed into government service and the year of the record matches with the year that Coutinho lent the ship to Governor Melo, even though it refers to the year of arrival in the Americas. None of the records with vessel name Curveta Rainha dos Anjos, listed for the years between 1797 and 1802, shows Coutinho’s name as owner or indicates the voyage outcome as thwarted by human agency from the owner’s point of view. 36 Miller, “Some Aspects,” 98; Miller, Way of Death, 399-401, 27-36. 37 Manolo G. Florentino, Daniel B. Domingues da Silva, and Alexandre Vieira Ribeiro, “Aspectos Comparativos do Tráfico de Africanos para o Brasil (Séculos XVIII e XIX),” Afro-Ásia 31 (2004), 97-103. 38 Miller, Way of Death, 252-55, 295-311. 39 Joseph C. Miller, “Slave Prices in the Portuguese Southern Atlantic, 1600-1830,” in Paul E. Lovejoy, ed., Africans in Bondage: Studies in

76 DANIEL B. DOMINGUES DA SILVA

Slavery and the Slave Trade (Madison: African Studies Program, University of Wisconsin, 1986), 54-55. 40 The classic work on the history of the Santa Casa de Misericórdia in the Portuguese Empire is A.J.R. Russell-Wood, Fidalgos and Philanthropists: The Santa Casa da Misericórdia of Bahia, 1550-1755 (Berkeley: University of California Press, 1968). For the Santa Casa de Misericórdia in Luanda, see António Brásio, “As Misericórdias de Angola,” Studia (Centro de Estudos Históricos Ultramarinos, Lisboa) 4 (1959), 106-49; Jofre Amaral Nogueira, “A Misericórdia de Luanda através dos Séculos,” Boletim do Instituto de Angola 3 (1954). 41 Eltis et al., Voyages, at www.slavevoyages.org. 42 Corcino Medeiro dos Santos, “Relações de Angola com o Rio de Janeiro (1736-1808),” Estudos Históricos 12 (1973), 9-25. 43 Dauril Alden, “Late Colonial Brazil, 1750-1808,” in Leslie Bethell, ed., The Cambridge History of Latin America (New York: Cambridge University Press, 1985) vol. 2, 627-35. 44 Alex Borucki, “Trans-Imperial Networks and Atlantic Warfare in the Making of the Slave Trade to the Río de la Plata, 1777-1812” (Ph.D. thesis, Emory University, 2008), 5. 45 António Carreira, As Companhias Pombalinas de Navegação, Comércio e Tráfico de Escravos entre a Costa Africana e o Nordeste Brasileiro (Bissau: Centro de Estudos da Guiné Portuguesa, 1969), 31-36, 243-52. 46 Carreira, Companhias Pombalinas, 48-50, 262. 47 Maurício Goulart, Escravidão Africana no Brasil: Das Origens à Extinção do Tráfico (São Paulo: Editora Alfa-Ômega, 1975), 197. For the conversion of palmos into meters see Fortunato José Barreiros, Memória sobre os Pesos e Medidas de Portugal, Espanha, Inglaterra e França (Lisbon: Typographia da Academia Real das Sciencias, 1838), 20. 48 Klein, “Portuguese Slave Trade,” 903-906; Klein, Middle Passage, 35; Horácio Gutiérrez, “O Tráfico de Crianças Escravas para o Brasil durante o Século XVIII,” Revista de História 120 (1989), 62. 49 That implied, for the Portuguese, that males achieved adulthood at 14 years of age and female at 12 years of age. See Sebastião Monteiro da Vide, Constituições Primeiras do Arcebispado da Bahia (São Paulo: Typographia de António Louzada Antunes, 1853), 109-10, Title 64. 50 Klein, “African Women,” 32. However, it should be noted that more recent research has criticized the attention historians have been giving to this ratio. See David Eltis and Stanley L. Engerman, “Fluctuations in Sex and Age Ratios in the Transatlantic Slave Trade, 1663-1864,” Economic History Review 46, 2 (1993), 321. REEXAMINING THE GEOGRAPHY AND MERCHANTS OF THE WEST CENTRAL AFRICAN SLAVE TRADE: LOOKING BEHIND THE NUMBERS

Filipa Ribeiro da Silva Netherlands Royal Academy of Sciences and Stacey Sommerdyk University of the Witwatersrand

he Slave Voyages Database is the most significant work on the quantification of the Atlantic slave trade to have appeared Tsince Philip D. Curtin produced his census of the commerce.1 As the first resource of its kind compiled by economic historians of the transatlantic slave trade the authors should be commended for realizing a project of such enormous breadth and vision. Yet due to the innovative nature of the resource, the creators have faced multiple unforeseen issues in the categorization and dissemination of this material. Equally, in light of the overwhelming conclusiveness of the study, economic historians are left asking themselves the question of what is next. Using the West Central African coast as a case study this paper will explore two key areas of the database which, with some modifications, could begin to inform and guide new research directions. First, this article will examine the geographical categorization of West Central Africa and will demonstrate how the flexibility of geographical interpretation in West Central Africa can be problematic. The authors of this article propose a more in depth definition of the geography, paying special attention to Dutch sources on the Loango Coast. Second, this article will discuss inherent problems within the database’s categories. Ship captains and ship owners have been categorized according to an English model and are incompatible with the Portuguese and Spanish documentation. This article will conclude with suggestions for uses of the database which move beyond simple enumeration of the slave trade into questions about the formation of merchant communities, the interconnectivity

African Economic History v.38(2010):77–105

78 F. R. DA SILVA AND S. SOMMERDYK of merchant networks, as well as the process of database creation and how it can inform historians in the creation of future databases.

I: A New Geographical Categorization for West Central Africa and Its Implications

Beginning with the issue of geography, regional definition is a problem that plagues all scholars of the transatlantic slave trade. This difficulty springs from a lack of geographical standardization in European shipping documents throughout the slave trade. This is particularly pronounced in the region of West Central Africa as the Dutch, Portuguese, English and French all possessed differing definitions of the coastal region. Although some scholars including Martin, Postma, and Manning clearly define the Loango Coast as a separate trading region,2 the overwhelming trend is to approach the West Central Coast as one trading community as exhibited by the Slave Voyages Database. Despite the many benefits of having an online resource with a vast amount of information, the formatting of the African Voyages Database results in some significant limits when studying specific regions within the vast definitions of West Central Africa. The current version requires the user to have an intimate knowledge of the sources in order to use the data and assumes that they are able to identify the geographic location of the twenty-three ports listed.3 Even with that level of awareness, the degree of flexibility within interpretations is extremely high. This is exhibited by Roquinado Ferreira’s Suppression of the Slave Trade and Slave Departures from Angola, 1830s–1860s, in which he defines the entire West Central African coast as “Angola” suggesting Portuguese dominance in the entire region.4 The category of West Central Africa encompasses a number of geographical regions with conflicting definitions. The primary terms used to describe sections of this coast are Angola, Kongo, and Loango. While Kongo and Loango are fairly linear categories - Kongo being an African community situated to the south of the Congo River and Loango being an African community to the North of the Congo River— the definition of Angola’s location and size differs from one group of European traders to the next and shifts significantly over time. REEXAMINING THE GEOGRAPHY AND MERCHANTS 79

In Warfare in Atlantic Africa 1500–1800, John K. Thornton defines three “military-diplomatic regions” in Atlantic Africa: Upper Guinea, Lower Guinea and Angola. He suggests that people within these regions “had more interactions among themselves than with those lying around them.”5 Within this context, Thornton defines Angola as “the whole of west central Africa, from the coast of Gabon to Angola.”6 This definition inherently possesses a difficulty: Angola is both a region and a distinct sub-region. Having defined the region, what is the definition of the sub-region? This is a difficult question to answer as definitions can shift over time. Using Portuguese sources, Thornton suggests that the initial Portuguese colony of Angola was formed in the 1580s. The Portuguese gained a foothold by inserting themselves in a war between the Kongo and the Ndongo peoples of this area. Thus, at the dawn of the seventeenth century, Angola was a small colony centered in Luanda and hemmed in by Kongo to the north and Ndongo to the east. Throughout the seventeenth century the Portuguese, using Imbagala allies, pressed steadily eastward gaining land and slaves from the Ndongo.7 The eighteenth century was characterized by less intense fighting, minimal expansion of the Angola colony, and frequent raids to supply the slave trade. The Portuguese colony exerted nominal power over its subjects in the vicinity of the colony who engaged in frequent “low key wars.”8 In Kingdoms of the Savanna, Jan Vansina aptly reminds us: “The conquest of Angola is something that lasted for centuries.”9 Using current maps of Angola, which reach from the Congo River in the north (including the hotly contested region of Cabinda to the north of the Congo River) to the Kalahari Desert in the south and reach roughly a third of the way to the eastern African coast and include Cabinda, it is almost impossible to project backwards and imagine the Angola of the transatlantic slave trade era. In addition to the problem presented by fluctuating definitions is the problem of variation within the European definitions themselves. The Dutch present a contrasting definition of Angola altogether. Postma addresses this issue in 1990 in The Dutch in the Atlantic Slave Trade, 1600–1815: “The northernmost portion of [the west central African Coast], between Cameroon and the Congo River, was generally referred to by the Dutch as Angola, although it has also been referred to as Loango after the seventeenth century African state 80 F. R. DA SILVA AND S. SOMMERDYK in that region. This mixing of historical and contemporary labels can be confusing.” 10 To avoid this confusion Postma chooses to use the term Loango-Angola in reference to the region north of the Congo. Postma goes on to explain:

Contemporary Dutch documents nearly always referred to [West Central Africa] as Angola, although after 1649 they really meant the area north of the Congo River. This area would be more appropriately referred to as Loango, after the dominant state of the region during the seventeenth and early eighteenth centuries.11

While Postma’s demarcation of the Dutch sphere of influence in West Central Africa is useful, Phyllis Martin’s terminology, which locates the Loango coast to the north of the Congo River and the Angola coast to the south of the Congo River, is the specific language favored by this article.12 Moving from the general to the specific, another example of confusing geography on the West Central African coast is found in the category “Congo North.” Masquerading under the title of “Port,” this first garnered attention as 88 percent of enslaved Africans from this “port” embarked on Dutch ships. This is by far the category with the highest Dutch involvement on the Loango Coast (as shown below in Table 1). Further research showed that Postma collected the data for 176 of the 238 voyages in this category. This raises several important questions: What was Postma’s original purpose in creating this category? Do the original sources contain more detailed information? And perhaps most importantly, how did this key information get lost in between the databases? This first question can be easily answered by looking at Postma’s 1990 analysis of his Dutch data. Here he explained his interpretation of the term “Angola” as meaning north of the Congo River or the Loango Coast (as explained above).13 However, the organizing principle of the database focuses on the slaving hinterland rather than on how Europeans defined the coast, or indeed how trade was conducted in various regions. Even so, Martin offers compelling evidence that, in addition to the overlapping slave supply from the Luanda hinterland via the Congo River, traders of the Loango Coast purchased numerous slaves from REEXAMINING THE GEOGRAPHY AND MERCHANTS 81 the same northern and eastern traders from whom they sought ivory and copper.14 Thus, even if slaving hinterlands are the criteria for separating the coast, the Loango Coast still stands out as a distinct region. This raises another issue that can be addressed in answering the last two questions. It was only upon searching the codes section of the first Transatlantic Slave Trade Database on CDROM that the location became clear: “Congo North (no dominant location, [Cape] Lopez to Congo).”15 This implies that this region, otherwise referred to as the Loango Coast, is a separate category from the Angolan coast. It is unclear why the authors chose to leave this as a distinct category rather than including it in the category of West Central Africa if the region is not of distinct significance. Along the 1200 km stretch of coast referred to as West Central Africa, Europeans traded at 19 known locations for the purchase of slaves. From north to south, these are the locations recorded by European traders from which slaves were embarked: Mayumba, Kiloango, Loango [including Boary], Malembo, Cabinda, Congo North, Congo River, Rio Zaire [Congo River], Mpinda, Ambriz, Rio Dande [Dande River], St. Paul de Loanda [Luanda], Salinas, Coanza River [Kwanza River], Ambona, Benguela Velho [Old Benguela], Nova Redonda, Quicombo [Kikombo], and Benguela. Table 1, below, provides significantly more detail on the boundary of the Loango and Angola coasts. Using the Slave Voyages Database to establish Mpinda as the northern most port, which traded a majority of its slaves to the Portuguese, the West Central African coast is split into two regions. This allows researchers to calculate the number of exports from the Loango coast alone. Of the 2.7 million enslaved Africans embarked from located ports on the West Central African coast, 870,000 slaves were embarked at the Loango coast and 1.8 million slaves were embarked at the Angola Coast (see Table 1). Scholars with more expertise in Portuguese Angola may even suggest a further separation of Benguela from Luanda, though this reaches beyond the boundaries of this article.16 Having extrapolated these numbers from the actual data, it is possible to further speculate that of the 3.2 million slaves who embarked from this coast, one third embarked from the Loango coast and two thirds from the Angolan coast. Even with this separation, the Angola coast remained the most significant supplier of slaves to the Atlantic slave trade, while the 82 F. R. DA SILVA AND S. SOMMERDYK

Loango coast emerged as a separate category which remained among the top four exporters of enslaved Africans exporting comparable numbers of enslaved Africans to the Bights of Biafra and Benin (see Table 2). Through observing patterns of European trade on the West Central African Coast, this division between north and south becomes more distinct (see Graphs 1 & 2 and Table 1). Although the Slave Voyages Database clearly indicates Portuguese dominance on the West Central African coast with the Portuguese trading 67 percent of the 3.2 million enslaved Africans,17 this information is misleading and gives the impression of Portuguese domination of the entire West Central African coast. By separating the ports of the Loango coast a very different profile of traders emerges. Using the known data, Table 1 clearly demonstrated the separation of the Loango and Angola coast (above). By focusing on the Loango coast data, Graph 1 illustrates that traders using the Portuguese flag had only nominal engagement in the slave trade on the Loango coast before 1810.18 Contrary to popular belief, this graph shows the Netherlands, Great Britain, and France as clearly dominating the trade between 1650 and 1810, while the Portuguese dominated the slave trade on the coast for a mere forty years between 1810 and 1850. Thus a pattern of free trade dominated the slave trade on the Loango coast, while the Angolan coast was distinct for its Portuguese dominated trade. However the degree to which these differing patterns of trade were dependent on local politics must remain the topic for an additional study. Dividing the West Central African coast in this manner provides a radically different pattern of European influence. This is exemplified when exploring the case of the Dutch on the Loango Coast. Switching from the hard numbers of the Slave Voyages Database to the estimates, the database suggest approximately 5.7 million slaves embarked from the West Central African coast of which Portugal/Brazil accounts for 4 million or 70 percent of these slaves while the Dutch account for a mere 200,000 or 3.6 percent.19 However, if the coast is divided at Mpinda as exhibited in Table 1, the concentration of Dutch activity on the Loango Coast is far more pronounced as is evident in Graph 3.

Graph 1: Slave Exports from the Loango Coast by Ship Flag 1514–1800 70000

60000

50000 Other Denmark / Baltic 40000 France

30000 U.S.A. Netherlands 20000 Great Britain

10000 Portugal / Brazil Spain / Uruguay 0

84 F. R. DA SILVA AND S. SOMMERDYK

Graph 2: Slave Exports from the Loango Coast by Ship Flag, 1800–1864

200000 180000 160000 140000 120000 100000 80000 60000 40000 20000 0

Spain / Uruguay Portugal / Brazil Great Britain Netherlands U.S.A. France Denmark / Baltic Other

For the Dutch, the Loango Coast was the most significant coast in the transatlantic slave trade, embarking almost 150,000 slaves, followed by 125,000 in the Bight of Benin and 100,000 on the Gold Coast. Interestingly, the Dutch transported 16 percent of all slaves embarked at the Loango Coast. If these numbers can be seen as indicators of European importance to the formation and evolution of merchant communities on the African coast, then this is an area where the Dutch played an unusually large role. Second only to their importance to the slave trade on the Windward Coast where they engaged in 23 percent of the slave trade but also where the supply of slaves was far less substantial: the Dutch exported an estimated 80,000 of a total 340,000 slaves.20

REEXAMINING THE GEOGRAPHY AND MERCHANTS 85

Graph 3: Flag of Ships Embarking Slaves at the Loango Coast

Great Britain 14% Portugal / Brazil 40%

Netherlands 16%

Spain / Uruguay U.S.A. 1% 6% Other 0%

Denmark / France Baltic 23% 0%

The radically different picture of Dutch influence which emerges through the re-categorization of the West Central African coast is only one of the many potential insights into understanding the complex and overlapping trades which existed on the Loango and Angolan coast. Evidence indicates that the Loango coast trade drew from a north eastern hinterland not accessible to Luanda traders. Differing trade patterns of European engagement further solidifies this separation. While acknowledging these differences, the redefinition of geographical regions to include these distinctions would also resolve the problem of the awkward category of “Congo North” as a “port.” It could be easily redefined as “Loango Coast undefined.” Additionally, it is important to note that no matter how West Central African regions are defined or where the boundaries are formed, to view these regional categories as fixed and immovable categories with contained systems would be short-sighted. Instead, the Loango and Angola coasts can best be understood as overlapping spheres with distinct but interconnecting systems, as are all of the

Table 1: Numbers of Enslaved Africans Embarked in the Primary West Central African Ports by Ship Flag

Port Spain / Portugal / Great Netherlands U.S.A. France Other Totals Uruguay Brazil Britain NORTH Cabinda 4,276 241,198 29,572 6,521 10,347 55,972 347,886 ∗ Congo North 140 1,034 4,179 78,057 2,635 2,859 100 89,004 Congo River 5,132 32,022 39,113 37,797 3,392 1,812 119,268 Rio Zaire 2,542 2,542 Kilongo 1,145 1,145 Loango 3,727 5,176 23,293 23,148 4,873 55,311 115,528 Malembo 63,066 24,572 25,742 77,843 191,223 Mayumba 210 212 716 149 1,309 2,596 Mpinda 2,554 2,554 SOUTH Ambona 1,016 1,016 Ambriz 1,895 61,486 11,646 4,853 4,665 84,545 Benguela 3,982 336,442 1,537 331 391 342,683 Benguela Velho 401 401 Nova Redonda 1,795 1,795 Quicombo 622 622 Salinas 572 572 Luanda 5,424 1,350,134 1,304 15,827 107 801 475 1,374,072 Total 2,677,452

∗“Congo North” is equivalent to “Loango Coast Undefined" as argued above.

Table 2: Total Documented Enslaved Africans Embarked for Shipment to the Americas Divided by Region

Embarkation Region A) Slaves B) Estimate Regional Totals Embarked (from B) West Africa Senegambia and offshore Atlantic 425,463 755,513 Sierra Leone 236,570 388,771 Windward Coast 181,358 336,868 Gold Coast 737,236 1,209,321 Bight of Benin 1,534,827 1,999,060 Bight of Biafra and Gulf of Guinea islands 1,084,413 1,594,560 6,284,093 West Central Africa Loango Coast 1,083,407 1,854,085 Angolan Coast 2,244,188 3,840,490 5,694,575 Other Southeast Africa and Indian Ocean islands 378,283 542,668 Other Africa 330,661 0 Asia & Africa 490 0 542,668 Total 8,236,896 12,521,336 12,521,336

Sources: http://www.slavevoyages.org/tast/assessment/estimates.faces and http://www.slavevoyages.org/tast/database/search.faces; Accessed 12 March 2010.

Table 3: Estimation of Slaves Embarked in Western Africa by Region and Flag

Spain/ Portugal/ Great Nether- Denmark/ Regions U.S.A. France Totals Uruguay Brazil Britain lands Baltic Senegambia & off- 122,088 221,612 226,637 9,205 43,791 124,247 7,933 755,513 shore Atlantic Sierra Leone 85,432 16,907 163,393 2,276 56,494 61,048 3,221 388,771 Windward Coast 10,558 9,248 200,905 79,102 13,375 23,681 0 336,868 Gold Coast 6,705 68,394 718,127 103,375 126,259 115,574 70,887 1,209,321 Bight of Benin 132,018 1,009,212 353,853 126,913 4,402 348,897 23,765 1,999,060 Bight of Biafra 188,288 156,167 1,030,582 28,677 7,037 182,284 1,525 1,594,560 West Central Africa 432,789 4,018,540 534,280 204,788 29,464 472,288 2,425 5,694,574 South-east Africa & 83,646 348,185 31,663 0 24,504 53,383 1,286 542,668 Indian Ocean Totals 1,061,524 5,848,265 3,259,440 554,336 305,326 1,381,404 111,041 12,521,336 REEXAMINING THE GEOGRAPHY AND MERCHANTS 89 regions of the Western African slave trade. The next section will highlight one aspect of this interconnectivity by exploring the presence of slave merchants within the Slave Voyages Database.

II: Categorization of the Slave Merchants: Problems, omissions and their implications

Having explored the geographical challenges highlighted in the discourse surrounding the Slave Voyages Database, we will now shift our focus to the implications of the database in the study of this hideous trade from the perspective of business history. The current version of the Slave Voyages Database available online offers a considerable amount of data about the names of the ships, their owners and captains. The availability of search options within these categories seems ideal if the reconstruction of the activities of a specific businessman, a merchant house or firm within a defined time frame is desired. For the regions being scrutinized in this article, this type of search might be extremely useful to reconstruct the activities of private merchants operating side-by-side with commercial companies, like the Middelburgsche Commercie Compagnie. The name search poses, however, multiple challenges to scholars interested in studying the organization of the trade and European, or American and African engagement. Using the question of identifying networks of slave merchants operating simultaneously in West Central Africa and in other regions of the Atlantic and the Indian Oceans, this section will explore the possibilities for future research while highlighting some of the issues with the definitions and limitations of the Slave Voyages Database’s organizational model.21 The main difficulties in utilizing the database become apparent when using the categories of ship captain and ship owner and also by the absence of crucial elements for the reconstruction of commercial activities and webs, such as freighters, insurers, credit providers, agents, and brokers, which are highlighted in the following pages. The categories of ship owner and ship captain used in the database present several problems, especially if researchers are examining non-English participation in the slave trade. These categories were created according to an English model, which does not neatly fit the multiple Portuguese and Spanish categories for crew leadership and

90 F. R. DA SILVA AND S. SOMMERDYK ship ownership. For example, Portuguese and Spanish distinctions between shipmaster, ship’s commander, and ship’s pilot are lost when conflated into the category of ship captain. This loss of information is evident when reexamining primary sources and multiple professional categories re-emerge, such as ship’s pilot, commander, captain, master, and so on. Each of these terms has distinct meanings in the context of Portuguese and Spanish maritime enterprises (see Table 4). The same principle applies to the multiple forms of rights that businessmen held over ships. Some of this information is lost due to the expanded categorization of ship owner. Owning a ship as an item implied a set of rights over the property (propriedade). Within the Iberian World, these rights were distinct from rights of someone owning the use of a ship (senhorio), as property. As a consequence, both legally and in practice, the ownership of a vessel as an object and the right to its use could either be in the hands of two separate merchants or more, or of a single merchant (see Table 5). The single field in the new Slave Voyages Database for vessels ownership does not address the distinctions mentioned above. Moreover, these two types of property rights also had implications on the freightage of ships. The merchants who possessed rights to the ship’s use were able to rent it out to a third party. However, those who owned just the ship/object and not its use, were only entitled to dispose of the ship as an object, but could not lease out the ship’s use to a third party. Fretador and armador are two other problematic terms (see Table 6). Their literal translation into English is freighter and skipper, respectively. While the translation of the former offers no problems, the latter is extremely complex. An armador could be indeed a skipper, but could also be a man in charge of equipping the ship. Again, the two functions could be held either by an individual, by multiple merchants, or even by the captain himself. These issues, however, cannot be addressed in the context of the Slave Voyages Database because freighters and skippers are altogether absent as categories. Another problem is the overlapping of professional categories and accumulation of multiple functions, as Table 5 clearly illustrates. As an example of Atlantic intercontinental commerce, the slave trade required large investments in insurance, freightage of ships,

Table 4: Variations on crew leadership: a selection as way of example

First known Name (modern year of Known activity Portuguese Known Areas of Activity (in spelling) activity in the slave trade terminology Africa and Indian Ocean) Voyage ID West Central Africa, St. Câncio, Jose ship's master, Helena, Southeast Africa, Joao 1818 ship's captain mestre, capitão Indian Ocean Islands 1053 ship's captain, second capitão, West Central Africa, St. Carrilho, João lieutenant, ship's segundo-tenente, Helena, Southeast Africa, Rodrigues 1806 commander comandante Indian Ocean Islands 144; 7240 Bight of Biafra, Gulf of Guinea ship's captain, Islands, West Central Africa, Chaves, José de ship's capitão, St. Helena, Southeast Africa, Freitas 1826 commander comandante Indian Ocean Islands 874; 1037; 734 ship's master, ship's mestre, Franco, Joaquim commander, comandante, West Central Africa, Southeast António 1826 ship's captain capitão Africa, Indian Ocean Islands 915; 1229; 3353 Gomes, ship's pilot, West Central Africa, St. Domingos ship's captain, piloto, capitão, Helena, Southeast Africa, António 1799 accountant caixa Indian Ocean Islands 49779; 7691 Source: http://www.slavevoyages.org22

Table 5: Variations on ships’ ownership: a selection as way of example

First known Know activity Name (modern year of in the slave Portuguese Known Areas of Activity (in spelling) activity trade terminology Africa and Indian Ocean) Voyage ID businessman, ship's owner, negociante, Almeida, António lease holder, proprietário, Bight of Benin, Southeast 47209; 7056; 49780; da Cruz e 1784 freighter senhorio, fretador Africa, Indian Ocean Islands 900054 West Central Africa, St. Helena, Bight of Biafra, Gulf of Almeida, Bernardo merchant, comerciante, Guinea Islands, Southeast Luís de 1817 ship's owner proprietário Africa, Indian Ocean Islands 76; 305 merchant, co- comerciante, co- West Central Africa, St. Alves, António ship's owner, proprietário, Helena, Southeast Africa, Ferreira 1811 ship's captain capitão Indian Ocean Islands 1095; 5007; 915 merchant, comerciante, West Central Africa, St. Caldeira, António ship's owner, proprietário, Helena, Southeast Africa, Jose da Silva 1811 lease holder senhorio Indian Ocean Islands 900086 Sá, José Bernardino West Central Africa, St. de (Baron and Helena, Bight of Biafra, Gulf of 2099; 2118; 2131; 2132; Viscount of Vila businessman, negociante, Guinea Islands, Southeast 2136; 2138; 2139; 2207; Nova do Minho) 1825 skipper armador Africa, Indian Ocean Islands 2245; 2310; 3408 businessman, negociante, West Central Africa, St. Silva Porto, João skipper, ship's armador, Helena, Southeast Africa, Alves da 1818 owner proprietário Indian Ocean Islands 1996

Table 6: Sale of Slaves by Prins Tom to the MCC ship Prins Willem V, November 1757

Port Date African Trader(s) Total No. of Slaves Malemba 1 November 1757 Prins Tom 17 Malemba 2 November 1757 Prins Tom 10 Malemba 3 November 1757 Prins Tom 2 Malemba 4 November 1757 Prins Tom 2 Malemba 5 November 1757 Prins Tom 3 Malemba 6 November 1757 Prins Tom 6 Malemba 7 November 1757 Prins Tom 5 Malemba 8 November 1757 Prins Tom 21 Malemba 10 November 1757 Prins Tom & Jan Clase 9 Malemba 11 November 1757 Prins Tom 8 Malemba 12 November 1757 Prins Tom 12 Malemba 13 November 1757 Prins Tom & Jan Clase 10 Malemba 16 November 1757 Prins Tom 1 Malemba 21 November 1757 Prins Tom 11 Malemba 22 November 1757 Prins Tom 2 Malemba 25 November 1757 Prins Tom 1 Sources: ZA, MCC, 985. 94 F. R. DA SILVA AND S. SOMMERDYK purchase of cargo, recruitment and payment of crews and commercial agents. Additionally, insurers, merchants and commercial agents needed to possess extensive commercial expertise and an extensive knowledge of supply and consumption markets in Western Africa, the Americas and Europe. This knowledge was necessary in order to build wide trading networks which covered several geographical areas. By including only vessels owners and captains in the Slave Voyages Database, key elements in the study of the slave trade have been lost. Information regarding the ownership of cargo, insurers, credit providers, supercargoes on board the ships and business conducted on shore is all extremely crucial and is at risk of being lost. Without this information, important direct and indirect investors and actors of slave trading have been left out of the picture. Several Portuguese Sephardi merchants based in Amsterdam well-known for their engagement in the Western African commerce and the slave trade (among other areas of investment) during the first half of the seventeenth century, are not included in the database mainly because most of the ships they used were freighted.23 The case of Diogo Nunes Belmonte is an excellent example. The ship De Engel Michiel, whose voyages are listed in the database, was freighted by Belmonte on 22 May 1613 to transport enslaved Africans from Luanda to the Spanish West Indies and return to Seville loaded with Spanish bullion, gold, silver and other goods.24 The same applies to the insurers of the slave ships and cargoes. These men were indirect investors in the trade, but they were essential in the operation of the business. In seventeenth-century- Amsterdam, Jan Jansz Smits, Claes Andriaesz, Albert Schuijt, Barent Sweets, Jan de Clerck, Pelgrom van Dronckelaer, Anthoni van Diemen, Hans van Soldt, Hans van Geel, Hendrick Voet, Willem Pauw, Van den Bogaert, Wijbrant Warwijck and Salomon Voerknecht were the most important entrepreneurs backing the insurance of commercial voyages to Western Africa and the slave voyages to the Americas.25 For instance, in 1614, Jan Jansz Smits, in association with Anthoni van Diemen, Pelgrom Van Dronckelaer, Hans van Soldt de Jonge, Hendrick Voet, Albert Schuijt, William Pauw, Van der Bogaert and many others non-Jewish merchants based in Amsterdam insured Diogo Nunes Belmonte, for a ‘cargo of slaves’ on board De Engel Michiel, skippered by Sebastião Ribeiro, as well

REEXAMINING THE GEOGRAPHY AND MERCHANTS 95 as the return cargo, which was to be gold, silver and other commodities. The ship was to sail on the route Luanda–West Indies– Seville.26 Again, their names do not appear in the slave trade database, though the voyage is listed. The examination of the participation seventeenth-century chartered companies in the slave trade through the lens of the slave trade database might also be misleading. The Dutch West India Company, for instance, appears as owner of various vessels operating in the slave circuits. However, this does not mean that the Company was a key player in the business. Often the ships were freighted by private merchants, and the cargoes were their property or the property of other parties. For example, on 1 August 1657, Henrico Mathias signed a contract with the directors of the West India Company to furnish slaves to Curaçao using the ship Den Coninck Salomon.27 Heerman Abrahamsen, Dirck Pietersz Wittepaert and Andries Sael signed an identical agreement with the West India Company in 1662, for the supply of slaves to Curaçao, by the ship Abrahams Offerande, which was the property of the aforementioned merchants.28 Other contracts followed, like the one signed between the West India Company and Marcus Broen and associates in 1675 also for the transportation of slaves to Curaçao.29 The analysis of the data available in the database might lead users less familiar with the primary sources to paint an incomplete portrait of the trade. Some of the merchants that appear in the slave trade database as vessel owners or captains held multiple roles, as mentioned earlier. For instance, Francisco Ferroni, a merchant in Amsterdam, was not only a ship owner. He was the representative of Domingo Grillo and Ambrósio Lomelin, holders of the Spanish asiento (1662–1669) in Amsterdam. In this capacity, Ferroni negotiated the agreements between the asientistas and the West India Company and appears in the notarial contracts referring to slave voyages as a party with interest in the trade. At the same time, in 1664, Ferroni granted power of attorney to Martin Noel to sign a contract with the agents of Grillo and Lomelin in Barbados for the supply of 600-1000 “Black Indians or Moors” to the Spanish West Indies. Martin Noel was a merchant in London and an investor in the Royal Company of the Adventurers of England.30 Credit providers only indirectly engaged in the entire operation are another important 96 F. R. DA SILVA AND S. SOMMERDYK group of investors in the slave trade that is also absent from the Slave Voyages Database. Additionally, non-European traders operating in Africa, the Americas and the Indian Ocean have also been neglected. Despite the rich data available in eighteenth century, the Middelburgsche Commercie Compagnie records regarding multiple trading partnerships with indigenous traders on the Loango Coast are incomplete. All data on indigenous traders has been excluded. The records list over 600 individual African traders trading on the Loango Coast alone. Prominent traders with the Middelburgsche Commercie Compagnie on the Loango Coast included Prins Tom, Jan Claase, and Tom Arij.31 A study of six voyages of the Middelburgsche Commercie Compagnie ship Prins Willem V to the Loango Coast mapped transactions between Captain Adriaan Jacobse and numerous African traders. In these records, Prins Tom emerges as the most prolific slave supplier. Between 21 August 1755 and 25 November 1757, Prins Tom engaged in 154 sales. Prins Tom supplied the Prins Willem V with a total of 267 slaves in only two voyages (see Table 6).32 The compilation of a list of these traders would increase not only our understanding of European/African trade relations on the coast but could also provide more solid information on the origins of the enslaved Africans which embarked upon European ships. However, the use of the voyage as the organizing principle has the unfortunate result of strengthening the Eurocentric bias of the data. This issue leads us to the relevant matter of networks within the slave trade business. Often captains, pilots, commanders, freighters and ships owners performed various tasks and roles related not only with sailing but also to business. Their tasks often included operating as accountants in charge of commercial transactions on board the ships and on the coast, where they would conduct trade with local traders whether they were African, Euro-African or European (see Table 7).33 Henrico Mathias, for instance, had connections in Europe, Western Africa and the American colonies to organize his participation in the slave trade. In Europe, Mathias appeared associated with Jacinto Vasques, a merchant in Seville with investments in the slave trade, as well as with Marcelo van der Goes and Philip van Hulten who were merchants in Amsterdam.34 On the Gold Coast and Curaçao, Henrico Mathias maintained regular contact REEXAMINING THE GEOGRAPHY AND MERCHANTS 97 with the representatives of the West India Company during the 1650s and 1660s.35 In Curaçao, Mathias and his partners (Guiljelmo Belin le Garde and Philip van Hulten, also merchants in Amsterdam) had Ghijsberto de Rosa, conducting trade on their behalf with some inhabitants of the island and the Company since the early 1660s.36 Edward Man and Isaac van Beeck, directors of the West India Company Chamber of Amsterdam, were also important contacts for Mathias’ business with Curaçao.37 These important links in the slave trade commercial chain cannot be retrieved from the data assembled in the Slave Voyages Database. The names of accountants, information about supercargoes on board the vessels, names of local merchants operating in Western Africa and the Americas on their own name or as commercial agents of others has not been gathered and made available for study. In addition to all the aforementioned difficulties, scholars examining the business activities of the men engaged with the slave trade also have to overcome many linguistic challenges. Many personal names in the database have language symbols associated with them, in particular the French, Portuguese and Spanish names. However, the names have not been standardized. For example, the name of António Pedroso de Albuquerque, owner of three slave vessels sailing between Brazil and the Congo River, appears written in various ways (see Table 8). These languages symbols also pose other challenges. During the transfer of the data into Microsoft Excel, these symbols are sometimes replaced by special characters making names difficult to decipher, especially for researchers who are not native speakers (see Table 8). Secondly, the same personal name can appear abbreviated or translated depending on the primary sources that were used to collect the data on the first place. Thirdly, identifying personal names is also made difficult by the existence of multiple spellings used for each name. This is a consistent problem within the categories of ships owners and ship captains regardless of the language used (see Table 8). These variations in form also make name identification unnecessarily complex. A fourth problem is the use of expressions such as Son, Father, Junior and Senior. They appear without specifying individual names making the recognition of each man more complicated. Fifth and finally, multiple spellings and the split of composite surnames, especially common in 98 F. R. DA SILVA AND S. SOMMERDYK

Portuguese, Spanish and French, make name searches even harder. As a solution to the difficulties pointed above, we suggest a single standard spelling for each person, using Modern Language and avoiding language symbols and special characters (see Table 8). Despite the problems of categorization and language highlighted earlier, business historians interested in analyzing the engagement of private merchants in the slave trade as well as in examining their modus operandi in the business can benefit from searching and using the information available on the Slave Voyages website. By shifting the attention from the involvement of individual states to private entrepreneurship, and by focusing on private involvement in the trade within the framework of state policies and regulations, a business history will add a new dimension to the study of the transatlantic slave trade. In the last two decades, several scholars have examined in detail merchant groups operating in the Atlantic, their economic activities and commercial and their financial networks. In addition, the economic strategies of private businessmen operating in the Atlantic either to cope with competition or promote cooperation with other mercantile groups have also been analyzed.38 By drawing on this scholarship and with the aid of the Slave Voyages Database, researchers studying private entrepreneurship in the transatlantic slave trade are poised to make an important contribution to our understanding of the mercantile dimensions of this dreadful trade and its relationship with other commercial activities in Western Africa, the Atlantic and in other geo-economic regions.

III: Conclusions and Avenues for New Research

This article has explored how the categories of the Slave Voyages Database could be refined to increase our understanding of both the geographical definition of the West Central African Coast and the contributions of the individuals involved. First we have argued for the re-categorization of the West Central Africa into two distinct regions: the Loango and Angola Coasts. Redefining these categories based on the spheres of influence and trade dominated by African polities allows us to see distinct patterns of European shipping and influence in the two regions. Secondly we highlighted the potential usefulness of the database in tracing the networks of individuals

Table 7: Agents for the slave trade: a selection as way of example

First known Known activity Name (modern year of in the slave Portuguese Known Areas of Activity (in Voyage spelling) activity trade terminology Africa and Indian Ocean) ID Fontes, José ship's master, West Central Africa, St. Helena, Joaquim de ship's captain, mestre, capitão, Southeast Africa, Indian Ocean Sousa 1817 supercargo sobrecarga Islands 3352 ship's captain, West Central Africa, St. Helena, Lopes, Francisco accountant, capitão, caixa, Southeast Africa, Indian Ocean da Silva 1813 ship's owner proprietário Islands 397 Moreira, José ship's captain, West Central Africa, St. Helena, 499; Lopes da Costa ship's master, capitão, mestre, Southeast Africa, Indian Ocean 1043; (Jr) 1823 supercargo sobrecarga Islands 47983 ship's captain, West Central Africa, St. Helena, 46468; Sousa, Francisco ship's master, capitão, mestre, Southeast Africa, Indian Ocean 7239; José de 1817 accountant caixa Islands 49039 ship's captain, Bight of Benin, West Central Silva, Manuel accountant, capitão, caixa, Africa, St. Helena, Southeast 395; Francisco 1814 ship's owner proprietário Africa, Indian Ocean Islands 515; 739 Source: http://www.slavevoyages.org39

100 F. R. DA SILVA AND S. SOMMERDYK

Table 8: Variations on spelling and other language issues: a selection as way of example

Voyage Year of Personal name Major region of slave ID Arrival (owner or captain) purchase Albuquerque, West Central Africa and 1028 1829 Antonio Pedroso de St. Helena Albuquerque, West Central Africa and 1158 1830 Antônio Pedroso de St. Helena Albuquerque, West Central Africa and 1161 1830 António Pedroso de St. Helena 2581 1837 Zulueta Sierra Leone West Central Africa and 4285 1858 Zulueta, Don Julian St. Helena 4799 1859 Zulueta, J 2224 1843 Zulueta, Juan Bight of Benin West Central Africa and 2012 1840 Zulueta, Julián St. Helena 2228 1843 Zulueta, Julian Bight of Benin Santos, Félix West Central Africa and 22 1817 José dos St. Helena Santos, Felix Jose West Central Africa and 414 1824 dos St. Helena Abrahamsen, West Central Africa and 98803 1661 Heerman St. Helena West Central Africa and 44141 1669 Abrahamsz, Hereman St. Helena West Central Africa and 11733 1670 Abrams, Heerman St. Helena Bight of Biafra and Gulf 11806 1658 Mathias, Henrico of Guinea islands Senegambia and offshore 44183 1672 Mathias, Henrique* Atlantic West Central Africa and 44187 1673 Mathijs, Henrique St. Helena West Central Africa and 98808 1663 Matias, Henrique* St. Helena Sources: http://www.slavevoyages.org

REEXAMINING THE GEOGRAPHY AND MERCHANTS 101 involved in the slave trade to the Americas which spanned both the Atlantic and Indian Oceans. We caution the user of the database to acknowledge the limitations of this data which is organized around English categories of captaincy and ownership. The absence of significant detail in the merchant data of the Slave Voyages Database presents an opportunity for the creation of a new resource. This is not to suggest that the compilers of Slave Voyages Database got it wrong, but rather that their project was so conclusive that the debate is evolving beyond the number of enslaved Africans which crossed the Atlantic on European ships. If we were to undertake such a challenge, the Slave Voyages Database provides a project template upon which slavery researchers can both expand and improve. First, by clearly defining geographical regions at the onset of the project we could avoid conflicting geographical organizations between researchers. This would mean taking into consideration local and foreign geographical, political, and economic factors and clearly presenting this information with the data. Second, we would attempt to avoid a linguistic bias in the description of job titles, including categories in English and the original language of primary sources. Fields to categorize other economic activities not directly related to the slave trade as well as political and military roles of each merchant should also be incorporated in this type of dataset. By moving from the voyage to the individual as the organizing model, we would open the opportunity to gather biographical and professional information for each merchant. It would also allow us to include non-European traders and commercial partners where the documents allow. Moreover, it would be possible to examine the transatlantic slave trade as part of a wider economic system encompassing the Atlantic, Indian and Pacific Oceans. This would put the study of this trade in a global perspective and show that it did not always have a major role or an isolated role in the economic growth/decay of certain areas as emphasized by existing historiography. Therefore, we believe that the data should not be limited to slave trading but rather that it should eventually be extended to all economic exchanges (at least to those taking place in the Atlantic World). We argue in favor of this solution because recent scholarship by us and others has shown that slave merchants invested in multiple businesses, and often financed their slave trade operations 102 F. R. DA SILVA AND S. SOMMERDYK through investments in other economic activities which were less risky and more profitable. As for the geographical borders of this potential merchant dataset, the fluidity of merchants, investors, insurers, and sailors between markets is a clear obstacle. However, this can be overcome by creating multiple geographical fields to trace each merchant’s connections to the commercial chain of the early modern inter- continental trade. This opens the opportunity to reconstruct financial and commercial networks and to analyze commercial interactions on both macro and micro scales. A potential database on merchants and their participation in the slave trade would obviously be connected to the Slave Voyages Database available online so that researchers may link biographical and professional information to actual slave voyages. This would build upon the strengths of the current database and allow scholars to optimize usage of both resources. This is only one of the many new possible directions of inquiry fueled by the innovations made possible through the Slave Voyages Database. The implication and importance of the Slave Voyages Database and its place in the historical record is something we are only beginning to comprehend. It will continue to inspire new ideas and debate and in doing so will reshape the way we understand the interconnectivity of the Atlantic, Indian and Pacific worlds.

1 Filipa Ribeiro da Silva is a post-doctoral fellow at the International Institute of Social History of the Netherlands Royal Academy of Sciences, and Stacey Sommerdyk is a post-doctoral fellow at the University of the Witwatersrand. The authors would like to acknowledge the Social Sciences and Humanities Research Council of Canada (SSHRC) and the European Union 7th Framework and the Project Slavery Trade, Slavery, Abolitions in European Histories and Identities (EURESCL) for their support of the research projects which form the basis of this article. 2 See Johannes Menne Postma, The Dutch in the Atlantic Slave Trade, 1600- 1815 (Cambridge: Cambridge University Press, 1990), 60-61; Patrick Manning, Slavery and African Life: Occidental, Oriental and African Slave Trades (Cambridge: Cambridge University Press, 1990), 10. 3 Alecuba, Ambona, Ambriz, Benguela, Boary, Bomara, Cabinda, Cape Mole, Coanza River, Congo North, Congo River, Rio Zaire, Grenada Point, Kilongo, Loango, Malembo, Mayumba, Mpinda, Nova Redonda, Penido, Quicombo, Salinas, Luanda, and Soyo. 4 Roquinaldo Ferreira, “The Suppression of the Slave trade and Slave Departures from Angola, 1830-1860,” in David Eltis and David Richardson, eds., Extending the Frontiers: Essays on the New Transatlantic REEXAMINING THE GEOGRAPHY AND MERCHANTS 103

Slave Trade Database. (New Haven, CN: Yale University Press, 2008), 313-34. 5 John K. Thornton, Warfare in Atlantic Africa 1500-1800 (New York: Routledge, 2003), 14. 6 Thornton, Warfare in Africa, 15. 7 Thornton, Warfare in Africa, 136-37. 8 Thornton, Warfare in Africa, 138-39. 9 Jan Vansina, Kingdoms of the Savannah (Madison: University of Wisconsin Press, 1966), 18. 10 Postma, Dutch Atlantic Slave Trade, 56-57. 11 Postma, Dutch Atlantic Slave Trade, 60-61. 12 Phyllis M. Martin, The External Trade of the Loango Coast 1576-1870: The Effects of Changing Commercial Relations on the Vili Kingdom of Loango (Oxford: Oxford University Press, 1972), 8-10. 13 Postma, Dutch Atlantic Slave Trade, 60-61. 14 Martin, External Trade of the Loango Coast, 116-30. 15 See Codes Section of The Transatlantic Slave Trade Database on CD- ROM, 1999. 16 Both Ferreira and Candido point to the distinct nature of the Benguela trade while stopping short of suggesting this separation. See: Roquinaldo Ferreira, “The Suppression of the Slave trade,” 31; and Mariana Candido, “Merchants and the Business of the Slave Trade at Benguela, 1750-1850,” African Economic History 35 (2007), 1-2. 17 Undefined. 2009. The Trans-Atlantic Slave Trade Database Voyages. http://www.slavevoyages.org/tast/assessment/estimates.faces. (accessed Query - Principal place of slave purchase: West Central Africa and St. Helena. 18 It is possible that records of Portuguese slave trade on the Loango Coast remain to be discovered. This could further change our understanding of trade volume for West Central Africa. 19 Undefined. 2009. The Trans-Atlantic Slave Trade Database Voyages. http://www.slavevoyages.org. 20 These figures are calculated based on the figures in Table 3. 21 Throughout the early modern period, in particular during the sixteenth and seventeenth centuries, businessmen’s participation in the slave trade as well as in other commercial branches shows low levels of specialization. As a consequence, merchants appear involved in various trades and operating in multiple regions simultaneously. The merchants engaged in the West Central African slave trade were also active in other regions of western Africa, as evidence presented in the following tables will show. For further information on the low levels of specialization of businessmen in the Atlantic trade and in the commerce with western Africa, see, for instance: Cátia Antunes, Globalisation in the Early Modern Period: The Economic Relationship between Amsterdam and Lisbon, 1640-1705 (Amsterdam: Aksant, 2004), chap. 3, 4, 5. Antunes, “Atlantic entrepreneurship: Cross- cultural business networks, 1580-1776,” Paper presented at the workshop, 104 F. R. DA SILVA AND S. SOMMERDYK

Transitions to Modernity, Yale University, Nov. 2007. Filipa Ribeiro da Silva, Dutch and Portuguese in Western Africa: Empires, Merchants and the Atlantic System, 1580-1674 (Leiden: Brill, 2011), chap. 5, 6. 22 Changes made by authors on Excel File as explained in the text. Information on known activity in the slave trade gathered from various primary sources and secondary literature, including José Capela, Dicionário de Negreiros em Moçambique: 1750-1897 (Porto: Centro de Estudos Africanos, Universidade do Porto, 2007). The information presented here is part of a larger database on businessmen involved in the trans-Atlantic and the Indian Ocean Slave Trades recently made available on the website of the Project Slave Trade, Slavery and Abolition in European Histories and Identities (EURESCL) funded by the European Union 7th Framework Programme: http://www.eurescl.eu/ 23 Among these were Diogo Nunes Belmonte, Miguel de Pas, Duarte and David de Palacios as well as Manuel Dias Henriques. For further details on the participation of these merchants in the Western African trade and the slave trade, in particular, see: Ribeiro da Silva, Dutch and Portuguese in Western Africa, chap. 6. Cátia Antunes and Filipa Ribeiro da Silva, “The African Trade and Slave Trade in the business portfolio of Amsterdam’s businessmen, 1580s-1670s,” Tijdschrif voor Sociale en Economiche Geschiedenis (forthcoming). 24 Stadsarchief van Amsterdam (former Gemeente Archief van Amsterdam, hereafter GAA), Notarieel Archief (hereafter NA) 258/81v: 1613-03-19; NA 254/188-188v: 1614-05-22. 25 Ribeiro da Silva, Dutch and Portuguese in Western Africa, chap. 6. 26 GAA, NA 254/188-188v: 1614-05-22. 27 GAA, NA 2118/...: 1657-08-01. 28 GAA, NA 1996A/113: 1663-04-28. 29 GAA, NA 322/675-699: 1675-04-27; NA 3221/695: 1675-04-27 30 GAA, NA 2231/82-89: 1669-09-09. For further details on the Grillo and Lomelin asiento, see: Marisa Vega Franco, El trafico de esclavos con America; asientos de Grillo Y Loemlín, 1663-1674 (Seville: Escuela de Estudios Hispano-Americanos, 1984), 194-202; Enriqueta Vila Vilar, “La sublevación de Portugal y la trata de negros,” Ibero-Americkanisches Archiv 2 (1976), 171-92; Johannes M. Postma, The Dutch in the Atlantic slave trade. 1600-1815 (Cambridge: Cambridge University Press, 1990), 33-38 and Appendix 3, 349-53. 31 Zeeuwsarchief (hereafter ZA), The archief van de Middelburgsche Commercie Compagnie (hereafter MCC), 216, 219, 335, 400, 410, 456, 488, 517, 524, 528, 533, 575, 677, 795, 825, 932, 938, 980, 985, 990, 1009, 1013, 1019, 1106, 1224, 1229, 1286, 1289 & 1308. 32 ZA, MCC 980 & 985. For a list of the African traders who engaged in trade with the Prins Willem V see: Stacey Sommerdyk, “Trans-Cultural Exchange at Malemba Bay: The Voyages of the Fregatschip Prins Willem V, 1755 to 1771,” in Circuits of Exchange: Slaves, Capital and Networks in REEXAMINING THE GEOGRAPHY AND MERCHANTS 105

Atlantic Commerce (16th-19th centuries), edited by Filipa Ribeiro da Silva and David Richardson (forthcoming). 33 For further details on slave trade networks, see: Filipa Ribeiro da Silva, “Networking Across Empires: Dutch, Sephardim and Portuguese Business and Commercial Webs for the Atlantic Slave Trade (1580-1674),” The Americas 68, 1 (Jul. 2011), 7–32. 34 GAA, NA 2117/161: 1656-11-23; NA 2715/207: 1660-04-10. 35 GAA, NA 2717/65: 1661-01-19. 36 GAA, NA 2211/140-142: 1661-07-26. 37 GAA, NA 2118/137: 1657-08-01. 38 See, for example: Peter A. Coclanis, ed., The Atlantic Economy during the Seventeenth and Eighteenth Century: Organization, Operation Practice and Personnel (Columbia, SC: University of South Carolina Press, 1999); John J. McCusker and Kenneth Morgan, eds., The Early Modern Atlantic Economy: Essays on Transatlantic Enterprise (Cambridge: Cambridge University Press, 2000); Diogo Ramada Curto and Anthony Molho, eds., Commercial Networks in the Early Modern World (Firenze: European University Institute, 2002). 39 Changes made by authors on Excel File as explained in the text. Information on the activities in the slave trade gathered from multiple primary sources and secondary literature, including José Capela, Dicionário de Negreiros em Mocambique: 1750-1897 (Porto: Centro de Estudos Africanos, Universidade do Porto, 2007). The information presented here is part of a larger database on businessmen involved in the trans-Atlantic and the Indian Ocean Slave Trades recently made available on the website of the Project Slave Trade, Slavery and Abolition in European Histories and Identities (EURESCL) funded by the European Union 7th Framework Programme: http://www.eurescl.eu/

THE REGISTERS OF LIBERATED AFRICANS OF THE HAVANA SLAVE TRADE COMMISSION: TRANSCRIPTION METHODOLOGY AND STATISTICAL ANALYSIS1

Henry B. Lovejoy University of California at Los Angeles

etween 1824 and 1841, the Anglo-Spanish Court of Mixed Commission in Cuba was responsible for creating passenger Blists for over ten thousand liberated Africans, or emancipados, found aboard forty-two different slave ships. The Registers of Liberated Africans are a unique historical source because they describe the following personal information for victims of the transatlantic slave trade: port and date of embarkation, register number, African name, Christian name, sex, age, nación (nation), height, physical descriptions (señales); and in some cases, the Christian name, nación and owner of the African-born interpreters who were used during the registration process.2 At present, David Eltis has been the leading expert working with these records in conjunction with a much larger sample of comparable registers from Freetown, Sierra Leone between 1819 and 1845.3 In the late-1970s, Eltis created “The African Names Database”— containing a total of 67,228 individual entries—using microfiche copies of these registers found in the Public Record Office, now National Archives. “The African Names Database” is searchable online as such or accessible via the “Additional Resources” link in The Trans-Atlantic Slave Trade Database.4 This article provides a critique of the existing database in relation to the Caribbean data and examines the methodologies of transcription and database construction. As such, the registers from the “Archives of the Havana Slave Trade Commission” will be examined as a single unit of analysis. However, the implications of this report are relevant to the much larger Sierra Leone dataset. Naming practices in most world cultures have rich cultural meanings which are usually passed on through family generations. It is, African Economic History v.38(2010):107–135

108 HENRY B. LOVEJOY for example, easy to identify the ethno-linguistic origins of the names John, Jean, Juan and João as being English, French, Spanish and Portuguese. This basic technique can therefore be applied to the analysis of the African names recorded in these registers.5 A native Yoruba- speaker from modern-day Nigeria, for example, will recognize Yoruba names from passengers aboard the ships leaving Lagos or Ouidah. Moreover, some of the listed Yoruba names are very specific to Yoruba sub-groups such as Ijẹbu, others to Ọyọ, Ẹgba, etc. Clearly, a native Yoruba-speaker will have more success isolating Yoruba names, even though s/he may be familiar with, yet not feel entirely comfortable identifying non-Yoruba names from neighbouring ethno-linguistic groups who also left from the Bight of Benin, such as Nupe, Borgu, Hausa, Fon, Edo, Mahi, etc. While the method of interpreting the African names is basic, the practice is inherently more complex because the forty-two ships in this collection left from nineteen different ports located between Bissau and Luanda. A Yoruba-speaker would presumably have much more difficulty recognizing, with any precision, names from the Upper Guinea Coast or West Central Africa. A complete analysis of the names will therefore require numerous volunteers from hundreds of different West African ethno-linguistic backgrounds. The compilers of The Trans-Atlantic Slave Trade Database began The African Origins Project, which is co-directed by Eltis, based out of Emory University and released in 2011. It is a scholar-public collaborative intending to trace the geographic origins of these victims of the transatlantic slave trade. According to the website, this interactive online research tool will solicit volunteers with knowledge of African languages, cultural naming practices, and ethnic groups to draw on their own expertise to identify the likely ethno-linguistic origins of these documented names.6 In due course, the results should provide an idea of the ethnic composition of the transatlantic slave trade in this period. The project will enable people to search for names according to specific African regions and view a transcription of the names stemming from the spelling at the time of the nineteenth century registration. It will also be possible to listen to a sound-bite of the presumed pronunciation to REGISTERS OF THE HAVANA SLAVE TRADE COMMISSION 109 hopefully clarify any colonial distortions in spelling. Users will then be able to submit their opinions as to the name’s possible ethno-linguistic origins, which a team of experts involved with The Africans Origins Project will verify each of the submissions for accuracy. Since so much weight will be placed on the African names, any mis- transcribed names could significantly affect the results of The African Origins Project. During the registration process, Spanish and British secretaries, who almost certainly had no fluency in many, if any, West African languages and dialects, spelled out these names phonetically as they heard them spoken. Furthermore, most non-Muslim, sub-Saharan cultures had not developed orthographies for their languages in the first half of the nineteenth century. To afford the registration process some credibility, African-born interpreters were present to assist the secretary. However, they were almost certainly illiterate and probably spoke some Spanish at a minimal level; hence, they could not have verified the secretary’s spelling or accentuation. It is unknown exactly where these interpreters came from in Africa and what languages they spoke because only their Christian names and nación were documented. The interpreters typically had the same nación as a group of people on a given boat. A perfect transcription of the African names is sometimes impossible because it cannot be known with any certainty exactly what the secretary was trying to spell. Fortunately, the documents are kept in excellent condition and the majority of names are entirely legible, if not always easy to transcribe. However, like most documentation from the colonial period, the records are handwritten and the ink has sometimes blurred and faded over time. A name like Guegue, for example, could just as easily read: Gueque, Quegueor Queque, all of which are plausible names. In other similar examples, arguments could be made for more than one possible spelling of the same name. This report examines the conditions in which the registers from “The Archives of the Havana Slave Trade Commission” were made in the past and the strategies implemented to minimize transcription errors in the present. The most recent transcription of the Caribbean registers was the result of a lot of thorough work and involved 110 HENRY B. LOVEJOY the collaboration of several historians to scrutinize the letters of each African name. This paper also presents the statistical data contained in these registers on gender, height and age; regions and ports of embarkation; and by the nación classifications in relation to regions in Africa.

Transcription Methodology and Database Construction

In the National Archives, England, the Colonial Office (CO) 313 series is labelled “The Archives of the Havana Slave Trade Commission.” In chronological order therein, numbers CO 56-62 are the registers originating in both Cuba and the Bahamas. The majority of the registers were made in Havana and composed in Spanish, except for the only two registers of the 1840s, the Jesús María and Secunda Rosario, which were created at Nassau and written in English. Repatriation of the collection to London probably occurred soon after the abolishment of the transatlantic slave trade in the 1860s, but it is unclear if the whole collection from Cuba went to the Bahamas at any point in time, or if the two English registers were added to the Spanish ones later on. In any case, all these voyages were originally bound for Cuba. Even though two registers were made in the Bahamas, the Mixed Commission in Havana remained open into the 1860s. After its closure, the records were most likely taken to England at this point in time. There are also handwritten duplicates of the original collection from the CO 313 series scattered throughout the Foreign Office (FO) 84 series. It is not known exactly when or where these duplicates were made, but the handwriting was different and contains twenty-five of the forty-two registers between 1828 and 1835. Outside of England, additional copies of these registers have not been located in the Archivo Nacional de Cuba or the Archivo General de Indias, Spain.7 During the re-transcription, I worked from Eltis’ database and digital copies I made of the CO 313 and FO 84 collections in 2002 and again in 2009. I used the documents from the CO 313 series exclusively, except for when an African name was entirely illegible or had a plausible alternative spelling, and if a duplicate was available, only REGISTERS OF THE HAVANA SLAVE TRADE COMMISSION 111 then would I consult the FO 84 series. Both collections found in London maintain the same structural format of rows and columns of the personal information per captive. Several key factors contributed to the necessity to review and update the data in “The African Names Database.” First, the African names recorded in the SPSS program of the 1970s did not include the accents originally recorded in the colonial documentation because computer technology at this time was not yet advanced enough to do so. In African tonal languages, accents could change the name’s pronunciation, interpretation and meaning; hence, the ethno-linguistic details about the origins of an individual. It is worth noting that there are no accents included in the majority of the registers written in English whether from the Bahamas or Sierra Leone. Second, the “African Names Database” is not organized in the same sequential order as the colonial records, but instead loosely around nación classifications. For some reason the original register numbers were not included in the “African Names Database,” which makes cross-referencing with the original record a logistical nightmare. Third, some of the numeric fields for age, sex, height and nación were sometimes left blank, miscopied or hit with what seems to have been some sort of computer glitch. Fourth, this database does not include the documented Christian name given to each captive during the registration process. This is a serious weakness because emancipados and the enslaved population were identified by their Christian names in colonial Cuba. Christian names in other forms of documentation which are linked to a registered ship may be used to identify an African name and hence the ethno-linguistic origins of the individual. Last, the total for the sample Eltis reached is missing thirteen individuals. In light of the aforementioned reasons, it would appear as if the “African Names Database” was not adequately edited and checked after it was first transcribed from microfiche. As Eltis recognizes, however, most of the names in the original documents are legible and any mis- transcribed African names probably represent a small percentage overall. Furthermore, neither “The African Names Database” nor The Trans- 112 HENRY B. LOVEJOY

Atlantic Slave Trade Database includes an inventory of the collection of registers from the Caribbean or indeed Sierra Leone. Even so, some African names were inevitably and unintentionally miscopied. Despite the mistakes, Eltis published this database online in order to give other scholars a chance to work with the data. My own research has attempted to meet this challenge and benefit from its online release.8 Before a revision of “The African Names Database” could begin, it became evident that an inventory of the original collection should be established to track progress. As noted previously, “The African Names Database” had thirteen names less than my total of 10,391. To explain this difference, sometimes the secretary wrote little notes into the ledgers instead of an African name. For example, the African name for captive 252 of the Julita reads, “He did not express his name to the interpreter because he is mute.”9 In other cases, it simply states, “does not say a name,” or “does not have one.” In any case, the notes written in the place of an African name apparently had no practical value to the aims of “The African Names Database” and indeed The African Origins Project. Otherwise, Eltis simply missed two names during his transcription. Rosanne Marion Adderley published an inventory of the collection from Cuba in a study unrelated to the aims of The African Origins Project. She used these data to examine the male to female ratio of the emancipado community from the Caribbean registers. She demonstrates that the proportion of males in the entire sample was about 70% and that the total was the same as mine.10 Appendix I is my updated inventory of the Cuba dataset and includes: register date, ship name, port of embarkation and total number of individuals per ship. It also includes the Voyage ID number for easy cross-referencing with The Trans-Atlantic Slave Trade Database. However, it should be noted that the “The African Names Database” and The Trans-Atlantic Slave Trade Database still does not take into consideration the further displacement of these liberated Africans from Cuba to other British colonies in the Caribbean. Since “The African Names Database” does not include accents for intonation, original register number and Christian names, I systematically REGISTERS OF THE HAVANA SLAVE TRADE COMMISSION 113 re-transcribed all this information, including the African names line-by- line, ship-by-ship. Once I had completed a ship, I inserted the updated information with the individual ID number that had been assigned to each person in “The African Names Database.” This task meant descrambling the order of “The African Names Database” so that it matched the original order of the colonial records. Once the two databases were linked, I copy-edited the columns of information for age, sex, height and nación. From my review, it seemed as if a computer glitch had somehow infected the 1970s SPSS programme because there were many blank fields in the height column which always correlated to someone who measured exactly three, four and five feet. From time to time, someone's age would also read 114, when indeed it could have been either 11 or 14. Through my careful revision, I took the liberty of filling in blank fields and rectifying any minor mistakes. The transcription process was tedious, time consuming and it ran the risk of being less thorough the more consecutive hours spent working. From beginning to end, it took over six months of full-time work to transcribe, descramble and edit the 10,391 entries from the Caribbean data. It could sometimes take several days to complete one ship depending on its legibility and the order in “The African Names Database.” From my experience, the repetitiveness of data entry increased the potential for spelling, copying and typing errors. Naturally, my mind wandered from time to time and I would turn my attention to the number of possible mistakes that I could have made along the way. Sometimes I thought about how hard it would have been for the secretaries registering these weary individuals in person by the boatload and how many careless mistakes they might have made. In various places in the original records it is possible to see where the secretaries had scratched out their own mistakes. I also began to feel an ever increasing respect for Eltis who had compiled a database with over 67,000 entries compared to the mere 10,000 I was now tasked with organizing. Since there is no way of clarifying any mistakes the secretary may have made in the past, the issue at hand is to minimize the number of transcription errors in the present. As I matched my re-transcription 114 HENRY B. LOVEJOY alongside Eltis’ version, I noticed how frequently we spelled the same African names, minus the accents, in noticeably different ways. Sometimes our renditions were so different I could only locate an individual’s position in “The African Names Database” through the process of elimination and then by cross-referencing the data of age, sex, height and nación. I estimate that Eltis and I have spelled names differently by one or more letter for about a quarter to one third of the entire collection. It should also be noted that there were also some minor discrepancies in spelling among the various nación classifications and sub-classifications, which also have accents for intonation. Despite these many difficulties, many revisions were inevitably required. My revision of the Caribbean registers has demonstrated that “The African Names Database” should still be edited in its entirety. As a result of my re-transcription, Eltis brought together a team of historians involved with The African Origins Project to scrutinize over the differences between the “African Names Database” and my updated version. Over several days in July 2010, Eltis, Oscar Grandío Moraguez, Daniel Domingues da Silva, Philip Misevich and I closely compared digital copies of the original documentation with my revised database and “The African Names Database.” Through this collaborative editing process, we debated spellings, lettering and accents. In the more debatable moments, we came to a consensus by vote even if some of us continued to agree to disagree. Given the nature of these data, I would still estimate that between five and ten percent of the African names in the updated transcription may still have alternative spellings or are still mis-transcribed. Nevertheless, this methodological strategy was necessary to reduce the number of transcription errors. In consideration of The African Origins Project, it would be extremely useful to include links to digital copies of the original records so that volunteers interpreting the names could verify the colonial handwriting for themselves. The updated database of registers from the Caribbean will be published with the release of The African Origins Project. The sound bites will be re-recorded to account for the accents and possible changes REGISTERS OF THE HAVANA SLAVE TRADE COMMISSION 115 in the plausible pronunciation. The full transcription of the Caribbean, and indeed Sierra Leone, dataset is still only partial because the columns pertaining to the physical descriptions and interpreters have yet to be fully transcribed. The physical descriptions (señales) contain information related to personal injuries, body and facial markings, tattooing, scarifications and evidence of small pox. Fortunately, this data is easier to transcribe since the writing uses recognizable Spanish and English words. In consideration of the Caribbean records alone, the physical descriptions amount to a two-to-three line description per individual, which I estimate, could take more than a year to transcribe, translate and edit. These descriptions are certainly a gold mine of information relating to ethnicity and other topics, such as small pox, but new methodologies will still need to be developed on how to sort through the material and organize it for analysis.

Preliminary Analysis and Statistical Data

Of all the transcribed data contained in the Caribbean registers, sex and height are probably the most trustworthy information. It is unlikely the secretaries made many mistakes when recording this type of personal data because all that was required was a simple assessment and measurement of the human body. In the sex column, people were labelled as either varón (male) or hembra (female). Height was measured in feet and inches; and sometimes down to the fraction of an inch. In this database, feet and inches were converted into inches alone in order to calculate averages more easily. Table 1 is the distribution of the registers by gender and average height. The height data is in inches and centimetres.

116 HENRY B. LOVEJOY

Table 1: Gender and Average Height of the Emancipados Gender Individuals Percentage Height (inches/cm) Male 7,509 72.3 56/142 Female 2,881 27.7 53/135 Total 10,391

The lower percentage of females is typical for this period and reflects the high demand for male labour in response to the expansion of the Cuban sugar industry in the nineteenth century. On average, males were just under five feet tall and about three inches taller than their female counterparts. In comparison to the sex and height data, the listed ages are unreliable as an historical source. Presumably, the secretaries guessed ages at a time when these people had just endured the lengthy trial of a transatlantic crossing. It is most likely that the ages provided were entirely different than the actual ages of each individual. Given the nature of slave trading patterns at this time, young healthy adults were in highest demand and the average age of the entire dataset was just over nineteen years old. Adults listed as forty years and older represented just 3% of the overall total. The oldest male was listed at seventy-two years old and the oldest female was sixty. No one was listed as three, four and five years old; children two years and younger were probably infants. The youngest infants were listed at less than 1 month old, meaning some children could have been born on slave ships or indeed during the trial. In total, there were twenty-nine infants (sixteen male and thirteen female), and on average, they measured about two feet (61 cm). As a general pattern observed in the age, sex and height columns, the order of people in the registers was typically men, boys, women and girls, and sometimes according to nación sub-classifications. Toward the ends of the registers, there were a number of loose stragglers of varying gender, ages and nación classifications. This ordering suggests that the secretaries were trying to make some sort of distinction between adults REGISTERS OF THE HAVANA SLAVE TRADE COMMISSION 117 and children. Price levels varied and were never uniform throughout the island in any time period. Between 1821 and 1835, there were many short term fluctuations in the cost of slaves, marked by a sharp decline by 1833 when prices hit levels prevalent in the 1790s. After 1836 until the 1850s, there was general stability in slave prices with no sharp fluctuations. In this period, the average prices per adult slave varied between 250 and 500 reales, while children could have cost anywhere between 150 and 250 reales.11 Even though the listed ages could be different than the actual age, European buyers of slaves along the coast of Africa had a history of using height to distinguish between adults and children. In 1795, Thomas Leyland, a notorious slave trader and three times mayor of Liverpool, instructed Captain William Young of the Spitfire that his “ship is intitled to carry 253 full grown [people], exceeding 4 Feet 4 Inches, and 169 small [people], under 4 Feet 4 Inches, and it is certain of such description will make the highest Average in the West Indies, particularly at the Havannah.”12 This measurement—henceforth called Leyland’s measurement (converted into 52 inches or 132 centimetres)—is a good indication of how slave traders in Cuba used height to differentiate between adults and children. Regardless, Leyland’s measurement did not take into account growing abnormalities during and after puberty; thus, there could have been many adults shorter, and indeed children taller, than 52 inches. Given the original order in the records, Table 2 identifies adults and children based on Leyland’s measurement—meaning adults are equal to and taller than 52 inches, while children are considered shorter than this measurement. Table 3 assumes the ages provided by the secretary are approximate and roughly assumes the age of adulthood at puberty to be about thirteen years for both sexes.13 Both tables provide the percentages of men, boys, women and girls in the total sample, as well as averages for height and age.

118 HENRY B. LOVEJOY

Table 2: Approximate Proportions of Men, Boys, Women and Girls based on Leyland’s Measurement (LM) with Averages for Age and Height

Sex/Height Individuals Percent Age Height (cm/in.) Men (≥LM) 5,588 53.8 23.9 150/59 Boys (

Table 3: Approximate Proportions of Men, Boys, Women and Girls based on Thirteen Years Old with Averages for Age and Height

Sex/Age Individuals Percent Age Height (cm/in.) Men (≥13) 5,520 53.1 24.3 150/59 Boys (6-12) 1,971 19.0 10.3 122/48 Women (≥13) 1,737 16.7 19.4 142/56 Girls (6-12) 1,131 10.9 10.3 125/49

Again, the methods used to obtain the proportion of adults and children in both tables are highly contentious and should be treated as such. Table 2, for example, does not take into account growth abnormalities during puberty, while Table 3 does not consider how the age of adulthood is defined differently in many cultures. If, for example, a different age was used to determine the age of adulthood in Table 3, the results would obviously be very different. However, the order in the original records suggests the secretaries were trying to make some sort of distinction. No matter how debatable the results of both tables may be, the totals differed by less than one hundred people and the percentages by less than 1 percent. Furthermore, the averages for age and height in both tables are nearly identical. Tables 2 and 3, in my opinion, provide a reasonable approximation of the proportion of adults in the entire REGISTERS OF THE HAVANA SLAVE TRADE COMMISSION 119 sample, which could have been upwards of 80 percent. The high percentage of young adult males also reinforces the type of labour required to support the rapidly expanding sugar industry in Cuba. The ports of embarkation for ships in the sample (see Appendix 1) can be placed into the eight broad African coastal regions as defined in The Trans-Atlantic Slave Trade Database, which are: Senegambia, Sierra Leone, the Windward Coast, Gold Coast, Bight of Benin, Bight of Biafra, West Central Africa and Southeast Africa. The three busiest ports represented in the registers were Ouidah, Bonny and Rio Pongo, which areall estuaries that involved several locations. Table 4 reflects the distribution of the Cuban registers according to the eight regions of embarkation as defined in The Trans-Atlantic Slave Trade Database. It also provides the estimates of the total number of slaves disembarking in Cuba between 1824 and 1841 and the percentages of how much the registers represent the overall volume of trade to Cuba in this period. In total, the Caribbean registers represent about 3% of the movement to Cuba in this seventeen year period. The region in the sample that best represented the overall trade was West Central Africa and the Bight of Benin. There are no slaves in the sample from the Gold Coast and Southeast Africa. In the 1820s, hardly any ships, if any at all, arrived to Cuba from the Gold Coast due to the increasing effectiveness of British patrols operating out of Sierra Leone and the closing of the notorious British and Dutch slave factories at Cape Coast and Elmina. Even though people from South East Africa went to Cuba in this period, none are represented in this sample probably because the British only had jurisdiction in the Atlantic Ocean north of the equator until the amendments of 1835. After, it would take the patrols some considerable time to effectively patrol the South Atlantic. The regions of embarkation as defined in the The Trans-Atlantic Slave Trade Database (TSTD2) overlap with the uses of the ten broad nación classifications used in the Caribbean registers. Table 6 reflects the Caribbean dataset according to nación and the African regions of embarkation. Based on the relatively small groupings of people boarding slave ships in Senegambia, Sierra Leone and Windward Coast, I elected 120 HENRY B. LOVEJOY to group these eleven voyages together under one region labelled the Upper Guinea Coast. This region includes ports located between Bissaud and Rio Pongo in upper Guinea. Since there were larger contingents of people leaving from the Bight of Benin, the Bight of Biafra and West Central Africa, these regions are the same as defined in the TSTD2. The percentages represent the proportion of nación classifications in the entire sample.

Table 5: Regional Origins of Emancipados with Estimates of the Overall Volume of the Transatlantic Slave Trade to Cuba, 1824– 1841

Region of Embarkation Registers Total Percent Senegambia 146 3,142 4.6 Sierra Leone 1,545 43,708 3.5 Windward Coast 207 5,575 3.7 Gold Coast 0 2,711 0.0 Bight of Benin 3,663 59,458 6.2 Bight of Biafra 2,514 129,831 1.9 West Central Africa 2,316 33,476 6.9 Southeast Africa 0 35,563 0.0 Total 10,391 313,464 3.3

The historiography related to when, where and how these colonial terms originated and what they may have meant throughout the entire history of the transatlantic slave trade is extensive.14 Colonial slave traders and owners subjectively classified people from the many different regions in Africa according to these broad colonial terms. They do not take into consideration the myriad of African ethno-linguistic groups. Furthermore, their historical interpretation in the recent past and present has generated a great deal of historical debate especially since these terms were used differently, or not at all, depending on British, French, Dutch, Danish, Spanish or Portuguese interests. As Robin Law demonstrates, Lucumí was typically used only in Spanish speaking REGISTERS OF THE HAVANA SLAVE TRADE COMMISSION 121 colonies, whereby Nâgo referred to people leaving from the Bight of Benin in Portuguese, French and British colonies.15

Table 6: Region of Embarkation and Nación Classifications in the Caribbean Registers Regions of Embarkation Individuals Percent Nación Classifications Upper Guinea Coast de Bissau 146 1.4 Sierra Leona 4 0 Arpongo 1 0 Gangá 863 8.3 Mandinga 883 8.3 Congo 1 0 Bight of Benin Mina 686 6.6 Arará 349 3.3 Lucumí 2,624 25.3 Mandinga 2 0 Carabali 2 0 Bight of Biafra Lucumí 104 .1 Carabali 2,410 23.2 West Central Africa Congo 2,316 22.3 Total 10,391

The purpose here is not to discuss the nación classifications in relation to African ethno-linguistic groups, but to demonstrate their geographic association with their respective regions and ports of embarkation. Understanding what these terms mean could have tremendous implications related to the study on slavery in Cuba because cabildos de nación formed in accordance with these broad classifications. The scholarship related to mutual aid societies in Cuba is 122 HENRY B. LOVEJOY also extensive.16 There are also other studies examining certain slave uprisings which were fought and organized by groups of people belonging to certain nación classifications.17 Evidence related to colonial uses of these nación terms can be found in the interpreter column and side notes of the registers. Sometimes terms were used for the name of a language, such as English or French, and African interpreters were chosen because they could “speak” Mandinga, Gangá, Arará, Mina, Lucumí, Carabali or Congo. However, the nación classifications de Bissau, Sierra Leona and Arpongoprobably represent the names of ports or rivers, such as Bissau, the Sierra Leone River and Rio Pongo. De Bissau was the only classification used for the people aboard la Caridad Cubana of 1839, which left Bissau from the Upper- Guinea Coast. Sierra Leona and Arpongo only total five people; all of whom arrived on the Secunda Rosario in 1841 from the Rio Pongo with many Gangá and Mandinga. The Mandinga-Gangá boundary is not clearly defined, but according to the registers would appear to be around Rio Pongo. Mandinga only left from Rio Pongo which is an estuary on the Atlantic Ocean near Boffa, in modern-day Guinea. In other historical sources, Mandinga typically left from places in the Senegambia region, including Bissau. Gangá left from Grand Mesurado, Cape Mount, Galinhas Islands, Sherbro and Rio Pongo. Gangá is a much more complicated classification to define because this term was only used in Cuba and generally nowhere else in the Americas. In the Caribbean registers, Gangá were found aboard slave ships, often with Mandingas at Rio Pongo. Mina, Arará and Lucumí all left from ports between Popo and Lagos in the Bight of Benin and likewise the boundaries are not clearly defined. Minas were mostly concentrated around Popo, but a small percentage also left from Ouidah. Arará only left from Popo and Ouidah. There were very few Mina and Arará in this collection probably as a result of the increased effectiveness of British maritime patrols off the Gold Coast, Sierra Leone and eventually around Ouidah in the late- 1830s. Lucumí, which were the most represented in the entire sample, REGISTERS OF THE HAVANA SLAVE TRADE COMMISSION 123 primarily left from Ouidah and Lagos, but small contingents left from as far west as Popo and as far east as Bonny in the Bight of Biafra. There were also two people classified as Mandinga leaving from Lagos, which seems rather out of place. However, these two Mandinga could have referred to some of the Muslim ethnic groups located in the Bight of Benin hinterland. Carabali typically arrived to Cuba from the following ports in the Bight of Biafra: River Brass, Bonny, Calabar and Cameroons River. There were however two individuals classified as Carabali leaving from Lagos in the Bight of Benin. Otherwise, Congo typically refers to places in West Central Africa. There were nine registers with only Congo people onboard. The ports of embarkation included Loango, Mayumba, Congo River, Ambriz and Luanda. The only Congo captive leaving from the Upper Guinea Coast seems entirely misplaced or was possibly some sort of clerical error from the past. Captive number 216, Melí, a.k.a. Dolores, was listed as a twenty-four year-old Congo woman found aboard the Segunda Rosario along with people classified as Mandinga, Gangá, Arpongo and Sierra Leona. Along with these ten broad classifications, the secretaries also used over two hundred different nación sub-classifications, which appear to signify—despite their colonial distortions—more specific African ethno- linguistic groups (see Appendix II for complete list of all sub- classifications). For example, the sub-classifications, Gangá Gorá and Gangá Conó, could have referred to Gola and Kono ethnic groups respectively. Or, the two Mandinga leaving Lagos have a sub- classification of Mandinga Fula, possibly meaning they could have been Fulani Muslims. It should be noted that about one third of the entries for these emancipados did not have any sub-classification at all. Some sub- classifications also appeared to have alternative spelling. For example, Lucumí-Ayó was also written as Eyó, Elló and Aylló, which the team of historians responsible for editing the database determined were almost certainly the same sub-classification. Ayó, and its alternative spellings, probably referred to Oyo-Yoruba.18Another example includes Congo Moyombe which was also written as Congo Mollomve. In Appendix II, 124 HENRY B. LOVEJOY the sub-classifications with alternative spellings are put in parentheses and grouped together. Last, some of the sub-classifications transcended the broader colonial nación classifications as the examples of Mandinga Conó and Gangá Conó demonstrate. The nación classifications and sub-classifications over-simplify the regional, ethnic and linguistic backgrounds of these liberated Africans. They were assigned, beyond the registers, based solely on the subject opinions of colonial slave traders, owners/masters and government officials. These classifications are therefore highly problematic and cannot be trusted on their own merit. The identification of the African names, which is the primary aim of The African Origins Project, will hopefully clarify where these people came from in the African interior beyond the port of embarkation. Once the entire sample of African names has been interpreted and the data released, historians should get a better sense of the meanings associated with the nación classifications in this period.

Conclusion: Implications for the Sierra Leone Dataset and The African Origins Project

In the aftermath of the transfer of Sierra Leone to the British crown from the Sierra Leone Company in 1807 and “The Act for the Abolition of the Slave Trade” was passed the year after, a court of Vice Admiralty was established in Freetown to enforce British maritime law. This was just one of well over a hundred such courts scattered across British possessions around the globe. In fact the court that adjudicated the greatest number of liberated Africans was in St. Helena. The High Court of Admiralty, located in London, was the appellate court. By 1818, Great Britain had pressured Portugal, France, the Netherlands and independent republics of the former Spanish empirein the Americas into signing anti- slave trading treaties. Soon after, Courts of Mixed Commission, like the one in Havana, were established in some of those non-British colonies as well. Hundreds of cases were processed in these courts of Vice Admiralty and Mixed Commission. Even though the courts managed to REGISTERS OF THE HAVANA SLAVE TRADE COMMISSION 125 free hundreds of thousands of people, an estimated three million enslaved people from Africa still went to the Americas from 1808 onward. Although other registers exist, it appears that the courts at Freetown, Havana and Nassau are the only ones to include African names. Since 2010, Paul Lovejoy, Susanne Schwarz, Neil Marshall, Jennifer Toews and others have been working on the digitization project funded by the British Library’s Endangered Archives Programme in Sierra Leone and have identified a collection of registers which is larger than the collection of registers housed at the British National Archives. The collection in Freetown begins in 1808, while the London collection begins in 1819. At present, it remains unclear which of the two collections are the original, but it is most likely the collection located in Sierra Leone. The registers in Sierra Leone may be incomplete because two volumes could not be located during either research trip, but duplicates of those volumes exist in the FO 84 series in London. At this point in time, the exact total of registered individuals at Sierra Leone remains unknown, but it totals more than 84,000, instead of the 57,000 or so entries already transcribed in “The African Names Database.” It also includes approximately 12,000 from 1808-1819, not recorded in London. An inventory of both the Freetown and London collections would certainly resolve any dispute concerning the overall total, demonstrate duplicates and possibly locate any of the missing volumes. Any missing registers will obviously need to be transcribed. Since a perfect transcription of the African names is sometimes impossible, the other 57,000 or so names from the Sierra Leone dataset should be edited to reduce the number of transcription errors. Fortunately, Eltis has done the bulk of the transcription work and it is far easier to edit than it is to start over. Still, a decision must be made as to which collection should be used before “The African Names Database” is edited and that transcription is re-released with The African Origins Project.

126 HENRY B. LOVEJOY

Appendix I: Inventory of the Registers from “The Archives of the Havana Slave Trade Commission”

TSTD Reg. Date Ship Name Port of Individuals/ Voyage Embarkation Ship ID 2358 1824 Relâmpago Grand Mesurado 149 2366 1825 Isabel Gallinhas 10 2374 1826 Mágico Popo 175 558 1826 Fingal Cape Mount 58 557 1826 Orestes Ouidah 212 Nuevo 561 1826 Campeador Calabar 211 Xerxes (a. 664 1828/06/22 Gerges) Bonny 385 668 1828/08/25 Intrepido Bonny 133 753 1828/12/10 Maria Congo River 1 756 1828/12/23 Firme Popo 483 Fortuna (a. 770 1829/04/28 Josefa) Gallinhas 202 Voladora (a. 776 1829/07/01 Mulata) Popo 330 777 1829/08/12 Midas Bonny 208 941 1829/12/05 Gallito Rio Pongo 135 960 1830/05/28 Santiago River Brass 100 Emilio (a. 963 1830/07/07 Cesar) New Calabar 187 1245 1832/05/04 Planeta Cameroons River 236 1249 1832/07/03 Aguila Loango 596 1250 1832/07/16 Indagadora Lagos (Onim) 134 1266 1833/01/05 Negrito Ouidah 477 1295 1833/12/03 Joaquina Bonny 318 1298 1833/12/31 Manuelita Lagos (Onim) 477

REGISTERS OF THE HAVANA SLAVE TRADE COMMISSION 127

1307 1834/02/10 Rosa Ouidah 289 1338 1834/12/07 Carlota Gallinhas 163 1355 1835/02/01 Maria Bonny 340 1361 1835/02/23 Julita Ouidah 336 La Joven 1367 1835/04/15 Reyna Congo River 254 1368 1835/05/07 Chubasco Rio Pongo 230 1372 1835/05/12 Marte Loango 326 1383 1835/07/23 Tita Ouidah 392 1396 1835/11/24 Amália Congo River 200 1403 1835/12/31 Diligência Mayumba 94 Ninfa (a. 1446 1836/02/04 Matanzera) Calabar 396 Ricomar (a. 1462 1836/03/15 Zafiro) Ouidah 186 1479 1836/07/08 Preciosa Rio Pongo 290 1469 1836/11/11 Empresa Luanda 407 1631 1837 Matilde Ambriz 255 1569 1838 Antonica Congo River 183 Sierra del 1856 1839 Pilar Lagos (Onim) 172 CaridadCuba 1860 1839 na Bissau 146 2071 1840 JesúsMaría Sherbro 234 Segunda 2078 1841 Rosario Rio Pongo 281 Total Ships 42 Total Individuals 10,391

Appendix II: Nación Classifications and Sub-Classifications in the Registers of “The Archives of the Havana Slave Trade Commission”

Lucumí Individuals Congo Individuals LucumíAyó, Eyó, EllóorAylló) 1,235 Congo (no sub-classification) 1,172 Lucumí (no sub-classification) 598 Congo Mondongo 354 LucumíEcumachó 470 Congo Bomá 182 LucumíTapá 82 Congo Luango 154 LucumíLlabú (Yavú) 68 Congo Musundí 143 LucumíChambá (Chamvá) 62 Congo Moyombe (Mollomve) 66 LucumíOtá 61 Congo Mongoma 45 LucumíEvá (Ebá) 60 Congo Bimgo 9 LucumíAgusá (Jausá) 36 Congo Damba 8 LucumíCacanda 21 Congo Musicongo 6 LucumíBasa 9 Congo Bamba 5 LucumíMosi 3 Congo Lemba 5 LucumíYesa 3 Congo Mobonva 5 LucumíDagñame 2 Congo Antpaango 4 LucumíEcuá 2 Congo Somo 4 LucumíIgara 2 Congo Baco 3 LucumíBogú 1 Congo Camba 3 LucumíDasá 1 Congo Chocho 3 LucumíEfú 1 Congo Gongo 3 LucumíEgruá 1 Congo Gumga 3

LucumíEllico 1 Congo Lombo 3 LucumíGuarí 1 Congo Real 3 LucumíLayí 1 Congo Soso 3 LucumíLayí (Llallí) 1 Congo Tando 3 Lucumí Llama 1 Congo Biri 2 LucumíOpu 1 Congo Boco 2 LucumíPové 1 Congo Buco 2 LucumíSabé 1 Congo Buere 2 LucumíVacúo 1 Congo Bullonde 2 LucumíYacó 1 Congo Canga 2 Total Lucumí 2,728 Congo Cango 2 Total Sub-Classifications 29 Congo Cansa 2 Congo Cay 2 Carabali Individuals Congo Cuní 2 CarabaliSuamo (Isuamo) 504 Congo Enchí 2 CarabaliCamaron 396 Congo Ensuca 2 CarabaliElugo 302 Congo Fula 2 CarabaliIbibí (Bibí) 266 Congo Lano 2 CarabaliBrican 171 Congo Lucutí 2 CarabaliBrícamo 122 Congo Manba (Manva) 2 CarabaliDuri (Induri) 117 Congo Matendi 2 CarabaliCicuato 101 Congo Muema 2 CarabaliApapá (Apápá) 80 Congo Noque 2 CarabaliOrú 61 Congo Quiso 2

Carabali Ibo 52 Congo Sombo 2 CarabaliApá 39 Congo Tamba 2 Carabali Isa 35 Congo Bafo 1 CarabaliBanfule 29 Congo Bandechendi 1 CarabaliOsosó 29 Congo Bansa 1 Carabali Bane 19 Congo Biabo 1 CarabaliUngua 14 Congo Bongela 1 CarabaliOtán 10 Congo Buchimpe 1 CarabaliMogo 9 Congo Cabinda 1 CarabaliObane 6 Congo Caí 1 CarabaliUqua 5 Congo Chiongo 1 CarabaliVende 5 Congo Chita 1 Carabali Bony 4 Congo Cimchí 1 CarabaliItepu 4 Congo Cimgo 1 CarabaliAchena 3 Congo Cocumbe 1 CarabaliAgó 2 Congo Cola 1 CarabaliCalabal 2 Congo Conche 1 CarabaliEsalá 2 Congo Cuma 1 CarabaliAquese 1 Congo Cuno 1 CarabaliAsa 1 Congo Cusa 1 CarabaliAtane 1 Congo Cutuide 1 CarabaliBasá 1 Congo Danval 1 CarabaliBó 1 Congo Decolo 1 CarabaliCuachú 1 Congo Densuso 1

CarabaliCuiüa 1 Congo Emagebo 1 CarabaliCunácuná 1 Congo Enlaza 1 CarabaliDoque 1 Congo Ensadi 1 CarabaliEculasu 1 Congo Ensomga 1 CarabaliEdú 1 Congo Esombe 1 CarabaliEpá 1 Congo Febo 1 CarabaliErí 1 Congo Femba 1 CarabaliEsa 1 Congo Fete 1 CarabaliEyó 1 Congo Ganda 1 CarabaliNenu 1 Congo Gimse 1 CarabaliNiconecha 1 Congo Guaguana 1 CarabaliOcoloba 1 Congo Guelé 1 CarabaliOmuma 1 Congo Guembo 1 CarabaliOnecha 1 Congo Jali 1 CarabaliOrobio 1 Congo Laba 1 CarabaliUbacaua 1 Congo Leque 1 CarabaliUgu 1 Congo Llanga 1 Total Carabali 2,412 Congo Lomica 1 Total Sub-Classifications 51 Congo Longo 1 Congo Lotala 1 Mandinga Individuals Congo Lufo 1 Mandinga (no sub-classification) 553 Congo Lumbi 1 MandingaSosó 110 Congo Lusanda 1 MandingaTeminé 101 Congo Lusanga 1

MandingaLogó 50 Congo Macará 1 MandingaLimbá 11 Congo Maganié 1 MandingaQuisí 11 Congo Melele 1 MandingaToma 10 Congo Mesa 1 MandingaLocó 9 Congo Mesara 1 MandingaCurangó 6 Congo Moache 1 MandingaBámbara 5 Congo Moamba 1 MandingaComiaca 5 Congo Mocanda 1 MandingaBagá 3 Congo Mongo 1 MandingaFulá 3 Congo Monlaso 1 MandingaConó 2 Congo Mopaso 1 MandingaSangara 2 Congo Mudimba 1 MandingaCranco 1 Congo Mulimba 1 Mandinga Gora 1 Congo Musimba 1 MandingaVaré 1 Congo Muyala 1 MandingaYeré 1 Congo Niense 1 Total Mandinga 885 Congo Noca 1 Total Sub-Classifications 18 Congo Ocama 1 Congo Ofó 1 Gangá Individuals Congo Pesa 1 Gangá (no sub-classification) 466 Congo Queta 1 GangáLongobá (Longová) 215 Congo Quiama 1 GangáBuché 37 Congo Quindonga 1 GangáConó 28 Congo Ruya 1

GangáBeré 27 Congo Sacala 1 Gangá Fai 24 Congo Sande 1 GangáGorá 21 Congo Say 1 GangáQuisí 21 Congo Sese 1 GangáFulá 10 Congo Simba 1 GangáManí 6 Congo Sita 1 GangáBahi 5 Congo Solón 1 GangáBumi 2 Congo Suca 1 GangáToma 1 Congo Sucuté 1 Total Gangá 863 Congo Tibo 1 Total Sub-Classifications 12 Congo Timga 1 Congo Totela 1 Mina Individuals Congo Untacala 1 Mina (no sub-classification) 273 Congo Vinda 1 Mina Fanti 101 Total Congo 2,317 Mina Popó 312 Total Sub-Classifications 129 Total Mina 686 No sub-classifications Individuals Arará Individuals De Bisao 146 AraráMagí (Magín) 147 Sierra Leona 4 Arará (no sub-classification) 130 Arpongo 1 LucumiArará 67 Total 151 AraráCuatro Ojos 5 Total Arará 349 134 HENRY B. LOVEJOY

1 This paper was not possible without funding from the Tabor Pre-Dissertation Fellowship, Latin America Institute and the Department of History (all from the University of California, Los Angeles). It was originally presented at The Slave Voyage Database and African Economic History: A Workshop, York University, Canada, (May, 2010). Special thanks to David Eltis, Liz Milewicz, Oscar Grandío Moraguez, Daniel Domingues da Silva, Phil Misevich, my father, Susanne Schwarz, David Richardson and the people involved with The African Origins Project at Emory University. 2 The ten nación classifications found in these registers are: De Bisau, Arpongo, Sierra Leona, Mandinga, Gangá, Mina, Arará, Lucumí, Carabalí and Congo. 3 National Archives (NA), London, Foreign Office (FO), 84, vols. 4, 9, 15, 21, 38, 63, 64, 76, 86, 87, 100, 101, 102, 116, 127, 166 and 212. Cited from David Eltis, “Nutritional Trends in Africa and the Americas: Heights of Africans, 1819-1839,” Journal of Interdisciplinary History 12, 3 (1982), 454. 4 David Eltis, David Richardson, Manolo Florentino, and Stephen D. Behrendt, Voyages: The Trans-Atlantic Slave Trade Database (www.slavevoyages.org), (Accessed in August 2010). 5After the online release of “The African Names Database,” Eltis and G. Ugo Nwokeji demonstrate how this methodology can be applied to ascertain where certain individuals came from in the Bight of Biafra hinterland. See G. Ugo Nwokeji and Eltis, “The Roots of the African Diaspora: Methodological Considerations in the Analysis of Names in the Liberated African Registers of Sierra Leone and Havana,” History in Africa 29, (2002), 365-79. 6 See The African Origins Project, www.african-origins.org/about (accessed January 2011). Currently, I am a member of the project’s development team and academic consultant for research on the Havana registers. 7 Evidence related to the Mixed Commission and some of the trials can be found in these non-British archives, but those records are sporadic at best. 8 Eltis, personal emails and conversations (May-July, 2010). 9 “No espresó su nombre pr. los interpretes que es mudo.” 10 Roseanne Marion Adderley, “New Negroes from Africa:” Slave Trade Abolition and Free African Settlement in the Nineteenth-Century Caribbean (Bloomington: Indiana University Press, 2006), 128-29. Adderley and I reached different totals regarding the number of males and females in the dataset. Adderley’s totals were: 7,498 males and 2,893 females. Regardless of these differences, the male to female ratio remains about the same. 11 See Laird W. Bergad, Fe IglesisasGarcía and Maríadel Carmen Barcia, The Cuban Slave Market, 1790-1880 (Cambridge: Cambridge University Press, 1995). 12 Cited and quoted as Leyland to Young, 15 June 1795, Ms 10/49, Leyland papers, Harold Cohen Library, Liverpool in Paul E. Lovejoy, “The Children of

REGISTERS OF THE HAVANA SLAVE TRADE COMMISSION 135

Slavery – The Transatlantic Phase,” Slavery and Abolition 27, 2 (2006), 207- 208. 13 The twenty-nine infants are not included in these tables. There were three male captives whereby an age was not provided in the original records: captive number 61 aboard the Marte; and captives number 145 and 146 aboard the Caridad Cubana. 14 Refer to Gwendolyn M. Hall, Social Control in Slave Plantation Societies: A Comparison of St. Domingue and Cuba (Baltimore; Johns Hopkins University Press, 1971); Linda M. Heywood, ed., Central Africans and Cultural Transformations in the American Diaspora (Cambridge: Cambridge University Press, 2002); and Stephen Palmié, Wizards and Scientists: Explorations in Afro- Cuban Modernity and Tradition (Durham: Duke University Press, 2002). 15 Robin Law, “Ethnicity and the Slave Trade: ‘Lucumi’ and ‘Nago’ as Ethnonyms in West Africa,” History in Africa 24, (1997), 205-19. 16 See Fernando Ortiz, Los cabildos y la fiesta afrocubanos del Día de Reyes (Habana: Editorial de Ciencias Sociales, 1992 [1921]) and María del Carmen Barcia, Los ilustres apellidos: Negros en la Habana colonial (Habana: Oficina del Historiador de la Ciudad de Habana, 2009). 17 See Manuel Barcia, Seeds of Insurrection: Domination and Resistance on Western Cuban Plantations, 1808-1848 (Baton Rouge: Louisiana University Press, 2008); and Juan Iduate, “Noticiassobresublevaciones y conspiraciones de esclavos: Cafetal Salvador, 1833,” Revista de la Biblioteca Nacional José Marti 73, 24, (1982), 117-52. In Iduate’s article, there are over 250 African names recorded in the original documentation. At some point, these names should be included in The African Origins Project. 18 See also Eltis “The Diaspora of Yoruba Speakers, 1650-1865: Dimensions and Implications,” in Toyin Falola and Matt Childs, eds., The Yoruba Diaspora in the Atlantic World (Bloomington: Indiana University Press, 2004), 19-39. In this chapter, Eltis grouped together several of the Lucumí sub-classifications that had alternative spellings.

EXTENDING THE AFRICAN NAMES DATABASE: NEW EVIDENCE FROM SIERRA LEONE*

Suzanne Schwarz University of Worcester

n 1999 the authors of The Trans-Atlantic Slave Trade: A Database on CD-Rom emphasised that one of the “basic limitations” of the Idata set was it contained “thousands of names of shipowners and ship captains, but … no names of the millions of slaves carried to the Americas.”1 In many respects, the anonymity of the African men, women and children forcibly transported to the Americas reflects problems inherent in business records generated by slave merchants and captains who saw no utility or interest in recording the names of Africans routinely dehumanised as items of cargo.2 By representing Africans as anonymous figures devoid of any personal or cultural identity, the abolitionist image of the slave ship Brooks featured on the front cover of the original database reinforces the importance of recovering from obscurity details of African lives shattered by the Atlantic slave trade.3 The availability of systematic quantitative evidence on the forced migration of millions of unidentified men, women and children in this groundbreaking database, however, has undoubtedly stimulated new research initiatives to trace the African identity of these individuals.4 In terms of retrieving the identities and origins of at least some of the Africans affected by the trade, an important development in the new expanded Slave Voyages database is the inclusion of the names of 67,228 African men, women and children derived from lists of liberated Africans held in the FO84 series at the National Archives at Kew.5 Although these individuals represent less than one per cent of an estimated 12.5 million Africans transported in the transatlantic slave trade between the early-sixteenth and the mid-nineteenth century, the evidence offers considerable potential to trace the features of the trade in the early nineteenth century from an African perspective.6 In contrast to

African Economic History v.38(2010):137–163

138 SUZANNE SCHWARZ the “mass of black human flesh” depicted in the image of the Brooks,7 these Registers of Liberated Africans provide rich details of the names, gender, appearance, age and height of enslaved Africans released at Sierra Leone and Cuba from illicit slaving vessels intercepted by Royal Navy patrols between 1819 and 1845. Over 80 per cent of the African Names Database is composed of the names of Africans landed at Freetown between 1819 and 1845, with a further 10,378 names derived from the registers of the Court of Mixed Commission at Havana between 1824 and 1841.8 The earliest entries in the Registers of Liberated Africans at Kew (and hence in the African Names Database) date only from 1819, eleven years after the commencement of policies of slave trade suppression by royal naval patrols stationed at Freetown.9 However, important evidence on the earliest groups of African recaptives landed at Freetown in the immediate aftermath of British abolition has recently re-emerged after a period of neglect during a collaborative British Library Endangered Archives project under the direction of Paul E. Lovejoy.10 Registers of Liberated Africans, containing details of enslaved Africans released by the Vice-Admiralty Court at Freetown in the early phases of suppression activity from 1808, were retraced in the Sierra Leone Public Archives at Fourah Bay College in Freetown in February 2010. Entries for 15,967 Africans are contained in nine registers spanning the period between 1808 and 1822, and they include Africans taken off intercepted slave vessels, Africans released as a result of naval attacks on slave barracoons on the coast, as well as a smaller number of “slaves seized in the colony.”11 Comparison between these nine early Registers of Liberated Africans and the African Names Database indicates that they provide the names of approximately 12,000 Africans adjudicated by the Vice- Admiralty Court who are not currently listed in the Database, thereby substantially increasing the total of known names of enslaved Africans.12 The registers spanning 1808 to 1819 potentially include 12,178 entries which are not currently listed in the African Names Database. However, the number of African names actually listed is lower as some entries are blank. For example, no names or descriptions are provided in the EXTENDING THE AFRICAN NAMES DATABASE 139 registers for recaptives numbered from 1,100–1,105, 1,121–1,126, 1,144, and 1,151. A pencil entry notes how number 4,143 is “a blank.”13 This paper examines the scope and nature of these early Registers of Liberated Africans, and their potential for establishing the identities, origins and experiences of the first groups of recaptured Africans released at Freetown by British anti-slave trade patrols in the immediate aftermath of British abolition. These early registers not only record patterns of forced migration associated with the earliest phases of suppression activity, but also provide a basis for tracing links between individuals’ point of embarkation in Africa and their subsequent life histories. This ability to reconstruct individual life histories is exceptionally rare in the historiography of transatlantic slavery, and is facilitated in part by the rich body of documentary evidence generated by colonial administrators in Sierra Leone. These records make it possible to trace the subsequent movements and experiences of a significant proportion of these men, women and children following their release, thereby shedding light on survival strategies adopted by enslaved Africans in the aftermath of abolition. Many of the individuals released at Freetown were entered into the service of the Royal African Corps and the Royal Navy, whilst others were apprenticed for up to fourteen years in a system attacked by contemporaries as another form of enslavement. As many of the recaptives remained within the Crown Colony and its hinterland, the identification of their areas of origin also informs understanding of the diverse influences which shaped cultural formation of the settlement in the first decades of the nineteenth century.14

II

These nine Registers of Liberated Africans comprise part of a much longer series of Registers of Liberated Africans in the Public Archives of Sierra Leone which extend to 1848 and, according to the numbering sequence in the volumes, account for 84,307 enslaved Africans released at Freetown. The front cover of the final register in the Freetown series indicates that the numbering sequence extends to 84,420, but the last 140 SUZANNE SCHWARZ entry relates to Marloryar, a girl aged two identified as recaptive number 84,307 who was released from the ship Bela Miquelina on 5 August 1848.15 In the 1960s, Richard Meyer-Heiselberg concluded that the overall total was higher and estimated that 94,329 individuals were released when references to recaptives in the Liberated African Department letter books between 1848 and 1861 were taken into account.16 This far exceeds the total of 56,850 names of recaptives released at Sierra Leone who are currently included in the African Names Database. Even allowing for the addition of approximately 12,000 names from the registers spanning 1808 to 1819, this indicates that the Freetown registers include far more names than are currently listed. Systematic comparison between the names in the Freetown series of registers and the names derived from the Registers of Liberated Africans at Kew is required to trace the identity of the outstanding individuals. Another reason for cross-referencing the Freetown registers with the Kew series is that a very cursory comparison of entries points to significant differences in how names and other details were recorded. In the Freetown register spanning the period from 1819 to 1822, for example, recaptive number 12,183 was recorded as Mafa, but the register at Kew records his name as Mufas. The extent to which such differences in transcription between the two registers affects the linguistic analysis of names requires further research, and highlights the importance of a systematic comparison of the Freetown and Kew registers where they begin to overlap from 1819. Further analysis is required to determine the precise relationship between the two series of registers and whether one series is derived from the other.17 Two registers are still missing in the Public Archives of Sierra Leone but, as they relate to the period after 1822, this does not affect analysis of the earlier series of Vice-Admiralty registers which form the main focus of this paper.18 The nine registers spanning 1808 to 1822 provide some of the earliest extant lists of the names of Africans enslaved in the transatlantic trade. The first surviving Register of Liberated Africans (1808-1812) predates by more than a decade inventories of recaptive Africans used by Misevich in his analysis of “The Origins of Slaves Leaving the Upper EXTENDING THE AFRICAN NAMES DATABASE 141

Guinea Coast,” which drew on registers from the Court of Mixed Commission in Havana between 1824 and 1841.19 A distinctive feature of the nine early Registers from the Vice-Admiralty Court in Freetown is that the lists of names were not duplicated in a series of registers returned to London. Hence, the African names identified in the early series of registers offer a unique record of the identities of the very first cohorts of recaptives released at Freetown.20 At various intervals in the early nineteenth century censuses compiled in Freetown list the names of some of these recaptives, but these do not reproduce the full range of details contained in the registers at Freetown.21 Although the existence of these early Registers of Liberated Africans was known to several scholars working in Freetown,22 they have clearly escaped the notice of many slave trade scholars in recent decades. In 1966, Meyer-Heiselberg compiled an inventory of the registers’ contents, including systematic notes on the age and gender distribution of the African recaptives. Although he commented on the difficult and cramped working conditions in the archives, he did not make any direct reference to the poor physical condition of the registers. Almost half a century later, these early registers are in a very fragile and endangered condition, and their deterioration has no doubt been exacerbated by the emergency re-location of the archives during the recent civil war in Sierra Leone.23 In a preliminary visit to the Public Archives of Sierra Leone in February 2010, the Endangered Archives project team identified the first two Registers of Liberated Africans for 1808–1812 and 1812–1814 among piles of badly damaged documents stacked on all available surfaces in the congested archive office. The first register is in a particularly fragile condition. Entries for up to 3,772 Africans released between 1808 and 1812 are listed on loose pages and the brittle, acidic nature of the paper has resulted in the edges breaking away (Figure 1). There is extensive damage from iron gall, with the result that a large number of entries are extremely difficult to decipher. The second register

142 SUZANNE SCHWARZ

Figure 1

Source: Register of Liberated Africans, 1808-1812, Sierra Leone Public Archives

in a bound volume provides details of up to 2,515 Africans with identification numbers ranging from 3,773 to 6,288, but the covers are missing and the binding is loose. A third register spanning the period from 1814 to 1815 lists up to 2,823 individuals identified by numbers ranging from 4,684 to 7,507. A further six registers include entries for individuals numbered through consecutively to recaptive number 15,967, identified as Ouner, a twelve-year old male released from the Spanish schooner Josefa (a) Maracagerca in October 1822.24 Three of the nine registers contain long series of duplicate entries, although the format of the entries is not identical.25 Although seven of the nine registers have recently been re-bound, they all show evidence of damage caused by acidification, water, mildew, iron gall and insects. The environmental conditions in the archives contribute to the endangered nature of these sources, as they are housed on open shelves in conditions of high humidity.26 EXTENDING THE AFRICAN NAMES DATABASE 143

In addition to providing a unique identification number for each of the Liberated Africans, the registers provide details of the name, age, gender, and stature (height) of the individuals listed. Several registers also include information on the “disposal” of recaptives, although the entries vary considerably in frequency, content and format.27 In contrast to the Havana series, details of the recaptives’ country of origin or nation are only recorded sporadically. The names of clerks and their African interpreters are not identified in the registers at Freetown.28 In common with the later FO84 series at Kew, the “description” column provides extensive information on scarification, tattoos and other physical features. Pen sketches of markings appear at infrequent intervals, and appear to have been used to capture some of the more elaborate, decorative or unusual examples. The description for Wosousoo, a man of 20 identified as recaptive number 2,669, included a sketch of marks on his “breast & belly” which were represented as two spherical shapes with 12 short lines radiating from the upper circle and 13 short lines radiating from the one below. It was also noted how his “forehead and temples scarred” and that he had “several black spots under ea[ch] eye.”29 One of the first examples of a drawing to appear in the registers relates to a Renga, a man age 20 of 5 feet 2 inches in height who was identified as recaptive number 1,030. He had a “large round scar on back, d[itt]o below right breast” and a simple sketch of a vertical line underscored by a short horizontal line was used to represent the appearance of a mark “under [his] left arm pit.”30 As comparatively little research has been done on scarification, there is scope for systematic analysis of the “cultural code” embedded in these markings and their relationship to meanings of identity, ethnicity, social status, kinship, as well as the origins of the recaptives.31 The descriptions also document the marks resulting from the violence endemic in the trade and the branding of enslaved Africans prior to their release at Freetown.32 A number of Africans disembarked from the Portuguese/Brazilian registered Principe da Beira in 1812 had been branded with the letters TR, with the T superimposed over the letter R.33 Soqua, a woman aged 32, was “branded on [her] arm” and Cawley, a man aged 35, had a “lumpified TR on [his] 144 SUZANNE SCHWARZ arm.”34 The circumstances which resulted in Ensiah, a man aged 30, receiving a “musquet shot through [his] left thigh” are unclear from the brief entry in the register, but this may indicate that he had resisted his enslavement prior to his release from the Marie Paul in 1808.35 The “description” column frequently includes reference to other physical characteristics including smallpox markings, the colour of skin, injuries and deformities, and the appearance of hair and teeth. Details of family relationships among enslaved Africans are occasionally documented. Fatima, a woman aged 30 released from the Cuba (a) Marianha in 1809 was described as being “with a child only 4 days old” and having “Moorish hair.” Embas, a woman aged 29, was described as the “mother of Senegal No 1019,” a child who was only months old.36 Following the transfer of the colony to the Crown and the formation of a Vice-Admiralty Court, the first vessel condemned and entered in the registers on 10 November 1808 was the Marie Paul, a French schooner in the command of Captain Debouney. Neither the Slave Voyages Database nor the Register of Liberated Africans indicates where this vessel had embarked its human cargo of 69 slaves. Of these Africans, nine had died by the time the court proceedings had been completed.37 The losses were no doubt exacerbated by a period of eleven weeks in the Middle Passage, although some vessels could show devastating levels of mortality in shorter periods.38 Macha, a man age 28 whose right ear had been partly cut off, was listed as recaptive number 1 and his release marked the beginning of policies leading to the eventual release of tens of thousands of Africans in the Crown Colony. The majority of Africans on this vessel were males, with 35 classified as adults ranging in age from 14 to 34. Ten males classified as boys ranged in age from ten to thirteen and the tallest among them was Baaree, identified as recaptive 45, entered in the register as 4 feet 8¾ inches.39 Although more than one- fifth of the Africans on this vessel were children, this was still significantly lower than the average proportion of children among the enslaved in the first quarter of the nineteenth century.40 Secree and Sochra, both aged ten, were girls of 4 feet 3 inches and 4 feet 7 inches respectively. Anta, the third girl, was aged one and described as a EXTENDING THE AFRICAN NAMES DATABASE 145

“sucking child … daughter of Adam No. 56.”41 As the ship had departed Africa in August 1808, she had been embarked as a nine-month old baby with her mother aged 26.42 The Africans listed in the early registers include individuals released as a result of a number of controversial attacks on coastal slave factories by royal naval patrols. As these were mostly in places close to Sierra Leone, the individuals released reflected a narrower geographical range than among slaves liberated from transatlantic vessels. As the Rio Pongo and the Iles de Los were outside British jurisdiction, the raids were of dubious legal authority.43 Recaptives numbered from 4,684 to 4,923 were “seized at Rio Pongus by the Expedition 1814.” A number of family relationships are highlighted in the description column. Amburee, a boy of 3, identified as recaptive number 4,853 was described as the “son of Sally no. 4874.” Tom, a boy of 2 years of age, was described as the “son of Peggy no. 4877.” A number of Africans of advanced age were listed among this group, and included Tangba a woman of 70, and Sera and Maria, women both aged 66.44 Two infants and their mothers were among a group of Africans “Received from the Isle de Los” in April 1814. Following the introduction of Courts of Mixed Commission at Sierra Leone in 1819, the names of Africans released at Freetown also begin to appear in a second series of registers in Foreign Office Papers (FO84) in the National Archives at Kew. The first of the African recaptives to appear both in the registers at Sierra Leone and in the series at Kew was Boorah, a man age 23 embarked at Little Bassa on the Windward Coast and landed at Freetown on 14 September 1819 after the NS de Regla was condemned for being in breach of international treaties.45 The identification number he was ascribed in the Freetown register was 12,179, although this numbering sequence was not reproduced in the registers at Kew.46 Boorah was presumably among some of the earliest recaptives to walk up the Old Wharf Steps at Freetown constructed the previous year.47 He was 5 feet 2½ inches in height and was “marked on the right side of [his] belly” and had “a large hole in the left ear.” Fabiana, a Spanish schooner seized at sea, was the 146 SUZANNE SCHWARZ next vessel condemned on 7 October 1819 for “breach of the treaty,” and is among the first group of vessels to be listed in both the Freetown registers and the series at Kew.48 In the ninth register of Liberated Africans spanning the period from February 1819 to October 1822, the first 270 names listed are not currently included in the African Names Database. Most of these were disembarked from the Sylphe, a Portuguese schooner, delivered to Freetown on 19 February 1819.49

III

The anti-slave trade patrols succeeded in releasing only a small fraction of the total number of Africans embarked on slave vessels in the nineteenth century.50 Between 1808 and 1863, only 198,710 out of a total of 3.2 million Africans embarked as slaves were released (6.2%). The proportion was lower still in the early years of suppression activity, as only 3.9% of enslaved Africans embarked on transatlantic vessels were released between 1808 and 1817 (23,239 of 603,000 Africans embarked).51 This reflects the small number of royal naval vessels stationed off the African coast, which numbered just two to four ships in the period between 1808 and 1815.52 Enslaved Africans released at Freetown in the first decade of suppression activity were taken off ships which had embarked slaves in seven of the eight main areas of coastal embarkation used in the Slave Voyages database. Of 110 vessels adjudicated by the Vice-Admiralty Court between 1808 and 1818,53 the largest groupings of vessels had embarked slaves on the Upper Guinea coast, the Bight of Benin and the Bight of Biafra. Among the 23 vessels which had embarked slaves in the Bight of Benin was the Dois Amigos, part-owned by Francisco Felix de Souza, a Brazilian trader resident in Ouidah.54 The Africans on board this intercepted vessel were embarked at Ouidah in 1816 and, in common with many others embarked at ports in the Bight of Benin, were originally destined for slavery in Brazil.55 Twenty of the vessels adjudicated in this period had embarked slaves in the Bight of Biafra. Grouping together vessels from Senegambia, Sierra Leone and the Windward Coast suggests that the largest number of EXTENDING THE AFRICAN NAMES DATABASE 147 intercepted vessels (29 of 110) had embarked slaves on the Upper Guinea Coast.56 None of the vessels condemned between 1809 and 1817 had obtained slaves at Bunce Island or the Iles de Los, areas which had dominated slave supply before 1807.57 This reflects the way in which suppression activities focused on the Sierra Leone estuary displaced the trade from the colony’s immediate hinterland and re-orientated slave supply to areas north and south of the colony.58 In 1812, British commissioners reported that the patrols had managed to cause sufficient “interruptions and annoyance” to disrupt slave exports from the rivers in the colony’s immediate hinterland.59 The astute adaptation of slave traders resulted in a shift in trade to areas which were more difficult for royal naval vessels to police and control.60 Six of the vessels adjudicated before the Vice-Admiralty Court had embarked Africans in West Central Africa, another important supply area for Brazil.61 Three vessels, the Donna Mariana, the Ana, and the Triunfo Africano, had embarked slaves on the Gold Coast.62 The regional designations of coastal areas used in the Slave Voyages database do not indicate, however, where the enslaved originally came from in Africa.63 There is some comparative data for Sierra Leone from the early 1820s based on Church Missionary Society (CMS) registers which document the origins of 638 recaptive schoolchildren in Sierra Leone. “Cosso” or Mende from the Upper Guinea coast constituted one of the largest groupings in the registers (16.9%), which is consistent with the high proportion of intercepted vessels from this area. Smaller groupings from the Upper Guinea coast in the CMS registers included Fula, Mandinka, Sherbro, Susu and Temne.64 Large groupings of children in the registers were identified as Igbo (15.7%), Ibibio/Efik (6.7%) and “Accoo” or Yoruba (6.9%). This indicates that Yoruba were numerically less significant than in Sigismund Wilhelm Koelle’s mid-nineteenth century survey, which revealed that approximately half of the recaptives “spoke dialects of Yoruba.”65 Jones argues that “in general, the figures seem to suggest that between the early 1820s and late 1840s the relative importance of 148 SUZANNE SCHWARZ

Yoruba- and Ewe-speaking groups in the Colony grew rapidly, while that of Igbo and of ‘nations’ from near Sierra Leone declined.”66 The potential for using the African names in the Registers of Liberated Africans as a way of tracing the geographic origins of enslaved Africans before they reached their coastal point of departure has been explored by G. Ugo Nwokeji and David Eltis for the Cameroons, and by Misevich in a study of slaves originating from the Upper Guinea coast.67 This method depends on name recognition by teams of African linguists and an ability to link the names to current areas of usage. This poses a number of methodological complexities, as the extent to which the verbal rendition of names by African recaptives resulted in an accurate representation of their form and sound in the registers is problematic. Even though details of their names and age were supplied by the recaptives, the information was conveyed through an African interpreter and entered into the register by an English speaking clerk who, unfamiliar with African pronunciation and linguistic forms, relied on different phonetic representations of the names.68 The methodology is also complicated by the existence of “multi-ethnic” names and Islamic names. However, in his study of recaptives originally embarked in Sierra Leone, Misevich is confident that the recorded names show such a clear association with those in common use in modern-day Sierra Leone that they can be used to identify the “interior origins of Africans exported from Sierra Leone.”69 There is scope to compare Misevich’s database of over six and a half thousand African names from Sierra Leone with the names of recaptives in the Vice-Admiralty registers released from vessels which had embarked slaves on the Upper Guinea coast. For the purposes of this article, I will undertake a preliminary comparison of Misevich’s database with the names of recaptives released from two vessels which had embarked slaves in different areas of the Upper Guinea coast. The Maria Josefa was one of five intercepted vessels that had embarked slaves in the Galinhas between 1808 and 1818.70 This was an “important export point … located about 150 miles leeward of the Sierra Leone peninsula” in the south-east of modern Sierra Leone.71 Out of a EXTENDING THE AFRICAN NAMES DATABASE 149 total of 63 Africans embarked in the Galinhas in 1814, 62 were disembarked at Freetown.72 Of the 62 names listed in the register, it is possible to identify three-quarters in Misevich’s database. Discarding ten names that were multi-ethnic and in common use throughout Sierra Leone, leaves 35 individuals for whom it is possible to identify origins. Almost half of the names can be identified as Mende or Sherbro, which indicates that many of the slaves embarked on this vessel were drawn from supply areas between 50 and 75 miles from the coast. Names that are identifiably Mende include Dugba and Yako, and names classified as Sherbro include Sese and Choma.73 This predominance of Mende and Sherbro names from southern Sierra Leone is entirely consistent with the pattern Misevich found in his later study of recaptives recorded in the Freetown and Havana registers, and it is also reflects the patterns of origin of Africans on board the Amistad. If the names of Temne and Vai origin are combined with the Mende and Sherbro names, this indicates that two-thirds were drawn from an area stretching approximately sixty miles inland from the coast. This is identical to the pattern found in Misevich’s study based on the later registers and he concludes that the area’s “shallow slaving frontier” contrasts with longer distances associated with other areas of slave embarkation.74 The Pennel was one of seven intercepted vessels that had embarked slaves in the the Rio Pongo between 1808 and 1818.75 This vessel which commenced its transatlantic voyage in Britain embarked 18 slaves at the Rio Pongo in 1809, all of whom were released at Sierra Leone. The limitations of name recognition are clear in the case of this vessel, as it is possible to identify the origins of only ten of the names listed. Although the importance of Muslim influence is reflected in the presence of six names of Islamic origin, including Mousa, Mamadoo, and Brima, these names are clearly not distinctive to any specific region. Given their presence on ships embarking from the Rio Pongo, it is possible that these individuals originated from the Futa Jalon, the main Islamic state in the adjacent interior.76 As early as 1793, employees of the Sierra Leone Company displayed an awareness of the importance of the Futa Jalon as an area of slave supply to the Rio Pongo. An expedition to the Futa Jalon 150 SUZANNE SCHWARZ in 1793 was undertaken as a means of developing a long-distance trade in “legitimate” commodities, but also as a means of persuading Muslim leaders to abandon slave trading and develop the natural resources of their land.77 In 1814, six years after the formation of the Vice-Admiralty Court, efforts were still being made to persuade Muslim traders to abandon slave trading. An abolitionist medal commissioned by Zachary Macaulay from the Soho Mint in Birmingham was intended to deter Muslim traders from bringing their slaves to the coast for sale into the transatlantic slave trade. This medal intended for distribution in Sierra Leone featured an Arabic inscription on the reverse stating that the “sale of slaves prohibited in 1807, Christian era, in the reign of George the Third; verily we are all brothers.”78

IV

Evidence relating to the “disposal” of these early recaptives is recorded in a number of the registers. Although there are substantial gaps in coverage, the entries still provide a valuable glimpse of the subsequent life histories of the first cohorts of recaptives to experience this forced migration to Sierra Leone. The register containing entries for recaptives numbered from 10,115 to 15,143 provides information on the allocation of individuals to different colony villages and areas of settlement, including Kissey Town, Regent Town, Charlotte Town, Bathurst Town and Kent. Cudjoe, a boy aged 8 entered as recaptive number 10,204, was among those allocated to the “C[hristian] M[issionary] Society’s Christian Institution Leicester Mountain.” This evidence provides a basis for reconstructing settlement patterns in the early phases of suppression activity, and as recaptives from different vessels were often sent in groups to different areas it also sheds light on features of community formation among Africans drawn from different provenance zones of the slave trade. The Register of Liberated Africans for 1812-1814 includes some reference to the occupations of individuals, but this information is unevenly distributed in the register. Pencil notations on the “disposal” of some recaptives are included in the register for 1814 to 1815. For EXTENDING THE AFRICAN NAMES DATABASE 151 example, the letters RAC are entered against the names of recaptives numbered from 4,765 to 4,799 indicating that they had been enlisted into the Royal African Corps. Other entries record references to occupations, and the death of individuals. The information on “disposal” contained in the first Register of Liberated Africans for 1808 to 1812 is particularly rich, and includes comments on the apprenticeship of recaptives. A large proportion of those for whom information is available were entered into the Royal African Corps or apprenticed within the colony. A single line in the register describes how African males numbered 359 to 389 embarked in the Rio Pongo and released from the American vessel the Lucia (a) Albert in 1810 had all “Entered His Majesty’s Land Service as soldiers in the Royal African Corps.” These included Balla a man of 29 whose name points to Koronko origins and Bara, whose name indicates Temne origins.79 Service at sea was the fate of Sennama, Jeddo, Maca and Coomba, four boys aged between 10 and 11 who were released from the Lucia (a) Albert in 1810, and “Entered on board His Majesty’s Ship Crocodile.”80 Sara, a boy of 6, whose name indicates Limba origins, was apprenticed to Lieutenant Scott of “His Majesty’s Ship Myrtle.”81 A number of recaptives were entered for service on board ships engaged in anti-slave trade patrols. Tom, a boy aged eight released at Freetown on 24 November 1808, was subsequently entered on board His Majesty’s sloop Derwent. Thong, a boy aged eight, was also entered on board this vessel in the command of Captain Frederick Parker, together with a man and a boy aged eight released from the Two Cousins on 25 November 1808.82 Other recaptives took the initiative to leave the colony and return home. Miah, a man aged 25, had “returned to his country at his own Request” and Banna was amongst a number of individuals who had “returned to their country at their own request.”83 The terms of the Act for the Abolition of the Slave Trade in 1807 made provision for the apprenticeship of recaptives and their enlistment into “His Majesty’s Land or Sea Service, as Soldiers, Seamen or Marines...”84 In October 1809, correspondence in Colonial Office papers included “suggestions for the disposal of the Captured negroes at Sierra 152 SUZANNE SCHWARZ

Leone.” Various recommendations were set out in response to the question “What course ought in future to be pursued with respect to the Slaves taken and condemned as prize at Sierra Leone?” It was argued that the best course for boys and girls released from slave ships was to bind them as apprentices, as this offered the opportunity of admitting them to “domestic intercourse with persons more civilized than themselves...” It was recommended that a small number of married adult males who possessed “a sufficient degree of intelligence to undertake at once the culture of a small farm” should be provided with a few acres of land. The “rest of the Men, with such exceptions, and limitations, as it shall appear proper to make may … be enlisted into His Majesty’s Service; and after the Garrisons of Sierra Leone, Goree, Cape Coast Castle &c. are supplied with an adequate number of men, the residue may be drafted, from time to time, into the West India Regiments.”85 Although fierce contemporary controversy surrounded the apprenticeship of recaptives on the basis that it represented another form of enslavement, it appears that the existing Nova Scotian and Maroon settlers saw apprenticeship as an opportunity to acquire additional labour.86 Thomas Johnson, described as Colonel of the Maroons, was amongst those who took as apprentices males and females released from the Marie Paul in 1808. Barre, a boy of 13, identified as recaptive number 39 was apprenticed to Johnson, and Scipio next in the list was apprenticed to George Ferrie, a gaoler. Two boys named Samba, aged 10 and 11 were both apprenticed to yeomen, and this may relate to the African name Sambo which suggests Fula origins as a name given to second sons.87 Phillis Hazeley, one of the first female settlers to take an apprentice from among the recaptives, took Barka a woman of 14. Sambo, a male aged 16 released from Lucia (a) Albert and Seaka, a male aged 14, were apprenticed to Martha Burden, a Nova Scotian shopkeeper who owned a large property on the corner of Wilberforce Street, and after her death they were transferred to her son-in-law.88 Despite his impassioned exchanges with Zachary Macaulay and William Wilberforce on the immorality of apprenticing recaptives, Thomas Perronet Thompson, the first governor of the Crown Colony, is recorded EXTENDING THE AFRICAN NAMES DATABASE 153 as apprenticing Sybell, a girl aged four from the San Joaquim.89 This reflected his obligation to enforce the provisions of Articles VII and XVI of the Act for the Abolition of the Slave Trade.90 Thompson also took a female apprentice Maradah, aged 16 from the Pennel, in 1809.91 Dr. Robert Thorpe, who launched a vitriolic attack on Macaulay and Wilberforce for sustaining slavery in the colony, is also recorded as receiving recaptives as apprentices.92 A number of entries give personal information about the recaptives which is not obtainable from other sources. In the case of Fatima, a woman of 24 identified as recaptive number 55, the register notes that she was “married to a Joliff man (a Moor).”93 A woman of 26 released from the ship Lucia (a) Albert was described as living with “a Foulah Man settled in the Colony.”94 Another woman aged 25 released from the same ship was noted to be living with “Jim Maloolin, a Foulah Man settled in the Colony.”95 Others were simply recorded as being “settled in the colony” or working with their countrymen. An African woman, identified as recaptive number 453 from the Lucia (a) Albert was described as “living with Mousa Kenta, No 273 in this Register.” The entry for “Musah Kenta” indicates that he was aged 20 and had been released from the schooner Cuba (a) Marianha a year earlier.96 A further list of recaptives which has come to light in Colonial Office correspondence at the National Archives at Kew provides information on the first 1,991 Africans landed at Freetown between 1808 and 1812 and adds significantly to the biographical information about these uprooted individuals.97 The numbering of Africans in the list corresponds directly with the first surviving register at Freetown, and nominal linkage between the two sources makes it possible to trace the subsequent movements of a large number of those landed. Of the enslaved Africans disembarked from the Marie Paul in 1808, only about half of the original number remained in the colony by the time the listing was completed. The entries indicate that 29 recaptives still living in the colony were pursuing a number of different occupations. These included Masamba, a man of 25, described as “a mason living in the colony” and Karafa, a “labourer living in the colony.” Two of the men, Yoro and 154 SUZANNE SCHWARZ

Barrick, had been “entered into the Royal African Corps.” No entries are included for 27 of the Africans originally listed in the Register of Liberated Africans, and a number of these appear to have left their apprenticeships.98 Sochra, a girl aged ten had been apprenticed to Mary Brown, a Maroon, in 1808 but by the time the list was compiled she was absent from the colony. Similarly, Barka (recaptive number 57) was apprenticed to the schoolmistress Phillis Hazeley in 1808 but was no longer present in the colony four years later. These gaps in numbering in the list are explained by how “those Captured Negroes that are not accounted for have deserted to Native Towns in the back parts of the country.” Sinnaba, a woman aged 17 identified as recaptive number 53 released from the Marie Paul in 1808, had “returned to her country.” Recaptives from other vessels also left the colony. This was a form of active resistance to their enforced migration to Sierra Leone, and their subsequent apprenticeship. Both the register of Liberated Africans and the list in Colonial Office records agree that Maria, recaptive number 202, had “Returned to the Croo country.”99 For those Africans enslaved in the hinterland of Sierra Leone, there was a greater opportunity to return home compared to those who had been enslaved further afield in the Bight of Benin or the Bight of Biafra, although many recaptives did return home to Yorubaland in the Bight of Benin from 1838 onwards.100

V

As a result of the Endangered Archives pilot project, a diverse range of manuscript sources relating to liberated Africans in Sierra Leone has been digitized and deposited in the Public Archives of Sierra Leone. The further award of a major grant by the British Library Endangered Archives project in 2011 will facilitate the on-going digitization of nineteenth-century sources by Sierra Leonean archivists. In addition to extending the scope of the African Names Database, the analysis of these early Registers of Liberated Africans from the Vice-Admiralty Court opens up the potential to trace the origins of some of the first individuals released at Freetown, as well as their later destinations. As the transfer of EXTENDING THE AFRICAN NAMES DATABASE 155 the colony to the Crown in 1808 resulted in the development of a more bureaucratic and carefully regulated administrative system, there is a wealth of documentary evidence which makes it possible to reconstruct the lives of African recaptives individually and collectively. For those recaptives apprenticed to colonial officials and Nova Scotian and Maroon settlers, the surviving records make it possible to locate them by reference to households, streets and individual land plots. There is considerable potential for tracing the life history of recaptive populations in the colony, including their medical history, by reconstituting information in the civil registers in conjunction with surviving parish registers. A methodological problem in this process of nominal linkage, however, is that many of the recaptives adopted or were ascribed European names. As in the CMS school registers analysed by Adam Jones, there is no systematic listing of African names and adopted names for individuals.101 There are some individual exceptions. Recaptive number 7 from the Marie Paul was identified as Sank in the Register of Liberated Africans, but in the “List of Captured Negroes” he was identified by the name Tom Wilson. In common with Samuel Ajayi Crowther, other recaptives kept or subsequently resumed their African names.102 Clearly, this reconstruction of biographical information for up to 100,000 recaptives released at Sierra Leone represents a huge agenda for research and highlights the importance of preserving the endangered archives of Sierra Leone.103

* An earlier version of this paper was presented at the Transatlantic Slave Voyages Database Workshop at the Harriet Tubman Institute for Research on the Global Migrations of African Peoples at York University, Toronto, on 3 May 2010. I would like to thank Katrina Keefer, Martin Klein, Robin Law, Henry Lovejoy, Paul Lovejoy, Philip Misevich and Silke Strickrodt for their helpful comments and suggestions on this paper. I would also like to thank Albert Moore, Senior Government Archivist in Sierra Leone, and his staff for their help and guidance during visits to the Public Archives of Sierra Leone. I am grateful to Philip Misevich for his generosity in allowing me to consult his extensive database of African names from Sierra Leone.

156 SUZANNE SCHWARZ

1 David Eltis, Stephen D. Behrendt, David Richardson, and Herbert S. Klein, The Trans-Atlantic Slave Trade. A Database on CD-Rom (New York: Cambridge University Press, 1999), 1-2. 2 The outlook of James Irving, a Liverpool slave ship captain and surgeon, was probably typical of other men engaged in the trade, as he described 526 Africans on board the Jane in 1786 as “disagreeable Cargo” and “Black Cattle.” Suzanne Schwarz, ed., Slave Captain. The Career of James Irving in the Liverpool Slave Trade, 2nd edition (Liverpool: Liverpool University Press, 2008), 22, 85-87. 3 Marcus Wood, Blind Memory. Visual Representations of Slavery in England and America 1780-1865 (Manchester: Manchester University Press, 2000), 27- 29. 4 The African Origins Project is a “scholar-public collaborative endeavour to trace the geographic origins of Africans transported in the transatlantic slave trade.” www.slavevoyages.org/tast/about/origins.faces 5 African Names Database, www.slavevoyages.org/tast/resources/slave.faces 6 Between 1501 and 1867, an estimated 12,521,000 Africans were transported into slavery. David Eltis and David Richardson, Atlas of the Transatlantic Slave Trade (New Haven and London: Yale University Press, 2010), xvii, 19. 7 Wood, Blind Memory, 27. 8 Recent analysis of the Havana registers indicates that a total of 10,391 Africans were released from 42 vessels by the Court of Mixed Commission. Henry B. Lovejoy, “The Shipping Registries of the Archives of the Havana Slave Trade Commission: Transcription Methodology and Statistical Analysis,” Paper presented at the Transatlantic Slave Voyages Database Workshop, Harriet Tubman Institute for Research on the Global Migrations of African Peoples, York University, Toronto, 3 May 2010. 9 The Kew series of registers relates to the work of the Courts of Mixed Commission from 1819 onwards. Anglo-Portuguese, Anglo-Spanish and Anglo- Dutch commissions established at Sierra Leone in 1819 replaced the British Vice-Admiralty Court. An Anglo-Brazilian Commission was added to those sitting in Freetown in 1828. Leslie Bethell, “The Mixed Commissions for the Suppression of the Transatlantic Slave Trade in the Nineteenth Century,” Journal of African History 7:1 (1966), 79-82; Christopher Fyfe, A History of Sierra Leone (London: Oxford University Press, 1962), 137-138. For a discussion of the work of the Vice-Admiralty Court and the legal framework within which it operated, see Tara Helfman, “The Court of Vice Admiralty at Sierra Leone and the Abolition of the West African Slave Trade,” The Yale Law Journal, 115 (2006), 1122-1156. 10 British Library Endangered Archives, Pilot Project EAP 284. 11 Recaptives numbered from 100 to 248 are identified as “slaves seized in the colony.” The second register also lists Africans “seized in the colony”, including three individuals numbered from 4,679 to 4,681. Public Archives of Sierra EXTENDING THE AFRICAN NAMES DATABASE 157

Leone, Register of Liberated Africans 1808-1812 and Register of Liberated Africans 1812-1814 [hereafter PASL, RLA]. 12 PASL, RLA, 1808-1848. 13 PASL, RLA 1808-1812, 1812-1814. 14 Adam Jones, “Recaptive Nations: Evidence Concerning the Demographic Impact of the Atlantic Slave Trade in the Early Nineteenth Century,” Slavery and Abolition 11:1 (May 1990), 42-57. 15 This vessel embarked 522 Africans at Lagos, of whom 517 were disembarked at Freetown. www.slavevoyages.org, voyage 3680. 16 As indicated earlier, the Registers of Liberated Africans at Freetown include Africans released from coastal slave barracoons, as well as slaves seized in the colony. The additional 9,909 entries identified by Meyer-Heiselberg, for example, include 1 woman, 17 boys and 4 girls released from a “cano captured by Thomas Ellis” and landed at Freetown on 2 October 1851. As a result, a straightforward numerical comparison with the totals of Africans disembarked from slave ships bound for the Americas at Sierra Leone in the Slave Voyages Database is not possible. R. Meyer-Heiselberg, Notes from Liberated African Department in the Archives at Fourah Bay College Freetown, Sierra Leone (Uppsala: The Scandinavian Institute of African Studies, 1967), 55-61. In the period between 1846 and 1863, 53 vessels were adjudicated by the Vice- Admiralty Court at Freetown after the work of the Courts of Mixed Commission came to an end. The cases accounted for approximately 15,000 liberated Africans. I am grateful to David Eltis for this information supplied in email correspondence. 17 Email correspondence with David Eltis, Paul Lovejoy and Philip Misevich, August-September 2010. 18 During visits to Sierra Leone in February and August 2010, members of the Endangered Archives team retraced the entire series of registers from 1808 to 1848 with the exception of three volumes spanning recaptives numbered 25,423- 30,708, 43,538-50,761 and 67,636-75,356. Richard Anderson subsequently located the volume spanning recaptives numbered 25,423-30,708. 19 Philip Misevich, “The Origins of Slaves Leaving the Upper Guinea Coast in the Nineteenth Century,” in David Eltis and David Richardson, eds., Extending the Frontiers. Essays on the New Transatlantic Slave Trade Database (New Haven/London: Yale University Press, 2008), chapter 5. 20 In contrast to the Court of Mixed Commission registers from Havana, these early registers from Freetown do not usually provide the Christian names of the recaptives. Lovejoy, “Shipping Registries.” 21 See, for example: TNA, CO 267/31, “List of Captured Negroes on Hand December 31st 1810 and of those Received, Enlisted, Apprenticed, Disposed of to December 31st 1812.” 158 SUZANNE SCHWARZ

22 Fyfe, History, 107; Daniel Augustine Vonque Stephen, “A History of the Settlement of Liberated Africans in the Colony of Sierra Leone During the First Half of the 19th Century,” MA thesis, University of Durham, 1963, II-III; Meyer-Heiselberg, Notes from Liberated African Department, I-XII, 1-11. Most recently, Gibril R. Cole has drawn attention to the Registers; see "Re-thinking the Demographic Make-up of Krio Society," in Mac Dixon-Fyle and Gibril Cole, eds., New Perspectives on the Sierra Leone Krio (New York: Peter Lang, 2006), 43. 23 The Public Archives of Sierra Leone were established by the Sierra Leone Public Archives Act in 1965. In the following year, a UNESCO report highlighted the importance of the surviving records relating to the liberated Africans. Albert S. Moore, “The Role of Archives in National Development: Sierra Leone Public Archives. A Case Study,” BA dissertation, University of Sierra Leone (1993), 9-14. 24 This vessel embarked 216 individuals principally at Bonny in the Bight of Biafra, of whom 184 arrived at Sierra Leone. www.slavevoyages.org, voyage 2351. 25 The sequence of registers includes volumes spanning entries numbered 1- 3,772; 3,773-6,288; 6,289-8,528; 4,684-7,507; 8,529-9,758; 7,508-9,758; 9,759- 11,908; 10,115-15,143 and 11,909-15,967. Meyer-Heiselberg, Notes from Liberated African Department, I, IV. 26 Inventory of Endangered Documents in the Public Archives of Sierra Leone, British Library Pilot Project Code EAP 284. 27 See below, section IV. 28 G. Ugo Nwokeji and David Eltis, “Characteristics of Captives Leaving the Cameroons for the Americas, 1822-37,” Journal of African History 43 (2002), 192, 209; Lovejoy, “Shipping Registries.” 29 PASL, RLA 1808-1812, recaptive 2,669. 30 PASL, RLA 1808-1812, recaptive 1,030. 31 Lovejoy notes how “there seems to have been a basic distinction between people from West Africa, who practiced scarification, with some qualifications, and people from west central and southeast Africa who did not, again with some qualifications.” Paul E. Lovejoy, “Scarification and the Loss of History in the African Diaspora,” in Andrew Apter and Lauren Derry, eds., Activating the Past Historical Memory in the Black Atlantic (Newcastle: Cambridge Scholarly Publishing, 2010). Katrina Keefer at York University, Toronto is currently undertaking research on patterns of scarification in the Sierra Leone registers. 32 This process continued in the Americas, as the “marks of slavery” replaced ritual forms of scarification. Facial and body scarification was “virtually absent among the creole populations of enslaved America.” Lovejoy, “Scarification and the Loss of History”. EXTENDING THE AFRICAN NAMES DATABASE 159

33 The area of slave embarkation is unknown. www.slavevoyages.org, voyage 7570. 34 PASL, RLA 1808-1812, recaptives 1,891, 2,016-2,022, and 2,035. 35 PASL, RLA 1808-1812, recaptive 27. 36 Senegal was listed as recaptive number 1,017. PASL, RLA 1808-1812, recaptives 313, 1,006 and 1,017. 37 This loss by mortality of 13% was slightly lower than the average rate of 18% for vessels adjudicated by the Vice-Admiralty Court at Sierra Leone between 1808 and 1818. www.slavevoyages.org 38 After embarking 308 Africans in the Bight of Biafra, the Santana de Africa spent just over seven weeks in the Middle Passage before arriving at Freetown on 21 August 1815. Only three Africans (1%) survived and were listed in the Vice-Admiralty register as recaptives numbered from 7,802-7,804. These three men were aged 16, 18 and 19. www.slavevoyages.org, voyage 7638; PASL, RLA 1815-1816. 39 The age of children was usually estimated on the basis of height. As Lovejoy points out, “the designation of slaves as ‘children’ is often unclear, but is sometimes assumed to be pre-pubescent, and hence roughly before age 13-14 and certainly before mid-teens.” Paul Lovejoy, “The Children of Slavery – the Transatlantic Phase,” Slavery and Abolition 27:2 (August 2006), 198-199; Audra A. Diptee, “African Children in the British Slave Trade During the Late Eighteenth Century,” Slavery and Abolition 27:2 (August 2006), 186. 40 www.slavevoyages.org, voyage 7539; Lovejoy, “Children of Slavery,” 200- 202. 41 PASL, RLA 1808-1812, recaptives 58-60. 42 The number of infants on board slave ships was usually very small. Lovejoy, “Children of Slavery,” 205-207; Diptee, “African Children,” 186, 190. 43 Fyfe, History, 120-122; Helfman, “The Court of Vice Admiralty,” 1145-1149. 44 PASL, RLA 1812-1814, recaptives 4,905, 4,909 and 4,911. 45 www.slavevoyages.org, voyage 2314. 46 In the Kew series, the numbering sequence begins at 1 in 1819 and extends to 56,935 in 1845. I am grateful to David Eltis for this information contained in email correspondence, May 2010. 47 A. Archer Betham, “The Old Wharf Steps, Freetown,” Sierra Leone Studies. A Reprint of Some of the Articles from the First Twenty Two Volumes Bearing on the Work of the Monuments and Relics Commission (The Government Printer: Freetown, 1953), 1-4; Fyfe, History, 134. 48 www.slavevoyages.org, voyage 2315. 49 Recaptives numbered 11,909-12,107 in the register for 1819-1822 were released from the Sylphe. This was a continuation of the list of recaptives (11,744-11,908) from this vessel contained in the previous register for 1816- 1819. In total, 364 Africans were released at Freetown. This schooner had 160 SUZANNE SCHWARZ

commenced its voyage in Guadeloupe and purchased 388 slaves at Bonny in the Bight of Biafra. PASL, RLA 1819-1822; www.slavevoyages.org, voyage 7675. 50 Robert Burroughs, “Eyes on the Prize: Journeys in Slave Ships Taken as Prizes by the Royal Navy,” Slavery and Abolition 31:1 (March 2010), 99-104. 51 Eltis and Richardson, Atlas, 273-275. 52 A.W.H. Pearsall, “Sierra Leone and the Suppression of the Slave Trade,” Sierra Leone Studies, New Series, 12 (1959), 211-213, 219-229. 53 The number of cases dealt with by the Vice-Admiralty Court was higher, but not all of the vessels carried enslaved Africans. Some of the cases also related to “slaves seized in the colony.” The National Archives, HCA 49/97/1, “Vessels, Cargoes & Slaves Proceeded against in the Court of Vice Admiralty at Sierra Leone Between June 1808 & March 1817.” 54 www.slavevoyages.org, voyage 7597; Robin Law, “Francisco Felix de Souza in West Africa, 1820-1849,” in José C. Curto and Paul E. Lovejoy, eds., Enslaving Connections: Changing Cultures of Africa and Brazil During the Era of Slavery (New York: Humanity Books, 2004), 187, 192. 55 Cuba was the other major destination for Yoruba and other enslaved Africans embarked from ports in the Bight of Benin; Lovejoy, “Scarification and the Loss of History.” Africans embarked in the Bight of Benin would have included individuals who spoke Gbe languages. In addition, the enslaved Africans would have included Yoruba and “an identifiable Muslim population from the far interior.” Yoruba who were disembarked in Brazil become known as “Nagô”, whilst speakers of Gbe languages from the Bight of Benin were known as Gege or Mina. Curto and Lovejoy, eds., Enslaving Connections, 12. See also Robin Law, “Ethnicity and the Slave Trade: ‘Lucumi’ and ‘Nago’ as Ethonyms in West Africa,” History in Africa 24 (1997), 205-219; “Ethnicities of Enslaved Africans in the Diaspora: On the Meanings of ‘Mina’ (Again),” History in Africa, 32 (2005), 247-267. 56 The blockading strategies adopted by royal naval patrols influenced the origins of the recaptives released at Sierra Leone. In the period between 1808 and 1827, the naval vessels “patrolled most frequently along the ‘Windward Coast’ on either side of Sierra Leone.” Philip D. Curtin and Jan Vansina, “Sources of the Nineteenth Century Atlantic Slave Trade,” Journal of African History 5:2 (1964), 187. 57 Misevich, “Origins of Slaves,” 161. 58 Allen M. Howard, “Nineteenth-Century Coastal Slave Trading and the British Abolition Campaign in Sierra Leone,” Slavery and Abolition 27:1 (April 2006), 27-28. 59 House of Commons Parliamentary Papers Online, Extracts from the Report of the Commissioners Appointed for Investigating the State of the Settlements and Governments on the Coast of Africa (1812), 3. EXTENDING THE AFRICAN NAMES DATABASE 161

60 Howard, “Nineteenth-Century Coastal Slave Trading,” 27-28; Philip R. Misevich, “On the Frontier of ‘Freedom’: Abolition and the Transformation of Atlantic Commerce in Southern Sierra Leone, 1790s to 1860s,” Ph.D thesis, Emory University, 2009, 4, 19, 60-61. 61 Curto and Lovejoy, eds., Enslaving Connections, 12. 62 www.slavevoyages.org, voyages 7552, 7546 and 7527. 63 Paul E. Lovejoy, “Extending the Frontiers of Transatlantic Slavery, Partially,” Journal of Interdisciplinary History 40:1 (Summer, 2009), 58-59, 64-67. 64 Jones, “Recaptive Nations,” 50-52. 65 Jones, “Recaptive Nations,” 46. 66 Jones, “Recaptive Nations,” 52. 67 Nwokeji and Eltis, “Characteristics of Captives,” 191-210; Misevich, “Origins of Slaves,” 155-175. 68 Misevich, “Origins of Slaves,” 158, 164; Nwokeji and Eltis, “Characteristics of Captives,” 192, 196, 208-209. 69 Misevich, “Origins of Slaves,” 157-158. 70 The vessel was restored on 12 August 1809. The other four vessels were the Resurreccion, Nueva Paz, Rosa and Dos de Mayo. www.slavevoyages.org, voyages 7523, 7532, 7594, 7562, 46559. 71 Howard, “Nineteeenth-Century Coastal Slave Trading,” 27. See also Adam Jones, From Slaves to Palm Kernels: A History of the Galinhas Country (West Africa) 1730-1890 (Wiesbaden: F. Steiner, 1983). 72 PASL, RLA 1812-1814 and 1814-1815, recaptives 5,926-5,987. 73 This information about origins of names is based on Misevich’s database. 74 Misevich, “On the Frontier of ‘Freedom,’” 5, 67, 86-88, 96. In the Cameroons estuary between 1822 and 1837, the names of the recaptives indicated that the “great majority of the captives originated within 200 miles of the coast.” Nwokeji and Eltis, “Characteristics of Captives,” 191, 200-201. 75 The other six vessels were the Eugenia, Joana, Lucia (a) Albert, Triunvirato (a) Dorset, Lucia (a) Rainbow, and the Laberinto. www.slavevoyages.org, voyages 7557, 7565, 7566, 7585, 7609, 46924, 7672. 76 Howard, “Nineteenth-Century Coastal Slave Trading,” 26, 30; Walter Rodney, “Jihad and Social Revolution in Futa Jallon in the Eighteenth Century,” Journal of the Historical Society of Nigeria 4: 2 (1968), 269-284. 77 Bruce L. Mouser, ed., Journal of James Watt: Expedition to Timbo Capital of the Fula Empire in 1794, African Studies Program, University of Wisconsin- Madison, 1994, xiii-xvi, 25-26, 33-34, 54-55. 78 National Maritime Museum, Greenwich, ZBA2808 (www.nmm.ac.uk /visit/exhibitions); Birmingham Museums and Art Gallery, Accession Number 1885N1541.88, Medal Commemorating the Abolition of the Slave Trade 1807. This medal, issued in 1814, was designed by John Phillip and engraved by G.F. Pidgeon. www.bmagic.org.uk/objects/1885N1541.88. 162 SUZANNE SCHWARZ

79 The information about origins of names is based on Misevich’s database. 80 PASL, RLA 1808-1812, recaptives 419-422. 81 PASL, RLA 1808-1812, recaptive 442. 82 PASL, RLA 1808-1812, recaptives 64, 67, 75-76. 83 PASL, RLA 1808-1812. 84 Helfman, “Court of Vice Admiralty at Sierra Leone,” 1143, note 79. 85 TNA, CO 267/31, “List of Captured Negroes on Hand December 31st 1810 and of those Received, Enlisted, Apprenticed, Disposed of to December 31st 1812.” 86 For a discussion of this controversy, see Michael J. Turner, “The Limits of Abolition: Government, Saints and the ‘African Question’, c. 1780-1820,” English Historical Review, 112: 446 (1997), 319-357. 87 I am grateful to Robin Law for this information. PASL, RLA 1808-1812, recaptives 43-44. 88 PASL, RLA 1808-1812, recaptives 401 and 417; Fyfe, History, 102-103. 89 PASL, RLA 1808-1812, recaptive 69. 90 Turner, “Limits of Abolition;” Helfman, “Court of Vice Admiralty at Sierra Leone,” 1143, note 79. 91 PASL, RLA 1808-1812, recaptive 98. 92 Robert Thorpe, A Letter to William Wilberforce … Containing Remarks on the Reports of the Sierra Leone Company, and African Institution: With Hints Respecting the Means by which an Universal Abolition of the Slave Trade Might be Carried into Effect, 4th edition (London: F.C. Rivington, 1815), passim. 93 Jolof in Senegal. The entry indicates that he was Muslim. 94 Fula. 95 PASL, RLA 1808-1812, recaptive 454. 96 PASL, RLA 1808-1812. 97 TNA, CO 267/31, “List of Captured Negroes.” 98 No information is provided about the location of recaptives numbered 4, 13- 15, 17, 21-22, 24-26, 31, 36-41, 43-44, 48-49, 51-52, 56-57 and 59-60. In the case of recaptive number 60, this one year old child had died in March 1809. TNA CO 267/31, “List of Captured Negroes;” PASL, RLA 1808-1812. 99 Kru in Liberia. 100 Jones, “Recaptive Nations,” 52; Curtin and Vansina, “Sources of the Nineteenth Century Atlantic Slave Trade,” 193; Andrew F. Walls, The Missionary Movement in Christian History. Studies in the Transmission of Faith (New York/Edinburgh: Orbis Books/T. & T. Clark, 1996), 105. 101 Jones, “Recaptive Nations,” 48. 102 A.F. Walls, Materials for the Study of Sierra Leone Church History, Unpublished Paper (March 1960). Koelle recorded both the African and European names of many of his informants. P.E.H.,“The Enslavement of Koelle’s Informants,” Journal of African History 6:2 (1965), 193-203. EXTENDING THE AFRICAN NAMES DATABASE 163

103 This research may make it possible to identify further biographical information about 179 recaptives who served as Koelle’s informants for his Polyglotta Africana in the mid-nineteenth century. Hair, “Enslavement of Koelle’s Informants,” 196-203.

CONTRIBUTORS

Daniel B. Domingues da Silva is assistant professor of African history at the University of Missouri, Columbia. He received his Ph.D. from Emory University in 2011. His research focuses on the history of slavery and the transatlantic slave trade between Angola and Brazil. He has received support from the Andrew W. Mellon Foundation and the Calouste Gulbenkian Foundation. He has published in Slavery and Abolition (2008) and Revista Afro-Ásia (2004), and is currently a consultant for Voyages: The Trans-Atlantic Slave Trade Database (www.slavevoyages.org) and the African Origins Portal (www.african- origins.org).

Henry B. Lovejoy is completing his Ph.D. thesis on "Oyo Influence on the Transformation of Lucumi Identity in Colonial Cuba" in the Department of History at the University of California, Los Angeles. He has published "The Transculturation of Yoruba Annual Festivals during the Día de Reyes in Colonial Cuba" in Christopher Innes, Annabel Rutherford and Brigitte Bogar (eds.), Carnival—Theory and Practice (2012) and is a contributing member to The African Origins Project (www.african-origins.org/about). He has been awarded a Post-Doctoral Fellowship from the Social Sciences and Humanities Research Council of Canada for 2012–2014 at the University of British Columbia.

Paul E. Lovejoy is Distinguished Research Professor in the Department of History, York University, and Fellow of the Royal Society of Canada. He holds the Canada Research Chair in African Diaspora History and is Director of the Harriet Tubman Institute for Research on the Global Migrations of African Peoples. He has published 28 books, including Transformations in Slavery: A History of Slavery in Africa (3rd ed., 2011). He is past member of the International Scientific Committee of the UNESCO “Slave Route” Project, Secteur du Culture, and has served as Associate Vice-President (Research) at York University from 1986 to 1990 and as a member of the Social Sciences and Humanities Research Council of Canada from 1990 to 1997, and Vice President in 1996 to 1997. He received a Killam Senior Research Fellowship from the Canada Council in 1994 to 1997 and was Visiting Professor at El Colegio de Mexico in 1999. In 2007, he was awarded an Honorary Doctorate by the University of Stirling, and in 2011 he received the Life Time Achievement Award from the Canadian Association of African Studies and the Teaching Award of the Faculty of Graduate Studies, York University.

Suzanne Schwarz is Professor of History at the University of Worcester, and an Honorary Fellow of the Wilberforce Institute for the study of Slavery and Emancipation at the University of Hull. She currently holds a Leverhulme Research Fellowship, and her recent research on Sierra Leone examines the origins, destinations and experiences of the first cohorts of Liberated Africans released at Freetown in the period between 1808 and 1819. She is working with Paul Lovejoy on a British Library Endangered Archives project to preserve rare and invaluable sources in the Public Archives of Sierra Leone. Her previous publications include Slave Captain. The Career of James Irving in the Liverpool Slave Trade (Liverpool: Liverpool University Press, 2008) and a volume entitled Liverpool and Transatlantic Slavery (Liverpool: Liverpool University Press, 2007), co-edited with David Richardson and Anthony J. Tibbles.

Filipa Ribeiro da Silva studied History and Portuguese overseas expansion at the New University of Lisbon in Portugal, where she obtained her BA honors (1996) and Master degree (2002). She received her Ph.D. at Leiden University in 2009, where she specialized in seventeenth-century Dutch and Portuguese settlement in western Africa. From 2009 to 2011, she held a Post-Doctoral fellowship at the Wilberforce Institute for the study of Slavery and Emancipation (WISE), University of Hull, as a member of the Project: Slave Trade, Slavery and Emancipation in European Histories and Identities financed by the European Union 7th Framework Program. Subsequently, she has been a post-doctoral fellow at the International Institute of Social History of the Netherlands Royal Academy of Sciences, where she is a member of the CLIO-INFRA Project and the Global Collaboratory on the History of Labor Relations. She has published various articles and Dutch and Portuguese in Western Africa: Empires, Merchants, and the Atlantic System (2011).

Stacey Sommerdyk received her M.A. in history from York University in 2007, where she was affiliated with the Harriet Tubman Institute. She received her Ph.D. in history from the University of Hull under the supervision of David Richardson at the Wilberforce Institute for the study of Slavery and Emancipation (WISE). Her thesis, “Trade and the Merchant Community of the Loango Coast in the Eighteenth Century,” reassesses the role of the Dutch in the slave trade. She has published “Rivalry on the Loango Coast: A Re-examination of the Dutch in the Atlantic Slave Trade” in Trabalho forçado africano: o caminho de ida, edited by Arlindo Manuel Calderia et al. Edições Húmus, Ribeirão, Portugal, 2009. Currently, she is a Post-Doctoral Fellow at the Department of History at the University of the Witwatersrand.

Jelmer Vos is Assistant Professor in the Department of History at Old Dominion University. He received his Ph.D. from the School of Oriental and African Studies and was a post-doctoral fellow at Emory University and the Universidade Nova de Lisboa. He is currently preparing a manuscript on the history of the Kongo kingdom, northern Angola, under early colonial rule. His recent publications include "Child Slaves and Freemen at the Spiritan Mission in Soyo, 1880–1885," Journal of Family History (2010) and "Of Stocks and Barter: John Holt and the Kongo Rubber Trade, 1906–10," Portuguese Studies Review (2011).

The Harriet Tubman Series on the African Diaspora

Publications in Association with The Harriet Tubman Institute for Research on the Global Migrations of African Peoples

General Editor

Paul E. Lovejoy FRSC Distinguished Research Professor Canada Research Chair in African Diaspora History Director, The Harriet Tubman Institute York University, Toronto Canada

Paul E. Lovejoy and Toyin Falola, eds., Pawnship, Slavery and Colonialism in Africa, 2003

Donald G. Simpson, Under the North Star: Black Communities in Upper Canada before Confederation (1867), 2005

Paul E. Lovejoy, Slavery, Commerce and Production in West Africa: Slave Society in the Sokoto Caliphate, 2005

José C. Curto and Renée Soulodre-La France, eds., Africa and the Americas: Interconnections during the Slave Trade, 2005

Paul E. Lovejoy, Ecology and Ethnography of Muslim Trade in West Africa, 2005

Naana Opoku-Agyemang, Paul E. Lovejoy and David Trotman, eds., Africa and Trans-Atlantic Memories: Literary and Aesthetic Manifestations of Diaspora and History, 2008

Boubacar Barry, Livio Sansone, and Elisée Soumonni, eds., Africa, Brazil, and the Construction of Trans-Atlantic Black Identities, 2008 Behnaz Asl Mirzai, Ismael Musah Montana, and Paul E. Lovejoy, eds., Slavery, Islam and Diaspora, 2009

Carolyn Brown and Paul E. Lovejoy, eds., Repercussions of the Atlantic Slave Trade: The Interior of the Bight of Biafra and the African Diaspora, 2011

Ana Lucia Araujo, Mariana P. Candido and Paul E. Lovejoy, eds., Crossing Memories: Slavery and African Diaspora, 2011

Ute Röschenthaler, Purchasing Culture in the Cross River Region of Cameroon and Nigeria, 2011

Ehud R. Toledano, ed., African Communities in Asia and the Mediterranean: Identities between Integration and Conflict, 2011

Editorial Board

Edward Alpers, UCLA Carolyn A. Brown, Rutgers Rina Cáceres, Universidad de Costa Rica Myriam Cottias, CNRS Mohammed Ennaji, Université Muhammad V Toyin Falola, University of Texas Naana Opoku-Agyemang, University of Cape Coast Mariza de Carvalho Soares, Universidade Federale de Fluminense Elisée Soumonni, Université Nationale du Bénin Ibrahima Thioub, Université Cheikh Anta Diop Ehud Toledano, Tel Aviv University David V. Trotman, York University