Extending the Frontiers: Essays on the New Transatlantic Slave Trade Database 9780300151749

Citation preview

This page intentionally left blank

Extending the Frontiers Essays on the New Transatlantic Slave Trade Database

Edited by David Eltis and David Richardson

Yale University Press New Haven & London

Published with assistance from the Annie Burr Lewis Fund and from the foundation established in memory of Philip Hamilton McMillan of the Class of 1894, Yale College. Copyright © 2008 by Yale University. All rights reserved. This book may not be reproduced, in whole or in part, including illustrations, in any form (beyond that copying permitted by Sections 107 and 108 of the U.S. Copyright Law and except by reviewers for the public press), without written permission from the publishers. Set in Adobe Garamond and Stone Sans by Binghamton Valley Composition. Printed in the United States of America by Thomson Shore. Library of Congress Cataloging-in-Publication Data Extending the frontiers : essays on the new transatlantic slave trade database / edited by David Eltis and David Richardson. p. cm. Includes bibliographical references and index. ISBN 978-0-300-13436-0 (cloth : alk. paper) 1. Slave trade—Africa— History—Sources. 2. Slave trade—Africa—History—Statistics. 3. Slave trade—Portugal—History—Sources. 4. Slave trade—Brazil—History—Sources. 5. Slave trade—Central America—History—Sources. 6. Slave trade—Europe— History—Sources. 7. Slave trade—America—History—Sources. I. Eltis, David, 1940– II. Richardson, David, 1946– III. Title: Essays on the new transatlantic slave trade database. IV. Title: Essays on the new trans-Atlantic slave trade database. V. Title: Trans-Atlantic slave trade HT1321.E98 2008 306.3'62—dc22 2007043711 A catalogue record for this book is available from the British Library. The paper in this book meets the guidelines for permanence and durability of the Committee on Production Guidelines for Book Longevity of the Council on Library Resources. 10

9 8 7 6 5 4 3 2 1

To Paul Lachance

This page intentionally left blank

Contents

Preface, ix Acknowledgments, xi List of Abbreviations, xiii Map of the Transatlantic Slave Trade, 1501–1867, follows page xiii A New Assessment of the Transatlantic Slave Trade, David Eltis and David Richardson, 1 1.

Part I: Origins and Destinations 2. The Foundations of the System: A Reassessment of the Slave Trade to the Spanish Americas in the Sixteenth and Seventeenth Centuries, António de Almeida Mendes, 63 3. The Slave Trade to Pernambuco, 1561–1851, Daniel Barros Domingues da Silva and David Eltis, 95

The Transatlantic Slave Trade to Bahia, 1582–1851, Alexandre Vieira Ribeiro, 130 4.

vii

viii

Contents

The Origins of Slaves Leaving the Upper Guinea Coast in the Nineteenth Century, Philip Misevich, 155 5.

6. The African Origins of Slaves Arriving in Cuba, 1789–1865, Oscar Grandío Moráguez, 176

Part II: National Slave Trades

The Significance of the French Slave Trade to the Evolution of the French Atlantic World before 1716, James Pritchard, David Eltis, and David Richardson, 205 7.

The Dutch in the Atlantic World: New Perspectives from the Slave Trade with Particular Reference to the African Origins of the Traffic, Jelmer Vos, David Eltis, and David Richardson, 228 8.

The Slave Trade of Northern Germany from the Seventeenth to the Nineteenth Centuries, Andrea Weindl, 250 9.

Part III: Some Wider Consequences and Implications of the New Data

The Slave Trade, Colonial Markets, and Slave Families in Rio de Janeiro, Brazil, ca. 1790–ca. 1830, Manolo Florentino, 275

10.

The Suppression of the Slave Trade and Slave Departures from Angola, 1830s–1860s, Roquinaldo Ferreira, 313

11.

The Demographic Decline of Caribbean Slave Populations: New Evidence from the Transatlantic and Intra-American Slave Trades, David Eltis and Paul Lachance, 335

12.

List of Contributors, 365 Index, 367

Preface

This book is the result of the great surge of information on the slave trade that has become available since 1999. In that year the editors were part of a team that published a CD-ROM containing information on 27,233 transatlantic slave voyages. As noted below, Latin American slaving expeditions were seriously underrepresented in this compilation, and between 2001 and 2005 a major research initiative was undertaken in Portuguese- and Spanish-language archives around the Atlantic basins to correct this deficiency. The current database now contains 34,850 voyages, well over half of which contain information that was unavailable in 1999. The essays included here are a first attempt to make sense of and interpret these new data. Preliminary versions of several of these essays were presented at a workshop at Emory University in December 2004. The database upon which the essays draw could not have existed without the financial support of the Arts and Humanities Research Board of the United Kingdom, the National Endowment for the Humanities of the United States, and Emory University. Their financial support made it possible for us to tap into the work of a great ix

x

Preface

many skilled and dedicated researchers with talents that we ourselves do not possess. Directly involved in constructing the database were Stephen Behrendt, Alex Borucki, Mariana Candido, Daniel Barros Domingues da Silva, Roquinaldo Ferreira, Manolo Florentino, Paul Lachance, Pedro Machado, Philip Misevich, Oscar Grandío Moráguez, Alexandre Vieira Ribeiro, Jelmer Vos, and Rik van Welie. Stephen Behrendt’s work, of course, was one of the foundations of the CD-ROM, and his inflow of new data has not paused since 1992, when he first integrated his own database into what became the CD-ROM. Jelmer Vos began working on the project in 2001 and has remained at the center of operations down to the present. The scholars who have voluntarily approached us with advice and data are unfortunately too numerous to mention. A complete list would come very close to including the names of everyone who has published on the slave trade and related topics in the last quarter century. We do, however, have to single out Manuel Barcia Paz, Franz Binder, Richard Birkett, Ernst van den Boogart, Jose Capela, Jerome Handler, Henk den Heijer, Robin Law, Jean-Pierre Leglaunec, Paul Lovejoy, James McMillin, António de Almeida Mendes, Joseph C. Miller, Johannes Postma, James Pritchard, Lorena Walsh, and Andrea Weindl.

Acknowledgments

Most of the information that underpins this book’s collection of essays has been drawn directly from a set of slaving voyages and estimates available at http://www.slavevoyages.org. The scholarly effort and funding that have made this site possible extend far beyond what the authors and editors could have supplied either from their own resources or from those of the institutions with which they are affiliated. In 2001 we began to address the obvious gaps that were apparent in the first published version of the database (The Transatlantic Slave Trade: A Database on CD-ROM [Cambridge, 1999]), particularly in the coverage of the Iberian branches of the slave trade, as well as to think about making the database available online. Two major grants enabled us to tackle these issues. The first came from the Arts and Humanities Research Board of the United Kingdom (2001–2004, ref: B/RE/AN7080/APN11756), and the second was from the National Endowment for the Humanities from 2006 to 2008 (PA-51985). Harvard University’s W. E. B. Du Bois Institute for African and African American Research provided additional support. Little would have happened without these donors. We would also xi

xii

Acknowledgments

like to acknowledge the generosity of Cambridge University Press, which gave us full and free access to the coding of the 1999 CD-ROM, thus facilitating the migration of the database to the Web, and of the University of Emory for providing funding for a conference held in December 2004, when preliminary versions of several of this volume’s essays were presented. The research that this financial support has made possible since 2001 has been directed not just by the editors of this volume, but also by Stephen Behrendt of the University of Victoria in Wellington, New Zealand, and Manolo Florentino of the Universidade Federal do Rio de Janeiro. All the major decisions have emerged from consultation with them, and of course we have drawn directly on their own substantial accumulation of data and knowledge. But the new database and the interpretations that build on it represent a collective effort in a much wider sense. It is hard to think of any scholar active in the field who has not opened up his or her personal store of data to us in the last few years, and we hope they are all mentioned in the subsequent pages. Their generosity has indeed been extraordinary, and we would like them to see these essays as providing something in return, even if they may not always agree with the uses we have made of their data. Finally, we would like to dedicate this book to a scholar who has spent most of his academic life studying slave societies in the Americas rather than the transatlantic slave trade that supplied them. Paul Lachance came to the project late, but his life has almost come to be dominated by it. Two years ago he was faced with a complicated system of spreadsheets that converted the raw data of slave voyages to a set of estimates of slave departures from Africa and another set of estimates of slave arrivals in the Americas. These contained inconsistencies, were not always transparent, and could not be used to generate estimates of the links between Africa and the Americas. Over many months and at the expense of his other interests, he has introduced consistency and clarity and, most important, has designed an amazingly flexible program that generates estimates of the slave trade for any combination of national carrier, embarkation region, disembarkation region, and year. Readers can discover this for themselves on the estimates page of http://www.slavevoyages.org, but we would like to point out here that most of the tables, and thus many of the essays, in this volume were generated directly from this program. Most of whatever merits the essays contain are founded on his meticulous contributions to both the estimates and database pages of the new Web site. The dedication of this volume to Paul is but modest compensation for the debt that we, as editors and fellow scholars, owe him.

Abbreviations

AGI AHMS AHNA AHU ANRJ APEB APEJE BFSP BNA BNL BNRJ GStA IAN/TT Slave Trade TSTD1

TSTD2

Archivo General de Indias, Seville Arquivo Histórico Municipal de Salvador Arquivo Histórico Nacional de Angola Arquivo Histórico Ultramarino, Lisbon Arquivo Nacional do Rio de Janeiro Arquivo Público do Estado da Bahia Arquivo Público Estadual Jordão Emerenciano, Recife, Brazil British and Foreign State Papers (London, 1812/14– 1967/68) British National Archives, Kew, London Biblioteca Nacional de Lisboa Biblioteca Nacional do Rio de Janeiro Geheimes Staatsarchiv Preußischer Kulturbesitz, Berlin Instituto dos Arquivos Nacionais/Torre do Tombo, Lisbon Irish University Press Series of British Parliamentary Papers. Slave Trade, 95 vols. (Shannon, 1967–71) David Eltis, Stephen D. Behrendt, David Richardson, and Herbert S. Klein, The Transatlantic Slave Trade: A Database on CD-ROM (Cambridge, 1999) David Eltis, David Richardson, Stephen D. Behrendt, and Manolo Florentino, The Transatlantic Slave Trade: An Expanded and Online Database (http://www .slavevoyages.org, launched 2008) xiii

This page intentionally left blank

Chapter 1 A New Assessment of the

Transatlantic Slave Trade David Eltis and David Richardson

The transatlantic slave trade was the largest transoceanic forced migration in history. Peoples throughout time had been forced to relocate in response to natural disaster, military defeat, or exhaustion of resources. Not quite as old was the forced movement of individuals to be sold in markets far removed from their homelands. Wherever slavery existed, and there were few parts of the globe where it was unknown, a slave trade was usually necessary to sustain the long-term viability of the institution. But at the start of the sixteenth century, the shipment of a few slaves from the Iberian Peninsula to the New World initiated a new phenomenon in human experience. Slaves were crowded together in small ships and dispatched thousands of miles to plantations, where they generated large quantities of produce—none of which was vital to the nourishment, clothing, or shelter of the consumers who devoured it. Relatively small improvements to the quality of life of a people on one continent—stemming from cheaper sugar, alcohol, and tobacco (labor for cotton did not come from the slave trade)—were made possible by the removal of others from a second continent, and their draconian exploitation on yet a third. In the 1

2

David Eltis and David Richardson

process this third continent, the Americas, was comprehensively repeopled, not initially with Europeans but with Africans. Prior to 1820, about four Africans arrived in the Americas for every single European. What made this unprecedented, coerced movement of peoples possible? It is almost inconceivable to those living today that for most of the Atlantic slave-trade era (and of course, for eons before this point), trading and forcibly moving people was in moral terms considered no different from buying and moving the merchandise such people might produce. From recorded time in most societies, slave trading had been as accepted an institution as slavery itself. If an expanded slave trade could reduce the price of sugar, then there was certainly no moral constraint to prevent the slave trade from expanding. What made the expansion of transatlantic slavery and the slave trade possible were new opportunities to obtain slaves rather than greater acceptance of servitude per se; such acceptance was a given. The new opportunities emerged first from a mastery of winds and ocean currents that greatly facilitated transoceanic transportation, and second from the collapse of the North American aboriginal populations as an alternative source of labor to migrants from Africa and Europe and the related emergence of large transatlantic differentials in labor productivity. Third and most important were intercontinental differences in conceptions of social identity. On the African coast, both Europeans and Africans traded in people that they looked upon as outsiders. Both sides in the transactions that put captives into the filthy below-deck environment of a slave ship had clear ideas of what made some peoples eligible for the status of captives, and what excluded others—by definition, “insiders.” The exclusively African composition of the transatlantic slave flow stemmed from the fact that Europeans saw all other Europeans as insiders and therefore not potential slaves, whereas the African conception of “insidership” covered a geographic region somewhat smaller than sub-Saharan Africa.1 Fifty years ago the transatlantic slave trade was seen as marginal to the development of the Americas and certainly to the major patterns of global history. It was viewed as an unfortunate and minor episode, even within the context of the global movement of peoples. Since the 1950s, however, it has moved steadily toward the center of the consciousness of professional historians and the general public alike. Print publications, museums, television documentaries, and Web sites dedicated to the subject proliferate. Why? As the above summary suggests, the striking features of the transatlantic slave trade in global history terms are its scale, its racial basis, the uniquely awful belowdeck conditions to which its victims were subjected, and the sheer distance it

A New Assessment of the Transatlantic Slave Trade

spanned. Yet in the face of the many other horrors of history, these features in themselves seem inadequate to explain the recent explosion of interest. Perhaps more important is that interest in the past is always driven by the concerns of the present, the disclaimers of historians notwithstanding. While no counterpart to the slave trade exists today, the disconnect between the values that made the slave trade possible and modern concern with human rights, as well as the burgeoning multicultural composition of modern societies, appears ever more stark and demanding of explanation. It is, after all, less than 150 years since the last slave ship crossed the Atlantic in 1867. Coupled with this compelling dichotomy between the present and a recent (yet ideologically remote) past is the fact that tools now exist for us to uncover so much about the subject. The evidentiary base is strong because the slave trade and slavery survived almost into the modern era. It is also strong because the slave trade was a business that involved accounts and careful record keeping, and it was a business, moreover, in which governments on both sides of the Atlantic took a keen interest. Far more records have survived for this branch of migration than for any other. We may know far more about the movement of Europeans to the Americas than we do about Africans in broad terms, but on the transoceanic phase of that movement, the situation is quite the reverse. The voyages of free migrants, convicts, and even indentured servants have by comparison largely disappeared from our view, at least in terms of their numbers, their direction, and what those on board endured, despite the fact that the last two of these three groups were in temporary servitude and thus subject to sale in the New World. The density of records on the slave trade is such that it is just conceivable that some documentary trace of every slaving voyage that occurred—at least after 1700, when 85 percent of the transatlantic slave trade took place—survives to the present. In the future it will be possible to retrieve ever larger shares of the information that has survived. It appears that more records on slavery and the slave trade remain than for most other institutions of historical interest. Interest in the slave trade has expanded not only as a result of the changing values of society, but more mundanely as a result of the evolution of personal computers and the large amounts of information that they can digest.2 Readers should be aware of the wider project to which the essays presented in this volume belong. Philip Curtin’s well-known 1969 book The Atlantic Slave Trade: A Census initiated the modern era of slave-trade studies and triggered a wave of research into slave-trading records in Europe, Africa, and the Americas.3 Almost forty years later, we are on the brink of a complete

3

4

David Eltis and David Richardson

reconstruction of the history of the transatlantic slave trade from the early sixteenth century through to its close in 1867. The level of detail now possible was unimaginable when Curtin published his book. Where Curtin sought to track slaving activities by centuries or quarter centuries, we can now do so on an annual basis, at least from the mid-seventeenth century onward. Where Curtin grouped ship departures to Africa by nationality, and embarkations and disembarkations of slaves by African coastal regions or American colonies, we can now do the same on a port-by-port basis. Where Curtin could only identify places of embarkation and disembarkation of slaves separately, we can now reveal links across the Atlantic and track how they changed through time. Where Curtin could only measure shipboard mortality through the percentage of losses of slaves in transit, we can now estimate shipboard mortality rates and the factors that helped to shape them. The evidential base for the study of the Atlantic slave trade (and the computational capacity for storing and interrogating it) has been revolutionized. The density and range of the new discoveries have meant that a full synthesis and reassessment of the transatlantic slave trade must be preceded by additional groundwork. Three preliminary steps are necessary: (1) the organization and publication of the massive amount of new data on slave-trade voyages; (2) the production of a convenient and accessible reference to summary statistics derived from these data; and (3) sustained scholarly attention to several branches of the slave trade that were previously known only in outline (to complement what has been learned about the British, French, and Dutch slave trades). The first two of these preparatory phases are now complete. The raw data on the slave voyages are available on an open-access Web site based at Emory and Hull universities in the form of a permanent database. This tool enables both scholars and the general reading public to view the building blocks of new interpretations of the slave trade, as well as make their own contributions of new data and corrections. The new open-access Web site and Yale University Press’s Atlas of the Slave Trade represent the second preliminary step, providing easily accessible statistical and cartographical representations of the Atlantic world’s involvement in the slave trade, all of it derived directly from the new database. The essays in this book complement the new database and the Atlas, and complete the foundations for new interpretations and a new version of Curtin’s Census. Each of the essays deals with a branch of the slave trade that the new data have allowed scholars to explore systematically for the first time. In recent years new research on the South Atlantic slave traffic has not kept pace with the research efforts devoted to the major northern slave powers. The

A New Assessment of the Transatlantic Slave Trade

present collection does much to redress this balance, especially for Englishlanguage readers. While this volume is primarily an intermediate and preparatory step toward what we hope will be a new era of slave-trade studies, in this introduction we nevertheless go beyond merely setting a framework for the new essays. We offer in addition a preliminary reexamination of the overall volume of the slave trade by national flags, African regions, and destinations of slaves in the Americas. In terms of the last reassessment of the slave trade, made with the data as they stood in 1999, this introduction thus constitutes an update of several essays in the special issue of the 2001 William and Mary Quarterly, especially the first essay.4 We also draw on the Eltis and Lachance chapter in this volume to attempt a reconciliation of the new estimates of the slave trade’s volume with population data in one major region of the Americas: the Caribbean. Such an integration of two branches of the historiography—slave demography and the slave trade—was first tried by Curtin in his 1969 book but by surprisingly few of the many scholars who subsequently revised his slave-trade estimates. We begin by reviewing the work carried out since the publication of the 1999 CD-ROM and the articles it made possible. Readers who wish to follow the earlier debates are referred to Per Hernaes’s excellent 1997 summary, but it should be noted that the main conclusion of that overview was that slave ships had carried off 12.8 million Africans to the Americas over nearly four centuries.5 The publication of The Transatlantic Slave Trade: A Database on CDROM in 1999, containing details of 27,233 voyages, formed the basis of a new overall assessment of the transatlantic slave trade.6 At that point we thought that 9.66 million Africans had arrived in the Americas, the survivors of 11.06 million carried from Africa.7 Thus, at first sight, the voyage data as they were in 1999 had apparently reversed the findings of thirty years’ research by reducing the consensus estimates of the mid-1990s, essentially suggesting that Curtin had been correct all along. Closer examination of our revision of nine years ago also showed that, while the new findings for the volume of the slave trade were similar to Curtin’s, the distributions of the slave trade over both time and space were markedly different from those of both Curtin and his critics. More important, however, our work on the CD-ROM database and since then has involved a shift in the way patterns of the slave trade, including its size, could be assessed. In broad terms, it is now possible to track these patterns with data on voyages alone, instead of with combinations of voyage information, contemporary opinion, and demographic research.

5

6

David Eltis and David Richardson

The essays in this book are all based on research completed since the publication of the 1999 database. They do not ignore either contemporary opinions on the size and direction of the trade or demographic data. Nevertheless, they could not have been written without recently discovered archival data. Collectively they represent a major shift toward reliance on shipping data. The reason for this shift lies in the strengths of the new database. Both versions of the transatlantic slavetrade database, the 1999 CD-ROM as well as the revised online version— henceforth called TSTD1 and TSTD2, respectively—are multisourced. By this we mean that the great majority of the voyages contain information that comes from more than one source.8 While hard-copy published lists of voyages that were multisourced first appeared in 1978, TSTD1 was the first electronic data set on the slave trade that employed this basic strategy. Thus, if the same voyage appears in different records created, say, in Europe, Africa, or the Americas, the information is integrated and the different sources are listed. Indeed, we have allowed for seventeen separate sources of information per voyage, although only eight voyages have all seventeen source variables completed. A great part of our research effort has been directed to eliminating double and triple counting of the same voyage. Over 80 percent of the voyage entries in TSTD2 have more than one source, and on average there are between three and four sources of information for each voyage.9 In short, the density and weight of the information now available on individual slave voyages have increased the database’s reliability and comprehensiveness. Population data, vital rates, and the opinion of contemporaries are now useful chiefly for the independent checks they might provide for voyage-based assessments or, as the present essays show, helping to fill in gaps in the shipping records for the pre-1700 era before the slave trade’s rapid expansion. What has been added to TSTD1 in the last few years? It was clear in 1999 that the largest hole in the data comprised missing voyages that had sailed under the Portuguese flag. The compilers of the 1999 set were also aware of weaknesses in the coverage of the Spanish trade, which in its transatlantic manifestation was active only in the first 150 and the last 85 years of the slavetrade era—either side of the trade’s eighteenth-century peak, in fact. Other potential gaps existed for the large trade based in London before 1662 and again between 1711 and 1779, the Dutch trade between the Brazilian interlude and the founding of the second West-Indische Compagnie in 1674, and the early French trade. Between 2001 and 2005, a group of scholars funded by an Arts and Humanities Research Board grant (UK) filled in or severely reduced most of these gaps.10 Archives in Luanda, Rio de Janeiro, Bahia, Lisbon, Havana, Madrid, Seville, Ghent, Amsterdam, Copenhagen, London,

A New Assessment of the Transatlantic Slave Trade

Figure 1.1

Portuguese and Brazilian Voyages with New Information since 1999

Source: Calculated from TSTD2.

and Middelburg, the Netherlands, as well as extensive eighteenth-century newspaper holdings in the Bodleian and British libraries, have been exploited. Scholars unconnected with the project have given generously of their time and the archival data that they themselves have collected.11 This research effort has resulted in the discovery of 8,232 additional slave voyages that we were not aware of in 1999. Just as important, it has allowed us to modify 19,729 voyages that were already included in the 1999 database.

7

8

David Eltis and David Richardson

The enormous advance in our knowledge of the Portuguese and Brazilian trade is shown in figure 1.1. Whereas the first edition of the database contained 6,183 records of voyages under these two national flags, the new edition has 11,382 such entries. Of the 6,183 voyages in the 1999 edition, over two-thirds have had new information added in the last five years. Figure 1.1 shows how the distribution of this new information is concentrated in the years 1676 to 1830—the period when the Portuguese and then (after independence) the Brazilian traffic was at its height. Thus over 60 percent of the slave voyages added to the database since 1999 are Portuguese and Brazilian crossings that occurred when that trade was at its peak. Each voyage in both TSTD1 and TSTD2 allows for 226 variables covering the identity of the ship, the major geographic and population characteristics of its voyage, and six different dates as the vessel tracked around the Atlantic. The structure of the database follows the voyage from its point of departure in Europe or the Americas through to its return to home base after first buying and then selling slaves.12 It need hardly be said that while large, the new database is not complete. It does not include all the transatlantic slave voyages that ever sailed, although what proportion of all voyages it does include forms a central part of our new assessment of the volume and direction. Nor does it include full information on the voyages that it does contain. Many of the variables remain blank, and no voyage in either the new or the old database includes information on all 226 variables. Indeed, a few entries have little more than a date and a single location recorded. Table 1.1 summarizes what the new database offers by showing the coverage for eighteen of the most important of the 226 variables. Geographic information is particularly well represented. Rows 9, 11, and 13 indicate that almost all voyages are associated with a particular location, and two-thirds have transoceanic connections in the form of two or more place names specified. The table also shows that that coverage of ownership, shipboard mortality, and the age and sex composition of captives, while less abundant than the geographic variables, is still extensive. Clearly, major information gaps remain for some voyages already in the database and for those voyages that scholars have yet to discover. The traffic to Rio de Janeiro and Pernambuco prior to 1700 and to Bahia de Todos os Santos before 1678 is thinly represented. A greater share of the slaving expeditions to the British Caribbean before 1662, and to the French before 1710, are in the new database, but the data are still deficient. Thus the remaining gaps in the data are generally for the earlier phases of the slave trade, when the flow of traffic was much smaller than it eventually became.

A New Assessment of the Transatlantic Slave Trade

Table 1.1 Select Summary Information Contained in the Revised Transatlantic Slave Voyage Data Set (TSTD2) Number of slave voyages in the data set Voyages with a record of the name of the vessel Voyages with name of captain(s) Voyages with name of at least one ship owner Number of Africans imputed to have embarked on slave ships Number of Africans imputed to have disembarked from slave ships Size of crew provided on one or more legs of voyage Voyages showing tonnage of vessel Voyages reporting place of ship departure in Europe or the Americas Date of ship departure in Europe or the Americas Voyages reporting place(s) of embarkation on the African coast Voyages reporting specific numbers of Africans embarked Voyages reporting place(s) of disembarkation Dates of arrival at place(s) of disembarkation Voyages reporting numbers of Africans disembarked Voyages reporting number of Africans died on board Voyages with age or gender of Africans disembarked reported Outcome of voyage indicated Evidence of slave revolt or attacks from shore-based Africans

34,808 33,207 30,755 20,978 10,125,456 8,733,592 13,253 17,592 28,505 25,265 26,939 8,547 28,985 23,478 18,473 6,382 3,570 31,077 530

Source: TSTD2.

To convert this mass of new data into new interpretations of the slave trade, it is necessary to make some assumptions and inferences. Two major types of inferences are required. Because few voyages in the historical record contain complete information on the routes taken and the captives carried, we have to surmise where the vessel went and how many slaves it carried. Often we know where the voyage intended to go but not whether it actually arrived. For other voyages we might know the numbers purchased but not the number sold or vice versa, or in some cases we know only the number the captain intended to buy. In the case of 3,600 voyages—10 percent of the new data set—we know only that the slaving venture set out for Africa. The first set of estimates—or more accurately, inferences—relate to missing information for a given voyage. The extent of the missing information for the key variables is indicated in table 1.1. It would be wasteful to set aside all such voyages, and we have chosen instead to create a number of additional variables that use historical data when they are available and inferences when they are not. Thus we assume that all vessels reached the Americas with slaves, unless information to the contrary exists. Similarly, on the geography

9

10

David Eltis and David Richardson

of the trade, we have assumed that a vessel actually reached the port to which it intended to sail, again unless information to the contrary has been found in the historical record. Finally, if no information on the number of captives put on board a vessel exists, an average mortality rate is used to infer that number on the basis of numbers disembarking. The same procedure is used if the missing number is for those disembarking when the number embarked is known. Where no information on numbers exists, an estimate is derived from either the rig of the vessel, or if that is not available, the route and time period of the voyage. To make this estimate as accurate as possible, we have distributed those voyages for which the number of slaves carried was available over 158 categories of different combinations of rig, route, and time period, calculating a separate average of number of captives embarked and disembarked for each one. Thus we have imputed variables for outcome of voyage, place of trade in Africa, place of disembarkation (usually in the Americas), and numbers of slaves who embarked, disembarked, and died on board. These “imputed” variables form the basis of the estimates presented below. The second type of estimate required—in addition to information missing about known voyages—is how many voyages are missing. To put this central question differently, how can we know what proportion of the total slave trade is represented by the voyages in the new data set? Answering this question is virtually the same as assessing the quality of the sources, but a good starting point is to recognize the significance of a multisource database. TSTD2 incorporates records created at the point of each voyage’s organization in Europe or the Americas; records created in Africa when the vessels purchased captives; and records at the point of sale in the Americas. The existence of diverse and dispersed sources for the same voyage encourages confidence in the internal consistency of the surviving records and reduces the possibility of missing data. Slave vessels did make voyages without leaving any historical trace of their activities, but after 1700 this did not happen often. A closer look at new research on some major branches of the slave trade will demonstrate these points. The two-volume Mettas–Daget catalogue of slave voyages sailing from French ports contains captains’ reports of sightings of other French slave ships. These sightings constitute as random a sample of French slave voyages as is possible to obtain. No less than 95 percent of the vessels named in these individual summaries are already in the database from other sources. We might conclude that the Mettas–Daget catalogue is therefore 95 percent complete. A second indication is the tiny number of new voyages

A New Assessment of the Transatlantic Slave Trade

that newly discovered sources now add to the data set. Thus, by far the largest destination of French ships carrying slaves to the Americas in the eighteenth century was the colony of Saint-Domingue. This traffic reached its peak between 1784 and 1791, when an average of eighty-four slave ships a year set sail for the colony. Since work on TSTD2 began, we have located 488 new references to voyages to Saint-Domingue in these nine years. Only seven of these turned out to have been voyages that were not already in the Mettas–Daget catalogue. In the British case, Stephen Behrendt has recently combed through the sailors’ and widows’ petitions at the Society of Merchant Venturers in Bristol, England.13 This is a source that covers voyages from Bristol to all parts of the world, not just Africa, and is not connected in any way to the major sources of shipping movements and records of state or the newspapers that support the bulk of the records in the data set. Again, it constitutes a random sample of slave-trade voyages. Of forty-six references to slave ships in its records, only one was to a vessel that was not already in the database. A third example is the massive Portuguese slave trade to Rio de Janeiro between 1795 and 1830 at the height of the coffee boom. Pre-1999 databases included 1,187 voyages arriving in Rio, assembled by Manolo Florentino for the years 1811 to 1830; 889 by Herbert S. Klein covering arrivals in 1795–1811 and 1825–30; and a list of 170 vessels leaving Angola for Rio between 1795 and 1808, put together by Corcino Medeiros dos Santos.14 After integrating these separate sources, we found that almost all vessels in the Angola list were also in the Rio de Janeiro records, and that all the Klein records for 1825–30 were also in the Florentino data set. In total, records of 1,536 slave voyages existed for Rio de Janeiro between 1795 and 1830. Since 1999 we have added new information to almost every one of these 1,536 voyages but have found only 357 completely new voyages for this branch of the slave trade. And many of these ships were captured or destroyed before reaching their destination. We think it highly unlikely that significantly more voyages sailed to Rio de Janeiro in this period than are contained in TSTD2. Sources explored since 1999 that have yielded significant numbers of new voyages usually helped to fill gaps that we always suspected. English newspapers for the period before 1750, as well as the Royal African Company’s duty books for 1698–1712, neither of which were included in TSTD1, turned up 178 new slave voyages leaving or returning to London between 1700 and 1750. London, indeed, was the Atlantic port that organized and dispatched the most slaving voyages anywhere between the mid-seventeenth and mid-eighteenth

11

12

David Eltis and David Richardson

centuries.15 How important was this new information? In 1999 we published records of 1,117 slaving expeditions based in London, but we were sufficiently aware of the gaps that we suggested that the CD was missing perhaps 10 percent of the actual total.16 In terms of actual voyages, the new information appears to have added 13.7 percent more voyages to what we knew about in 2001. Our estimate of the missing voyages was thus on the low side, but not by much. A similar situation exists for arrivals in the United States after 1782, where we also suspected that our database was somewhat deficient. James McMillin’s careful combing of newspaper sources in the post-1782 period has yielded only 147 voyages not previously known—or a 9 percent increment over those already in the database in 2004, when his database was published.17 In both these cases, however, the London and U.S. newspapers turned up many hundreds of voyages that were already in the TSTD1 database. The pattern of new research into independent sources uncovering relatively small increments of previously unknown voyages is now a well-established feature of ongoing investigations into the slave trade. Nevertheless, if we want an accurate estimate of the size and direction of the full slave trade and not just the sample of it in TSTD2, then a carefully calibrated port-by-port, region-by-region evaluation of how large that sample is becomes necessary. Our reassessment has been constructed around the national flags that slave traders used to cross the Atlantic. Overall, at the time of this writing, the data set identifies the national affiliations of 25,569, or 73 percent, of the voyages in the set. For a further 7,711, the context of the voyage and the name of the ship, owner, or captain make possible inferences about place of registration. An imputed variable of national flag—which contains national affiliations for 33,229 voyages, or 95 percent of the voyages in the database—was accordingly created. Almost all the voyages that lack an indication of the flag under which they sailed plied their trade either at the very beginning of the slave trade, when the probability that the vessel was Spanish or Portuguese was extremely high, or at the very end, when the slave trade was illegal. Illegality meant that owners and captains went to great lengths to avoid or to change national affiliation as part of an attempt to avoid detention. We thus distributed totals for national flags across regions in Africa and the Americas according to the ratios calculated from TSTD2. While discussions of our judgments of completeness for the larger branches of the trade form a central part of the present volume, if we were to explain all the details here, the rest of this introduction would read even more like a manual for a piece of computer hardware than it already does. All these judgments are posted on the slave-trade

A New Assessment of the Transatlantic Slave Trade

Web site and are accompanied by an interface that permits the derivation of a series for any combination of years, national flag, African region of departure, and American region of arrival.18 Scholars who disagree with the allowances we have made for missing voyages for particular years, regions, or ports will be able to see exactly what allowance we have made and to use our posted work as a basis for making their own calculations. Almost all the essays in this book either provide or draw on voyage information from TSTD2 or the estimates page that is an extension of TSTD2. The earliest transatlantic slave traders were the first Europeans to assume control of parts of the New World. The Spanish brought in the first slaves to be carried as part of a commercial venture, probably in 1501. The vessel set sail from Seville and embarked its slaves not in Africa but in Spain, which had absorbed several thousand African captives in the second half of the fifteenth century. The ship discharged its human cargo in the New World on the island of Hispaniola, well before the inception of the Spanish mainland empire. The first transatlantic slaving voyage (if we ignore the several hundred Amerindians that Columbus brought to Europe when he returned from his second voyage) was thus bilateral rather than triangular. The broad patterns of the subsequent movement of slaves across the Atlantic to the early Spanish Americas are described in António Mendes’s chapter and will not be repeated here. Suffice it say that it is not possible to separate the early Spanish from the early Portuguese voyages, although it is apparent that, especially during the union of the Iberian crowns (1580–1640), the Portuguese came to dominate this branch of the slave trade. The first column of table 1.2 shows a series of slave departures from Africa to the Spanish Americas in ten-year groupings taken directly from TSTD2. These were slaves for whose departure some historical record exists. The second column does the same for Mendes’s estimates of what the true volume of departures was likely to have been—in other words, it incorporates an allowance for slave departures missing from the historical record. Both series constitute upward revisions of Enriqueta Vila Vilar’s thirty-year-old work.19 The last two columns simply divide these estimates between Portuguese and Spanish traders. A systematic and steady slave trade to Brazil began only in the 1560s, several decades after its Spanish American counterpart. From the beginning it was entirely Portuguese, becoming Brazilian for a time in the aftermath of Brazilian independence, and operating sometimes without a flag of any kind during the decade before it finally shut down in 1851. Its volume before the Dutch conquest of Pernambuco in 1630, and indeed before 1700, has attracted little

13

14

David Eltis and David Richardson

Table 1.2 Slaves Carried from Africa to the Spanish Americas on Portuguese and Spanish Slave Vessels by Decade, 1501–1641

Decades 1501–10 1511–20 1521–30 1531–40 1541–50 1551–60 1561–70 1571–80 1581–90 1591–1600 1601–10 1611–20 1621–30 1631–41 1501–1641

Slave Departures for the Spanish Americas in TSTD2

Estimated Slave Departures for the Spanish Americas

Estimated Share of Column 2 Carried in Portuguese Vessels

Estimated Share of Column 2 Carried in Spanish Vessels

0 237 227 833 434 172 1,527 926 2,785 33,834 22,724 47,929 47,931 41,743 201,303

1,900 8,807 10,990 14,409 27,395 5,651 37,500 26,080 47,390 59,812 50,664 61,957 72,340 54,655 479,550

950 4,404 5,495 7,204 13,697 2,826 18,760 13,040 23,695 29,906 25,332 30,979 36,170 27,327 239,785

950 4,404 5,495 7,204 13,697 2,826 18,760 13,040 23,695 29,906 25,332 30,979 36,170 27,327 239,785

Sources: Calculated from estimates page at http://www.slavevoyages.org and António de Almeida Mendes, “The Foundations of the System: A Reassessment of the Slave Trade to the Spanish Americas in the Sixteenth and Seventeenth Centuries,” this volume (chap. 2), tables 2.4 and 2.5, adjusted for mortality on the voyage of 30 percent of those embarked.

new work beyond the secondary sources such as Goulart and Lopes that Philip Curtin used in 1969 to support an estimate of 50,000 slaves transported to Brazil before 1600 and 120,000 in the first three decades of the seventeenth century, all carried under the Portuguese flag.20 As already indicated, TSTD2 contains insufficient records of voyages to form an estimate of any branch of the slave trade to Brazil prior to 1678. For one of the four branches of the Brazilian slave trade, however—that to Pernambuco—there is some documentation on the scale of the traffic. As this port was the most active of the pre-1630 connections with Africa, it follows that there is some basis for using it as a template for the rest of the early Brazilian traffic. The Domingues and Eltis chapter in this volume presents the details of the estimates for Pernambuco from 1561 to the early 1850s, and they are incorporated into the aggregate slave-trade estimates used here without further discussion. Note, however, that the data in table 1.3 for Pernambuco are for departures from Africa rather

A New Assessment of the Transatlantic Slave Trade

than for arrivals in the Americas as shown in the appendix 3.1, and therefore table 1.3 allows for mortality on the ocean crossing.21 Space constraints dictated that only the decadal series could be shown in the table. A comparison of the totals in the TSTD2 and the “estimated” columns for Pernambuco suggests that just over half of our estimates of slave departures for this region are rooted in voyage data. The estimates for the rest of Brazil shown in table 1.3 require a little more discussion. For Bahia, abundant new data on voyages are presented in Ribeiro’s chapter in this volume, although almost all of them are for post1677. Obviously, our discussion here attempts to go beyond these data and allows for what might be missing from these new sources. The first slaves brought into Bahia arrived as early as 1538 and were set to work cutting wood.22 We think that the Recôncavo of Bahia, the second major sugarproducing captaincy, as well as the smaller São Vicente and Rio de Janeiro producing areas to the south, began a steady slave trade from 1580. Appropriately, TSTD2’s first arrival of slaves in Pernambuco, in 1574, predates its Bahian counterpart by eight years, though voyage data are generally scarce for these early years.23 Africanization of the work force of American indigenes proceeded a bit more quickly in Bahia and the southerly regions, and was complete in both Pernambuco and Bahia at least by 1620, according to Stuart Schwartz.24 Pernambuco had quickly become the premier sugar-producing area of the world in these years, and the other regions lagged behind prior to the Dutch invasion. Our estimated departures from Africa for Bahia are set at four-fifths of the Pernambuco figure through to the Dutch invasion in 1630, and for the southern regions of São Vicente and Rio de Janeiro, departures are considered to have been 70 percent of the Bahia estimates.25 Thus the estimates for Brazilian regions in table 1.3 are centered on Pernambuco, as developed in chapter 3. Before 1620, when the first reliable assessment of the level of slave arrivals by a contemporary observer becomes available, the estimate series assumes a constant annual growth of arrivals from 100 in 1560 (1580 in the case of Bahia and Rio de Janeiro) to 4,640 in Pernambuco (3,631 and 2,541 in Bahia and Rio de Janeiro, respectively). The total volume of slaves leaving Africa for Brazil prior to the Dutch invasion by this reckoning would have been 262,400—about 50 percent greater than the estimate Curtin derived from his survey of Portuguese and Brazilian historians. The Dutch invasion of the Portuguese Atlantic affected more than just Pernambuco. The attack on Bahia in 1624 was repulsed, but Dutch occupation of Luanda in the 1640s and attacks on Portuguese shipping in the Atlantic

15

Table 1.3 Slaves Carried from Africa to Non–Spanish American Destinations on Portuguese and Brazilian Vessels, Major Importing Regions by Decade, 1561–1860: TSTD2 Sample Compared to Our Estimated Totals Amazonia

Pernambuco

Bahia

Southeast Brazil

Decades

TSTD2

Estimate

TSTD2

Estimate

TSTD2

Estimate

TSTD2

1561–70 1571–80 1581–90 1591–1600 1601–10 1611–20 1621–30 1631–40 1641–50 1651–60 1661–70 1671–80 1681–90 1691–1700 1701–10 1711–20 1721–30 1731–40 1741–50 1751–60 1761–70

0 0 0 0 0 0 0 0 0 0 0 0 0 1,264 431 1,384 1,892 987 589 6,344 13,838

0 0 0 0 0 0 0 0 0 0 0 283 0 1,761 1,202 2,045 2,552 1,648 1,359 6,091 13,839

0 451 0 0 334 0 1,450 0 0 1,794 0 0 118 7,787 623 6,948 41,347 22,088 30,137 32,821 28,912

1,623 3,110 5,959 11,415 21,871 41,734 48,679 1,469 0 12,000 22,078 24,838 27,780 41,064 54,350 42,573 40,520 21,374 30,471 32,822 29,163

0 0 166 0 244 118 1,241 0 4,776 1,300 1,762 5,847 13,497 50,063 54,841 82,691 95,587 99,059 99,430 78,180 68,908

0 118 1,944 4,582 10,810 25,484 36,310 36,310 26,753 42,976 48,665 38,465 34,423 63,819 68,501 91,607 102,812 103,846 104,663 82,296 72,534

0 0 0 334 0 0 0 0 0 0 382 0 504 3,574 2,697 9,379 36,999 25,065 53,640 17,635 37,410

Estimate 0 82 1,361 3,209 7,566 17,837 25,414 25,414 18,725 30,083 36,179 27,378 24,094 44,675 46,939 62,334 55,629 72,479 69,364 91,959 95,364

Africa, Europe, Other TSTD2

Estimate

0 0 0 183 0 969 284 0 3,378 4,042 3,437 2,659 316 2,207 410 1,119 3,815 1,236 3,017 2,766 386

0 0 0 183 0 969 284 0 3,378 4,042 3,437 2,659 316 2,207 410 1,119 3,815 1,236 3,017 2,766 386

1771–80 1781–90 1791–1800 1801–10 1811–20 1821–30 1831–40 1841–50 1851–60 1561–1860 TSTD2 %

14,334 18,345 17,669 33,891 22,537 12,968 2,414 2,320 0 151,208 93.2

14,470 18,940 20,123 35,674 23,822 13,399 2,534 2,443 0 162,182

22,344 18,399 14,009 37,322 83,338 59,858 21,667 12,751 1,642 439,443 47.6

23,270 28,625 39,208 57,628 83,336 58,620 42,670 24,345 1,642 873,234

79,937 84,247 99,243 107,163 127,628 105,378 14,513 72,241 1,023 1,349,082 78.5

Sources: TSTD2, and calculated from estimates page at http://www.slavevoyages.org. *Includes 673 to Europe in 1511–20.

84,144 87,806 104,577 112,272 126,371 104,577 39,896 72,240 1,146 1,730,055

17,135 31,976 104,271 157,257 246,297 372,883 204,474 223,580 7,110 1,552,604 60.6

87,938 119,080 131,015 157,259 246,297 377,153 329,837 382,387 7,228 2,594,280

211 0 11,769 26,833 30,585 27,139 23,889 35,663 497 187,299* 100.0

211 0 11,769 26,833 30,585 27,139 23,889 35,663 497 187,299*

18

David Eltis and David Richardson

throughout this period disrupted (though never came close to shutting down) the slave trade to the whole of Brazil. Arrivals in Bahia are set at 2,000 (or 2,353 departures) a year between 1641 and 1648; after the Dutch were expelled from Angola, it is assumed that the slave trade increased once more. Almost all slaves arriving in Bahia until 1678 came from Angola. TSTD2 has occasional records of Bahia vessels sailing to Southeast and West Africa in the 1660s and 1670s, but the data set indicates that continual trade with West Africa began only in 1678. None of the English or French records for the Gold and Slave Coasts of West Africa mention vessels from Brazil prior to this point. For the years 1666 to 1672, a document in the AHU, cited by Luiz Felipe de Alencastro, lists annual totals of slave departures from Angola on Portuguese ships, and the Bahian share of this series, set at 55 percent of Angolan departures, is taken as the complete slave traffic into Bahia for the years covered by the document.26 For 1650 to 1665, when the Angola series began, and for 1673 to 1677, the years immediately after it ended, we use the annual average for 1666–72 as calculated above. From 1678, Bahia vessels tapped a hitherto unknown demand for rolls of tobacco coated in molasses, and by the years 1698–1700, the Costa da Mina, as the Portuguese called the Gold Coast, Slave Coast, and Bight of Biafra, was supplying two-thirds of the slaves destined for Bahia.27 The influence of Pierre Verger’s work on the cultural connections between Bahia and the Bight of Benin is so great that historians have often assumed that Bahia drew almost all its slaves from the Costa da Mina. In fact, Angola continued to supply about one-third of Bahia’s slaves through to the end of the slave trade—one reason why Bahia’s slave trade has been underestimated in the past.28 For most years after 1677, two and usually three separate recorded series of ships leaving Bahia for Africa have survived. Vessels leaving Bahia for the African coast had to obtain a license to export tobacco, given the normal mercantilist requirement that colonial produce could be exported only to the mother country—in this case, Portugal—but vessels sailing to Angola were also recorded in one of these series. From 1698 on, these records begin to mesh with those kept at São Tomé in the Gulf of Guinea, where vessels sailing from the Costa da Mina normally called, as well as with the scattered reports of slave vessels disembarking slaves in Bahia.29 We assume that some part of the voyage of 90 percent of the slave ships coming into Bahia from 1698 to 1728 entered the historical record and is thus included in the database. After 1728, when stronger information on arrivals begins, we assume that the database includes 95 percent of all actual voyages. For 1811 to 1830, the best-documented

A New Assessment of the Transatlantic Slave Trade

period of the Bahia slave trade, our assumption is that all slave vessels are included, and for the illegal period we rely on the estimates of the British consul and other contemporary observers, as discussed in an earlier essay.30 This leaves the period 1678 to 1698, when the coverage of the sources is less certain. For 1678 to 1689, the TSTD2 estimates are doubled, and from 1690 to 1697, voyage data are divided by 0.7 to arrive at an estimate; both procedures reflect our assessment of the gradual improvement in the range and complementary nature of the sources during the last two decades of the seventeenth century. The end result of this mixture of assumptions and hard data is that we have voyage-based data of 1,349,082 slaves leaving Africa for Bahia and, on the basis of this, project an actual total of 1,715,888 captives.31 Thus the voyage data comprise nearly 79 percent of the final estimates. Amazonia was the last of the major Brazilian regions developing a slave trade direct from Africa, and the earliest to bring it to an end. Maranhão and Pará were not settled until the late seventeenth century, with the first slaves recorded arriving directly from Africa in 1680 and the last in 1846. Slaves coming from other parts of Brazil likely preceded the first transatlantic voyage and possibly continued after the last arrivals from Africa. Indeed, of the four major Brazilian importing regions, Amazonia drew most heavily on the intraAmerican slave trade. Much of the small early export trade was in cacao, collected in the wild rather than cultivated and channeled through Pará. Rice expanded between the 1750s and 1770s, but it was cotton from the 1760s that pulled Amazonia into the Atlantic economy, a development that came only after one of the Pombalinas companies, the Companhia Geral do Grão-Pará e Maranhão, began to bring in slaves after 1755 as part of an attempt to kickstart significant commodity exports. The slave trade reflects commodity trade patterns. No more than a few hundred captives a year, and often none at all, arrived in either Pará or Maranhão before the 1750s, and not until the 1760s did the annual average rise above 1,000 slaves.32 Amazonia ports were remote from the sites of the first Brazilian gold discoveries and were not a conduit for the interior until the late eighteenth century. In addition, like most minor slave-importing regions of the Americas, Amazonia was able to draw on the intra-American slave trade via Pernambuco. It is highly unlikely that slaves arrived in the region directly from Africa before 1680, and from this point to the mid-1750s we have assumed that, in those years for which TSTD2 records no slaves, 100 slaves arrived. The Pombalinas company years are well documented, and from 1755 to 1778, when the company monopoly ended, TSTD2 probably includes all

19

20

David Eltis and David Richardson

arrivals. However, the Companhia Geral do Grão-Pará e Maranhão remained involved in the slave trade until 1788, and company accounts record the traffic in slaves carried on by “private” traders. When the company ceased activities, other sources from the Torre do Tombo and the AHU, Maranhão, and Guiné became available. This was the only branch of the traffic to Brazil in which Lisbon continued to play a major role as a port of organization, and voyages in this period are often documented in Europe, in Portuguese Guinea in Africa, and in the Amazonian ports.33 As a consequence, from 1778 through to the end of the trade, we feel justified in arguing that TSTD2 contains some record of 95 percent of the slaves who actually disembarked in Maranhão and Pará. Table 1.3, adjusted for voyage mortality as shown in the spreadsheet linked to the slave voyages’ Web site estimates page, reflects this judgment. Overall, table 1.3 also shows that Amazonia is the best-represented Brazilian region in TSTD2, with recorded departures of slaves leaving Africa for the region amounting to 94 percent of the estimated departures. For Rio de Janeiro, the voyage-based documentation is very strong for the nineteenth century but weaker than the rest of Brazil for the seventeenth century. It is comprehensive enough for us to assume 100 percent coverage somewhat earlier than for the traffic to Bahia—in 1793. During the illegal period of slaving, after 1830, British Foreign Office officials and their spies paid close attention to Rio, so that assessments of the illegal Rio traffic are more complete than for other parts of Brazil.34 Before 1710, however, albeit a period when Rio was the least important of the three major Brazilian slave-importing centers, TSTD2 contains only thirty-one slave voyages connected in any way to the port. Based on André João Antonil’s assessment of the relative importance of these three, we estimate annual arrivals into early Rio de Janeiro at 70 percent of the Bahian figures between 1620 and 1709, as noted above. For 1710 to 1793, we lack data on either arrivals or departures for Rio de Janeiro, but the good quality of the information on vessels and slaves leaving Angolan ports provides an alternative residual route to an estimate. In short, if we know the total departures from Angola and the numbers arriving from Angola in Bahia, Pernambuco, and eventually the Amazonian ports, then the size of the slave traffic to Rio de Janeiro (except for the small share of it coming from West Africa) must approximate the Angolan total minus the totals for the other three regions—adjusted, of course, for shipboard mortality. Brazilian-bound slaves left from only two Angolan ports during this period: Luanda, and later Benguela. For Luanda, José Curto has compared separate series constructed by Birmingham, Klein, and Goulart with figures he himself

A New Assessment of the Transatlantic Slave Trade

extracted from the archives, and has selected his own preferred series of annual departures for 1710 to 1810. Curto has also constructed a separate series for Benguela.35 Except for Klein’s estimates, none of these data are based on voyages but rather on the annual reports of officials stationed in the two Angolan embarkation points. A comparison of voyage data from TSTD2 with the reports of the colonial officials shows that, except for 1710–22, the officials’ series is about 10 percent greater. Our own preferred series for Angolan exports uses voyage data from 1710 to 1722, a period when the voyage data seem more complete, and Curto’s preferred series thereafter, with some minor adjustments to allow for voyages known to be missing. The number of slaves carried off to other Brazilian ports is calculated above, and departures from Angola to Rio de Janeiro are the residual after deducting the sum of the traffic to other ports and allowing for voyage mortality.36 However, Angola was not the only source of slaves coming into Rio de Janeiro, and Nireu Cavalcanti has recently presented a breakdown of slaves arriving in Rio from Angola and the Costa da Mina for the 1731–35 period. His figures indicate that one in ten Rio slaves arrived from regions other than Angola—mainly the Costa da Mina.37 There are occasional voyages in TSTD2 from the Costa da Mina into Rio throughout the eighteenth century, and Cavalcanti’s ratio is accepted for 1710 to 1792. The final series for departures and arrivals into Rio is shown in table 1.3, where a comparison of TSTD2 and our estimated totals indicates a voyage record for six out of ten slaves included in our estimate.38 Finally, there are records of Portuguese voyages to four other groupings of destinations. A few Portuguese slave voyages went to Europe, and many more went to parts of the Americas other than Brazil and the Spanish Americas. A much larger third group of slave voyages disembarked slaves in Africa, mainly in Sierra Leone, after having been captured. This group is discussed more fully below in the section on the Spanish trade. Finally, for a fourth group, we know only that the vessel embarked or intended to embark slaves, but we do not as yet have any information on port of arrival (or intended port of arrival) beyond, in a few cases, Brazil unspecified. Slaves carried on all these four categories of voyages are grouped together in the final columns of table 1.3 and total almost 150,000. Adding this to the totals for Brazilian regions shown in the rest of table 1.3, together with the Portuguese share of the traffic to the Spanish Americas before 1642 shown in column 3 of table 1.2, gives us an aggregate of close to 5.8 million over the whole period. The early slave trade to the Spanish Americas was surpassed not only by the Brazilian trade but also by a massive new traffic to the Caribbean organized by

21

22

David Eltis and David Richardson

northern Europeans. The Dutch led the North European incursion, but the basic fact is that they had no sugar plantation colonies for much of the seventeenth century. Their attempt to take Pernambuco from the Portuguese was temporarily successful, but they were never able to generate sugar exports (nor to attract slave arrivals) at levels that matched those achieved by the Portuguese before the conquest. After their final expulsion from Brazil in 1654, the Dutch spent a decade without a major plantation colony, and even after acquiring Surinam from the English in 1664, they brought in only 35,000 slaves during the rest of the century to the Dutch Guianas, compared with the 150,000 that the British carried into Barbados alone during the same period. Of course, historians have traditionally defined the Dutch role as carriers and suppliers for the other early Caribbean empires, particularly the French and the English, but such a role probably never existed, and if it did, Colbert’s exclusif policy and the English Navigation Acts had effectively excluded the Dutch from the Caribbean sugar complex by the last third of the seventeenth century.39 Between 1673 and 1698, the Dutch carried almost no slaves into the French and British Caribbean, and for every slave they sold in the Dutch Guianas, they carried two to Spanish markets, most via Curaçao. When the asiento (see chapter 2) went to Portugal, France, and then England in turn between 1698 and 1713, the Dutch first attempted to run a slave bazaar for the whole Caribbean. When this strategy had clearly failed by the 1720s, they concentrated on supplying their South American colonies for the rest of the slavetrade era.40 TSTD2 permits a firmly rooted reckoning of the relative importance of the Dutch in the transatlantic slave trade.41 Prior to the conquest of Pernambuco, the Dutch carried few slaves to the Americas. Except for a single voyage in 1630, their slave trade to Pernambuco operated only between 1637 and 1650. TSTD2 contains only three Dutch slave voyages in these years linked with any part of the Americas other than Pernambuco, and we think TSTD2 includes all, or almost all, Dutch voyages before midcentury. From 1650 on, we rely heavily on the database assembled by Franz Binder, which is both extensive and multisourced.42 Further research into new sources, both Dutch and non-Dutch, has generated additional information on some of Binder’s voyages but very few that were not already in the Binder database. We feel safe in assuming that TSTD2 includes 90 percent of slave voyages sailing under the Dutch flag between 1650 and 1674. In 1674 the Dutch established a second West India Company that, until 1730, held the sole right to trade on the African coast—at least as far as Dutch citizens were concerned. The documentation

A New Assessment of the Transatlantic Slave Trade

from this period is dense, and we accept Postma’s contention that some information about every Dutch voyage that sailed has survived. After 1730, the era of “free trade,” the documentation is almost as comprehensive as before: the archives of the Middelburg Company and the Coopstad and Rochussen partnership are perhaps the most complete slave-trade archive in existence. In summary, we think TSTD2 includes all Dutch slave-trading voyages with the exception of a very few before 1630 and about 10 percent that are missing from the 1651–73 period. How many slaves did the Dutch carry? As described in this book’s Vos, Eltis, and Richardson essay (chapter 8), we estimate 3,200 taken from Africa before 1631. During the Pernambuco venture, the Dutch carried away an estimated 30,900 slaves from Africa and disembarked 25,500 in the Americas for the thirteen years through to 1650—less than 2,500 a year on average.43 From 1651 to the operational beginnings of the second West India Company in 1674, the annual average increased substantially to just over 4,000, or nearly 5,000 in the 1660s alone, as the Dutch supplied slaves to most markets in the Americas, ranging from New Amsterdam in the north to Río de la Plata in the south. This was the period, before the exclusif policies were fully in place, in which the Dutch came closest to the role assigned to them in the traditional historiography, but they never threatened Portuguese dominance of the slave trade and were soon overtaken by the English.44 Between 1674 and 1730, West India Company vessels and various interlopers carried off 176,400 slaves by our reckoning, or 3,100 a year. This was actually one-quarter fewer than the annual average between 1651 and 1673, despite the fact that the West India Company era was one of rapid expansion in the slave trade as a whole. Our total for both the West India Company and free-trade periods is similar to Postma’s recently revised estimate,45 so that the new material is confined mainly to the pre-1674 era.46 Overall we project a volume of 554,300 slaves carried from the African coast in Dutch vessels in 1596–1829. This contrasts with Postma’s latest aggregate estimate of 501,409 slaves. The peak of the Dutch slave trade came in the fifteen years before the American Revolution, when the Dutch average reached 6,200 a year. This was not greatly different from the 1660s a century earlier. The largest of the North European national traders, the English, initially faced the same problem as the other colonial powers in the early seventeenth century: a shortage of slave markets in the Americas. The largest English slave colony in the 1630s, Providence Island in the western Caribbean, was populated with slaves taken from the Spanish rather than with Africans brought in

23

24

David Eltis and David Richardson

on English slave ships, and in any event the island refused to take additional slaves well before the Spanish drove out the English in 1641.47 Apart from the well-known John Hawkins voyages of the 1560s, there is no hard evidence of vessels leaving England to buy slaves, as opposed to gold and produce, until the Star showed up in Barbados in 1641.48 Like the French, the English established one monopoly chartered company after another.49 Unlike the French (and indeed the Dutch), private English merchants dispatched many voyages in defiance of the state monopoly once Barbados was established as a major slave market. The best known of these was the Royal African Company, but its charter was issued by the Crown rather than Parliament and was of uncertain legal standing. The Glorious Revolution of 1688–89 destroyed any hope of parliamentary sanction, and when Parliament finally took action, it was to destroy the monopoly rather than enforce it. From 1698 any English citizen could trade in Africa upon payment of 10 percent of the value of the outbound cargo to the Royal African Company, a requirement that was removed in 1711. The company began a long decline and ceased trading slaves in the 1720s. Where London had been the only important slave trading port before 1700, Bristol now expanded rapidly and then gave way to both Liverpool and, later, a resurgent London. Lancaster and Whitehaven were briefly important. The English plantation sector was already the largest in the Americas by 1700, and the annual value of British plantation produce increased between fourand fivefold in the 1700–1770 period (around fourfold in real terms).50 But English slave traders also dominated markets in the Spanish and, eventually, Dutch Americas. Perhaps as many as one-quarter of the slaves on British vessels finished their journeys in the non-English Americas.51 Except for a few thousand slaves sold in the Río de la Plata region by Luso-Brazilian traders from Rio de Janeiro, no other European slaving power played this intraAmerican trade role between 1730 and 1790. The English, in short, dominated the slave trade when it was at its peak. The British slave trade estimated here encompasses voyages originating in the British Isles, the British Caribbean, and foreign ports conducted under the British flag. Voyages from mainland North America are counted as part of the U.S. slave trade. TSTD2 contains 12,029 British ventures. The twenty-one of these voyages that are scattered over the 1556–1640 period are accepted as a complete record, partly because the size of the trade—and thus any error generated by such an assumption—could only be very small, and partly because it is not certain that all these vessels were seeking slaves as opposed to African produce.52 TSTD2 contains eighty-seven British slave voyages from 1641 to

A New Assessment of the Transatlantic Slave Trade

1661. Nevertheless, this is the period where the evidentiary basis of the English trade is weakest. The slave population of Barbados grew from perhaps 400 in 1640 to an estimated 27,000 in 1660, and there can be no doubt that natural rates of population growth were negative at this time (despite the relatively balanced sex ratios). British slave vessels at this time left Africa with only 154 slaves on average, and one-quarter of the above TSTD2 sample of eightyseven went to markets other than Barbados. These vessels could not have accounted for as many as 10,000 slaves arriving in Barbados. Given that the Dutch supplied few slaves to the British Caribbean, then the English slave trade must have been substantially greater than the TSTD2 sample suggests. A negative rate of natural population growth of −4 to −5 percent a year would have meant arrivals in Barbados of three times the level that TSTD2 suggests, implying an annual average of about 2,300, or 3,000 a year leaving Africa to all destinations.53 From 1662 the British Caribbean expanded strongly beyond Barbados, and English merchants began to make major inroads into Spanish American markets for the first time. For the period 1662–97, TSTD2 includes most English slave voyages, but certainly not all. A full discussion of the adjustments for missing voyages and the resulting estimates for this period is presented elsewhere.54 Suffice it to say here that 312,000 slaves are estimated to have been carried from Africa on English slave ships in this thirty-six-year period, or 8,700 per year. From 1698 to 1712, as already noted, owners of English ventures to Africa (slave and produce vessels alike) were required to pay 10 percent of the outbound cargo to the Royal African Company. A complete record of the revenues from this levy has survived, including all those ventures setting out from colonial as well as home ports, broken down by vessel and intended destination. It is therefore likely that every slave vessel under the English flag in these years is included in TSTD2. As we might expect, given the continued expansion of the British Caribbean at this time, especially Jamaica and Antigua, the English slave trade jumped to even higher levels. Departures from Africa in English vessels in 1698–1712 are estimated at 227,116 or just over 15,000 a year, a 70 percent increase from the previous period. Further expansion occurred in the aftermath of the Treaty of Utrecht. From the 1720s onward, the sources for the English slave trade, like that of its French counterpart, are rich indeed. Port books, a range of London newspapers reporting shipping movements, colonial sources, records of forts on the African coast, seamen’s sixpences records (a levy on wages paid into a fund to support infirm seamen), registers of Mediterranean passes (issued to protect

25

26

David Eltis and David Richardson

vessels from the Barbary powers), and many private papers both overlap and compensate for each other when these record series are discontinuous. It is safe to assume that TSTD2 contains all slave vessels originating from 1713 to 1779 in Bristol, and 95 percent of the ventures leaving all other English ports. For the final period, from 1780 to 1807, the assumption we have made is that TSTD2 includes all British slave voyages from all ports. The result of these procedures is that 2,643,460 slaves are estimated to have left Africa on British vessels between 1713 and 1807. The annual average increased from just over 15,000 in the first decade of the century to 22,200 in the second quarter, and to nearly 32,500 in the years 1751–1807. Levels fell off during the frequent periods of wars, and the peak of the British trade came as abolition approached, when in 1799–1803 British ships carried off 46,300 slaves a year. A summary of the three preceding paragraphs produces an overall estimate of the British slave trade. We estimate that, for the whole period from 1556 to 1810 inclusive, British owners carried off over 3.2 million slaves from Africa. Included in this estimate are all voyages from the British Caribbean and foreign ports. Voyages leaving from the mainland North American colonies that became the United States are allocated to the U.S. flag. Documentation on the French slave trade divides into two periods: before 1710, when coverage is less than comprehensive, and then from 1710, when it surpasses what is available for the British slave trade. French plantation colonies, founded about the same time as those of the English, grew relatively slowly, so that at the end of the seventeenth century they produced only onethird of the exports of the British Caribbean.55 Attempts to establish a systematic and permanent traffic at this time produced a catalogue of missteps and financial disasters. In fact, the Dutch probably supplied most slaves brought to the French Americas before 1673, and many others came on enemy slave ships captured in wartime. If the French could not supply their own colonies, they were not likely to make inroads into non-French slave markets. Only one of the many government-sponsored companies formed at this time, the Guinea Company—transformed into the Royal Asiento Company in 1701—supplied significant numbers of slaves outside the French Americas, though the 16,000 slaves it carried off to the Spanish Americas between 1703 and 1713 was less than half of the 38,000 it contracted to supply. Independent merchants began sending out slaving expeditions from Nantes in the early eighteenth century, and in 1713 this activity was formally sanctioned when the Guinea trade was effectively opened to all.56 The ancien régime supported the slave trade with subsidy schemes to such an extent that foreign ships registered

A New Assessment of the Transatlantic Slave Trade

in France in the 1780s to take advantage of them. Shut down by war in 1793, the slave traffic revived briefly in 1802–4 during the Peace of Amiens, and then, after disappearing again, experienced a much larger revival between 1813 and 1831, this time without subsidy and, for the most part, without legal sanction. The result of all this activity in terms of the number of unwilling people brought across the Atlantic is well documented, though less well for the seventeenth than for later centuries.57 For the pre-1716 period, Pritchard, Eltis, and Richardson in this volume (chapter 7) estimate that close to 96,000 slaves were carried off on French vessels, and that close to 72,000 were disembarked. Growth over time was dramatic, with only 1,800 slaves embarking before 1651, rising to 7,100 in the third quarter and to 29,500 in the fourth quarter of the seventeenth century. But even in the fourth quarter, the volume of departures was still only just over 1,000 slaves a year, increasing to 3,800 in the first years of the eighteenth century. Both these figures are well short of later levels. From 1716 to 1793, as already noted, we argue that 95 percent of French slave voyages are included in TSTD2. Allowing for the missing 5 percent, French vessels carried off 1,081,000 slaves and landed 916,400 in the Americas. Exports increased from an average of 6,400 before 1726 to 10,400 between 1726 and 1750, 13,100 in the third quarter of the century, and an extraordinary 24,000 a year between 1776 and 1793. The nineteenth-century French slave trade is more complicated. For its 1802–4 revival, we rely on Eric Saugera’s work, supplemented by other sources.58 Saugera suggests sixty-five slaving voyages from metropolitan France, but about one-third of these headed for the Mascarene Islands and thus lie outside the mandate of our analysis. TSTD2 points to 10,900 slaves carried from Africa to transatlantic markets and 9,700 arrivals in these years. The French slave trade resumed even before Napoleon reached the island of Elba. Serge Daget’s catalogue of the last phase of the traffic is not quite as comprehensive as its eighteenth-century Mettas–Daget counterpart, even when supplemented by sources that Daget did not consult.59 Our assumption for the post-1813 era is that 10 percent of the voyages are missing. Allowing for these, it is possible to compute an average of 11,300 slaves per year embarked on French ships from Africa between 1814 and 1830, the peak years of the French nineteenth-century trade. Thereafter, the French traffic never attained as many as 1,000 a year for any five-year period. With arrivals in the French Caribbean effectively closed in 1832, slaves went to Brazil or Cuba with ships that used the French flag as a subterfuge to escape detection, rather than because the

27

28

David Eltis and David Richardson

vessel was authentically French. Overall, from 1796 to the end of the traffic, 203,900 are estimated as departing Africa—152,000 in the 1820s alone. For the whole period, the French took just over 1.38 million slaves from Africa, of whom 1.16 million reached the Americas. The trade organized in the ports of mainland North America (not to be confused with the slave trade that delivered slaves there) is not as well documented as in the rest of the English Atlantic world. Conducted in vessels that were among the smallest in the slave trade—carrying on average only 118 captives as they left Africa—and operating outside the major record-gathering institutions of the large cities, its dimensions are less certain than that of any branch of the traffic except the early Brazilian trade.60 It breaks into three distinct phases. The first began in Boston in the mid-1640s and ended shortly after independence. Based mainly in New England, it drew heavily on the Gold Coast and carried slaves mainly to the sugar plantations of the British Caribbean. A second phase began with the gradual shutting down of the slave trade by state laws after U.S. independence and continued to the federal abolition of the slave trade in 1807. This period saw South Carolina emerge as a major organizing center—though New England continued to predominate—and a shift away from the Gold Coast to Upper Guinea and Angola as a source of slaves. On the Americas side, South Carolina, Georgia, and Cuba replaced the British Caribbean markets. Indeed, U.S. slave vessels went anywhere slaves were sold, except for Brazil. A third phase began in the 1830s, when ships delivering slaves to Cuba and Brazil often displayed the U.S. flag in order to discourage British naval vessels from interfering with their activities. Companies owning slave ships certainly operated from U.S. soil at this time, and slave ships flying the U.S. flag certainly plied the Atlantic—usually on the voyage to Africa rather than on the return haul when slaves were on board—but in most cases these ships were actually owned by Cuban or Portuguese slave traders.61 However, the true nationality of slave vessels becomes so unclear in these late years that TSTD2 makes no attempt to assign nationality on a ship-by-ship basis after 1835. Instead, slave vessels sailing to Brazil are assumed to have been Portuguese, and those going to Cuba are taken as Spanish. With the exception of two voyages, the U.S. slave trade is treated as having ended by 1836. Estimates of the U.S. slave trade are thus required only for pre-1836. For this era, TSTD2 contains records of 1,923 U.S. voyages trading in slaves across the Atlantic. These include the 920 Rhode Island ventures that Coughtry listed. Indeed, the Rhode Island trade is far better documented than that from any other region in the Americas.62 TSTD2 also includes 148 voyages from a

A New Assessment of the Transatlantic Slave Trade

recently published set of 1,781 ventures sailing in the 1783–1808 period that includes slave voyages arriving in the United States under foreign flags, as well as vessels plying the intra-American trade that the present analysis ignores.63 There are two central questions on the overall volume of the U.S. trade. The first is, how complete is our knowledge of Rhode Island–based voyages? The second is, how big was the rest of the North American slave trade relative to that based in Rhode Island ports? On the first, the new data set incorporates some sources that are not specific to any single port and that appear equally likely to have recorded voyages from any port in the English-speaking Americas. Thus the second largest European presence in West Africa from 1670 and for most of the eighteenth century was at Cape Coast Castle. Probably more than 80 percent of the slave vessels from the English Americas called at or sailed past this Gold Coast fort, and the fort logs that recorded such movements have survived for some years.64 London newspapers did not record the movements of all such vessels, but there was no known geographic bias toward one North American port over another among those that were recorded. Finally, TSTD2 draws heavily on sources created in the Caribbean—the destination for most U.S. slave ships.65 All such sources are quite independent of most of the North American–based information that Coughtry and McMillin used, and the extent to which they contain evidence of voyages that these scholars did not pick up provides a check on the completeness of their data. Strikingly, new sources from around the Atlantic have added further information on existing Rhode Island voyages, but little about previously unknown voyages. Apart from the corrections noted above, only twenty-eight additional Rhode Island voyages have emerged from these and other sources upon which the new TSTD2 data set has drawn. The twenty-eight are just 3 percent of Coughtry’s total. Thus, while not every Rhode Island voyage is included in the set, it seems highly unlikely that the complete total for Rhode Island could have exceeded Coughtry’s list by much. The new Atlantic-based sources also suggest that for every slave voyage organized on the North American mainland before 1730, there were five that set sail from the British Caribbean. After 1730, by contrast, this ratio was reversed. In other words, the sparse records of mainland colonial slave voyages before 1730 are the result of few voyages, not few records. A second conclusion suggested by the new sources is that Rhode Island was clearly not “synonymous with the North American mainland slave trade,” as Coughtry suggested. If only 948 slaving ventures out of 1,923 in TSTD2 under the U.S. flag left from Rhode Island (920 from Coughtry plus 28 new voyages), then Rhode Island must have sent

29

30

David Eltis and David Richardson

out less than half the traffic from all U.S. ports.66 The range of sources now available is such that the voyage sample for ports outside Rhode Island is likely almost as complete as that for Rhode Island. But it remains less than 100 percent for either group. We have assumed a 10 percent allowance for missing voyages during the 1730–76 period. Before 1730 the sources are less rich, and after independence several states in the Union banned the trade well in advance of the 1807 federal prohibition. Some slave vessel owners therefore had an incentive to make their activities less obvious. For pre-1730 and post1776, we have increased the allowance for missing voyages to 20 percent.67 Two further adjustments are necessary for the postabolition period. Effective January 1, 1808, Congress made the introduction of slaves into the United States illegal and subject to forfeiture, confiscation, and fines. Punishment beyond this was left up to the states. In 1820, however, Congress took the further step of declaring slave trading by U.S. citizens anywhere a capital offense. Just prior to this, in 1819 Amelia Island near the Georgia-Florida border, which had acted as a conduit for smuggled slaves, came under U.S. control with the cession of Florida.68 Between these measures and even beyond this period, some slaves were introduced illegally by U.S. citizens. There is no evidence in nineteenth-century U.S. Census documents of significant numbers of Africanborn individuals consistent with large number of arrivals.69 Nor is there much indication of slave vessels heading for the United States in the abundant reports from either slaving establishments on the African coast or naval patrols after 1808. We thus have allowed for 1,000 slaves a year carried off from Africa to the United States between 1809 and 1820, of which 500 arrived in U.S. vessels. After 1820 there are documented cases of U.S. slavers sailing to Cuba in the late 1820s, and several instances of captives arriving directly from Africa— the Wanderer in 1859 and the Clotilda in 1860 being the best known. Some additional slaves may have arrived in the United States from other parts of the Americas, but they are not accounted for in TSTD2. We can now summarize the estimate of the mainland North American carrying trade. As column 8 in table 1.4 shows, before 1701 North American mainland vessels carried off just 4,150 slaves, mainly in the 1640s in vessels departing from Boston and at the very end of the century in New York–based vessels trading in Mozambique. The colonial-based slave trade remained at low levels for the next quarter century, accounting for only 3,280 departures. Rhode Island’s involvement from 1730 expanded the traffic to 34,000, or 1,360 a year on average, in 1726–50, jumping to 84,600 in the third quarter of the eighteenth century. The Revolution initiated a shutdown of the traffic lasting

A New Assessment of the Transatlantic Slave Trade

almost eight years, after which, despite an increasing degree of legal proscription, the traffic recovered to its pre-Revolution level and was responsible for a further 67,450 unwilling individuals leaving Africa between 1783 and 1800. The overall peak of the U.S. trade, however, was during 1804–7, when South Carolina reopened its ports to the transatlantic trade at the same time that Cuba was attracting huge numbers of captives. The volume of the U.S. trade surged to an average of over 20,000 a year, including 36,200 in 1807 alone. Overall, U.S. vessels carried off from Africa an estimated 305,350 slaves, of which voyages we think TSTD2 contains about 82 percent. The trade rooted in northern Europe, principally Denmark and to a lesser extent northern Germany, may be dealt with more quickly. At the same time that larger western European powers such as England and France broke Spanish control of the Caribbean and established their own plantation complexes complete with slave routes from Africa and sugar production, several Baltic polities attempted a similar strategy with less success. In the half century after 1640, the Duke of Courland (with territories now mainly in Latvia), the Elector of Brandenburg, and the Swedish and Danish crowns tried at various times to ship captives across the Atlantic as part of their colonial ambitions. Each country established monopoly companies, each sought forts on the Gold Coast—the Danes at Christiansborg, the Brandenburgers at Gross Friedrichsburg just west of Cape Three Points—and each acquired access to territory in the Virgin Islands on which they grew sugar, all in miniature emulation of the large western European powers.70 Only the Danes, however were able to maintain a transatlantic plantation structure for any length of time. Brandenburg’s American presence was the island of St. Thomas, access to which they rented for thirty years starting in 1685. The Danish West India Company survived longer than any of its counterparts in western Europe, and the Danish Caribbean islands of St. Thomas, St. Croix, and St. John remained a slave market not only for Danish vessels, but also for North German and Dutch slave vessels, for most of the eighteenth century. Indeed, before 1730, the Baltic companies were, to a degree, a front for Dutch interlopers.71 Estimates of the scale of this activity are well supported by archival work over the last two decades. For the Danish trade, Svend Holsoe, focusing on Caribbean sources, and Per Hernaes, working quite independently on primarily Danish and African sources, have collected what they consider to be close to comprehensive data on this branch of the traffic.72 Encouragingly, their estimates are similar. They include not only Danish but also Brandenburg vessels, most of which sold their slaves at St. Thomas. Hernaes estimates 97,850 slaves

31

Table 1.4 Slaves Carried from Africa under the Flags of the British, French, Dutch, U.S., and Minor Slave-Trading Powers by Decade: TSTD2 Sample Compared to Our Estimated Totals

British

French

American Colonies/ United States

Dutch

Minor Powers

Decades

TSTD2

Estimate

TSTD2

Estimate

TSTD2

Estimate

TSTD2

Estimate

TSTD2

1551–1630 1631–40 1641–50 1651–60 1661–70 1671–80 1681–90 1691–1700 1701–10 1711–20 1721–30 1731–40 1741–50 1751–60 1761–70 1771–80 1781–90 1791–1800 1801–10

2,062 381 11,057 6,928 36,377 49,765 102,413 92,123 151,938 162,317 220,385 237,255 173,607 252,938 356,357 297,681 277,274 337,058 284,263

2,062 381 33,174 26,720 67,469 71,689 112,193 116,495 151,877 167,409 226,192 243,929 175,232 255,345 360,786 301,323 277,276 385,927 283,959

59 0 1,644 636 2,261 8,238 12,853 8,965 24,476 55,762 70,625 95,694 111,601 94,171 131,303 156,519 269,700 69,382 10,393

66 0 1,826 707 2,513 9,149 14,281 9,961 27,195 60,279 74,353 100,730 117,477 99,128 138,216 164,757 283,897 72,983 10,942

3,186 6,452 24,572 23,382 43,732 42,413 40,481 26,881 35,505 22,466 32,009 29,647 37,607 41,039 59,795 47,706 16,772 7,775 1,604

3,520 6,452 24,952 25,982 48,589 43,953 40,478 27,367 35,589 22,464 31,794 29,455 37,607 41,039 59,795 47,706 16,772 7,775 1,604

0 0 660 0 0 499 358 1,383 96 1,499 3,463 15,523 11,046 20,874 33,703 22,184 12,169 35,510 82,016

0 0 824 0 0 623 974 1,729 120 2,020 5,365 17,504 12,273 23,066 37,444 24,838 16,330 50,344 103,922

0 0 1,053 653 0 316 3,364 19,344 4,316 652 1,563 1,090 2,455 6,863 5,359 6,156 17,470 13,023 11,786

Estimate 0 0 1,053 653 0 316 3,729 21,640 4,319 652 1,561 1,178 2,915 8,157 6,108 6,541 18,307 17,597 16,316

1811–20 1821–30 1831–40 1841–50 1851–60 1551–1860 TSTD2 %

0 0 0 0 0 3,052,178 92.6

0 0 0 0 0 3,259,440

34,860 137,565 1,609 0 0 1,302,577 94.2

38,744 152,595 1,609 0 0 1,381,404

734 686 0 0 0 544,458 98.2

734 686 0 0 0 554,336

Source: TSTD2. For estimates, see estimates page at http://www.slavevoyages.org, and text of present essay.

819 1,758 0 0 476 244,036 79.9

5,276 2,197 0 0 476 305,326

0 0 0 0 0 95,463 86.3

0 0 0 0 0 111,041

34

David Eltis and David Richardson

leaving Africa for the Danish Americas, which likely comprised the entire slave trade to Danish territory in the years 1660–1806, and we have incorporated most of Hernaes’s adjustments into our own estimates.73 This figure includes some but not all vessels sailing under the Brandenburg flag. A more complete representation of Brandenburg voyages is provided by Andrea Weindl in this volume, and her estimates are also incorporated into the minor carriers section of table 1.4.74 TSTD2 has 68,391 captives carried from Africa under the Danish flag, 25,169 in Brandenburg vessels, and 2,239 in Swedish and Courland vessels, for a combined TSTD2 total of 95,798 during the whole period of the slave trade. We assume that all Brandenburg, Swedish, and Courland voyages are included in the new data set, and we largely accept Hernaes’s adjustments for the Danish traffic. These decisions generate a total for the Danish slave trade of 83,618 and for the slave trade of all the minor powers of 111,041 slaves carried from Africa over the entire trade period. Thus TSTD2 contains records of 86.3 percent of our estimate of the slave trade of the minor powers. After allowing for voyage mortality, 91,743 are estimated to have reached the Americas. Compared to our estimates for all other national groupings, the series for the minor carriers in table 1.4 displays a highly unusual bimodal distribution. This supports an argument made earlier about the changing structure of the trade over time. The minor carriers were chiefly significant when the traffic was expanding most rapidly in the last quarter of the seventeenth century and again in the late eighteenth century. The second peak came with the general easing of trade restrictions in the Atlantic world, which applied to slaves just as much to goods. The impact of the Bourbon reforms on the Spanish Americas from 1788 on is one example of this effect. Abolition put an end to this minor surge. Between 1700 and the 1780s, the role of the minor carriers diminished considerably, in response to the exclusionary economic policies of the larger European powers. We now return to the Iberian traffic, which we had left in the midseventeenth century just as the rest of the European maritime powers were making inroads into the transatlantic slave traffic for the first time. The Spanish bookended 367 years of slave trading. They initiated the business, albeit from Europe rather than from Africa, and they also carried the last African slaves into the Americas. But for a century and a half, when the slave trade was reaching its zenith, they were little involved in either the African or the transatlantic phase of the business. After 1640 and down to the 1648 Treaty of Westphalia, the Spanish crown issued some contracts, but none as far as we know were to Portuguese, Dutch,

A New Assessment of the Transatlantic Slave Trade

or English nationals. Yet of only nine voyages to the Spanish Americas included in TSTD2 for this decade (compared with 191 between 1631 and 1640), four were English, one was Dutch, and two were Portuguese. Clearly the volume of the traffic declined, and unlicensed trading increased. With the breakup of the Iberian union, Spain was shut out of the Portuguese factories on the African coast, and Spanish involvement in the traffic fell. Between 1648 and 1662, no asiento arrangements of any kind appear to have been in place.75 TSTD2 contains records of seventy-one slave vessels sailing to the Spanish Americas between 1650 and 1662—between five and six ventures a year, bringing in about 1,500 slaves. Just over half were Dutch, but there were thirteen Portuguese and fourteen Spanish, as well as other Swedish and English voyages. In 1663 a new asiento with the Genoese partnership of Grillo and Lomelin provided for the delivery of 3,500 slaves a year, of which less than two-thirds actually arrived. The volume of the trade between 1642 and the end of 1662 was certainly below the 3,100 a year brought in during the final five years of the pre-1641 Portuguese period, and perhaps was also below the 2,100 that Grillo and Lomelin supplied. For the 1642–62 period, we estimate 2,000 slaves a year arriving in the Spanish Americas, or 2,677 leaving Africa, on average, after allowing for shipboard mortality. Thus total departures for the Spanish Americas from Africa in this period were likely to have reached 56,000, as opposed to the just over 8,000 supported by TSTD2.76 We have assumed that one-third of the 56,000 were carried in Spanish vessels, except for three years—1651, 1654, and 1656—when TSTD2 shows departures in excess of the annual estimate provided by our procedures. Our estimated Spanish exports for 1642–62 are thus 19,705, or one-third of all slaves carried to the Spanish Americas, plus an additional 1,499 for 1651, 1654, and 1656.77 After 1662, direct voyages from Africa to the Spanish Americas became uncommon through to the end of the century, given that Grillo and Lomelin and succeeding asientistas drew on markets within the Americas to meet their asiento commitments—mainly the new Dutch entrepôt at Curaçao. We have records of only eight transatlantic voyages to the Spanish Americas between 1663 and 1674, none of them Spanish.78 From 1674 to the early nineteenth century, the Dutch, Portuguese, French, and English in turn dominated this branch of the slave trade, sometimes sailing directly from Africa and sometimes from Curaçao, Jamaica, Barbados, and other intermediary slave markets in the non-Spanish Americas. The occasional Spanish transatlantic voyages that show up in TSTD2 both in this period and subsequently to 1788 are accepted as the sum total of Spanish involvement in the transatlantic traffic.79

35

36

David Eltis and David Richardson

The transatlantic traffic picked up again in 1766, when the Companhia Gaditana began its ultimately unsuccessful attempt to channel all slaves to the Spanish Americas through the gateway of Puerto Rico, but disappeared once more between 1770 and 1788, despite the Spanish government’s purchase of the island of Fernando Po from Portugal and their dispatch of merchants to London and Liverpool to learn the latest techniques.80 The trade-liberalizing Bourbon reforms allowed foreign slave merchants virtually free access to the Spanish Americas from 1789 onward, and subsidies from the Spanish government to Spanish slavers eventually brought the Spanish flag back into the traffic in 1792, an event usually associated with the successful voyage of El Cometa from Havana in 1792 (voyageid 13,252). But the Spanish made no significant new inroads into the transatlantic slave trade until war, revolution, and abolition removed first the French and then the Danes, English, and Americans from the slave-trading scene between 1791 and 1808. It was only in 1808 that the Spanish finally reentered the business in any significant way, and even then the operational base of activities was in Cuba rather than Spain itself. The Spanish shared the large traffic to Cuba with the French when the latter came back into business between 1814 and 1831. Thereafter Spanish traders had the Cuban trade to themselves, though they certainly used the Portuguese flag as a cover in the Cuban trade after the 1835 Anglo-Spanish treaty made life more complicated for them. The sources for TSTD2 are strong on both sides of the Atlantic for the period before 1820, when a treaty made Spanish slave trading subject to interference by British antislavery patrols. With a small exception, we think it likely that TSTD2 contains all the Spanish slave ventures between 1808 and 1820. The exception is Puerto Rico, for which TSTD2 contains few primary records before 1832, when the British opened their first consular office in San Juan. Francisco Scarano’s research on Ponce, the main sugar-producing area of the island is quite helpful, but in the absence of complete documentation for the whole island, we have set the volume of the Puerto Rican arrivals at 5 percent of those of Cuba between 1811 and 1842, and we have assumed that all these captives arrived in Spanish vessels. The central question for the nineteenth-century Spanish trade is, how complete are the records for its illegal phase, beginning in 1821 and ending in 1867? Spanish slave vessels disembarked African slaves in four parts of the Atlantic world during this period. The major market was obviously Cuba. Here TSTD2 has drawn on extensive Spanish, Cuban, British, and African sources, and while we cannot fully identify every ship that arrived with slaves,

A New Assessment of the Transatlantic Slave Trade

we are confident that we have some record of almost all landings. Indeed, given that owners went to great lengths to disguise their activities, the risks of double counting are such that we may have overestimated the number of slaves disembarked in some years. However, an upper-limit estimate of the volume can be derived by dividing the TSTD2 total by 0.9 for the first ten years of arrivals in the illegal period, 1821–30, thus assuming that the TSTD2 series is 10 percent short of the true number. After 1830 the range of sources improves, and the TSTD2 series can be divided by 0.97, thus assuming a 3 percent shortfall. A second entry point in the Americas was the British Caribbean, where Spanish slaving vessels regularly shipwrecked, or to which they were conducted if captured in the Americas by a British naval vessel. TSTD2 likely contains all such cases. A third entry point was Montevideo, where in 1835 and 1836 Spanish, or more accurately Uruguayan, merchants introduced African “colonists” that were really slaves.81 Again we believe that all such arrivals are included in TSTD2. Finally, most Spanish slave vessels that the British captured were taken to courts in Sierra Leone, the Cape of Good Hope, and St. Helena; the court and naval records are such that all these vessels are included in TSTD2. One problem remains: after the 1835 Anglo-Spanish treaty, which made the use of the Spanish flag more difficult, Spanish owners used a number of flags of convenience, especially those of Portugal and the United States, and sometimes no flag at all. Few if any slaves were actually brought into the Spanish Americas by Portuguese or U.S. owners. Between 1836 and 1851, the British navy captured about two slavers sailing to Brazil for every one going to Cuba and Puerto Rico. We have therefore assigned 35 percent of all such captured vessels to the Spanish, and 65 percent to Portuguese/Brazilian owners.82 Table 1.5 summarizes these various decisions into a decadal series of departures of slaves from Africa on Spanish vessels between 1642 and 1867, and compares the results with hard data in TSTD2. The latter contribute 94.1 percent of the estimated total. Table 1.6 integrates the estimates columns from table 1.2 through table 1.5 to present our assessment of the overall size of the trade in decadal groupings from 1501 to 1867. We estimate that all national carriers together took just over 12.5 million slaves from Africa. This is over one and a quarter million more captives than one of us estimated after the first edition of the data set was published in 1999.83 TSTD2 has thus resulted in a 13 percent increase over what we believed to have been the size of the transatlantic slave trade six years ago.

37

38

David Eltis and David Richardson

Table 1.5 Slaves Carried from Africa by Spanish Vessels, 1642–1867 Years 1642–50 1651–60 1661–70 1671–80 1681–90 1691–1750 1751–60 1761–70 1771–80 1781–90 1791–1800 1801–10 1811–20 1821–30 1831–40 1841–50 1851–60 1861–67 1642–1867 Col. 1/col. 2

TSTD2 1,613 5,469 879 4,001 2,294 0 284 3,855 0 510 5,906 13,419 124,236 95,262 231,389 77,004 157,085 52,859 776,065 94.1%

Estimated 8,001 10,389 1,778 4,001 2,294 0 284 3,955 0 510 5,906 13,419 124,236 105,847 245,849 79,464 161,633 54,191 821,755

Source: TSTD2. For estimates, see estimates page at http://www .slavevoyages.org, and text of present essay.

Most of the rise derives from our increased estimates of the Portuguese traffic (11.2 percent), particularly in the pre-1700 era, and from a doubling in the size of the Spanish traffic—amounting to about an extra half million slaves in each case. It should be noted, however, that most of the Spanish increase is the result of reclassifying the nationality of slave ships that we had thought were Portuguese during the pre-1642 and post-1830 eras. The new estimated totals for the much smaller U.S., Danish, and North German traffics are also substantially higher (25 percent on average), but these increases explain only a small part of the rise in the overall total. The estimates for French, British, and Dutch involvement are not greatly different from what they were after 1999, and in the British case the small increase is explained largely by new information on British vessels leaving not from England, but rather from the British Caribbean. The gaps that have been filled in here are gaps that we knew about and made some, though apparently not enough, allowance for six years ago.

A New Assessment of the Transatlantic Slave Trade

Overall, the picture of shifts between slave-trading nations over the entire slave-trade era is similar to what it was at the time of TSTD1, except that the pattern can now be seen to have been even more pronounced. As the consolidated database developed in the early 1990s, it became apparent that, despite the lack of exploitation at the time of sources for Portuguese shipping, the English were not in fact the leading shippers of slaves in the era of the slave trade, as many had supposed.84 Brazil, defined as both colony and nation, received far more slaves than any other polity in the Americas. It now appears that the British dominance of the slave trade was confined to eight of the thirteen decades between 1681 and 1807. Such dominance was sandwiched between two long periods of Portuguese preeminence, during which the British slave trade was trivial. Portuguese hegemony, both before and after the British interlude, was even greater than we thought five years ago. As noted, of course, there were in fact two Portuguese slave trades, determined by the wind patterns of the Atlantic.85 Together they accounted for double the volume of slaves carried by vessels leaving metropolitan British ports. Liverpool has often been viewed as the quintessential slave-trading port, but in fact the ports of Bahia and Rio de Janeiro were individually responsible for far more slaves reaching the Americas. Moreover, within the British slave-trading sector, London now appears to have been twice as important as Bristol and not far behind Liverpool, albeit with a trade that endured over a longer period. A second significant finding relates to the neglected topic of the relationship between independent estimates of departures of slaves from Africa and those for arrivals—mainly but not exclusively—in the Americas (over 200,000 slaves sent out to the Americas actually disembarked in Africa and the Atlantic islands). As explained in the Eltis and Lachance essay in this volume (chapter 12), Curtin used shipping records to estimate that the British, Portuguese, and French, the three major carriers, carried off nearly 5.5 million slaves from Africa during the peak slave-trade period—the “long” eighteenth century (ca. 1701–1810).86 He then derived two independent pictures of arrivals in the Americas, the first by deducting losses in transit from his estimated departures, and the second from a combination of shipping data and estimates of arrivals implied by demographic evidence. He uncovered a major discrepancy between these two methods of estimating arrivals. While for Brazil, supplied entirely by the Portuguese, the difference between the two was negligible, for the Caribbean, voyage records indicated that the French carried half a million fewer than the arrival figure estimated from demographic data, while the English carried over 400,000 more. When Curtin allowed for

39

Table 1.6 Slaves Carried off from Africa for Transatlantic Markets by National Flag and Decade*

Spanish 1501–10 1511–20 1521–30 1531–40 1541–50 1551–60 1561–70 1571–80 1581–90 1591–1600 1601–10 1611–20 1621–30 1631–40 1641–50 1651–60 1661–70 1671–80 1681–90 1691–1700 1701–10

950 4,404 5,495 7,197 13,693 2,823 18,760 13,040 23,695 29,906 25,332 30,979 36,170 25,089 10,240 10,389 1,778 4,001 2,293 0 0

Portuguese 950 5,041 5,495 7,204 13,697 2,826 20,374 16,350 32,959 49,295 65,716 117,003 146,857 88,278 51,275 91,236 109,188 92,660 86,613 162,368 175,140

English 0 0 0 0 0 94 1,591 0 237 0 0 0 141 381 33,173 26,720 67,469 71,689 112,193 116,495 151,877

French 0 0 0 0 0 0 0 66 0 0 0 0 0 0 1,827 706 2,512 9,149 14,280 9,961 27,196

Dutch 0 0 0 0 0 0 0 0 0 1,365 878 951 326 6,452 24,951 25,983 48,592 43,953 40,482 27,363 35,589

Mainland North America/ USA 0 0 0 0 0 0 0 0 0 0 0 0 0 0 824 0 0 623 974 1,730 120

Minor Powers

Total

0 0 0 0 0 0 0 0 0 0 0 0 0 0 1,053 653 0 316 3,729 21,640 4,319

1,900 9,444 10,990 14,402 27,390 5,742 40,725 29,456 56,891 80,566 91,926 148,932 183,494 120,199 123,342 155,687 229,539 222,391 260,564 339,557 394,241

1711–20 1721–30 1731–40 1741–50 1751–60 1761–70 1771–80 1781–90 1791–1800 1801–10 1811–20 1821–30 1831–40 1841–50 1851–60 1861–67 1501–1867

0 0 0 0 284 3,955 0 510 5,905 13,419 124,236 105,847 245,849 79,464 161,633 54,191 1,061,527

200,583 209,128 205,206 221,086 215,934 212,655 210,497 254,899 307,875 393,392 516,854 594,421 438,826 517,078 9,309 0 5,848,268

167,409 226,192 243,929 175,232 255,346 360,785 301,323 277,276 385,928 283,959 0 0 0 0 0 0 3,259,439

60,279 74,353 100,730 117,477 99,127 138,216 164,756 283,897 72,983 10,942 38,744 152,595 1,609 0 0 0 1,381,405

22,465 31,793 29,457 37,607 41,044 59,797 47,712 16,775 7,775 1,605 734 687 0 0 0 0 554,336

2,021 5,364 17,504 12,272 23,066 37,444 24,838 16,331 50,344 103,922 5,276 2,197 0 0 476 0 305,326

Source: Estimated columns of tables 1.2 to 1.5 and estimate page of Voyages Web site. Totals may not add up due to rounding. *Prior to 1526, African slaves carried to the Americas left from Europe, not Africa.

651 1,563 1,178 2,915 8,157 6,109 6,542 18,304 17,597 16,316 0 0 0 0 0 0 111,042

453,408 548,392 598,003 566,589 642,958 818,960 755,667 867,993 848,407 823,554 685,843 855,747 686,284 596,542 171,418 54,191 12,521,334

42

David Eltis and David Richardson

estimated imports into Spanish America, the Danish and Dutch colonies, and the United States after independence during this period, and then compared it with the size of the slave trades of these minor carriers, he arrived at over 1.2 million more slaves coming into the Americas than could have been carried by the British “surplus.”87 Added to the half-million “shortfall” in the French case, this figure comprised 20 percent of his estimate of the total slave trade in the long eighteenth century. In assessing the size of the trade, Curtin had to choose between totals taken from the African side (using shipping records) and totals from the American side (using a combination of shipping and demographic data). He chose the latter and then projected what the departures from Africa must have been if mortality had averaged 15 percent of those taken on board. On this basis, he estimated just under 9.6 million arrivals in the Americas and then about 11.25 million departures from Africa. Curiously, the debate and extensive research activity that Curtin’s book triggered concentrated primarily on his total estimates, rather than on the implicit challenge he had thrown down to scholars to reconcile the differences between his arrivals and departures estimates. One of the main contributions of recent research on shipping data, both transatlantic and intra-American, has been to eliminate these discrepancies between African departures and American arrivals. First, more detailed information on individual voyages has shown significant losses of ships before they embarked slaves at the African coast. Such losses stemmed from natural hazard and from the actions of Africans, privateers, pirates, and European rivals.88 Second, much more is now known about the forced movement of slaves after their arrival in the Americas.89 The largest intra-American movement of slaves before the nineteenth century was in Brazil and was by land, but in the Caribbean the British organized a large water-borne traffic that also redistributed slaves from the British to the Spanish and French Americas. Thanks to the new data in TSTD2, discrepancies in shipping data between departures and arrivals no longer exist. Scholars may now use migration estimates derived from demographic data as an independent check on the voyagebased data, rather than as a method for filling in gaps in the shipping records— at least after 1700. A summary of our assessment of the completeness of the current database is presented in figure 1.2, which lays out a crude time profile of the transatlantic slave trade over three and a half centuries. The top function in figure 1.2 is taken from the new estimates in table 1.6, grouped in twenty-five-year rather than tenyear intervals. The lower function, by contrast, is based on the data taken directly

A New Assessment of the Transatlantic Slave Trade

from TSTD2. The difference between the two functions represents our present assessment of the voyages that occurred but which have not left a record. The major gaps in the coverage of TSTD2 are shown to be in the sixteenth and seventeenth centuries, when it appears that TSTD2 may contain less than half the slave voyages that actually occurred. A smaller gap appears in the nineteenth century during the illegal phase of the traffic. But nearly two out of every three slaves forced to leave Africa left between 1700 and 1825, and for this period we believe that TSTD2 probably includes over 95 percent of all slave voyages. Thus the new estimates—as shown in the final column of table 1.6 and in figure 1.2— support a rather flatter time profile of the slave trade than was apparent in 2001. The pre-1700 era now accounts for 17 percent of the trade, as opposed to 13.5 percent. The post-1800 traffic has also gained slightly in relative importance. For the entire slave-trade era, 1501 to 1867, TSTD2 provides evidence of 9,654,319 slaves leaving Africa for the Americas, compared with a total of 12.5 million estimated in table 1.6. We conclude that, at the present time, TSTD2 contains some record of 77.2 percent of all slaves leaving Africa for the Americas.

2,500,000 TSTD2 data Estimated total

Number of slaves

2,000,000

1,500,000

1,000,000

500,000

1851–1875

1826–1850

1801–1825

1776–1800

1751–1775

1726–1750

1701–1725

1676–1700

1651–1675

1626–1650

1601–1625

1576–1600

1551–1575

1526–1550

1501–1525

0

Years

Figure 1.2 Slaves Leaving Africa by Quarter Century: Estimated, Compared to Recorded in TSTD2 Sources: Table 1.6 and TSTD2.

43

44

David Eltis and David Richardson

Table 1.6 provides summaries of national participation. Portuguese vessels are now thought to have accounted for 46.5 percent of the traffic, and the British around 26 percent. The difference between the two is largely explained by the late arrival of the British into the trade and their early departure with the 1807 abolition act. As long as they continued in the business, they dominated the northern wind and current system almost as much as the Portuguese dominated the traffic south of the equator. Of the others, the Spanish pattern of involvement was the inverse of that of the British. After major involvement with the Portuguese prior to the separation of the Iberian crowns in 1640, the Spanish carried almost no slaves in the eighteenth century prior to the last decade. They then reentered the traffic to the point where they came to dominate the northern wind and current routes as effectively as the British had in the previous century. For the French, the major determinant of participation was war. From 1689 to 1831, the French flag frequently disappeared altogether during hostilities. This apart, the trajectory of their trade was similar to the British, except that it continued for a further quarter century after British withdrawal. The Dutch, by contrast, were most important in the seventeenth century and carried few slaves after the British destroyed a good part of their merchant marine between 1780 and 1784. The annual breakdowns for the participation of each national group across 350 years also provide a basis for estimating the departure of slaves from Africa and the arrival of slaves in the Americas. As table 1.1 shows, the database provides information on where voyages obtained slaves in Africa as well as where they sold slaves in the Americas. While this information is less than complete, for most years—or in the case of the African regions, for five-year periods—it provides a good basis for deriving shares of a nation’s slave traffic with each of eight African and thirty-two American regions that dispersed or received slaves between 1525 and 1867. The national totals can then be distributed across regions in Africa and the Americas according to these regional shares. Departures for any given African region, or arrivals in any given American region, are thus the sum of the disaggregated national totals. Table 1.7 shows the results of this procedure for the eight African regions that are conventionally used to evaluate the slave trade. Table 1.8 does the same for the thirty-two regions in which slaves disembarked, one of which, it should be noted, was in fact Africa rather than the Americas, given the large numbers of detained vessels taken into Sierra Leone, Liberia, Luanda, and St. Helena in the abolitionist era.90 Again the breakdown is provided for decadal intervals.

A New Assessment of the Transatlantic Slave Trade

Such a distribution of captives on both sides of the Atlantic returns us to this chapter’s point of departure. How do the new totals compare with the 2001 attempt to provide an overall assessment of the slave trade’s size?91 As already noted, in 2001 one of us estimated total slave departures from Africa at 11.062 million. We now think that 12.521 million were carried off—an increase of 13 percent. At the time of this writing, aggregate arrivals in the Americas are thought to have been 10.703 million, an 11 percent increase over 2001’s estimate of 9.657 million. On the African side, the infusion of new material has not changed the regional distribution by much. Upper Guinea and the Gold Coast are now seen to have been slightly more important relative to the rest of Africa, and the Bights of Benin and Biafra slightly less. West Central and Southeast Africa, by contrast, remain in the same relative positions. On the American side of the Atlantic, the relative distribution has shifted toward Bahia and away from the British Caribbean, especially Jamaica, with most of the other major regions keeping the shares that they held in 2001. The new data thus have not threatened the dominance of West Central Africa and Brazil, but this is hardly surprising. The major benefit of the new information has been to increase our understanding of the links between Africa and the Americas and, of course, to provide much more detail about those links. Space constraints mean that a fuller explication of this facet of the trade must be reserved for a different occasion. Our major hope is that the revised data set will be used for purposes far beyond estimating the size of the slave trade. Most debates over issues of agency, identity, cultural patterns, gender, and resistance would not exist without an often implicit quantitative component. The present volume of essays is in three parts. The first, “Origins and Destinations,” groups five chapters that provide new information and interpretations for branches of the trade that have received less attention than they should. In addition to the Ribeiro, Domingues and Eltis, and Mendes essays already mentioned, this section includes Philip Misevich’s analysis of slaves from Sierra Leone. He uses the new database to identify the coastal origins of vessels that were captured and condemned in the courts of mixed commission in Sierra Leone and Havana, and then exploits detailed information now available on the Africans onboard these vessels to construct a profile of enslaved peoples not hitherto possible. In an essay that overturns much of the traditional historiography, Oscar Grandío Moráguez argues for the central importance of Angola as a source of slaves arriving in nineteenth-century Cuba.

45

Table 1.7 Departures of Slaves from Major African Regions

Senegambia 1501–10 1511–20 1521–30 1531–40 1541–50 1551–60 1561–70 1571–80 1581–90 1591–1600 1601–10 1611–20 1621–30 1631–40 1641–50 1651–60 1661–70 1671–80 1681–90 1691–1700 1701–10

1,900 8,807 10,990 12,229 23,257 4,796 32,277 22,206 25,448 5,370 9,991 8,541 6,652 4,562 24,476 17,723 6,407 13,267 21,927 22,558 16,344

Sierra Leone 0 0 0 0 0 0 1,168 0 237 0 0 0 0 0 1,372 752 154 0 1,894 2,671 1,217

Windward Coast 0 0 0 0 0 0 0 0 0 2,482 0 0 0 0 0 351 0 0 0 999 3,059

Gold Coast 0 0 0 0 0 0 0 0 0 0 0 68 0 0 2,429 1,437 19,193 28,835 16,274 40,443 81,144

Bight of Benin 0 0 0 0 0 0 0 0 0 0 0 1,873 1,655 1,988 4,092 12,163 29,926 29,813 79,890 108,412 136,943

Bight of Biafra 0 0 0 719 1,361 282 1,867 1,883 0 2,346 0 1,142 2,247 1,630 31,442 24,791 37,668 34,394 21,709 31,299 21,979

West Central Africa

Southeast Africa

All Regions Combined

0 637 0 1,453 2,771 664 5,412 5,367 31,206 70,368 81,936 137,308 172,595 112,020 59,530 95,382 126,758 108,966 109,373 130,939 133,434

0 0 0 0 0 0 0 0 0 0 0 0 345 0 0 3,088 9,432 7,116 9,497 2,237 120

1,900 9,444 10,990 14,402 27,390 5,742 40,725 29,456 56,891 80,566 91,926 148,932 183,494 120,199 123,342 155,687 229,539 222,391 260,564 339,557 394,241

1711–20 1721–30 1731–40 1741–50 1751–60 1761–70 1771–80 1781–90 1791–1800 1801–10 1811–20 1821–30 1831–40 1841–50 1851–60 1861–67 1501–1867

22,669 34,933 44,816 24,210 50,555 52,405 51,267 37,944 28,043 53,702 29,166 13,073 4,626 8,375 0 0 755,512

3,114 9,419 1,468 8,004 17,419 42,296 36,551 31,378 51,119 42,627 22,624 43,543 43,926 21,023 4,795 0 388,771

4,365 4,532 9,392 25,202 44,083 76,521 65,186 36,067 21,176 25,241 7,190 7,867 3,155 0 0 0 336,868

97,287 113,877 106,723 61,626 88,174 108,658 112,562 135,036 109,441 75,746 1,712 5,362 3,293 0 0 0 1,209,320

149,463 194,430 145,805 108,220 122,566 110,383 109,887 113,692 93,197 95,428 74,093 59,250 73,081 108,943 22,528 11,339 1,999,060

34,615 41,830 56,583 93,891 93,294 146,542 109,997 151,242 154,642 140,385 65,870 163,525 97,829 27,554 2 0 1,594,560

Source: TSTD2. For estimates, see estimates page at http://www.slavevoyages.org. Totals may not add up due to rounding.

131,867 145,437 231,989 245,436 223,830 280,240 267,293 333,888 371,789 339,975 407,491 441,968 343,464 387,008 113,927 42,852 5,694,573

10,029 3,934 1,226 0 3,036 1,916 2,924 28,746 19,000 50,450 77,697 121,158 116,910 43,640 30,167 0 542,668

453,408 548,392 598,003 566,589 642,958 818,960 755,667 867,993 848,407 823,554 685,843 855,747 686,284 596,542 171,418 54,191 12,521,334

48

David Eltis and David Richardson

Table 1.8

Arrivals of Slaves in Major Atlantic Markets

Decades

Europe

Northern Mainland Colonies/ States Chesapeake

1501–10 1511–20 1521–30 1531–40 1541–50 1551–60 1561–70 1571–80 1581–90 1591–1600 1601–10 1611–20 1621–30 1631–40 1641–50 1651–60 1661–70 1671–80 1681–90 1691–1700 1701–10 1711–20 1721–30 1731–40 1741–50 1751–60 1761–70 1771–80 1781–90 1791–1800 1801–10 1811–20 1821–30 1831–40 1841–50 1851–60 1861–67 1501–1867

0 452 0 0 0 0 0 0 0 188 0 85 0 0 0 0 916 1,503 0 477 0 0 1,081 2,640 405 1,090 0 0 23 0 0 0 0 0 0 0 0 8,860

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 434 682 56 193 1,461 96 823 709 5,725 5,521 5,140 5,367 149 0 260 338 0 0 0 0 0 0 26,954

0 0 0 0 0 0 0 0 0 0 0 0 100 0 0 630 1,607 2,399 2,228 5,223 12,808 9,025 21,588 28,374 12,094 14,589 12,953 3,874 106 0 68 0 0 0 0 0 0 127,666

Carolinas/ Georgia

Mississippi/ Alabama/ Florida

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 227 2,223 7,851 27,860 2,982 22,856 28,044 24,905 13,363 13,085 66,682 95 0 0 0 303 0 210,476

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 607 6,386 0 222 362 327 2,256 829 672 5,733 4,190 91 0 0 110 0 21,785

Guadeloupe Martinique 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 632 1,560 890 268 0 215 1,008 563 104 218 11,557 16,532 1,573 2,461 9,552 8,054 8,243 9,109 333 0 0 0 72,872

0 0 0 0 0 0 0 0 0 0 0 0 0 0 545 1,222 5,422 2,556 4,427 4,367 8,130 13,788 24,871 20,717 32,692 21,752 7,245 5,378 3,472 22,253 6,942 7,932 22,760 441 0 0 0 216,912

A New Assessment of the Transatlantic Slave Trade

Saint Domingue 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 491 2,855 1,578 2,154 26,207 28,724 61,066 65,782 64,144 104,712 137,258 236,848 40,916 0 808 0 0 0 0 0 773,543

French Guiana 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 853 268 563 862 440 625 1,796 151 854 913 2,320 2,248 886 3,430 2,076 2,300 10,014 0 0 0 0 30,599

Dutch Caribbean

Dutch Guianas

Río de la Plata*

Spanish Caribbean Puerto Mainland* Rico*

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4,083 18,429 28,258 20,964 12,295 13,601 7,473 12,544 5,148 3,365 4,857 6,283 9,199 2,382 1,194 0 0 0 0 0 0 0 150,075

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4,578 5,879 6,162 12,568 10,942 11,234 8,886 12,947 20,795 30,471 31,373 47,883 33,503 12,999 19,077 21,012 812 3,532 0 0 0 0 294,653

0 0 0 0 0 0 0 0 0 0 0 0 0 0 699 2,560 2,929 192 0 0 3,885 6,838 8,878 5,541 2,758 940 0 0 2,180 4,522 21,521 0 2,142 1,659 0 0 0 67,244

0 0 0 655 1,249 324 2,610 1,240 9,719 34,313 32,091 39,250 45,827 32,118 9,311 15,042 3,685 3,969 3,868 6,703 10,863 7,029 1,571 510 2,234 193 357 236 1,154 741 285 353 0 0 0 0 0 267,500

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 305 0 0 0 0 0 0 0 147 236 0 9,757 0 0 477 544 1,238 3,925 9,310 943 0 0 26,882

49

Cuba*

Amazonia

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 336 0 0 623 538 1,559 350 338 0 8,386 0 14,516 41,723 54,167 115,188 136,381 186,179 54,309 126,823 37,124 778,540

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 250 0 846 1,072 1,041 600 669 799 5,559 12,144 13,547 17,540 18,767 31,938 21,229 11,776 2,285 2,169 0 0 142,231

(continued )

50

David Eltis and David Richardson

Table 1.8 (continued)

Decades 1501–10 1511–20 1521–30 1531–40 1541–50 1551–60 1561–70 1571–80 1581–90 1591–1600 1601–10 1611–20 1621–30 1631–40 1641–50 1651–60 1661–70 1671–80 1681–90 1691–1700 1701–10 1711–20 1721–30 1731–40 1741–50 1751–60 1761–70 1771–80 1781–90 1791–1800 1801–10 1811–20 1821–30 1831–40 1841–50 1851–60 1861–67 1501–1867

Pernambuco

Bahia

0 0 0 0 0 0 0 0 0 0 0 0 1,365 0 2,612 100 5,005 1,652 9,589 3,895 18,658 9,189 35,202 21,659 42,494 30,860 6,721 30,860 18,964 22,949 11,392 36,530 19,992 42,433 22,379 32,693 25,000 29,975 45,721 56,325 52,861 60,798 40,000 81,357 38,714 91,074 23,176 91,494 29,427 91,322 30,207 74,749 28,166 66,645 21,799 78,639 27,256 82,622 37,730 97,204 53,869 101,933 81,460 115,337 69,092 97,926 35,158 34,133 19,475 65,022 350 981 0 0 853,834 1,550,355

Southeast Brazil

Danish West Indies

Africa

Other Americas* Barbados St. Kitts

0 0 0 1,340 0 0 0 6,170 0 0 0 7,693 0 0 0 9,428 0 0 0 17,922 0 0 0 3,697 0 0 0 24,770 70 0 0 17,070 1,157 0 0 23,619 3,543 0 0 7,838 6,432 0 0 3,369 15,163 0 0 5,675 21,600 0 0 5,056 21,600 0 0 3,337 15,917 0 172 4,597 25,571 0 1,970 5,396 31,118 0 206 6,052 23,273 196 281 3,651 20,981 2,152 0 4,046 39,428 15,798 493 1,350 41,609 4,061 0 4,341 55,360 1,941 0 2,979 51,207 2,887 259 5,611 64,017 1,013 166 2,266 69,268 2,672 90 3,651 81,391 7,921 308 6,091 84,673 7,180 92 10,687 79,410 6,513 29 9,330 109,660 12,301 281 17,879 119,965 22,118 1,093 16,041 140,860 14,185 6,254 26,952 223,161 0 13,809 15,483 337,888 7,782 38,216 12,513 265,909 277 32,054 5,744 308,114 0 41,800 5,429 5,568 0 8,987 0 0 0 9,011 0 2,263,913 108,997 155,571 307,073

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 25,877 762 16,212 0 32,496 984 29,831 270 40,296 1,463 36,877 0 34,366 166 48,169 2,286 34,233 13,211 27,817 21,770 24,536 18,088 37,666 17,639 49,837 27,082 22,558 19,727 4,711 5,129 20,434 2,892 6,813 2,678 0 0 0 0 433 0 0 0 0 0 0 0 493,162 134,147

Source: TSTD2. For estimates, see estimates page at http://www.slavevoyages.org. * “Other Americas” in part 2 of this table includes voyages arriving in broad regions in the Americas such as Thus, the pre-1641 series presented here for “Spanish Caribbean mainland,” “Puerto Rico,” and “Cuba” are category in this early period. Table 2.7 provides some indication of the true distribution of captives arriving in

A New Assessment of the Transatlantic Slave Trade

Montserrat/ Nevis 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 214 7,111 10,704 3,950 5,061 4,491 5,401 3,820 1,076 3,009 359 1,046 0 151 450 0 0 0 0 0 0 46,843

Antigua

Jamaica

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 97 0 0 0 0 0 0 0 0 0 94 0 12,139 1,234 14,491 1,826 28,659 2,924 35,623 11,956 53,947 8,901 51,442 16,268 75,467 14,735 72,365 10,908 69,977 21,232 84,857 26,389 81,290 6,488 106,047 8,324 97,184 2,034 164,626 1,755 68,901 3,063 0 0 0 0 2,390 0 0 0 0 0 0 138,037 1,019,596

Dominica

Grenada

St. Vincent

Trinidad/ Tobago

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 17,658 30,804 40,735 12,276 8,573 0 0 0 0 0 0 110,046

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 230 0 0 0 233 0 217 462 0 1,100 30,629 27,888 30,791 31,443 4,596 0 0 1,100 0 0 0 128,689

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3,826 10,754 14,612 21,279 8,436 0 0 0 0 0 0 58,907

0 0 0 0 0 0 0 0 0 0 470 0 0 0 0 391 1,083 0 0 0 0 0 0 0 0 0 0 4,682 7,579 8,395 20,128 0 1,078 196 0 0 0 44,002

British Guiana

51

Total

0 1,340 0 6,622 0 7,693 0 10,083 0 19,171 0 4,021 0 28,745 0 21,092 0 41,152 0 59,366 0 70,306 0 117,034 0 145,937 0 94,636 0 99,793 0 127,043 0 188,909 0 182,740 0 213,037 0 283,242 0 334,747 0 383,038 0 466,218 0 502,899 0 481,986 0 551,495 0 696,825 0 659,841 0 767,823 30,647 764,997 41,725 727,471 0 614,700 0 764,222 0 577,601 314 497,574 0 143,122 0 46,135 72,686 10,702,656

the Spanish Americas and Brazil, especially before 1700, without any further indication of port or region. biased downward because such regions in reality received most of the slaves assigned to the “Other Americas” the Spanish Americas before 1685.

52

David Eltis and David Richardson

The book’s second part, entitled “National Slave Trades,” reassesses the organization and structure of some European-based national traffics. James Pritchard, Eltis, and Richardson attempt new estimates of the early French slave trade. Jelmer Vos, Eltis, and Richardson reevaluate the Dutch role in the early modern Atlantic world. Andrea Weindl explores the little-known North German slave trade. All the essays in these first two parts provide underpinning for the estimates presented in this introduction. Part III, “Some Wider Consequences and Implications of the New Data,” examines some of the broader influences on and effects of the slave trade. Manolo Florentino places the Brazilian slave traffic within a broader geographic context by tracking the shifting patterns of African origins on one side of the Atlantic and then tracing the long slave routes within Brazil that slaves followed to their ultimate destinations. There is much need for similar work for slaves arriving in the Caribbean and the North American mainland, where captives have tended to disappear from the view of the historian after disembarking. Roquinaldo Ferreira traces the impact of abolition on the traffic from Angola. Finally, Eltis and Paul Lachance combine the new estimates of the transatlantic traffic with recent research on the intra-American slave trade and population data to reevaluate the demographic experience of Caribbean slave populations. These essays constitute no more than a beginning to the task of digesting the new data. A host of other issues await analysis, including the implication of huge numbers of black females relative to white taken to the Americas before 1820; the vast numbers of children drawn into the traffic; the striking, indeed close to absolute, separation of the North and South Atlantic systems, which was determined not by politics but by the natural environment; and the closer interplay that is now possible to track between fluctuations in the transatlantic slave trade and events in both Africa and the Americas. Another striking feature of the new database is the degree to which it has been built and reinterpreted by a generation of new scholars. Whereas Eltis, Lachance, Pritchard, and Richardson have been combing the archives and publishing on slavery and the slave trade for thirty years and more, and Manolo Florentino is an established scholar still in midcareer, more than half the contributors to the present volume are either publishing for the first time or publishing for the first time in English. These essays thus herald a changing of the guard, an Atlanticization of the pool of talent working on the slave trade, and a shift in the focus of interest from the more northerly branches of the slave trade to the South Atlantic. Many of the major participants in the

A New Assessment of the Transatlantic Slave Trade

earlier debates have now retired or moved on to other topics. Indeed, most of the new contributions of data before the 1999 CD-ROM appeared were made in the 1968–88 period. Even those whose work had focused on reinterpreting the slave trade rather than bringing new data to the table did not return to the subject after the mid-1990s and, as a group, have not engaged in either evaluation or reinterpretation of the masses of new information now available.92 With the appearance of a renewable voyage-based data source, and the major infusion of younger blood as represented in the present essays, we hope that a new era of engagement, debate, and critical evaluation is about to begin. But the issues of this new work will no doubt differ from past concerns. It may seem that the major thrust of the research presented here is to add weight to Curtin’s critics. This is, of course, a central part of the present volume, although we expect that the Hernaes and Inikori estimates of close to or in excess of 13 million transported slaves are too high to find support in voyage-based evidence. The main contribution of the new work in quantitative issues should be viewed as reconciling the discrepancies to which Curtin pointed between the estimates of arrivals in the Americas and estimates of departures from Africa—an issue that few of Curtin’s critics engaged. For the first time in forty years, it is possible to see a broad picture that offers mutually reinforcing data on both sides of the Atlantic. We argue that there is a now a high degree of internal consistency and reliability about the aggregate estimates of the slave trade. More important, we would like to think that TSTD2 will shift attention away from the overall assessment of the slave trade and toward a tighter focus on individual branches of the slave trade and the thousands of ports that were involved in the business. Researchers now have the means with which to both draw on and add details to voyages for all branches of the slave trade. In the process, they will find it possible to address issues of far greater import than merely the size of the slave trade. Perhaps this book will be the last to devote a major part of its thrust to assessing the overall size of the slave trade. NOTES

1. For a fuller presentation of this argument, see David Eltis, The Rise of African Slavery in the Americas (Cambridge, 2000), chap. 3. 2. For those of similar age to the authors, it is a constant source of amazement that small flash drives can store the complete records of all transatlantic slave voyages that historians have discovered in the archives since 1966. 3. Philip D. Curtin, The Atlantic Slave Trade: A Census (Madison, Wis., 1969).

53

54

David Eltis and David Richardson

4. David Eltis, “The Volume and Structure of the Transatlantic Slave Trade: A Reassessment,” William and Mary Quarterly 58 (2001): 17–46. See also the chapter by David Eltis, Stephen D. Behrendt, and David Richardson, “National Participation in the Transatlantic Slave Trade: New Evidence,” in José C. Curto and Renée SoulodreLaFrance, eds., Africa and the Americas: Interconnections during the Slave Trade (Trenton, N.J., 2005), 13–42, for fuller information on the derivation of the 2001 national breakdowns. 5. Per O. Hernaes, Slaves, Danes, and African Coast and Society: The Danish Slave Trade from West Africa and Afro-Danish Relations on the Eighteenth-Century Gold Coast (Trondheim, 1997), 129–71. 6. David Eltis, Stephen D. Behrendt, David Richardson, and Herbert S. Klein, The Transatlantic Slave Trade: A Database on CD-ROM (Cambridge, 1999), henceforth referred to as TSTD1. 7. Eltis, “Volume and Structure.” 8. David Eltis, Stephen D. Behrendt, David Richardson, and Manolo Florentino, The Transatlantic Slave Trade: An Enhanced and On-line Database. The revised database is available via an open-access Web site at http//www.slavevoyages.org and will be periodically updated as new data become available. 9. These figures take into account that sources such as Mettas–Daget (Jean Mettas, Répertoire des expéditions négrières françaises au XVIIIe siècle, ed. Serge Daget, 2 vols. [Paris, 1978–84]) and David Richardson, Bristol, Africa, and the Eighteenth-Century Slave Trade to America, 4 vols. (Bristol, 1986–96), call on more than one source per voyage, but when we entered the volumes in our database we simply use “Mettas–Daget” or “Richardson” as the reference, rather than listing all the references that such volumes draw on. 10. These scholars were Jelmer Vos, Roquinaldo Ferreira, Mariana Candido, Alexandre Vieira Ribeiro, Oscar Grandío Moráguez, Daniel Barros Domingues da Silva, Philip Misevich, Pedro Machado, and Richard Birkett. 11. These scholars have included Franz Binder, Ernst van den Boogaart, Henk den Heijer, Johannes Postma, James Pritchard, Andrea Weindl, António de Almeida Mendes, Manuel Barcia Paz, and José Capela, drawing on archival records in the Netherlands, Spain, Cuba, Portugal, and Mozambique. 12. For a list of the variables, see http://www.slavevoyages.org. 13. Bristol Record Office, SMV/9/3/3. 14. Described in Manolo Florentino, Em costas negras: Uma história do tráfico atlântico de escravos entre a Africa e o Rio de Janeiro (séculos XVIII e XIX) (São Paulo, 1997), 12–13; Herbert S. Klein, The Middle Passage: Comparative Studies in the Atlantic Slave Trade (Princeton, 1978), 51–94; and Corcino Medeiros dos Santos, “Relações de Angola com o Rio de Janeiro (1736–1808),” Estudos Históricos 12 (1973): 7–68. 15. The duty books recorded all departures to Africa from every British port in the Atlantic world between 1698 and 1712. They were a direct result of the 1698 act that destroyed the Royal African Company’s monopoly and allowed all British investors access to the slave trade on payment of a duty worth 10 percent of the outgoing cargo. They may be found in the British National Archives, series T70, pieces 349–58.

A New Assessment of the Transatlantic Slave Trade

16. Eltis, “Volume and Structure.” 17. James A. McMillin, The Final Victims: Foreign Slave Trade to North America, 1783–1810 (Columbia, 2004). McMillin’s database comprises 1,764 voyages but does not distinguish between intra-American vessels and transatlantic expeditions. Our own database is concerned only with the latter, and we have attempted to separate the two categories in the McMillin collection. A larger problem with Final Victims is the extensive double and triple counting of voyages. Among the transatlantic voyages alone, 254 voyages are entered twice (in most cases once when the vessel left port and once when it returned), a further twenty-two are entered three times, and a single voyage appears as five separate entries. If the intra-American data in the set exhibit the same characteristics, then this is not a minor problem. Though not always recognized as such by their authors, duplicate entries are the largest single problem with any multisource database (which historians are now creating in increasing numbers) and for our own database— both first and second editions—elimination or at least reduction of such double counting absorbed more resources than the initial collection of the data. We cannot claim to have eliminated all. 18. To encourage transparency, Paul Lachance and David Eltis have constructed a set of spreadsheets that generate annual estimates for the major ports and countries involved in the trade, and have written an essay explaining the development of the worksheets. The estimates page of the database Web site contains the details of the new estimates and links to supporting material. 19. Enriqueta Vila Vilar, Hispanoamérica y el comercio de esclavos (Seville, 1977). For our detailed calculations of the early Spanish trade and the Portuguese and Spanish shares, see the Web site worksheets and essay. 20. Curtin, Atlantic Slave Trade, 116–17. See also Daniel Barros Domingues da Silva and David Eltis, “The Slave Trade to Pernambuco, 1561–1851,” this volume (chap. 3). 21. For estimated arrivals, see the estimates page of the Web site and the supporting essay and spreadsheets by Lachance and Eltis. 22. Affonso de Escragnolle Taunay, “Subsídios para a história do tráfico africano no Brasil colonial,” in Anais do Terceiro Congresso de História Nacional, vol. 3 (Rio de Janeiro, 1941), 533. 23. Voyageids 40,788 and 49,522, respectively. The first record of a Rio de Janeiro slave vessel in TSTD2 is not until 1597. 24. Stuart B. Schwartz, “A Commonwealth within Itself: The Early Brazilian Sugar Industry, 1550–1670,” in Stuart B. Schwartz, ed., Tropical Babylons: Sugar and the Making of the Atlantic Worlds, 1450–1680 (Chapel Hill, 2004), 156–200. 25. These procedures are described more fully in Domingues and Eltis, “Slave Trade to Pernambuco.” For the Excel files that lay out the calculations, see the worksheets for Pernambuco and Bahia on the Web site. 26. Luiz Felipe de Alencastro, O trato dos viventes: Formação do Brasil no Atlântico Sul, séculos XVI e XVII (São Paulo, 2000), 378. The document is cited as AHU, Angola, caixa 10/64, documento estudado por A. L. A. Ferronha, “Angola,” vol. 1, 119–20. According to Antonil, Bahia was much the largest sugar-producing region in Brazil by this time, and we have assigned 55 percent of the Angola figure to Bahia. See André João Antonil,

55

56

David Eltis and David Richardson

Cultura e opulência do Brasil (São Paulo, 1982; first published 1711), 140–43. We thank Daniel Domingues da Silva for drawing this to our attention. 27. Calculated from TSTD2. The African origins of Bahian vessels are reasonably complete for these years. 28. Calculated from TSTD2. See Pierre Verger, Flux et reflux de la traite des nègres entre le Golfe de Benin et Bahia de Todos os Santos, du XVIIe au XIXe siècles (Paris, 1968). For an overestimate of the Costa da Mina role, see Eltis, “Volume and Structure.” 29. Arquivo Público do Estado da Bahia (henceforth APEB), cod. 439; Biblioteca Nacional do Rio de Janeiro (henceforth BNRJ), Documentos históricos da Biblioteca Nacional do Rio de Janeiro, 110 vols. (Rio de Janeiro, 1929–55), vol. 58, 132. 30. David Eltis, “The Nineteenth-Century Transatlantic Slave Trade: An Annual Time Series of Imports into the Americas Broken Down by Region,” Hispanic American Historical Review 67 (1987): 109–38. 31. Alexandre Ribeiro’s essay in this volume (“The Transatlantic Slave Trade to Bahia, 1582–1851,” chap. 4) has 1,349,724 slaves going to Bahia based on an earlier version of TSTD2. 32. This paragraph is based on the Web site’s Amazonia worksheet 9 and Daniel Barros Domingues da Silva, “The Atlantic Slave Trade to Maranhão, 1684–1846: Volume, Roots, and Organization,” Slavery and Abolition, forthcoming. 33. Readers are invited to inspect the source variables of TSTD2 on the Web site to see the basis for this comment. 34. Eltis, “Nineteenth-Century Transatlantic Slave Trade.” 35. David Birmingham, Trade and Conflict in Angola: The Mbundu and Their Neighbours under the Influence of the Portuguese, 1483–1790 (Oxford, 1966), 137, 141, 154–55; Herbert S. Klein, “The Portuguese Slave Trade from Angola in the Eighteenth Century,” Journal of Economic History 32 (1972): 898, 917; Klein, Middle Passage, 254–55; Maurício Goulart, Escravidão africana no Brasil: Das origens à extinção do tráfico (São Paulo, 1975), 203–5; José C. Curto, “A Quantitative Reassessment of the Legal Portuguese Slave Trade from Luanda, Angola,” African Economic History 20 (1992): 1–25; and idem, “The Legal Portuguese Slave Trade from Benguela: A Quantitative Reappraisal,” Africa, 16 (1993): 101–16. Only Klein has provided voyage data for his series. 36. See the relevant worksheets on the TSTD2 Web site for the detailed derivation of this series. 37. “O comércio de escravos novos no Rio setecentista” (unpublished paper, March 2003), table VIII. We thank Nireu Cavalcanti for permission to cite this work. 38. The complete derivation of the series is shown in the Rio de Janeiro worksheet on the TSTD2 Web site. 39. Russell R. Menard, Sweet Negotiations: Sugar, Slavery, and Plantation Agriculture in Early Barbados (Charlottesville, 2006), 49–66. 40. Han Jordaan, “The Curaçao Slave Market: From Asiento Trade to Free Trade,” in Johannes Postma and Victor Enthoven, eds., Riches from Atlantic Commerce: Dutch Transatlantic Trade and Shipping, 1585–1817 (Leiden, 2003), 219–59; Henk den Heijer, “West African Trade of the Dutch West India Company,” in Postma and Enthoven,

A New Assessment of the Transatlantic Slave Trade

Riches from Atlantic Commerce, 161. The distribution of slaves carried on Dutch vessels was calculated from TSTD2. 41. The estimates of the Dutch traffic presented here are more fully developed in Jelmer Vos, David Eltis, and David Richardson, “The Dutch in the Atlantic World: New Perspectives from the Slave Trade with Particular Reference to the African Origins of the Traffic,” this volume (chap. 8). For their detailed derivation, see the relevant worksheet on the TSTD2 Web site. 42. See Johannes Menne Postma, The Dutch in the Atlantic Slave Trade (Cambridge, 1990), chap. 1, for a discussion of the significance of Binder’s work. See also the source variables of TSTD2 for its impact on our work. 43. The disembarkation number differs slightly from the total of 26,286 imported slaves calculated by Ernst van den Boogaart and Pieter C. Emmer, “The Dutch Participation in the Atlantic Slave Trade, 1596–1650,” in Henry A. Gemery and Jan S. Hogendorn, eds., The Uncommon Market: Essays in the Economic History of the Atlantic Slave Trade (New York, 1979), 369. The differences are accounted for first by the fact that we included in our calculations all slave ships reported on the African coast but for which there is no further information on the arrival side, while Van den Boogaart and Emmer counted purely on the basis of recorded imports in Brazil. Second, unlike them we have excluded the 1,326 slaves captured from Portuguese ships in 1630 and 1636. 44. Wim Klooster recently argued that the Dutch dominated the Atlantic slave trade between 1663 and 1688. See Klooster, “An Overview of Dutch Trade with the Americas, 1600–1800,” in Postma and Enthoven, Riches from Atlantic Commerce, 376. 45. See Johannes Postma, “A Reassessment of the Dutch Atlantic Slave Trade,” in Postma and Enthoven, Riches from Atlantic Commerce, 115–38. 46. The differences are explained in Vos, Eltis, and Richardson, “The Dutch in the Atlantic World.” 47. Karen Ordahl Kupperman, Providence Island, 1630–1641: The Other Puritan Colony (Cambridge, 1993). 48. See voyageid 21,876, although the Dragon cleared London for Angola in 1633 on a voyage that, if successful, must have targeted slaves (voyageid 99,021). 49. Eltis, Rise of African Slavery, 140–42; Hilary Jenkinson, “The Records of the English African Companies,” Transactions of the Royal Historical Society 6 (1912): 185–220. 50. David Eltis, “The Slave Economies of the Caribbean: Structure, Performance, Evolution, and Significance,” in General History of the Caribbean—UNESCO, 2nd ed., Franklin W. Knight, ed., vol. 3, The Slave Societies of the Caribbean (New York, 2007). 51. See David Eltis and Paul Lachance, “The Demographic Decline of Caribbean Slave Populations: New Evidence from the Transatlantic and Intra-American Slave Trades,” this volume (chap. 12). 52. The most recent primary research shows that slaves were present in Barbados prior to the sugar boom in the 1640s, but they may have arrived from some other part of the Americas. See Menard, Sweet Negotiations, 46–47. 53. In the 1660s (when the earliest solid data on sex ratios are available), the proportion of male slaves going to Barbados was 54 percent. For anecdotal evidence of a balanced sex ratio among Barbados slaves in the mid-seventeenth century, see Hilary

57

58

David Eltis and David Richardson

McD. Beckles, Natural Rebels: A Social History of Enslaved Black Women in Barbados (New Brunswick, N.J., 1989). For the estimate here, we have assumed a negative rate of natural increase of 4.5 percent and that 10 percent of new arrivals died in their first year in Barbados. 54. David Eltis, “The British Transatlantic Slave Trade before 1714: Annual Estimates of Volume and Direction,” in Robert L. Paquette and Stanley L. Engerman, eds., The Lesser Antilles in the Age of European Expansion (Gainesville, 1995), 191–215; David Eltis, “The Volume and African Origins of the Seventeenth-Century English Transatlantic Slave Trade: A Comparative Assessment,” Cahiers d’Etudes Africaines 138 (1995): 617–27. Note that the estimates in these publications for the years between 1698 and 1712 have been superseded by new research into T70/352–56, mentioned above. 55. This paragraph is based on James Pritchard, David Eltis, and David Richardson, “The Significance of the French Slave Trade to the Evolution of the French Atlantic World before 1716,” this volume (chap. 7); James Pritchard, “An Incidental Slave Trade: The French in Africa during the Seventeenth Century” (unpublished paper, 2004); James Pritchard, In Search of Empire: The French in the Americas, 1670–1730 (Cambridge, 2004), 215–20, 369; and TSTD2. 56. Robert Louis Stein, The French Slave Trade in the Eighteenth Century: An Old Regime Business (Madison, Wis., 1978), 13–15. 57. See the estimates page of the TSTD2 Web site for the annual breakdowns. 58. Eric Saugera, “Pour une histoire de la traite française sous le Consulat et l’Empire,” Revue Française d’Histoire d’Outre-Mer 76 (1989): 203–29. 59. Mettas’s Répertoire des expéditions négrières françaises is the foundation of TSTD2 for the eighteenth-century French trade; Serge Daget, Répertoire des expéditions négrières françaises à la traite illégale (1814–1850) (Nantes, 1988) has the same status for the nineteenth century. 60. TSTD2, number of cases = 646, standard deviation = 66.8. 61. The data set contains records of 251 U.S. registered slave ships between 1811 and 1861. 62. Jay Coughtry, The Notorious Triangle: Rhode Island and the African Slave Trade, 1700–1807 (Philadelphia, 1981), 241–84, lists 930 voyages. As the author has never provided the voyage-by-voyage sources on which his appendix is based, his work must be taken on trust. Further, his list obviously includes some imputed numbers of slaves, as opposed to data extracted from the historical record. Thus, for example, the number of 114 slaves carried appears far more often in the list than one would expect if it had an evidentiary base. The team of scholars working on TSTD2 has verified some of these voyages, and indeed added information to 709 of Coughtry’s voyages, but we have not been able to verify all. In addition, we have identified ten voyages that are included in Coughtry’s appendix twice and have therefore merged them, reducing Coughtry’s list to 920 voyages. 63. McMillin, Final Victims, CD-ROM. 64. In the T70 series of the BNA, vols. 1540–56. 65. These include the CO28 series (BNA) of Barbados treasurer’s reports of duties paid on slaves disembarked on the island, as well as the better-known Naval Office shipping lists and, for the seventeenth century, the reports of Royal African Company agents

A New Assessment of the Transatlantic Slave Trade

stationed in the Caribbean. See the source variables listed for U.S. voyages in TSTD2 for precise references. 66. This ratio is almost identical to that estimated by David Richardson from a Senate document for the years 1804–7. See Richardson, “Slave Exports from West and WestCentral Africa, 1700–1810: New Estimates of Volume and Distribution,” Journal of African History 30 (1989): 8–9. 67. See the U.S. worksheet on the TSTD2 Web site for the details of this calculation. 68. One of the best-documented cases was the Jesus Nazareno, Capt. Mariano Ferrar (id 41,884), ninety-five captives from which were taken into Georgia in late 1817 via Amelia Island. See the Georgia State Archives, Governor Clayton Papers, 678.364, 3710. See also W. E. B. Du Bois, The Suppression of the African Slave Trade to the United States of America, 1638–1870 (New York, 1896) for the role of Amelia Island, though our own assessment of the size of this trade is well below that of Du Bois. 69. See Sylviane Diouf, Dreams of Africa in Alabama: The Slave Ship Clotilda and the Story of the Last Africans Brought to America (New York, 2007). 70. Adam Jones, Brandenburg Sources for West African History, 1680–1700 (Stuttgart, 1985), 1–11. The Gross Friedrichsburg fort was virtually moribund for the first two decades of the eighteenth century, before the Dutch took it over. Note that Courland briefly attempted a factory in the Gambia. 71. See Jones, Brandenburg Sources, and Andrea Weindl, “The Slave Trade of Northern Germany from the Seventeenth to the Nineteenth Centuries,” this volume (chap. 9). For the Swedish slave trade, see Göran Skytte, Det kungliga svenska slaveriet (Stockholm, 1986). We would like to thank Eva Lachance for assistance in translating sections of this book. 72. Hernaes, Slaves, Danes, 170–233; Svend Holsoe, personal communication (note that Holsoe’s work on this has not been published). See also Weindl, “Slave Trade of Northern Germany.” 73. Not all these voyages had known outcomes, and many have incomplete information on numbers of slaves carried. Slightly different assumptions on both outcomes and imputed numbers on these voyages have resulted in TSTD2’s generating estimates that vary slightly from those of Hernaes. Hernaes did not include in his total those Dutch vessels that carried slaves to St. Thomas for both Brandenburg and Danish companies, apparently under the Dutch flag. These vessels are included as Danish company ships, however, in Waldemar Westergaard, The Danish West Indies under Company Rule (New York, 1917). In the present study, such vessels are included under the Dutch designation. 74. Note that Holsoe’s focus is region of arrival in the Americas, whereas Hernaes’s concern is the size of the Danish trade. The minor carriers section of table 1.4 is not directly comparable with Hernaes’s estimates. 75. See António de Almeida Mendes, “The Foundations of the System: A Reassessment of the Slave Trade to the Spanish Americas in the Sixteenth and Seventeenth Centuries,” this volume (chap. 2). 76. For all counts of voyages see TSTD2, and for estimates of slaves carried see the estimates page of the slave voyages Web site and the accompanying spreadsheets. Note that

59

60

David Eltis and David Richardson

Portuguese slave ships were still arriving in 1641 under the asiento that had expired in 1640, so that 3,134 are estimated for that year. 77. The roughly two-thirds of the slave trade to the Spanish Americas that was not Spanish between 1642 and 1662 is taken to have been Dutch, Portuguese, and English and is accommodated in the estimates for those flags elsewhere in this essay. 78. Mendes, “Foundations of the System.” 79. The many African sources for this period contain little evidence of Spanish transatlantic slave vessels. 80. Bibiano Torres Ramírez, La isla de Puerto Rico (1765–1800) (San Juan, 1968). For a summary of Spanish attempts to reenter the trade, see ibid., 111–18, and David Eltis, Economic Growth and the Ending of the Transatlantic Slave Trade (New York, 1987), 51. 81. Alex Borucki, “The ‘African Colonists’ of Montevideo: New Light on the Illegal Slave Trade to Rio de Janeiro and Río de la Plata (1830–1842),” forthcoming. 82. See the “Africa >1830” worksheet of the estimates page of the slave voyages Web site, which applies this assumption for captured vessels; also the Spanish worksheet, which integrates the resulting estimates of slaves removed from captured Spanish vessels with the estimates of other branches of the nineteenth-century Spanish traffic. 83. Eltis, “Volume and Structure.” 84. For a recent restatement of the older position by Saugera, however, see the Web document La traite des Noirs en 30 questions par Eric Saugera, at http://ww3.ac-creteil.fr/ hgc/spiparticle.php3?id_article=284 (posted January 5, 2003). 85. Domingues da Silva, “Atlantic Slave Trade to Maranhão.” 86. Curtin, Atlantic Slave Trade, 211, table 1.63. 87. Ibid., 212, table 1.64; 219–20. 88. It now seems that about one in ten ventures that actually left port failed to return after selling slaves in the New World. The lowest loss ratios—at least before the privateering activity associated with nineteenth-century independence movements in the Iberian Americas—were on the South Atlantic routes between Angola and Brazil. For the importance of this issue, see Stephen D. Behrendt, “The Annual Volume and Regional Distribution of the British Slave Trade, 1780–1807,” Journal of African History 38 (1997): 187–211. 89. Greg O’Malley, “The Intra-American Slave Trade: Forced African Migrations within the Caribbean and from Islands to the Mainland” (paper presented at the annual meeting of the American Historical Association, Philadelphia, 2006). 90. For a full discussion of the procedures that have generated tables 1.7 and 1.8, see the slave voyages Web site. 91. Eltis, “Volume and Structure.” 92. Some scholars have turned to world history, others have switched to the burgeoning industry of microhistorical studies of individual slave voyages and case studies of single slaves, and a third group has treated the pre-1999 work as set in amber, using it to pursue links between the slave trade and other historical phenomena such as industrialization and demographic trends in populations of African descent.

Part I Origins and Destinations

This page intentionally left blank

Chapter 2 The Foundations of

the System: A Reassessment of the Slave Trade to the Spanish Americas in the Sixteenth and Seventeenth Centuries António de Almeida Mendes

In several senses the slave trade to the Spanish Americas in the sixteenth and seventeenth centuries was the foundation stone of 350 years of forced migration from Africa to the Americas. It was the first major branch of the transatlantic traffic, and it probably retained this premier position in the aftermath of the beginning of the slave trade to Brazil—at least until 1640, when the Portuguese withdrew as the major carriers of slaves to the Spanish Americas. Even after this, the Spanish Americas continued as a major slave market for many years, but for the century and a half before the late eighteenth century, it was supplied via Dutch and English entrepôts in the Caribbean rather than directly from Africa. The first systematic trade in black slaves across the Atlantic was from Seville in Spain, from which the first African slaves were sent in 1501. In 1505 seventeen African slaves were sent under a Crown license to Hispaniola (Santo Domingo in Portuguese documentation) for use in the new copper mines.1 Later in the same year, a further hundred slaves were sent to the island’s gold mines. By 1510 more black slaves had arrived to supplement the primarily Indian labor in 63

64

António de Almeida Mendes

the mines and to work as domestics. We know that in 1508, in different parts of the Caribbean, Nicolás de Ovando had three slaves in his service, Diego Colón had ten, and his wife, María de Toledo, had eight.2 In 1518 Charles I, the future Holy Roman Emperor Charles V, assigned licenses to carry 4,000 black slaves to the Spanish Americas to Lorenzo de Gouvenod, comte de Bresa, over a period of seven years. He, in turn, sold the contract to a partnership of four Italians (Agustin de Vivaldo, Fernando Vasquez, and the brothers Tomas and Domingo de Forne).3 In the same year, the Spanish government authorized Dom Jorge de Portugal (1481–1550), bastard son of the Portuguese King João II, to send four hundred slaves to Hispaniola on his own account. As all these slaves traveled to the Americas from the Iberian Peninsula, it is obvious that a trade in slaves from Africa to Mediterranean Europe (Italy, Spain, and Portugal) long preceded the transatlantic traffic. Six years later, in 1524, the same four Italian traders dispatched three hundred slaves in what was probably the first transatlantic slave voyage from Africa to the Americas—from the Portuguese island of São Tomé to Santo Domingo.4 The switch from Europe to Africa as a source of black slaves occurred quickly. For several of the following decades, the traffic to Cartagena and Vera Cruz provided the opportunity for addressing both the complex issues of international exchange and the myriad day-to-day practical problems associated with moving thousands of unwilling people across a vast ocean. In establishing the massive trade in slaves to the Caribbean, the Spanish thus dominated the inception of black migration from the Old World to the New, just as they had its white counterpart—though in the case of black migration, the Spanish were neither the unwilling participants nor the immediate organizers. This chapter focuses on the size, direction, and salient characteristics of the early stages of four centuries of transatlantic slave trading. Before evaluating the existing literature on the topic and what the data collected in the last four years contribute, it will be helpful to sketch the broad outlines of the emerging market for slaves in the Atlantic world. Five sources of demand for African slaves developed in the fifteenth and sixteenth centuries. One was the offshore Atlantic islands, where Portuguese and Spanish conquests from the 1420s, together with the subsequent shift of the sugar complex from the Mediterranean, created a labor market, part of which was supplied by slaves from the African mainland. The second was the domestic Iberian market, where Muslim slaves had formed a basic feature of both medieval Portuguese and Spanish societies. The arrival in Lisbon of the first slave vessel from Africa in 1441 began a small trade that endured to 1759, initially as a supplement to supplies of

A Reassessment of the Slave Trade to the Spanish Americas

Muslim captives. Obviously, Iberian incursions into the Americas dramatically restructured the Atlantic market for slaves, but the details of that restructuring are not well known. The third source of demand for slaves in the Iberian Atlantic was centered on Santo Domingo, where gold was discovered in Cibao in the late 1490s.5 If African slaves were a supplement to white (Eastern European) and Islamic slaves in Iberia, they were also initially a supplement to indigenous slaves in the Americas, although the horrific mortality among Amerindians soon meant that Africans came to form the dominant part of the American market for slaves. Individual African slaves had arrived in the Americas probably on the first or second expeditions of Columbus, and the buying and selling of Africans, as opposed to indigenes, no doubt began at the same time; but the first systematic trade in African slaves to the Americas began in 1519. Slaves were sent from Castile to Santo Domingo to work in the new mines. The gold deposits in Santo Domingo were exhausted by the 1540s, but a fourth source of demand for slaves almost immediately materialized from the silver mines of the mainland. Slaves still arrived in Hispaniola, but many were now dispatched subsequently to the mainland. A fifth source developed from the nascent export sugar sector, which developed on the island of Hispaniola; this demand peaked in the third quarter of the sixteenth century.6 A small Caribbean demand for slaves was sustained into the next century as sugar cultivation moved to Cuba. The shifting patterns of Old World supplies of slaves, which are slightly better known, broadly followed the incursions of the Portuguese into the African sphere of the Atlantic. The first seaborne African slaves, captured as a result of raids launched from Portuguese vessels, arrived in Lagos from the current Mauritanian coast in 1441.7 An orderly market, centered on Arguim (on the Saharan coast), began a decade later and lasted into the 1540s. All slaves sold in Arguim came from the trans-Saharan slave trade. As explained below, at its peak in the 1520s and 1530s, Arguim dispatched about 2,000 slaves a year, most of them going initially to Lagos in the Algarve and, later in the period, to Lisbon, with a few sent to the Canaries and Madeira.8 A company was formed in 1520 to dispatch slaves directly from Arguim to the Caribbean, in what might have become the first transatlantic trade in slaves between Africa and the Americas, but this company apparently never became operational.9 Supplies from Arguim were first supplemented and then replaced by slaves from the so-called southern rivers of Senegambia, well before the Portuguese established Cach˜eu and Bissau. Santiago and Fogo in the Cape Verde Islands

65

66

António de Almeida Mendes

acted as bulking centers for such slaves. The main Cape Verde factory was established in 1466, but it grew slowly as a source of slaves, only 174 arriving there from Africa in 1491–93. Rapid expansion of the trade dates from the early 1510s, when 3,160 slaves arrived in two and one half years following July 1513.10 This expansion was driven principally by demand from Algarve and Andalusia for domestic slaves and ultimately, perhaps, from the Caribbean, where many of these slaves were eventually sent. But sugar planters on Madeira and the Canary Islands also bought these slaves. When Arguim could no longer gain access to trans-Saharan slaves in the 1540s, the Cape Verde Islands temporarily became the major source of all slaves in the Atlantic slave markets.11 After 1534 the Portuguese king, João III, allowed systematic direct exportation from São Tomé and the Cape Verde Islands to the Americas. Santiago sent seventy-one transatlantic slave ships to the Americas between 1544 and 1550, and others went to the Canary Islands.12 The new transatlantic link had the effect of cutting the slave traffic to Europe from Africa by more than half, although it should be noted that, apart from the specialist slave traffic, ships returning to Portugal with African commodities in these years often carried a few slaves for sale that are harder to track. Further south a different set of circumstances created a third source of slaves. Portuguese sugar plantations in the Gulf of Guinea islands of São Tomé and Príncipe drew slaves from Benin starting in the 1480s, a traffic that continued until relations between Portugal and the king of Benin broke down in 1540, resulting in Benin’s withdrawal from the traffic. While that trade lasted, the Portuguese also took slaves from Benin to African buyers in Elmina. São Tomé began receiving slaves from Angola early in the sixteenth century, a few years after the Portuguese established formal relations with the king of the Kongo in 1493. There were enough Kongo slaves on the island in 1534 to supply the human cargo of the first slave vessel to sail from São Tomé to the Americas—to Hispaniola and no doubt the Cibao gold mines. But it is also clear that some slaves were carried to Portugal itself on vessels returning from São Tomé with the year’s sugar crop, as well as from São Tomé to São Jorge da Mina, where they were exchanged for gold. For several decades after 1540, it appears that all of São Tomé’s slaves came first from the Congo River and later from ports north of the river. São Tomé was thus as much an entrepôt and bulking factory as Santiago in the Cape Verde Islands, as well as a major consumer of slaves in its own right. What is still not clear from the sources is the date of the first direct shipment of slaves from the Congo River to the Americas. This is particularly unfortunate, given that Angola was the dominant

A Reassessment of the Slave Trade to the Spanish Americas

source of slaves in the Atlantic world by the final quarter of the sixteenth century, and was to remain the dominant source for the whole of the subsequent slave-trade era. In summary, the first 150 years of slave trading in the Atlantic saw the transatlantic traffic gradually overwhelm the trade to Europe and the Atlantic islands, becoming the main—in effect the only—human commerce in the area. The supply side was dominated by two major entrepôts, one located in the Cape Verde Islands and one in the Gulf of Guinea. The ultimate source of slaves shifted from the Upper Guinea coast first to the Bight of Benin and eventually to Angola. For the most of the next century—that is, from about 1560 to 1650—Angola supplied at least two-thirds of the slaves that crossed the Atlantic. With this backdrop sketched in, we can now narrow the focus to the transatlantic slave trade itself. Despite its seminal importance to the history of the Atlantic world, the traffic to the early Spanish Americas is probably one of the most understudied and least well known of all major branches of the slave trade. The lack of archival work—relative to other transatlantic slave routes to the Americas in other periods—stands out in particular. As a consequence, many questions that have already been answered for other branches of the traffic have yet to be addressed for the two first centuries of the Iberian slave trade. Much of the current historiography is still based on frequently consulted travelogues and chronicles (by Gomes Eanes de Zurara, Ca’ da Mosto, Duarte Pacheco Pereira, Valentim Fernandes, and Diogo Gomes), and some letters of quittance from the Crown officers published by António Brásio and Braamcamp Freire.13 More broadly, although slavery has generated more interest in Portuguese and Spanish studies in recent years, much of it has focused on the sociological and economic aspects of slavery in the Iberian Peninsula itself and has not encompassed the transatlantic aspects of the topic.14 On the issue of the size of the slave trade, Philip Curtin’s 1969 work for the sixteenth and seventeenth centuries has been the subject of some recent reevaluations, especially for the sixteenth-century traffic into Portugal, but nearly forty years after publication it is still the basic work of reference.15 For the organizational aspects of the trade, scholars still rely on Vitorino Magalhães Godinho’s Os descobrimentos e a economia mundial, and Enriqueta Vila Vilar’s Hispanoamérica y el comércio de esclavos, both very important works but both now over twenty years old.16 Consistent with this, almost all of the voyage entries included in the first edition of the transatlantic slave trade database of

67

68

António de Almeida Mendes

1999 came from Vila Vilar’s tables presented in the final part of her book, which lists the entry of slaves into Hispanic America for the years 1595–1640, the so-called period of the Portuguese asientos. For tracking individual voyages and for constructing overall estimates of the size of the early slave trade derived from such records, Vila Vilar’s book remains without doubt the most important contribution of the last few decades. For the same period of time, we also have the venerable but still invaluable study of Georges Scelle, which provides the legal, diplomatic, and organizational background to the first contracts and licenses for the transport of slaves to the Americas, and Huguette and Pierre Chaunu’s work on the organization of the Sevillian traffic for Hispanic America. The recent study of Marisa Vega Franco supplies much new information for the final part of the Hispanic traffic during the asientos of Grillo and Lomelin (1663–74). For the first decades of the traffic—the period of the individual licenses—the major reference is by Frederick Bowser, supplemented by Maria da Graça Mateus Ventura’s master’s thesis on two individual contracts from 1541 and 1556, respectively. This work also draws on two little-cited but valuable articles by Lutgardo García Fuentes and by Esteban Mira Caballos.17 This scholarship has greatly increased our knowledge of the early slave trade to Spanish America, yet much more can be done with the archival sources. For too long, historians have assumed that the Lisbon earthquake of 1755 eradicated all trade records when it destroyed the Casa da Mina e da Índia (the central Crown agency responsible for Portuguese overseas trade), with its registers of the entries of slaves. The remaining sources are complex and less abundant than they might have been, but those contained in the national archives of Lisbon (Instituto dos Arquivos Nacionais da Torre do Tombo—IAN/TT) and in the Archivo General de Indias (AGI) in Seville, in particular, enable us, by deduction and much cross-referencing, to reconstruct the size and organization of the early Atlantic slave trade. More specifically, about 135 legajos in the Contratación series (focusing on departures from the Old World) and Contaduría series (records of arrivals in the Americas) have enabled the piecing together of many previously unknown slave voyages and the addition of new information on those voyages that were already included in the 1999 data set.18 Such integration of several different sources for the same voyage must be done with great care. For the majority of the early Spanish slave voyages, the captain at the point of departure was different from the one who arrived in Cartagena or Vera Cruz. Given the frequency with which these changes happened, it is likely that the first captain did not die but rather that a change of

A Reassessment of the Slave Trade to the Spanish Americas

command occurred during the ship’s stay in Africa. In almost all of these cases, the archival record of the voyage reads “captain X arrived in place of Y,” while in the case of a captain’s death, the record would explicitly mention this fact. This suggests a specialization of function—perhaps, in the early years, individuals had skills appropriate to one side of the Atlantic rather than the other. Also challenging (as well as unusual in light of later practices) were the exchanges between ships under the same contract (that is, slaves or goods transferred in toto from one vessel to another), which in the majority of cases probably occurred on the African coast, in particular on the Cape Verde and São Tomé islands—the two major centers of supply in the slave trade before 1570. The rewards from this detailed work are large, however. While the 1999 database contained only 505 voyages to the Spanish Americas before 1680 (1.8 percent of the estimated total of 27,233), these new sources support 1,213 voyages, or more than double the 1999 figure. The 1999 database was particularly weak for the years 1526–95, containing only 25 voyages as against 480 for the years 1595–1640. By contrast, the new database now contains 266 records for pre-1595. Just as important is the deepening of information for those vessels already researched by Vila Vilar. Data on crew size, tonnage, and second and third ports of disembarkation, which did not appear in the appendix to her 1977 book, are now available for the first time. Of course, as for any researcher following the archival trails of others, some correction of errors is possible. Vila Vilar’s main preoccupation was arrivals of slaves in the Spanish Americas. From the AGI she collected details of 474 slave voyages, only 180 of which contained any information about their departure from the Iberian Peninsula. Vila Vilar did not include voyages that left Portuguese and Spanish ports to take in slaves on the African coast but then left no record of subsequently having arrived in the Americas. It is now possible to integrate about 250 new departure records of such voyages with Vila Vilar’s records of arrivals. A further 410 voyages now entered in the database contain departure information alone. For these ships we have only the captain and ship names, the vessel tonnage, its armament, details of its license (including the date of issue), and the port of departure, as well as, in some cases, the exact date of departure and the members of the crew, with their respective names, ages, geographic origin, and duties on board. It is highly probable, however, that most of these voyages extracted from the registers of the Casa de Contratación in Seville completed the triangular Atlantic circuit, or at least one part of that route. Even if these were not completed voyages, we know that they at least set out

69

70

António de Almeida Mendes

with the intention of bringing African slaves to the Americas and in many cases actually did embark slaves in Africa. Many of these ships, in turn, would have disembarked slaves in the Americas, perhaps in ports other than Vera Cruz and Cartagena, even though it was strictly illegal to do so. In summary, the new data provide a more complete picture of the slave trade than did Vila Vilar’s appendix, they cover a much wider span of years, and they permit a more reliable estimate of the upper limit of the volume of the slave traffic to the early Spanish Americas than has hitherto been possible. From an organizational standpoint, the traffic to the early Spanish Americas can be separated into two distinct phases. The first might be called the period of the individual licenses, lasting from the beginning of the sixteenth century to 1595; the second, lasting from 1595 to 1640, was dominated by the Portuguese and is often termed the Portuguese asiento period. Before 1595, contracts were issued for individual voyages or for periods of time; after 1595, the asiento was typically granted for a period of years and, of course, for much larger numbers of slaves. The first organizational phase straddles the switch in the main direction of the seaborne slave trade from Portugal to the Atlantic world and the diversification of slave provenance away from Arguim and toward the Cape Verde Islands and locations further south. Complete data on slave arrivals in Portugal from Arguim alone exist for twenty-seven months spanning 1499–1501 and for 1508–11. These show the volume of the traffic from Arguim alone to have been 250 slaves a year. For 1518 and 1519, records of arrivals in Lisbon from all regions are available, and these show 3,000 African slaves a year entering the Portuguese capital by sea. Additional slaves arrived in Lagos and in Spanish ports, particularly Seville. Between 50 and 60 percent of arrivals in Portugal between 1515 and 1535 embarked in Arguim and therefore had previously crossed a good stretch of the Sahara. Another 30 percent passed through the trading center of Santiago and had been acquired in the southern rivers of Guinea, and the remaining 10 percent passed through the island factory of São Tomé.19 As already noted, these slaves derived principally from Benin in this early period, although it should be kept in mind that the great bulk of Benin slaves went to São Jorge da Mina rather than Portugal, or were held in São Tomé to work on sugar plantations.20 This early trade to Lisbon displayed some interesting characteristics that are relevant here, given the fact that the first transatlantic slaves came from the Iberian Peninsula. As can be inferred from appendix 2.1, in 1512–28 there were no more than three or four ships involved in the Arguim-Portugal trade, each one making several voyages a year, one leg of which took three weeks on average.

A Reassessment of the Slave Trade to the Spanish Americas

Similarly, only thirteen vessels and ship captains appear to have been involved in the Arguim-Portugal traffic between 1511 and 1528. As discussed below, these early voyages had high ratios of both females and children. Nevertheless, the shipboard mortality rate averaged only 4.4 percent overall. Out of a total inflow of perhaps 100,000 slaves arriving in Iberia between 1450 and 1550 (a direct trade from Africa to the Caribbean opened in 1526), how many remained in the Iberian Peninsula?21 Lisbon in middle of the sixteenth century had a servile population of around 10 percent of the general public, or 10,000 black slaves.22 A few thousand more would have worked in rural southern Portugal. In Seville, Alfonso Franco Silva indicates that in 1497 there were no more than 250 black slaves.23 Clearly, most slaves in the Iberian Peninsula were employed in domestic duties or in mixed agriculture.24 On the Iberian Peninsula, nothing existed comparable to the inhumane working conditions in the mines and on the plantations of the Americas or to the epidemiological environment of the low-lying areas of the Caribbean that favored sugar production. In addition, female and child ratios, as already noted, were high. Thus neither a high death rate nor a low birth rate account for the small black population of Iberia relative to the high level of arrivals. More probable is that most of the slaves who arrived at Lisbon were reexported to Andalusian ports and after that to Spanish America. Nevertheless, until 1526 the transatlantic traffic from Iberia remained small, amounting to no more than a few hundred a year, with the slaves distributed over a large number of ships whose main function was carrying merchandise and passengers.25 The Spanish Crown issued a decree enabling a formal slave trade to the Indies in 1510, and no doubt the first ships carrying primarily slaves left the Iberian Peninsula shortly afterward. The system became more formalized with the inception of licenses in 1513.26 These measures shaped the beginning of the transatlantic slave trade. The switch from Europe to Africa as the major provenance for slaves took perhaps fifteen years and can be explained by several factors. The Spanish authorities were particularly keen to exclude the Moors from the Indies. They forbade the entry of Moors into the Americas in 1501 and expelled those already present in the Indies in 1504. From the Spanish perspective, slaves transported directly from Africa were less likely to come under Moorish influence. In 1503 a revolt exploded at the Ayati mines in Hispaniola. Black slaves were accused of “teaching bad habits to the aborigines and encouraging them to run away” (cimarronaje).27 The governor of the West Indies, Nicolás de Ovando (1501–8), asked Queen Isabella to stop sending black slaves to the Americas. Isabella died

71

72

António de Almeida Mendes

in 1504, however, and her husband, Ferdinand, who had sugar interests in Aragon, permitted the traffic to start again. Further complaints emerged in 1518, when the inhabitants of Santo Domingo complained of the disobedience of slaves who had come from Castile and asked the king to allow them to obtain “boçais,” or possibly more compliant, slaves direct from Africa.28 In 1525 the Italian merchants mentioned above sent the first ship from São Tomé with three hundred slaves. The following year, other members of the Italian merchant community, Lorenzo Vivaldo and Antonio Pinelo, sent the first 145 slaves from Santiago de Cap Verde to Cuba and Hispaniola, where Vivaldo’s brothers (Justiano and Agustin Vivaldo) had two sugar mills.29 In 1528 the contract reverted to German merchants, the Welsers, who were authorized to export 4,000 slaves over five years. The Welsers’ factors were the merchants Enrique Ehingher and Jeronimo Sayller. By the 1530s the traffic had assumed a systematic character. The asientista who bought the licenses from the Crown in turn sold the piezas (or units of slave labor, commonly related to age) to the proprietor of the ship (almost always a Portuguese captain), who would carry the slaves to the Americas.30 In the 1580s the asientista obtained between 3,000 and 4,000 réis for each pieza from the actual carrier, a potential profit of between 2,000 and 3,000 réis for each slave. Table 2.1 summarizes details of the first voyages sailing to Spanish America from Africa. It shows that, for the first ten years of the transatlantic traffic from Africa between 1526 and 1535, and probably from the time of the first transatlantic slave voyage from Iberia, the ports of disembarkation were exclusively in the Caribbean: Puerto Rico, Santo Domingo, Jamaica, and Cuba. There was close cooperation between the Spanish and Portuguese crowns, and careful oversight by both over slave-trade activities. Initially, slaves were brought from Portuguese entrepôts in Africa to mainly Portuguese ports in Europe, and then most entered Spanish territory either in Iberia or in the Americas. In addition, some Spanish vessels obtained slaves in Portuguese Santiago and carried them directly to Spain. Indeed, during the years 1517–20, the captain-governor of the Arguim factory was Antonio Portocarrero, Marquis of Cádiz, Lord of Huelva, and the descendant of an illustrious Castilian family. On the other side of the Atlantic, Indian slaves were sent from Brazil to Hispaniola by Portuguese traders in the 1520s.31 Portuguese regulations governed the trade. The register (regimento) of São Tomé stipulated that the slaves sent to the West Indies had to be between eighteen and forty years of age (likewise the slaves sent to Mina) and that one-third of the slaves taken onboard had to be female. At arrival, the captains had to deliver the slaves to the factors of the

Table 2.1 Slave Voyages to the West Indies from São Tomé, 1526–35 Captain

Date of Departure from São Tomé

Slaves at Departure

Slaves Arrived

Port of Disembarkation

Place of Purchase

São Maria de Bogoña —

Pero Monteiro

1525

300

—

Santo Domingo

—

André Ferreira

—

1526

300

—

Antillas

—

São João São Maria da Luz São António

Fernam Diaz João Guizado Martim Afonso

1532 1532 11–18–1532

198 202 175

João Eanes

2–11–1534

187

Santo Domingo — Puerto Rico and Santo Domingo Santo Domingo

— — —

São António

—

São Miguel

João Guizado

12–11–1534

173

Santo Domingo

—

André Ferreira

Conceição

Pero Monteiro

6–11–1534

231 260 201 (134 males; 67 females) 200 (134 males; 66 females) 201 (134 males; 67 females) 250 (167 males; 83 females)

Jacome Pais Mateus Fernandes de Abreu André Ferreira — Jacome Pais André Ferreira André Ferreira

167

Jamaica

Ship

Kongo

Factor

—

Sources: IAN/TT, II, 131, 54; II, 131, 54; II, 205, 13; II, 205, 13; II, 180, 19; II, 180, 21; II, 205, 13; II, 187, 82; II, 190, 50; II, 209, 85; II, 196, 47; Fernand Braudel, La Méditerranée et le monde méditerranéen à l’époque de Philippe II (Paris, 1966), vol. I, p. 173.

74

António de Almeida Mendes

Portuguese Crown, one of whom resided in Puerto Rico and the other in Santo Domingo. After confirming the number of slaves alive, the factor would give the captain a certificate that had to be delivered to the Casa da Mina in Lisbon when the ship returned. Slave vessels sold their slaves for gold at this time in the Caribbean, at São Jorge da Mina, and, whenever possible, in Spain.32 Indeed, during the years 1493–1530, the Cibao area furnished three-quarters of the gold that arrived from overseas.33 In 1535 cooperation between the crowns apparently broke down when the Portuguese factor, André Ferreira, was expelled from Santo Domingo by the Spanish administration. Six years later, however, in 1541, the king of Castile signed a contract with two Portuguese brothers named Torres and, in 1556, another with Manuel Caldeira to carry 2,000 slaves from the Cape Verde Islands to the Americas. In 1566 a new contract was signed with merchants resident in Santiago.34 These contracts anticipated the asientos that operated between 1595 and 1640. Under the first, seventy-one ships left Seville and Cádiz and carried 10,669 slaves from Upper Guinea to the Spanish Americas between November 1544 and 1550, an average of 1,700 a year.35 This is not the place to list all the names of the asientistas and their known licenses. According to Lutgardo García Fuentes, however, licenses for 119,377 slaves had been issued for the Spanish Americas by 1599, corresponding to a total of 131,314 slaves taken onboard. Table 2.2 shows the breakdown by decade. No one has attempted to estimate illegal arrivals for this period—unlike for the period after 1595. This may be

Table 2.2 Licenses for the Movement of Slaves to the Spanish Americas, 1510–99 Years 1510–19 1520–29 1530–39 1540–49 1550–59 1560–69 1570–79 1580–89 1590–99 Total

Number of Slaves Licensed 4,865 1,711 9,530 15,345 242 25,213 18,486 28,667 15,318 119,377

Sources: AGI, Contaduría 2, 238, 257 (A), (B); Contratación 5756–5763; Indiferente General 2766.

A Reassessment of the Slave Trade to the Spanish Americas

because the practice of smuggling was more difficult in the sixteenth century. It would be quite mistaken to think of the African coast as wide open for sales of slaves in this early era. Slave supply mechanisms in the interior were not yet in place, and large numbers of slaves were available in only two entrepôts situated off the coast. Both the king of Benin and the king of the Kongo refused to supply slaves at various times, and the former required the king of Portugal to take action when an illegal trader sailed up the Benin River and carried out raids in his territory as a way of acquiring slaves. Nonetheless, illegal voyages certainly occurred. The English and French sold slaves in the Indies from the 1550s to the 1570s, but always under conditions of uncertainty. The second organizational phase began in 1595 and lasted until the Portuguese and Spanish crowns separated in 1640. Whereas before 1595 the crown issued licenses or permits to sell slaves in the Americas, after this date the crown required the merchant to sign a contract undertaking to deliver a set number of slaves within a given period. This was the period of the so-called Portuguese asientos, and as it has received much more attention in the literature than the pre-1595 era, it receives less attention here. Table 2.3 lists the

Table 2.3 Portuguese Asientos, 1595–1640 Years 1595–1600 1601–5, 1606, 1608–1611 1607 1615–22 1622–29 1630 1631–40

Total

{

Asientista

Terms of Asiento

Slaves Arrived (Vila Vilar)

Pedro Gomez Reinel João Rodrigues Coutinho & Gonçalo Rodrigues Coutinho Coroa Antonio Fernandes de Elvas Manuel Rodrigues Lamego Coroa Melchior Gomes Angel & Cristovão Mendes de Sousa

4,250 por 6 anos = 25,500 licenças

30,568

4,250 por 3 anos = 12,750 licenças 4,250 por 5 anos = 21,250 licenças

38,970

— 3,500 × 7 = 24,500 licenças

350 31,646

3,500 × 8 = 28,000 licenças

20,330

— 2,500 × 8 = 20,000 licenças

1,200 20,934

132,000 licenças

143,998

Source: Calculated from Enriqueta Vila Vilar, Hispanoamérica y el comércio de esclavos (Seville, 1977), 23–58.

75

76

António de Almeida Mendes

asientos and compares Vila Vilar’s estimate of arrivals with the numbers of slaves contracted with each asientista. The first of these asientos was granted to the newly Christian Pedro Gomez Reinel. It was the only asiento in the Portuguese period that provided for slaves coming from the Cape Verde Islands as well as from São Tomé and Angola. All subsequent asientos were for São Tomé and Angola alone as the African sources of slaves. The inception of the new system thus coincided with a major shift south from Upper Guinea and toward Angola as a source of African slaves. Relative to other coastal regions, Upper Guinea would not again resume its role as a major supplier of slaves in the Atlantic until 150 years later, and even then only briefly.36 Over the entire era of the slave trade, in fact, only 10 percent of the slaves arriving in the Americas left from Upper Guinea’s shores, whereas Angola was easily the largest single regional source in any period after 1600. On the American side, the vast majority of slaves in the Portuguese era were disembarked in Vera Cruz and Cartagena. Whereas the Greater Antilles had been the dominant markets in the early sixteenth century, the mainland ports of Vera Cruz and Cartagena dominated the market even more strikingly by the first half of the seventeenth century, Hispaniola and Cuba providing the only minor alternatives in the Caribbean. The slave trade in the Portuguese period increased strongly compared to earlier years. If, in fact, the number of licenses between 1505 and 1599 reflects the number of slaves actually carried to the Spanish Americas, then on average between 1,000 and 2,000 slaves a year were carried across the Atlantic between 1510 and 1599 (see table 2.2). For the Portuguese asiento era, taking into account illegal voyages and slaves smuggled on legal voyages, the 132,000 licensed slaves provided for in the asiento over the entire period becomes, according to Vila Vilar, a total of 268,664 slaves arriving in Spanish America between 1595 and 1640; at an average of 6,000 a year, this constitutes a fourfold increase from the previous century.37 Behind this large expansion were surging production of gold and silver on the mainland and the prosperity that this activity generated. In the Potosí mines, at least, the major source of labor was always indigenous rather than African, but there was certainly increased demand for slaves in Peru for nonsilver occupations, demand that would not have existed without the mining activities. The rupture of the Luso-Castilian Union in 1640 had profound consequences for the structure of the traffic, the main one being the loss of the monopoly held by Seville’s House of the Contratación. Some contracts were made in the 1640s, but the 1650s inaugurated an “anarchical and illegal time.”

A Reassessment of the Slave Trade to the Spanish Americas

In 1663 a new asiento was concluded with two Genoese merchants, Grillo and Lomelin, that promised the supply of 3,500 slaves annually over seven years, or 24,500 slaves in total. But, as Marisa Vega Franco reports, during the years of 1663–74 only 15,210 slaves actually entered the Spanish Americas.38 During this period the direct trade from Africa ended, and slaves began to arrive principally from other islands in the Caribbean, especially Curaçao, Barbados, and Jamaica (by now British). Cartagena and Vera Cruz continued as the principal destinations, but a traffic to Portobelo also developed. A crude summary of the volume of slaves delivered to the Spanish Americas is thus possible from the existing literature for the period 1510–1674, though no one previously has actually combined sources in this way. Table 2.2 suggests about 120,000 for the years to 1599. To this we can add Vila Vilar’s estimate of 268,000 for the Portuguese era to total 368,000 arrivals through 1640. There is no existing estimate for 1641–62, and the Grillo and Lomelin contracts after 1662 might properly be excluded on the grounds that most, if not all, of their slaves came from other ports in the Caribbean rather than from Africa.39 The new data add much to the picture of the early traffic to the Spanish Americas. For the period 1525 to 1641, TSTD2 contains 1,209 voyages extracted from the Spanish and Portuguese records of ships that either sailed, intended to sail, or can reasonably be assumed to have intended to sail to the Spanish Americas. Nearly 60 percent of these voyages actually arrived in the Americas. A further 250 sailed with the intention of disembarking in the Spanish New World, though the outcome of the voyage is unknown, and 236 voyages were slave voyages that left the Old World for which a Spanish New World destination has been assumed. We have many records of voyages that failed to disembark slaves (because of shipwrecks or enemy action, for example), but where the outcome is not specified, we have made the further assumption in the analysis that follows that the voyage did in fact disembark slaves according to plan. For the period before 1600, it is instructive to compare the licenses with the new data. Table 2.4 compares five-year breakdowns of licenses issued with the slaves delivered according to TSTD2. Two preliminary comments are necessary. Fuentes provides data for 1518 through to 1599, but in 1595 the asiento system was introduced, inducing a steep fall in the number of licenses issued, although the two systems coexisted for at least five years. At the same time, the coverage of the new database improves greatly beginning in 1595, but it is impossible to separate vessels with licenses, on one hand, from those sailing under an asiento, on the other. To facilitate comparisons between the licenses and voyages, only the years before 1594 are therefore included in the table. It should

77

78

António de Almeida Mendes

also be noted that almost 60 percent of the licenses (accounting for 10 percent of the slaves) were issued for groups of slaves of fifty or fewer, and nearly onethird of the licenses were for groups of fewer than twenty slaves. Unless merchants combined licenses, such low numbers would not normally be sufficient for a slaving voyage, and there is no doubt that many of these small groups of Africans were sent out from Iberian ports in vessels that were primarily carrying merchandise and passengers to the New World. Some of these small groups would indeed have been mariners. Thus, there were many more licenses than there were voyages, and columns 2 and 4 in table 2.4 are not strictly comparable. The transatlantic slave-trade database would not normally be able to track and include merchant vessels carrying slaves as an addition to general cargo. Nevertheless, it is obvious from table 2.4 that the license series offers a more complete record of slave movements in the Atlantic than the new database. Table 2.4 Licenses Issued and Voyages and Slaves Arriving in the Americas, 1511–94 Licenses Years 1511–15 1516–20 1521–25 1526–30 1531–35 1536–40 1541–45 1546–50 1551–55 1556–60 1561–65 1566–70 1571–75 1576–80 1581–85 1586–90 1591–94 All years

TSTD2

Number of Licenses

Number of Slaves

Number of Voyages

Number of Slaves

Preferred Series (Slaves)

0 11 1 57 118 66 83 26 4 2 77 64 76 104 84 98 18 889

0 4,900 2 6,280 5,193 4,048 12,570 2,240 61 272 16,342 9,906 9,222 8,871 15,279 17,784 5,838 118,808

0 0 0 4 7 0 15 57 0 3 9 12 2 40 62 17 12 240

0 0 213 374 1,154 0 1,320 6,591 0 377 833 748 402 6,572 5,842 2,649 2,015 29,090

2,000 5,500 1,413 6,280 6,017 4,048 12,570 6,591 961 2,987 16,342 9,906 9,222 9,039 15,389 17,784 5,838 131,887

Sources: Lutgardo García Fuentes, “Licencias para la introducción de esclavos en Indias y envíos desde Sevilla en el siglo XVI,” Jahrbuch für Geschichte von Staat, Wirtschaft und Gesellschaft Lateinamerikas 19 (1982): 1–46; TSTD2; and chapter text.

A Reassessment of the Slave Trade to the Spanish Americas

The series records permission for nearly 119,000 slaves to enter the Spanish Americas, as opposed to 28,000 documented arrivals, and even if we include in the comparison only those licenses large enough to sustain a slave voyage and exclude all licenses prior to 1526, when the first TSTD2 voyage from Africa occurs, the license series still accounts for more than three times the slaves tracked by TSTD2. Yet TSTD2 and other sources can be used to fill gaps in the license series. There are licenses for only ten slaves from 1547 to 1550, while the new database shows 5,894 slaves arriving in the Americas in these years. From 1554 through 1558, the license series is blank, nor is there anything in the new database. But, as previously noted, one major license was in fact issued to Manuel Caldeira for 2,000 slaves in 1556.40 The pattern in the Fuentes license series over the century is one of a few major licenses, interspersed with a steady stream of licenses for much smaller groupings. Thanks to recent research on this period, it is likely that all the major licenses are now known. Where gaps remain, we can assume that the missing licenses are for smaller groupings and that at most these amount to a few hundred per year. The preferred series in table 2.4 (column 5) thus comprises the higher values of the license versus the TSTD2 series from columns 2 and 4 (in fact TSTD2 is the greater of the two only for the five-year periods of 1546–50 and 1556–60), together with some additions. The additions stem from the Caldeira license and from an allowance of three hundred slaves arriving in small groups for each year for which no license or TSTD2 information exists of any kind. (There are only eleven such years: 1517, 1521, 1523–25, 1532–33, 1552, 1554–55, and 1558.) Finally, an allowance of 2,000 arrivals total is made for all years preceding 1516. These procedures yield 131,887 slaves arriving in the Spanish Americas prior to 1595. How robust is this estimate for the sixteenth century? Its foundation, obviously, is licenses, not voyages, and this immediately raises two questions: first, how many slaves were sent without licenses—in other words, smuggled—and second, how many licenses were issued but not used? Generally, slaves were not as easily available on the African coast in the sixteenth century as in later centuries. The Portuguese bulking centers in Cabo Verde and São Tomé constitute a recognition of this fact. The Portuguese control of these centers also meant that illegal voyages were more likely to have sailed from the African coast than from either of these islands. Voyages routed from the Iberian Peninsula to the Americas via either of these islands were subjected to rigorous checks and cross-checks.

79

80

António de Almeida Mendes

In TSTD2 there are three indications of possible illegal status. One of these is the presence of the AGI’s Escribania series as a source for some voyages. Such a source indicates the existence of an official inquiry into the voyage’s circumstances. The second clue is the occasional absence of any entry in the variable indicating the number of slaves that a vessel intended to carry (variable = slintend); more rarely, no entry appears if a vessel was recorded as carrying many more slaves than the number intended. “Slaves intended” in this context means the number of slaves that the vessel was licensed to carry. The third indicator occurs when the slave ship arrived under a flag other than Spain’s or Portugal’s. By these criteria, TSTD2 contains twenty-seven illegal voyages out of a total of 244 in the database before 1595, and eighteen of these were foreign vessels—all but one English. Most of the foreign voyages were not successful, and it is difficult to see how foreign shippers were a significant source of slaves in the early Spanish Americas. Only nine in the illegal group were Iberian. Of course, there would have been more illegal voyages by Spanish and Portuguese vessels, mainly those that carried more slaves than permitted, but the controls were such that the smuggling ratio was not likely greater than 10 percent of the legal total before 1595. Offsetting this downward bias in the estimate are those licenses that were not used. Sometimes this happened when a vessel was not able to complete its voyage because of a natural hazard or because of hostile human intervention— slave resistance or attacks from other Europeans. In the eighteenth century about one in twenty of all oceangoing ventures failed to complete their voyages as intended, and the ratio of failure was no doubt even greater in the sixteenth century. In addition, the length of time between the license and the departure date was often several months, and for various reasons a project might abort. In such cases, the license would be reissued.41 This would result in an upward bias to the estimate. The assumption here is that the downward bias resulting from smuggling at this time is offset by the upward bias from aborted voyages. The new total estimate of 132,000 slaves arriving before 1595 represents about a 17 percent increase over the 1982 Fuentes estimate on the basis of licenses alone (120,000 slaves arriving before 1599, as opposed to before 1595). For the 1595–1641 period, table 2.5 compares the voyages in Vila Vilar’s appendix (as opposed to the asientos she discusses in the main body of her book) with the latest breakdown from TSTD2. Columns 1 and 2 present her appendix material in five-year groupings. As already noted, Vila Vilar’s assessment of the volume of the trade between 1595 and 1640 was based on slaves that the asientistas undertook to deliver to the Americas, as shown in

A Reassessment of the Slave Trade to the Spanish Americas

table 2.2. She increased the total provided for in those contracts by one-third to allow for smuggling, in order to reach an estimate of the trade’s volume in the Portuguese asiento period. Column 3 of table 2.5, by contrast, applies her allowance of one-third to her voyage-based material—in other words, column 2 is multiplied by 1.33. Columns 4 and 5 present five-year groupings of voyages and slaves disembarked as calculated from the revisions to her voyage data presented in this chapter. As previously explained, these revised data are much more comprehensive, taking into account asiento voyages that sailed from the Old World but then left no record of their arrival in the Americas, in addition to the illegal voyages defined above. Vila Vilar included only vessels that were recorded as arriving in the Americas; moreover, at least ten of these were investigated for illegal activity (as opposed to twenty-eight among the new data). It thus seems reasonable to base a new estimate of the slave trade’s volume in this era on the voyage material rather than the asiento contracts, not all of which, as historians of the early trade well know, were fulfilled. It also seems reasonable to reduce Vila Vilar’s allowance for slaves carried illegally to the Americas. Column 6 of table 2.5 thus presents a new assessment of trade volume based on voyages, not on slaves contracted for, and adding one-quarter to the voyage-based number as an allowance for smuggling. Thus column 6 is column 5 multiplied by 1.25. The new overall estimated total for 1595–1641 is 203,800, or about 4,300 slaves a year on average. This reduces by just under 25 Table 2.5 Estimates of Slaves Carried to the Spanish Americas, 1595–1641 Vila Vilar Years 1595–1600 1601–5 1606–10 1611–15 1616–20 1621–25 1626–30 1631–35 1636–41 All years

TSTD2

Voyages

Slaves

Final Estimate

Voyages

Slaves

Preferred Estimate

138 36 28 25 68 75 46 28 37 481

24,777 6,756 4,985 3,641 12,685 14,259 6,144 3,803 5,350 82,400

33,035 9,008 6,647 4,855 16,913 19,012 8,192 5,071 7,133 109,866

162 59 102 40 161 169 73 96 101 963

28,819 10,665 17,706 5,851 28,845 30,447 10,063 15,563 15,043 163,002

36,024 13,331 22,133 7,314 36,056 38,059 12,579 19,454 18,804 203,754

Sources: Calculated from the appendixes in Vila Vilar, Hispanoamérica; TSTD2 (see text).

81

82

António de Almeida Mendes

percent the number Vila Vilar estimated on the basis of the asiento contracts (including her allowance for illegal voyages). Between 1641 and 1663, there were no further asientos, and for forty years after the reinstitution of the system in 1663, beginning with Grillo and Lomelin, slaves were brought into the Spanish Americas from other parts of the Caribbean rather than directly from Africa. Two completely new environments for shipping slaves to Spanish America dominated the sixty years after 1640, the second replacing the first in 1663. The first, characterized by lack of regulation, saw slave arrivals into the Spanish Americas decline to well below the more than 4,000 a year that held for the first four decades of the century.42 The sugar sector in the Spanish Caribbean bowed out of export activities for more than a century, and silver production declined on the mainland from the mid-seventeenth century. Gauging the extent of the decline in the slave trade is an imprecise exercise. TSTD2 has records of eighty-seven vessels introducing slaves (or in some cases intending to introduce slaves) in the Spanish colonies between 1642 and 1662. The absence of licenses and contracts meant that not only Spanish (sixteen) and Portuguese (thirteen) vessels were involved, but also English (four), Swedish (two), and, most prominently, Dutch (thirty-five), which accounted for half of those vessels with known national affiliations. It might be surmised, however, that most vessels that cannot be assigned a nationality were either Portuguese or Spanish. As we might expect, given Dutch preoccupations with Brazil in the 1640s, Iberian vessels were prevalent in the 1640s, while the Dutch predominated in the 1650s. The Spanish requirement that ships should carry only one slave per ton was probably harder to enforce in this era, and the number of slaves disembarked in this period jumped to its highest level since the Spanish American trade began: 259 slaves per ship on average. As a consequence, the eighty-seven voyages in the new database are estimated to have disembarked just under 21,000 slaves, or about 1,000 per year. The actual mean between 1642 and 1662 no doubt fell somewhere between 1,000 and 4,300; given the trends in Spanish American precious metal and produce exports at this time, it was probably nearer the low end of this range. The guess offered here is 2,000 annually, or 42,000 in total. Combining 42,000 slaves between 1642 and 1662 with the totals in tables 2.4 and 2.5 yields an overall estimate of just over 377,000 slaves introduced into the Spanish colonies directly from Africa from the inception of the transatlantic trade in the early sixteenth century to 1663—the beginning of a significant intraAmerican traffic in slaves. Because of the lack of any previous estimate for

A Reassessment of the Slave Trade to the Spanish Americas

post-1640, comparisons with existing estimates of the volume of the Spanish trade must be confined to pre-1642. For that period the new estimate is about 9 percent below that of Vila Vilar and Fuentes combined. Built largely on newly discovered voyages, this new estimate is also more solid. Moreover, the data set from which it derives affords the basis for a broader investigation than was previously possible of aspects of the slave trade beyond volume. One of the benefits of working with an enlarged data set is throwing new light on the direction of the trade. Before 1641 the main European ports of departure for a transatlantic slave voyage were Lisbon and Seville. These two ports, of roughly equal importance, accounted for four out of five of all voyages for which we have records. In the first quarter of the seventeenth century, Cádiz briefly played an important role because the Spanish Crown, in an attempt to limit the illegal traffic, banned slave ships from leaving Lisbon. This restriction was in force between 1610 and 1622, after which Lisbon returned to its former prominence. In Seville or Lisbon, the captain chartered the ship and presented to the royal officers the authorization for the voyage, which included the license and its date of issue. Two authorized judges from the trade house of Seville or Lisbon then inspected the ship, recorded all the names of the crew, their functions onboard the ship, their geographic origins as well as their physical characteristics (color of hair and eyes, scars, and so on), and the tonnage of the ship and weapons embarked. The pilot was questioned by another pilot on his technical knowledge and his sailing skills. Most pilots were Portuguese, because they were “mais practicos e destros para Angola y sus Rios de Guinea que los castellanos per la continua navegacion que hacen a la dichas partes.”43 Prior to 1615 there were three inspections of the vessel (only one thereafter). The first was made in Seville, the second and third in San Lúcar de Barrameda, the outport of Seville. In Lisbon the process was quite similar. These elaborate procedures, as well as the concentration of the Iberian slave trade in just two ports, reflected in part the Crown’s attempt to assert its control and, of course, to restrict smuggling. On the African side, the two major bulking centers of Santiago and São Tomé drew small numbers of slaves from a range of African coastal outlets. In Upper Guinea, there were small coastal communities on the island of Arguim, at Portudal and Rufisque, and on the southern rivers of Guinea (the Rio Grande, Rio Buguba, and so on). These provided bases for the lançados (literally, “the thrown-out ones”), initially the offspring of Portuguese men and African women. These were first mentioned in the 1490s but quickly became vital to the trading function of the communities. None of the small centers

83

84

António de Almeida Mendes

had forts in the sixteenth century, though Rufisque had what was likely the first synagogue in sub-Saharan Africa. The three most northerly centers normally had ambassadors of the African powers in residence. São Tomé, by contrast, drew on Angola to the south and on Benin in the north, along with the five rivers in the Niger delta over which the king of Benin had influence (the Rios Primeiro, Formosa, dos Forcados, dos Ramos, and Escravos). In Benin captains traded directly with Africans, and there were no counterparts to the lançado communities of Upper Guinea. Angola had two major slave markets. One was centered in the Kingdom of Kongo, from which slaves were carried to São Tomé via the Congo River, with a small but steady traffic in operation in the second half of the 1520s. The first transatlantic shipment of Kongo slaves (albeit sailing from São Tomé) may well have been the Conceição in 1534 (voyageid 46,480). The first mention of “Angola” as opposed to “Kongo” comes in 1535, when João Dias journeyed from São Tomé to a region south of Pinda (Cabinda) in what was likely a combined voyage of exploration and slaving. A regular traffic in slaves was established by the 1540s or 1550s. For most of this period and certainly until the late sixteenth century, the Portuguese operated from their two island centers of Santiago and São Tomé rather than conducting a direct trade from the mainland to the Americas, although illegal traders did carry some slaves directly from Africa to the Americas. Table 2.6 summarizes the African origins of slaves carried to the Spanish Americas in the first century and a half of the direct trade from Africa. The 139,000 slaves in the sample represent close to 36 percent of all slaves carried to the Spanish Americas, as estimated earlier in this essay. It is unfortunately not possible to separate out, in any systematic fashion, those slaves originating in the Kongo from those leaving from further south. In addition, many of the slaves identified in the table as being from Angola and the Bight of Benin embarked on their transatlantic voyage at the island of São Tomé, having been brought there from mainland Africa. The corollary of this is that most of the 4,300 slaves shown as leaving from São Tomé were slaves whose precise African origins cannot be identified. The data for the early period are not abundant, but the table does show the early predominance of Senegambia and the increase in slave departures from that region at the end of the sixteenth century. This increase was associated with the Portuguese defeat in the 1580s of the Buramos led by Sampessão and with the subsequent establishment of a trading base at Cach˜eu in 1595. The slave trade from this area had declined strongly prior to these events.44 But table 2.6 also shows that this increased activity was

A Reassessment of the Slave Trade to the Spanish Americas

Table 2.6 Origins of African Slaves Carried to the Spanish Americas before 1685 (in thousands; row percentages in parentheses) Years 1526–75 1576–1600 1601–25 1626–50 1651–75 1676–85 All years

Senegambia* 1.2 (67.5%) 1.5 (12.4%) 9.6 (13.7%) 3.2 (8.6%) 3.9 (31.5%) 2.8 (46.4%) 22.1 (15.8%)

Bight of Benin

São Tomé**

West Central Africa

Southeast Africa

—

0.4 (22.2%) 0.2 (1.5%) 1.4 (2.0%) —

0.2 (10.2%) 10.3 (86.1%) 57.2 (81.7%) 34.4 (90.6%) 2.6 (21.0%) 2.2 (37.8%) 106.9 (76.4%)

—

1.7

—

11.9

0.1 (0.2%) —

70.0

—

12.3

—

5.9

— 1.6 (2.3%) 0.3 (0.8%) 4.5 (36.5%) — 6.4 (4.6%)

1.3 (10.9%) 0.9 (15.8%) 4.3 (3.1%)

0.1 (0.1%)

Total during Period

37.9

139.8

Source: TSTD2. Totals may not add up due to rounding. *Includes 821 slaves embarked at Sierra Leone. **Includes 504 slaves embarking at Bonny and New Calabar.

quite minor when compared with the expansion of the trade in Angola. From 1576 to 1640, eight or nine out of every ten slaves arriving in the Spanish Americas came from Angola. Dutch aggression and the reemergence of an autonomous Portugal in 1640 disrupted this pattern. After 1640, Spanish America drew on all parts of Africa for slaves, a pattern that continued when the direct link between Africa and the Americas was broken in the last quarter of the seventeenth century. This time the intra-American market rather than what might be termed the intra-African market intervened to disrupt the link. Beginning in the 1660s, the Spanish instituted a 130-year dependence first on the Dutch and then on the British Caribbean for their slaves. In the Americas there were fewer options, given that during the asiento period all vessels were required to disembark their slaves at one of two custom houses, one at Cartagena and the other at Vera Cruz. If vessels carried more slaves to the New World than their licenses or asientos permitted, it was standard practice for these vessels to stop at a Spanish Caribbean port and sell those additional slaves before proceeding to the official markets on the mainland. Thus most of the available information on arrivals of slaves in

85

86

António de Almeida Mendes

the Spanish Caribbean at this time arises from judicial proceedings in which Spanish officials pursued captains who introduced slaves illegally. It is a reasonable assumption that much, if not most, of the allowance for illegal slave trading that has been incorporated into our overall estimate of 377,000 slaves comprised slaves introduced into the Spanish possessions in this fashion. For Vera Cruz, we know of fifty-five voyages that made one stop in the Caribbean before reaching their destination. None of these provide information on the Africans disembarking at this stop, but it is a reasonable assumption that slaves were in fact sold. Jamaica and Campeche were the ports of call in fortythree of these cases (twenty-nine in Jamaica and fourteen in Campeche). By contrast, only two vessels sailing to Cartagena are recorded as making such unexplained stops. Table 2.7 summarizes the data for slaves with known destinations. Once more the sample comprises less than half of those thought to have arrived in the Spanish Americas (and perhaps as little as one-third) at this time. But apart from the likelihood of many illegal slaves having been disembarked prior to the arrivals in Cartagena and Vera Cruz, there are no known biases. The table shows 88 percent of all arrivals disembarking at these two ports—over 90 percent if we include Portobelo, which became a third official port of entry late in the period. After adjusting for a disproportionate share of illegal arrivals in Caribbean ports, it is still likely that over 80 percent of the new arrivals in Spanish America entered through Cartagena and Vera Cruz. Hispaniola, with its early gold-mining sector and later modest exports of sugar and hides, constituted the largest market among the Caribbean islands, accounting for only 6 percent of the sample. Clearly Cartagena, with the rapidly expanding market for slaves in Peru and Chile behind it, was by far the largest single port of disembarkation in the Spanish Americas. Still unexplained is the rapid rise and then decline of Vera Cruz as a rival for Cartagena in the first quarter of the seventeenth century. In summary, for the better part of a century the traffic to the Spanish Americas was unusually concentrated geographically at both the African and American ends of the slave trade (and, indeed, at its point of origin in Europe, too). No other branch of the transatlantic slave trade had this characteristic, with the possible exception of the strong connection that evolved between the Bight of Benin and Bahia in a later period. The traffic to the Spanish Americas was unusual in several other respects as well. It stands out as the first trade to be subject to government regulation. Beginning in 1586, vessels in the asiento trade were restricted to carrying no more than one slave per ton (after an allowance for shipboard voyage mortality

Table 2.7 Destinations of Slaves Carried to the Spanish Americas before 1685 (row percentages in parentheses)

Years

Cartagena

Portobelo

Vera Cruz

1526–50

168 (14.3%) 168 (24.7%) 20,067 (78.8%) 18,920 (37.4%) 18,287 (78.8%) 3,280 (52.9%) —

—

—

—

—

—

—

—

3,163 (12.4%) 27,847 (55.1%) 4,310 (18.6%) —

—

1551–75 1576–1600 1601–25 1626–50 1651–75 1676–85 All years

60,889 (55.2%)

— — 2,125 (34.3%) 629 (20.8%) 2,754 (2.5%)

Source: TSTD2. Totals may not add up due to rounding.

930 (30.8%) 36,249 (32.9%)

Río de la Plata

181 (0.4%) 460 (2.0%) — — 641 (0.6%)

Hispaniola 624 (53.2%) 513 (75.3%) 1,058 (4.2%) 2,839 (5.6%) 135 (0.6%) 259 (4.2%) 1,458 (48.3%) 6,887 (6.2%)

Puerto Rico

Other Spanish Caribbean

89 (7.6%) —

293 (25.0%) —

1,173

580 (2.3%) 168 (0.3%) 21 (0.1%) 207 (3.3%) —

582 (2.3%) 567 (1.1%) —

25,450

331 (5.3%) —

6,202

1,064 (1%)

1,773 (1.6%)

110,257

Total

681

50,522 23,212

3,017

88

António de Almeida Mendes

that varied between 10 and 40 percent), a requirement that was presumably abandoned or fell into disuse after the Spanish switched to Dutch and British Caribbean markets for their slaves.45 This level of restriction was certainly never adopted by any other nation before the end of the eighteenth century. In 1685 the Portuguese instituted a 2.5 slave per ton limit, and a hundred years later the British instituted a series of controls beginning with Dolben’s Act of 1788. Despite the Spanish regulation, asientistas used relatively small vessels prior to 1640. The mean tonnage of 388 slave ships sailing before 1640 was only 111, with only seven in the sample exceeding 200 tons.46 This contrasts with the 455 tons of the five authorized vessels of the Grillo and Lomelin asiento between 1663 and 1674.47 These regulations did not prevent exceptionally high shipboard mortality in the early Spanish trade. Just under 30 percent of those embarked died in the Middle Passage for the period as a whole (number of cases = 114; standard deviation = 20.1 percent). Because most of the information is from the first quarter of the seventeenth century (ninety-two cases), there is no real indication of a trend over time, but it is at least clear that mortality was much higher than in the seventeenth-century Dutch and English slave trades, though somewhat lower than the horrendous 39 percent that occurred on the early voyages from São Tomé to Lisbon. One explanation for this is likely the length of the voyage from Angola (the major port of embarkation) to Cartagena (the major port of disembarkation)— almost a month longer than the passage from Angola to Barbados. In the broadest context, it is difficult not to view the early slave trade to the Spanish Americas as experimental. The world had never before seen the forced transoceanic migration of thousands of people held in appalling conditions below deck for months on end. There was no template or pool of knowledge from which to draw. From the Spanish and Portuguese perspective, the trade was clearly a success or it would not have endured for over a century, and this success guaranteed a succession of imitators who carried the volume of the traffic to many times the 2,000 to 3,000 a year that on average arrived in the Spanish Americas in the 150 years before the 1680s. The huge expansion of the transatlantic traffic that kicked in at the middle of the seventeenth century was undoubtedly predicated in part on the preceding Spanish and Portuguese experience. Later European traders continued to prefer islands or castles, where these were available, on which to base their trade, but those islands came to be located much closer to the sources of slaves than were Santiago and São Tomé. The later traders used larger vessels, carried more slaves per ton than did the early Iberian-based traders, and came to experience lower

A Reassessment of the Slave Trade to the Spanish Americas

shipboard mortality even after allowing for the shorter distances that they sailed in comparison with the early contratantes and asientistas. Ultimately, of course, the later traders came to operate within an environment that closely resembled free trade, something that the Spanish themselves could never contemplate until the late eighteenth century. But on this issue, as on many practices in the trade that cannot be mentioned here, the slow Iberian beginnings accumulated the experience that made possible the eventual dispatch of over 12 million Africans across the Atlantic.

NOTES

1. José Antonio Saco, Historia de la esclavitud desde los tiempos más remotos hasta nuestros días (Paris, 1875), vol. 1, p. 95. 2. Lynne Guitar, “No More Negotiation: Slavery and the Destabilization of Colonial Hispaniola’s Encomienda System,” Revista Interamericana 29 (1999), http://www.sg.inter .edu/revista-ciscla/volume29/guitar.html. 3. In 1527 a contract was signed in Burgos, Spain, with Francisco de los Cobos and some others to supply 600 slaves for the gold mines. AGI, Patronato 246, N. 2, R. 14. 4. IAN/TT, Parte II, Bundle 135, doc. 96. 5. Fascinated by Marco Polo’s description of the Cipango goldmines in Asia, Columbus named Hispaniola’s sites of riverine gold Cibao after them. 6. Genaro Rodríguez Morel, “The Sugar Economy of Española in the Sixteenth Century,” in Stuart B. Schwartz, ed., Tropical Babylons: Sugar and the Making of the Atlantic World, 1450–1680 (Chapel Hill, 2003), 85–114. 7. Gomes Eanes de Zurara, Crónique de Guiné (Paris, 1994). 8. António de Almeida Mendes, “Traites ibériques entre Méditerranée et Atlantique: Le noir au cœur des empires modernes et de la première mondialisation (c. 1435–1550),” Anais de História de Além-Mar 6 (2006): 351–88. 9. Alfonso Franco Silva, La esclavitud en Andalucía, 1450–1550 (Granada, 1992), 55. 10. Maria Emília Madeira Santos and Maria Manuel Torrão, “Subsídios para a história geral de Cabo Verde: A legitimidade da utilização de Fontes escritas Portuguesas através da análise de um documento do início do século XVI (Cabo Verde ponto de intercepção de dois circuitos comerciais),” Instituto de Investigação Científica Tropical, Série Separatas no. 218 (Lisbon, 1989), 527–51. 11. The head of the Sherifian dynasty of Morocco expelled the Portuguese from North Africa beginning in the 1530s. The Portuguese thus had reduced access to the Morisco textiles that were essential to purchasing slaves in Arguim. IAN/TT, I, 85, 102. 12. Archivo General de Simancas, Consejo y Juntas de Hacienda, 23: “Asiento dos escravos negros enviados a América entre 1–1–1544 e 31–12–1550 a partir de Sevilha e de Cádiz (a partir dos livros e registos da Casa de Contratação de Sevilha, 1551).” 13. António Brásio, Monumenta missionaria africana (Lisbon, 1963–91), 21 vols.; Braamcamp Freire, Cartas de quitação de El-Rei D. Manuel I, in AHU, vols. 1–9.

89

90

António de Almeida Mendes

14. See A. F. C. Saunders, A Social History of Black Slaves and Freedmen in Portugal, 1441–1555 (Cambridge, 1982), and more recently Didier Lahon, “Esclavage et confréries noires au Portugal durant l’Ancien Régime (1441–1830)” (PhD thesis, Ecole des Hautes Etudes en Sciences Sociales, Paris, 2001); Jorge Fonseca, Os escravos em Évora no século XVI (Evora, 1997); idem, Escravos no sul de Portugal: Séculos XVI–XVII (Évora, 2002); and Alessandro Stella, Histoires d’esclaves dans la péninsule ibérique (Paris, 2000). 15. Philip P. Curtin, The Atlantic Slave Trade: A Census (Madison, Wis., 1969); Ivana Elbl, “The Volume of the Early Atlantic Slave Trade, 1450–1521,” Journal of African History 38 (1997): 31–75; and Paul Lovejoy, “The Volume of the Atlantic Slave Trade: A Synthesis,” Journal of African History 23 (1982): 473–501. 16. Vitorino Magalhães Godinho, Os descobrimentos e a economia mundial (Lisbon, 1984), and Enriqueta Vila Vilar, Hispanoamérica y el comércio de esclavos (Seville, 1977). 17. Georges Scelle, La traite négrière aux Indes de Castile: Contrats et traités d’assiento, 2 vols. (Paris, 1906); Huguette and Pierre Chaunu, Séville et l’Atlantique: 1504–1650 (Paris, 1955–60); Frederick Bowser, The African Slave Trade in Peru, 1524–1650 (Stanford, 1974); Marisa Vega Franco, El tráfico de esclavos con América (Seville, 1984); Maria da Graça Mateus Ventura, Negreiros portugueses na rota das Indias de Castela (1541–1556) (Lisbon, 1999); Lutgardo García Fuentes, “Licencias para la introducción de esclavos en Indias y envíos desde Sevilla en el siglo XVI,” Jahrbuch für Geschichte von Staat, Wirtschaft und Gesellschaft Lateinamerikas 19 (1982): 1–46; and Esteban Mira Caballos, “Las licencias de esclavos negros a Hispanoamérica (1544–1550),” in Revista de Indias 54 (1994): 273–97. 18. The following sources have been consulted systematically: AGI, Contratación, Legajos 2876 to 2896. This lists asientos registered “in the House of contratación of Seville and the ships that had left San Lúcar de Barrameda, Seville, Cádiz, Lisbon and Canaries islands to load slaves in the Cape Verde islands, Guinea, S. Tomé, Angola, Mina and its Rios” between 1584 and 1658. AGI, Contaduría, Legajos 882–885B, covering the entrances in the port of Vera Cruz between 1600 and 1640. Contaduría, Legajos 1397–1405, ships arrived in Cartagena from 1622 to 1640. 19. António de Almeida Mendes, “Portugal e o tráfico de escravos na primeira metade do século XVI,” Studia Africana 7 (2004): 13–30. 20. Between 1522 and 1527, the books of the factory of São Tomé show a total of 12,260 slaves disembarking in the island. From this total, 2,060 slaves left the island for São Jorge del Mina in eight months alone between June 1528 and February 1529. 21. António de Almeida Mendes, “Esclavages et traites ibériques entre Méditerranée et Atlantique (XV–XVII siècles): Une histoire globale” (PhD thesis, Ecole des Hautes Etudes en Sciences Sociales, Paris, 2007). 22. Cristovão Rodrigues de Oliveira, Lisboa em 1551: Sumário em que brevemente se contêm algumas coisas assim eclesiásticas como seculares que há na cidade Lisboa (Lisbon, 1987). 23. Alfonso Franco Silva, La esclavitud en Sevilla (Seville, 1997), 37. 24. Fuentes, “Licencias,” 16. Twenty-six slaves are recorded as working in the mines of Guadalcanal (Sierra Morena), Andalusia, in 1559 (Alessandro Stella, “L’esclavage en Andalousie à l’époque moderne,” Annales ESC 1 [January–February 1992]: 35–64). 25. Fuentes, “Licencias,” 16.

A Reassessment of the Slave Trade to the Spanish Americas

26. Scelle, La traite négrière. 27. “Negros enseñaban malas costumbres a los aborigenes y practicaban el cimarronaje,” cited in Germán Peralta Rivera, Los mecanismos del comercio negrero (Lima, 1989), 18. 28. AGI, Santo Domingo, Junta de Procuradores en la España, 1518, doc. 77, cited in Caballos, “Licencias,” 275. 29. Wilfrid Brulez, “Marchands italiens dans le commerce américain au XVIe siècle,” Bulletin de l’Institut Historique Belge de Rome 44 (1974): 90–99. AGI, Justicia, 701, provides information regarding relationships between the Vivaldo brothers and other Italian families, but this document is not available for consultation. 30. Different interpretations exist of the pieza. For Philip Curtin, “a pieza was a measure of potential labor, not of individuals . . . the very young, the old and females were defined as fractional parts of a pieza de India” (Atlantic Slave Trade, 22). For Vila Vilar, one pieza taken from Africa represented one slave (Hispanoamérica, 186–93). I am inclined toward Vila Vilar’s assessment. In the first decades (first half of the sixteenth century), in São Tomé only slaves between eighteen and forty years old were embarked. The Royal Regiment of 1529 stipulated that only “slaves in good health, good and fat” could be imported to America as well to São Jorge da Mina. The rejected slaves (escravos de refugo) were absorbed within the São Tomé sugar plantations or imported to Portugal. Voyage mortality in the trade between São Tomé and Portugal for the years 1519–35 averaged 39 percent of those embarked in a voyage averaging two months (fourteen voyages embarking 973 slaves). By comparison, the mortality on the first voyages to the Antilles (duration of about three months) was 18 percent of those embarked (we have only four records, however). A distinction should be made between the values of the pieza in Africa and America: the reduction of a pieza de India would have been made in the Americas during the sale of the slaves, based on the perceived potential labor output of the individual slave: a slave 15–36 years = 1 pieza; a slave 8–15 years or 36–45 years = 2/3 pieza; a slave 4–8 years = 1/2 pieza. Children less than four years old were sold with their mothers. 31. AGI, Indiferente General, 420, L. 8, fol. 177r–178r. 32. In November 1531, 6 marcos, 7 onças, and 1 1/2 oitavas of gold (about 2 kg) were registered at Lisbon House (Casa da Mina) by a slave ship returning from Santo Domingo, where it had disembarked slaves. Camâra Municipal de Lisboa, Livro da receita e despesa do Tesoureiro de 1531, fol. 12. 33. Pierre Chaunu, Conquête et exploitation des nouveaux mondes (Paris, 1969). 34. IAN/TT, CC I, 108, 23. 35. Archivo General de Simancas, Consejo y Juntas de Hacienda, doc. 23. All these voyages are included in TSTD2. 36. David Eltis, “The Volume and Structure of the Transatlantic Slave Trade: A Reassessment,” William and Mary Quarterly 58 (2001): 17–46. 37. Enriqueta Vila Vilar, “The Large-Scale Introduction of Africans into Veracruz and Cartagena,” in Vera Rubin and Arthur Tuden, eds., Comparative Perspectives on Slavery in New World Plantation Societies (New York, 1977), 202–9. 38. Franco, Tráfico de esclavos, 18.

91

92

António de Almeida Mendes

39. There is a five-year overlap between the Fuentes and Vila Vilar series, but given the roughness of the estimate, an attempt to adjust for this seems redundant. 40. Ricardo Escobar Quevedo, “Inquisition et judaïsants en Amérique espagnole (1569–1649): Carthagène des Indes au temps des réseaux” (PhD thesis, Ecole des Hautes Etudes en Sciences Sociales, Paris, 2005), 292–93. 41. In September 1591, for example, a license was issued for seventy-seven slaves in the San Buena Ventura. The vessel was lost on its voyage to the Cape Verde Islands, and the license was reissued for the San Antonio in 1592 (voyageid 29,791). It is not always possible to trace duplications. 42. See David Eltis and David Richardson, “A New Assessment of the Transatlantic Slave Trade,” this volume (chap. 1). 43. AGI, Contratación, Legajo 2878. 44. André Alvares de Almada, Tratado breve dos rios de Guiné do Cabo Verde, ed. Luís Silveira (Lisbon, 1945). Original manuscript published in 1594. 45. Vila Vilar, Hispanoamérica, 195. 46. Calculated from TSTD2. One tonelada (tonelada de frete) was equivalent to forty cubic feet in England in the sixteenth century. 47. Franco, Tráfico de esclavos.

A Reassessment of the Slave Trade to the Spanish Americas

93

Appendix 2.1 Slaves Disembarked in Lisbon from Arguim, 1512–28 Date of Departure (Arguim)

Captain

October 1511 August 1512 April 1513

Antonio Lopes Rodrigo Afonso Rodrigo Afonso

August 1513

Rodrigo Afonso

November 1513

Rodrigo Afonso

March 1514

September 1517 December 1517 January 1518 February 1518 March 1518

Pero Anes de Leiria Pero Anes de Leiria Pero Anes de Leiria Pero Anes de Leiria Pero Anes de Leiria Gonçalo Pirez Pero Anes de Leiria João Rodriguez Pero Ribeiro Pero Ribeiro João Rodriguez Pero Anes de Leiria Pero Ribeiro Alvaro Afonso Pero Anes de Leiria Pero Anes de Leiria Pero Ribeiro Pedro Ribeiro Pedro Ribeiro Alvaro Afonso Pero Ribeiro

April 1518

Alvaro Afonso

July 1514 June 1515 January 1516 February 1516 March 1516 April 1516 May 1516 November 1516 December 1516 January 1517 March 1517 April 1517 June 1517 July 1517 August 1517

Ship

Number of Slaves Departed

Number of Slaves Arrived

Mortality (%)

102 111 161

97 109 156

4.9 1.8 3.1

145

143

1.4

137

135

1.5

Ildefonso São Sebastião Santa Maria de Guadalupe Santa Maria do Rosario Santa Maria do Rosario Santa Maria do Rosario Conceição

23

23

0.0

146

146

0.0

Conceição

139

137

1.4

Conceição Conceição Santa Cruz Conceição

160 210 150 220

157 204 128 167

1.9 2.9 14.7 24.1

Santa Cruz Conceição Santiago Santa Cruz Conceição

128 89 136 55 80

110 88 133 54 80

14.1 1.1 2.2 1.8 0.0

Santiago Santiago Conceição

100 59 87

100 59 86

0.0 0.0 1.2

São Miguel

47

47

0.0

114 110 98 117 150

112 106 95 115 148

1.8 3.6 3.1 1.7 1.3

120

116

3.3

Anunciada Santiago Santiago Conceição Santa Maria da Luz Conceição

(continued)

94

António de Almeida Mendes

Appendix 2.1 (continued) Date of Departure (Arguim) April 1518 May 1518 June 1518

Captain

April 1519 May 1519

Afonso Martins Pedro Ribeiro Pero Anes de Leiria Afonso Martins Pedro Fernandes Pero Ribeiro Alvaro Afonso Afonso Martins Pero Ribeiro Francisco Rodrigues Pedro Fernandes Marcos Rodrigues

May 1519 June 1519

Afonso Martins Luis Soeiro

July 1519

Pero Anes de Leiria Alvaro Afonso André Afonso Alvaro Martins Alvaro Afonso Pedro Fernandes Pedro Ribeiro

July 1518 August 1518 August 1518 November 1518 December 1518 March 1519 April 1519

July 1519 July 1519 October 1519 November 1519 December 1519 January 1520 March 1520 May 1520 August 1520 October 1520 May 1528 August 1528 October 1528 Total

Pedro Fernandes Afonso Martins Pedro Fernandes Afonso Martins Joam Gonçalves Joam Gonçalves Joam Gonçalves

Ship

Number of Slaves Departed

Number of Slaves Arrived

Mortality (%)

Santiago Santo Espírito Conceição

80 150 130

80 148 129

0.0 1.3 0.8

Santiago Santo Espírito Santo Espírito Conceição Santiago Santa Cruz São Jorge

80 102 155 166 120 180 203

80 101 152 134 97 178 197

0.0 1.0 1.9 19.3 19.2 1.1 3.0

Santiago Santa Maria da Guadalupe Santiago Santa Maria da Ajuda Nazaré

162 201

160 192

1.2 4.5

100 170

97 166

3.0 2.4

110

105

4.6

111 110 136 147 115 110

109 106 130 144 112 108

1.8 3.6 4.4 2.0 2.6 1.8

70 120 142 36 151 150 101 6,802

69 120 140 34 145 138 94 6,516

1.4 0.0 1.4 5.6 4.0 8.0 6.9 4.2

Santo Espírito São Miguel Santiago Santo Espírito São Miguel Santa Maria do Rosario São Miguel São Tiago São Miguel São Tiago Anunciada Anunciada São Miguel

Source: IAN/TT (reconstruction from various bundles).

Chapter 3 The Slave Trade to

Pernambuco, 1561–1851 Daniel Barros Domingues da Silva and David Eltis

Historians know less about the slave trade to Pernambuco than about any other major branch of the traffic. A more or less continual slave trade began here very early, probably in 1560, and ended very late, in 1851. During these three centuries, few years passed without several hundred slaves arriving. Thousands of slave voyages were organized in and left from the port, and hundreds of thousands of slaves disembarked there. The hinterland of the port was the bridgehead in the Americas for the sugar complex and remained the world’s leading sugar-producing area for several decades. Two hundred and fifty years later, sugar exports drove the port’s slave trade to the highest level it was ever to attain. In addition to supplying extensive plantations in the hinterland, Pernambuco became an entrepôt for other slave-using areas in Brazil, including the mining areas in the interior and some coastal areas of Amazonia. In the seventeenth century, the very prominence of Pernambuco in the slave Americas triggered a Dutch invasion when the Netherlands was at the peak of its imperial power. In this chapter we attempt to provide some foundations for a more intensive exploration of the Pernambucan slave trade than has 95

96

Daniel Barros Domingues da Silva and David Eltis

hitherto been available in the literature. Our ultimate goal is to pave the way for bringing scholarship on this branch of the slave trade into line with its better-known counterparts in the Americas. We first address the volume of the traffic, before turning to the geography of its organization and the coastal regions of Africa on which it drew. New estimates of slave arrivals in Pernambuco over the approximately three centuries that slaves disembarked at the port have been based partly on records of actual arrivals and departures of vessels and partly on estimates of wellplaced officials. TSTD2 provides the voyage data, and the TSTD2 Web site provides the details of our estimates that have built on these data.1 It contains records of 1,370 slave voyages that either arrived in Pernambuco, left Pernambuco, or were recorded at some point on their out-and-back voyage to Africa as intending to disembark slaves in Pernambuco. Information on these voyages has come from archives located in virtually all parts of the Portuguese Atlantic world. Of the 1,370 records, only 524 contain information on actual arrivals in the port, with the rest specifying only an intention to sell slaves at the port. Comparisons within the larger TSTD2 data set of all voyages recording both an intended arrival and an actual arrival indicate that over 95 percent of these voyages did actually arrive in the port that they had earlier stated as their intended destination. Thus the assumption is made here to treat an intention to sail to Pernambuco as equivalent to an actual arrival at the port. The sources dictate the division of three centuries of slave arrivals in Pernambuco from Africa into eleven separate periods: six when the data are sufficiently complete or when good estimates by contemporaries are available that sustain a hard estimate of the size of the trade; and five bridging periods when the empirical base is much weaker or nonexistent. Into the first of these two categories we can group the years 1620–23, 1630–54, 1720–84, 1801–6, 1811–30, and 1831–51. If we take the beginning of a more or less continual slave trade to Pernambuco as dating from circa 1560, then 1561–1619, 1624–29, 1655–1719, 1785–1800, and 1807–10 are the five periods when shipping data are either incomplete or nonexistent. For these periods we are dependent not on counts of vessels, but rather on broad economic trends, known movements of slaves from Pernambuco to other Brazilian areas, and analogies with other ports where hard data do exist. These tools enable us to construct bridges over the data gaps. It should be noted, however, that except for most of the sixteenth century and the years 1655 to 1697, shipping data do exist for these gap years, and while we may be fairly certain that the data are not complete, shipping data do provide a basis for minimum estimates of slave arrivals. By contrast,

The Slave Trade to Pernambuco

all previous estimates of the volume of the Pernambuco trade have drawn on only economic trends—usually sugar exports or counts of engenhos (sugar mill complexes)—and analogies with other areas. No one has hitherto pulled together the available shipping data for the Pernambuco slave trade except for the Dutch period and a twenty-five-year period prior to formal abolition in 1830, when a couple of series based on single sources of shipping exist. For the latter period, the data presented here are certainly more complete. We focus first on the six periods where estimates are relatively robust, beginning with 1620–23, before turning to the more problematic intervals. The first reliable estimate of the overall scale of slave arrivals derives from a director of the Dutch West India Company, Johannes de Laet, who immediately after the Dutch conquest of Pernambuco (and using documents no longer available) noted: “In order to keep these [sugar] Mills working, a very large number of Blacks are recruited; so that yearly many Ships from Angola and other regions of Africa, sail to and from. One can find in the Registers kept thereof, that from Angola alone in the years 1620, 21, 22, 23, being four years, to the Captaincy of Pernambuco have been disembarked 15,430 Blacks, from whom the King of Spain has received very great utility.”2 The question is, how extensive were arrivals from “other regions of Africa”? There is no easy answer to this from the Portuguese sources for Brazil, but as noted below there is now solid evidence of the African origins of another large branch of the Portuguese transatlantic slave trade at this time—that to Spanish America. The Angolan share of the Portuguese asiento trade in 1620–23 was 83.2 percent.3 If the same ratio held for the traffic to Brazil, then total arrivals in Pernambuco for the four years 1620–23 can be estimated to have been 18,550 (15,430/0.832), or 4,640 per year. The second period for which we have data is the Dutch interlude in Brazil, which began in 1630 and ended with the Dutch withdrawal from the port of Recife in 1654. The records for this period on both sides of the Atlantic are plentiful, and it is highly likely that some documentary trace of every slave ship carrying slaves to New Holland has survived. The Dutch invasion had disastrous consequences for sugar production and the slave trade, both at the beginning of the period when the Dutch were attempting to extend their control from the port of Recife to the whole colony, as well as at the end when the Portuguese gradually expelled the Dutch. Dutch exports of sugar became significant only in 1636, after six years of fighting, and during this period neither the Dutch nor the Portuguese brought in many slaves to New Holland. Indeed, pacification of the new sugar colony intensified Dutch efforts to conquer Portuguese possessions in Africa, because in the aftermath of a destructive war the

97

98

Daniel Barros Domingues da Silva and David Eltis

normalization of sugar production was not possible without ready access to the slave-trading areas of Africa. Sugar exports from New Holland surpassed 1,000 tons a year only for eight years, 1638 to 1645, and again declined precipitously after 1645. Even at their peak, it is unlikely that sugar exports ever matched the levels attained in the first two decades of the seventeenth century. Given Dutch control of the coastline and the ocean, sugar exports from (and, indeed, slave imports into) the part of Pernambuco outside Dutch control could not have been significant in the Dutch period. The slave-trade trends, based on shipping data first analyzed by Van den Boogaart and Emmer and now incorporated into TSTD2, follow the pattern of sugar exports.4 Table 3.1 compares the TSTD2 series with those of earlier scholars. Van den Boogaart and Emmer are more complete than Hermann Wätjen, and the differences between their estimates and TSTD2’s are quite small.5 Ninety percent of the slaves that the Dutch carried to Brazil arrived between 1636 and 1645—the very years when sugar exports approached normalcy. There was only a single shipment in 1630, another possible arrival of a slave vessel in 1633, and, as the Portuguese revolt against Dutch rule gathered strength after 1645, only a few hundred slaves a year from 1646 to 1651, when arrivals ceased altogether. The Dutch obtained a foothold in West Africa only with the taking of Elmina in 1637, and they had access to Angola only from 1641. The period between the Dutch obtaining of access to the main slave markets, especially Angola, and the revival of war and destruction in Pernambuco was extremely brief—four years from 1642 to 1645. During these years the Dutch slave trade averaged 4,000 per year—close to the level estimated for the Portuguese traffic prior to the Dutch invasion. During these years too, the Dutch drew just as heavily on West Central Africa (which accounted for four out of five of all slaves) as had the Portuguese between 1620 and 1623. The great majority of slaves that the Dutch carried from West Africa (as opposed to Angola) traveled before 1642, when Dutch access to West Central Africa became severely limited. Overall, we estimate that 26,687 Africans disembarked in Pernambuco from Dutch vessels between 1630 and 1654, but the annual average of 1,160 was far below that of the preceding Portuguese period. The initial impact of the Dutch was thus to reduce the incidence of Atlantic slavery rather than to increase it. Paradoxically, not until they left Brazil can it be said that the Dutch brought about strong increases in the size of slave trade and the extent of African slavery in the Americas. After the Dutch withdrawal there followed a long period from which both shipping data and contemporary estimates are sparse. From 1720 to 1784 the

The Slave Trade to Pernambuco

Table 3.1 Comparisons of Estimates of Arrivals of Slaves in Pernambuco in the Dutch Period Years

Wätjen

van den Boogaart/ Emmer

Domingues/ Eltis

1630 1633 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1649 1650 1651 1652 Total

— — 1,031 1,580 1,711 1,802 1,188 1,437 2,312 3,948 5,565 2,589 — — — — — — 23,163

280 — 1,046 1,557 1,752 1,796 1,316 1,359 2,378 4,014 5,465 3,773 275 — 490 — 785 — 26,286

734 300 805 1,229 1,508 1,590 1,289 1,606 2,400 5,009 4,932 3,579 570 6 414 448 0 269 26,687

Sources: Hermann Wätjen, O domínio colonial holandês no Brasil: Um capítulo da história colonial do século XVII (São Paulo, 1938), 487; Ernst van den Boogaart and Pieter C. Emmer, “The Dutch Participation in the Atlantic Slave Trade, 1596–1650,” in Henry A. Gemery and Jan S. Hogendorn, eds., The Uncommon Market: Essays in the Economic History of the Atlantic Slave Trade (New York, 1979), 369.

situation improves dramatically, allowing us to designate the third period of the slave trade into Pernambuco for which we have data. For these years the sources on shipping data include incomplete records of ships’ departures from and arrivals into Pernambuco; vessels from Pernambuco entering Luanda and Benguela in Angola as well as Elmina on the Gold Coast (where captains were required by treaty to pay dues to the Dutch authorities); ships leaving and entering São Tomé, where almost all Brazilian vessels trading for slaves on the Mina Coast called; and two separate series of slave-ship departures from Angola, which usually supplied the intended port of arrival in Brazil. While none of these series is complete in itself, the range and density of the sources generate some confidence in their coverage. Other evidence supports this assessment.

99

100

Daniel Barros Domingues da Silva and David Eltis

In 1778 José César de Menezes, the governor of Pernambuco, provided a careful overview of ship and slave arrivals during the preceding thirty-six years. By the end of 1777, the Pernambuco Pombalina Company (Companhia Geral de Pernambuco e Paraíba) had been operating with a monopoly for eighteen years. The governor compared a range of trade data for these years (1760–77) with equivalent data for the eighteen years before the monopoly went into effect (1742–59). Slave arrivals and African coastal origins were among the details he provided in his comparison, and he gave a separate breakdown for the year 1760.6 In short, he presented summaries of vessels and slaves arriving from the Costa da Mina and Angola for 1742–59, 1760, and 1761–77. A direct comparison is possible between the governor’s summary and a voyage-based total for the same years from TSTD2. Table 3.2 presents the comparison. Both in terms of vessels and slaves arriving, Menezes’s summary is remarkably similar to the counts extracted from the new shipping database, with TSTD2 generating a slightly larger estimate. This is to be expected. The data set includes voyages that are known to have set out from Pernambuco or to have arrived on the African coast, but for which no subsequent information exists. Although about 5 percent of all round-trip long-distance voyages failed to reach their intended destinations at this time, one of the conventions of the data set is to assume that such vessels landed slaves in the Americas. Hence, an imputed number of slaves arrived is always allowed for these cases, even though the reason for the lack of information for the later part of the voyage may simply have been that the vessel failed to disembark any slaves. In addition, because a slave voyage often lasted more than a year, small variations in dates are possible between the governor’s data and ours. It is also possible that the data set’s allowance for missing numbers of slaves (that is, those vessels that disembarked an Table 3.2 Slave and Slave-Ship Arrivals in Pernambuco, 1742–77: Summary Report of the Governor of Pernambuco Compared with TSTD2 José César de Menezes Costa da Mina Years 1742–59 1760 1761–77 All years (1742–77)

TSTD2

Angola

Costa da Mina

Angola

Vessels

Slaves

Vessels

Slaves

Vessels

Slaves

Vessels

Slaves

63 1 25 89

16,322 294 7,661 24,277

111 10 69 190

35,593 3,194 27,008 65,795

61 0 25 86

17,984 0 7,601 25,586

116 9 91 216

35,133 2,250 35,154 72,537

The Slave Trade to Pernambuco

unknown number of slaves) may be slightly high. These considerations mean that, for five of the six combinations of regions and year groupings shown in table 3.1 (Costa da Mina, 1742–59, 1760, 1761–77; and Angola, 1742–59, 1760), the two series are similar—indeed, in most of these cases almost identical. For the sixth, Angola between 1761 and 1777, the difference is modest. From 1777 to 1784 the Pombalina Company continued to carry slaves to Pernambuco. (Although the company lost its monopoly in 1780, it continued to trade in competition with other Portuguese slave merchants until 1784.) Surviving shipping records of both company and private traders are relatively diverse and numerous through to 1784, encouraging some confidence in their completeness. Apart from the company archives at the Torre do Tombo, TSTD2 draws on the Angola sources described above; the records that the Dutch kept of Portuguese vessels calling at Elmina to pay their treaty dues; the Junta do Commércio series also at the Torre do Tombo; various AHU series (São Tomé, Pernambuco, Angola); and an incomplete record of arrivals at Pernambuco (Arquivo Público Estadual Jordão Emerenciano [hereafter APEJE], Registro de Passaportes, Passaportes de Embarcações, D.I.-7, 1760/1802) in the Recife public archive, which historians have not previously used. The TSTD2 annual totals are therefore accepted as complete for 1778–84. Total slave arrivals for these seven years are estimated at 15,633. Overall, 176,830 slaves are estimated to have arrived in Pernambuco direct from Africa over the years 1720 to 1784, or just over 2,700 per year. Table 3.3 presents our new annual estimates for 1720–84, broken down by the same broad African regions of embarkation that Governor Menezes used. We now jump to the fourth and fifth periods for which we have data, 1801–6 and 1811–30. The more solid ground from 1801 comprises a series for vessel departures from Pernambuco (ANRJ, códice 157) to supplement the series for arrivals and departures from Angola. TSTD2 supports imputed slave arrivals in Pernambuco of 21,927 in 1801–6, or about 3,700 a year, and these are taken as complete as far as Angola is concerned. There are, however, few records of arrivals from the Costa da Mina, or indeed departures from the Costa da Mina, except for the single year 1803. There seems little doubt that the Costa da Mina was less important to Pernambuco slave traders in the early nineteenth century than it had been half a century earlier, but it is unlikely that Pernambuco received no slaves from West Africa. A sample for 1811–14, a period for which Costa da Mina information does exist, indicates that 16 percent of arrivals were from that region. We have accepted this ratio for 1801–6 and divided the annual series of arrivals from Angola by 0.84 to obtain a total of 26,710

101

102

Daniel Barros Domingues da Silva and David Eltis

Table 3.3 Estimated Slave Arrivals in Pernambuco by Region of Africa, 1720–84 Year 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759

Costa da Mina 2,786 2,287 3,482 1,428 3,319 3,829 3,940 5,492 1,994 3,570 2,960 1,071 1,428 3,035 2,601 3,035 1,428 1,071 322 0 1,071 357 1,126 0 551 744 775 907 1,785 714 986 357 1,205 1,498 1,162 1,439 1,929 1,135 1,253 0

Angola 0 209 0 412 845 633 285 723 1,102 0 0 276 0 0 757 0 714 676 653 986 1,009 941 1,283 1,482 3,088 1,521 2,019 1,705 2,602 1,510 2,616 1,405 1,569 2,429 1,195 1,405 1,830 1,617 3,043 2,819

Total* 3,143 2,496 3,482 1,860 4,893 4,462 4,582 6,262 3,096 3,570 2,960 1,347 1,428 3,035 3,358 3,035 2,142 1,747 975 986 2,080 1,298 2,409 1,482 3,639 2,265 2,794 2,612 4,387 2,224 3,796 1,762 2,774 3,927 2,357 3,068 3,759 2,752 4,295 2,819

The Slave Trade to Pernambuco

Table 3.3 (continued) Year 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 All years (1720–84)

Costa da Mina

Angola

Total*

0 0 0 422 0 284 744 358 1,051 338 942 568 789 523 690 0 310 582 267 286 241 230 200 0 0 76,541

2,250 2,123 2,167 3,070 1,695 2,958 2,327 3,104 2,632 709 1,862 1,119 2,226 1,184 2,315 2,865 1,544 1,254 926 1,488 1,018 2,498 2,377 3,669 2,110 96,843

2,250 2,123 2,167 3,492 1,695 3,242 3,071 3,462 3,683 1,047 2,804 1,687 3,015 1,707 3,005 2,865 1,854 1,836 1,515 1,775 1,259 2,728 2,577 3,669 2,110 175,973

Source: TSTD2. *Totals include some slaves from Upper Guinea but exclude those from unknown regions.

slaves arriving in these six years in Pernambuco. Thus we can say that the slave trade to Pernambuco doubled from a mean of 2,250 a year between 1778 and 1784 to 4,500 a year between 1801 and 1806.7 From 1811 onward, the new series of departures from Pernambuco fits well with the two series of voyages leaving Africa that continued in this period. TSTD2 thus generates an imputed arrival series that we accept as being essentially complete. The departure series continues after 1818 in APEJE, R.P.2.2.4., and from 1820 onward is supplemented by a major additional source that lists arrivals of slaves in Pernambuco (APEJE, 04.06.18). Up until official Brazilian abolition of the slave trade in

103

104

Daniel Barros Domingues da Silva and David Eltis

1830, there are thus independent sources of departures, records of ship movements in Africa, and documentation of arrivals in Pernambuco, which all reinforce each other and provide a complete record of the Pernambuco slave trade totaling 129,711 arrivals, or about 6,500 slaves a year. Table 3.4 shows TSTD2’s annual series for 1801–6 and 1811–30, broken down by two broad regions of African origin. The two further decades of illegal slave arrivals into Pernambuco—the Brazilian government having abolished the slave trade in 1830—constitute the sixth period for which we have reliable data. These are not, however, all from Table 3.4 Estimated Annual Slave Arrivals in Pernambuco by Broad Region of Africa, 1801–6 and 1811–30 Year

Angola

Rest of Africa

Total

1801 1802 1803 1804 1805 1806

4,514 4,591 2,878 3,977 2,883 3,059

860 874 1,185 758 549 353

5,374 5,465 4,063 4,735 3,432 3,412

1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830

2,337 4,735 4,456 4,087 5,253 6,897 6,621 7,497 5,138 9,044 8,590 4,656 4,303 2,480 6,156 5,420 5,537 3,555 5,967 3,172

677 2,733 1,418 4,536 3,208 3,101 653 1,421 160 1,119 930 0 455 353 335 432 444 0 582 1,143

3,014 7,468 5,974 8,623 8,461 9,998 7,274 8,918 5,298 10,163 9,520 4,656 4,758 2,833 6,491 5,852 5,981 3,555 6,549 4,315

Sources: Col. 1, TSTD2; col. 2, see text.

The Slave Trade to Pernambuco

the new database. TSTD2 data imply imputed arrivals of 31,837. Recently, however, Marcus de Carvalho has published a new series that draws not only on the British Parliamentary Papers but also on Portuguese diplomatic documents in the Torre do Tombo (Coleção do Ministério dos Negócios Estrangeiros, Pernambuco) not yet incorporated into TSTD2.8 Table 3.5 compares Carvalho’s series and the series from TSTD2. Carvalho’s post-1830 series totals 45,041 arrivals. However, for the years 1831, 1832, 1840–42, 1845, 1847, and 1851, the TSTD2 series yields estimates in excess of Carvalho’s, and the two series are accordingly spliced together, with the new series pulling together the higher annual totals from both sources. The preferred combined series and the resulting aggregate for 1831–51 of 55,394 are shown in column 3 of table 3.5.

Table 3.5 Estimated Annual Slave Arrivals in Pernambuco by Source, 1831–51 Year 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 All years (1831–51)

TSTD2

Carvalho

Preferred

3,553 1,713 301 421 0 1,219 837 444 5,201 5,683 4,539 2,234 907 130 1,143 829 1,101 0 0 0 1,350 —

2,841 1,400 0 700 1,400 3,500 6,650 5,950 5,250 3,500 2,100 1,750 1,400 2,100 350 700 350 1,750 1,050 2,300 0 —

3,553 1,713 301 700 1,400 3,500 6,650 5,950 5,250 5,683 4,539 2,234 1,400 2,100 1,143 829 1,101 1,750 1,050 2,300 1,350 55,394

Sources: TSTD2 and Marcus J. M. de Carvalho, Liberdade (Recife, 2003), 112, 134.

105

106

Daniel Barros Domingues da Silva and David Eltis

Tables 3.1 through 3.5 have shown the series we have for the six periods with reliable indications of the volume of arrivals into Pernambuco (table 3.4 contains two of those periods). We now turn to the five periods for which the shipping data are inadequate to the task of deriving an annual series. We begin with the earlier periods: before 1620 and 1624–29. The establishment of an African labor force in Pernambuco spanned from 1560 to approximately 1620. As might be expected, data for the earliest phase of the transatlantic slave trade are sparse. The first engenhos to have appeared in the Pernambuco captaincy in the midsixteenth century yield no evidence of African slaves. Nor is there any indication of African involvement in the earliest production of brazilwood and hides there, the export of which predated that of sugar. The Portuguese queen regent issued a royal decree in 1559 allowing the governor of São Tomé to sell up to 120 Kongo slaves to each senhor de engenho, in what was the first official reference to a transatlantic slave traffic, but no record of the first arrivals has survived.9 At this time Brazil was exporting hides, dyewood, and sugar, but there is little doubt that sugar production drove the slave trade, as elsewhere in the Americas. The short-lived French colony in Brazil exported only dyewood in the mid-sixteenth century and apparently brought in no African slaves.10 The first estimates of relative values for produce exports from Pernambuco are not available until the Dutch interlude in the next century, but at that point dyewood amounted to only 5 percent of the value of total sugar and dyewood exports combined.11 The first solid indication of an African presence in the labor force of an engenho comes from 1577, when a prominent mill located in the center of the best sugargrowing region of Pernambuco was stated to have had fifteen African and twenty-five Indian workers.12 We have only a few scattered shipping records for the period before the Dutch arrived—three slave-ship arrivals in the 1570s and two more in the early seventeenth century. Other sources indicate that the Africanization of the engenho workforce in Pernambuco proceeded rapidly after 1560, but two-thirds of that workforce was still Indian in 1580. Africanization was complete by only 1620, according to Stuart Schwartz.13 Shipping data for 1561 to 1619 yield evidence of only 1,018 slaves arriving in Pernambuco. To derive a more realistic estimate, we have assumed that 100 slaves arrived in Pernambuco in 1561 and that thereafter the slave trade grew at a constant rate to 4,640 arrivals in 1620 (this estimate is derived from De Laet). This implies an annual 6.72 percent average rate of growth over fifty-nine years. Under these assumptions, the total arrivals from 1560 to 1619 would have amounted to 67,501. Combining our estimates for 1560–1619 with those for 1620–23 from De Laet and allowing 4,640 per year for the remainder of the

The Slave Trade to Pernambuco

1620s—amounting to 27,840 for the six years—generates a total of 113,901 African slaves arriving in Pernambuco until the Dutch invasion in early 1630. There are two independent comments from well-informed observers for part of this initial era of slave imports from Angola. In 1591 Domingos de Abreu de Brito claimed that he had inspected the revenue book from the Luanda slave factory and stated that no less than 52,053 “pieces” of slaves had departed from Angola between 1575 and 1590 to “this Kingdom [Portugal], parts of Brazil and the Indies of Castile.”14 Angola was easily the primary African source of slaves by the late sixteenth century; as well as Pernambuco, Bahia, and parts of Brazil south of Bahia, the Spanish Americas also obtained the bulk of their slaves from Angola. Our estimate for arrivals in Pernambuco between 1575 and 1590, shown in appendix 3.1, is 6,770. A further 1,500 slaves are estimated to have arrived in the rest of Brazil in these years, and after allowing for mortality between Angola and Brazil, we estimate 11,600 slave departures for Brazil between 1575 and 1590.15 But the Spanish Americas, according to António de Almeida Mendes, were the principal market for slaves from West Central Africa in these years. According to Mendes, about 63,000 Africans were taken for the Spanish Americas between 1575 and 1590, and of these at least 85 percent, or 53,500, came from Angola. Thus recent estimates of departures from Angola, derived from voyage-based data for 1575–90, total 65,000 (11,600 plus 53,500), compared to Brito’s number of 52,053 “pieces.” Because there is still debate about the number of slaves included in a “piece”— supposedly a measurement of prime male equivalency—these two independent assessments reinforce each other. The second independent comment was from a slightly later period. In the years following Brito’s assessment, the slave trade from Angola increased further, especially to the Iberian Americas. Between 1617 and 1619, Father Alonso de Sandoval, for example, was able to observe from Peru that vessels “loaded with more than fifteen thousand blacks each year” left the river Bengo near Luanda, on their way “not only to the Indies, but to Brazil and other places as well.”16 For this period, Mendes estimates 10,300 slaves leaving from all of Africa for the Americas, or about 9,000 from Angola, while the spreadsheets for the first chapter of this volume suggest that the Brazilian slave trade had grown to the point where about the same number were leaving for Brazil. By this time the Bengo River had become the embarkation point for the great majority of Angolan slaves bound for the Americas, but it was not the only departure point. The number of slaves taken onboard elsewhere is unknown, but the voyage-based estimates of Angolan departures of about 18,000 compare

107

108

Daniel Barros Domingues da Silva and David Eltis

well with Sandoval’s estimate from the Bengo River alone. After appropriate adjustments for voyage mortality and for the slave trade to other parts of Brazil and the Spanish Americas, there is indeed some similarity between De Laet’s estimate for Pernambucan arrivals in the early 1620s, Sandoval’s comments for the late 1610s, and the estimates provided here that derive ultimately from shipping movements. The next interval for which good data are lacking is the period after the Dutch left Pernambuco: from 1655 to 1719. The literature contains no hard information whatsoever on the size of the traffic, and TSTD2 thus contains many years without a single voyage either to or from Pernambuco. In total, TSTD2 contains evidence of only 10,860 slaves in the slave trade to Pernambuco over seventy-five years, and informed contemporary estimates are also scarce. In their absence we are dependent on the known patterns of economic growth and particularly of developments in sugar production and the mining economy of Minas Gerais. For Pernambuco, two major long-run factors shaped slave arrivals. The first, dominating the second half of the seventeenth century and most of the eighteenth century, was the slow recovery of the sugar sector to reach the level of its prosperity in the early 1600s. Recovery from the Dutch invasion was hampered by low sugar prices, the devastation of sugar mills and plantations in the years 1645–54, increasing competition from the new and rapidly expanding Caribbean sugar islands, and the insecurity of land tenure in Pernambuco, stemming from ownership claims generated by the war and subsequent peace settlements. A further four-year war between Portugal and the Netherlands disrupted Portuguese shipping until the Treaty of The Hague in 1661, but a second treaty was necessary in 1669 to settle land claims in the former New Holland.17 The significance of this last point is that the claims of the former owners displaced by the Dutch, as well as those of the Dutch themselves, hampered land transfers and hence recovery. Stuart Schwartz, relying on André João Antonil’s assessment, sees little difference between Brazilian capacity and output at the beginning of the seventeenth century and the beginning of the eighteenth century, and the relative position of the Pernambuco industry in Brazil probably declined in this period.18 The second major factor was the discovery of gold in Minas Gerais between 1693 and 1695. All three major Brazilian ports—Pernambuco, Bahia, and Rio de Janeiro—reexported to Minas Gerais slaves that they had received from Africa, with Pernambuco and Bahia making use of the Rio São Francisco as the main link with the interior. Gold sustained the slave trade to Brazil until the mid-eighteenth century and enabled the Portuguese to outbid their English

The Slave Trade to Pernambuco

and French competitors for slaves in West Africa. The strong Brazilian market in these years attracted foreign slavers and slave merchants for the first time since Dutch withdrawal. Both British and French slave merchants attempted, with limited success, to deliver slaves to the major Brazilian ports.19 Most of the greatly increased slave trade after 1693 supplied markets in the interior, not the sugar plantations. About one-quarter of the slaves brought into Pernambuco between 1742 and 1777 were subsequently moved on to Rio de Janeiro or Bahia.20 Others were carried to Minas Gerais via the Rio São Francisco route. In short, net imports of slaves into Pernambuco in the mid-eighteenth century—and probably in any year after the mid-1690s—were well below 2,000 a year and thus far less than they had been at the likely high point of the trade in the early seventeenth century. Using counts of engenhos in the captaincy, Dauril Alden argues that the sugar sector experienced close to zero growth between the early 1720s and the early 1760s. There was some expansion in the later 1760s and 1770s, but the Luso plantation sector is widely recognized to have experienced prolonged economic malaise in the middle decades of the century.21 In the absence of information on vital rates and black population trends we cannot be certain, but it is entirely possible that the number of slaves entering the Pernambuco sugar sector between the discovery of gold in the late seventeenth century and the “agricultural renaissance” of the later eighteenth century was insufficient to maintain the captaincy’s slave population. What do these broad trends suggest about the size of the slave trade to Pernambuco between 1655 and 1719? It is generally recognized that Pernambuco supplied fewer Africans to Minas Gerais than did either Bahia or Rio de Janeiro and that sugar remained Pernambuco’s main source of income from the Atlantic world throughout this period.22 We have reasonably secure estimates of slave arrivals for the early 1620s, through the Dutch period, and again just as the Minas Gerais gold boom was moving to its peak in the 1720s and 1730s.23 Cumulatively, these suggest that slave arrivals after the conversion of the Pernambuco slave labor force from Indians to Africans, and before the disruptions of the Dutch invasion, averaged between 4,500 and 5,000 slaves a year—a figure that the Dutch were never able to match, even between 1642 and 1645, when their access to African slave supply regions was greatest. At the other end of the period, we know that slave arrivals in the 1720s and 1730s averaged about 3,700 per year, well below the pre-Dutch volume. Between 1655 and 1694—before the gold discoveries—the straitened circumstances of the Pernambuco sugar sector suggest that slave arrivals must have stayed below the levels at either end of this attempted bridge over the unknown.

109

110

Daniel Barros Domingues da Silva and David Eltis

One unfocused snapshot is possible from the middle of this period. A document in the AHU, cited by Luiz Felipe de Alencastro, lists annual totals of slave departures from Angola on Portuguese ships between 1666 and 1672.24 The annual average was 10,500. There was no trade to Amazonia at this time, and the Spanish Americas were drawing mainly on the intra-American slave traffic, so all these slaves would have been going to Pernambuco, Rio de Janeiro, or Bahia. Antonil stated that Bahia was the largest Brazilian sugar-producing territory at this time, and as geography had largely insulated the southern sugar-producing areas of São Vicente and Rio de Janeiro from Dutch depredations (except for the brief Dutch occupation of Luanda in the 1640s), then Pernambuco probably lagged behind the other two Brazilian regions in slave imports in these years. If the Pernambuco share of these Angolan departures was 20 percent, then arrivals in Pernambuco (after allowing for shipboard mortality of 10 percent) would have averaged 1,900 slaves a year. Our estimate of annual arrivals in Pernambuco is thus 1,800 between 1655 and 1665, 1,900 between 1666 and 1672, and—allowing for some continued recovery of the Pernambucan sugar sector—2,500 from 1673 to 1695, the beginning of the first Brazilian gold rush.25 Starting in the mid-1690s, the volume of arrivals would certainly have increased yet further. TSTD2 provides evidence of nearly 10,000 slaves coming into Pernambuco between 1696 and 1719, including 3,100 and 1,900 in 1698 and 1699 alone. But thirteen of the years between 1696 and 1719 lack any information whatsoever on slave-ship arrivals. For the other ten years that do contain data, the mean is 1,000 per year, with almost three-quarters arriving from the Costa da Mina. The relatively high TSTD2 figure for 1698 suggests that the initial effect of the gold discoveries was to increase the volume of slave arrivals substantially over the levels we suggest to have been likely between 1655 and 1694. Our estimates for the post-1719 period support our assessment that this surge in slave arrivals did not last. Our previous calculations show about 3,600 slaves arriving in the 1720s, with a decline to half this level in the 1730s. We estimate an average of 5,700 arrivals each year from 1695 to 1710, declining to 4,000 a year in the second decade of the century.26 This yields a rather speculative total for arrivals in Pernambuco between 1655 and 1719 of 212,305 slaves. We now jump to the fourth interval without solidly based information on slave arrivals. From 1785 to 1800 some of the sources for pre-1785 fall away. The company records end, information on Pernambuco arrivals becomes very intermittent, the Dutch were no longer collecting treaty dues from Portuguese slavers trading on the Costa da Mina, and there is no comprehensive

The Slave Trade to Pernambuco

source for departures from Pernambuco. We are left once more with the Angolan records of ship movements, which provide some data on the intended ports of arrival in Brazil, but no sources for vessels moving into and out of Pernambuco itself. A glance at the source variables in TSTD2 for Pernambuco voyages in these years shows how thin the informational foundations are compared with pre-1784, even though the new database has imputed arrivals of nearly 19,000 slaves for 1785–1800—a mean of 1,180 a year. The Brazilian “agricultural renaissance” of the last quarter of the eighteenth century came late to Pernambuco and was centered on cotton rather than sugar. Strong expansion began in the later 1780s, and by the early 1790s output growth matched that in Maranhão, usually seen as the heartland of early Brazilian cotton production. At some point in the 1790s, cotton surpassed sugar as Pernambuco’s most valuable export, and by 1805 Pernambuco had permanently overtaken Maranhão as Brazil’s principal cotton region, though ultimately, like most other cotton-growing places in the world, it could not compete with the U.S. South.27 There is less information on sugar than on cotton at this time, but in 1791 (the only year between 1785 and 1793 for which such data exist) Pernambuco sugar exports ranked well below the mean for the 1770s. As we might expect, given the collapse of Saint-Domingue exports and increasing demand stemming from British industrialization, sugar exports grew strongly after 1791. There are figures for 1801 and 1805 that suggest a doubling over the next ten years. Even when surpassed by cotton, sugar remained easily the second most important crop, and cotton and sugar accounted for 80 percent by value of the captaincy’s exports. All this suggests an expanding demand for labor beginning in the late 1780s, with most of the growth occurring in the 1790s as war, revolution in Saint-Domingue, and accelerating industrialization in Europe took effect. For an estimate of slave arrivals in 1785–1800, we have resorted to fitting a trend line between the average arrivals in 1778–84 and 1801–6. This procedure generates a total for slave arrivals between 1785 and 1800 inclusive of 51,888 slaves—or 176 percent more than the 19,000 slaves suggested by TSTD2 alone. This aggregate estimate for 1785–1800 can then be distributed across individual years, in proportion to such annual shipping data that are available from TSTD2, to derive an annual series for the period. This leaves the years 1807–10 as the final period without good information on arrivals. Like the 1785–1800 period, the major sources in TSTD2 are shipping movements into and out of Angolan ports. A new source for vessel departures from Pernambuco becomes available in 1808 (APEJE, D.I.-9, 1808–18, v.1), but

111

112

Daniel Barros Domingues da Silva and David Eltis

there are obvious gaps in this new series until 1811. TSTD2 includes sufficient shipping data to support an incomplete count of 8,492 arrivals in these four years (1807–10). To compensate for missing data in these years, we have employed once more a trend line from the average for 1801–6 to the average for 1811–16. This yields a total for 1807–10 of 23,328 slaves or 6,000 per year. Table 3.6 presents summary estimates for the eleven separate periods discussed above. Over nearly three centuries a total of 814,617 African slaves are estimated to have disembarked in Pernambuco according to this series, or about 2,800 slaves a year. The annual series behind these summaries is presented in appendix 3.1, alongside the equivalent numbers derived from TSTD2. Over the same period, the appendix shows TSTD2 generating records of 425,075 slaves—about 52.2 of the new estimated series. For some periods, noticeably 1560–1619 and 1655–1696, when almost no voyage-based information is available, the estimates are particularly speculative. For others, such as 1698–1719, the voyage data are weak, and the estimates must not be regarded as robust. Nevertheless, the overall series is anchored on periods when the voyage-based data are fairly solid, including the Dutch era, the second and third quarters of the eighteenth century, and two decades in the nineteenth century. Whatever the shortcomings of the estimates, the reliability of the overall series compares well indeed with those of earlier writers.

Table 3.6 Estimates of Slave Arrivals in Pernambuco for Groupings of Years, 1561–1851 Years 1561–1619 1620–23 1624–29 1630–54 1655–1719 1720–84 1785–1800 1801–6 1807–10 1811–30 1831–51 All years (1561–1851)

Domingues/Eltis

TSTD2

67,501 18,560 27,840 26,687 212,305 176,238 51,888 26,710 23,328 129,703 53,856 814,617

1,018 176 521 26,687 9,675 176,238 18,919 24,517 8,492 129,703 30,838 425,075

Sources: Tables 3.1 to 3.5, text, and TSTD2 for non-Portuguese voyages.

The Slave Trade to Pernambuco

According to the new series, the trade expanded gradually in the late sixteenth century, reached 4,000–5,000 a year by the time the plantation labor force became entirely African in 1620, and then was severely curtailed by the Dutch incursion. Arrivals probably did not again match pre-Dutch levels until the last years of the seventeenth century. The peak level of arrivals in Pernambuco occurred late in the trade’s legal period, in the late 1810s and early 1820s, which coincided with the withdrawal of the British and Americans from the slave-trade business as well as with rising demand for plantation output. Indeed, 28 percent of all slaves ever carried into this part of Brazil arrived during the nineteenth century, a period comprising only one sixth of the time span of one of the longest-lasting branches of the transatlantic slave trade. For Brazil as a whole, the nineteenth-century share of the total era’s trade would be even greater—over 40 percent. Appendix 3.1 also shows that in the seventeenth century there were several years when no slave vessels from Africa arrived—a pattern that also probably existed in the sixteenth century after regular traffic got under way in 1561.28 Because no one has previously attempted an overall assessment of the Pernambucan slave trade, comparisons between the new estimates and those of earlier authors are possible for only certain periods. Comparisons for the Dutch period have already been made in table 3.1. For the eighteenth century, there are estimates from Joseph Miller of a total of 7,473 slaves for 1723–28, 1731, 1734, 1738, and 1740–41, which compare with TSTD2’s sum of 7,636 for the same years.29 However, Miller’s data are for departures from Angola to Pernambuco, whereas the TSTD2 figures are for arrivals in Pernambuco from Angola; after adjusting for voyage mortality, the difference between the two series is over 15 percent more than the nearly identical figures that the unadjusted comparison suggests. Moreover, the transatlantic voyage lasted about a month, and therefore some differences between the estimates arise because the year of departure was not necessarily the same as the year of arrival. TSTD2 is based on multiple sources, and we would argue that our series is to be preferred. For the post-1741 period, table 3.7 compares, first, Miller’s figures for slave departures from Angola to Pernambuco with our own estimates for arrivals in Pernambuco from Angola only (including only those years for which Miller provides data) and, second, António Carreira’s estimate for departures for Pernambuco from all African regions with our own estimate for all arrivals. As with the scattered pre-1742 years, Miller’s estimates are not greatly different from ours, although the difference is larger if we include an allowance for voyage mortality. Once more, however, we

113

114

Daniel Barros Domingues da Silva and David Eltis

Table 3.7 Comparisons of Estimates of the Slave Trade to Pernambuco in the Eighteenth Century, Selected Years All African Regions Years 1742 1744 1747 1748 1749 1758 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 All years

Angola Only

Carreira Departures

Domingues/ Eltis Arrivals

Miller Departures

Domingues/ Eltis Arrivals

— — — — — — 2,298 2,299 4,080 1,849 2,808 3,063 3,630 2,549 1,377 2,034 1,717 623 1,826 1,939 2,521 2,906 1,564 1,268 808 360 1,396 1,862 2,239 0 206 385 47,607

— — — — — — 2,123 2,167 3,491 1,695 3,242 3,071 3,482 3,683 1,047 2,804 1,687 3,015 1,707 3,005 2,865 1,853 1,836 1,515 1,775 1,259 2,728 2,577 3,669 2,110 1,065 1,798 61,269

1,096 1,515 2,188 2,661 1,455 3,235 — 1,666 2,689 1,834 3,217 2,380 2,649 — 758 1,685 1,704 1,580 — 2,082 2,110 — — — — — — — — — — — 36,504

1,283 3,088 1,705 2,602 1,510 3,043 — 2,167 3,070 1,695 2,958 2,327 3,104 — 709 1,862 1,119 2,226 — 2,315 2,865 — — — — — — — — — — — 39,648

Sources: António Carreira, As companhias Pombalinas de navegação, comércio e tráfico de escravos entre a costa africana e o nordeste brasileiro (Bissau, 1969), 261; Joseph C. Miller, “The Numbers, Origins, and Destinations of Slaves in the Eighteenth-Century Angolan Slave Trade,” in Joseph E. Inikori and Stanley L. Engerman, eds., The Atlantic Slave Trade: Effects on Economies, Societies, and Peoples in Africa, the Americas, and Europe (London, 1992), 77–115; appendix 3.1.

The Slave Trade to Pernambuco

would argue that our data are to be preferred because they are multisourced. In contrast to Miller, Carreira provides estimates of departures for Pernambuco from all African regions, based upon his own independently generated voyage data set. Our estimates for 1761–86 are 35 percent greater than his, but much of the difference can be explained by our more complete data for the so-called free traders who moved into the trade when the Pombalina Company lost its monopoly in 1780. Again, we would argue that our series is to be preferred. For the nineteenth century (shown in table 3.8), as with earlier periods, scholars have not provided information for every year. Moreover, the data in the existing literature are only for slaves leaving Angola for Pernambuco, or slaves arriving in Pernambuco from Angola. Unfortunately, some scholars have confused the two or, worse, have assumed that Pernambuco drew all its slaves from West Central Africa. They have also tended to cite one another rather than to investigate the primary sources. This last problem is dealt with in table 3.8 by boldfacing the figures that appear to us to have been the original estimate; thus, if two or more estimates in the same row are identical, the boldfaced figure is the original source. The two most widely cited sources for before 1831 are Goulart and Rebelo. Rebelo’s source, covering only a few years, was for exports to Pernambuco from Angola and came from a contemporary summary.30 But Goulart’s series came from Lopes, and although Goulart described his series as arrivals at Recife, his series, in reality, documents departures from Angola.31 In 1975 Miller accepted Goulart’s series for what Goulart claimed it was, adding an imputed allowance for voyage mortality to derive estimates of departures from Angola. Thus Miller’s 1975 series should be set aside, and Goulart’s series should be read as departures from Angola. Miller’s 1992 series used Goulart’s series correctly and added new material for years that Goulart did not cover. Carvalho also draws on Rebelo and Goulart for different years before 1831, but he confusingly calls all his data “arrivals in Pernambuco,” even though Rebelo, at least, made clear that his figures were departures from Angola. Thus the only valid series in table 3.6 for 1800–31 arrivals at Pernambuco is our own. There are, in fact, no estimates in the existing literature for arrivals of slaves in Pernambuco from Africa as a whole. The situation improves after 1830. While Carvalho’s work looks quite confusing for the legal period, for the illegal slave-trade period he called on new materials for the years 1835–49. These included not only the British Parliamentary Papers but also Portuguese diplomatic documents. As previously noted, we have incorporated his estimates into our own series.32 Overall, from the

115

116

Daniel Barros Domingues da Silva and David Eltis

Table 3.8 Comparisons of Estimates of the Slave Trade from Angola to Pernambuco in the Nineteenth Century, Selected Years Arrivals in Pernambuco from Angola

Departures from Angola to Pernambuco

Years

Goulart

Miller 1975

Carvalho

Domingues/ Eltis

Lopes

Rebelo

Miller 1992

1802 1803 1804 1805 1809 1810 1811 1812 1813 1815 1816 1817 1818 1819 1820 1822 1823 1824 1825 1826 1827 1828 1829 1830

— — 3,325 1,401 2,494 1,254 — 2,489 3,265 3,911 5,499 5,932 7,702 7,802 — — 4,824 2,683 2,384 — — — — —

— — 3,477 1,529 2,804 1,411 — 2,747 3,437 4,237 5,958 6,448 8,606 8,362 — — 5,143 2,880 2,509 — — — — —

— — — — — — — — — 3,911 5,499 5,932 7,702 7,802 7,816 3,203 4,824 2,683 3,620 2,515 2,384 2,400 5,020 1,448

4,591 2,878 4,396 2,883 2,538 924 2,337 4,735 4,456 4,952 6,738 6,621 7,497 5,138 9,044 4,656 4,303 2,487 6,156 5,420 5,537 3,555 5,967 3,172

— — 3,325 1,401 2,494 1,254 — 2,489 3,265 3,911 5,499 5,932 7,702 7,802 — — 4,824 2,683 2,384 — — — — —

— — — — — — — 2,489 — 3,911 — 5,932 — — — 3,203 2,108 1,845 3,620 2,515 — — — —

3,622 4,013 3,325 4,401 2,492 1,254 2,010 2,489 3,265 3,911 5,499 5,932 7,702 6,863 7,816 3,203 — — — — — — — —

Sources: Maurício Goulart, Escravidão africana no Brasil (São Paulo, 1975), 269; Joseph C. Miller, “Legal Portuguese Slaving from Angola: Some Preliminary Indications of Volume and Direction, 1760–1830,” Revue Française d’Histoire d’Outre-Mer 62 (1975): 171–72; Marcus J. M. de Carvalho, Liberdade (Recife, 2003), 112, 134; TSTD2; Edmundo Correia Lopes, A escravatura (subsídios para a sua história) (Lisbon, 1944), 141; Manuel dos Anjos da Silva Rebelo, Relações entre Angola e Brasil, 1808–1830 (Lisbon, 1970), table 2; and Joseph C. Miller, “The Numbers, Origins, and Destinations of Slaves in the Eighteenth-Century Angolan Slave Trade,” in Joseph E. Inikori and Stanley L. Engerman, eds., The Atlantic Slave Trade: Effects on Economies, Societies, and Peoples in Africa, the Americas, and Europe (London, 1992), 77–115. Note: Where two entries are of equal value in the same row, the boldfaced entry is the original source for the estimate.

The Slave Trade to Pernambuco

end of the Dutch period to 1830, the existing literature contains no counterpart to the series presented here. Our overall estimate of just over 800,000 African arrivals leads to the question of where the traffic was based. On average, each slave vessel discharged just under 300 slaves in Pernambuco.33 A division of this mean into aggregate arrivals of 814,617 suggests that 2,725 transatlantic voyages entered Recife over three centuries. In which ports were these voyages organized? The answer to this question appears straightforward. TSTD2 provides the ports of departure for 997 voyages associated in some way with Recife—a sample size of just below two-fifths. Eighty-two percent of this sample left from Recife itself, with another 10 percent from other ports in Brazil. Europe accounted for the remainder, including seventeen vessels or nearly 2 percent from the Netherlands during the Dutch interlude, and a few more from other northern European ports (the latter being mainly vessels driven off course that entered Pernambuco in distress on their way to the Caribbean). Lisbon, the capital of the Portuguese Empire, sent out just sixty-two slave voyages to Pernambuco, or 6 percent of the sample. There may well be some downward bias in the Lisbon figure, because the sixteenth and seventeenth centuries are poorly represented in TSTD2; it would be reasonable to expect a larger role for the financial and trading center of the Luso world in this early period, when port and financial facilities in the Americas must have been primitive compared to Europe’s. Even Cádiz may have played a role during the period of the union of the Portuguese and Spanish Crowns. In this context, it should be noted that only 6 percent of the aforementioned large sample of ports of departure fall within the period before 1700, compared with 94 percent after. Yet Pernambuco still represents nearly half of this small pre-1700 sample (because of the Dutch invasion, the Netherlands accounts for most of the rest). If we count only the sixteenth century, there is almost no evidence regarding ports of departure (only two voyages, both originating in Spain), but it is not unreasonable to assume a significant involvement for Lisbon in this early period. Overall, however, such an involvement would not alter the assessment that the slave trade to Recife was overwhelmingly based in Recife. The African end of the Pernambuco slave trade has received rather more attention than the American terminus. This situation stems from the fact that most attempts to construct estimates of the arrivals of Africans in Recife have started with estimates of departures from Angola to Pernambuco. Nevertheless, much new information on African embarkation points is now available,

117

118

Daniel Barros Domingues da Silva and David Eltis

summarized in table 3.9. For the period before the Dutch arrived in 1630, as noted, little information has survived even from Angola. Fortunately a proxy exists in the form of a large sample of data for the early period of another major branch of the Portuguese traffic—that to the Spanish Americas after 1580. For thirty years historians have drawn heavily on Frederick Bowser’s work on the ethnicities of slaves in sixteenth-century Peru as an indicator of the African origins of the early transatlantic slave trade.34 Over half the slaves in Bowser’s sample left from the Upper Guinea coast. By contrast, TSTD2’s sample of the African coastal origins of 92,354 slaves purchased by the Portuguese for the Spanish Americas before 1630 provides quite a different distribution and casts doubt on the representativeness of the Peru sample. TSTD2 shows only 12 percent of Africans leaving from the Upper Guinea coast, with 83 percent from Angola (or 87 percent if we add São Tomé in the Bight of Biafra, almost all of whose slaves originated in Angola at this time). Other parts of Africa, including West Africa south of Upper Guinea, contributed only trivial quantities of slaves to the Americas before 1630. Prior to 1600, in fact, there were two widely separated systems of generating slaves in the Atlantic world. The first and the most northerly used the Cape Verde Islands as an entrepôt, though drawing slaves in small numbers from a wide range of the Upper Guinea coast; it supplied the Iberian Peninsula as well the Spanish Americas. The second centered on Angola and used São Tomé, a plantation colony in its own right at this time, as a minor entrepôt. Almost all the slaves collected by these two systems were sent to the Americas. In the course of the sixteenth century, the Angolan system overtook its northern counterpart. Parts of the thousands miles of coast between these two systems gradually entered the slave trade after 1600. Pernambuco entered the transatlantic slave trade when the transition from Upper Guinea to Angola was well advanced, and there is no evidence of any slaves from Cape Verde or, later, Cach˜eu and the southern rivers of Senegambia entering Pernambuco. Table 3.9 displays the voyage-based evidence that supports this new interpretation. During the Dutch period, quite a different pattern to that described above emerged. Dutch involvement in the slave trade before 1630 had been confined mainly to capturing Portuguese slave vessels on their way to the Americas. When the Dutch invaded Pernambuco, they had no established slave-supply structure on the African coast. They captured Elmina, more important for gold than for slaves, in 1637 but were not able to draw heavily on Angola until they took Luanda in 1641. The Portuguese were able to pressure the court of Sonho to halt slave sales to the Dutch in Mpinda—then the main

The Slave Trade to Pernambuco

Table 3.9 The African Coastal Origins of Slaves Disembarked in Pernambuco, 1636–51, 1720–85, and 1801–51 Years

Upper Guinea

1636–40 1641–45 1646–51 All years %

0 0 0 0 0

Gold Coast/ Bight of Benin 1,739 1,621 291 3,651 17.4

Bight of Biafra

West Central Africa

Southeast Africa

Dutch Period 1,279 1,975 290 3,544 16.9

1,470 11,275 1,037 13,782 65.7

0 0 0 0 0

1720–25 1726–30 1731–35 1736–40 1741–45 1746–50 1751–55 1756–60 1761–65 1766–70 1771–75 1776–80 1781–85 All years %

919 0 0 0 0 357 0 0 0 0 0 324 0 1,600 0.9

Eighteenth Century 16,774 357 17,600 357 10,456 0 3,535 0 2,421 0 5,168 0 5,661 265 4,316 0 705 0 3,433 0 2,571 0 1,685 0 456 0 74,781 979 42.8 0.6

2,100 2,100 1,032 4,037 8,315 10,451 8,004 11,462 11,919 10,634 9,708 6,235 11,172 97,169 55.7

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

1801–5 1806–10 1811–15 1816–20 1821–25 1826–30 1831–35 1836–40 1841–45 1846–51 All years %

0 353 411 217 653 0 0 0 0 109 1,743 0.9

Nineteenth Century 0 1,525 887 2,482 948 7,501 160 730 335 156 1,078 97 0 188 135 0 0 91 500 0 4,043 12,770 2.1 6.8

18,844 7,179 20,868 35,198 26,192 23,651 5,032 10,962 7,325 2064 157,315 83.3

300 412 3,595 5,348 629 804 0 1,180 754 0 13,022 6.9

Source: TSTD2.

119

120

Daniel Barros Domingues da Silva and David Eltis

slave embarkation point in the Kingdom of Kongo.35 As already noted, the Dutch hold on Angola was never secure and was subject to Portuguese interference in the interior, particularly on slave supplies from Matamba. The Dutch were expelled from Luanda in 1648. The distribution of the coastal origins of slaves in the Dutch period shown in table 3.9 was thus much more heavily oriented to West Africa than was the preceding Portuguese traffic. Upper Guinea disappeared as a source of slaves, the Bights of Benin and Biafra supplied most slaves, and Angola was important only for the few years at the beginning of the Dutch occupation of Angola. After the Dutch withdrawal, it is likely that the traditional dependence of Pernambuco on West Central African slaves reestablished itself. There is, however, no hard information until the end of the seventeenth century, when, as already noted, shipping records point to the arrival of 4,700 slaves between 1697 and 1699, almost all from Angola. Thereafter a major if temporary shift in the African sources of Pernambuco slaves seems to have taken place. Portuguese traders based in Bahia and Pernambuco were able to use increasing quantities of illegally exported Brazilian gold to squeeze out English slave traders from the Costa da Mina slave markets.36 The gold greatly reinforced the advantage that Pernambucan and Bahian traders had already garnered from using tobacco rolls as trade goods in the Bight of Benin beginning in the 1670s.37 As table 3.9 shows, hard voyage-based evidence for this shift appears only after 1720, but the shift was already complete at that point. Between 1720 and 1735, seven out of every eight slaves arriving in Pernambuco from Africa originated on the Costa da Mina, mainly the Bight of Benin. Starting in the mid-1730s, as gold production leveled off and began to decline, the old pattern of slave supply reappeared. Between 1736 and 1740 a small majority of arrivals came from Angola, and after 1740 this majority increased steadily until, by the 1770s, three-quarters of Pernambuco’s slaves came from Angola. It is thus likely that the first third of the eighteenth century was the one period when the Costa da Mina dominated the supply of slaves to Pernambuco. For the nineteenth century, table 3.9 shows a further decline in the Costa da Mina share, even as the volume of the slave trade expanded. Southeast Africa emerged as a more important source of slaves, but not one that ever came close to replacing West Central Africa. After 1807 new West Central African sources emerged as Ambriz, Cabinda, Molembo, and Loango gradually supplanted Luanda and Benguela. As a consequence, West Central Africa increased its share of slaves entering Recife to close to 90 percent, a ratio probably not seen since the early seventeenth century. This ratio is probably the

The Slave Trade to Pernambuco

product of some upward bias, however; beginning in 1810, a series of AngloPortuguese anti-slave-trade conventions gradually restricted Portuguese slavetrading activity. The earliest treaties outlawed slave trading north of the equator, which excluded virtually the whole of West Africa. Slave trading on the Costa da Mina continued but is more difficult to detect, since vessels from both Pernambuco and Bahia made official declarations that they had obtained slaves at points near the mouth of the Congo River when they had actually obtained some of their slave cargos in West Africa. TSTD2 contains twentythree voyages from Pernambuco captured by the British navy after 1809 that either traded or intended to trade in West Africa rather than Angola. Many of these carried official papers that specified Angola as the point of trade. Except for the rising presence of Southeast African slaves—probably at the expense of Africans from the Costa da Mina—it is likely that the true ratio of slaves carried into Pernambuco in the nineteenth century was not much different from what it had been in the 1770s. A second striking pattern that emerged in the nineteenth century, at least in West Africa, was the apparent inability of Pernambuco slave traders to maintain a presence at the major Costa da Mina ports of Ouidah, Porto Novo, and Onim (Lagos). Of the thirty-nine Pernambuco slave voyages that we can link with the Costa da Mina after 1810, only six traded or intended to obtain slaves from these three major embarkation points. Instead, Pernambucan slave traders moved into the major Bight of Biafra ports after the British pulled out. Old Calabar appears to have become the major West African embarkation point for Pernambuco arrivals. Recife slave traders were much more likely to draw on Benin, Formosa, Gabon, and Cape Lopez than on Ouidah and Onim. This contrasts sharply with the West African origins of Bahia traders at this time, who drew heavily on these major ports. Havana slave traders, too, had much greater access to Lagos, Ouidah, and Porto Novo than their Pernambuco counterparts. In summary, except for a few decades in the first half of the eighteenth century, Pernambuco’s connection with West Central Africa, though less than most historians have assumed, was nevertheless only slightly less strong than that of southeastern Brazil. Indeed, Pernambuco was likely second in importance only to Rio de Janeiro as the destination for slaves leaving all African points of embarkation south of Cape Lopez. As with Rio de Janeiro, no doubt locally produced cachaça was a major product exchanged for these slaves. While several decades lack data, it is probable that it was only in the Dutch period and then again in the first third of the eighteenth century—when gold

121

122

Daniel Barros Domingues da Silva and David Eltis

and tobacco were important trade goods—that these patterns were temporarily broken. Preliminary breakdowns suggest that, after 1800, southeast Africa was a much less important source for Pernambuco than for Bahia and Rio de Janeiro. In the nineteenth century, and possibly earlier, the designation “Costa da Mina” referred to a much wider range of ports for Recife slave traders than for their Bahian counterparts. Overall, how significant are these new findings that TSTD2 has made possible? If Recife sent out about 2,000 or more slaving voyages during the slavetrade era, then the port would rank as the fifth or sixth largest organizational center in the transatlantic slave-trading world. It would rank behind Rio de Janeiro, Liverpool, Bahia, and London. Admittedly, these 2,000 voyages were spread over a much longer period than those of the two English ports. Nevertheless, Recife’s trade was about as big as that of Bristol, the third-largest English slave-trading port. It was larger than that of any slave-trading organizational center on the continent of Europe. Its trade was one-sixth larger than that of Nantes, the most prominent French slaving port. Recife generated more slaving voyages than Bordeaux, La Rochelle, or indeed all the rest of the French ports put together. Its traffic was also greater than all the voyages that originated on the North American mainland. It is therefore odd that tiny centers such as Lancaster in England (122 voyages) and Saint-Malo in France (217 voyages) have received more scholarly attention than the slave trade of this important port. It is probable that more has been written on the single voyage of the Diligent from Vannes in France or the passage of a single slave through Pernambuco than on the 2,000 or so slave voyages that left Recife.38 The investors and coordinators of these voyages lived in Pernambuco, much of the merchandise exchanged for slaves was produced there (rolls of tobacco, cachaça), the vessels themselves were fitted out in Recife, and it is hard to see the profits accumulating anywhere else. NOTES

1. All the data and the assessments based on them used in this essay may be found on the estimates page of the TSTD2 Web site at http://www.slavevoyages.org/tast/assessment/ estimates.faces. 2. Johannes de Laet, Historie ofte Iaerlijck verhael van de verrichtingen der Geoctroyeerde West-Indische Compagnie in derthien boecken. Tweede deel: Boek IV–VII (1627–1630), ed. S. P. L’Honoré Naber (’s-Gravenhage, 1932; first published 1644), 139. Similar numbers and wording appeared in Casper van Baerle, Nederlandsch Brazilië onder het bewind van Johan Maurits, grave van Nassau, 1637–1644 . . . trans. and ed. S. P. L’Honoré Naber

The Slave Trade to Pernambuco

(’s-Gravenhage, 1923; first published 1647), 50–51. We thank Ernst van den Boogaart and Rik van Welie for their help in locating and translating these references. It should be noted that Daniel P. Mannix and Malcolm Cowley appear to be drawing on these sources in their widely read Black Cargoes: A History of the Atlantic Slave Trade (London, 1962), when they erroneously refer to the Dutch having carried thousands of slave to Brazil before their 1630 conquest of the colony (p. 62). They apparently transpose a Dutch source into a Dutch slave trade. 3. Calculated from TSTD2. See also note 34 below. 4. Ernst van den Boogaart and Pieter C. Emmer, “The Dutch Participation in the Atlantic Slave Trade, 1596–1650,” in Henry A. Gemery and Jan S. Hogendorn, eds., The Uncommon Market: Essays in the Economic History of the Atlantic Slave Trade (New York, 1979), 353–76. Our count of ships and our allowance for the number of slaves carried on vessels that lacked head-count data vary slightly from Van den Boogaart and Emmer’s careful assessment of the size of the Dutch slave trade at this time, although it is based on essentially the same material. See Jelmer Vos, David Eltis, and David Richardson, “The Dutch in the Atlantic World: New Perspectives from the Slave Trade with Particular Reference to the African Origins of the Traffic,” this volume (chap. 8). We thank Ernst van den Boogaart for making available his data set of Dutch voyages. 5. These small differences arise as a function of minor variations on dates and different assumptions about the outcome of a few voyages for which information exists only for the African end of the business. 6. AHU, Pernambuco, caixa 130, doc. 9823, José César de Menezes, July 3, 1778, enc. 2, “Manuel Lopes de Santiago Corrêa, proprietário da abertura dos despachos e descargas na Alfândega do Recife de Pernambuco.” 7. Joseph C. Miller, “Legal Portuguese Slaving from Angola: Some Preliminary Indications of Volume and Direction, 1760–1830,” Revue Française d’Histoire d’Outre-Mer 62 (1975): 171. Miller estimates departures of 2,500 a year from Luanda to Pernambuco in the years 1804 and 1805. 8. Marcus J. M. de Carvalho, Liberdade: Rotinas e rupturas do escravismo no Recife, 1820–1850 (Recife, 2003), 112, 134. 9. Mauricio Goulart, Escravidão africana no Brasil: Das origins à extinção do tráfico (São Paulo, 1975); Gerald Cardoso, Negro Slavery in the Sugar Plantations of Veracruz and Pernambuco, 1550–1680 (Washington, D.C., 1983). 10. Elizabeth Donnan, Documents Illustrative of the History of the Slave Trade to America, 4 vols. (Washington, D.C., 1931–35), 1:13–14. Clarence J. Munford, in The Black Ordeal of Slavery and Slave Trading in the French West Indies, 3 vols. (Lewiston, N.Y., 1991), 1:141 n. 48, cites Donnan and describes the French as carrying slaves to Brazil, but we have been able to find no reference to such activity in this source. 11. Calculated from Henk den Heijer, “The Dutch West India Company, 1621–1791,” in Johannes Postma and Victor Enthoven, eds., Riches from Atlantic Commerce: Dutch Transatlantic Trade and Shipping, 1585–1817 (Leiden, 2003), 88–89. 12. This was the Engenho São Pantaleão do Monteiro. See Francisco Augusto Pereira da Costa, Arredores do Recife (Recife, 2001), 126.

123

124

Daniel Barros Domingues da Silva and David Eltis

13. Stuart B. Schwartz, “A Commonwealth within Itself: The Early Brazilian Sugar Industry, 1550–1670,” in Stuart B. Schwartz, ed., Tropical Babylons: Sugar and the Making of the Atlantic Worlds, 1450–1680 (Chapel Hill, 2004), 156–200. 14. Domingos de Abreu de Brito, Um inquérito à vida administrativa e econômica de Angola e do Brasil em fins do século XVI (Coimbra, 1931; first published Lisbon, 1591), 30. 15. Slave arrivals in the rest of Brazil and the mortality calculation to convert these arrivals into departures from Africa are shown in the worksheet of the TSTD2 Web site. 16. Alonso de Sandoval, De instauranda aethiopum salute (Bogotá, 1956; first published Seville, 1627), 88. However, Sandoval must have written his work in Lima between 1617 and 1619; see Angel Valtierra, S.J., “El Padre Alonso de Sandoval, S.J.” in ibid., xxi. 17. Evaldo Cabral de Mello, O negócio do Brasil: Portugal, os Países Baixos e o Nordeste, 1641–1669, 3rd ed. (Rio de Janeiro, 2003). 18. Schwartz, “Commonwealth within Itself,” 163–64. 19. In TSTD2, see the Eagle, voyageid 21,388 in 1697, to Pernambuco; the Windsor, voyageid 25,658 in 1714, and the Broughton, voyageid 23,006 in 1718, both sailing into Bahia; and several unnamed English vessels (voyageids 23,002, 23,003, 23,004, and 23,006) into both ports in 1714, 1723, and 1738. In addition, Humphry Morice, a major London trader, sent at least one batch of slaves on a Portuguese vessel, NS do Rosário S José e S Domingos, in 1722, and the Royal African Company put several other groups on other vessels; see Pierre Verger, Trade Relations between the Bight of Benin and Bahia from the 17th to 19th Century (Ibadan, 1976), 33–34. The French vessels Reine, voyageid 30,038 in 1714; the Comte de Toulouse, voyageid 32,868 in 1724; the Saint-Jean-Baptiste, voyageid 33,124 in 1721; and Notre-Dame de l’Epine, voyageid 33,518 in 1703, carried slaves into Bahia and Rio de Janeiro. 20. AHU, Pernambuco, caixa 130, doc. 9823, José César de Menezes, enc. 3, “Vitoriano Gomes Maciel Silva, escrivão da Receita e Despeza da Tesouraria Geral, Deputado da Junta da Administração e Arrecadação da Real Fazenda de Pernambuco.” In addition, TSTD2 contains records of thirteen vessels that set sail from Pernambuco but sold slaves or intended to sell slaves in Bahia and Rio de Janeiro. These have not been included in the present analysis. 21. Dauril Alden, “Late Colonial Brazil, 1750–1808,” in Leslie Bethell, ed., The Cambridge History of Latin America, 4 vols. (Cambridge, 1984–1986), 2:627–31. 22. Schwartz, “Commonwealth within Itself,” 163–64; Alden, “Late Colonial Brazil,” 628. 23. Virgilio Nóia Pinto, O ouro brasileiro e o comércio anglo-português, 2nd ed. (São Paulo, 1979), 114. 24. See Luiz Felipe de Alencastro, O trato dos viventes: Formação do Brasil no Atlântico Sul, séculos XVI e XVII (São Paulo, 2000), 378. We have not been able to locate the document, which is cited as AHU, Angola, caixa 10/64, documento estudado por A.L.A. Ferronha, “Angola,” vol. 1, 119–20. 25. See the worksheet for Pernambuco at the TSTD2 Web site for incorporation of the Alencastro source, broken down by our assessment of their intended destinations in the Americas. 26. A pattern of rapid rise in slave imports from the mid-1690s and gradual decline after 1710 is not totally consistent with the pattern of gold production in Brazil. Initial

The Slave Trade to Pernambuco

expansion in the mid-1690s was rather slow, and the most rapid output growth came between the 1720s and the 1730s. Peak output for Minas Gerais occurred in the late 1730s, and for Brazil as a whole (after the fuller development of the Goiás and Mato Grosso discoveries), peak output came in the early 1750s. See Nóia Pinto, O ouro brasileiro, 114; A. J. R. Russell-Wood, “The Gold Cycle, c. 1690–1750,” in Bethell, ed., Cambridge History of Latin America, 2:593–95. There is, of course, abundant evidence of slaves pouring into the gold-mining areas throughout this period, but Pernambuco clearly lost its role as a major conduit for such slaves early in the eighteenth century. 27. Alden, “Late Colonial Brazil,” 635–39. By 1801 cotton accounted for two-thirds of the value of Pernambuco’s exports (AHU, Pernambuco, caixa 232, doc. 15638). 28. The figures for 1560–1619 are estimates based on a trend line from 1560 to 1620 and are not intended to reflect actual arrivals in any given year. 29. Miller, “Legal Portuguese Slaving from Angola,” 171–72, and idem, “The Numbers, Origins, and Destinations of Slaves in the Eighteenth-Century Angolan Slave Trade,” in Joseph E. Inikori and Stanley L. Engerman, eds., The Atlantic Slave Trade: Effects on Economies, Societies, and Peoples in Africa, the Americas, and Europe (London, 1992), 77–115. 30. AHU, Angola, conta corrente do balanço de todas as receitas e despesas da Alfândega de Angola, caixas 61–62, 64–67, and AHU, Angola, balance geral da receita e despesa da Tesouraria da J. A. F. Do Reino de Angola, maço 16. 31. Edmundo Correia Lopes, A escravatura (subsídios para a sua história) (Lisbon, 1944), 139–41. Lopes does not cite a source, but his context suggests it was the same as that Rebelo used: Manuel dos Anjos da Silva Rebelo, Relações entre Angola e Brasil (1808–1830) (Lisbon, 1970), table 1 (after p. 81). But Goulart also used Afonso de E. Taunay’s História do café no Brasil (Rio de Janeiro, 1939), 4:228. Taunay cites a “Resumo da Importação da Província de Pernambuco” (Summary of the Imports of the Province of Pernambuco), dated 1823. According to this document, shown to him by the historian Alcides Bezerra, “130,418 slaves arrived there in twenty-three years, from 1801 through 1823, or an annual rate of 5,670” (p. 269). 32. IAN/TT, Coleção do Ministério dos Negócios Estrangeiros, Pernambuco, caixas 1–3. For the year 1851, Carvalho used Leslie Bethell, The Abolition of the Brazilian Slave Trade (London, 1970), 343. 33. Years 1560–1600 1601–1700 1701–1800 1801–1851

Mean Number of Slaves

Number of Observations

Standard Deviation

138.33 259.36 316.18 299.92

3 22 50 251

127.359 241.291 122.097 159.987

34. See António de Almeida Mendes, “The Foundations of the System: A Reassessment of the Slave Trade to the Spanish Americas in the Sixteenth and Seventeenth Centuries,” this volume (chap. 2), for an explanation of procedures. These new figures for the Portuguese asiento are radically different from the currently accepted standard description

125

126

Daniel Barros Domingues da Silva and David Eltis

of African origins, which is based on Frederick Bowser, The African Slave in Colonial Peru, 1524–1650 (Stanford, 1974), 39–41, and Jean-Pierre Tardieu, “Origins of the Slaves in the Lima Region in Peru (Sixteenth and Seventeenth Centuries),” in Doudou Diène, ed., From Chains to Bonds: The Slave Trade Revisited (Paris, 2001), 43–54. These scholars, whose analyses are based on a small number of inventories in Peru, estimate that the majority of slaves carried to Peru at this time were from Upper Guinea. For conclusions based on this work, see Walter Hawthorne, Planting Rice and Harvesting Slaves: Transformations along the Guinea-Bissau Coast, 1400–1900 (Portsmouth, N.H., 2003), 69. 35. Van den Boogaart and Emmer, “Dutch Participation,” 361–64. 36. An observer based in Barbados commented in 1714 that “the Portuguese have very much of late enhanced the Price of Slaves on this Coast by bringing Gold from Brazil and giving from four even to seven ounces of Dust for a slave.” In “Observations on the Trade of Africa and Angola by the Reverend Mr. Gordon” (1714), Huntington Library, San Marino, Calif., Stowe ms, ST 9, pp. 48–49. See also the Papers of Humphrey Morice, microfilm edition, Reel 2, Morice to Lawrence Prince, July 20, 1719, for one of several letters at this time directing English slave captains on the Slave Coast to sell their slaves to Portuguese vessels and thus to realize “a much better price to sell at upon the Coast, than to carry them to Jamaica.” 37. See Verger, Trade Relations, 33–35. The Royal African Company instructed its agents on the coast to treat “Portugueze [sic] vessels from Brazil . . . civilly and try to encourage the gold trade with them for the accounts of the company” (BNA, T70/53, pp. 34–35). The Slave Coast’s share of slaves carried away from Africa on British ships declined from 35 percent in 1676–1700 to 24 percent in 1701–25 and to 7 percent in 1726–50 (calculated from TSTD2). 38. See Alain Roman, Saint-Malo au temps des négriers (Paris, 2000); Melinda Elder, The Slave Trade and the Economic Development of Eighteenth-Century Lancaster (Halifax, 1992); Robin Law and Paul Lovejoy, eds., The Biography of Mahommah Gardo Baquaqua: His Passage from Slavery to Freedom in Africa and America (Princeton, 2001); and Robert W. Harms, The Diligent: A Voyage through the Worlds of the Slave Trade (New York, 2002).

The Slave Trade to Pernambuco

Appendix 3.1 Slaves Landed in Pernambuco by Portuguese and Brazilian Vessels, 1561–1851: Imputed from TSTD2 and Estimated Years

TSTD2

1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599

0 0 0 0 0 0 0 0 0 0 0 0 0 48 284 83 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Estimated Years 100 107 114 122 130 138 148 158 168 180 192 204 218 233 249 265 283 302 322 344 367 392 418 446 476 508 542 579 618 659 703 751 801 855 912 973 1,039 1,109 1,183

1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638

TSTD2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 521 0 0 0 0 0 0 0 0 0 0 0

Estimated Years 1,263 1,347 1,438 1,535 1,638 1,748 1,865 1,990 2,124 2,267 2,419 2,582 2,755 2,940 3,138 3,348 3,573 3,813 4,070 4,343 4,640 4,640 4,640 4,640 4,640 4,640 4,640 4,640 4,640 4,640 0 0 0 0 0 0 0 0 0

1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677

TSTD2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Estimated 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1,800 1,800 1,800 1,800 1,800 1,800 1,800 1,800 1,800 1,800 1,800 2,295 2,358 2,317 1,833 2,189 1,176 1,203 2,500 2,500 2,500 2,500 2,500 (continued)

127

128

Daniel Barros Domingues da Silva and David Eltis

Appendix 3.1 (continued) Years

TSTD2

Estimated Years

1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717

0 0 0 105 0 0 0 0 0 0 0 0 0 0 0 0 515 0 0 335 3,311 1,683 523 211 0 0 0 0 0 0 0 357 0 0 0 0 0 714 357 0

2,500 2,500 2,500 2,500 2,500 2,500 2,500 2,500 2,500 2,500 2,500 2,500 2,500 2,500 2,500 2,500 2,500 2,500 5,234 7,201 8,235 5,350 6,755 5,485 5,465 5,464 4,976 4,149 4,641 7,583 4,849 5,100 5,149 4,000 4,000 4,000 4,000 4,000 4,000 4,000

1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757

TSTD2 701 1,355 3,143 2,496 3,482 1,840 4,893 4,462 4,582 6,262 3,096 3,570 2,960 1,347 1,428 3,035 3,358 3,035 2,142 1,747 975 986 2,080 1,298 2,409 1,482 3,639 2,265 2,793 2,613 4,387 2,224 3,796 1,762 2,774 3,927 2,357 3,068 3,759 2,752

Estimated Years 4,000 4,000 3,143 2,496 3,482 1,840 4,893 4,462 4,582 6,262 3,096 3,570 2,960 1,347 1,428 3,035 3,358 3,035 2,142 1,747 975 986 2,080 1,298 2,409 1,482 3,639 2,265 2,794 2,612 4,387 2,224 3,796 1,762 2,774 3,927 2,357 3,068 3,759 2,752

1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797

TSTD2

Estimated

4,295 2,819 2,250 2,123 2,167 3,492 1,695 3,242 3,071 3,462 3,683 1,047 2,804 1,687 3,015 1,707 3,005 2,865 1,854 1,836 1,516 1,774 1,259 2,728 2,577 3,669 2,110 383 644 2,204 1,106 649 971 1,294 1,452 1,851 324 1,623 607 1,558

4,295 2,819 2,250 2,123 2,167 3,492 1,695 3,242 3,071 3,462 3,683 1,047 2,804 1,687 3,015 1,707 3,005 2,865 1,854 1,836 1,516 1,774 1,259 2,728 2,577 3,669 2,110 1,069 1,798 6,154 3,088 1,812 2,711 3,613 5,054 5,168 905 4,532 1,695 4,350

(continued)

The Slave Trade to Pernambuco

Appendix 3.1 (continued) Years

TSTD2

Estimated Years

TSTD2

1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815

2,296 636 985 4,815 4,591 4,064 4,573 3,063 3,412 1,673 1,577 1,362 4,234 3,014 7,469 5,974 8,624 8,460

6,411 1,776 2,750 5,374 5,465 4,063 4,735 3,432 3,642 4,596 4,329 3,742 10,661 3,014 7,469 5,974 8,624 8,460

9,999 7,274 8,918 5,298 10,164 9,520 4,656 4,332 2,833 6,156 5,516 5,981 3,555 6,549 4,139 3,365 1,713 301

Source: TSTD2 and text.

1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833

Estimated Years 9,998 7,274 8,918 5,298 10,164 9,520 4,656 4,332 2,833 6,156 5,516 5,981 3,555 6,549 4,139 3,365 1,713 301

TSTD2

Estimated

1834 421 700 1835 0 1,400 1836 1,069 3,500 1837 837 6,650 1838 444 5,950 1839 5,201 5,250 1840 5,683 5,683 1841 4,539 4,539 1842 1,850 1,850 1843 816 1,400 1844 130 2,100 1845 888 888 1846 829 829 1847 993 993 1848 0 1,750 1849 0 1,050 1850 0 2,300 1851 1,350 350 All years 397,600 786,100

129

Chapter 4 The Transatlantic Slave

Trade to Bahia, 1582–1851 Alexandre Vieira Ribeiro

The pace of research on the transatlantic slave trade has accelerated recently, a pattern that perhaps derives from an increasing awareness of the importance of the topic for understanding societies on both sides of the Atlantic. The impact of the slave trade on African societies is central to understanding precolonial African history, and the enormous number of Africans who disembarked in the New World helped to shape the new societies of the Americas. In the Portuguese Americas, quite apart from the social and cultural ramifications of the trade itself, African slaves constituted the main labor force of a complex economic system, the sustainability of which was possible only via the Atlantic slave trade. The port of Salvador in Bahia performed an important role in both the reception of slave arrivals from Africa and their redistribution to the regional markets in the northeast and other areas of Brazil. What do the new data tell us about the size of the slave trade and the role of Bahia in that traffic? We start first with a survey of the existing knowledge before presenting new findings. Several studies have estimated the volume of the slave trade to Bahia. Luís Viana Filho estimated the disembarkation of slaves from 130

The Transatlantic Slave Trade to Bahia

Africa in Bahia during the sixteenth century at 20,000, with a further 205,000 arriving in the following century.1 Maurício Goulart considered that 2,000 Africans were imported every year through the port of Salvador in the first half of the 1600s, with the annual average rising slightly after 1650 before surging in the 1690s in response to the discovery of gold in the Brazilian hinterland.2 Goulart’s total for the seventeenth century was 225,000. In a recent study, David Eltis estimated 15,000 arrivals during the sixteenth century and approximately 180,000 in the seventeenth century.3 In contrast to the early period, documentation from the eighteenth century records slave arrivals in spans of several decades. Table 4.1 shows the numbers historians have provided so far. The differences between them can be explained in part by the type of documentation used by each historian. Moreover, in several cases, hard information contained in scattered serial or quantitative documentation was combined with qualitative sources, such as commercial reports or letters, or with data available from other studies. Pierre Verger’s estimates in column 4 are based only on the trade between Bahia and the Bight of Benin. In addition to these studies, TSTD1, published in 1999, resulted from the intention of its authors to track every ship that crossed the Atlantic to acquire slaves at the African coast.4 Voyages to some regions of the Americas such as Brazil, however, were underrepresented, and a second phase of research was initiated to focus on new sources situated in the Luso-Atlantic world. The final column in table 4.1 is in part a product of this new research. The new data about the slave trade to Bahia were drawn mainly from the Alvarás para navegar, that is, the licenses to sail to the African continent, located in the Arquivo Público do Estado da Bahia (APEB), in the Arquivo Nacional do Rio de Janeiro (ANRJ), and in the Biblioteca Nacional do Rio de Janeiro (BNRJ). These licenses allowed traders from Bahia to sail to Africa and obtain slaves in return for a range of commodities, especially tobacco, between 1678 and 1822. Besides the probable date of departure to Africa, this documentation also provides the ship name, the rig of the vessel, the names of captains and owners of the vessels, and the intended ports of embarkation. In addition, the new data drew on two other important sources. The first was the Livros de Visitas da Saúde (literally, “health visit books”) for the years 1776–99, 1809–10, and 1822–24, located at the Arquivo Histórico Municipal de Salvador (AHMS). The second is the newspaper A Idade d’Ouro do Brasil, located at BNRJ, which provides a short series for the years between 1811 and 1822.5 Both sources provide the date of arrival, the African port or region from which the slaves were shipped, the ship name, the rig, the names of the captains and owners, the

131

Table 4.1 A Comparison of Scholars’ Estimates of Slaves Disembarking in Bahia, 1681–1855 Years 1681–90 1691–1700 1701–10 1711–20 1721–30 1731–40 1741–50 1751–60 1761–70 1771–80 1781–90 1791–1800 1801–10 1811–20 1821–30 1831–40 1841–50 1851–55

Alden (by Schwartz)

Viana Filho

Goulart

Verger

Verger (by Manning)

Eltis

Santos

Ribeiro

— — — — — — — 63,500 29,500 31,500 24,000 39,000 — — — — — —

— — — — — — — — — — — 63,850 65,708 50,975 70,247 120,000 120,000 —

— — 70,000 70,000 66,256 47,520 46,016 38,416 41,446 29,816 20,233 62,301 54,900 55,000 55,000 55,000 — —

— — — — 14,250 (a) 47,500 41,468 (b) 24,615 (c) 19,267 (c) 15,554 (c) 12,234 (c) 40,842 (b) 38,339 (d) 55,352 72,066 1,675 (e) 63,046 785 (e)

17,200 60,800 86,400 67,200 63,400 49,000 39,200 34,400 36,000 30,000 32,700 53,100 72,900 59,000 51,800 54,800 63,000 —

17,747 55,687 76,868 85,993 69,451 32,712 39,160 33,913 43,852 34,506 34,918 59,689 87,635 70,700 71,600 32,500 66,100 1,900

— — — — — — — — — 47,032 50,933 74,524 74,151 — — — — —

13,344 43,089 53,303 67,240 53,207 38,517 46,795 36,421 50,522 57,435 56,796 69,406 72,262 71,951 75,529 — — —

Sources: Stuart Schwartz, Segredos internos (São Paulo, 1995), 283; Luís Viana Filho, O negro na Bahia (Rio de Janeiro, 1988), 155–57; Maurício Goulart, Escravidão africana no Brasil (São Paulo, 1975), 215, 216, and 272; Pierre Verger, Fluxo e refluxo (São Paulo, 1987), 661–63; Patrick Manning, “The Slave Trade in the Bight of Benin, 1640–1890,” in Henry A Gemery and Jan S. Hogendorn, eds., The Uncommon Market (New York, 1979), 136–38; David Eltis, “The Volume and Structure of the Transatlantic Slave Trade: A Reassessment,” William and Mary Quarterly 58 (2001): 36; Corcino Medeiros dos Santos, “A Bahia no comércio português da Costa da Mina e a concordância estrangeira,” in Maria Beatriz Nizza da Silva, ed., Brasil: Colonização e escravidão (Rio de Janeiro, 2000), 236–37; Alexandre Ribeiro, “O tráfico atlântico de escravos e a praça mercantil de Salvador, c. 1680–c. 1830” (master’s dissertation, UFRJ/ PPGHIS, Rio de Janeiro, 2005), app. 4.2, pp. 114–18. (a) Data for three years; (b) data for nine years; (c) data for six years; (d) data for eight years; (e) data for only one year.

The Transatlantic Slave Trade to Bahia

volume of slaves disembarked, the number of slaves who died in the Middle Passage, and the length of the voyage in days. Additional data were also obtained from the specialized literature, such as the studies mentioned in table 4.1, and from commercial letters, reports, and other qualitative sources. Researchers working at European and African archives have also contributed new data on the Bahian slave trade. Collectively, the new data provide a much more complete profile of the Bahian slave trade. Figure 4.1 compares the series of figures calculated from TSTD1 with a series derived from the data now available (TSTD2). The estimated number of African slaves imported in Bahia has grown approximately 500 percent between 1999 and 2006: from 223,699 to 1,349,724. Although some gaps remain in the series, we may consider the results observed in TSTD2 as being the most complete offered so far. Table 4.2 groups the data into decadal intervals, except for the first 118 years of the traffic. Overall, the 1.3 million slaves imported in Bahia according to table 4.2 represent about one-third of the entire African contingent disembarked in Brazil during the Atlantic slave-trade era.6 The first voyage to Bahia recorded in TSTD2 occurred in 1582, when the transition from indigenous to African slave labor was just beginning to get under way in the sugar mills of northeast Brazil.7 From the beginning of colonization, the economy of Bahia was based on sugar exports. Although there is evidence of an African presence in Bahia by 1560, for almost a century the Amerindians were a major source of labor on sugarcane plantations. The Portuguese had a strongly positive perception of Africans’ working capacity and had already had experience in using them as slaves when they began to colonize Brazil. Africans were usually 20

Number of slaves disembarked

18

TSTD1 TSTD2

16 14 12 10 8 6 4 2

Figure 4.1 Comparison of Slaves Disembarking in Bahia in TSTD1 and TSTD2, 1582–1855 (in thousands) Sources: TSTD1 and TSTD2.

45

36

Years

18

27

18

18

18

09

18

00

18

91

18

82

17

73

17

64

17

55

17

46

17

37

17

28

17

19

17

10

17

01

17

92

17

83

16

65

16

16

15

82

0

133

134

Alexandre Vieira Ribeiro

employed as domestic servants, urban traders, or craft workers on the Iberian Peninsula and on the Atlantic islands. Many of them possessed knowledge of the techniques for sugar production, as the Portuguese had established sugar plantations in São Tomé and the Madeira Islands well before settling Brazil. In Africa itself, of course, Africans were perfectly familiar with agriculture in general as well as ironwork. In addition to such considerations, the Indian population had experienced demographic catastrophe as a result of epidemiological contact with immigrants from the Old World. Further, the Jesuits insisted that the Crown abolish indigenous captivity. The first law against Indian slavery was officially published as early as 1570. Later, two more laws were promulgated in 1595 and 1609. Although Indian labor was not completely eliminated in the Portuguese Americas, the legislation, in addition to the great mortality of Indians, made African labor a more attractive alternative for the Portuguese. Nevertheless, this transition was completed only by the mid-seventeenth century. In 1587 Father Manuel Soares thought that 60 percent of the slaves in Bahia were Indians.8 Stuart Schwartz provides data on the engenho (sugar mill) Sergipe, and

Table 4.2 Estimates of Slaves Disembarked in Bahia, 1582–1851 Years

Number of Slaves

1582–1700 1701–10 1711–20 1721–30 1731–40 1741–50 1751–60 1761–70 1771–80 1781–90 1791–1800 1801–10 1811–20 1821–30 1831–40 1841–51 Total Source: Calculated from Appendix 4.1.

106,066 85,719 109,283 106,962 89,985 87,694 75,833 66,751 73,267 76,539 93,259 89,066 113,376 99,437 12,142 64,329 1,349,708

The Transatlantic Slave Trade to Bahia

they suggest a gradual growth of African labor in the Bahian plantations. In 1572, for example, only 7 percent of the 280 adult slaves working at Sergipe were Africans. By 1591 they represented 37 percent of the 103 slaves. Finally, by 1638 all the eighty-one slaves working in the mill were either Africans or crioulos.9 Table 4.3 summarizes the increasing presence of African slaves in the Bahian sugar mills, consistent with the foregoing narrative. The growth of slave arrivals appears particularly pronounced in the middle of the seventeenth century. Nevertheless, it is necessary to draw attention to gaps yet to be filled in our series of slave arrivals in Bahia between 1600 and 1675. As planters developed a preference for African as opposed to indigenous slaves, the price of the former grew to be three times that of the latter. Despite a problematic fiscal and commercial relationship with Portugal, the Bahian economy prospered during the seventeenth century. Sugar prices were high enough to allow sugar mill owners to afford the costs of buying African slaves and still generate profits. In the face of such prosperity, the Portuguese state permitted the settlers themselves to deal with the question of how to reproduce the labor force. They became increasingly dependent on the flow of African captives to renew their labor force. Among such captives, planters preferred adult slaves to children because the latter were not immediately useful in the sugarcane fields or mills. At the same time, African rulers did not want to sell female slaves. The slave trade was thus dominated by adult males. Table 4.3 Number of Slave Ships and Slaves Disembarked in Bahia, 1581–1700 Years

Number of Vessels

Number of Slaves Disembarked

1 1 1 6 4 4 5 13 16 33 66 151

166 356 361 2,134 1,423 1,428 1,771 4,626 5,532 11,734 23,515 53,012

1581–85 1606–10 1611–15 1646–50 1651–55 1661–65 1671–75 1676–80 1681–85 1686–90 1691–95 1696–1700 Source: Appendix 4.1.

135

136

Alexandre Vieira Ribeiro

By 1680 Brazilian sugar prices had fallen sharply. The sugar colonies of the French, English, and Dutch in the Caribbean were offering strong competition to the Luso-Brazilians. Europeans could now enjoy sources of sugar separate from Brazil. If in the 1630s some 80 percent of the sugar traded in London came from Brazil, in the 1670s this percentage fell to 40 percent, and by 1690 it was only 10 percent.10 With sugar now produced on a global scale, the Luso-Brazilians lost their international market share. Caribbean competition certainly squeezed the profits of Bahian sugar growers. It raised the price of slaves in Africa, and at the end of the seventeenth century the situation for the Bahian planters worsened when the Paulistas discovered gold mines in the interior of Brazil and added further upward pressure on prices of slaves passing through the port of Salvador (see figure 4.2). Easy access to gold meant that, over three decades, slave prices increased by 400 percent. Before the gold boom, a captive could be acquired for somewhere between 40 and 50 milreis. In 1730 the same slave cost more than 200 milreis. Given this situation, slave traders were able to accumulate considerable profits, receiving payment in gold. Meanwhile, the Bahian sugarcane planters were able to pay for their slaves only in sugar and, consequently, often had to borrow on promissory notes against future harvests.11 Landowners began to complain about the lack of slaves, and the Portuguese Crown decided to intervene in 1701 with a decree restricting the number of slaves that could be sent from Bahia and Pernambuco to the gold mines of Minas Gerais.12 This legislation quickly showed itself to be ineffective, as several Bahian merchants sought to divert slaves to the mines rather than to the sugarcane plantations. The drive for profits was stronger than fear of the law. For the sugar planters, slaves were increasingly out of reach, and while land remained abundant, sugar production fell. In 1702 Bahia sent 507,609 arrobas (over 16 million pounds) of sugar to Portugal—the output of 249 sugar producers, 100 of whom were planters rather than owners of engenhos.13 By 1710 Bahia was still supplying 507,500 arrobas of sugar, made in 146 mills, to Lisbon.14 Ten years later, however, in 1720, sugar production had fallen to around 420,000 arrobas (about 13 million pounds). Clearly the expanding slave trade to Bahia was supplying more labor to the mineral regions of the Brazilian hinterland than to the sugar-producing areas of the northeast. What routes did the slaves of this era follow after their arrival in Brazil? After leaving the port of Salvador, the slaves were distributed inland via the São Francisco and Velhas rivers, traveling more than 1,000 kilometers to their final destination at Minas Gerais.15 Prior to 1700, Bahian slave traders were

The Transatlantic Slave Trade to Bahia

Figure 4.2 Five-Year Averages of Slaves Disembarked in Bahia, 1582–1855 (in thousands) Source: Appendix 4.1.

the main suppliers of slaves to the gold mines of Minas. However, Bahian dominance in this trade came to be challenged by Rio de Janeiro merchants operating along the newly opened Caminho Novo (literally, “new way”).16 Via the Caminho Velho (the “old way”), which linked Rio to the mines through the port town of Paraty, a merchant caravan would take between forty-three and ninety-nine days to reach its destination, depending on the number of stops. Such trips were not competitive with the Bahian route. When the Caminho Novo was opened in 1711, however, the eighty leagues between Rio and Minas could be covered in only ten to twelve days.17 Although Bahia lost its dominant position in the slave market of Minas Gerais, it still performed an important complementary role in the internal slave trade. Between 1739 and 1759, Rio de Janeiro was responsible for about 65 percent of the 6,000 slaves that arrived every year in Minas. Bahia sent the remainder (around 2,100 workers every year).18 Between 1760 and 1770, as gold output went into decline, the volume of slaves sent to the mines fell. About 920 captives arrived in Minas annually from Bahia in that period, although Minas remained the principal destination of the slaves exported from Bahia (about 60 percent of the total).19 Certainly such slaves were employed not only in the gold mines but also in other economic activities ultimately dependent

137

138

Alexandre Vieira Ribeiro

on gold.20 Throughout the eighteenth century, the economic development of Minas Gerais remained the major center of demand for African slaves from Bahia.21 The flow of slaves leaving Bahia for regions other than Minas Gerais suggests a diversification of the colonial economy. New regions such as Goiás specialized in supplying foodstuffs to the internal market. Indeed, while Brazilian exports declined some 60 percent by value between 1760 and 1776, figure 4.2 shows the slave trade holding up reasonably well. The crisis in the mining regions was not, therefore, symptomatic of a general economic decline. In fact Portuguese America as a whole underwent a process of reallocating the factors of production, as other activities took the place of gold. During the first years of Pombal’s government, the Bahian slave trade certainly declined in the face of competition from Rio de Janeiro, in response to the crisis of the northeast agricultural sector, and because of the exhaustion of the gold deposits. Although the number of sugar mills in the Bahia-Sergipe region tended to grow, sugar production fell from 507,697 arrobas in 1702 to about 400,000 in 1758. The new policies of the Marquis of Pombal, including tax reductions and the abolition of the fleet system in 1765, did little to change this situation. Beginning in the early 1780s, however, war and revolution in North America and in the Caribbean transformed the situation for planters and mill owners in Brazil. A rebirth of agricultural and stock-herding activities in Bahia, resulting in a surge in the production of leather and tobacco as well as sugar, triggered a fresh inflow of slaves into agriculture. A central element in this rebirth was the 1791 slave revolt in Saint-Domingue, until then the largest sugar producer in the world. The resulting civil war, which ended in Haitian independence in 1804, eventually removed the island from sugar production permanently. Bahian planters and traders were able again to buy large numbers of African captives as sugar exports increased. Stuart Schwartz has characterized this period as the renaissance of Bahian agriculture. The number of mills producing sugar rose from 180 in 1758 to 260 in 1798, 340 in 1820, and 583 in 1834.22 Annual sugar exports are estimated at 10,000 boxes in 1770, increasing in the following decades to an average of 16,000 from 1796 to 1811, and the number of slaves disembarking in Bahia recovering to early eighteenth-century levels (see figure 4.2). During the nineteenth century, the slave trade to Bahia was shaped by politics as well as economics. The British began to exert pressure to abolish the Atlantic slave trade early in the century. They seized several vessels from Bahia along the African coast between 1811 and 1812, leaving relations between

The Transatlantic Slave Trade to Bahia

Lisbon and London tense and Bahian merchants anxious. The British effort to suppress the Atlantic slave trade resulted initially in a treaty signed in 1815 with Portugal that put the slave trade in African ports located above the equator off-limits to Portuguese vessels. A second treaty in 1817 permitted the naval vessels of one country to detain the slave ships of the other and established a Court of Mixed Commission to adjudicate such detentions. As these measures interdicted the large traffic from the Costa da Mina, Bahian slave traders created mechanisms to continue the trade without being detected. Many vessels left Salvador with false passports to buy slaves at Molembo and Cabinda, in West Central Africa, but in reality they headed to Ouidah, Onim (Lagos), Porto Novo, and Badagry in West Africa. About fifty-four ships were seized by the British in the Bight of Benin between 1822 and 1830 and judged in Sierra Leone, and forty of them were caught with passports to Molembo.23 Usually, passports to Molembo or Cabinda allowed Bahian captains to call at São Tomé or Princes Island and therefore to cross the equator northward. When stopped by the British above the equator, captains explained their presence by showing the British their passports to Molembo or Cabinda, but many were in fact going to or leaving traditional ports of slave supply in West Africa. Thus the ship Estrela, which had permission to call at these islands en route to Molembo, was seized in 1822 with a full complement of slaves embarked in Badagry.24 Some merchants even took two licenses for the same ship; one for Molembo or Cabinda, and the other for the Costa da Mina on the West African coast. Anacleto José Barbosa, for example, supposedly owned two vessels: the Leal Portuense and the Furão. Both had licenses issued on July 8, 1829. The first vessel had a license to trade commodities other than slaves at the Costa da Mina, but it never left the port of Salvador. The second, however, arrived in Salvador on November 6, 1829, disembarking 568 slaves from Cabinda and making an entrance in the documentation as “Furão”–Leal Portuense.25 Pierre Verger traced twenty-five combinations of real names and nicknames used by slave traders to avoid the terms of the 1815 treaty.26 As figure 4.2 shows, the Bahian slave trade continued to thrive despite the British efforts. The slave trade declined only between 1822 and 1824, when Bahia experienced conflicts related to Brazilian emancipation. In 1822 England recognized the new country of Brazil but pressed for a further treaty to put an end to the whole of the transatlantic slave trade to Brazil. These efforts produced the treaty signed on November 23, 1826, which stated that at the end of three years no slaves from Africa could be introduced into Brazil. Even with this treaty, however, Africans continued to disembark in Salvador with

139

140

Alexandre Vieira Ribeiro

the complicity of the local authorities until as late as 1850. Itaparica Island, Frade Island, and Itapuã Beach, all outside the limits of the port of Salvador, became sites of major disembarkations of slaves. Indeed, between 1830 and 1850, about 75,872 Africans disembarked in Bahia.27 In many Brazilian minds, the slave trade was still seen as underpinning Brazilian prosperity. As the British increased their naval pressure, Bahian merchants successfully adjusted their tactics, as the activities of Joaquim Pereira Marinho and José de Cerqueira Lima—two of the major slave traders of that time—demonstrated. By the end of the 1840s, pressures to abolish the slave trade appeared within Brazil. Apparently, British efforts eventually did have some impact on Brazilian public opinion about the trade in African people. Another factor—and maybe the most important—was the several rebellions that occurred in Brazil in the first half of the nineteenth century, especially in Bahia. Brazilians were afraid that a slave revolt, similar to that in Saint-Domingue, could take place in Brazil. Thus, in 1850 a law known as the Eusébio de Queiroz law formally abolished the Atlantic slave trade in all Brazilian ports. Finally, TSTD2 now allows us to identify the major African ports and regions of slave supply to Bahia. Table 4.4 shows that, during the first century of the transatlantic traffic, West Central Africa was the main source of slaves to Bahia. The Dutch, however, captured the ports of this region from the Portuguese in the 1640s, and Pernambuco, which the Dutch had captured in 1630, took over from Bahia as the major port of entry for slaves from that region. In 1648 military forces sailed from Rio de Janeiro under Salvador Correia de Sá’s command and restored Angola’s capital city, Luanda, to the Portuguese. From this moment on, political and economic ties were formed between these two cities in the South Atlantic. By the end of the seventeenth century, Portuguese governors at Luanda began to privilege the shipment of slaves to Rio.28 Such a strategy reduced the number of slaves from West Central Africa going to Bahia and other areas of northeast Brazil, and as a consequence these areas of Portuguese America directed their trade to other African regions. Besides Angola and Pernambuco, the Dutch West India Company also captured the São Jorge da Mina castle, or Elmina, the principal stronghold of Portuguese trade at Costa da Mina, in 1637. The Dutch sought to establish control over Portuguese trade in that region. According to Verger, despite the attentions of the Dutch and the later British restrictions, trade relations between Bahia and the West African ports, especially in the Bight of Benin, prospered for three reasons. First, this region was the only place where Bahian

Table 4.4

Slaves Disembarked at Bahia by African Coastal Region of Embarkation, 1581–1855 Senegambia Row Percentage

Years 1581–1680 1681–90 1691–1700 1701–10 1711–20 1721–30 1731–40 1741–50 1751–60 1761–70 1771–80 1781–90 1791–1800 1801–10 1811–20 1821–30 1831–55 Total

Gold Coast

— — — — 6,872 — 261 876 183 — 352 — — — 390 745 1,478 11,157

— — — — 5.6 — 0.3 0.9 0.2 — 0.4 — — — 0.3 0.6 3.2 0.8

Bight of Benin

Row Percentage — 878 439 — 5,508 3,744 1,348 — — — — — 1,070 510 1,081 — — 14,577

— 4.5 0.5 — 4.5 3.2 1.3 — — — — — 1.0 0.5 0.9 — — 1.0

Bight of Biafra

Row Percentage 1,617 14,181 62,686 87,342 80,826 76,873 60,139 53,183 54,520 46,388 37,735 52,137 67,043 71,361 54,383 14,609 33,268 868,293

Source: TSTD2 as of November 2006. Totals may not add up due to rounding.

11.9 72.1 72.9 90.0 66.2 65.0 59.6 53.9 66.4 63.2 47.4 62.3 65.5 73.4 44.9 12.7 71.9 59.5

West Central Africa

Row Percentage 2,021 501 9,271 7,680 17,257 6,599 5,385 4,740 1,287 — 317 685 902 1,514 7,724 798 2,856 69,538

14.9 2.5 10.8 7.9 14.1 5.6 5.3 4.8 1.6 — 0.4 0.8 0.9 1.6 6.4 0.7 6.2 4.8

Southeast Africa

Row Percentage 9,081 3,638 13,538 1,977 10,786 30,532 33,701 39,925 24,902 26,678 41,164 30,423 33,349 23,474 46,492 94,628 8,657 472,944

66.8 18.5 15.7 2.0 8.8 25.8 33.4 40.4 30.3 36.4 51.7 36.3 32.6 24.2 38.4 82.3 18.7 32.4

Row Percentage 868 464 — — 868 464 — — 1,248 337 41 464 — 297 10,935 4,136 — 20,125

6.4 2.4 — — 0.7 0.4 — — 1.5 0.5 0.1 0.5 — 0.3 9.0 3.6 — 1.4

142

Alexandre Vieira Ribeiro

traders could sell locally produced third-quality tobacco, which did not find acceptance in Portugal. Second, the Dutch United Provinces, which dominated this region in the early period, tolerated the Portuguese slave trade in order to obtain access to Brazilian tobacco on the coast. Finally, the competition Bahian merchants in West Africa faced from other parts of Brazil was reduced, since non-tobacco-producing areas like Rio de Janeiro were forbidden to trade at the Bight of Benin.29 A royal decree of November 12, 1644, allowed slave traders to carry third-quality tobacco direct from Salvador to the Bight of Benin to buy slaves, since Angola was still under Dutch control.30 Thus a bilateral trade system was established between Bahia and West Africa, instead of the classical triangular system (Europe–Africa–America–Europe). The African elite had slaves, and Bahian merchants had tobacco. But tobacco from anywhere was not sufficient to sustain trade. The tobacco had to be from Brazil. Although the Bahian tobacco intended for Africa was of a low quality, it was different from any other. It was prepared with pure sugarcane syrup, for which Africans developed a taste, and it became an indispensable commodity for the export slave trade of parts of West Africa. It is estimated that Bahia exported annually 13,000 arrobas (about 416,000 pounds) of tobacco to the Bight of Benin at the beginning of the eighteenth century.31 Table 4.4 shows that the large expansion of the Bahian slave trade, and especially the traffic from the Costa da Mina, coincided with the establishment of the traffic in tobacco. The main ports where Bahian merchants traded at the Bight of Benin were Grand Popo, Ouidah, Apa, Jakin, and Offra. The prevalence of these ports was dictated by a Luso-Dutch treaty that required Bahian-based Portuguese merchants to pay a duty to the Dutch at Elmina amounting to 10 percent of their trading cargoes. In return for this levy, the Dutch allowed the Portuguese to trade without Dutch interference at five ports in the Bight of Benin. Captains from Bahia normally paid the duty prior to beginning their quest for slaves, and the Dutch received about 300 rolls of tobacco per year.32 Nevertheless, Bahian traders also took illegal goods to trade at the Bight of Benin, such as rum and gold dust. By 1720, however, Brazilian trade in the Bight of Benin had reached its peak. French traders began to compete against the Portuguese in the slave markets of West Africa and came to dominate the region’s slave trade for most of the eighteenth century. British and Dutch merchants were also present in the region, but they did not represent as great a threat to Bahian traders. Political conflicts among the African kingdoms in the interior were also a destabilizing factor in West African slave trading. After playing the role of a mercenary state, the Dahomey kingdom attacked its neighbors in a process of

The Transatlantic Slave Trade to Bahia

expansion. The political conflicts among several African kingdoms eventually reached the coast, with the invasion of Ouidah in 1727 by Agaja, king of Dahomey. In an attempt to weaken Dahomey’s power, the Oyo Empire, which dominated the slave trade in the interior of the Costa da Mina in the first decades of the eighteenth century, attacked Dahomey between 1726 and 1730 to obtain tributes and also to protect Oyo commercial routes to the coast. In response, in 1732 Dahomey attacked and conquered Jakin, a port that was under Oyo control. The Oyo Empire was thus obliged to look for new ways to reach the sea.33 The wars unleashed by Dahomey’s aggression created serious obstacles for Bahian activities along the African coast. By the end of the 1720s, the Portuguese viceroy in Brazil wrote that vessels which had sailed from Bahia and Pernambuco more than a year earlier still had not returned home. Vessels that did arrive took sixteen months to load slaves in Africa and get to Brazil, an operation that had previously taken just over six months.34 In effect, the delay was caused by the political conflicts among the different African kingdoms, which blocked the commercial routes that linked the interior to the coast. Clearly, African traders experienced many difficulties in bringing slaves from the interior to the coastal slave markets. At the same time that the slave trade decreased in areas under Dahomey’s control, new ports were opened on the eastern coast of the Bight of Benin, such as Porto Novo, Badagry, and Onim. In these places the captains could trade with little difficulty, and they received the help of João de Oliveira,35 an ex-slave from Pernambuco, who opened up direct trading in 1758 with Porto Novo, as well as on the island where the state of Onim was situated, a site that became known as Lagos.36 With these eastern ports open, Badagry overtook Ouidah as the most important slave exporter in the region by 1770. After Dahomey attacked Badagry in 1783, Porto Novo assumed the leading position. At the beginning of the nineteenth century, Lagos became the region’s principal slave port. Between 7,000 and 10,000 Africans may have embarked from this port to the Americas.37 Luso-Brazilians could count on the help of the ruler of Onim—the ologun—who preferred them to the English and other Europeans trading in the region.38 In 1770, even before Lagos rose as a trading port, the king of Onim sent an embassy to Salvador to establish initial trading contacts with the Bahian authorities. The political conflicts in Africa, together with the European instability caused by the Seven Years’ War (1756–1763), the American War of Independence, the French Revolution, and the struggles that gave rise to Haitian

143

144

Alexandre Vieira Ribeiro

independence, resulted in Bahia drawing on new African trading ports after 1760 and strengthening its slave-trade ties with the Bight of Benin. This process was further encouraged when, in 1791, the Dutch stopped extracting the 10 percent tribute from Luso-Brazilian vessels.39 Table 4.4 shows that the number of slaves disembarked in Bahia from this region increased between 1780 and 1810, with a decline in the decade following 1810 attributable to British efforts to suppress the slave trade above the equator. The trade between Bahia and the Bight of Benin reached its lowest level after 1820. Other areas of West Africa also traded slaves with Bahia. Coastal regions such as Senegambia and the Gold Coast did not contribute the same volume as the Bight of Benin, because other European powers such as the English and the Dutch had a strong presence in those areas, reducing the trading capacity of Bahian merchants. West Central Africa, by contrast, ranked second in the Bahian supply of slaves during the eighteenth century. Perhaps the ties developed by Bahian merchants at the Bight of Benin diverted them from developing Angolan trading networks as strong as those of traders from Rio de Janeiro. Nevertheless, table 4.4 shows two periods when shipments of slaves from West Central Africa to Bahia exceeded those to Rio de Janeiro. The first was the decade of the 1770s, when the slave trade in West Africa was experiencing conflicts owing to the Dahomey wars of expansion, and when the northern Europeans who traded heavily in northern Angola were embroiled in conflict. The second was after 1815, when the British imposed restrictions on the slave trade above the equator. In 1820 the volume of slaves disembarking from Kongo and Angola constituted about 80 percent of the total slaves disembarked in Salvador. However, as previously mentioned, one should be careful with these numbers, because while many ships arrived in Bahia with documents recording Molembo or Cabinda as the source of their captives, it is highly possible that they in fact embarked their slaves in West Africa. Despite the fact that some slaves arrived from Southeast Africa as early as 1660, exports of East African captives to Bahia remained irregular until the end of the eighteenth century, and before 1811 the participation of Indian Ocean ports in the Bahian slave trade was only occasional.40 Two factors shaped this pattern: the distance and the high shipboard mortality of the captives. According to André de Melo e Castro, the Count of Gouvêas, slave voyages to Mozambique were not worth the effort because of ocean hazards and the weak condition of the survivors of such voyages.41 It is possible that Southeast African voyages began in earnest after 1815 as a reaction to the

The Transatlantic Slave Trade to Bahia

British attempt to suppress the slave trade, when Bahians sought an alternative to embarkation points north of the equator. Entrepreneurial risks, as well as risks to human life and welfare, were an inevitable characteristic of the complex business of selling human beings. The possibilities were constant of slaves escaping, stealing, and above all dying. Moorish pirates and French, English, Spanish, Dutch, North American, and eventually Argentinean corsairs constantly attacked Bahian slaving vessels, stealing millions of tons of sugar, rum, leather, tobacco, cotton, and amber, as well as hundreds of slaves. This loot was then sold in ports in the Americas, Africa, and even along the Mediterranean. Bahian traders were frequently attacked while awaiting shipments. Thus, in 1794 a French squadron sent to Africa under the flag of the revolutionary National Assembly captured slave vessels of various nationalities. This expedition, with six ships, 1,040 crewmen, and 131 cannons, swept the coast between Sierra Leone and Grand Popo until December 1794, capturing nine ships from Liverpool, four from London, one from Bristol, three from Holland, and seven whose nationality was not recorded. In three days, December 6–8, 1794, this squadron also took eight Luso-Brazilian slavers between Ouidah and Porto Novo.42 In August 1797 two French ships and two sloops under the English flag captured two Bahian slave vessels at Ouidah— the S. João Nepomuceno e S. Francisco de Paula and the Nossa Senhora da Graça. Soon after, one of the French ships appeared at the port of Apê, where it captured the Zabumba, under Master Raimundo Justino da Graça’s command; later it captured the sloop Paraíso at Porto Novo.43 On October 15, 1798, the English vessel Christopher from Liverpool captured the bergantim Nossa Senhora da Vitória e São José, which had sailed from Bahia to Elmina, and took it to Cape Coast. The bergantim’s master, Manuel Duarte da Silva, who was trading onshore at the time of the capture, traveled to Cape Coast to try to regain his ship. There, the English captain, John Watson, kidnapped him and took him to Onim. Part of the cargo from the Bahian vessel was then negotiated by Watson as if it had been his own. After sixty-four days he took Duarte da Silva, with two pilots and some other members of his crew, to Cape Coast, leaving behind eleven men and the bergantim. In Cape Coast, the master and his crew ran away to Elmina, where they reported the case and embarked on a canoe to Ouidah. Only at Ouidah was the Portuguese master finally able to journey back to Bahia, in the brig Americano.44 British maritime attacks increased after the British abolished their own slave trade. They demanded that Portugal also abolish its Atlantic slave traffic.

145

146

Alexandre Vieira Ribeiro

Bahian traders resisted the English efforts, because the West African coast supplied most of the slaves imported through Salvador. The situation worsened when the British prohibited the trade above the equator in 1815. Between 1811 and 1830, about eighty-five vessels from Bahia were captured by British forces.45 Apart from conflicts on the African coast and during the Middle Passage, shipboard mortality contributed further risks to all involved. Deaths always occurred, but the rates varied widely across expeditions. Félix da Costa Lisboa’s brig Correio da Guiné lost 230 out of the 340 slaves that he had bought in the Bight of Biafra at the end of 1805.46 Such losses could bankrupt a slavetrade company. That happened in 1809 to Felipe Justiniano Costa Ferreira after a disastrous voyage to Mozambique, when 192 out of the 557 slaves he bought died in the Middle Passage.47 Because he had just purchased his vessel, Costa Ferreira was forced to sell his other ship, the brig Flor da Bahia, ten days after it returned from its own expedition.48 There is no more record of his participation in the slave trade after this occurrence. At the other extreme, the brig Cipião Africano, under Domingos José Lima’s command, arrived in Salvador in 1807 with all the 135 Africans bought at the Gulf of Guinea having survived the crossing. On October 15, 1789, the brig Netuno Pequeno was visited by medical agents from the senatorial chamber of Salvador, whose task was to inspect ships for diseases. The vessel had arrived in Salvador with a small cargo of twenty-nine slaves after a month’s voyage from Costa da Mina, with a short layover in Recife. No deaths had been reported among its cargo, and its captain and owner, Pedro Gomes Ferreira, even informed the port bureaucrats that “two new slaves were born in the shelter of the brig.”49 This was an extremely unusual occurrence in the transatlantic slave traffic. During the Middle Passage, Africans were much more likely to die than to give birth. The awful conditions under which the Africans traveled perhaps could account for the mortality during the Middle Passage. Epidemiological factors seem to have played an important role in the number of slaves dying in the transoceanic voyage. Dissemination of diseases onboard was doubtless related to the facts that slaves were kept chained, between decks with little ventilation, and in what were probably the filthiest conditions in the history of longdistance transportation. Voyages that took longer than average were particularly likely to experience elevated mortality, since food and water would be rationed. Another major explanatory factor could also lie within Africa, given that prior to embarkation the captives had already experienced psychological trauma and physical misery of the kind common among defeated soldiers.

The Transatlantic Slave Trade to Bahia

Not infrequently, they arrived at the coast already weakened and infected by diseases, resulting in elevated mortality during the voyage. These factors might have been at the root of the increase in loss of slaves among Bahian ships operating on the Angolan route during the first decade of the nineteenth century. From 4.5 percent in 1803–5, the mortality level climbed to 13.6 percent in 1806 and 16.2 percent in 1807.50 There are records of outbreaks of smallpox occurring in Luanda in 1805, 1807, and 1808, and when further outbreaks occurred in Angola in 1811 and 1814, Bahian traders apparently tried to avoid embarking slaves in the region. Whereas thirty-seven ships had used this route during the first five years of the nineteenth century, in the following five years this participation fell to twenty-three ships, and ventures remained at this reduced level between 1811 and 1815, when just twenty-five ships chose the route.51 Disease not only made the life of the slaves miserable; it also affected the crew. On March 9, 1810, the bergantim Ligeiro arrived in Salvador from the Bight of Benin. It had lost 84 of the 466 slaves embarked during the voyage, as well as part of the crew and even the captain.52 These ships usually faced such diseases of the time as smallpox, scurvy, chicken pox, scabies, measles, eye infections, and diphtheria.53 The slave trade brought the epidemiological environments of the Old World and the New into contact, causing large numbers of deaths not only in the vessels but on both sides of the Atlantic as well.54 The populations of the port cities of Brazil suffered continually from diseases introduced by the slave ships. One contemporary source claimed that ships were the main vehicles of infectious diseases in Salvador.55 The governor of Bahia received frequent requests to prevent the spread of contagious diseases in Salvador attributed to the arrival of vessels from Africa. Citizens demanded that newly arrived slaves and crew members spend up to forty days in quarantine. The government eventually agreed with these requests, choosing the Ponta do Monserrat as the quarantine station for both groups.56 In general, slaves died in direct proportion to the length of the transatlantic voyage, with distinct patterns of “losses in transit” according to the African region in which they were embarked. Table 4.5, developed from TSTD2, calls attention to the geographical differences in slave mortality and to the variation in the passage time of the vessels that brought slaves into Salvador between 1790 and 1819. Mortality, of course, did not end with the disembarkation of the slaves. Many Africans who had survived the Middle Passage were so weakened and sick that they had to remain in quarantine a long time after their arrival.

147

148

Alexandre Vieira Ribeiro

Table 4.5 Slave Deaths as a Percentage of Slaves Embarked on Vessels Arriving in Bahia, 1790–1819, by Coastal Region of Embarkation (number of slaves in parentheses) Years 1790–99 1800–1809 1810–19

West Africa

West Central Africa

East Africa

8.2 (34,576) 6.0 (36,940) 7.0 (32,722)

7.8 (26,910) 8.9 (10,931) 5.6 (11,328)

— — 28.1 (3,827)

Source: TSTD2 as of November 2006.

Slaves in this condition incurred losses for traders, and even after they were sold, sick Africans had to be given special care. Medical data from Salvador’s Ponta do Monserrat between 1822 and 1824 enable us to estimate the effects of mortality among recent slave arrivals. About 8 percent of the Africans arriving alive in Bahia remained in quarantine. Research indicates that such slaves remained in quarantine from two weeks to two and a half months on average. More than half of the slaves held in quarantine died—suggesting an overall mortality in the first few weeks of approximately 5 percent of the captives disembarked alive in Bahia. Table 4.5 shows mortality according to the regions of embarkation. Between 1790 and 1819, deaths among slaves from West Africa were usually higher than among slaves sailing from Angola. As the length of voyage between West Central Africa and Salvador was shorter than that from West Africa (see table 4.6), regional mortality differences were no doubt partly due to voyage length differences.57 However, the mortality was higher on the Angolan route during 1800 and 1809, owing essentially to problems related to smallpox in this region. But much higher again than either of the west coast branches of the traffic was the mortality among slaves carried from Southeast Africa (ca. 28 percent). The average number of days to complete a voyage on this route was about fiftyeight, so that once more voyage length was undoubtedly important. Indeed, the decrease in mortality on vessels leaving from Angola observed in the second decade of the nineteenth century might well be because of the decrease in voyage length observed in table 4.6 over this period. The reason for the shorter voyages after 1810 is not clear at this stage and calls for further investigation. Overall, the traffic to Bahia was one of the most important branches of the transatlantic slave trade. It constituted an enduring transatlantic connection between the Old World and the New, and more slaves passed through Bahia than

The Transatlantic Slave Trade to Bahia

Table 4.6 Average Length of Voyage in Days between Africa and Bahia by African Region of Origin, 1790–1824 (number of voyages in parentheses) Years 1790–94 1795–99 1800–1804 1805–9 1810–14 1815–19 1820–24

West Africa

West Central Africa

Southeast Africa

46 (61) 45 (76) 47 (35) 46 (106) 41 (105) 40 (14) —

37 (44) 40 (26) 37 (11) 36 (17) 35 (10) 28 (50) 35 (53)

— — — — 47 (1) 62 (10) 63 (2)

Source: TSTD2 as of November 2006.

through any other New World port, with the single exception of Rio de Janeiro. Nevertheless, the level of the traffic varied substantially over nearly three centuries. Sources of demand for slaves within Brazil shifted markedly from sugar to mining and then back to sugar again. The sources of supply shifted from West Central Africa to West Africa before becoming more dispersed from the Bight of Benin to Southeast Africa in the nineteenth century. Less susceptible to change were the conditions that the slaves experienced during the transatlantic crossing. While the voyage length shortened after 1810, mortality, including that occurring before and after the voyage, remained high. Finally, throughout the nearly three centuries of its existence, the slave trade to Bahia remained a business whose risks remained higher than in most other entrepreneurial endeavors. These and other aspects of the traffic to Bahia may now be subjected to greater in-depth exploration than ever before. As a result of recent research, the Bahian slave trade is now the best-documented branch of the Brazilian trade from the late seventeenth to the mid-nineteenth centuries. NOTES

1. Luís Viana Filho, O negro na Bahia (Rio de Janeiro, 1988), 158. 2. Maurício Goulart, Escravidão africana no Brasil (São Paulo, 1975), 113. 3. David Eltis, “The Volume and Structure of the Transatlantic Slave Trade: A Reassessment,” William and Mary Quarterly 58 (2001): 36. 4. David Eltis, Stephen D. Behrendt, David Richardson, and Herbert S. Klein, The Transatlantic Slave Trade: A Database on CD-ROM (Cambridge, 1999), henceforth referred as TSTD1.

149

150

Alexandre Vieira Ribeiro

5. Raymundo Nina Rodrigues was the first scholar to call attention not only to the newspaper A Idade d’Ouro do Brasil but also to the books found in APEB as potential sources to estimate the trade of African slaves to Bahia. Although Rodrigues did not make full use of these documents, he stressed the higher number of slaves imported from Western African compared with other African regions. See Raymundo Nina Rodrigues, Os africanos no Brasil (São Paulo, 1982), 22–26. 6. In past studies, Viana Filho guessed this number. According to him, 1.3 million slaves had entered into Bahia during the whole traffic period. See Viana Filho, O negro na Bahia, 159. 7. It seems that the movement of captives crossing the two sides of the Atlantic had started before. According to Affonso de Escragnolle Taunay, in 1538 Jorge Lopes Bixorda, contractor of the brazilwood trade, inaugurated the transatlantic traffic by carrying some black slaves to Bahia. See Affonso de E. Taunay, “Subsídios para a história do tráfico africano no Brasil colonial,” Anais do Terceiro Congresso de História Nacional, vol. 3 (Rio de Janeiro, 1941), 533. Goulart also suggested that the traffic to Bahia began before 1582. According to him, a group of African slaves arrived in Salvador in 1550. See Goulart, Escravidão africana no Brasil, 58. 8. Goulart, Escravidão africana no Brasil, 100. 9. Stuart B. Schwartz, Segredos internos: Engenhos e escravos na sociedade colonial 1550–1835 (São Paulo, 1988), 68. 10. Ibid., 162; Herbert S. Klein, A escravidão africana: América Latina e Caribe (São Paulo, 1987). 11. Schwartz, Segredos internos, 166. 12. APEB, Ordens Régias, 20–6–1703. This decree determined also that people from São Paulo could only acquire 200 workers from Angola every year through the harbor of Rio de Janeiro. 13. Schwartz, Segredos internos, 149. 14. According to Antonil, 14,000 boxes of sugar were exported to Portugal. See André João Antonil, Cultura e opulência do Brasil (São Paulo, 1976), 140. 15. Ibid., 186–87. 16. From Rio de Janeiro, the route to Minas Gerais was by ship to the port of Nossa Senhora do Pilar (future port of Estrela) at the back of Guanabara Bay and, from there, upriver to Morobaí, where the road passed through the mountains to the mines. Alternatively, one could take a land route through Irajá, eventually arriving at the foot of the mountains and the road. Regarding the Caminho Novo to the mines, see Antonil, Cultura e opulência, 184–86. 17. Ibid.. See also Guillaume François Parscau, “A invasão francesa de 1711,” in Jean Marcel Carvalho França, ed., Outras visões do Rio de Janeiro colonial: Antologia de textos (1582–1808) (Rio de Janeiro, 2000), 135. 18. Goulart, Escravidão africana no Brasil, 170. 19. Alexandre Vieira Ribeiro, “O tráfico atlântico de escravos e a praça mercantil de Salvador, c. 1680–c. 1830” (master’s dissertation, Universidade Federal do Rio de Janeiro, Programa de Pós-graduação em História Social, Rio de Janeiro, 2005), 95. 20. Regarding the changes in Minas Gerais’s economic activities at the end of the eighteenth century and the beginning of the nineteenth century, see Roberto Borges

The Transatlantic Slave Trade to Bahia

Martins, “Minas Gerais, século XIX: Tráfico e apego à escravidão numa economia não-exportadora,” Estudos Econômicos 13 (1983): 181–209. 21. The second region that received the most slaves from Bahia was Goiás. Approximately 13 percent of the slaves disembarking in Salvador were sent in the direction of this region. In the villages, farms, and mines of Goiás and Mato Grosso, between the end of the eighteenth century and the first decades of the nineteenth century, the slaves from West Africa were the majority group among the Africans. See Mary C. Karasch, “Central Africans in Central Brazil, 1780–1835,” in Linda M. Heywood, ed., Central Africans and Cultural Transformations in the American Diaspora (Cambridge, 2002), 117–52. 22. The statistics relative to the plantations and their production in Bahia are taken from Schwartz, Segredos internos, 150, 169, 337, and 340; and from B. J. Barickman, Um contraponto baiano (Rio de Janeiro, 2003), 74–75 and 346. 23. Pierre Verger, Fluxo e refluxo (São Paulo, 1987), 107. 24. TSTD2. 25. Ibid. 26. Verger, Fluxo, 440. 27. TSTD2. 28. Regarding the establishment of commercial ties between Rio de Janeiro and Angola, see Luiz Felipe de Alencastro, O trato dos viventes: Formação do Brasil no Atlântico Sul, séculos XVI e XVII (São Paulo, 2000). 29. Verger, Fluxo, 19–20. 30. Ibid., 21. 31. Antonil, Cultura e opulência, 157. 32. Verger, Fluxo, 35. 33. Paul E. Lovejoy, A escravidão na Africa: Uma história de suas transformações (Rio de Janeiro, 2002), 136–37. 34. “Carta do vice-rei português no Brasil para Lisboa, em 29 de abril de 1730,” in Verger, Fluxo, 146–49. 35. AHU, Bahia, Col. Castro e Almeida, caixa 44, doc(s). 8244–51. 36. Alberto da Costa e Silva, Um rio chamado Atlântico: A África no Brasil e o Brasil na África (Rio de Janeiro, 2003), 119. 37. TSTD2. 38. Silva, Um rio chamado Atlântico, 124. 39. Verger, Fluxo, 228. 40. TSTD2. 41. APEB, Col. Ms., Ordens Régias, vol. 35, p. 54. 42. “Correspondência dos governadores,” in Anaíza Vergolino-Henry and Arthur Napoleão Figueiredo, A presença africana na Amazônia colonial (Belém, 1990), doc. 80, pp. 128–29 (original in the Arquivo Público do Pará). 43. APEB, maço 193, doc. 33. 44. APEB, maço 193, doc. 43. 45. TSTD2. 46. TSTD2. 47. TSTD2.

151

152

Alexandre Vieira Ribeiro

48. BNRJ, Idade d’Ouro do Brasil, BA, January 1, 1819, and January 18, 1819. 49. APEB, códice 178.1, p. 96. 50. TSTD2. 51. AHMS, códice 182.1; José Curto and Raymond R. Gervais, “A dinâmica demográfica de Luanda no contexto do tráfico de escravos do Atlântico Sul, 1781–1844,” Topoi 4 (2002): 122. 52. TSTD2. 53. Luís dos Santos Vilhena, A Bahia no século XVIII (Salvador, 1969), vol. 1, p. 156. 54. See Philip D. Curtin, The Atlantic Slave Trade: A Census (Madison, Wis., 1969), 283–86. 55. Vilhena, Bahia, 156. 56. Ibid. 57. TSTD2.

The Transatlantic Slave Trade to Bahia

Appendix 4.1 1582–1851

Estimated Number of Slaves Disembarked in Bahia,

Year

Number of Slaves

Year

Number of Slaves

Year

Number of Slaves

Year

Number of Slaves

1582 1610 1611 1650 1651 1652 1653 1661 1664 1665 1671 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702

166 356 361 2,134 356 711 356 356 711 361 1,778 717 711 711 708 1,778 474 356 356 1,857 2,490 1,067 2,134 1,778 2,490 4,265 5,335 5,019 2,846 4,980 5,335 7,509 10,671 13,803 8,909 12,121 8,904 8,181

1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740

8,181 7,120 6,047 7,825 13,200 8,121 9,248 8,892 9,604 14,256 13,872 13,927 10,082 10,418 10,400 8,441 9,405 8,878 8,548 7,228 9,368 17,106 11,775 12,883 9,777 12,963 11,877 5,437 12,965 7,287 8,038 10,199 7,767 9,028 9,022 9,290 7,084 9,305

1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778

3,798 10,192 5,580 12,214 4,403 6,908 12,761 9,244 10,390 12,204 7,355 9,639 9,794 5,842 7,034 8,462 7,267 7,516 5,459 7,465 7,102 6,357 5,826 3,381 6,220 9,772 7,327 7,535 7,561 5,670 7,441 4,600 5,568 7,861 8,950 7,799 3,497 8,140

1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816

6,402 13,009 12,583 11,869 8,509 9,385 6,470 7,270 2,127 4,867 7,104 6,355 6,707 6,965 9,062 13,034 11,504 8,139 8,100 9,147 9,816 10,785 7,255 4,125 8,130 9,886 9,323 11,448 10,350 6,535 8,163 13,851 8,791 11,153 10,962 8,779 11,792 9,213 (continued)

153

154

Alexandre Vieira Ribeiro

Appendix 4. 1 (continued) Year

Number of Slaves

Year

Number of Slaves

Year

Number of Slaves

Year

Number of Slaves

1817 1818 1819 1820 1821 1822 1823 1824 1825

12,841 13,016 16,792 10,037 9,040 13,997 4,177 4,692 6,571

1826 1827 1828 1829 1830 1831 1832 1833 1834

14,248 16,197 7,378 15,784 7,353 352 352 1,253 456

1835 1836 1837 1838 1839 1840 1841 1842 1843

1,401 1,055 702 1,275 2,828 2,468 1,471 4,452 3,111

1844 1845 1846 1847 1848 1849 1850 1851 Total

6,601 2,935 7,504 11,019 7,119 10,056 9,461 600 1,349,708

Source: TSTD2 as of November 2006.Totals may not add up due to rounding.

Chapter 5 The Origins of Slaves

Leaving the Upper Guinea Coast in the Nineteenth Century Philip Misevich

Although the Atlantic slave trade has captured the popular imagination for centuries and attracted much scholarly attention since the 1960s, details of its workings on the African continent remain less clear than for the Americas. This chapter seeks to contribute to the understanding of the African end of the slave trade through an analysis of the Registers of Liberated Africans in conjunction with the revised edition of the Transatlantic Slave Trade Database (TSTD2). The linguistic identification of nearly 1,000 African names in these registers makes possible a partial reconstruction of the geographic origins of slaves embarked on vessels trading along the coast of Sierra Leone between 1824 and 1841. Most significant, this analysis facilitates an estimation of the distances between their regions of enslavement in the interior and their points of sale on the Atlantic coast. Ultimately, what results is a new picture of the sources of supply for the Atlantic slave trade from the hinterland of Sierra Leone in its final phase of existence. The Registers of Liberated Africans were produced following the British abolition of the slave trade in 1807. To prevent further 155

Map 5.1 Upper Guinea Coast

The Origins of Slaves Leaving the Upper Guinea Coast

shipments of slaves from crossing the Atlantic, the British signed treaties with the governments of those nations most active in the trade. One result of the treaties was the establishment of courts of mixed commission.1 Eventually numbering eight, these courts were employed to adjudicate cases concerning vessels suspected of taking captives illegally on board. Each case was presided over by two officials, one British and the other representing the nation from which the captured slave trader originated. In many cases, the vessels detained by the British antislavery squadron had full slave cargoes still onboard, and in such cases the African captives were “liberated,” ultimately being absorbed into the colony where the court was established. TSTD2 allows us to identify the place on the African coast where the African captives had embarked. Prior to their liberation, the Africans were asked to provide information about themselves that was recorded by colonial officials in ledger books.2 Each page of these books was divided into seven columns and each “recaptive” was given one line on the page. The details recorded by the officials included a unique identity number, name, sex (man, woman, girl, or boy), age, height, and a description of any distinctive body markings. In the case of the Sierra Leone and Havana courts, where the large majority of vessels were tried, an additional column for “Country of Origin” was included, although the Sierra Leone officials eventually discontinued this practice.3 In addition, the registers clearly indicate that this information was provided through an African translator who was previously freed by the courts, and who came from the same region as the slaves on board the vessel under adjudication, thus partially mitigating the normally severe linguistic barriers that existed between European officials and African captives. To my knowledge, the registers represent the earliest extant list of indigenous African names. These records provide an unparalleled opportunity to determine the geographic origins of slaves exported in the second quarter of the nineteenth century. More specifically, this chapter focuses on vessels that embarked slaves along the coast of Sierra Leone and were captured in the Caribbean, drawing conclusions from only those names included in the Havana Registers. Vessels that were captured before crossing the Atlantic were adjudicated in the Sierra Leone court and remain the subjects for a future study. From 1824 to 1841, the names of more than 1,600 slaves liberated from nine separate vessels by the British antislavery squadron and originating in the larger Sierra Leone region were recorded in the Havana Registers.4 In a large majority of the cases, the names are obviously indigenous and remain prevalent in parts of modern-day

157

158

Philip Misevich

Sierra Leone and surrounding countries. By linking the indigenous names of the liberated Africans as recorded in the Havana Registers with the geographic regions in which they remain current today, estimates can be made of the points of the recaptives’ enslavement. Recent work by scholars making use of the registers in a similar way has generated new conclusions about the role of the African interior as a source of supply for the Atlantic slave trade, but for a different African region.5 This project is the first to employ the registers to determine the interior origins of Africans exported from Sierra Leone.6 The details in these registers were recorded by Spanish colonial officials, and it is clear that they had much difficulty transcribing the names provided by recaptives. Most commonly, the names were written as approximations based on the conventions for writing Spanish. But while colonial officials’ inability to spell what must have been quite unfamiliar names is evident, the association of the names in the registers with those found in contemporary Sierra Leone is nonetheless apparent. To facilitate their identification, this study worked backward from the way in which the records were generated. Because the names were transcribed in Cuba using Spanish spelling conventions, assistance was sought from a native Cuban, who tape-recorded the likely pronunciation of each recaptive’s name. Each name was then played in Sierra Leone for a group of informants, who attempted to identify the name itself and the geographic region in which it is commonly used.7 When disagreements between consultants occurred, the informants discussed with each other their reasoning for the identification they suggested, attempting to reach a consensus. In cases where they were unable to reconcile their differences, the name was discarded from the final analysis. These identifications were also checked against the available scholarly literature on naming practices among the region’s various ethnic groups.8 The results of this procedure were promising. Of the 1,603 names recorded in the Havana Registers, some form of identification was made for 1,197 (74.7 percent). There were some cases, however, in which names were identified as common throughout all parts of Sierra Leone; these names were discarded. In other cases, names were found to be indistinguishable between several ethnic groups that are concentrated within shared or neighboring regions. In the latter situation, the regions were combined for analysis.9 With these adjustments, a total of 987 names from the registers were identified, representing approximately 61 percent of the entire Havana sample. It is important to note that while a name might be identified as, for example, a Mende name among informants, no conclusions are made regarding how that individual recaptive may have self-identified. A name represents just

The Origins of Slaves Leaving the Upper Guinea Coast

one element of a person’s identity; other elements are clearly significant too. The complexity of African identities makes it impossible to suggest that a Mende name necessarily implies a Mende person, although it is of course possible that this may have been the case.10 The ethnolinguistic continuity of a particular region does, however, make it possible to draw conclusions regarding the geographic origins of a recaptive with a name identified as Mende. As Paul Hair concluded in a seminal article in 1967, “If we . . . compare the ethnolinguistic inventory of [Africa] today with that of the period before 1700, we find a striking continuity. In the particulars compared, the ethnolinguistic units of the Guinea Coast have remained very much the same for three, four, or five centuries. . . . This continuity is striking because it contrasts with the impression of wholesale disturbance, and hence discontinuity, given in the standard history texts (under such headings as ‘Slave Trade’ and ‘Imperialism’) or in the oral traditions (where sagas of unrelenting migration are relieved only by lists of rival units examined en route).”11 While the identity of African captives in the slave trade remains a central element in understanding the cultural development of the Atlantic world, it is with this geographic phenomenon that this chapter is mostly concerned. A final point to address is the extent to which the process of enslavement itself might have influenced the names provided by recaptives. More specifically, is it possible that slaves exported from the Sierra Leone region during this period were renamed at some point prior to their liberation? Answering this question is particularly tricky given the lack of evidence concerning slave experiences prior to their disembarkation in the Americas. However, records of the individuals carried on board the schooner Amistad offer a chance to address this issue. The events surrounding the capture of the Amistad are well known and need not be repeated here.12 For this chapter’s purposes, it is relevant only to note that the captives onboard this vessel were embarked at the Gallinas region in southern Sierra Leone in the late 1830s. There is no reason to believe that their collective experiences were markedly different than those of other slaves taken from this region in the nineteenth century. The Africans onboard the Amistad were held in Connecticut beginning in 1839, and over the subsequent year they were periodically visited by Josiah W. Gibbs, a linguistics professor at Yale University. The information Gibbs collected during these visits was used to sketch brief biographies of thirty-five Amistad captives.13 As a linguist, Gibbs was particularly interested in African names, often annotating the meanings attached to the slaves’ names and sometimes even the names of their children. It can be assumed that if any captives

159

160

Philip Misevich

had mentioned taking a new name, Gibbs’s biographies would have included such information. But none of the captives seems to have described such an occurrence. Moreover, the amount of time spent as a slave in Africa seems to have had no impact on a slave’s name. Several captives testified that they had been held in Africa for periods ranging anywhere between two months and ten years before being sent off to Cuba.14 Some even recalled the ethnicity of their owner. In all but a single case, however, the origins of the names that the individuals provided were consistent with their statements concerning their homelands.15 Thus, for example, captives with recognizably Mende names also described their births and points of enslavement as occurring in “the Mendi country.”

THE SLAVE TRADE OF SIERRA LEONE IN THE NINETEENTH CENTURY

The slave trade from Sierra Leone never came close to matching the export figures of the more active slaving centers south of the equator. Until the 1750s, slave departures from the region remained relatively trivial. In this decade, however, departures increased tenfold, from 500 to more than 4,000 annually. This volume was more or less maintained for nearly a century, declining only in the 1840s, despite mounting pressure after 1807 from the British to end the slave trade.16 After abolishing their nation’s slave trade, the British government devoted significant funds to end the trade globally, signing treaties with nations to prohibit the transshipment of African slaves to the Americas and establishing an antislavery squadron that patrolled the West African coast in search of captains breeching such treaties. Although the British initiative did not slow the pace of slave exports from Sierra Leone, it did have significant consequences for the way the trade operated along the coast. Before 1790 the most active embarkation points in this region were Bance Island in the Sierra Leone estuary and the Iles de Los, off the coast of modern Guinea.17 These islands were highly valued because of the protection they provided against indigenous inhabitants in cases of commercial disputes. It is thus not surprising that the Sierra Leone estuary would soon represent the heart of the British thrust to end the region’s slave trade, beginning in 1787 with the settlement of Freetown and continuing with the establishment of the Crown colony there in 1808.18 As the British presence in the Sierra Leone River increased, the protection that the islands once offered to slave traders became a liability. While the fortifications of the island settlements were capable

The Origins of Slaves Leaving the Upper Guinea Coast

of repulsing merchant vessels, they were easily destroyed by the more heavily armed British cruisers. With the British attempt to suppress the slave trade, then, the slave-exporting centers of Sierra Leone slowly moved inland, as traders sought to regroup in regions beyond the reach of vessels’ cannons.19 This trend is easily traceable through the use of the slave-trade database. In the quarter century from 1776 to 1800, nearly one out of every two slaves purchased along the coast of Sierra Leone was embarked at Iles de Los or Bance Island. Over the following twenty-five years, this proportion dropped to less than one in ten slaves.20 The reorganization of the trade from Sierra Leone thus brought new parts of the coast into the slaving mix. This was particularly true for the Rio Pongo, in the northern part of the region, and for the Gallinas, a region reaching about thirty miles from the coast into the interior, bounded by the Mano River on the southeast and stretching just northwest of the Moa River. After 1800 these were increasingly active areas from which slaves were exported.21 A similar shift inland would most likely have occurred from Bance Island to points along the Sierra Leone River, if not for the British settlement of Freetown. Without the strong British influence in this region, an additional export of between 500 to 1,000 slaves annually might reasonably have been expected. The reorganization of the slave trade is underscored by the evidence found in the Havana Registers. Of the 1,603 recaptives documented in the ledger books between 1819 and 1841, just over 85 percent were sold from inland-based ports. Seven of the nine vessels captured embarked their slaves from either the Rio Pongo or the Gallinas. Out of the remaining two, one vessel, the Fingal, purchased its slaves from Grand Cape Mount and the other, the Jesus Maria, from Sherbro Island.22 Because all nine vessels were captured in the Caribbean, the sample is presumably unbiased. Put another way, there is no reason to believe that a British cruiser would have been more likely to catch a slave trader coming from one part of Sierra Leone over another. Overall, the new security that the shallow creeks in the Gallinas and the Rio Pongo provided against British naval cruisers seems to have compensated for any disruption caused by the relocation from the islands to the African mainland. In fact, taken at five-year intervals, the highpoint of the slave trade from Sierra Leone seems to have been the period 1816–20, when an estimated 28,100 captives were exported. The 1820s witnessed an estimated 38,800 slaves departing from Sierra Leone, climbing to nearly 40,000 in the 1830s. From there, exports dropped to just over 22,000 in the following decade before ceasing completely in the mid-1850s.23

161

162

Philip Misevich

Between 1824 and 1841—the period during which each of the nine vessels taken to the Havana Court of Mixed Commission was captured—an estimated 67,520 slaves were exported from Sierra Leone. The 1,603 names recorded in the Havana Registers thus represent approximately 2.4 percent of the entire slave trade from Sierra Leone for the period. It is important to remember, however, that each of these vessels was intending to disembark slaves in Cuba. Indeed, Cuba was the intended destination for the majority of vessels that purchased slaves along the coast of Sierra Leone in the second quarter of the nineteenth century. For such voyages in which a region of disembarkation is known, it represented the point of sale for nearly eight out of every ten captives who were lucky enough to survive the Middle Passage.24 Using information from the new data set, the total volume of the slave trade from Sierra Leone to Cuba between 1824 and 1841 is estimated at 53,009. Returning to the registers, we can thus conclude that just over 3 percent of the volume of the slave trade from Sierra Leone to Cuba is represented in the sample for the period under consideration.

COUNTRY OF ORIGIN

Throughout the period that the Havana Registers were kept, a separate column in the ledger book recorded the “country of origin” for each recaptive. A total of forty-one different labels were provided by Spanish officials to describe the origins of each of the 1,603 liberated slaves, using either one- or two-term identifications. Such descriptions must be used with care; it is well known that Europeans possessed only a questionable understanding of the African interior.25 But to discard the information they recorded based on this point would ignore the potentially rich insight it can offer into what “country of origin” may have meant to a colonial official. At first glance, the labels do not appear to have been very systematic. Some have ethnic connotations and can be recognized among the groups found in Sierra Leone today. Thus “Quisi” or “Cranco” can be easily identified as the modern-day ethnic groups Kissi and Koronko. Others are less obvious, as in the cases of “Ganga” and “Longová.” Labels were also combined in a variety of ways, such as for those recaptives called “Mandinga, Temene.” The key to this puzzle lies in a careful analysis of such two-part identifications. Particularly interesting is the prevalence of the label “Mandinga, Mandinga,” given to 81 of the 1,603 liberated Africans, or approximately 5 percent of the total. Tempting as it may be to identify these recaptives as Mandingo, an ethnic

The Origins of Slaves Leaving the Upper Guinea Coast

group found throughout large parts of West Africa, such a presumption must be avoided. Indeed, among those for whom this label was given, only 11 percent have been identified by my informants as having Mandingo names. What these recaptives did have in common, however, is that they were all exported from the Rio Pongo. This suggests that the identifications made by colonial officials regarding country of origin were geographically based. With this in mind, the column begins to make more sense. In fact, from a geographic perspective, the colonial officials were remarkably consistent in their recording of recaptive origins. The most common label given was the single term “Mandinga.” This was provided as the country of origin for 474 recaptives who, after removing five names for which no details were provided on origins, represented almost 30 percent of the entire sample. In each of these cases, the “Mandinga” slave was purchased from the Rio Pongo region. Furthermore, if we add to this sample those recaptives whose country of origin was described in two parts, and in which the first was “Mandinga” (as, for example, with “Mandinga, Cranco”), the count is increased to no less than 882, all still exported from the Rio Pongo. Clearly, then, in those cases with two-part descriptions, the first label had a geographic (and not ethnic) significance. Mandinga, in this case, was a way to indicate that a particular recaptive departed from the Rio Pongo. It would be a mistake to assume that a particular label given for the country-of-origin column was used to describe all slaves from the same port. “Ganga,” the second most frequent origin given, appears 465 times in the registers. For these recaptives, the point of embarkation was split almost equally between the Gallinas and Sherbro Island. Both of these slave-trading centers were located in the southern part of Sierra Leone, within a short distance of each other. “Ganga” can thus have been intended to identify recaptives that were shipped from one of the ports in southern Sierra Leone. Indeed, among all of the 662 slaves coming from these southern ports—including the Gallinas, Sherbro Island, and Grand Cape Mount—643 (97.1 percent) were labeled “Ganga.”26 Based on this evidence, it seems safe to conclude that, in general, the country-of-origin category in the Havana Registers described the geographic region from which a recaptive was exported. In one way or another, 98 percent of the recaptives provided with a country of origin in the registers were labeled as either “Mandinga” or “Ganga.”27 This still leaves unanswered questions about names for which the country of origin appears to have suggested an ethnicity in its second part. For now these questions must remain unresolved. Although recaptives’ names that

163

164

Philip Misevich

were labeled “Cono” in the country-of-origin column were at times identified as Kono names by my informants, the pattern was far from conclusive.

GEOGRAPHIC ORIGINS OF RECAPTIVES

By providing a linguistic identification of recaptive names from the Havana Registers and pinpointing the geographic regions in which they are currently used, an estimation of the point of enslavement for these recaptives is possible. Such a procedure draws heavily on the work of linguists with knowledge of African names and naming practices. As several Africanist linguists have demonstrated, diversity among the languages of Sierra Leone is considerable. Broadly speaking, most languages have been grouped under the Mande or Mel language families. The former represents the most widespread family of languages in Sierra Leone and can be divided into northwestern Mande and southwestern Mande. Languages classified as northwestern Mande include Susu, Yalunka, Kono, Vai, and Koronko. Southwestern Mande languages include Mende and Loko. Of course, many Mande languages are spoken outside of Sierra Leone and indeed throughout West Africa.28 Mel languages can also be subdivided based on their relationships to one another. According to the standard breakdown, northern Mel languages include Temne, Bullom, and Kissi, while Gola is the lone southern Mel language spoken in Sierra Leone. Bullom has been even further divided, with those dialects spoken south of the Sierra Leone estuary referred to as Sherbro, while north of the estuary the name Bullom is kept. Although this linguistic breakdown is useful in providing a broad classification of Sierra Leone’s languages, it should be remembered that the relationships among the various spoken languages are not always clear and that, even within the Mande or Mel families themselves, not all languages are mutually intelligible.29 The diversity of this region’s languages was underscored by my attempt to linguistically identify the names of African recaptives from the registers. Ninety different ethnolinguistic combinations were used to identify a list of 1,603 names. But this overall picture was misleading. After removing from the sample those names that are commonly used throughout all parts of Sierra Leone, and combining several regions in which names were found to overlap, the labels applied were considerably fewer. Table 5.1 shows the various designations used to identify all recaptives that were exported from Sierra Leone after making these modifications, including those embarked from the Rio Pongo, the Gallinas, Sherbro Island, and Grand Cape Mount.

The Origins of Slaves Leaving the Upper Guinea Coast

Table 5.1 Linguistic Identifications of Sierra Leone Recaptives from the Havana Registers, 1824–41 Ethnolinguistic Unit Mende Koronko/Kissi/Kono Mandingo Susu/Yalunka Temne Fula Limba Loma From outside region Minor groups Sherbro Total

Number of Recaptives

Column Percentage

188 171 130 98 96 92 55 51 40 34 32 987

19.0 17.3 13.2 9.9 9.7 9.3 5.6 5.2 4.1 3.4 3.2 100.0

Source: Copies of Havana Registers of Liberated Africans, held in BNA, FO 313, vols. 56–62.

What is immediately clear from the table is the high concentration of recaptive names among a small group of classificatory units. Taken together, nearly 50 percent of these names were identified among the top three designations. Mende names make up nearly one out of every five on the list. The remaining half are spread among eight additional groupings. For those recaptives originating from outside this region, there was a heavy bias in favor of the hinterland of Guinea: half of these forty names were identified as Gbese, and an additional nine were recognized as Konyaka.30 Designations with less than twenty applications, in this case representing Vai and Loko, were placed in the “minor groups” category.31 What is equally apparent is the lack of geographic concentration in this table. The distance between the Mende and Mandingo groups, which make up more than 30 percent of the names identified, was quite considerable (see map 5.1). It should be remembered, however, that this sample includes slaves shipped from several different ports lying along a lengthy stretch of coast. Tables 5.2 and 5.3 thus provide further breakdowns of the sample, based on the location of the port of embarkation along the coast of Sierra Leone. Table 5.2 summarizes the identification of slaves exported from the Rio Pongo, in the northern part of Sierra Leone. It shows a near halving of the proportion of Mende slaves being exported when compared with table 5.1.

165

166

Philip Misevich

Table 5.2 Linguistic Identification of Havana Recaptives Leaving from the Rio Pongo, 1824–41 Ethnolinguistic Unit Koronko/Kissi/Kono Mandingo Susu/Yalunka Temne Mende Fula Limba Outside region Loma Total

Number of Recaptives

Column Percentage

125 94 70 65 61 59 41 28 27 570

21.9 16.5 12.3 11.4 10.7 10.4 7.2 4.9 4.7 100.0

Source: Same as table 5.1.

The combined grouping of Koronko/Kissi/Kono names represents the most common identification for recaptives, followed by Mandingo and the combined unit of Susu/Yalunka. Alexander Gordon Laing’s observations among the Kissi-speaking peoples substantiate the high proportion of Kissi names among liberated Africans in Havana. Passing through this region in the mid1820s, Laing commented on the deleterious impact of the slave trade, noting that “the people of Kissi have no trade except in slaves, which they sell to the people of Sangara for salt, tobacco, and country cloth; and, in such a savage state of wretchedness and barbarism are they, that without the least compunction they will dispose of their relatives, wives, and even children.”32 Given the abolitionist sentiments of the period, it is possible that British travelers would have applied such comments to many parts of the African interior, and they often did. The presence of liberated Kissi slaves in Freetown was large enough, however, to justify the establishment of a separate Kissi village, made up of “natives of the district of Kissy, lying between Falaba and the sources of the Niger . . . having been captured from slave ships by British men-of-war, it was considered desirable to locate them in one place.”33 The slave trade from the Rio Pongo also drew on regions beyond the Koronko/ Kissi/Kono-speaking interior. The twenty-eight names identified in table 5.2 as originating from an outside region were almost all found to have come from what would today be located in Guinea, as might well be expected given the proximity of the Rio Pongo to that country.

The Origins of Slaves Leaving the Upper Guinea Coast

Table 5.3 Linguistic Identification of All Havana Recaptives Leaving from the Gallinas, Sherbro Island, and Grand Cape Mount, 1824–41 Ethnolinguistic Unit Mende Minor groups Mandingo Koronko/Kissi/Kono Fula Temne Loma Susu/Yalunka Total

Number of Recaptives

Column Percentage

127 52 36 34 33 31 24 20 357

35.6 14.6 10.1 9.5 9.2 8.7 6.7 5.6 100.0

Source: Same as table 5.1.

Table 5.3 provides a look at the interior origins of captives exported from ports located in the southern part of Sierra Leone, combining the Gallinas, Sherbro Island, and Grand Cape Mount. Broken down this way, a total of 357 recaptives were identified by ten different ethnic combinations. For this region, Sherbro, Vai, Limba, Loko, and Gola constitute the “minor groups” category, since such identifications were each made for less than twenty names. The prevalence of Mende names among this sample is striking with more than one in every three being recognized as common among Mende people today. Also notable is the relative decline in exports of the Koronko/Kissi/Kono grouping, dropping from more than 20 percent of the recaptives exported from the Rio Pongo (table 5.2) to less than half that amount coming from ports in the south. The predominance of slaves with Mende names in southern Sierra Leone provides an opportunity to cross-check the data from the sample with contemporary observations made by individuals who had some knowledge of this region’s slave trade. In the 1840s, as the British campaign to suppress the slave trade became increasingly expensive for the government to maintain, abolitionists began rethinking their approach to achieving their central objective. They conducted interviews with hundreds of individuals with experiences in the transatlantic trade to assess the effectiveness of their previous measures. Interviewees included former slave traders, merchants engaging in agricultural commerce, missionaries, and even a handful of liberated Africans. In one detailed interview, the Reverend James Frederick Schön, who resided at Freetown

167

168

Philip Misevich

from 1832 to 1848, recounted a trip he had taken to Shebar, a center of the slave trade on Sherbro Island. While there, the reverend witnessed a small cargo of slaves being sold to one of the island’s principal slaving merchants. Schön claimed that all the slaves were of the “Cossoo” (Mende) nation. Although he was surprised by the low price paid for the captives, the slave dealer explained that “they were cheap, but that they were Cossoos, and that Cossoos were mere cattle, and more should not be paid for them.”34 The high ratio of Mende captives was also noted by Try Norman, a Yoruba recaptive who was enslaved during a trip to collect a debt from a man named Lewis, who resided in the Gallinas. Before she could collect her money, Norman was taken prisoner by Prince Manna, the most powerful African slave dealer in all of the Gallinas at the time, on account of a debt Norman’s mistress was said to have incurred. Norman was held in a slave barracoon for three months before being liberated by a British naval officer. In her interview with the British committee, Norman claimed that the slaves with whom she was held at the barracoons in the Gallinas were all from the Cossoo country.35 When the names data are combined with observations by contemporary observers, the high proportion of Mende slaves coming from southern Sierra Leone is undeniable. In fact, among the captives embarked at the Gallinas who were captured onboard the Amistad, an even higher proportion of Mende slaves were found. Based on their own testimonies, self-identifications have been provided for twenty-two of these individuals. Of the twenty-two, fourteen (64 percent) were born in the Mende-speaking interior.36 Such high proportions of Mende names were much less common among recaptives from the Rio Pongo. Taken together, tables 5.1 to 5.3 reveal interesting similarities and differences between the origins of recaptives exported from the northern part of Sierra Leone and those who embarked at southern ports. The top two designations from table 5.1 provide an opportunity to contrast the two regions. Among these, Mende, the most common identification, provides the most dramatic case. In the south, more than one-third of the recaptives had recognizably Mende names. Moving north to the Rio Pongo, this was the case for less than one in ten. A drastic change in the names identified under the combined grouping of Koronko/Kissi/Kono is also apparent, but this time in the opposite direction: 21.9 percent of the recaptives departing from the Rio Pongo were acknowledged to have a name so identified, while in the south this grouping represented less than half of that percentage. Geography inevitably accounts for some of the differences between regions. The region in which Mende was spoken was no further than fifty or sixty

The Origins of Slaves Leaving the Upper Guinea Coast

miles from the coastline. It would have been easier and less costly to transport a slave from this region to the southern ports of the Gallinas, Sherbro Island, or Grand Cape Mount than to take the journey 150 miles north to the Rio Pongo. Similarly, it is not surprising that recaptives with identifiably Susu or Yalunka names were more than twice as likely to end up being sold from the Rio Pongo (see map 5.1).37 Given the significance of the relationship between the point of enslavement for African slaves in the interior and the Atlantic ports to which they were transported for sale to ship captains, tables 5.4 and 5.5 present the data differently, in terms of the distance from the estimated region of enslavement to the point of embarkation. The regions of enslavement are based on rough calculations of the central point of each linguistic unit that has been used to identify a recaptive’s name.38 Indeed, it is impossible to know how many recaptives would have started their journeys to the coast from the chosen central point; few probably did. Because of the difficulties in making such a determination with accuracy, the figures are presented in increments of 50 miles, combining all those regions beyond 150 miles from the coast in the final column. Table 5.4 shows the distances from points of origin for recaptives exported from the Rio Pongo. Given the above presumptions, it might be expected that the majority of the sample would come from within fifty miles of the coast. But this does not appear to have been the case. There is no substantial bias among any of the particular increments. On average it appears that, for this region, the journey from the interior to the coast was above 100 miles. The homelands of the majority of recaptives from these four categories were between

Table 5.4 Groupings of Average Distances between Regions of Enslavement and Embarkation Points for Recaptives Embarked at the Rio Pongo, 1824–41 Distance Less than 50 miles Between 50 and 100 miles Between 100 and 150 miles More than 150 miles Total

Number of Recaptives

Column Percentage

129 106 186 121 542

23.8 19.6 34.3 22.3 100.0

Source: Same as table 5.1. Note: The total in col. 1 is smaller than the total sample for table 5.2 because of the elimination of recaptives from outside Sierra Leone.

169

170

Philip Misevich

100 to 150 miles from the point of embarkation, although the crudeness with which this distance was calculated leaves open the possibility of a significant margin for error. However, these findings are broadly consistent with Barry Higman’s sample of the origins of British West Indian slaves embarked from Upper Guinea, in which an estimated 44 percent came from within 100 miles of the shoreline although, as the tables 5.4 and 5.5 show, there could be significant differences between ports.39 Comparing the Rio Pongo results with the distances traveled by slaves from southern Sierra Leone ports leads to some noteworthy contrasts between the north and south. As table 5.5 demonstrates, among those slaves departing from southern Sierra Leone ports, the majority had much closer homelands. Fully 44 percent of the recaptives lived within fifty miles of the coast. The case of James Campbell, a liberated Mende slave, can serve as an example of the experiences that many southern Sierra Leone slaves would have undergone. Campbell was enslaved during a war in the interior and taken to a barracoon on the coast by a native chief, a trip which took just three days. While he recalled enduring many hardships during his forced migration to the coast, including regular whippings and a chain kept tightly around the neck, his health remained good throughout the experience.40 This no doubt reflected the close proximity of the Mendespeaking hinterland to the southern Sierra Leone coast. While similar experiences of enslavement would have been encountered by other Mende slaves, table 5.5 also shows that this was not the case for all recaptives in the region. About 50 percent seem to have originated from beyond 100 miles of their Atlantic points of embarkation, meaning that one of every two recaptives traveled more than 100 miles to reach the port from which they were taken to the New World. Table 5.5 Groupings of Average Distances between Regions of Enslavement and Embarkation Points for Recaptives Embarked in Southern Sierra Leone, 1824–41 Distance Less than 50 miles Between 50 and 100 miles Between 100 and 150 miles More than 150 miles Total Source: Same as table 5.1.

Number of Recaptives

Column Percentage

157 31 86 83 357

44.0 8.7 24.1 23.2 100.0

The Origins of Slaves Leaving the Upper Guinea Coast

CONCLUSIONS

Perhaps the most striking feature that emerges from these new data is the similarities between the two regions of embarkation. Taking the top six out of the eight designations provided in table 5.1, we find significant correspondence in tables 5.2 and 5.3. Put another way, the six most common identifications provided for recaptives in the entire sample were statistically significant both in the northern and southern Sierra Leone ports. This is not to suggest that they were equally represented. As we have seen, Mende slaves were much more likely to be exported from the Gallinas, Sherbro Island, or Grand Cape Mount than from the Rio Pongo. But it is nonetheless notable that sixty-one slaves with Mende names were transported to the Rio Pongo for export—enough to represent more than 10 percent of the sample in table 5.2. In sum, the provenance of the slave trade that drew on Upper Guinea between 1824 and 1841 was extremely diverse. This is a diametrically opposing pattern to the hinterland of the Cameroons during a comparable period, when interior supply sources were found to have been highly concentrated.41 The fact that such diverse inland sources fueled the Sierra Leone slave trade raises some intriguing questions about the organization of the trade in this region, particularly concerning the relationship between interior regions that supplied slaves and the points of embarkation along the littoral, where they were sold. One explanation suggested by these data is that the various ports along the coast represented different markets for the same sources of supply. This would not have been the first region in which slave traders demonstrated remarkable flexibility in their ability to transport slaves to a variety of places along the coast, seeking ultimately the most advantageous prices for their human cargo. Philip Curtin documented a similar process among traders in Senegambia throughout the eighteenth century.42 But a major difference between these two regions is the much greater distance between the ports traveled by African slave merchants in the Sierra Leone slave trade. The northern Rio Pongo and the southern ports are separated by more than 200 miles of coastline. Still, based on studies of other regions involved in the slave trade, the conclusion remains tempting and not unlikely. In West Central Africa, the slave trade drew on peoples from as far as 700 miles beyond the coast. Although this was the most extreme case, the shifting slaving frontiers in other African regions also demonstrate that transporting captives for 200 miles was not out of the realm of possibilities, even if it was not perhaps a merchant’s first

171

172

Philip Misevich

choice.43 At this point, however, the best that this research can hope to do is to open the issue for debate. Most important, however, this work demonstrates the potential that culturally significant evidence such as African names can have for shedding further light on the African end of the slave trade. NOTES

1. For the operational aspect of the courts, see Leslie M. Bethell, “The Mixed Commissions for the Suppression of the Transatlantic Slave Trade in the Nineteenth Century,” Journal of African History 7 (1966): 79–93. 2. The original Sierra Leone Registers are held in the Sierra Leone Government Archives at Fourah Bay College in Freetown. Copies can be found scattered through the FO 84 series of the BNA. The Havana Registers are housed in BNA, FO 313, vols. 56–62. 3. The courts of mixed commission presided over more than 600 cases in total. Occasionally ships captured with slaves onboard were acquitted. At the Sierra Leone court, 29 of the 485 vessels under adjudication were cleared of charges. For the Havana court, 7 out of 48 were released. 4. Additional information for each vessel can be found in TSTD2. The database is available via an open-access Web site at http://www.slavevoyages.org. The ship names, followed by their identity numbers in the data set, are as follows: La Isabel (2366), Fingal (558), Gallito (941), Josefa alias Fortuna (770), Carlota (1338), Chubasco (1368), Preciosa (1479), Jesus Maria (2071), and Segunda Rosario (2078). 5. Ugo Nwokeji and David Eltis, “Characteristics of Captives Leaving the Cameroons for the Americas, 1822–37,” Journal of African History 43 (2002): 191–210. The methods used in this article are further explained in Ugo Nwokeji and David 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. Adam Jones hinted at the potential for such a project, finding that although “these registers do not record the slaves’ tribes . . . it should be possible to draw conclusions from their names alone.” See Adam Jones, From Slaves to Palm Kernels (Wiesbaden, 1983), 54. A similar conclusion was reached by Richard Meyer-Heiselberg, but with a more skeptical tone. He suggested that “certain characteristic names or eventually tattooings may yield some information, but still this cannot be regarded as sure means of establishing the [recaptives’] origin.” Quoted in Richard Meyer-Heiselberg, Notes from the Liberated African Department in the Archives at Fourah Bay College, Freetown, Sierra Leone (Uppsala, 1967), iii. A pioneering effort to link names with specific African ethnic groups was made by Lorenzo Turner, Africanisms in the Gullah Dialect (Chicago, 1949). 7. The informants included representatives from the Susu, Yalunka, Koronko, Mandingo, and Temne ethnic groups and a Mende linguist from Njala University College, the University of Sierra Leone. The author extends his gratitude to all the informants, and particularly to Taziff Koroma, for their assistance with this project. The author’s debt to Oscar Grandío Moráguez for recording the pronunciation of each recaptive name is also gratefully acknowledged.

The Origins of Slaves Leaving the Upper Guinea Coast

8. There is an overall paucity of research on naming practices among Africans in Upper Guinea. Numerous short entries began to appear in the late 1910s in the journal Sierra Leone Studies, recording a handful of names used by a particular ethnic group in Sierra Leone. This continued with the publication of the Sierra Leone Language Review, with the added benefit of more rigorous analysis by the contributors. For one example, see Gordon Innes, “A Note on Mende and Kono Personal Names,” Sierra Leone Language Review 5 (1966): 34–38. 9. The overlapping of names used among Susu and Yalunka peoples makes it impossible to differentiate one group from the other. This was also the case with names identified as Koronko, Kissi, or Kono. While this situation makes conclusions regarding ethnic origins difficult, it does not hinder efforts to identify the geographic regions in which the names are used. Susu and Yalunka peoples are heavily concentrated in the northern part of Sierra Leone, occupying much of the same territory. Koronko, Kissi, and Kono peoples are settled in the eastern region of Sierra Leone. Although the Koronko, Kissi, and Kono languages are not all closely related to one another, these ethnic groups nonetheless share numerous personal names in modern Sierra Leone. 10. The problems with making definitive claims regarding ethnicity and identity in the African diaspora have recently received increasing attention. See David Northrup, “Igbo and Myth Igbo: Culture and Identity in the Atlantic World, 1600–1850,” Slavery and Abolition 21 (2000): 1–20, and Peter Caron, “ ‘Of a Nation Which the Others do not Understand’: Bambara Slaves and African Ethnicity in Colonial Louisiana, 1718–60,” Slavery and Abolition 18 (1997): 98–121. For a treatment that stresses the fluidity of ethnicity along the Upper Guinea coast, see Allen M. Howard, “Mande and Fulbe Interaction and Identity in Northwestern Sierra Leone, Late Eighteenth through Early Twentieth Centuries,” Mande Studies 1 (1999): 13–39. For a broader theoretical critique of the analytical value of the term “identity,” see Frederick Cooper and Rogers Brubaker, “Identity,” in Cooper, Colonialism in Question: Theory, Knowledge, History (Berkeley, 2005), 59–90. 11. P. E. H. Hair, “Ethnolinguistic Continuity on the Guinea Coast,” Journal of African History 8 (1967): 247. An additional linguistic breakdown that lends further support to the stability of the region’s ethnolinguistic distribution can be found in David Dalby’s contribution, “Languages,” in John Innes Clarke, Sierra Leone in Maps (London, 1966), 15. 12. For a recent Sierra Leonean perspective on the historical and contemporary relevance of the story, see Iyunolu Folayan Osagie, The Amistad Revolt: Memory, Slavery, and the Politics of Identity in the United States and Sierra Leone (Athens, 2003). 13. Printed in John Warner Barber, A History of the Amistad Captives (New Haven, 1840), 9–15. 14. It is impossible to know the length of time that Africans later liberated by the Havana mixed court remained in Africa prior to being taken onboard a slave ship. Using evidence that S. W. Koelle gathered from 179 liberated Africans in Freetown around 1850, P. E. H. Hair notes that “few of the informants had spent much time as slaves . . . twenty-nine had spent periods of years in Africa, mainly as slaves to Africans. The remainder had reached Sierra Leone shortly after enslavement . . . that is, they were enslaved in their home district and immediately taken down to the coast . . . and were

173

174

Philip Misevich

shortly afterwards captured aboard a slave ship and brought straightway to Freetown.” These figures, however, included slaves drawn from many different parts of West, West Central, and even Southeast Africa. See P. E. H. Hair, “The Enslavement of Koelle’s Informants,” Journal of African History 6 (1965): 195. 15. The exception was Kwong, who was said to have been born at Mambui, “a town in the Mendi country.” Kwong, however, was identified as a Bullom (Sherbro) name. See Barber, Amistad Captives, 11. 16. David Eltis, David Richardson, and Stephen Behrendt, “The Transatlantic Slave Trade from Africa,” unpublished paper, 2000. The figures can be derived by using TSTD1. 17. A recent treatment of Bance Island within an Atlantic context is David Hancock, Citizens of the World: London Merchants and the Integration of the British Atlantic Community, 1735–1785 (New York, 1995), chap. 6. The role of the Iles de Los in the slave trade is detailed in Bruce Mouser, “Iles de Los as Bulking Center in the Slave Trade, 1750–1800,” Revue Française d’Histoire d’Outre-Mer 83 (1996): 77–90. 18. Although much research has been undertaken regarding the settlement of Sierra Leone, the standard work remains Christopher Fyfe, A History of Sierra Leone (Oxford, 1962). For a different emphasis on the factors accounting for the placement of the antislavery settlement, see P. E. H. Hair, “Aspects of the Prehistory of Freetown and Creoledom,” History in Africa 25 (1998): 111–18. 19. David Eltis, Economic Growth and the Ending of the Transatlantic Slave Trade (New York, 1987), 165–66, 181–83. 20. Calculated from TSTD2. 21. For the Gallinas, see Jones, From Slaves to Palm Kernels. The involvement of the Rio Pongo in the slave trade is described in Bruce Mouser, “Trade and Politics in the Nunez and Pongo Rivers, 1790–1865” (PhD diss., Indiana University, 1971), and in idem, “Trade, Coasters and Conflict in the Rio Pongo from 1790 to 1808,” Journal of African History 14 (1973): 45–64. 22. Although Grand Cape Mount is traditionally considered part of the Windward Coast in slave-trade studies, here it has been included within Sierra Leone. With only fiftyeight slaves onboard this vessel, it unlikely to have had a significant impact on the overall export figures. 23. The figures presented here are dramatically changed from those given by TSTD1. These numbers are based on TSTD2’s new version of the 1999 data set. 24. Calculated from TSTD2. 25. See in particular Leroy Vail, ed., The Creation of Tribalism in Southern Africa (Berkeley, 1989). The potential consequences of Europe’s misunderstanding of African ethnicity have been described in Mahmood Mamdani, When Victims Become Killers: Colonialism, Nativism, and the Genocide in Rwanda (Princeton, 1999). 26. “Buche” was the country of origin given for the remaining recaptives in this sample. It should be noted that all nineteen slaves identified by this term were exported from Grand Cape Mount. 27. This proportion includes any two-part identification beginning with either of these two labels, as well as those cases using “Ganga” and “Mandinga” singly.

The Origins of Slaves Leaving the Upper Guinea Coast

28. This point underscores the need, in this case, to work with regions undefined by colonial boundaries. Claiming that ethnolinguistic units were stable over time should not suggest that colonial powers took them into account during their occupation of Africa. Though the following statistics are based on a language survey of postindependence Sierra Leone, the wider analysis of the geographic origins of recaptives exported from Sierra Leone transcends the definition of the modern nation, as shown in map 5.1. 29. Dalby, “Languages.” While Dalby claims that the distribution of several of Sierra Leone’s languages is changing, this is a more recent phenomenon and would not have had a significant effect on the linguistic patterns in the early to mid-nineteenth century. 30. The Mandingo informant for this project was particularly knowledgeable concerning names that were common outside Sierra Leone, having spent a considerable period living in modern Guinea. 31. It is quite possible that there is overlap between Mende and Vai names. These groups were not combined, however, because it was determined that many Mende and Vai names could be recognized independently from each other. 32. Alexander Gordon Laing, Travels in the Timanee, Kooranko and Soolima Countries in Western Africa (London, 1825), 280–81. 33. John Joseph Crooks, A History of the Colony of Sierra Leone, Western Africa; with Maps and Appendices (Northbrook, Ill., 1972), 88. 34. Rev. James Frederick Schön to the Select Committee on the Slave Trade, April 11, 1848, in Irish University Press series of British Parliamentary Papers (Shannon, 1967–71), Slave Trade, vol. 4, “First Report,” p. 182 (henceforth Slave Trade). 35. Try Norman to the Select Committee on the Slave Trade, March 28, 1848, in Slave Trade 4:63. 36. Barber, Amistad Captives, 9–15. 37. The various ports from which Sierra Leone recaptives departed make this analysis considerably different from that for the Cameroons slave trade, which embarked slaves from one region. See Nwokeji and Eltis, “Characteristics.” 38. To calculate distance, two methods were employed. First, the political units from map 5.1 were used as the central point from which a measurement to the various ports was made. To account for those regions in which some languages are particularly widespread—where a distance could vary considerably from one part of a region speaking the language to another—the calculation was then double-checked with the linguistic distribution found in Dalby, “Languages.” 39. Barry Higman, Slave Populations of the British Caribbean, 1807–1824 (Baltimore, 1984), 442–57. 40. James Campbell to the Select Committee on the Slave Trade, March 28, 1848, in Slave Trade, 4:78–80. 41. Nwokeji and Eltis, “Characteristics.” 42. Philip Curtin, Economic Change in Precolonial Africa: Senegambia in the Era of the Slave Trade (Madison, Wis., 1975). 43. Joseph C. Miller, Way of Death: Merchant Capitalism and the Angolan Slave Trade, 1730–1830 (Madison, Wis., 1988), 148; for Senegambia, see Curtin, Economic Change, 190.

175

Chapter 6 The African Origins of

Slaves Arriving in Cuba, 1789–1865 Oscar Grandío Moráguez

The legacy of the Atlantic slave trade continues to be one of the strongest influences over the evolution of identity in the Americas. The largest forced migration in human history uprooted more than 12 million Africans from their home communities and scattered them throughout the Americas.1 A lot of information is now available, both about the slave trade and about slave experiences in the Americas. Much less is known, however, about the ethnicity of the African peoples who came to the Americas as slaves, and we know much less again about how these people became captives in Africa, and why thereafter they entered the transatlantic slave trade. While sources and discussions on the ethnicity of slaves within the Americas, by contrast, are relatively abundant, they are not always reliable. Planters, lawyers, colonial authorities, churches, and notaries generated lists and descriptions of slaves—for parish records, property records of individual estates, marriages, baptisms, runaway notices, bills of slave sales, and the like—many of which identified Africanborn slaves by geographic or linguistic criteria or even by imputed nationality. The people who compiled these lists recognized cultural 176

The African Origins of Slaves Arriving in Cuba

variants, but they did not know much about Africa itself, and the typology they used was neither that of contemporaneous Africans nor of present-day ethnographers.2 According to Philip Curtin, two main tendencies were at work in the creation of these records. The first was the European tradition of identifying “nationalities” customarily shipped from a particular African port by the name of the port. The second was the tendency to pick one ethnic or linguistic term to identify a much larger group. This was the case, for example, with “Congo,” originally referring to a subgroup of the great Bantu family. The term meant Bakongo in the early sixteenth century, but in nineteenth-century Cuba it came to include all Bantu-speaking peoples and perhaps all slaves shipped from West Central Africa. A similar pattern was apparent with the frequently used term “Mondongo,” equivalent to present-day Mondonga, which could refer to people coming from the interior, roughly to the north and east of the mouth of the Congo River. Similarly, the term “Angola” referred vaguely in the Americas to the region south and east of the Congo.3 At the African end of the slave trade, with the exception of Koelle’s survey and the Sierra Leone census of 1848, the information on ethnicities is meager, consisting mainly of reports made by Europeans. Attempts to establish African identities that were not influenced by foreigners are almost nonexistent.4 In summary, the historiography concerning which groups of people were captured and sent to the Americas is not always clear. In recent years, scholarship on the slave trade has turned away from elaborating and revising Curtin’s figures—although this continues as well—to focusing on tracing the African origins and American destinations of slaves.5 As a result of collaborative efforts to construct reliable data sets, scholars can now more easily avoid using the generic nondescriptive terms “Africa” and “Africans” and can identify more precisely the origins of slaves and their New World destinations.6 In the particular case of Cuba, the debate on the origin of the Afro-Cuban population is almost a century old and still continues.7 Many historians have tried to trace connections between specific homelands in Africa and particular places in Cuba with varying levels of success, and some broad ethnic terms have been studied more extensively than others. A comprehensive study of the different ethnic groups that comprised the African component of the Cuban nation has yet to appear, however. In the nineteenth century, authors such as Esteban Pichardo, José María de la Torre, and José Miguel Macías made a few attempts to inquire into the origins of African slaves in Cuba, although the results were quite general and did

177

178

Oscar Grandío Moráguez

not provide many clues on the subject.8 Pichardo’s work was the best of these early efforts. Although his book focused on the study of the Cuban linguistic tradition and devoted many pages to the influence of African words on the Cuban Spanish spoken during the nineteenth century, he made an attempt to summarize what he considered the main characteristic of each African “nation” in Cuba: “Congos, although well appreciated because of their loyalty, are lazy”; the Lucumies “are really appreciated for being hard workers . . . although have a tendency to commit suicide”; the Carabalies “have a rebel character [but] are good workers”; while the Araras “are easy to identify because all have more facial scarifications than the rest of the Africans.”9 At the end of the nineteenth century, a French anthropologist, Henri Dumont, wrote the first study devoted exclusively to discussing the different African “nations” among the Cuban blacks. Dumont’s study, not published until 1915 and racist in character, basically underlined the stereotypes prevalent among Cuban whites on the main characteristics of the so-called “African nations” in Cuba, and it did not provide much hard evidence on ethnicity or cultural traits. According to Dumont, each “African nation” had intrinsic characteristics that made its members different from the rest. One example was the Lucumí, who were characterized by their obedience, “marked by a certain inclination to suicide,” while the Congo, “strong, but timid and extravagant, had a strong inclination to rest in excess and insubordination.”10 The beginning of the twentieth century brought a new generation of academics and more serious and provocative scholarship on slaves’ ethnic identities. These scholars’ contributions, rooted in an anthropological approach, were better substantiated, more carefully researched, and richer in sources. Manuel Pérez Beato’s article, “Procedencia de los Negros en Cuba,” published in 1910, represented a major advance and was the first attempt to enumerate the different African ethnic groups among the slaves in Cuba.11 A few years later, Fernando Ortiz published an article that came to be seen as a path-breaking study on the subject. Ortiz reviewed ten of the ethnic denominations of African slaves in Cuba and linked them to a precise geographical location in Africa. “Here,” wrote Ortiz, “I could not list all the diverse provenances of the black cargoes; besides, to list all of them is impossible, only limiting myself to indicate the names of the African regions which I have seen quoted at different works from Cuban authors and old documents, all consulted by me, and which provide some clarifications and observations regarding their location in Africa, commonly unknown by many of us, because nobody before me has dealt publicly with this basic and interesting issue for Cuban anthropology.”

The African Origins of Slaves Arriving in Cuba

Ortiz even mentioned that in colonial Cuba the purchase of an African slave (a negro de nación) was highly influenced by his or her origin: “For a purchaser, it was not the same thing, at the psychological level, a Lucumí (Yoruba), a Congo or a Mandinga.” However, Ortiz followed Dumont’s path in characterizing certain African ethnic groups as superior to others. For him, the Yoruba (or Lucumí) were the most intelligent slaves, and the ones more easy to civilize, while Congo slaves were simply “inferior.”12 During the 1920s and until his death in 1968, Ortiz pursued other impressive and voluminous studies on the Afro-Cuban population. He generally repeated his thesis that, of all the Africans arriving in Cuba, the Yoruba came in larger numbers and brought a superior civilization; they were not only the most civilized of Africans, they also shared a superior religious system. Following Ortiz’s lead, one of his disciples, the folklorist Lydia Cabrera, expanded the research on Afro-Cuban cultural and religious traditions, producing several studies during the 1950s and 1960s on the Yoruba and Congo religions (Regla de Ocha and Palo Monte) and on the Carabalí brotherhoods (Abakuá).13 Rómulo Lachatañeré, a Cuban ethnologist, also carried out research from the late 1930s to the late 1950s, publishing numerous essays and monographs on Yoruba, Congo, and Carabalí influences on the Afro-Cubans’ religious systems. The triumph of the Cuban revolution in 1959 marked a watershed for AfroCuban studies. Fidel Castro’s Afrocentric foreign policy fostered additional ethnographic research on the Afro-Cuban population. Studies of Afro-Cuban religions proliferated in the 1960s and 1970s, led by scholars such as Rogelio Martínez Furé and Miguel Barnet.14 A major new wave of scholarship occurred in the 1980s with Enrique Sosa’s major work, Los ñáñigos, focusing on a male brotherhood comprising people of mainly Carabalí descent. For Sosa, the arrival in Cuba of large numbers of slaves from the Calabar coast produced a lasting impact on the customs, folklore, popular language, and traditions of the island, a contribution exemplified by Cuba’s Abakuá secret society, which appears to have been a legacy of the Egbo society of the Ekoi and Efik.15 Sosa’s work was closely followed by Rafael López Valdés’s 1985 study of the ethnic background of African slaves in Cuba. Valdés did not support his research with hard data, limiting himself to a synthesis that drew heavily on Ortiz’s work and restated the importance of the Yoruba component in the Afro-Cuban cultural ensemble.16 But the scholarship also began changing direction in the 1980s. Alejandro de la Fuente, in an article published in 1986, analyzed the different ethnic groups among the African slaves imported to Cuba between the sixteenth and seventeenth centuries.17 De la Fuente, in fact, opened up a new path in the

179

180

Oscar Grandío Moráguez

field by focusing his research on the Africans’ regions of provenance, using data collected at the Cuban National Archives and at the historical archive of Havana’s cathedral. His findings were impressive: forty-two ethnic groups were described in the article, and his conclusions differed radically from those of Ortiz and Valdés. Up to the seventeenth century, he argued, most African slaves residing in Cuba belonged to three main ethnic groups: Angola (with 233 individuals reported), Arará (184), and Congo (128).18 At about the same time, Nery Gómez Abréu and Manuel Martínez Casanova added weight to these findings based on their widely cited work on African baptismal records from a parish in Placetas, in Central Cuba, for the years 1817–86. They pointed to the strong presence of eight main “African nations”: Congo, with 32.4 percent of all slaves residing in Placetas; Gangá (10.7 percent); Mandinga (6.8 percent); Lucumí (3.8 percent); Ibo (3.4 percent); Guinea (3.04 percent); Carabalí (2.43 percent); and Mina (1.4 percent).19 Applying a similar methodology at the national level, Jesús Guanche Pérez estimated in 1995 the ethnic composition of Africans in Cuba from 1851 to 1860. His pioneering study synthesized a large quantity of data about African ethnic and linguistic groups, including their origins, numbers, gender, and age. As Abréu and Martínez had before him, Guanche worked with parish records, this time from two Cuban regions: Pinar del Río and Sancti Spiritus. His results showed nine principal ethnic denominations among African slaves, with the Congo comprising the largest group (see table 6.1). However,

Table 6.1 Ethnic Composition of the African Population in Cuba, 1851–60 Denomination Congo Lucumí Gangá Carabalí Macuá Mandinga Mina Arará Other

Percentage 34.81 22.83 13.22 9.53 4.45 3.81 1.60 1.31 8.44

Source: Jesús Guanche Peréz, “Contribución al estudio del poblamiento africano en Cuba,” Revista do Centro de Estudos Africanos 18–19 (1995–96): 132.

The African Origins of Slaves Arriving in Cuba

Guanche surprisingly concluded that, although Congo slaves were more numerous that any other “African nation” in nineteenth-century Cuba, the Lucumí nation ended up having the biggest influence among the Cuban slaves, simply because of its “higher socio-cultural development.”20 It is clear from the preceding review of the literature that the African backgrounds of Cuban slaves were diverse. They did not share a “unique” culture, except for the common denominator of being racially excluded.21 Africans in Cuba were not brought en masse from a culturally unified African continent, and their ethnic traditions were neither totally lost in transit nor retained merely as submerged fragments.22 On the contrary, ethnicity remained especially important, as slave populations in the New World seem to have been drawn disproportionately from relatively few of the peoples of Africa.23 Slave departures from Africa were concentrated in a dozen or so ports along the African coast.24 From the African side of the exchange of slaves for merchandise, slaves entered these ports from regional hinterlands, where ethnic groups were numerous.25 As a consequence, slaves who would not normally have identified with each other mixed together in the coastal holding areas. Once in Cuba they sought to preserve and highlight cultural similarities with people that might belong to their own particular ethnic background, but in some cases such interaction would have included people belonging to a similar, though not identical, cultural background. Identities were first reshaped when they were moved from the interior to the coast and while awaiting shipment. On the vessels that brought them to Cuba, further bonds were secured along spontaneously generated ethnic lines and communicated through a common language, most of the time overcoming dialectal differences.26 Thus a new sense of ethnic identity was already evolving before they landed in the Americas. Once in Cuba, slaves shipped from the same ports or regions were given ethnic labels that identified them as members of certain African “nations,” a process that apparently tended to cement the new bonds already forged. Once on the plantations, enslaved Africans began to establish additional contacts, both with other Africans perceived as culturally related and also with others of completely different cultural, ethnic, and linguistic backgrounds. Perceived affinities with certain peoples, and separateness from others, reinforced their sense of belonging to particular “affinity groups” or intentional communities with shared proper names, languages, cultural identities, links to a homeland, collective memories, and, just as significant, “shared amnesias.”27 The scale of linguistic and cultural diversity within particular African regions from which Cuban slaves were drawn must be taken into account in this

181

182

Oscar Grandío Moráguez

analysis. The Bight of Biafra hinterland, for example, contained at least four major languages—Igbo, Ibibio, Edo, and Ijo and their respective dialects— together with many other minor languages including Efik, which was spoken by many slaves who came to Cuba.28 To point to the predominance of the Bight of Biafra as a region of origin for Cuban’s slave population is not therefore to say that all slaves from that region shared much of an identity or a mutually intelligible language. Many Africans from the Bight of Biafra, who had never heard the name “Ibo” in their own lands and identified themselves instead by their villages or districts, nevertheless came to recognize, at least to some degree, the term in the Americas. They may even have incorporated people and cultural traits from places far removed from the Bight of Biafra.29 A similar example could be used for the slaves labeled in Cuba as “Lucumí,” most of whom came from the Yorubaland yet also included people sold by the Yoruba but who were not Yoruba speakers. So non-Yoruba people like the Hausa or Fon were integrated within the so-called Lucumí nation.30 Any study dealing with the forced African repopulation of Cuba requires some basic information on how many immigrants came from where and over what periods. Given advances in the study of African history and the history of the slave trade over the last decades, it is now possible to reconstruct the backgrounds of these broad ethnic categorizations in Cuba. Shipping data are now considered vital to such inquiries, since they offer clues regarding the precise regions from which the slaves were shipped. The use of this kind of source is not new in the Cuban slave-trade historiography. In 1827 Alexander von Humboldt used Havana customhouse returns to estimate the volume of slaves imported into Havana during the last years of the legal slave trade.31 Six years later, José Antonio Saco, using the same hard data, published a similar study.32 Thereafter there was little until 1978, when Herbert S. Klein, using fragmentary data from the same sources used by Humboldt and Saco but collected in Seville, produced a database that incorporates many of the slave vessels that brought slaves legally into Havana between 1789 to 1821.33 The publication of the Transatlantic Slave Trade Database (TSTD1) on CD-ROM in 1999 opened up a new path for the study of the Cuban slave trade. Its more than 27,000 transatlantic voyages provided a basis for modifying accepted estimates of the size of the trade and also provided the foundations for a more precise reassessment regarding embarkation points on the African coast and where in the Americas slaves were taken.34 The database confirmed that the structure of the transatlantic slave trade tended to concentrate rather to

The African Origins of Slaves Arriving in Cuba

disperse people from broad regions into a limited number of points on the African coast. For Cuba, in particular, the implications were enormous. For the first time, estimates were produced on the African origins of slaves arriving to Cuba. The new data set used mostly British sources—basically Foreign Office records from 1821 to 1867—and added the data used by Klein in 1978 for the period between 1789 to 1821. The comparative perspective suggests that, of all the major importing areas, Cuba received the greatest mix of African peoples at the same time. No single part of Africa supplied more than 28 percent of arrivals to the island, and the only major region not well represented was the Gold Coast. The database also shows that the region supplying the greatest number of slaves to Cuba was West Central Africa.35 The 1999 database nevertheless omitted many voyages because its pre-1821 Spanish sources were confined to data from the Klein database, itself far from complete. The post-1820 British sources were quite complete indeed. The British commissioners at the Havana Court of Mixed Commission sent many reports on illegal slave landings in the vicinity of Havana throughout the period.36 Moreover, British consular authorities in Cuba, mostly at Santiago, also reported many imports. In addition, the British commissioner stationed at Sierra Leone reported all the Cuban slavers that came to his attention. Nevertheless, much information on Cuban voyages can be found in the Spanish and Cuban records and was not generally accessible to the British. These records are scattered among various Spanish archives, mainly in the collections deposited at the Archivo General de Indias in Seville and at the Archivo Histórico Nacional in Madrid.37 This chapter attempts to fill this gap by generating new estimates of the African origin of slaves arriving in Cuba between 1790 and 1867, when the island received over 95 percent of all Africans estimated to have arrived in Cuba. New data were collected at the two archives mentioned above and consolidated with TSTD1’s records to generate new estimates of the African regional participation in the Cuban slave trade. Consistent with the procedures outlined in chapter 1 for the overall slave trade, the consolidated data were grouped first under the national flag of the carrier from which the estimates of slaves transported are derived, and then these estimates were distributed across the eight broad coastal regions of sub-Saharan Africa.38 Table 6.2 provides five-year estimates of slave departures bound for Cuba from these eight regions between 1790 and 1867. Analysis of the data clearly suggests that West Central Africa, followed by the Bight of Biafra–São Tomé and the Principe Island–Cameroon–Gabon regions, was the largest source of slaves for Cuba in this period.39

183

184

Oscar Grandío Moráguez

Table 6.2 Arrivals of Slaves in Cuba by Major African Coastal Regions and Five-Year Periods, 1791–1867

Years

Senegambia

1791–95 1796–1800 1801–5 1806–10 1811–15 1816–20 1821–25 1826–30 1831–35 1836–40 1841–45 1846–50 1851–55 1856–60 1861–65 All years

150 0 1,443 163 250 636 570 0 0 1,062 1,320 0 0 0 0 5,594

Sierra Windward Leone Coast 376 698 342 650 312 924 2,949 3,680 4,117 5,860 1,923 0 540 485 0 22,856

0 45 236 158 489 341 853 136 0 747 0 0 0 0 0 3005

Bight of Benin

Bight of Biafra

West Central Southeast Africa Africa

1,783 1,766 815 70 4,046 0 71 0 0 0 0 419 500 2,259 0 2,846 639 6,783 500 6,087 0 4,532 0 150 0 1,870 0 3,491 0 3,526 8,354 33,799

1,924 464 5,246 534 1,107 1,166 8,207 13,454 17,007 8,550 500 402 0 0 0 58,561

3,509 299 3,975 1,796 735 3,923 205 744 6,731 4,817 1,464 1,719 8,173 21,346 8,198 67,634

Gold Coast

1,787 0 1,523 0 0 1,359 628 0 660 11,054 711 0 5,723 1,430 0 24,875

Source: TSTD2.

Table 6.3 regroups table 6.2 into longer periods to highlight broad trends over time. Before 1806, West Central Africa emerges as the leading region, with 25.5 percent of the total of slaves identified in the sample. The Bight of Biafra supplied just over one-quarter of the slaves arriving, but the Gold Coast was not far behind—the only period of the Cuban slave trade when that region was important. On the African coast the slave trade was a highly specialized business. Particular European nations generally dominated the trade with entire African regions.40 Before 1807, Cuban slave markets were supplied by British and to a lesser extent by U.S. carriers, so it is not a surprise that most slaves bound for Cuba before 1806 came from the three areas dominated by British and American merchants.41 West Central African slaves usually came from ports north of Luanda, where the British business was centered.42 In 1807 Great Britain prohibited all British involvement in the transatlantic slave trade, committing significant resources to stopping the slave trade worldwide. The government signed treaties with many nations to outlaw the embarkation of African slaves to the Americas. A British squadron was sent to patrol the West African coast as a deterrent to the slavers still operating in the

Table 6.3

Arrivals of Slaves in Cuba by Major African Coastal Regions and Fifteen- or Twenty-Year Periods, 1790–1865 1790–1805

Region of Embarkation Senegambia Sierra Leone Windward Coast Gold Coast Bight of Benin Bight of Biafra West-Central Africa Southeast Africa Source: TSTD2.

1806–20

1821–40

1841–65

All Years

Number

Col. %

Number

Col. %

Number

Col. %

Number

Col. %

Number

Col. %

1,593 1,416 281 6,644 1,836 7,634 7,783 3,310

5.2 4.6 0.9 21.8 6.0 25.0 25.5 10.9

1,049 1,886 988 71 419 2,807 6,454 1,359

7.0 12.5 6.6 0.5 2.8 18.7 42.9 9.0

1,632 16,606 1,736 1,639 17,975 47,218 12,497 12,342

1.2 15.6 1.7 1.5 16.1 41.8 11.1 11.0

1,320 2,948 0 0 13,569 902 40,900 7,864

2.0 4.4 0.0 0.0 20.1 1.3 60.6 11.6

5,594 23,856 3,005 8,354 33,799 58,561 67,634 24,875

2.5 10.2 1.3 3.7 15.0 26.1 30.1 11.1

186

Oscar Grandío Moráguez

region.43 The presence of the British squadron significantly slowed the pace of slave exports in some West African regions, while in others it had important consequences for the way the trade was structured. After British slave-trade abolition, most of the slaves exported to the Americas from West Africa were embarked at points that were out of sight and almost out of reach of the British navy. These developments are reflected in the Cuban data’s next period (1806–20), when some of the slavers operating in West Africa moved their activities south to West Central Africa, avoiding the British squadron. The Gold Coast, with its British forts and no mangrove swamps to aid concealment, would disappear as a slave supplier for the Cuban market, whereas West Central Africa, largely beyond British patrols at this time, sharply increased its share of the Cuban trade, accounting for 43 percent of all arrivals from known African regions. The Bight of Biafra witnessed a slight reduction of its slave exports to Cuba, accounting for just under one in five of arrivals, while Sierra Leone surprisingly expanded its participation. The first region was not heavily affected by the British repression, because most of its slave exports were boarded through the maze of creeks and lagoons that formed the deltas of the Niger and Cross rivers, out of reach of the British cruisers. Sierra Leone managed to expand its exports even under the nose of the British squadron; it reorganized the structure of its trade by bringing new slaving ports—in the Gallinas and the Rio Pongo, located on shallow creeks—into the business, while closing down others more exposed to British interference.44 West Central African dominance faltered only during the years 1821–40. In these years the Bight of Biafra expanded its participation as a supplier of the Cuban slave market, exporting two-fifths of all slaves arriving during the period. Most of this expansion, however, can be explained by the fact that many voyages attributed to the region were reported as coming from São Tomé or Principe Island. In addition, the Bight of Benin shows up for the first time as one of the leading exporters to the Cuban slave market, embarking 16.1 percent of all slaves bound to Cuba. During those years, Sierra Leone—the Gallinas and Rio Pongo regions—also experienced a high level of departures for Cuban ports, accounting for 15 percent of arrivals. West Central African numbers declined sharply, while Southeastern Africa supplied one in ten of Cuban bozales (recently arrived slaves). The persistence of the Bights and Sierra Leone as viable slave exporting areas to Cuba until the 1840s indicates how ineffective the British squadron

The African Origins of Slaves Arriving in Cuba

was in coping with the Cuban slave trade. The reasons are various. First, the security that shallow creeks and lagoons at secluded embarkation points provided against the British navy apparently offset any disruption caused by the squadron activities. In the 1820s and 1830s, slavers generally anchored near the places where they intended to acquire slaves and, after landing the goods used in the trade, stood out from shore until the cargo was ready. If no British cruisers were near, they stood in again and took the slaves onboard. Second, treaty restrictions made it difficult for the British squadron to capture slavers near Cuba. Prior to the 1830s, only a few slavers bound for Cuba were seized. The length of the Cuban coast and the thousands of small islands and keys around it made it almost impossible to patrol the Cuban perimeter effectively. Moreover, since more than 1,500 vessels entered Cuban ports annually, it was difficult to distinguish the slavers, and errors were frequent. The problems were heightened by the fact that the Spanish interisland slave trade was still legal.45 In addition, a clause of the 1817 treaty stated that slavers could only be seized if there were slaves on board. The British minister to the Spanish court, George Lamb, drew attention to the limitations faced by the cruisers. Lamb even asked the Spaniards to add a new article to the 1817 treaty that would permit the seizure of slavers waiting for their slave cargoes, a proposal that Madrid categorically dismissed.46 As a substitute, however, Madrid issued a new royal decree in March 1830 that allowed the Cuban captain-general to impose severe fines “upon so inhuman [a] traffic [so] that no longer would there be any importation of Negroes into the island.”47 The new decree, couched in such general terms, was not even published in Cuba, and slavers continued to enter the island’s ports with complete impunity. The British blockade became more effective after the new Anglo-Spanish treaty of 1835. Previously, as long as the British squadron could only seize vessels with slaves onboard, it was necessary to cruise at a considerable distance from the coast, out of sight of the slavers. This pattern changed after 1835, when slavers could be seized on outward voyages under the equipment article. It thus became possible for the squadron to cut slavers off from the African coast by establishing an “in-shore” blockade.48 As a result, when ships equipped for the slave trade tried to penetrate the blockades, cruisers would seize them. This was a factor causing slave exports from the Bight of Biafra and Sierra Leone to Cuba to decline considerably during the 1840s and 1850s and to disappear completely in the 1860s, which left West Central Africa accounting for more than three out of every five slaves delivered to the island during the last

187

188

Oscar Grandío Moráguez

quarter century of the trade, with the Bight of Benin and Southeast Africa supplying 20.1 percent and 11.6 percent, respectively.49 The slave trade along the African coast was dominated by a few embarkation points, although the identity of these ports changed slowly over time. The Congo River, Cabinda, Bonny, and Ouidah, among others, seem to have dominated their respective regions from the beginning of the Atlantic slave trade, while ports such as the Gallinas and Lagos emerged as major players over time. Table 6.4 shows a breakdown of embarkation points on the African coast that supplied slaves to Cuban ports. It reveals that more than three slaves out every five landing in Cuba from the late eighteenth century to 1865 set sail from just six embarkation points: the Congo River, Mozambique, the Gallinas, Bonny, Table 6.4 Departures of Slaves to Cuba from Individual African Ports, 1790–1865 Port of Embarkation Congo River Mozambique* Gallinas Bonny Ouidah Lagos Ambriz Cabinda Luanda Benguela Calabar Loango Rio Pongo Popo Cape Coast Kilongo New Calabar Sherbro Bissau Keta Madagascar Zanzibar

Number of Slaves

Port of Embarkation

Number of Slaves

29,897 21,758 15,126 14,854 12,670 9,615 7,040 6,766 5,251 4,556 4,349 4,259 3,766 2,939 2,757 2,003 1,941 1,522 1,239 1,109 1,072 1,008

Aghway Rio Brass Bance Island Accra Inhambane Gambia Senegal Corisco Gabon Anomabu, Adja, Agga Porto Novo Iles de Los Grand Cape Mount Gorée Bimbia Elmina Cameroons River Cameroons Malembo Mayumba Lahou (Cap) Coanza River

950 915 864 855 840 583 577 558 538 534 444 435 408 342 340 273 238 215 205 120 91 26

Source: TSTD2 *Mozambique figures are biased upward by the nature of the historical records (where this region often means “Southeast Africa”).

The African Origins of Slaves Arriving in Cuba

Ouidah, and Lagos. The fact that these six venues were widely scattered underlines the diversity of the Cuban slave population, though without denying the centrality of the Angolan-Cuban connection. Until 1807 Bonny, the Congo River, and New Calabar embarked the highest volume of slaves to Cuba. After 1807 Luanda became the leading slaveexporting port to Cuba, although Bonny retained its importance, and Loango, now occupied by Portuguese agents, became one of the most important slave exporters to Cuba. For the period 1821–40, after the Cuban slave trade was declared illegal, the Gallinas (where Cuban-based traders had established slave factories) and ports along the Mozambican coast emerged as the most important shipping points. Bonny retained its leading role, but its dominance did not last. In the final period (1841–65), the Congo River became the leading port, while Mozambique retained its position, followed by a resurgent Ouidah and Benguela. Four more West Central African ports—Ambriz, Cabinda, Kilongo, and Loango—reinforced the importance of the Congo coast as the leading exporting area to Cuba. Is it possible to develop estimates of the ethnicities of slaves departing from Africa on the basis of this new information? Some connections between ethnicity and port are possible. In general, evidence suggests that most of the slaves traded at the coast came by the most direct route. It does seem that slaves from specific African ethnic or language groups tended to leave from the port closest to where that group was located. For example, most of the slaves departing from Lagos were Yoruba speakers; most of the people leaving from Bonny in the Bight of Biafra spoke Ibo or Ibibio; and a majority of the people embarked from the mouth of the Congo River were Kikongo speakers or spoke a language close to it. Within the Upper Guinea region, encompassing Senegambia, Sierra Leone, and the Windward Coast, two different embarkation patterns with implications for ethnicity are apparent. The first, evident before 1805, was characterized by the active involvement of several embarkation points around Gambia, Sierra Leone, Iles de Los, Senegal, and Gorée. As already noted, a dramatic change would occur after 1806, with the Rio Pongo and Gallinas regions emerging as the leading departure points for slaves bound to Cuba, before the Gallinas alone dominated the region’s slave trade after 1820. Cuban-based slave traders opened several slave factories in the Gallinas, the most notorious one run by Pedro Blanco.50 Blanco traded extensively with local chiefs from the hinterland, especially among the Dei and the Gola, who were at war with each other, obtaining prisoners from both parties. He also purchased many slaves further south, from Digbe and the St. Paul River area,

189

190

Oscar Grandío Moráguez

transporting them overland to markets in the Gallinas.51 By the beginning of the 1840s, the slave trade around the Gallinas seemed to taper off. Blanco left the coast, and his agent, Theodore Canot, maintained the Gallinas factory. After a series of British naval assaults in the 1840s, the Gallinas was almost free of the slave trade by 1850.52 Slaves embarked there were commonly labeled in Cuba as “Gangá,” a term that also included any slave shipped from slave-trading points in southern Sierra Leone. Other people were identified, once in Cuba, as “Mandinga,” a label that grouped all people dispatched north of Sierra Leone from the Rio Pongo coast. These markers did not say much, however, about the true ethnic identities of the individuals embarked from either shipping point. The Sierra Leone interior is quite diverse—ethnically and linguistically— but, in general, most ethnic groups inhabiting the region have been grouped under two language families: Mande and Mel.53 It is thus likely that most of the slaves coming from this area spoke a language belonging to one of these two families, which no doubt facilitated their integration within a broad new diasporic identity once in Cuba. On the Gold Coast and in the Bight of Benin (considered here as a single region), the Gold Coast ports dominated Cuban imports up to 1806, with only a few slaves leaving from the Bight of Benin ports. For the next decade and a half, few slaves were reported as coming from the region, with most of them leaving from the Bight of Benin. After 1820 Lagos and Ouidah would recover and dominate as embarkation points to the Cuban market for the next thirty years, although Lagos would disappear from the trade in 1851 after its capture by the British. Ethnicity in this region is relatively easy to identify. Lagos was perfectly located to be a major slave exporter. It was close to a major source of slaves: the collapsing Oyo Empire. It was probably the major generator of transatlantic slaves embarking from the Slave Coast after 1825. Lagos’s lagoon system permitted traders to move slaves relatively quickly to embarkation points beyond the knowledge of patrolling cruisers.54 In the case of Ouidah, its middleman position in the transatlantic slave trade had declined somewhat since the first quarter of the eighteenth century, and it was eventually superseded by Lagos as the main slaving port in the Bight of Benin area.55 During most of eighteenth century, captives bound for the Americas from Ouidah traditionally were obtained in Dahomey’s slave catchment zone and tended to be Ewe-Fon and Aja. As noted, Ouidah would recover to some degree after 1850, on the basis of more slaves originating beyond Dahomey’s borders. In fact, slaves exported from Ouidah and Lagos during most of the nineteenth century were generally Yoruba speakers, although slaves belonging

The African Origins of Slaves Arriving in Cuba

to the Ewe-Fon/Aja cultural ensemble also found their way to Cuba. The Yoruba wars that destroyed the Oyo Empire supplied many of these slaves. Abeokuta, Ibadan, Ilorin, Ijebu Ode, Ife, and Ketu, all originally part of Oyo, lost many people as slaves to be exported on the coast, while lesser numbers of mostly Hausa speakers came from the savanna to the north.56 In the Bight of Biafra, the degree of concentration at first looks to have been extreme, with São Tomé and Principe Island accounting for about 50 percent of all voyages to Cuba reported as originating in that region. The high percentage of São Tomé and Principe voyages, however, can be explained by the fact that most of these voyages were reported after the 1835 Anglo-Spanish treaty took effect. Almost all of them sailed under the Portuguese flag, and no doubt reflected the fact that Cuban slavers were looking for a Portuguese base on the African coast to use as a cover, given that a Portuguese ship sailing between Portuguese ports could claim to be carrying slaves from one Portuguese possession to another and thus avoid British interference. Hence many slavers at this time left Cuban ports for São Tomé and Principe Island, where they picked up Portuguese papers before heading for West Central African, Bight of Biafra, or Bight of Benin ports, where slaves would be purchased. The Gulf of Guinea islands are thus not listed in table 6.1.57 Leaving aside São Tomé and Principe Island, one single port from the Bight of Biafra prevailed: Bonny.58 Calabar and New Calabar were also of some importance for the Cuban market. All three ports, located at the deltas of the Niger and Cross rivers, were controlled by local merchants who brought slaves from the huge markets in the interior dominated by the Aro’s commercial networks, mainly covering the Ibo-Ibibio area.59 Such markets in the interior could last for several days at a time. The Bende fair, not far from Bonny, was an example of a megamarket where Ibo slaves were gathered to be sold to the coastal traders of Bonny and New Calabar. Most of the enslaved people brought to such markets were victims of kidnapping, intervillage raiding, and enslavement though oracles.60 Many of them were brought to the coast via a labyrinth of creeks and lagoons along the river deltas. Once on the coast, the slaves were immediately boarded on vessels because the deltas’ unhealthy environment did not allow the maintenance of coastal depots. British slave traders before 1807 and Cuban-based merchants after this date dominated the slave trade from the Bight of Biafra, in partnership with local agents. Cameroon and Gabon, by contrast, did not export significant quantities of captives to Cuba.61 Between 1790 and 1865, West Central Africa experienced an explosion in exports that reaffirmed the region’s position as the largest single supplier of

191

192

Oscar Grandío Moráguez

the slave trade to the Americas. Most of the West Central African slaves arriving in Cuba during this period came from the Congo River, Ambriz, and Cabinda, areas that expanded their participation in the trade in the late eighteenth century after the decline of Loango as a major export center.62 The trade expanded gradually from the beginning of the nineteenth century along the Congo River and its tributary embarkation points to the north, with many slaves coming from the Malebo Pool and the savanna to the south and east. Vili traders, subjects of the state of Loango, an absolute monarchy inhabited by the Vili, controlled the slave routes supplying the Congo River, Ambriz, Cabinda, and Loango exporting outlets.63 They acted as middlemen between the Europeans and the Bobangi, who were experts in long-distance trade. During most of the eighteenth century and the beginning of the nineteenth, the Bobangi traders acquired slaves deep in the interior of the Congo basin, especially from the Tio.64 The period between 1807 to 1820 saw Loango and points in the Congo estuary emerge as the most important embarkation points for Cuba’s West Central African slaves. Starting in 1815, Luanda began to supplant them, as the Cuban slave trade expanded strongly in the five years before 1820. Many slaves arriving in Luanda during the 1810s came from the interior of Angola, mostly from the regions of Encoge and Dembos via the copper-mining district of Bembe. The Congo River seems, during this period, to have vanished as an important exporting conduit. For several years after 1821, when the treaty prohibiting the Cuban slave trade came into effect and the traffic to Cuba temporarily declined, imports of slaves from West Central Africa disappeared altogether. In the late 1820s, however, embarkation points connected with the Congo reemerged and accounted for five out of six slaves leaving the region for Cuba. The smaller trade from further south—Luanda, Benguela, and the Cuanza River—seems to have been reactivated by Portuguese and Brazilian traders operating in partnership with Cuban merchants.65 After the 1840s, when British cruisers started systematic patrols south of the equator and when antislave trade treaties became more extensive, smuggling and secrecy gained new importance, causing the slave trade to intensify along the lower Congo River, with its many hidden creeks and favorable conditions for smuggling.66 The data reflect this development, with embarkation points in or close to the mouth of the Congo accounting for over half of the slaves from all African regions sent to Cuba after 1841. Here slavers were able to carry on the traffic because of the absence of European authorities.67 They dealt with a variety of African middlemen who supplied

The African Origins of Slaves Arriving in Cuba

them with slaves. Presents to the local chiefs gave them access to trade facilities. Mboma, just fifteen miles from Punta da Lenha on the Congo River, was a major slave center during the 1850s, with its demand for slaves being supplied from a major trade route coming from the Malebo Pool, which in turn was fed by upriver markets. In Mboma, slaves from the interior were collected and sold to Portuguese and Spanish agents in Punta da Lenha; they were then taken to the different embarkation points downstream or to points such as Cabinda.68 The Southeast African data seem more problematic. They show that slave exports to Cuba were heavily concentrated in one area, Mozambique, which accounted for 88 percent of all the estimated slaves from this region disembarked between 1790 and 1865. Mozambique’s involvement in the Cuban slave trade began in 1793. Prior to 1835, shipments were occasional, and there were long periods when it sent no slaves to Cuba. Departures to Cuba then expanded rapidly and were significant between 1835 and 1841 and then again in the 1850s, before ending in 1859.69 In the historical record, however, the term “Mozambique” often means any shipping point in Southeast Africa, rather than the island of Mozambique itself. This traffic was always an appendage to the much larger Brazilian traffic, but it does reflect in part the relative scarcity of British antislave patrols in the Indian Ocean before the 1840s. The broader picture includes the decline of the Mozambican high-quality ivory trade and a fast regional switch in the slave trade to the Mascarene Islands, as well as Arab, Brazilian, and Cuban markets.70 Slaves came from the Mozambique hinterland as a consequence of extensive slave raiding among two ethnic groups, the Yao and Makua.71 Some important conclusions emerge from this survey. Slaves leaving for Cuba from particular regions after 1790 were designated with broad country terms such as Gangá, Mandinga, Lucumí, Carabalí, Congo, or Macua. Each of these ethnic designators had a geographic definition and represented a hybridized ethnicity that did not exist in Africa. A Congo, for example, was someone who came from the hinterland of West Central Africa, but not necessarily an individual who would have identified himself in Africa as Congo. What we know is that most of the slaves identified in Cuba as Congo arrived from embarkation points in or close to the Congo River. These slaves were supplied from the near hinterland. Many of the areas on the Congo coast supplying the Cuban market were indeed Kikongo, but many were not. Teke or Tio slaves, who spoke a language close to Kikongo, were transported through the Congo

193

194

Oscar Grandío Moráguez

River or from land routes to slave markets such as the Malebo Pool or Mboma, where they were transferred to shipping points extending from Loango to Ambriz. Similarly, many Mbundu people, who also spoke a language mutually intelligible with Kikongo, were embarked at the Congo coast. Slaves labeled in Cuba as Lucumí were generally Yoruba speakers, although it is clear that many people classified as Lucumí were non-Yoruba. It is true that the routes supplying Lagos after 1790 were mostly linked to Yoruba areas, severely affected by the Oyo civil wars, but important contingents of non-Yoruba slaves were shipped from Lagos as well, especially Ewe-Fon/Aja people or Hausa. Other “nations” in Cuba, like the Carabalí, were identified as coming from the hinterland of the major Bight of Biafra ports. Many captives passed through Bonny, one of the biggest suppliers of Cuban slaves. Such slaves originated in the immediate hinterland and were for the most part Ibo or Ibibio speakers, but they also included many Efik, Ekoi, or even Yoruba, brought from the far interior. Slaves labeled as Gangá or Mandinga corresponded with those people coming from the Sierra Leone–Liberia hinterland. Most of the slaves embarked in the Gallinas or Rio Pongo, the main shipping points in the region after 1807, were automatically labeled as Gangá or Mandinga once they disembarked in Cuba, but these terms obscure the original ethnic identities of the slaves embarked there. What we know is that most belonged to two language families: Mande and Mel. The data provided in this chapter also underline wide disparities among the eight traditionally defined African regions in the supply of slaves to Cuban slave markets between 1789 and 1865. One of these eight—West Central Africa—was clearly the major supplier of slaves. The regional data available modify our understanding of the relative importance of certain regions as suppliers of slaves to Cuba. Some scholarly works have deemed the Bight of Benin the most important African supply region because of the apparently heavy Yoruba component in the Afro-Cuban population. The quantitative information analyzed here, however, suggests that the region’s share of slave exports to Cuba was considerably lower than the exports from West Central Africa. Furthermore, the data show an uneven distribution of slave departures within the eight regions, with some ports exporting more slaves than others. A very few embarkation points tended to dominate their respective regions, though the identity of these ports changed slowly over time. Estimates indicate that certain ports were considerably more important than has been usually assumed. This is particularly evident in the cases of the mouth of the

The African Origins of Slaves Arriving in Cuba

Congo River in West Central Africa and Bonny in the Bight of Biafra. Many of the slaves embarked at the latter port were probably brought from the interior of the Niger River basin and from the area inhabited by the Ibo and Ibibio linguistic clusters. Around the Congo River, during the period analyzed, many captives seem to have traveled routes heading to the interior of the Congo River basin, where Kikongo or related languages predominated. In the Bight of Benin, Lagos and Ouidah dominated the overall regional export to Cuba, although their traffic was below the levels of the Congo and Bonny ports. Between 1836 and 1850, however, Lagos’s and Ouidah’s exports surpassed those from the Congo and Biafra. Certain problems remain. Why, for example, if the Bight of Biafra or West Central Africa exported more slaves to Cuba than the Bight of Benin during the boom years of the Cuban slave trade, does Yoruba culture seem to be more predominant in Cuba? The answers are likely to be complex, and will require substantial additional research.

NOTES

1. For a summary of the extensive literature on the volume of the slave trade, Per O. Hernaes, Slaves, Danes, and African Coast Society: The Danish Slave Trade from West Africa and Afro-Danish Relations on the Eighteenth-Century Gold Coast (Trondheim, 1995), 129–71. For a more recent reassessment, see David Eltis and David Richardson, “A New Assessment of the Transatlantic Slave Trade,” this volume (chap. 1). 2. A huge variety of sources created by Europeans in the Americas identified Africans with certain “nations.” These sources included slave traders’ accounts and papers, official colonial and ecclesiastical records, fugitive slave advertisements and former slave narratives, collective memories set in slave-derived institutions such as secret societies and brotherhoods and folk religions, and vernacular performances of songs, dances, and the like, as well as numerous contemporary reports and investigations. See examples in Elizabeth E. Donnan, ed., Documents Illustrative of the History of the Slave Trade to America, 4 vols. (Washington, D.C., 1930–35; Buffalo, 2002 ed.); A. De Sandoval, De instauranda aethiopum salute: El mundo de la esclavitud negra en América (Bogota, 1956); C. G. A. Oldendorp et al., C. G. A. Oldendorp’s History of the Mission of the Evangelical Brethren on the Caribbean Islands of St. Thomas, St. Croix, and St. John (Ann Arbor, 1987); Edward Long, The History of Jamaica; or General Survey of the Ancient and Modern State of that Island (London, 1970); S. W. Koelle, Polyglotta Africana (London, 1854); Lathan A. Windley, Runaway Slave Advertisements: A Documentary History from the 1730s to 1790 (Westport, 1983); Gwendolyn Midlo Hall, ed., Databases for the Study of Louisiana History and Genealogy (Baton Rouge, 1999); and Daniel C. Littlefield, Rice and Slaves: Ethnicity and the Slave Trade in Colonial South Carolina (Baton Rouge, 1981).

195

196

Oscar Grandío Moráguez

3. Philip D. Curtin, The Atlantic Slave Trade: A Census (Madison, Wis., 1969), 183–90. 4. On Koelle’s survey, see Philip D. Curtin and Jan Vansina, “Sources of the NineteenthCentury Atlantic Slave Trade,” Journal of African History 5 (1964): 185–208; P. E. H. Hair, “The Enslavement of Koelle’s Informants,” Journal of African History 6 (1965): 193–203. 5. Paul E. Lovejoy, “The Impact of the Atlantic Slave Trade on Africa: A Review of the Literature,” Journal of African History 30 (1989): 365–94; David Richardson, “Slave Exports from West and West-Central Africa, 1700–1810: New Estimates of Volumes and Distribution,” Journal of African History 30 (1989): 1–22; P. D. Morgan, “The Cultural Implications of the Atlantic Slave Trade: African Regional Origins, American Destinations and New World Developments,” in David Eltis and David Richardson, eds., Routes to Slavery: Direction, Ethnicity and Mortality in the Transatlantic Slave Trade (London, 1997), 122–45; David Eltis, “The Slave Trade: The Story of the Atlantic Slave Trade, 1440–1870,” William and Mary Quarterly 55 (1998): 622–23; Anthony J. Barker, review of Eltis and Richardson, eds., Routes to Slavery, International Journal of African Historical Studies 32 (1999): 559–60; Paul E. Lovejoy, “Identifying Enslaved Africans in the African Diaspora,” in Paul E. Lovejoy, ed., Identity in the Shadow of Slavery (New York, 2000), 1–29. 6. David Eltis, Stephen D. Behrendt, David Richardson, and Herbert S. Klein, The Transatlantic Slave Trade: A Database on CD-ROM (Cambridge, 1999). 7. See Martin Lienhard, Le discours des esclaves: De l’Afrique à l’Amérique latine: Kongo, Angola, Brésil, Caraïbes (Paris, 2001). 8. Esteban Pichardo, Diccionario provincial casi-razonado de voces cubanas (Havana, 1849); José María de la Torre, Compendio de geografía, física, política, estadística y comparada de la isla de Cuba (Havana, 1854); idem, Nuevos elementos de geografía e historia de la isla de Cuba (Havana, 1868); José Miguel Macías, Diccionario cubano, etimológico, crítico, razonado y comprensivo (Coatepec, 1888). 9. Pichardo, Diccionario, 6. 10. Henri Dumont, “Antropología y patología comparada de los negros esclavos (Memoria inédita referente a Cuba),” Revista Bimestre Cubana 10 (1915–16): 11. 11. Beato’s study, however, did not link African ethnic groups in Cuba with any African region. Manuel Pérez Beato, “Procedencia de los negros de Cuba,” Revista Bimestre Cubana 5 (1910): 161–63. 12. Fernando Ortiz, Los negros esclavos (Havana, 1975 [1916]), 74. 13. Lydia Cabrera, La sociedad secreta Abakuá, narrada por viejos adeptos (Havana, 1959); idem, Anagó: Vocabulario lucumí (el yoruba que se habla en Cuba) (Miami, 1970); idem, Reglas de Congo: Palo Monte Mayombe (Miami, 1979); idem, Vocabulario congo: El bantú que se habla en Cuba (Miami, 1984); idem, La Regla Kimbisa del Santo Cristo del Buen Viaje (Miami, 1986). 14. Miguel Barnet analyzes the different African cultural retentions in contemporary Cuba but focuses on only two African groups: Yoruba and Congo. He writes that Yoruba culture was very rich at a “superstructural” level, whereas Congo culture was more “flexible” and that this “permeability” would have allowed the Congo culture to be dominated by the Yoruba. As elsewhere, this valorization of the Lucumís seems to be

The African Origins of Slaves Arriving in Cuba

determined by the fear generated by the Congo rituals with death. Miguel Barnet, Cultos afrocubanos: La Regla de Ocha, la Regla de Palo Monte (Havana, 1995), 7. Furé, an anthropologist, has investigated the different African ethnic nations and associations still surviving in Cuba. Rogelio Martínez Furé, Diálogos imaginarios (Havana, 1979) 15. Enrique Sosa, Los ñáñigos (Havana, 1982). 16. Rafael López Valdés, Componentes africanos en el etnos cubano (Havana, 1985). 17. Alejandro de la Fuente, “Denominaciones étnicas de los esclavos introducidos en Cuba: Siglos XVI y XVII,” Anales del Caribe 6 (1986): 75–96. 18. Ibid., 94–96. 19. Nery Goméz Abréu and Manuel Martínez Casanova, “Contribución al estudio de la presencia de las diferentes etnias y culturas africanas en la región central de Cuba: Zona de Placetas (1817–1886),” Revista Islas 85 (1986): 114–20. 20. Jesús Guanche Pérez, “Contribución al estudio del poblamiento africano en Cuba,” Africa: Revista do Centro de Estudos Africanos 18–19 (1995–96): 119–38. 21. In Cuba, ethnic identities were reconstructed with new names for each “nation,” and it seems that this process was common to the experience of enslaved Africans elsewhere in the Americas. This terminology was part of the repertoire of geographical names used by European and American slave traders when referring to Africa, some of which drew on the names of the African groups delivering slaves to the coast rather than of the slaves themselves. Frequently the orthographic corruption of the words further hinders their identifications. Too often the origins of the slaves were designated within a general location that prevents us from determining their specific ethnic filiations. In this sense, enslaved Africans in Cuba reconstituted themselves as broad and not always clear diasporic “nations” that were called Lucumí (Yoruba), Congo (West Central Africa), Arará (Ewe-Fon), Mandinga (Upper Guinea, including Senegambia), Gangá (Southern Sierra Leona), Carabalí (Bight of Biafra), and Macua (South-East Africa). 22. See Sidney Wilfred Mintz and Richard Price, The Birth of African-American Culture: An Anthropological Perspective (Boston, 1992); George A. De Vos, “Ethnic Pluralism: Conflict and Accommodation,” in Lola Romanucci-Ross and George A. De Vos, eds., Ethnic Identity: Creation, Conflict, and Accommodation (Walnut Creek, Calif., 1995), 15–47. 23. See David Eltis, “The Nineteenth-Century Transatlantic Slave Trade: An Annual Time Series of Imports into the Americas Broken Down by Region,” Hispanic American Historical Review 67 (1987): 109–38. 24. David Eltis and David Richardson, “The ‘Numbers Game’ and Routes to Slavery,” in Eltis and Richardson, eds., Routes to Slavery, 1–15; John Kelly Thornton, Africa and Africans in the Making of the Atlantic World, 1400–1800 (Cambridge, 1998). 25. See Paul E. Lovejoy, Transformations in Slavery: A History of Slavery in Africa (Cambridge, 2000). 26. Lovejoy, “Identifying Enslaved Africans.” 27. Shared amnesias refer to when enslaved people often “forgot” strictly local idioms, customs, and definitions in the diaspora. See Ernest Gellner, Culture, Identity, and Politics (New York, 1987); Manning Nash, “The Core Elements of Ethnicity,” in John Hutchinson and Anthony D. Smith, eds., Ethnicity (Oxford, 1996), 24–28.

197

198

Oscar Grandío Moráguez

28. For languages spoken by the people of Calabar descent in Cuba, see Rafael A. Núñez Cedeño, “The Abakua Secret Society in Cuba: Language and Culture,” Hispania 71 (1988): 148–54. 29. For an interesting debate on Igbo cultural traits, see David Northrup, “Igbo and Myth Igbo: Culture and Ethnicity in the Atlantic World, 1600–1850,” Slavery and Abolition 21 (2000): 7–14; Douglas B. Chambers, “The Significance of Igbo in the Bight of Biafra Slave Trade: A Rejoinder to Northrup’s ‘Myth Igbo,’ ” Slavery and Abolition 23 (2002): 101–20. 30. Robin Law, “Ethnicity and the Slave Trade: ‘Lucumi’ and ‘Nago’ as Ethonyms in West Africa,” History in Africa 24 (1997): 205–19. 31. Alexander von Humboldt, Ensayo político sobre la isla de Cuba, trans. José López de Bustamante (Paris, 1827) 32. José Antonio Saco, Colección de papeles científicos, históricos, políticos y de otros ramos sobre la Isla de Cuba (Havana, 1960). 33. Herbert S. Klein, “Slave Trade to Havana, Cuba, 1790–1820,” University of Wisconsin Data and Program Library Service, distributor (Madison, 1978). The data file contains information on Cuban port of arrival, date of arrival, type of ship, nationality of ship, number of slaves landed, African sailing date, slave mortality, number of slaves by sex, and captain of ship. 34. For detailed conclusions on the results produced by the database, see David Eltis, “The Volume and Structure of the Transatlantic Slave Trade: A Reassessment,” William and Mary Quarterly 58 (2001): 17–46. 35. This conclusion would challenge Ortiz’s approach to the issue and would raise many questions regarding the ethnic components of the Afro-Cuban population. 36. According to Eltis, it is likely that, in every year from 1821 to 1868, they recorded over 90 percent of every landing made by slavers that had any connections with the Havana area. See Eltis, “Nineteenth-Century Transatlantic Slave Trade.” 37. AGI, Secciones de Santo Domingo, Contratación, Indiferente General, Contaduría, Santa Fe; Archivo Histórico Nacional, Madrid, Secciones de Estado (Expedientes Esclavitud, Correspondencia con el Foreign Office, Correspondencia con el Capitán General) y Ultramar (Expedientes Esclavitud, Cuba, Cuba Secretaría de Gobierno). This new evidence has been incorporated into the second edition of the database (TSTD2). 38. The data contain estimates of vessels arriving from Africa, but also they have vessels departing from Cuban ports that intended to go to pick up slaves in Africa. Also they include vessels captured by British after taking slaves in Africa but that never landed in Cuba. 39. The Bight of Biafra–São Tomé and Principe–Cameroon–Gabon are probably overstated in these breakdowns. While São Tomé, Principe, and the Cape Verde Islands accounted for a total of 58 percent of all intended voyages from the Bight of Biafra–São Tomé and Principe–Cameroon–Gabon to Cuba, there is no certainty that the vessels actually took slaves onboard there; they may have used the islands simply as a cover to avoid British interference. I thank David Eltis for his comments on this issue. 40. David Eltis and Lawrence C. Jennings, “Trade between Western Africa and the Atlantic World in the Pre-Colonial Era,” American Historical Review 93 (1988): 936–59.

The African Origins of Slaves Arriving in Cuba

41. Richardson, “Slave Exports from West and West-Central Africa.” 42. 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, 1972), 93–117. 43. David Eltis, Economic Growth and the Ending of the Transatlantic Slave Trade (New York, 1987), 81–101. 44. Ibid., 164–84. Also see Philip Misevich, “The Origins of Slaves Leaving the Upper Guinea Coast in the Nineteenth Century,” this volume (chap. 5). 45. BFSP, vol. IX, Jameson to Clanwilliam, September 1, 1821, p. 141. 46. BFSP, vol. XIV, Lamb to Infantado, February 19, 1826, p. 261. 47. BFSP, vol. XVIII, Royal Order, March 4, 1830, pp. 499–505. 48. Eltis, Economic Growth, 125–44. 49. The key event in West Africa appears to have been the reorganization of the trade around the Congo River, where Cuban traders opened trade establishments in partnership with Portuguese and Brazilian dealers. 50. Lino Novás Calvo, Pedro Blanco, el negrero (Barcelona, 1999); William Renwick Riddell, “Observations on Slavery and Privateering,” Journal of Negro History 15 (1930): 337–71. 51. Svend E. Holsoe, “A Study of Relations between Settlers and Indigenous Peoples in Western Liberia, 1821–1847,” African Historical Studies 4 (1971): 331–62. 52. Patrick S. Caulker, “Legitimate Commerce and Statecraft: A Study of the Hinterland Adjacent to Nineteenth-Century Sierra Leone,” Journal of Black Studies 11 (1981): 397–419; Eltis, Economic Growth, 120–21. 53. On the linguistic diversity of the region, see P. E. H. Hair, “Ethnolinguistic Continuity on the Guinea Coast,” Journal of African History 8 (1967): 247–68. 54. Robin Law, “Trade and Politics behind the Slave Coast: The Lagoon Traffic and the Rise of Lagos, 1500–1800,” Journal of African History 24 (1983): 321–48; idem, “Between the Sea and the Lagoons: The Interaction of Maritime and Inland Navigation on the Pre-colonial Slave Coast,” Cahiers d’Etudes Africaines 29 (1989): 209–37. 55. For an overview of Ouidah’s role in the nineteenth-century slave trade, see Elisée Soumonni, “The Administration of a Port of the Slave Trade: Ouidah in the Nineteenth Century,” in Robin Law and Silke Strickrodt, eds., Ports of the Slave Trade (Bights of Benin and Biafra): Papers from a Conference of the Centre of Commonwealth Studies, University of Stirling, June 1998 (Stirling, 1999), 48–54. 56. Robin Law, “Slave-Raiders and Middlemen, Monopolists and Free-Traders: The Supply of Slaves for the Atlantic Trade in Dahomey, c. 1715–1850,” Journal of African History 30 (1989): 45–68. 57. Clarence-Smith mentions that the British navy presence near São Tomé and Príncipe, after the slave trade became illegal for Portugal, made it increasingly risky to transport slaves to the islands. Many slavers bound for Cuba still called there for water and provisions or for repairs, and acquired Portuguese colors and passports. William Gervase Clarence-Smith, The Third Portuguese Empire, 1825–1975: A Study in Economic Imperialism (Manchester, 1985), 39–41. 58. The Bight of Biafra was the most densely populated region along the West African coast during the eighteenth and nineteenth centuries. It obtained its slaves from sources both close to the shore and in the far interior. See Ebiegberi Joe Alagoa, A History of the Niger

199

200

Oscar Grandío Moráguez

Delta: An Historical Interpretation of Ijo Oral Tradition (Ibadan, 1972); A. J. H. Latham, Old Calabar, 1600–1891: The Impact of the International Economy upon a Traditional Society (Oxford, 1973); and David Northrup, Trade without Rulers: Pre-colonial Economic Development in South-Eastern Nigeria (Oxford, 1978). 59. According to Lovejoy, the vast majority of the captives sold at Bonny, New Calabar, and elsewhere along the Bight of Biafra came from the interior, not the coast itself. See Lovejoy, Transformations, 59–60. On the networks, see David Northrup, “The Ideological Context of Slavery in Southeastern Nigeria in the 19th Century,” in James L. Watson, ed., Asian and African Systems of Slavery (Berkeley, 1980), 101–8. 60. Oracles like the Ibinukpabi of the Aro, the Kamalu at Ozuzu, and the Niri required slaves as payments for misbehaviors and services rendered. See Northrup, Trade without Rulers, 133–34; Simon Ottenberg, “Ibo Oracles and Intergroup Relations,” Southwestern Journal of Anthropology 14 (1958): 295–317; S. J. S. Cookey, “An Ibo Slave Story of the Late Nineteenth Century and Its Implications,” Ikenga 1 (1972): 1–9; and Lovejoy, Transformations, 147–49. 61. On ethnicities of slaves arriving to Cuba from Cameroon ports, see G. Ugo Nwokeji and David Eltis, “Characteristics of Captives Leaving the Cameroons for the Americas, 1822–37,” Journal of African History 43 (2002): 191–210. 62. During the eighteenth century and the first decade of the nineteenth century, slavers arriving or leaving Cuba from West Central Africa usually declared “Angola” as their main destination but picked up slaves north of Ambriz and not in Portuguese Angola. That explains why “Angola” shows up in almost 50 percent of all slaves’ estimated departures to Cuba between 1790 and 1805. West Central Africa, which covers a vast coastline from south of Cabo López to the vicinity of Cabo Frio, extends far to the continent’s interior. This is a region encompassing many ecological zones, forests, rivers, savannas, and mountains, fostering different types of economic activities and therefore varied ways of life. The nineteenth-century web of ethnic groups, subgroups, languages, and dialects in the area was bewildering, with political units ranging from kingdoms to simple villages fragmented into several autonomous wards, and kinship systems that were patrilineal, matrilineal, and double-descent. For the slave trade in West Central Africa, see Martin, External Trade; Robert W. Harms, River of Wealth, River of Sorrow: The Central Zaire Basin in the Era of Slave Trade and Ivory Trade (New Haven, 1981); Joseph C. Miller, Way of Death: Merchant Capitalism and the Angolan Slave Trade, 1730–1830 (Madison, Wis., 1988); and Eltis, Economic Growth, 63–76, 159–77. 63. Jan Vansina, “Long-Distance Trade-Routes in Central Africa,” Journal of African History 3 (1962): 375–90. 64. On the Tio, see Jan Vansina, The Tio Kingdom of the Middle Congo, 1880–1892 (London, 1973); idem, Kingdoms of the Savanna (Madison, Wis., 1966), 98–113; Susan Herlin Broadhead, “Trade and Politics on the Congo Coast: 1770–1870” (PhD diss., Boston University, 1971), 68–70; Harms, River of Wealth, 27–37, 80. 65. For transatlantic links among Portuguese and Cuban-based slave traders during the nineteenth century, see Roquinaldo Amaral Ferreira, “Dos sertões ao Atlântico: Tráfico ilegal de escravos e comércio lícito em Angola, 1830–1860” (MA thesis, Universidade Federal do Rio de Janeiro, 1996).

The African Origins of Slaves Arriving in Cuba

66. Louis-Edouard Bouët-Willaumez, Commerce et traite des noirs aux côtes occidentales d’Afrique (Paris, 1848), 214; George E. Brooks, Yankee Traders, Old Coasters and African Middlemen: A History of American Legitimate Trade with West Africa in the Nineteenth Century (Brookline, Mass., 1970), 113–14; Martin, External Trade, 145. 67. Mary C. Karasch, “The Brazilian Slavers and the Illegal Slave Trade, 1836–1851” (PhD diss., University of Wisconsin, 1967), 49–50. 68. Norman Robert Bennett and George E. Brooks, New England Merchants in Africa: A History through Documents, 1802 to 1865 (Boston, 1965), 10–11. 69. Gerhard Liesegang, “A First Look at the Import and Export Trade of Mozambique, 1800–1913,” in Gerhard Liesegang, Helma Pasch, and Adam Jones, eds., Figuring African Trade: Proceedings of the Symposium on the Quantification and Structure of the Import and Export and Long Distance Trade in Africa, 1800–1913 (Berlin, 1986). 70. By 1809 ivory exports in Mozambique “had declined to about 7300 arrobas; in 1810 they had sunk to less than 4000 arrobas. By the following year, revenue collected on the exportation of slaves was more than five times that accruing from importation of ivory and other goods.” See Edward A. Alpers, Ivory and Slaves in East Central Africa: Changing Patterns of International Trade to the Later Nineteenth Century (London, 1975), 210–11. See also Curtin, Atlantic Slave Trade, 229–30. 71. Alpers, Ivory and Slaves, 209.

201

This page intentionally left blank

Part II National Slave Trades

This page intentionally left blank

Chapter 7 The Significance of

the French Slave Trade to the Evolution of the French Atlantic World before 1716 James Pritchard, David Eltis, and David Richardson

The transatlantic slave trade carried on by the most powerful nation of seventeenth- and eighteenth-century Europe presents a striking paradox. It is the best documented of all the national slave trades after the Treaty of Utrecht; before 1714, by contrast, we have only fragmentary information about French slaving activities. Historians have recognized this paucity but have still argued that early French slave trading “was formative and displayed all the major characteristics of its eighteenth century successor.”1 Paul Hair has documented more than one voyage a year from Le Havre alone declaring for Sierra Leone and the Americas between 1571 and 1610, and he argues that slaves formed the major component of this trade.2 A historian of the slave trade has argued that “serious participation began with the development of the monopoly trading companies in the second half of the seventeenth century.”3 Another has claimed that the French carried 35,000 slaves from the Bight of Benin alone between 1671 and 1710.4 Others have been more cautious. Abdoulaye Ly acknowledged that sailors from the United Provinces of the Netherlands remained the unchallenged masters of the slave trade to the West Indies until 205

206

James Pritchard, David Eltis, and David Richardson

1664, and evidence drawn from his book suggests that something less than a robust French slave trade emerged during the final quarter of the century. Jean Meyer asserted that the first known French slaver sailed only in 1672. Yet the current consensus is perhaps summarized by David Geggus’s assessment that “much of the [evidence for] seventeenth-century French traffic is missing,” thereby implying that the presence of an early French transatlantic slave trade remains a difficult proposition to accept.5 When we move beyond such general statements and search for well-grounded discussions of the direction and composition of the early French trade, the literature is almost silent. For the sixteenth century, a transatlantic slave trade carried on in French vessels is unlikely. The French had no possessions of their own in the Caribbean, and their isolated and temporary holdings in Brazil produced some wood but no plantation produce. No hard evidence exists of slaves crossing the Atlantic in French ships, and the single reference to slaves is from a Portuguese author pointing to French disruption of the slave trade in the Sierra Leone rivers. Smuggling of slaves into Portuguese or Spanish possessions may have been possible, but Brazilian and Latin American sources yield no indications of such activity, except for the Fleur-de-Lys in 1571. Perhaps the French sold slaves on the African side of the Atlantic—at São Tomé or the Cape Verde Islands—but no conclusive evidence has yet been produced of a sustained transatlantic business in the sixteenth or early seventeenth centuries. For the later seventeenth century, when French plantation colonies appeared, we have to decide what we mean by a “French slave trade.” Does it involve slaves arriving in the French possessions in the Americas and indeed in France itself, or should it be defined as slaves carried across the Atlantic in French vessels? Employing subsidies that lasted down to the Revolution, the French certainly focused on supplying their own colonies, but in the early period it is clear that the French did not themselves supply the majority of the slaves arriving in the French Caribbean. We now have many records of Dutch and English traders selling slaves to the French islands throughout the seventeenth century. Further, many non-French slave vessels fell into French hands during the extensive wars of the period, and French raids on enemy territory (particularly the British sugar islands) furnished considerable numbers of other slaves. Both activities augmented the slave populations under French control. Nor were French slave ships exclusively directed to French territory. Between 1702 and 1713, the French delivered thousands of slaves to the Spanish Americas under the asiento, and some slavers heading for the French sugar colonies were captured and diverted to enemy territory—though the French

The Significance of the French Slave Trade

did capture many more foreign ships than foreigners captured from the French during this period. A proper evaluation of either of these definitions of a French slave trade requires that they be considered together. Attempts to estimate the scale of transoceanic movement of humans around the globe fall into one of two categories: (1) direct observation in the form of shipping records and the impressions of contemporaries, or (2) inferences from population trends in the regions of origin or, more usually, arrival. For the period before 1716 we will employ both types of analysis. The 1999 CD-ROM (TSTD1) contained 159 records of French slaving voyages before 1716 and another twelve captured English slavers taken into French territories with slaves onboard. Archival work since 1999 has brought information to light on a further 165 slaving voyages under the French flag or disembarking in French ports, so that we now have a data set of 336 voyages with which to reassess the early French slave trade. We also have added new information for seventy (almost half ) of the voyages that were included in TSTD1. Nevertheless, unlike for the French slave trade after 1715, we still lack a reliable method of assessing how many voyages may have sailed without leaving any evidence behind. We do know that the vast majority of French slaving voyages delivered their slaves to French territories, however, and thus any estimate of the size of the early trade must begin with a review of population patterns in the early French Caribbean. We begin with a demographic approach. In order to derive an estimate of the numbers of slaves brought into the French Caribbean before 1716 from population data we use the following formula: Equation 1

M = P1 − P0 − P0R (1 − S) 2 + R 2

where M = Net number of arrivals, P1 = Population at end of the period, P0 = Population at the beginning of the period, S = Proportion of arrivals dying during the first year after arrival, and R = Rate of natural increase over the interval between the population counts.6 The derivation of the values of the four variables on the right side of equation 1 is shown in table 7.1, but the basic method employed in this derivation is to impute the rate of natural increase on the basis of the period 1716–50 and then to calculate estimated arrivals for the pre-1716 era. The justification for this approach is that, for any period between 1715 and 1803, there are now reasonably good data on both the populations of the French colonies and on slave arrivals, including those from the intra-American trade (unlike the pre-1716

207

208

James Pritchard, David Eltis, and David Richardson

era, when one of these variables—total slave arrivals—is unknown). If we solve Equation 1 for R, we derive equation 2: Equation 2

R = [M − M S − P1 + P0] / [0.5M S − 0.5M − P0]

Equation 2 enables the calculation of natural rates of population change when migration is known. For the post-1715 era, assuming a seasoning rate of 10 percent and values as shown in tables 7.1 (a) and 7.1 (b), we can thus use population data and slave arrivals to calculate R values for two broad slave-importing regions in the French Americas. These calculations, developed fully by Eltis and Lachance (this volume, chap. 12), yield some surprising results. While the literature stresses that black populations in the Caribbean improved their rates of natural increase over the course of the eighteenth century, it now appears that these rates did not in fact change much, and indeed may have worsened in the later period. Moreover, in neither period were negative values as severe as the earlier literature had taken to be the norm. Our purpose here, however, is not to dwell on this issue but to derive a value for R that we can use for the period before 1716. For this purpose, it would seem most useful to call into service the estimate for the period closest to pre-1716: 1716–50. To derive estimates of slave arrivals for the pre-1716 period, therefore, the estimate for 1716–50 can be substituted within equation 1, along with the population estimates for the pre-1716 era.7 Tables 7.1’s three parts present the results of these procedures. Table 7.1 (a) provides the population data for the French colonies for both 1640–1715 and 1716–50. Table 7.1 (b) employs equation 2 to estimate the rates of population decline for 1716–50 for the two groupings of French slave colonies: those in the eastern Caribbean (including Cayenne), and in Saint-Domingue. For the eastern Caribbean, the decline of the black population is taken to have averaged −0.7 percent per year, while that for Saint-Domingue is assumed to have been higher at −2.0 percent. Table 7.1 (c) completes the estimation process. It shows estimates of net migration into these colonies for the pre-1716 period calculated from equation 1 and using the values of R estimated from 1716 to 1750 and the population values in table 7.1 (a). As might be expected, the inflow of slaves into the eastern Caribbean was well ahead of that into SaintDomingue before 1685, of similar size between 1685 and 1700, and slightly less than that of Saint-Domingue in the first fifteen years of the eighteenth century. Overall, 106,155 slaves are estimated to have been added to the population of the French Americas before 1716, although, as we shall see, by no means all of these came directly from Africa or sailed in the holds of French ships. This compares to around a million who arrived between 1716 and 1793,

The Significance of the French Slave Trade

Table 7.1 Values Used to Estimate the Early French Slave Trade from Population Data and Estimates of the Volume of Slave Arrivals in the French Americas (a) Black Populations of the French Americas, 1640–1750 Years

Martinique

Guadeloupe

SaintChristophe

Cayenne

SaintDomingue

All Regions

1640 1670 1685 1700 1715 1750

100 6,393 10,611 15,266 29,247 67,300

100 4,482 5,257 6,855 14,200 46,274

100 4,901 5,294 722 132 —

— 50 1,507 1,418 2,563 5,542

— 100 2,939 9,082 35,451 169,600

300 17,596 27,293 35,043 83,308 290,466

(b) Annual Rates of Black Population’s Natural Increase in the French Americas, 1716–50 Windward Islands and Cayenne

Saint-Domingue

All Regions

−0.7%

−2.0%

−1.3%

Natural increase implied by table 7.1(a)

Note: In calculating rates of population increase, we estimated that slave arrivals in Martinique, Guadeloupe, Saint-Christophe, and Cayenne totaled 88,300 in 1716–50. Arrivals in Saint-Domingue were 178,200.

(c) Estimated Slave Arrivals, 1640–1715 Years 1640–70 1671–85 1686–1700 1701–15 1640–1715

Martinique

Guadeloupe

SaintChristophe

Cayenne

SaintDomingue

All Regions

7,270 5,415 6,280 17,441 36,406

5,172 1,278 2,294 9,063 17,807

58 1,686 26 1,443 3,213

5,549 873 −4,822 −619 981

0 3,586 8,534 35,629 47,749

18,049 12,838 12,312 62,957 106,156

Sources: See text.

a period of similar length. Of course, such an estimate is sensitive to changes in the estimated rate of attrition, but not excessively so. If we change the value of R in either direction by 50 percent for both groups of colonies, the effect is to raise or lower estimated slave arrivals by just under 5,000 slaves, or about 5 percent of the above estimate. Of course, if we assume the value of R was in

209

210

James Pritchard, David Eltis, and David Richardson

the range accepted by the earlier literature for attrition of the early Caribbean slave populations—say −5 percent per year—then the effect would be to raise estimates of arrivals threefold, a result that, as argued below, is not consistent with the wide range of sources we have consulted on shipping data. It would mean that three times the number of French slave ships that we know about would have made long voyages passing through several jurisdictions without leaving any traces of their movements. Finally, changing the value of S—the so-called seasoning rate—has a much smaller impact again. We can now turn to the second edition of the data set (TSTD2) and compare the volume of the transatlantic slave trade estimated from population data with that suggested by the surviving historical record of voyages. The new data set contains 142 French slave voyages that disembarked or intended to disembark slaves in French possessions. In addition to voyages that can be identified as sailing to French territory, seventy French voyages sold or could have sold slaves in the Americas, but no record of where they sold them has survived. It is assumed here that all these voyages went to the French Americas. We offer two reasons for this assumption. First, the best-documented part of the early French slave trade is undoubtedly the traffic to the Spanish Americas. This traffic was concentrated in the later part of the early period, and we have been able to draw on a variety of Spanish as well as French sources in our construction of the new database. It is not likely that any French voyages to the Spanish Americas occurred without some trace of those voyages coming to our attention, and thus the new database contains almost all French voyages carried out under the asiento. Second, French slaving voyages to parts of the Americas outside those under the control of the French or the Spanish must have been few indeed. French slave traders of this (or indeed, any other time) were clearly unable to make inroads into non-French slave markets in the Americas without official protection of the kind that the asiento provided. Indeed, the same point could be made even of the traffic to the French colonies, where French slave merchants received major subsidies through most of the ancien régime.8 Two voyages did venture into the huge Brazilian market in the early eighteenth century, when gold discoveries triggered a boom in slave prices, but one was subsequently detained by the Portuguese authorities. The only French ships that disembarked slaves in the English Americas were those vessels captured by the English on the high seas, and there is no recorded instance of a French slaver sailing into the English Americas on any basis other than its prior capture, though we would not argue that this never happened.9 In addition to the 212 French voyages, a further sixty-three slave ships arrived

The Significance of the French Slave Trade

in the French Americas in this period under the flag of nations other than France. Of this latter group, thirty-four were Dutch, twenty-five were English, three were Portuguese, and one sailed under the Brandenburg flag. Almost all the English, Portuguese, and Brandenburger vessels were conducted into French ports as a result of being captured while sailing for parts of the Americas that were not French. In sum, the database shows a total of 275 slave vessels disembarking slaves in French territory before 1716 (142 French definites, 70 French probables, and 63 non-French voyages). How many slaves did these vessels carry? Table 7.2 presents a breakdown of estimated numbers distributed by the nationality of the carrier and the period before and after the exclusif policy became effective in the Caribbean. These are based partly on records of actual arrivals and partly on imputed numbers where the total of slaves onboard was not recorded or where the outcome of the voyage is not known. Consistent with the discussion above, the French share is divided into two: one for voyages that we know disembarked or intended to disembark slaves in French territory, and the second for voyages that left no records of their destinations and that we have assumed disembarked slaves in the French Americas. For the whole of the pre-1716 period, Table 7.2 Slave Voyages and Slaves Disembarked in French Territory Prior to 1716 by Nationality of Carrier Nationality

Estimated Slaves Carried

Voyages

Before 1673 French recorded French assumed Dutch English Total

10 11 31 6 58

2,430 2,690 7,881 891 13,892

French recorded* French assumed Dutch English Portuguese Brandenburg Total

132 59 3 19 3 1 217

1673–1715 32,450 14,070 550 4,050 500 190 51,810

Source: TSTD2. *Includes two voyages selling slaves in France.

211

212

James Pritchard, David Eltis, and David Richardson

voyage records support an estimate of 65,700 slaves arriving in the French Americas, with foreign slavers playing a much larger role in the early period than the later—a question to which we will return. We have thus arrived at two independent estimates of the pre-1716 slave trade to the French Americas. One, based on estimates of the black population and attrition rates, is 106,156; the other, which rests on shipping-based data, totals 65,702. Which of these two estimates is closer to the reality, and are they in fact incompatible? We make no claim to having obtained information on every slave voyage that disembarked slaves in the pre-1716 French Atlantic world. For this early period, no simple test of the completeness of our shipping data is possible that would compare with what can be done with the post-1715 era. We do argue, however, that our combing over the last three decades of French, Spanish, English, Dutch, and Portuguese sources has been extensive. As an inspection of the source variables in the TSTD2 database will indicate, our research encompasses the records of every major port that dispatched slave voyages, as well as every major documentary series in the national archives of each of the major countries with an Atlantic presence. We think it most unlikely that the size of the French slave trade could have been more than 50 percent greater than what our shipping-based estimates will support—the ratio necessary to bring the two estimates into alignment. Much less do we think it possible that the size of that traffic was three times larger than our shippingbased estimate suggests—the ratio that would make the shipping data consistent with an attrition rate of −5 percent among the late seventeenth-century French Caribbean slave population, which some historians have posited. Rather, we believe the difference between the population-based and the shipping-based estimates can be explained not by hundreds of transatlantic voyages for which there is now no surviving record, but by a movement of slaves into the French Americas that originated in other parts of the Americas. It is not possible to derive an independent estimate of the size of the flow of slaves into the French Caribbean from other parts of the Americas, but enough anecdotal evidence exists to suggest a movement of several thousands. There were two sources for such flows: one was the capture of slaves in raids and the other was an intra-Caribbean traffic through which the French bought slaves from the Dutch and the English. A broadsheet published four years after the events claimed that the French had captured 15,000 slaves in 1666 from Antigua, Montserrat, and the English half of St. Kitts “which they carried to their own sugar Plantations.”10 While St. Kitts took years to recover, its labor force was mainly white at this time, and 15,000 was at least three times larger

The Significance of the French Slave Trade

than the black population of the three islands put together. One-tenth of this figure is more likely.11 In December 1693, French raids on the north coast of Jamaica resulted in the loss of 2,006 slaves, for which compensation was eventually paid.12 In Montserrat in 1712, six French cruisers carried off 1,500 slaves, two years after a similar attack had resulted in the loss of several hundred other slaves.13 Further raids on Nevis in 1706 and on the Bahamas two years later resulted in smaller transfers. French attacks on Essequibo in 1708, St. Eustatius in 1709, and Suriname in 1712 brought in several more hundred, though of these three Dutch possessions, it should be noted, only Suriname held large numbers of slaves.14 In total, the French obtained perhaps 7,000 to 8,000 slaves through attacks on English and Dutch possessions. There appears to have been no significant return flow from reciprocal attacks. The English and Dutch tended to replenish their slave-labor supplies from Africa, not from the rest of the Americas. French purchasing of slaves from the Dutch, English, and Danish islands was certainly a more important source of slaves than French military activity, although evidence of its size remains largely anecdotal. In 1692 the governor of Saint-Domingue allowed the planters of the island to buy slaves from any source. In 1713 the governor of Martinique threw open that island to foreign slave ships, and a year later an English estimate claimed that 3,800 of 6,100 Africans carried to the French colonies came from English and Dutch ships.15 These were wartime measures. Peacetime breaches of French colonial regulations were significant but are difficult to quantify. We know that St. Eustatius was a general emporium for the non-Dutch Leeward and Windward islands, and that its role was much bigger for the French than for the English islands.16 One route to an assessment is to ask how many Dutch slave vessels from Africa disembarked in St. Eustatius. The records for the Dutch slave trade are excellent for this period thanks to the work of Franz Binder, among others. They show that, in the half century prior to 1715, only 2,700 slaves from Africa were disembarked on the island, with an additional few hundred possible for those slavers without a known port of disembarkation. The Dutch carried some slaves to St. Eustatius from Curaçao, but the main purpose of Curaçao was, of course, to supply the Spanish Americas with slaves.17 Over a half-century period, St. Eustatius cannot have received as many as 8,000 slaves from all sources, and some of these must have either died or been sold to English planters. We can be certain that few indeed remained on St. Eustatius itself. The French slave colonies may have received 7,000 slaves from St. Eustatius, and perhaps more than this in the form of purchases from English islands, but given the prevalence of war

213

214

James Pritchard, David Eltis, and David Richardson

and the absence of records suggesting a large outflow on the English side, it could not have been much more. We suggest a total of 9,000. Finally, as Andrea Weindl observes in her chapter in this volume, the French also went to Danish St. Thomas for slaves. Buyers from Saint-Domingue obtained 900 slaves there between 1692 and 1694. Planters from Saint-Christophe purchased others. As the Brandenburgers established an agent in SaintDomingue in 1694, these were probably not the only slaves shipped from St. Thomas to Hispaniola. Overall, more slaves arrived from Africa at St. Thomas than at St. Eustatius, but on the Danish island the great majority of them remained to grow sugar for their Danish and Brandenburger owners. Outflows to the French from St. Thomas no doubt fell below the level of French purchases in St. Eustatius, perhaps amounting to 4,000 slaves. In summary, we allow for an upper-limit estimate of slave inflows into the French Caribbean from the rest of the Americas of 28,000 before 1716. This total comprises 8,000 from French raids, 7,000 purchases from St. Eustatius, another 4,000 from St. Thomas, and 9,000 purchased from English colonies—a large portion of which arrived in Martinique in 1713. The addition of this figure to our shipping-based estimate of 65,700 yields an estimated total of 93,700 slave arrivals in the early French Caribbean. It is thus possible to nearly reconcile the demography-based estimate of slave arrivals in the French Americas (106,156) with the shipping-based estimate of 65,702 slaves arriving directly from Africa (shown in table 7.2), if we allow for the additional movement of slaves into the French Caribbean from other parts of the Americas. We are certain that most early French slave-trading voyages are included in the new TSTD2 data set, but as we make no claim to completeness, it is possible that the actual total of slaves brought into the French Americas directly from Africa was 10 to 20 percent higher than the new data suggest. It is, however, more likely that movements of slaves from other parts of the Americas into the French islands made up a greater share than the one-third of total arrivals suggested here, and that the missing transatlantic data are accordingly not highly significant. Before leaving the issue of arrivals in the French Americas, it is worth asking what the new data may imply about patterns of development in the individual French colonies. Table 7.3 shows the destinations of all slave carriers arriving in the French Americas and provides a breakdown of arrivals in the period before and after the exclusif was fully effective. As we might expect, the eastern French Caribbean islands set the early pace, with Martinique apparently the leading importer. It is worth noting, however, that over two centuries of slave arrivals, Guadeloupe received most of its slaves

The Significance of the French Slave Trade

Table 7.3 Slaves Disembarked in the French Americas under All National Flags Prior to 1716 by Colony and Groupings of Years Before 1673 Colony Martinique Saint-Domingue Guadeloupe French Guiana Grenada Saint-Christophe Unspecified Total

1673–1715

Voyages

Slaves

Voyages

Slaves

27 3 9 3 1 4 11 58

6,554 213 2,533 759 200 943 2,690 13,892

78 51 6 11 1 9 59 215

19,814 12,645 1,203 1,976 210 1,860 14,070 51,778

Source: TSTD2. Note: The total for “French Territory” in 1673–1715 in table 7.2 includes two vessels arriving in France that are omitted here.

via Martinique, whose relative importance in the direct trade from Africa was greater in these years than in any other period, with the exception of the short period of British occupation in the early 1760s. Despite the large increase in slave arrivals at Saint-Domingue after Spain ceded the colony to France in the Treaty of Ryswick, Martinique remained the leading destination for slave vessels in the pre-1716 era, though Saint-Domingue assumed its overwhelmingly dominant role in the French slave trade to the Americas shortly thereafter. Other regions in the eastern French Caribbean were of small significance; most of them received their slaves from Martinique as part of the intra-American traffic, much as Guadeloupe did. A possible exception to this pattern was Saint-Christophe, which before 1670 received only slightly fewer slaves than Martinique according to table 7.1 (c), though in the shipping-based distribution shown in table 7.3 it lagged behind Guadeloupe. The likely explanation is that, while it was an important early market, Saint-Christophe received relatively few slaves directly from Africa. We shall now examine the size of the French slave trade to all parts of the Atlantic world. The new database provides evidence of fifty voyages under the French flag that carried slaves or intended to carry slaves into the Spanish Americas before 1716, with all but two of these sailing after 1697—most of them while the French held the asiento between 1701 and 1713. As previously noted, at least two French vessels tried Brazilian markets, and two carried slaves back to France for sale to the French government and to work in the galleys.

215

216

James Pritchard, David Eltis, and David Richardson

Table 7.4 summarizes the destinations of all French slave voyages to the Americas prior to 1716. It draws on table 7.2 for the French voyages to French territory and adds fifty-one voyages and 12,800 slaves that were carried in French vessels to the rest of the Atlantic world. We do not know if all these voyages actually arrived in the Americas with slaves, though all set out with the intention of delivering slaves to non-French parts of the Americas. Nevertheless we have assumed that all these vessels completed their voyages in the interest of deriving an upper-limit shipping-based estimate of the volume of slave trading by early French traders. Table 7.4 suggests that the French carried off 86,800 slaves from Africa before 1716 and disembarked 64,900 of these in various parts of the Americas (65,000 if we include the two small groups taken to France). If we allow for the missing voyages suggested by the total arrival figure based on vital rates—about 10 percent—then 86,800 departures from Africa and 65,000 arrivals become 95,800 and 71,600, respectively. Table 7.4 Slave Voyages and Slaves Embarked and Disembarked by French Transatlantic Slave Vessels Prior to 1716 by Region of Disembarkation Embarked Region

Voyages

Slaves

Disembarked Voyages

Spanish America (asiento) 9,888 27 5,413 17 705 2

Slaves

Spanish Main Río de la Plata Cuba

27 17 2

Martinique Saint-Domingue Guadeloupe French Guiana Grenada St. Kitts France Unspecified

71 49 7 10 1 2 2 74

French Territory 24,280 71 15,912 49 2,141 7 2,382 10 275 1 537 2 73 2 23,627 70

18,351 12,397 1,633 1,790 210 422 67 16,766

Bahia English America São Tomé Total

2 2 1 267

Other Regions 490 2 844 2 240 1 86,807 263

374 774 155 64,915

Source: TSTD2.

7,693 3,738 545

The Significance of the French Slave Trade

Finally, we need to draw attention to one other possible source of undercounting in our reckoning of the early French slave trade. In addition to the vessels that bought or were intent on buying slaves before 1716, we have records of eighteen voyages that were involved in produce trade. Vessels returning directly to France with African produce were a rarity before the nineteenth century. Indeed, we know of only twenty-seven such voyages during the entire period prior to abolition of the slave trade in 1807, compared with nearly 1,000 English merchant ships that were so engaged (and for which we have documentary evidence) over the same period. Eighteen of these twenty-seven known produce vessels sailed before 1716, with no less than fifteen of these voyages occurring between 1670 and 1685. Three-quarters of the voyages whose African place of trade can be identified traded in Senegambia. The significance of this activity is twofold. Like other European slave-trading powers, the French were initially interested in gold, hides, and gum, and it is likely that, as with the Dutch and the English, the value of produce traded exceeded that of slaves for a few years. Yet French exclusion from the Gold Coast heartland of the gold trade ensured that their dependence on slaves to sustain trade with Africa evolved quickly, predating the points at which English and Dutch trade with Africa switched from produce (including gold) to slaves. The second point is that we cannot always be certain whether a vessel that set out from a French port for Africa and left no further trace in the historical record was trading for produce or for slaves. Whenever possible, we have used the previous and subsequent records of the vessel or the captain to identify the intent of such voyages, but there remains an element of doubt for as much as 8 percent of our sample. Of the eighteen early produce voyages, six are known to have returned directly to France, but the outcomes of the remaining twelve are not known. Some scholars may want to add these twelve to the slave-ship pool, but our primary point is that these judgment calls cannot have a major impact on our conclusions. Implicit in these estimates of departures and arrivals is a high rate of shipboard mortality. French slavers left Africa with similar numbers of slaves onboard before 1716 as they did after that year (326 on average, compared with 328 after 1715), but the earlier group experienced almost double the number of shipboard deaths than did their later counterparts.18 Almost one-quarter of those embarked before 1716 died before reaching the Americas, compared with one in eight of those carried off between 1716 and 1792. This discrepancy partly reflects declining shipboard mortality over time, but the declining overall trend in mortality was not so large that it can explain all of the differential,

217

218

James Pritchard, David Eltis, and David Richardson

and it seems that the latter may also have been due to the lack of experience of French crews at this time compared with Portuguese or English slave traders. Such a rate is similar to that found on Portuguese and Dutch vessels in the second quarter of the seventeenth century. Typically, a French slaver arriving in the Americas before 1716 carried an average of only 242.8 slaves (number of observations = 165; standard deviation = 131.5). The relative inexperience of the French is underlined if we return to table 7.2 and note that their early slave colonies were supplied with slaves by the Dutch. Yet the sugar revolution came to Martinique and Guadeloupe about the same time as it arrived in Barbados, and while it progressed more slowly, by 1700 the French eastern Caribbean islands were producing about one-third of the sugar of their English counterparts. For the period as a whole, of course, the French dominated the supply of slaves to their own colonies, supplying about threequarters of the slaves brought to the French Americas before 1716. A further one in ten arrived in foreign vessels that had been captured by the French. Yet these ratios disguise two important patterns. First, in comparison with the English and Portuguese colonies, the overall ratio of seven out of ten is relatively low. Both the major transatlantic slave powers relied far less on captured slave vessels for their slaves than did the French, and both drew on foreign slave ships for trivial quantities—less than 1 percent of the slaves they used even in the first years of settlement. Second, there is a pronounced temporal pattern in the arrival of foreign slave vessels in the French Americas. Prior to 1664, all slaves brought to the French Americas directly from Africa arrived in Dutch vessels, except for two groups brought on French ships. The inception of Colbert’s exclusif did not change this situation immediately— not drastically, at least. Between 1664 and 1669, one in eight slaves still arrived on Dutch vessels according to our estimates, and two vessels under the French flag in these years are known to have set out from a Dutch port. Substantive change began in 1670. In that year and in 1671, French slave traders for the first time assumed the majority role in the supply of slaves to their own colonies. From 1673 on, the only foreign slave vessels carrying slaves into French ports were captured vessels. In this, too, the French were quite different from their major rivals. The British were already dominating the supply of slaves to their own colonies before the 1660 Navigation Act, and the Portuguese excluded foreign vessels from Brazil from their first settlement. It is quite possible that the early French Caribbean fell somewhere between the English and Spanish in the pattern of slave arrivals and the pace of development of their respective sugar sectors. Next to the Portuguese, whose

The Significance of the French Slave Trade

possessions were the closest of all plantation regions in the Americas to major African slave markets, the English had the cheapest possible access to slaves from Africa, simply because their slave traders were more efficient initially than their French counterparts. At the other extreme, the evolution of the Spanish Caribbean as a sugar producer was delayed until the late eighteenth century by a combination of major limitations on slave supplies, dependence on the asiento, and a premodern land tenure system. In the eastern Caribbean, the French controlled natural resources to rival those of the English islands, but they were unable to exploit them quickly, owing to a lag in the development of an efficient national slave trade. The exclusion of Dutch slave ships after 1670 appears to have been of greater significance to French planters than the similar thrust of the Navigation Acts was to their English counterparts. Perhaps, in the absence of the exclusif, the gap in plantation output between the English and the French Caribbean would have been much smaller than two-thirds, and the French planters would not have been expected to produce sugar with relatively restricted access to labor from Africa. This last point is illustrated by a simple comparison of the numbers of slaves disembarked in the Caribbean by the English and French before 1716. English slave traders were already selling nearly 3,000 slaves a year in Barbados in the 1640s. By the 1660s they had raised that number to an average of 4,500 and, led by the much-maligned Royal African Company (which could be considered a model of efficiency in view of the operations of the various French chartered companies), were disembarking an average of 10,000 a year in the early 1680s. While the numbers fell back to 8,500 for the remainder of the century, the English then carried 180,000 Africans across the Atlantic, or nearly 12,000 a year between 1701 and 1715. In contrast, the French managed an occasional voyage before 1670 and, despite the exclusion of the Dutch, averaged less than 1,000 slaves a year disembarked between 1671 and 1700. Thereafter the French slave trade increased dramatically, but even when including the numbers of Africans delivered to French and Spanish colonies together—as previously mentioned, the French carried nearly 14,000 slaves to other parts of the Americas after 1697—we calculate an annual average between 1701 and 1715 of 4,600 per year, much less than half of the English transatlantic slave trade.19 The same picture of French inability to supply the labor that their planters required to compete internationally emerges from a closer comparative look at some of the ratios associated with profitability. We have no direct measure of the financial returns at this point, although it is clear from the constant

219

220

James Pritchard, David Eltis, and David Richardson

reorganizations of the various French companies charged with operating a transatlantic slave trade in the seventeenth century that profitability, except for the occasional voyage, remained a chimera. On average, French ships carrying slaves in the seventeenth century were 82 percent larger than their English counterparts, were manned by crews more than 40 percent greater, and they carried 50 percent more guns (all differences statistically significant at the 5 percent level). Yet the French vessels disembarked on average only 15 percent more slaves than their English counterparts.20 French slaving voyages from France to Africa, the Americas, and back to France lasted slightly longer than English voyages sailing from London, even though over half the French slave vessels sailed by far the shortest of the triangular trade’s round trips—the one via Senegambia—and most sailed from ports outside the English Channel (a distinct advantage given that the prevailing westerlies in the channel added many weeks to London-based voyages). Many of the French slave ships were in fact naval vessels seconded into service on behalf of one of the chartered companies. It is thus not surprising that French slave ships had slavesper-crew and slaves-per-ton ratios below what was standard at this time on English slave ships, the great majority of which belonged to or were chartered by the supposedly inefficient Royal African Company. The problems of the early French slave trade are also reflected in the geographic distribution of the French ports from which slave vessels departed. Prior to 1670 there was no single French port that could compare with London, Lisbon, Bahia, or Pernambuco as centers of organizational expertise. English and Portuguese slave traders concentrated their operations in such ports and by 1670 had dispatched hundreds of voyages from them to Africa and the Americas.21 By contrast, Dieppe, Honfleur, La Rochelle, and Le Havre had all sent out just one or two voyages by 1670, while, as noted, other French slavers left from Dutch ports. With the attempted establishment of a series of monopoly companies after 1670 and heavy pressure from Colbert and his successors, the center of the French slave trade became La Rochelle.22 Between 1671 and 1700, 60 percent of all voyages under the French flag left from La Rochelle, although this still amounted to an average of just over one voyage per year. As direct state involvement declined after 1700 and the French slave trade underwent rapid expansion, Nantes quickly took over as the leading center, dispatching an average of four slaving voyages a year between 1701 and 1715. La Rochelle, by contrast, took no part in the expansion of the business and continued to send out ventures at its pre-1700 rate. During the rest of the slave-trade era, La Rochelle was overtaken by Bordeaux and Le Havre as well

The Significance of the French Slave Trade

as Nantes and overall was responsible for about one in ten of French slaving voyages after 1700. Almost all western European governments were involved to some degree in the seventeenth-century transatlantic slave trade, but one would have to examine the records of the Duchy of Courland or the Danish crown to find a match for the degree of involvement and, indeed, the extent of misdirection on the part of the French government. The state also had a major influence over where in Africa the early French slave ships obtained slaves. Table 7.5 shows the regions of trade for the slaves who were embarked from Africa for the French colonies, according to the revised data set. Once again, we have included the intended region of African trade where this was the only information available, and we have assumed that voyages with unknown outcomes did in fact leave Africa with slaves. Before 1670, and prior to effective intervention by the French state, the Dutch slavers that supplied the French colonies ranged over all the major markets on the western coast of Africa. Table 7.5 indicates that one-third came from the Slave Coast of the Bight of Benin, but generally the slaves carried to early Martinique and Guadeloupe must have come from extremely diverse backgrounds. This pattern once more separates the French transatlantic slave system from its English and Portuguese counterparts.23 With the inception of the French monopoly companies and their focus on Senegal, there came a pronounced shift north in the provenance of slaves sent to the French colonies that lasted from 1671 to 1700, though with a gradually decreasing effect over time. Senegambia, which in the Dutch-dominated era had supplied only 18 percent of Frenchbound slaves, now accounted for more than half the slaves disembarked on French territories. Of the major slave-importing regions of the Americas, only early Virginia came close to matching this concentration on the Upper Guinea coast. Table 7.5 shows the Slave Coast supplying a further 27 percent, so that for thirty years four out of five slaves came into the French Americas from two relatively short, though widely separated, stretches of the West African coast. The weakening state controls in the post-1700 era brought yet a third pattern to the African supply of French slaves. The Slave Coast, whose importance had grown steadily from the pre-1670 era both for the French and indeed for all transatlantic slave traders, now became the dominant regional source, supplying over half the slaves leaving for the French territories between 1701 and 1715. It is, however, worth noting that the Slave Coast’s share of all African slave departures was also more than half the peak share the region attained in any fifteen-year period during the entire era of the transatlantic slave trade. Because 1702 to 1715 was the only period when the French sent significant numbers of

221

222

James Pritchard, David Eltis, and David Richardson

Table 7.5 Distribution of Slaves Arriving in the French Americas Prior to 1716 by African Region of Origin Region Senegambia Sierra Leone– Windward Islands Gold Coast Bight of Benin Bight of Biafra West Central Africa Total

Before 1670

1671–85

1686–1700

1700–15

All Years

2,002 226

5,443 0

3,630 180

3,409 1,353

14,484 1,759

0 4,488 2,608 2,700 12,024

706 1,845 712 0 8,706

400 2,700 326 1,076 8,312

2,184 11,885 1,805 2,331 22,967

3,290 20,919 5,451 6,107 52,010

Source: TSTD2. Totals may not add up due to rounding.

slaves to the non-French Americas (under the asiento), these slaves too were drawn overwhelmingly from the Bight of Benin. Almost 60 percent of the Africans that the French carried to the Spanish Americas embarked on the Slave Coast. A further 12 percent—mainly those taken to Buenos Aires rather than to Central America and the Caribbean—embarked in Angola. Thus the 1701–15 distribution of slaves arriving in both the French colonies and the Spanish Americas was not markedly different from those arriving in other major slave-importing areas of the Americas at this time. Within the Bight of Benin, Ouidah was by far the most dominant embarkation point; in this, too, the French trade reflected the pattern that held for the wider traffic carried on by other national groups of traders. Between 1701 and 1715, Ouidah was close to reaching its maximum share of total departures from both the Bight of Benin and from Africa as a whole to all the Americas. Probably 90 percent of the French vessels trading on the Slave Coast obtained their slaves there at this time.24 But it is also true that this was the period when the French established their long-term connection with the Slave Coast. In the aftermath of the Dahomean conquest of Ouidah in 1727, the Dutch and the English drew not only on other ports but also on other regions. The French, by contrast, continued to use the Slave Coast as a major source of slaves down to the outbreak of the Saint-Domingue revolution in 1791. Before 1700, Ouidah had been much less important to the French, although it was still the second most frequented single African embarkation point. Senegambia was easily the dominant region in this early period, but we lack information on the specific ports of trade within this region. Of those we know about, Saint-Louis, usually designated as Senegal in the records at this

The Significance of the French Slave Trade

time, accounted for 20 percent of all slaves on French vessels. Gorée accounted for another 8 percent, but the French were also drawing on the Southern Rivers region—Bissau and Cach˜eu—as well as the Gambia River and Portudal. This was probably the only African provenance zone where the French offered a serious rivalry to English slave traders. Finally, during the Dutch period before 1670, slaves coming into the French Americas were carried from those ports where the Dutch were dominant. Ardra in the Bight of Benin supplied 41 percent of all slaves, with the rest distributed evenly around ports where the Dutch had cultivated trading contacts in Senegambia, the Bight of Biafra, and West Central Africa. No other single port of embarkation in this early period came close to matching Ardra’s importance. To summarize what we now know about the African coastal provenance of slaves in the early French slave trade, it seems that before 1670 and after 1700 the French pattern was about what we would expect in terms of the supply structure of the larger transatlantic slave market, though only after 1670 were the French themselves carrying the slaves. From 1671 to 1700, on the other hand, the slave provenance pattern in the French slave trade was rather exceptional when set against the larger picture. Not coincidentally, given the French focus on the relatively minor provenance zone of Senegambia, the volume of slaves delivered by the French grew quite slowly, if at all, between 1670 and 1700. Overall, the attenuated form that the early French slave trade now appears to have taken offers interesting insights into some larger issues associated with the rise of slavery in the Americas. From an imperialist perspective, it is hard not to conclude that the French state missed a major opportunity to catch up and surpass their English rivals during this period—something that eventually happened in the second half of the following century. While the consequences were not as severe for the French planters as Spanish policies were for plantation development in the Spanish Caribbean, it seems clear that in counterfactual terms a more liberal French policy toward the foreign slave trade would have substantially increased the slave labor supply to the French Caribbean and accelerated the growth of plantation output. The contrast between before and after 1670 suggests that it would not have taken much on the part of the French government to have kept the slave trade on a strong growth path in the last thirty years of the seventeenth century. We would argue, indeed, that given the heavy Dutch involvement before 1670, no action at all would have ensured such an outcome, though such a conclusion must remain tentative in the absence of good data on relative slave prices in the pre-1700 English and French Caribbean islands. If our argument is valid, we would expect slave

223

224

James Pritchard, David Eltis, and David Richardson

prices to be somewhat higher behind the protective walls that Colbert placed around the French slave trade than in the open-market English Caribbean. NOTES

1. Robert Louis Stein, The French Slave Trade in the Eighteenth Century: An Old Regime Business (Madison, Wis., 1979), 10. 2. P. E. H. Hair, “A Note on French and Spanish Voyages to Sierra Leone, 1550–1585,” History in Africa 18 (1991): 137–38. See also idem, “Some French Sources on Upper Guinea, 1540–1575,” Bulletin de l’Institut Français d’Afrique Noire 31B (1969): 1030–34. 3. Herbert S. Klein, The Middle Passage: Comparative Studies in the Atlantic Slave Trade (Princeton, 1978), 175. 4. Patrick Manning, “The Slave Trade in the Bight of Benin, 1640–1890,” in Henry A. Gemery and Jan S. Hogendorn, eds., The Uncommon Market: Essays in the Economic History of the Atlantic Slave Trade (New York, 1978), 117. 5. David Geggus, “The French Slave Trade: An Overview,” William and Mary Quarterly 58 (2001): 119–38. 6. David Galenson derived this formula in White Servitude in Colonial America: An Economic Analysis (Cambridge, 1981), appendix G, and used it to calculate levels of immigration for Europeans and Africans until 1780. 7. These two propositions are argued at length in David Eltis and Paul Lachance, “The Demographic Decline of Caribbean Slave Populations: New Evidence from the Transatlantic and Intra-American Slave Trades,” this volume (chap. 12). This chapter is also the source of our estimate of the rate of population decline in the French colonies between 1716 and 1750. The best discussion of the rates at which West Indian slave populations decreased is by J. R. Ward, British West Indian Slavery, 1750–1834: The Process of Amelioration (Oxford, 1988), 121–29. Eltis and Lachance argue that the rates Ward calculated for the late eighteenth century probably hold for the earlier period as well. 8. See Stein, French Slave Trade, 14, 33, and 40–41; James Pritchard, In Search of Empire: The French in the Americas, 1670–1730 (Cambridge, 2004), 216–18. 9. See David Eltis and David Richardson, “Productivity in the Slave Trade,” Explorations in Economic History 32 (1995): 465–84, for an analysis of the relative efficiency of French and English slave traders. 10. Anon., “The State of the Case of the Sugar Plantations in the Americas” (London, n.d. but ca. 1670), in Rawlinson Collection, Bodleian Library, Oxford, A 478, p. 63. 11. For comments on St. Kitts at the time of the raid, see “An Acct given by Sir Charles Wheeler late Governor of ye Leeward’s Islands” (n.d. but ca. 1672), BNA, CO153/1, p. 60. 12. Nigel Tattersfield, The Forgotten Trade: Comprising the Log of the Daniel and Henry of 1700 and Accounts of the Slave Trade from the Minor Ports of England, 1698–1725 (London, 1991), 280, 436. 13. Regarding the 1712 raid, see William Frye, Montserrat to the Royal African Company, August 15, 1712, BNA, T70/2, p. 36. 14. Pritchard, In Search of Empire, 265, 317–18, 380–83, and 389–90. 15. Ibid., 223; David Eltis, The Rise of African Slavery in the Americas (Cambridge, 2000), 216.

The Significance of the French Slave Trade

16. Wim Klooster, Illicit Riches: Dutch Trade in the Caribbean (Leiden, 1998), 18–71; idem, “Dutch Trade, Capital and Technology in the Atlantic World, 1595–1667” (paper presented at the annual meeting of the American Historical Association, Seattle, 1998). 17. Calculated from TSTD2. None of these arrivals occurred in the aftermath of the exclusion of the Dutch transatlantic slave ships from French islands in the 1660s. The first recorded direct disembarkation from Africa was not until in the mid-1680s. 18. Some information about numbers of slaves onboard exists for 165 vessels (127 at point of arrival, 83 at point of departure, and 45 with information on both). From the latter group we have calculated a voyage mortality rate of 24.6 percent (standard deviation = 24.1). All statistics for post-1715 have been calculated from the 1999 CD-ROM edition of the database. 19. The French averages have been calculated from the last column of table 7.1 (c), with the 1701–15 average derived by adding together row 4 of table 7.1 (c) and row 2 of table 7.2. The English data have been taken from Eltis’s chapter in Robert Paquette and Stanley L. Engerman, eds., The Lesser Antilles in the Age of European Expansion (Gainesville, 1995), 191–215, as later modified in Eltis, “The Volume and Structure of the Transatlantic Slave Trade: A Reassessment,” William and Mary Quarterly 58 (2001): 17–46. 20. The comparisons between the English and the French are summarized in the following table. For conversions of tonnages to an approximately common standard, see the introduction to the CD-ROM (TSTD1). French

English

Variables

Number Standard Mean of Voyages Deviation

Number Standard Mean of Voyages Deviation

Tonnage Crew at outset Guns Slaves disembarked

317.5 48.7 21.2 261.5

173.9 34.7 14.1 228.1

28 42 26 36

102.0 29.5 10.0 160.9

408 109 228 589

99.7 16.2 8.2 136.0

Source: TSTD2.

21. According to TSTD2, London and Lisbon alone had dispatched at least 403 slaving voyages by 1670. 22. For an evaluation of the impact of the French state on the slave trade at this time, see James Pritchard, “An Incidental Slave Trade: The French in Africa during the Seventeenth Century” (unpublished paper, 2004). 23. See Eltis, Rise of African Slavery, chap. 9, for the English colonies, and Daniel Barros Domingues da Silva and David Eltis, “The Slave Trade to Pernambuco, 1561–1851,” this volume (chap. 3), for early Pernambuco patterns. 24. A more precise estimate is difficult because French records at this time began to use “Gold Coast” when designating the African point of departure, and this term could apply to almost any part of West Africa south of Sierra Leone. Certainly some and possibly most French vessels reported from the “Gold Coast” did in fact receive their slaves at Ouidah.

225

226

James Pritchard, David Eltis, and David Richardson

Appendix 7.1 Slave Voyages in TSTD2 Used in the Derivation of Tables 7.2, 7.3, and 7.4, Listed by the Unique Voyage Identity Number (see http://www .slavevoyages.org) French Arriving in French Territory 21,559; 21,560; 21,561; 21,562; 30,003; 30,004; 30,005; 30,007; 30,020; 30,021; 30,022; 30,023; 30,031; 30,032; 30,033; 30,034; 30,041; 30,042; 30,043; 30,044; 30,050; 30,051; 30,052; 30,053; 30,061; 30,062; 30,065; 30,067; 31,839; 31,882; 31,883; 32,023; 32,030; 32,451; 32,452; 32,454; 33,113; 33,343; 33,364; 33,365; 33,623; 33,628; 33,630; 33,631; 33,653; 33,654; 33,684; 33,725; 33,746; 33,747; 33,748; 33,749; 33,756; 33,757; 33,758; 33,760; 33,768; 33,769; 33,771; 33,780; 33,881; 33,882; 33,883; 33,885;

21,593; 30,008; 30,024; 30,035; 30,045; 30,054; 30,068; 32,024; 32,849; 33,525; 33,632; 33,731; 33,750; 33,761; 33,782; 33,886;

21,871; 30,010; 30,027; 30,036; 30,046; 30,056; 30,075; 32,025; 32,850; 33,527; 33,641; 33,732; 33,751; 33,762; 33,783; 33,889;

21,872; 30,013; 30,028; 30,037; 30,047; 30,057; 31,831; 32,027; 33,005; 33,528; 33,642; 33,733; 33,752; 33,763; 33,785; 33,890

21,873; 30,014; 30,029; 30,039; 30,048; 30,058; 31,832; 32,028; 33,092; 33,531; 33,650; 33,743; 33,754; 33,766; 33,814;

30,002; 30,015; 30,030; 30,040; 30,049; 30,060; 31,833; 32,029; 33,093; 33,579; 33,651; 33,745; 33,755; 33,767; 33,835;

French, No Destination 11,605; 11,608; 21,591; 33,634; 33,635; 33,636; 33,646; 33,647; 33,681; 33,734; 33,735; 33,736; 33,786; 33,787; 33,788; 33,797; 33,798; 33,799; 33,887; 33,888; 33,894; 33,904; 33,906; 33,907;

30,001; 33,638; 33,683; 33,739; 33,790; 33,803; 33,898; 33,909;

30,018; 33,639; 33,726; 33,740; 33,791; 33,831; 33,899; 34,718;

30,019; 33,643; 33,727; 33,744; 33,792; 33,837; 33,900; 44,266

31,838; 33,644; 33,729; 33,781; 33,793; 33,880; 33,901;

33,633; 33,645; 33,730; 33,784; 33,796; 33,884; 33,902;

21,860; 33,637; 33,682; 33,737; 33,789; 33,800; 33,896; 33,908;

Non-French to French Territory 9,564; 9,589; 9,593; 9,949; 10,028; 11,360; 11,379; 11,390; 11,393; 11,396; 11,400; 11,584; 11,590; 11,610; 11,644; 11,786; 11,809; 14,931; 15,079; 16,049; 16,059; 20,419; 21,009; 21,103; 21,115; 21,136; 21,143; 21,195; 21,259; 21,369; 21,558; 21,971; 21,998; 21,999; 22,000; 33,652; 33,838; 33,839; 33,893; 44,115; 44,123; 44,127; 44,138; 44,150; 44,217; 44,228; 44,229; 44,237; 44,240; 44,243; 44,248; 44,254; 44,256; 44,257; 44,262; 44,268; 44,271; 44,284; 44,285; 50,339; 94,841; 98,836; 98,837; 98,838 (continued)

The Significance of the French Slave Trade

Appendix 7.1 (continued) French to Non-French Territory 11,353; 21,878; 30,006; 30,009; 30,038; 32,026; 32,450; 33,501; 33,508; 33,509; 33,510; 33,511; 33,517; 33,519; 33,624; 33,625; 33,773; 33,775; 33,776; 33,777; 33,818; 33,834; 33,891; 33,892; Source: TSTD2 as of November 2006.

30,011; 33,503; 33,512; 33,626; 33,778; 33,903;

30,012; 33,504; 33,513; 33,627; 33,779; 33,905

30,016; 33,505; 33,514; 33,655; 33,801;

30,025; 33,506; 33,515; 33,657; 33,804;

30,026; 33,507; 33,516; 33,772; 33,810;

227

Chapter 8 The Dutch in the

Atlantic World: New Perspectives from the Slave Trade with Particular Reference to the African Origins of the Traffic Jelmer Vos, David Eltis, and David Richardson

The historiography of Dutch overseas expansion often appears to embody two somewhat inconsistent positions. Among mainly Dutch scholars, certainly until recently, the Dutch role in Asia has received the most attention and has been seen as far more important than its Atlantic counterpart.1 In Asia the Dutch displaced the Portuguese relatively quickly, and from the late sixteenth century dispatched large volumes of people (albeit less than half of them Dutch and few of them female) as well as capital to establish an informal empire of trade unrivaled in European terms until the development of British India. Many non-Dutch scholars, by contrast, have seen the Dutch role in the early modern Atlantic world as being, if not more important than its Asian counterpart, critical to the development of the export-based economies of the Americas. Thus, after their massive and ultimately unsuccessful attempt to wrest Brazil from Portuguese control, the Dutch applied the expertise they had thereby obtained to the rest of the American colonies. From this perspective, the Dutch introduced the plantation complex into the Caribbean by making their technological skills, capital, and merchant fleets available 228

The Dutch in the Atlantic World

to the British and the French. They also quickly became the chief suppliers of slave labor to the Spanish Americas and established a flourishing set of plantation colonies of their own centered on Suriname. These two positions are not mutually exclusive and perhaps differ more in emphasis than in substance. The early modern Dutch economy probably had the largest per capita income in early modern Europe and was able to harness labor resources and expertise from a population that extended far beyond the relatively small pool of Dutch speakers.2 Amsterdam was the world’s leading capital market, and the Dutch merchant fleet was likewise the largest and most efficient anywhere. Advantages such as these made Dutch preeminence possible in both Asia and the Atlantic world at the same time. A fresh perspective on these issues has been made possible by substantial new data on the Dutch transatlantic slave trade. The Dutch were the first non-Iberian power to enter the business in any significant way, and while carrying slaves to the New World was only one aspect of establishing the plantation complexes of the Caribbean, it was of central importance. More important, given the surviving documentation, it is relatively easy to measure. Along with the post-1710 French traffic, the Dutch slave trade was the first to be documented in close to its entirety thanks to the extensive work of Johannes Postma.3 In this chapter we build on this work in three ways: First, we have new information for 95 percent of the data set that Postma published in 1990, including the merging and deleting of 93 of Postma’s original voyages and the addition of 122 new ones. Many of these decisions are consistent with changes that Postma himself has recently made to his data set. Indeed, we have worked together with Postma on many of these changes. Second, it should be noted that much of the new data pertains to new information on the African trading locations of these Dutch slaving vessels. The original Postma data set provided some information on African places of trade for about half the slaves carried on Dutch vessels sailing after 1674. The present data set offers such information for 75.4 percent of the Dutch slaving voyages in this period. Third and most important, we have added almost 400 new voyages for the period before 1674. The resulting new set will enable scholars to shed new light on ongoing debates in the history of the Atlantic slave trade. The new data have been drawn from a wide variety of sources. To a large extent we have built on the research of other scholars. Ernst van den Boogaart provided us with more than 100 slaving voyages to Dutch Brazil, while Franz Binder shared the results of his exhaustive work on the Dutch Atlantic trade in the quarter century from 1650 to 1675.4 These two scholars have thus provided

229

230

Jelmer Vos, David Eltis, and David Richardson

the core of the new information we now have on the seventeenth-century Dutch slave trade. Ruud Koopman offered his transcriptions of documents from the 1650s and 1660s from the Amsterdam municipal archives, adding substantially to the data provided by Binder. Finally, Henk den Heijer allowed us to use his original data set of slaving voyages of the second West India Company (1674–1740), hereafter WIC.5 In addition, we have consulted many published sources, among which the works of Klaas Ratelband, Albert van Dantzig, and the sometimes neglected Cornelis Goslinga were especially useful.6 Our own archival research has predominantly focused on the records of two eighteenth-century trading companies—the Rotterdam-based firm of Coopstad and Rochussen and the Middelburgse Commercie Compagnie (MCC)—although major state archives have also been consulted. The records of the MCC have not only enabled us to track down details of the slaving voyages this company organized, but also yielded rare information on the trading locations of numerous noncompany ships. This material is extensive, and work on it continues. In 1990 Postma published details of 1,211 Dutch slave voyages that sailed after 1673 divided into WIC vessels—broadly to 1730—and free-trade ships, mostly after 1730. He has since revised his database. Through eliminating voyages that had been counted twice and identifying vessels that went to Africa but then returned directly to the Netherlands with African produce rather than slaves, he recently argued for a total of 1,001 Dutch slaving voyages from 1674 onward.7 We have modified this total further after additional checking and drawing on a few new sources, and we now have a Dutch database of 1,173 voyages from 1674 onward. For the WIC period we have basically come to the same results as the new Postma compilation, since we and Postma both have drawn heavily on Den Heijer’s carefully constructed database of WIC slave and produce traders. We differ from Postma, however, in the period after 1730. In 1990 Postma counted as a slaving voyage any ship setting out for Africa for which he had no record of arrival in the Americas, if the voyage was projected to last longer than sixteen months. As subsequent research by Den Heijer and Reinders Folmer-van Prooijen showed, many of these vessels were actually employed in the African produce trade and never carried slaves to the Americas.8 In 2003, therefore, Postma swapped his former “maximalist” approach for a “minimalist” one: from his database he deleted all Africa-bound voyages for which he has found no record of arrival in the Americas, on the assumption that they may have been produce traders. This amounted to a total of 207 voyages. By contrast, in cases of doubt our method was to consider

The Dutch in the Atlantic World

both the history of the ship or captain in question, if known, and the African destination of the vessel. We assumed that all Dutch vessels sailing to Angola were going there at least in part for trade in slaves. We have strong reasons to believe that in the eighteenth century few if any vessels sailed to the Angola region to trade in produce only. Den Heijer has pointed out that, in the period 1674–1740, the WIC sent only four vessels to the Loango coast to purchase something other than slaves (mostly dyewood); the last of these sailed in 1722.9 Furthermore, it seems that the MCC and Coopstad and Rochussen, the only two free-trade companies whose records have survived more or less intact, did not send any vessel to Angola specifically for produce trade. We have also been able to identify American destinations for some of the vessels in question. All in all, therefore, the volume of the Dutch slave trade after 1730 comes out larger in our database than in Postma’s. For the period before 1674, we had records of only sixteen slaving expeditions to include in the 1999 CD-ROM (TSTD1). To these we have now added 383 new entries, creating a total of 399 voyages for the era preceding the formation of the second Dutch WIC. Adding these to the post-1673 material means that we are working with a total of 1,572 records of slaving voyages that sailed under the Dutch flag between 1596 and 1829.10 This data set is the first to attempt to cover the full span of Dutch Atlantic slaving activities and the first to draw together and consolidate the databases and the superbly detailed work of Binder, Van den Boogaart, Postma, and Den Heijer, conducted over the past forty years. To maximize the analytical potential of these new data, we have made some assumptions that should be spelled out. If a vessel set out on a voyage and then slipped from the historical record, we have made the assumption that it reached the Americas with slaves.11 Further, we know the number of slaves onboard on arrival in the Americas for only 1,065 voyages, and for embarkation in Africa we have precise numbers for only 629 expeditions. The 1999 CD-ROM calculated a single average for all vessels leaving from Europe and used this as the imputed value where the actual numbers were missing. TSTD2 computes a more refined mean by calculating separate figures for various combinations of rigs and periods. Where this information is not available, the imputed figures for arrivals and departures are calculated for four periods: prior to 1650; 1650–1673; 1674–1730; and post-1730.12 This procedure certainly produces more accurate estimates of missing data. Finally, it seems to us that while for most of the slave-trade era, particularly 1630–50 and after 1673, the coverage of sources justifies the assumption that some record of

231

232

Jelmer Vos, David Eltis, and David Richardson

almost all Dutch slaving ventures has found its way into the new database, this is probably not the case for the period from 1650 to 1673, effectively a period of free trade with no large companies operating. Binder drew on sources from around the Atlantic world to construct his database for these years, but we have nevertheless increased the annual figures generated by his database by 10 percent to reflect this opinion. What do our new data tell us about the Dutch slave trade? Overall, we estimate a total volume of 554,300 slaves carried from the African coast in Dutch vessels in 1596–1829. This contrasts with Postma’s latest aggregate estimate of 501,409 slaves, with the difference deriving partly from our new information, partly from our differing estimating procedures for missing data, and partly from our slightly more aggressive assumptions about which vessels actually engaged in slave trading as opposed to just intending to engage.13 There appear to have been five distinct periods of Dutch slave trading activity. These periods are differentiated not only by the way the Dutch organized their trade but also by major shifts in American destinations and African origins, as well as several other features of the trade, such as types of vessel and shipboard mortality. The first period predates the Dutch attack on Brazil and begins with the voyage of the Fortuin under captain Cornelis Jansz Boer, carrying fifty-eight slaves to Portugal in 1596, and ends with the establishment of a regular Dutch slave trade to Pernambuco in 1637 by the first WIC.14 Some scholars contend that no transatlantic Dutch slave trade existed in this period. The evidence of the new database is ambivalent on this issue. Only ten voyages appear to have intended to carry or actually carried slaves to the Americas before 1637, with the destinations split evenly between Brazil and the Spanish Americas. Of these, one ship sank and another experienced a voyage-ending crew mutiny, in both cases before taking on slaves. More strikingly, there are no less than seven cases at this time of Portuguese slavers that were captured by the Dutch, who either turned the slaves loose on the Brazilian coast or, after looting the vessels of their valuable nonhuman commodities, allowed the vessel to continue with its human cargo intact.15 The clear implication is that the Dutch had no slave markets of their own in the Americas and frequently found unattractive the option of trying to dispose of slaves in the territories of Spain and Portugal. This was the period when the Dutch brought slaves from Portuguese vessels into Virginia and even into Holland itself—both marginal markets at best.16 Capturing a slave vessel during this period was not a certain route to riches. If a Dutch trade did exist at this time, it cannot have carried off more than a few thousand people from Africa over four decades.

The Dutch in the Atlantic World

That being said, further research in Dutch notarial archives might reveal a wider participation by the Portuguese Jewish community in the Netherlands in the early seventeenth-century slave trade than currently appears. The few collected records of Dutch slaving before 1635 certainly point to a strong Portuguese involvement in these voyages.17 Much has been written about the role played in Dutch expansion by the Sephardic diaspora, which linked Amsterdam with the Iberian Peninsula, Brazil, and other locations in the Americas.18 But the importance of Portuguese Jewish merchants for the development of the Dutch trade with Africa is a theme that still needs exploration. Some scholars have already pointed out that a number of Amsterdam-based Jewish merchants participated in the African slave and produce trade in the 1610s and even set up companies to that end.19 It is unlikely that future research on this early period will revise our assessment of the volume and direction of the Dutch slave trade, but it may lead to new ideas about early Atlantic history.20 The conquest of Pernambuco in 1630 opened up an accessible market in the Americas for the first time and was, from the standpoint of the slave trade, a development more important than Dutch incursions into Africa. During this second period, spanning the years 1637–50, the Dutch focused almost exclusively on Pernambuco, a point of some significance when we consider the Dutch role in the growth of the sugar complex in the English and French Caribbean starting in the 1640s.21 In fact, as table 8.1 suggests, the TSTD2 data set contains only three slave voyages in these years linked with any part of the Americas other than Pernambuco. The Dutch carried away an estimated 30,900 slaves from Africa and disembarked 25,500 in the Americas during the thirteen years through to 1650.22 This was not an insignificant level of activity, but its average level of less than 2,500 slaves per year was well below what the Dutch would carry over the next century and a half, periods of war excepted. The third period, ranging from 1650 to the operational beginnings of the second WIC in 1674, is the least known and in some ways most interesting of the five Dutch slave-trading periods. This quarter century largely predated the enforcement of the French exclusif policy, whereby French colonies were reserved for French slave merchants. It also coincided with the final peace between Spain and the Netherlands (in 1648) and came just a decade after the independence of Portugal—the dominant supplier of slaves to the Spanish Americas for as long as the crowns of the two countries were united. The Dutch took full advantage of this opportunity, and by the early 1660s they had become the dominant suppliers of slaves to the asientistas. Over half the slaves

233

Table 8.1 Principal Regions of Disembarkation of Slaves Carried on Dutch Transatlantic Slave Vessels by Twenty-Five-Year Intervals, 1600–1800 (in thousands)

Years 1601–25 1626–50 1651–75 1676–1700 1701–25 1726–50 1751–75 1776–1800 All years* Row %

Brazil, Pernambuco

Brazil, Other

Spanish Antilles

Spanish Mainland America

Spain

English Americas

French Caribbean

Dutch Caribbean

Dutch Guianas

Danish America

All Other Regions

Total

0.3 24.8 0.2 — — — — — 25.3 5.8

— 0.3 0.5 — — — — — 0.8 0.4

0.8 0.3 3.6 — — — 0.4 0.3 5.4 1.4

0.5 — 9.8 3.3 2.6 — 0.2 0.3 16.9 4.1

— — 0.9 1.0 — — — — 1.9 0.4

— 0.5 3.1 0.9 0.4 — — 0.8 5.7 1.4

— — 9.6 — 0.6 — 0.3 0.3 10.8 2.5

— — 31.4 42.5 28.4 9.0 13.9 3.3 128.5 29.4

— — 7.2 23.4 24.0 54.7 94.0 23.7 227.0 52.0

— — — 2.8 2.3 — — — 5.1 1.2

— — 4.6 0.4 — — 0.3 0.5 5.8 1.5

1.6 26.0 70.8 74.2 58.3 63.7 109.2 29.2 433.0 100

Source: TSTD2. *Totals include 1,160 slaves disembarked before 1601 (of whom 861 went to Brazil) and 2,665 disembarked after 1800 (of whom 443 went to Cuba, 279 to the English Americas, and 1,684 to the Dutch Guianas). There are also 1,004 slaves in the English Americas category that went to New York before it became English in 1664, and 855 in the “All Other Regions” category that went to St. Croix, which was French at the time of their arrival.

The Dutch in the Atlantic World

brought to the New World in this quarter century went to the Dutch Caribbean, mainly to Curaçao and thence to the Spanish Central American mainland. Yet, as table 8.1 shows, the Dutch ranged far beyond the Spanish American heartland. They sent several shipments of slaves to New Amsterdam—about as far north as slaves were sold in the Americas—as well as to Río de la Plata at the extreme southern range of American slave markets. They were the major transatlantic suppliers of slaves to the nascent plantation complex in the French Windward Islands during the pre-Colbert era; they carried slaves to their own colony of Suriname after taking it from the English in 1667; and they even managed to sell nearly 3,000 slaves in Jamaica and Barbados, the only time in the history of the slave trade when non-English transatlantic slave traders made even minor inroads into those markets. After expanding their traffic through the early 1650s, Dutch slavers carried off at least 4,600 slaves a year from Africa in 1658–74, double the number they had carried off in the Pernambuco era. This was the period when the Dutch came closest to the role assigned to them in the traditional historiography.23 The fourth era of the Dutch slave trade, dominated by the second Dutch West India Company, is the best known of the five, and we need say little about it here. The number of slaves carried increased only slightly in the final quarter of the seventeenth century, but then declined by a quarter in the early eighteenth. This was a period of rapid expansion in the slave trade as a whole, yet the Dutch continued to carry annually about the same number between 1674 and 1730 as they had in the twenty years before the formation of the second WIC. Moreover, as table 8.1 shows, the Dutch largely supplied just two markets in the Americas between 1674 and 1730, delivering about the same number of slaves to each. One comprised the Dutch plantation colonies on the northern edge of South America, principally Suriname but later extending to Demerara, Berbice, and Essequibo. The other was Spanish America, largely supplied through Curaçao. These areas apart, the Dutch withdrew at this time from all other markets in response to the increasing mercantilist restrictions of the major colonial powers. Some smuggling into French and English colonies through St. Eustatius certainly occurred, but there is no suggestion that the Dutch were supplying these colonies through St. Eustatius with more than a fraction of their labor needs at this time. Compared with the efficient English traders of this era, the ships of the second WIC were large, spent a long time on the coast obtaining their slaves, and had to carry the large overhead of the Gold Coast castle system as the gold trade declined.24 The Dutch apparently retained little if any competitive advantage at this time.

235

236

Jelmer Vos, David Eltis, and David Richardson

A final major shift in the structure of the Dutch slave trade occurred with the removal of the WIC’s monopoly over slave trading between 1730 and 1734. Within a decade, independent merchants dominated the Dutch involvement in the slave trade.25 Their involvement helped first to restore Dutch slave carrying to levels close to those attained in 1651–75 and then to new heights after 1750. Annual Dutch slave shipments averaged 5,300 a year in 1750–75 and reached their highest ever level of 6,100 a year in the fifteen years before the American Revolution. Almost all the slaves shipped by the Dutch after 1730 were sold in the Dutch plantation territories. A final attempt to supply slaves to foreign colonies in the 1720s through the free port of St. Eustatius was abandoned as a money loser, though not before some 10,000 slaves had passed through the colony.26 After 1776, moreover, Dutch control of the slave supply to their own colonies came under threat, principally from British and U.S. traders, and by 1815 all the Dutch Caribbean colonies other than Suriname had fallen under British economic and political control. The Dutch involvement in the slave trade had already dwindled and, except for a brief resurgence during the Peace of Amiens, largely disappeared after 1803. Tables 8.1 and 8.2 reveal the continuing existence of Dutch slaving activity after 1803, but these were in reality foreign-owned ships that employed the Dutch flag as a flag of convenience.27 By the early nineteenth century, the financing of slaving voyages by Dutch merchants had ended. In summary, there was a bimodal pattern over time in the Dutch slave trade, probably unique among national groupings of slave traders, with the peaks occurring a century apart in the 1660s and 1760s, and a slightly longer cycle in destinations that began in Dutch territories (Pernambuco), briefly exploded across the Americas, and then gradually returned to the original pattern with the focus now on Suriname. For most of the slave-trade era, the Dutch supplied only their own possessions with slaves. The one significant exception to this was their large share of Spanish markets between the late 1650s and 1698, but even in this period they were never able to exclude the English from supplying these markets through intercolonial trafficking. For a couple of decades after 1698, the Dutch themselves participated in the intercolonial trade, but by 1720 the English had taken over completely. Thus, for as long as the Dutch could sell to the Spanish—in other words, until English competition for that market became intense—the transit trade to Spanish America was always more important than supplying slaves to their own plantation colonies. The secular trends in the Dutch slave trade described above provide only part of the answer to the issues posed in this chapter’s introduction. Much of

The Dutch in the Atlantic World

the rest comes from placing these patterns against the activities of the other transatlantic slave traders. Figure 8.1 provides a broad comparative perspective. The relatively small role of the Dutch in the transatlantic slave trade is immediately apparent in the figure, though this is scarcely a new finding. The figure also shows the bimodal distribution over time. The Dutch traffic declined in relative terms in the last quarter of the seventeenth century (though it remained steady or expanded in absolute terms), and when it expanded once more in the eighteenth century, it did so by far less than that of any other slavetrading power, including the Danes. As a consequence, the relative position of the Dutch as slave traders (and perhaps not unrelated, in broad economic terms as well) declined steadily from the 1660s onward. In addition, the Dutch relative position was greatest during 1651–75, when they accounted for 101,000 slaves out of a total of over 500,000 in the trade as a whole. There may well be some undercounting of Dutch slave vessels in this period, but it is certainly the case that they accounted for much less than half the total trade when their relative share was at its greatest. In no other quarter century or indeed in any period shorter than twenty-five years did they come anywhere near this ratio. More interestingly, at the time that the Dutch attained their peak share of the slave trade—from 1651 to 1675—that trade was only one-tenth of what it was to become a century later. Figure 8.1 illustrates that the Dutch did nothing to lead the way in this expansion. At best the Dutch offset a decline in the third quarter of the seventeenth century, brought about by a decline in the Brazilian sugar sector.28 The expansion of the slave trade was led by the English, the Portuguese (at least after the gold boom in Minas Gerais), and eventually the French. Good indications of the size of the Spanish trade are lacking between 1640 and 1650, but in the preceding twenty-five years, 1616–40, the revised database indicates that the Portuguese delivered more slaves to the Spanish Americas than did the Dutch between 1651 and 1675 and only slightly fewer than the Dutch managed in the final quarter of the seventeenth century.29 A closer examination of the main rivals of the Dutch suggests that the latter did not seek to penetrate other slave markets in the Americas until the Pernambuco episode closed at the end of the 1640s, by which time the Barbados sugar revolution was already well under way, thanks to the slaves brought by English ships. The new database contains records of forty-seven English slave voyages in the decade of 1641–50, compared to eighty-six under the Dutch flag—all but six, as noted, to Pernambuco. Pernambuco was of course a bigger market for slaves, but there is no evidence of significant Dutch involvement in the Barbados trade either in the 1640s or at any subsequent period. Moreover,

237

238

Jelmer Vos, David Eltis, and David Richardson

Figure 8.1 The Dutch Share of the Transatlantic Slave Trade Source: TSTD2.

while the new database shows the Dutch increasing their lead over the English in the 1650s, with seventy-five slaving expeditions sent out compared against thirty, we should note, first, that the 1650s remain the most underresearched decade in the history of the English slave trade, and second, that there is already evidence in the 1650s of the English carrying slaves to Brazil and Spanish America. It was thus not only the Dutch who saw a window of opportunity in the slave-trade business in the years after their expulsion from Brazil. They did make inroads into the French market for slaves after 1650, but the majority of slaves arriving in the French colonies in the seventeenth century came from other sources, even after allowing for the small intercolonial traffic from St. Eustatius in the late seventeenth century.30 Table 8.1 suggests that the Dutch sold five times more slaves to the Spanish than they did to the French before 1675. All this points to a scenario far removed from the dominant role that the Dutch have been assumed to have played in the establishment of the early modern American slave colonies.31 A similar message emerges from an examination of Dutch patterns of trading for slaves on the African coast. Table 8.2 shows that the Dutch followed the

The Dutch in the Atlantic World

Table 8.2 African Region of Embarkation of Slaves Carried on Dutch Vessels by Twenty-Five-Year Intervals, 1626–1800 (in thousands)

Years 1601–25 1626–50 1651–75 1676–1700 1701–25 1726–50 1751–75 1776–1800 All years* Row %

Senegambia

Sierra Leone

0 0 4.9 3.2 0 0.3 0.4 0 8.8 1.9

0 0 0.7 0 0 0 0.3 0.3 1.3 1.5

Windward Coast 0 0 0 0 0.1 7.1 44.2 13.5 64.9 14.0

Gold Coast

Bight of Benin

Bight of Biafra

0 0.6 6.5 9.9 8.4 27.3 21.2 15.8 89.7 19.2

0 4.2 27.3 31.3 32.2 12.3 1.8 0 109.1 23.4

0 4.7 16.1 2.4 0.4 0 0.9 0 24.6 5.3

West Central Africa

Total

1.3 14.6 25.5 31.5 23.5 14.9 47.9 7.3 165.2 35.9

0 24.0 80.9 78.3 64.7 61.9 116.6 36.9 463.3 100

Source: TSTD2. Totals may not add up due to rounding. *Includes 1,100 slaves embarked before 1601 and 1,100 embarked after 1800.

Portuguese in drawing mainly on Angola before 1650. But unlike the Portuguese, they obtained one-third of their slaves from West Africa—all from the Bight of Benin and the Bight of Biafra. Thus, as a result of the Dutch incursion, more West Africans began to arrive in Pernambuco, including what were likely the first Igbo. The Dutch drew no more than a few hundred slaves a year from the Gold Coast in this early period, but then neither did any other slavetrading group.32 Thus, for a half century after 1650, the Dutch maintained a significant presence in all the major regions of Western Africa in which slaves were actually obtainable. Their prime source was the Slave Coast, although even after they were expelled from Luanda, they continued to tap into Angolan slave supplies via embarkation points north of the Congo River. Starting in 1700, the English began trading slaves more than gold on the Gold Coast, and two decades later the Dutch followed suit.33 But the Dutch never replicated in the Gold Coast slave trade their earlier dominance in the gold trade. The overriding pattern after 1675, however, is one of the Dutch being squeezed out of most of the major African coastal slaving regions. They were replaced first in the Bight of Biafra by the English, and when they switched from gold to slaves on the Gold Coast, they lagged behind their English counterparts despite their extensive holdings of forts in the region. Further, the Dutch steadily lost ground in the first half of the eighteenth century to all the major national groupings of traders in the Bight of Benin.

239

240

Jelmer Vos, David Eltis, and David Richardson

The so-called free-trade era of Dutch slaving beginning in 1730 was characterized by three key developments in the African side of the business. First, the free traders were permanently eliminated from the Slave Coast after 1750, their share of the trade being taken by first the French and then the Portuguese. Second, squeezed out of all the major West African markets, the Dutch now took advantage of an expanding market for slaves on the Windward Coast, which in the third quarter of the eighteenth century temporarily became a significant source of slaves.34 No other major slave-trading nation at this time relied as heavily on the Windward Coast as did the Dutch, who drew almost 40 percent of their slaves from the area between 1750 and 1775. Indeed, when the Windward Coast returned to its role as a peripheral slavetrading region after 1775, the Dutch trade fell into permanent decline. Third, the Dutch free traders orchestrated a comeback in northern Angola through their connections with African trade brokers in Malembo. Table 8.3 shows that Malembo alone supplied no less than 40 percent of all slaves carried off in Dutch vessels in the third quarter of the eighteenth century. The Windward Coast and the Malembo-Loango region supplied almost four out of five of the slaves arriving in the Dutch Guianas during this boom period of Dutch plantation agriculture. In the long run, however, the French undermined and eventually took over the Dutch position in Malembo. Where table 8.2 applies a broad brush, table 8.3 examines the new data by individual embarkation points. All slave-trading nations drew on a surprisingly small number of individual ports on the African coast, but the Dutch present the most extreme variant of this pattern. Between two-thirds and three-quarters (72.1 percent) of all slaves on Dutch vessels who embarked in known ports left from Elmina on the Gold Coast; Ardra and Ouidah on the Slave Coast; Luanda; and Loango and Malembo just north of the Congo River. The Windward Coast is omitted from table 8.3 because more research is required to identify the many small embarkation points where the Dutch and other national groupings of slave traders obtained small numbers of slaves. So far the records suggest that Cape Lahou stands out among these ports, although only in twenty-four cases does it rank as the principal port of slave purchase. As previously noted, trade on the Windward Coast was largely confined to the post-1730 period; almost all the 63,000 slaves from the Windward Coast on Dutch slavers shown in the final column of table 8.2 embarked after 1730. Among the six major ports outside the Windward Coast, a pronounced sequential pattern is apparent in table 8.3: at any given time, Dutch slavers were mostly drawing on only three embarkation points. Thus, Ouidah came

Table 8.3 Major Ports of Embarkation for Slaves Carried from Africa on Dutch Slave Ships by Twenty-Five-Year Intervals, 1626–1800 (in thousands) Ports

1626–50

1651–75

1676–1700

1701–25

1726–50

1751–75

1776–1800

All Years

Column %

Gorée Lahou Elmina Cape Coast Castle Kormantine Ouidah Jakin (Jacquin) Ardra (Offra) Badagry New Calabar Old Calabar** São Tomé Loango (Boary) Malembo Cabinda Soyo Luanda Largest Five Ports All Identified Ports

0 0 0.3 0 0 0 0 4.2 (2) 0 0 2.3 (3) 0.7 (4) 0.1 0 0 0.4 (5) 13.7 (1) 21.3 21.7

2.2 (5) 0 4.9 (4) 0 0 0.5 0 22.5 (1) 0 8.5 (3) 0.3 0 18.2 (2) 1.5 0 2.1 1.8 56.3 62.5

0 0 0.7 0 0 3.3 (2) 0 24.7 (1) 0 0 2.1 (3) 0 1.0 (5) 0.9 (4) 0 0 0 32.0 32.7

0 0.1 5.3 (3) 0 0 27.0 (1) 1.4 1.8 (5) 0 0 0.4 0 5.7 (2) 2.8 (4) 0 0 0 42.6 44.5

0 1.3 19.7 (1) 0 0.2 0 6.3 (2) 0 5.2 (3) 0 0 0 1.7 (4) 1.5 (5) 0.6 0 0 34.4 36.5

0.3 4.1 (4) 7.9 (2) 0 3.3 (5) 0.8 0 0 0 0 0 0 2.7 18.5 (1) 5.9 (3) 0 0 39.7 43.5

0 1.0 (2) 14.4 (1) 0.3 (5) 0.2 0 0 0 0 0 0 0 0.4 (3) 0.3 (4) 0 0 0 16.4 16.6

2.5 6.5 53.2 0.3 3.7 31.6 7.7 53.2 5.2 8.5 5.1 0.7 29.8 25.5 6.5 2.5 15.5 242.7 284.9

0.9 2.3 18.7 0.1 1.3 11.1 2.7 18.7 1.8 3.0 1.8 0.2 10.5 9.0 2.3 0.9 5.4 85.2 90.7

Source: TSTD2. *Numbers in parentheses show the ranking of the largest five ports in each time period. **Includes Rio Real and Calabari.

242

Jelmer Vos, David Eltis, and David Richardson

to replace Ardra in the late seventeenth century (the Dutch were the last of the major European slaving powers to abandon Ardra); south of the equator, Loango replaced Luanda following the Dutch expulsion from Angola, and it was in turn replaced by Malembo in the free-trade era.35 Moreover, four of the six major embarkation ports may be paired in the sense that, arguably, the same mix of peoples passed through the pairings. Ardra and Ouidah were alternative outlets for Aja-Fon captives, while Loango and Malembo were similarly situated for Vili traders drawing on the Congo basin and the Angolan hinterland.36 When the Dutch switched from Ardra and Loango to Ouidah and Malembo, respectively, the ethnicity of the captives themselves may not have changed much. The major innovations in the African side of the Dutch traffic were induced by the closing down of the Biafra trade, the switch from gold to slaves on the Gold Coast, and the late inception of the Windward Coast trade. The implications of the TSTD2 data for defining the transatlantic links maintained by the Dutch and for tracking the African origins of the slaves they delivered to the Americas may now be spelled out. As in other branches of the trade—including the extensively discussed links between the ricegrowing regions of Africa and South Carolina in the eighteenth century—it was not America’s demand for slaves from particular African regions that shaped such links, but rather geography and power relations both among Europeans and among Africans and Europeans. In Pernambuco, the Dutch incursion brought a greater share of slaves from West Africa, in particular from the Bight of Biafra, than had hitherto held. This pattern was even more pronounced for Spanish America, the dominant market for the Dutch for the balance of the seventeenth century. Compared to the pre-1640 Portuguese traffic to Spanish America, the Dutch transatlantic link meant the introduction of large numbers of West Africans, in particular peoples passing through Ardra, most of them probably Aja-Fon speakers.37 Just over half of the Dutch trade originated from this area, a major break with the Spanish American past and the beginning of the large diversification in the origins of African arrivals that eventually characterized the entire Spanish Caribbean and Central America. For the other major area supplied by the Dutch, Suriname, a variation on the above pattern is apparent.38 Table 8.4 shows that the small “founding” or charter populations of Africans were drawn exclusively from the Bight of Biafra and West Central Africa. Not shown in table 8.4 are the four large English slavers that carried 1,700 slaves into the region before the Dutch takeover of

The Dutch in the Atlantic World

Suriname in 1667.39 These too came from the Bight of Biafra and northern Angola. From the assumption of Dutch sovereignty to the beginning of the free-trade period in 1730, however, three out of four arrivals came from the Bight of Benin and northern Angola, most passing through the ports of Ardra and Ouidah on the Slave Coast and, presumably, Loango and Malembo north of the Congo.40 Elmina Castle supplied most of the remainder. After 1730 the Windward Coast replaced the Slave Coast as one of the principal regional suppliers, alongside Angola and the Gold Coast. No other major slave-trading power resorted to the Windward Coast to the extent that the Dutch did in this period, and it is likely that the Dutch accounted for up to half of the greatly expanded Windward Coast slave trade at this time. It is hard not to link this shift with the exclusion of the Dutch from the major Bight of Benin ports. Before the trade became illegal in the nineteenth century, the regions of the Americas that drew heavily upon the Windward Coast (indeed, the whole of Upper Guinea) were the minor plantation regions of the Americas. Also not shown in table 8.4 is the resurgent non-Dutch traffic to the Dutch Guianas, which developed after the collapse of the Dutch slave trade in the 1780s. English and U.S. slave traders in particular brought in slaves from a wide variety of African regions in the late 1780s and 1790s.41 In summary, it is difficult to discern a dominant region of provenance for Suriname’s African population. But during the major period of the colony’s expansion, the half century after 1725, we can at least say that the great majority of forced migrants were passing primarily through three ports—Elmina, Loango, and Malembo—in addition to a number of smaller outlets on the Windward Coast. Some of these locations were, of course, widely separated from each other, but within these widely separated groupings, slaves must have found a good deal of shared cultural elements, including languages, among fellow captives both in the vessel that carried them to the New World and in the communities where they found themselves thereafter. On the other hand, the larger picture shows that major cultural fissures must have prevailed initially on the American side.42 The Dutch influence on the makeup of African communities in the Spanish Americas and the Dutch Guianas, and, more specifically, Dutch weakness relative to the other major slave empires in the Atlantic world, which in effect shaped that influence, bring us back to the question posed at this chapter’s outset. Much of the recent debate on the role of the Dutch in European overseas expansion has focused on the question of which was of greater importance to the Dutch: Asia or the Americas? The most current research suggests

243

244

Jelmer Vos, David Eltis, and David Richardson

Table 8.4 African Regional Coastal Origins of Slaves Arriving in Suriname by Twenty-Five-Year Intervals, 1651–1825 (in thousands)

Years 1651–75 1676–1700 1701–25 1726–50 1751–75 1776–1800 1801–25 All years Row %

Senegambia

Sierra Leone

Windward Coast

Gold Coast

Bight of Benin

Bight of Biafra

West Central Africa

Total

0 1.1 0 0 0.3 0.2 0 1.5 0.8

0 0 0 0 0.2 1.8 1.6 3.6 1.8

0 0 0 5.7 29.2 5.8 1.4 42.1 21.4

0 0.9 1.5 15.7 15.4 9.5 2.6 45.6 23.2

0 7.2 13.2 9.1 1.5 0 0.5 31.4 15.9

2.9 1.4 0 0 0 1.5 6.4 12.3 6.2

3.0 8.6 3.8 8.9 29.0 2.6 4.4 60.3 30.6

5.9 19.2 18.5 39.4 75.6 21.4 16.8 196.8 100

Source: TSTD2. Totals may not add up due to rounding.

that it was the Americas.43 There is, however, little tension between this conclusion and the picture of the Dutch role in the Atlantic world presented here. The Dutch contributed to but did not lead the way in the slave trade’s rapid expansion beginning in the middle of the seventeenth century. Before 1650 the Dutch slave trade concentrated almost exclusively on Brazil. The English Caribbean was established without significant Dutch involvement in the supply of its slaves. Dutch slave traders were far more active in the early French Caribbean but were quickly excluded in the 1670s, and even before this there were other sources of slaves available for planters in Martinique and Guadeloupe. It is hard to believe that, without the 10,000 or so slaves the Dutch supplied, the development of the French sugar complex would have been seriously delayed. None of this, of course, reverses the conclusions of recent research on the relative importance of Asia and the Americas to the Dutch, though it does perhaps reflect the general decline in Dutch power in Europe at this time. A final cautionary note is called for. The focus of this chapter has obviously been the slave trade. It has said nothing about a possible Dutch role in transferring sugar-making technology from Brazil to the Caribbean and little about the role of Dutch capital in establishing plantations in non-Dutch Caribbean colonies. The fact remains, however, that sugar production was established in the Caribbean while the Dutch were preoccupied with Pernambuco and still had prospects of permanence there. Detailed examination of Barbados land

The Dutch in the Atlantic World

registry records has shown little evidence of Dutch involvement in the sugar revolution of the 1640s.44 The balance of the evidence at the moment is toward revising the traditional interpretation that places the Dutch at the center of the early Caribbean sugar revolution.

NOTES

1. Charles Boxer is, of course, the major exception to this generalization. 2. See the writings of Jan Lucassen, such as “The Netherlands, the Dutch, and LongDistance Migration in the Late Sixteenth to Early Nineteenth Centuries,” in Nicholas Canny, ed., Europeans on the Move: Studies in European Migration, 1500–1800 (Oxford, 1994), 153–91. 3. Johannes Menne Postma, The Dutch in the Atlantic Slave Trade, 1600–1815 (Cambridge, 1990). 4. Franz Binder’s original note cards were recovered with his assistance from the Municipal Archives of Amsterdam and are now on deposit at the new Wilberforce Institute for the study of Slavery and Emancipation at the University of Hull, UK. 5. These data were published in concise form as an appendix to Den Heijer’s book Goud, ivoor en slaven: Scheepvaart en handel van de Tweede Westindische Compagnie op Afrika, 1674–1740 (Zutphen, 1997). 6. Klaas Ratelband, ed., Vijf dagregisters van het kasteel São Jorge da Mina (Elmina) aan de Goudkust (1645–1647) (’s-Gravenhage, 1953); idem, Nederlanders in West-Africa, 1600–1650: Angola Kongo en São Tomé (Zutphen, 2000); Albert van Dantzig, The Dutch and the Guinea Coast, 1674–1742: A Collection of Documents from the General State Archive at the Hague (Accra, 1978); Cornelis Goslinga, The Dutch in the Caribbean and the Guianas, 1680–1791 (Assen, 1985). 7. See Johannes Postma, “A Reassessment of the Dutch Atlantic Slave Trade,” in Johannes Postma and Victor Enthoven, eds., Riches from Atlantic Commerce: Dutch Transatlantic Trade and Shipping, 1585–1817 (Leiden, 2003), 115–38. 8. Corrie Reinders Folmer-van Prooijen, Van goederenhandel naar slavenhandel: De Middelburgse Commercie Compagnie, 1720–1755 (Leiden, 2000). 9. Henk den Heijer, “The West African Trade of the Dutch West India Company, 1674–1740,” in Postma and Enthoven, Riches from Atlantic Commerce, 149. 10. In addition, we have created a separate file of 360 voyages under the Dutch flag, most of which were clearly trading in African produce but some of which are cases of genuine doubt. Such voyages may or may not have been slave ships, though on balance we think they were not. This set, it must be said, can easily be expanded. Not yet included, for example, is Den Heijer’s collection of WIC produce voyages from 1674 to 1740. 11. By contrast, the new Postma database assumes that such vessels did not reach the Americas. 12. The new database thus calculates separate means for 152 different types of vessels and periods, as opposed to only three categories in the 1999 database. 13. Postma, “Reassessment,” 137. Our own total is derived from TSTD2.

245

246

Jelmer Vos, David Eltis, and David Richardson

14. For the Fortuin voyage, see TSTD2, voyageid 11,355. 15. See TSTD2, voyageids 11,227, 11,228, 11,231, 11,232, 11,234, 11,238, and 11,367. 16. In 1596, 130 slaves from a captured Portuguese vessel (voyageid 46,482) were brought to Middelburg, Zeeland, which caused some confusion and embarrassment among the local notables. See Victor Enthoven, “Early Dutch Expansion in the Atlantic Region, 1585–1621,” in Postma and Enthoven, Riches from Atlantic Commerce, 40 n. 73. For the 1619 Virginia delivery (voyageid 29,252), see Engel Sluiter, “New Light on the ‘20 Odd Negroes’ Arriving in Virginia, August 1619,” William and Mary Quarterly 54 (1997): 395–98. 17. At present the database contains records of eleven Dutch slaving voyages between 1596 and 1615, nine of which hint of strong Portuguese involvement. Eight voyages set out to Angola (a ninth to São Tomé). Four of these were destined for Brazil, the other four for the Spanish Americas. In four cases we know that a Portuguese captain commanded the vessel. 18. See, for example, the analyses and literature cited by Jonathan Israel and Daniel Swetschinski in J. C. H. Blom, R. G. Fuks-Mansfeld and I. Schöffer, eds., The History of the Jews in the Netherlands, trans. Arnold J. Pomerans and Erica Pomerans (Oxford, 2002), 77–79, 85–112, 399–408. For a clear statement, see Wim Klooster, “Sephardic Migration and the Growth of European Long-Distance Trade,” Studia Rosenthalia 35 (2001): 121–32. See also Christopher Ebert, “Dutch Trade with Brazil before the Dutch West India Company, 1587–1791,” in Postma and Enthoven, Riches from Atlantic Commerce, 65–67; Wim Klooster, “Curaçao and the Caribbean Transit Trade,” in ibid., 207; and idem, “An Overview of Dutch Trade with the Americas,” ibid., 368–69. 19. Ernst van den Boogaart and Pieter C. Emmer, “The Dutch Participation in the Atlantic Slave Trade, 1596–1630,” in Henry A. Gemery and Jan S. Hogendorn, eds., The Uncommon Market: Essays in the Economic History of the Atlantic Slave Trade (New York, 1979), 354; Enthoven, “Early Dutch Expansion,” 43. 20. It should not be forgotten that other trade diasporas, like those based in the southern Netherlands in the 1560s and 1580s, may have played an equal if not more important role in Dutch commercial expansion. See Jonathan Israel, The Dutch Republic. Its Rise, Greatness, and Fall, 1477–1806 (Oxford, 1995), 307–27, and Charles Boxer, The Dutch Seaborne Empire, 1600–1800 (London, 1965), 20. One of the first major Dutch traders in West Africa was Balthasar de Moucheron, from Zeeland, whose name suggests a stronger connection with the francophone world than the Iberian Peninsula. See Enthoven, “Early Dutch Expansion,” 42. Several of the owners of Dutch slaving ventures in the 1650s, 1660s, and 1670s were Portuguese, but perhaps even more notable are the many names of French origin. 21. During the first six years of Dutch occupation of northern Brazil, only one cargo of a captured slaver from Angola was delivered in Recife. Not until 1635, when the country was sufficiently pacified to resume sugar production, did the WIC organize its first proper slaving voyage, which arrived in Brazil in 1637 (voyageid 11,288). In addition, a little more than 1,000 slaves were captured from the Portuguese at sea and landed in Brazil in 1636 (voyageids 11,224, 11,225, and 11,226). See Van den Boogaart and Emmer, “Dutch Participation,” 357–58.

The Dutch in the Atlantic World

22. The disembarkation number differs slightly from the total of 26,286 imported slaves calculated by Van den Boogaart and Emmer (“Dutch Participation,” 369). The differences can be accounted for, first, by the fact that we included in our calculations all slave ships reported on the African coast but for which there is no further information on the arrival side, while Van den Boogaart and Emmer counted purely on the basis of recorded imports in Brazil. Second, unlike them we have excluded the 1,326 slaves captured from Portuguese ships in 1630 and 1636. 23. Wim Klooster recently argued that the Dutch dominated the Atlantic slave trade between 1663 and 1688. See Klooster, “Overview of Dutch Trade,” 376. 24. Postma calculated that the coasting period of free traders was roughly double that of WIC slave ships, causing a higher mortality rate among the purchased slaves while on the African coast. See Postma, Dutch in the Atlantic Slave Trade, 249. 25. Our records show that the WIC still dispatched at least fourteen slave ships to the Americas between 1741 and 1793. 26. Den Heijer, “West African Trade,” 161. 27. Pieter Emmer, “The Abolition of the Abolished: The Illegal Dutch Slave Trade and the Mixed Courts,” in David Eltis and James Walvin, eds., The Abolition of the Atlantic Slave Trade: Origins and Effects in Europe, Africa, and the Americas (Madison, Wis., 1981), 177–92. 28. This picture is consistent with a broader assessment of the Dutch role in the Atlantic world. See Pieter Emmer, “The Dutch and the Slave Americas,” in David Eltis, Frank D. Lewis, and Kenneth Sokoloff, eds., Slavery in the Development of the Americas (Cambridge, 2004), 70–86. 29. The totals for Dutch deliveries in 1651–75 and 1676–1700 have been derived by combining the arrivals in the Dutch Caribbean, the Spanish Antilles, and the Spanish American mainland shown in table 8.1. For the earlier Portuguese trade, see António de Almeida Mendes, “The Foundations of the System: A Reassessment of the Slave Trade to the Spanish Americas in the Sixteenth and Seventeenth Centuries,” this volume (chap. 2). 30. See James Pritchard, David Eltis, and David Richardson, “The Significance of the French Slave Trade to the Evolution of the French Atlantic World before 1716,” this volume (chap. 7), and Pritchard, “An Incidental Slave Trade: The French in Africa During the Seventeenth Century” (unpublished paper, 2004). 31. The key texts stressing the Dutch role in the creation of the plantation complex of the British Caribbean are still Richard Dunn, Sugar and Slaves: The Rise of the Planter Class in the English West Indies, 1624–1713 (New York, 1972), 59–67, and Carl and Roberta Bridenbaugh, No Peace beyond the Line: The English in the Caribbean, 1624–1690 (New York, 1972), 61–100. 32. There are no recorded Portuguese or English voyages to or from the Gold Coast in this period. 33. David Eltis, “The Volume and African Origins of the Seventeenth-Century English Transatlantic Slave Trade: A Comparative Assessment,” Cahiers d’Etudes Africaines 138 (1995): 617–27. 34. The Windward Coast is defined as Cape Mount up to and including the Assini River.

247

248

Jelmer Vos, David Eltis, and David Richardson

35. On Ouidah, see Robin Law, Ouidah: The Social History of a West African Slaving Port, 1727–1892 (Athens, Ohio, 2004), 30. 36. Phyllis Martin, The External Trade of the Loango Coast, 1576–1870: The Effects of Changing Commercial Relations on the Vili Kingdom of Loango (Oxford, 1972). 37. See Mendes, “Foundations of the System.” 38. In fact, the Dutch supplied the whole of the Dutch Guianas, but Berbice, Demerara, and Essequibo received relatively few slaves directly from Africa before 1780, and the analysis here focuses on Suriname. 39. Postma, Dutch in the Atlantic Slave Trade, 289–90. English planters settled in Suriname in 1651, but by the end of the 1670s all but thirty-one of them had left. See Johannes Postma, “Surinam and Its Atlantic Connections, 1667–1795,” in Postma and Enthoven, Riches from Atlantic Commerce, 289. 40. The information on specific ports in northern Angola is not as complete before 1730 as it is afterward. 41. According to Postma, only 120 slaves arrived from American vessels before 1789, but between then and 1794, Americans carried 7,011 slaves to Suriname in 82 ships. See “Surinam and Its Atlantic Connections,” 306–8. For the end of the Dutch slave trade to Suriname, and thereby of the Dutch traffic in general, see Pieter Emmer, De Nederlandse slavenhandel, 1500–1850 (Amsterdam, 2000), 65–66. 42. This breakdown in time and space of the African origins of the Suriname slave population demonstrates that the slave trade database is a productive tool for scholars trying to “historicize creolization,” as recently advocated by Richard Price. See his “The Concept of Creolization,” in David Eltis and Stanley L. Engerman, eds., The Cambridge World History of Slavery, vol. 3 (Cambridge, forthcoming). Naturally, the same applies to American slavery scholars of an Afrocentric disposition. In general, the database could prove critical in creating new links between African studies and African diaspora studies. See Patrick Manning, “Africa and the African Diaspora: New Directions of Study,” Journal of African History 44 (2003): 487–506. 43. Johannes Postma and Victor Enthoven, “Introduction,” in Postma and Enthoven, Riches from Atlantic Commerce, 10. 44. See, most recently, John J. McCusker and Russell R. Menard, “The Sugar Industry in the Seventeenth Century: A New Perspective on the Barbadian ‘Sugar Revolution,’ ” in Stuart Schwartz, ed., Tropical Babylons: Sugar and the Making of the Atlantic World, 1450–1680 (Chapel Hill, 2004), 295.

Appendix 8.1 Estimates of Slaves Carried from Africa under the Dutch Flag, 1596–1830 Years 1596–1600 1601–5 1606–10 1611–15 1626–30 1636–40 1641–45 1646–50 1651–55 1656–60 1661–65 1666–70 1671–75 1676–80 1681–85 1686–90 1691–95 1696–1700 1701–5 1706–10 1711–15 1716–20 1721–25 1726–30 1731–35 1736–40 1741–45 1746–50 1751–55 1756–60 1761–65 1766–70 1771–75 1776–80 1781–85 1786–90 1791–95 1801–5 1806–10 1816–20 1821–25 1826–30 Total

Number of Slaves

Number of Vessels

1,365 334 544 951 326 6,452 22,233 2,669 6,655 19,327 17,956 30,633 25,953 18,000 20,893 19,596 13,380 13,987 19,378 16,211 10,658 11,808 15,762 16,032 19,279 10,176 17,984 19,623 17,588 23,451 29,039 30,758 31,488 16,218 9,865 6,908 7,775 1,297 307 734 330 356 554,300

5 1 1 3 1 30 76 10 19 54 52 83 61 41 44 42 27 29 42 35 24 29 32 30 42 21 57 58 59 63 92 110 111 53 35 25 32 5 2 3 2 2 1,543

Source: Estimates page of http://www.slavevoyages.org. Note: No voyages were recorded for 1616–25, 1631–35, 1796–1800.

Chapter 9 The Slave Trade of

Northern Germany from the Seventeenth to the Nineteenth Centuries Andrea Weindl

Throughout the seventeenth century, the transatlantic slave trade of the northern European states was largely organized by national companies, but the history of slaving activities by these companies cannot properly be told in national terms alone. Early modern intercontinental trade involved participation in global markets, and the financial capital for such trade transcended national boundaries and institutions. Consequently, it is sometimes difficult to discuss the slave trade purely in terms of national flags. German slave traders played at most a minor part in the history of the transatlantic slave trade. We know that subjects of the German states were employed in service to the great European naval powers, especially the Netherlands, and that some German sailors, surgeons, carpenters, and locksmiths served onboard slave ships owned by other nationals. Such activities are not, however, the focus of this chapter, which concentrates instead on German political and financial involvement in the slave trade from the seventeenth century onward. The political fragmentation of the German states in the early modern era makes tracing German involvement in the slave trade difficult. 250

The Slave Trade of Northern Germany

Most German duchies were small and had little impact on international politics. This was true, for example, of the maritime cities of the Hanse, which failed to develop a centralized state of international significance. It is not by chance, therefore, that despite efforts by various scholars to trace German involvement in slaving, only the duchy of Brandenburg, which later became Prussia, has left clearly traceable evidence of participation in the slave trade.1 Some states failed to become involved either because of inadequate resources or because of international opposition. In other cases, the fact that slaving was carried on illegally by private traders means we lack the source materials to study it properly. This chapter will describe the efforts of German states and private merchants to participate in the transatlantic slave trade. Its emphasis will be on the ventures of Brandenburgers, which have attracted the attention of several historians and for which source materials exist. I have drawn on studies by Schück, Kellenbenz, Jones, and Brübach, as well as on my own research in the Geheimes Staatsarchiv Preußischer Kulturbesitz (GStA) in Berlin. Since 1993–94 the latter archive has been merged with documentary materials formerly deposited in a small East German town. Combining these materials with TSTD2 enables us to improve our understanding of the German role in the slave trade.

COURLAND INITIATIVES AND BRANDENBURG-PRUSSIA ENTERPRISE

Although there is evidence that a German slave trader purchased 150 slaves in Angola in 1841 with the intention of sailing to Brazil (the ship was captured by the Dutch), the initial German efforts to acquire colonies in West Africa and the West Indies date from the middle of the seventeenth century.2 In 1651 Jacob (James) Kettler, Duke of Courland, established a factory at the mouth of the Gambia River and two years later claimed ownership of Tobago in the Caribbean. Some years later, Courland lost its territories in Gambia and Tobago to rival European states and, despite frequent international negotiations, was unable to enforce its claims to them until the eighteenth century.3 It is unlikely that any slave voyages between Gambia and Tobago occurred, but in 1683 Captain Cornel Gildem and the Sereine left Amsterdam under the Courland flag and took on slaves in Gambia and Gorée with the declared intention of carrying them to Guadeloupe. We do not know the outcome of this voyage, but it is probable that Courlanders continued to take an interest in the

251

252

Andrea Weindl

slave trade in the second half of the seventeenth century, though actual voyages, if any, would have been only occasional. Unfortunately, little evidence relating to the overseas commerce of Courland has survived. The fairly good documentary sources generated by the English in Gambia have so far turned up nothing on Courland activity.4 By contrast, the efforts of the Duke of Kurbrandenburg to gain his own factories and promote overseas trade have left much stronger traces in the historical record and therefore have received considerable attention from scholars. A primary motivation of Brandenburg enterprises was Great Elector Friedrich Wilhelm’s desire for international recognition. In pursuing his goals, the elector was influenced by his uncle, the Duke of Courland, who had grown up in England and based his own plans for trade on English ventures. In Friedrich Wilhelm’s case, the model followed was that adopted by the Netherlands, where he had grown up. During the Thirty Years’ War, he tried to establish trading companies in both the East and the West Indies, but these efforts failed for political or financial reasons. It was not until Middelburg merchant Benjamin Raule joined the services of Brandenburg and the duchy founded its navy that a trading company could be successfully established in the 1680s. The initial bases for Brandenburg trading enterprise were Königsberg and Pillau, but neither proved particularly advantageous, and in 1682 the elector took advantage of a local dispute in East Frisia to transfer the headquarters of his commercial and naval enterprises to Emden, a port then considered one of the best in Europe. Even before this move, however, the first expedition to Africa under the Brandenburg flag had taken place in 1680 at the instigation of Raule and other Dutch merchants. For this venture the elector provided the crew for two ships and his flag; Raule and his partners provided the finance and carried the other costs.5 The ventures were to trade for gold, ivory (or “teeth”), corn, and slaves on the coasts of Guinea and Angola and to sell the slaves at Cádiz, Lisbon, and the Canary Islands, or secretly at some islands of the West Indies. For the Brandenburg court, the elector requested “half a dozen young and handsome slaves of 14, 15, and 16 years of age.”6 Despite some conflicts with ships of the Dutch West India Company, this first expedition succeeded in making trade agreements with three caboceers based between Axim and Cape Three Points.7 It was in Raule’s interest to reinforce this connection, and so in 1682 the Trading Company of the Coast of Guinea was created.8 Later known as the Brandenburg African Company (BAC), the company was founded by two charters issued in March and

The Slave Trade of Northern Germany

October 1682, with an intended capital of 50,000 thaler.9 The company proved to be unprofitable. Building and supporting trading posts in Africa and America proved expensive, and those involved in ventures seem often to have traded on their own behalf rather than on that of the company. In Africa agents traded with pirates, and money disappeared. The company also suffered from the hostility of other nationals. Brandenburg ships were repeatedly seized and recovered, if at all, only after extended proceedings, with the result that the company had to submit requests for new ships. By 1692 the BAC was bankrupt.10 Four years before the BAC became insolvent, however, the elector had chartered another American company with a permit for trading in slaves.11 Created at the instigation of some Scottish merchants, Spanish opposition and the outbreak of the Nine Years’ War in 1689 precipitated an early end to this initiative. In 1692 Friedrich Wilhelm transferred the rights of the former BAC through a new charter to the newly created Brandenburg African American Company (BAAC). Most of the new stockholders were Dutch merchants. Because they were not allowed to contract with other powers for overseas trade, their investments in the BAAC were registered in the name of a member of the Brandenburg navy council. Later the assets were transferred into stocks with a declaration of assignment.12 The charter of the new company was similar to its predecessor’s, although its organization resembled that of the Dutch West India Company. In common with other European states, Brandenburg tried to secure connections with the Gold Coast by creating forts and factories. In 1682 two ships were dispatched to Africa to renew treaties and build a fort. With local assistance, a fort named Großfriedrichsburg was built on the mountain Mamfort or Mumfort at Cape Three Points.13 A year later a similar arrangement was concluded with the caboceers of Akwida, nearly two and a half miles farther east. There the Brandenburgers built the fort called Dorotheenschanze, also acquiring a factory with access to the region’s granary and the only navigable port along the Gold Coast.14 They gained another post five miles farther east in 1685 but lost it two years later. To offset this loss and to protect connections between Großfriedrichsburg and Dorotheenschanze, in 1694 the Brandenburgers built another small trading post east of the former.15 Farther north, Cornelis Reers, sailing with the Zeelander under the Brandenburg flag, negotiated a treaty in 1687 with the king of Arguim, an island off modern-day Mauritania. The treaty conceded an exclusive trading right in exchange for rebuilding the European fort there. Raule held these trading rights until 1692,

253

254

Andrea Weindl

when they passed to the BAAC. From 1694 onward the Brandenburgers also maintained a small base on the Slave Coast at Ouidah, which they used to purchase slaves. From their earliest voyages to Africa, the Brandenburgers explored the possibility of slave trading. The Dutch West India Company impeded the trade of the two ships sent to Africa in 1682 by the Brandenburgers, but two years later Brandenburgers delivered 254 slaves to Berbice. In preparing for these arrivals, Raule concluded an agreement with the owner of the Dutch colony, Abraham van Pere. Even earlier, Raule had tried to obtain contracts to supply slaves. First, he negotiated with the Coymans brothers (Balthasar and Caspar) and with Josua and Pedro van Belle, who, together with the Dutch West India Company, were suppliers of slaves to the Spanish under the asiento. In 1683, while Spanish needs were temporarily being supplied under licenses issued by the Casa de Contratación, Raule negotiated directly with the Spanish Crown for consignments of 2,000 to 3,000 piezas de Indias per year. In doing so, Raule was motivated not only by profit but also by a desire to eliminate the hated Dutch company.16 These various negotiations failed because Brandenburg did not possess a base in the Caribbean through which to conduct trade, and neither the Coymans brothers nor the Spanish were interested in allowing an unregulated trade in slaves in Brandenburg ships. In the early 1680s the BAC sought to negotiate the lease or purchase of a base from the French or the Danes, and in November 1685 Raule signed a treaty with Denmark to rent St. Thomas in the West Indies for thirty years. This agreement allowed the BAC to maintain a post on the island and to own as much land as it could farm with 200 slaves. The treaty also established the rules under which the Brandenburg slave trade might be conducted. Trade was allowed only to the Danes and to the BAC. Should foreign merchants seek to sell slaves at St. Thomas, it was agreed that the Danish governor of the island and the director of the BAC would be permitted to purchase them in equal shares.17 The treaty also provided in vague terms for mutual support between the Danes and Brandenburgers in the purchase of the slaves at Guinea, in cases where their ships might come into contact.18 Differences in interpretation of the treaty and constant smuggling by Brandenburgers soon caused problems with the Danes, however, and during the late 1680s the search began for another West Indian base. In 1689 Brandenburg seized the island of St. Peter, two miles south of Tortola, which, in addition to serving as a potential base for slaving, also became a base for privateering and for tax avoidance by traders.19

The Slave Trade of Northern Germany

We possess only limited evidence of how the Brandenburgers conducted their slave trade in Africa. Dutch merchants conducted most of the trade, and in the seventeenth century the market organization for slaving was developed in partnership with African trading partners. In this respect, the Brandenburg slave trade in the African sphere worked like that of other European nations. Although Brandenburgers expected to buy slaves at Arguim in North Africa, it is uncertain whether they achieved regular deliveries from that place to the Americas. In the fifteenth century, Portuguese traders had diverted the caravans designed for the trans-Saharan slave trade to Arguim and had taken these slaves to Portugal. When the French came to Arguim in 1638, the transSaharan routes that the Portuguese had drawn on had shifted further east. In 1691–92, however, it was reported that several hundred slaves were brought from Arguim to the West Indies, and in the 1699 minutes of the nine directors of the BAAC there are instructions that suggest attempts to introduce slave shipments from Arguim to the Canaries.20 But evidence also exists of plans by Brandenburg to exchange the African base for a French island in the Caribbean, suggesting that from the Brandenburg perspective, Arguim had yielded little of significance in terms of developing the slave trade.21 A Prussian merchant nevertheless estimated in 1709 that 100 slaves a year could be purchased at Arguim, though whether this was realized is unclear.22 Similarly, there is little clarity in the sources regarding the importance of the fort at Großfriedrichsburg for the Brandenburg slave trade. It is probable that the number of slaves held for sale at the fort was usually no more than twelve at one time, although the number of working slaves at the fort might exceed this.23 Because of high maintenance costs, the BAAC decided in 1699 to reduce the number of working slaves at the fort from 216 to 60 and to send the rest to St. Thomas in the West Indies. Probably some of the 421 slaves brought to St. Thomas the following year came from Großfriedrichsburg.24 During the 1692 voyage of the Fredrik Wilhelm (voyageid 21,950), the only Brandenburg voyage for which detailed accounts survive, slaves were purchased for the first time at Accra, about 110 miles to the northeast of Cape Three Points.25 At that time Accra was reputed to be one of the best places to acquire slaves because local wars supplied captives constantly.26 African points of trade are not as well depicted in the sources as their American counterparts, but, as with the Dutch and the English in the late seventeenth century, those flying the Brandenburg flag bartered for the majority of their slaves at the ports of the Slave Coast. At about the time the Brandenburgers started slave trading, the center of the trade at the Slave Coast was moving from the kingdom of Allada to the smaller

255

256

Andrea Weindl

coastal state of Ouidah, formerly a province of Allada.27 Whether the pattern of Brandenburg trade in the region reflected these trends is unclear, but it appears that Allada remained the major source of supply for the slaves carried by Brandenburg ships.28 TSTD2 shows Allada dispatching 85,000 slaves before 1700, with the peak occurring in the 1680s, but how many of these left on Brandenburg ships is unclear. The Danish forts and Ouidah appear to have been the major embarkation points. Some of the Brandenburg slaves originated in the Bight of Biafra. Known as Calabari (likely Igbo) and probably taken from the Cross River, slaves from the Bight of Biafra were not well regarded by contemporary buyers in the Americas, who were concerned about their desire for freedom and resistance to slavery.29 In 1699 the BAAC gave the order not to bring any more slaves from Calabar for fear of “spoiling” the rest of the slave cargo.30 In the 1690s the BAAC also tried to extend its trade farther south in Angola. Three passports were issued for trade to this region between 1692 and 1698.31 The company’s accounts for 1694, moreover, show Angola to have been the most important single destination for buying slaves. Why the BAAC sought to extend its slaving activities to Angola at this time is unclear, but it may have been linked to simultaneous negotiations with the Portuguese Companhia de Cache˜u over supplying slaves under the asiento.32 The procedures of the BAC and BAAC for embarkation and transport of slaves were similar to those followed by the Dutch West India Company. Physicians assisted Brandenburg traders in negotiating slave purchases with local dealers and in selecting slaves for purchase. Once bought, the slaves were branded, in the case of the BAAC, with the letters “CABC” prior to boarding ship.33 After leaving their place of trade, most Brandenburg slave transports stopped at São Tomé in the Gulf of Guinea to take in fresh water and provisions before crossing the Atlantic.34 Records of shipboard mortality for only four voyages have survived, and these suggest an average mortality of 13.4 percent of those taken onboard on the African coast, but even with this small a sample the range of mortality was wide. Both the average and the range are not inconsistent with what has been observed in other branches of the trade during this period. The merchants of the BAAC actually allowed for a mortality loss of 8 percent.35 This appears low by contemporary standards. Brandenburgers adopted three main strategies for selling slaves once their ships reached the Americas. One method involved negotiating contracts and supplying slaves in lots. Another was to smuggle slaves to other parts of the Caribbean, while a third involved selling slaves to the planters of St. Thomas.36

The Slave Trade of Northern Germany

The last probably constituted the smallest part of the Brandenburg trade, but its significance to the growth of the plantation sector in St. Thomas should not be understated. If the Brandenburg transatlantic system was an attempt at a miniature replica of those of the larger European powers, then St. Thomas was in some senses a tiny version of Barbados. It had a small plantation sector that grew explosively in the seventeenth century and produced sugar steadily into the nineteenth century, but it was also an entrepôt that supplied other plantation regions of the Americas. While St. Thomas was never a Brandenburg possession, Brandenburg traders established the base slave population in the 1690s, and as with Barbados, most of those first slaves came from the Gold Coast. Their arrival made possible the island’s boom in the 1690s and early eighteenth century. The colony’s first census in 1688 revealed a population of 317 whites and 422 blacks, but by 1697 an English report suggested there were about 1,500 “working negroes and seven sugar-works” at St. Thomas, attributing this expansion to Brandenburg slave suppliers. By 1715 there were said to be 547 whites and 3,042 blacks living on the island, close to its peak population before the nineteenth century.37 Brandenburg traders were not responsible for all the post-1690 growth in St. Thomas’s slave population, but in these early days they were the major suppliers. The entrepôt role of the island becomes clear when we recognize that 13,000 slaves arrived under the Brandenburg flag between 1686 and 1703, while the slave population did not reach 4,000 until 1720. Other traders, including Danes and interlopers, supplied slaves to the island thereafter. The terms of the 1685 treaty negotiated with Denmark brought gains both to the island’s Danish governor and to the Brandenburg director of the BAAC. Permitted under the treaty to buy slaves equally and sell them at market prices, the governor and director appear to have earned profits of between 25 percent and 100 percent from such activities.38 For most of the time, however, Brandenburg activity in the Americas centered on supplying slaves to purchasers outside St. Thomas, whether on contract or through smuggling. A primary goal of the BAC and BAAC was to sell slaves in Spanish America. The Spanish authorities, however, treated Brandenburg’s entry into the slave trade with suspicion. In February 1686, even before company employees arrived at St. Thomas, the Spanish king ordered the governor of Puerto Rico to be particularly vigilant about the activities of Brandenburgers and emphatically forbade the unloading of Brandenburg ships at Spanish ports in the Americas.39 Efforts by the Spanish to prevent Brandenburg infiltration of the Spanish American slave market were only partially successful, however. A Dutch merchant working in the service

257

258

Andrea Weindl

of the Germans traveled on one of the first company ships to St. Thomas and later traded along the shore of the Spanish West Indies.40 Furthermore, the decline of Dutch control over the asiento following the death of Balthasar Coymans in 1686 and the ensuing prolonged struggle for the contract among the Dutch, English, and French created opportunities for smaller trading countries to engage in supplying slaves to the Spanish colonies. Within four years of Coymans’s death, his plenipotentiary in the Americas, Pedro van Belle, had entered the service of Brandenburg.41 Between 1690 and 1693, Brandenburgers were able to benefit from the confusion over the asiento by trying to meet some of the demand for slaves in the Spanish territories.42 One arrangement allowed them to deliver 600 slaves four times a year to the Spanish.43 It appears that only in 1693 were they able to meet this target, and we do not know how successful they were in subsequent years or if, in attempting to meet it, they resold slaves from interlopers and pirates. In 1694 the BAAC made an agreement with the holder of the asiento, under which the company was allowed to supply 2,000–3,000 slaves annually to the asientistas.44 The Brandenburgers also bargained in vain for contracts with the French as well as with the Portuguese Companhia de Cache˜u, which had temporarily taken over delivery of slaves from the West India Company, and when opportunities to work with the holders of the asiento were lacking, they tried to infiltrate Spanish American markets by dealing with the interlopers of Zeeland.45 Those trading under the Brandenburg flag also sought to supply slaves to non-Spanish markets in the Americas. One potential outlet was the English West Indies. In 1697 Governor Codrington of the English Leeward Islands noted the contribution of Brandenburg traders to the thriving slave trade at St. Thomas and warned that this might entice poor planters in the Leeward Islands and Barbados to move to Tortola, where they could access slaves from Brandenburg suppliers on highly favorable terms.46 The reverse—that Brandenburg traders might supply slaves to the Leeward Islands—was also possible and seems to have been an option that Codrington himself chose to exercise.47 Planters and traders sometimes journeyed to St. Thomas when a slave delivery from Guinea was expected. In other cases, middlemen redistributed slaves, particularly when war made the sea routes dangerous. There are references to trade connections as far north as Carolina.48 The English also kept an eye on the smuggling of slaves from St. Thomas to the Spanish colonies, fearing competition for Jamaica, which was beginning to supersede Curaçao as a supply center for asientistas.49 In 1693 the governor of Jamaica reported agents of

The Slave Trade of Northern Germany

the asiento taking slaves from St. Thomas rather than Jamaica, noting that because there were at Jamaica “no negroes. . . . to supply the Assiento, Sir James Castile50 sent four sloops to St. Thomas with about £300,000 in money in hopes of securing negroes there.”51 Other Europeans received slaves from those flying the Brandenburg flag. Brandenburgers were said to have near exclusive control of the slave supply to the Danish island of St. Croix between 1690 and 1696.52 This connection was likely to have built on links between St. Croix and St. Thomas dating from the 1670s, when Dutch merchants had sold slaves from Curaçao via St. Thomas.53 There is evidence, too, of connections with the French Caribbean, notably through Martinique and Saint-Domingue, especially Petit-Goâve in southern Saint-Domingue.54 The French were unable to meet the demand for slaves in their colonies at this time, and the problem was compounded by war in Europe. In 1692 the planters of Saint-Domingue were given permission to buy their slaves wherever they could.55 Two years later, Ducasse, the governor of Saint-Domingue, rejoiced in the fact that “for two years this colony has shown a brilliant luster . . . 8 or 900 Blacks have come from St. Thomas; one no longer has to fear a shortage.”56 In the same year, Brandenburgers opened a branch at Saint-Domingue, accepting indigo as advance payment for slave deliveries, and were probably instrumental in supplying slaves from St. Thomas to French planters in Saint-Christophe.57 In seeking to trade with the colonies of other nations, Brandenburg slave traders became embroiled in international conflicts when they spread to the West Indies. The effects were mixed. In 1691 the English requested support from three Brandenburg ships in an expedition against the French.58 Three years later, the Dutch seized two Brandenburg ships trading at Curaçao.59 The damaging effects of war on the Brandenburg fleet were sometimes offset by benefits, as in 1691–92: after the French plundered the English islands of Antigua, Nevis, and Montserrat, the English sought to restock their colonies with slaves by acquiring eight shiploads at St. Thomas, some doubtless supplied by Brandenburg traders.60 By the close of the century, however, the complaints by Brandenburg interests about the difficulties they had with trading in the French and English islands seem to have grown, as the French and English sought to exclude foreign vessels trading with their colonies. The measures taken to inhibit Brandenburg infiltration included restrictions on debt collection, with the result that by 1699 Van Belle, a BAAC employee, complained of difficulties in making return cargoes because his ships were prevented from calling at foreign ports.61

259

260

Andrea Weindl

Van Belle’s complaints were part of a wider catalogue of challenges facing Brandenburg traders that were to intensify and ultimately ruin them. Colonial ventures seem to have brought nothing but losses to the duchy in the main. This may have been more apparent than real, for the BAAC accounts were kept so poorly that Raule was suspected of fraud and eventually went to prison.62 Profits made on voyages were, in any case, distributed to BAAC shareholders rather than reinvested. Some losses were nevertheless real. Between 1694 and 1697, a French pirate attack on St. Thomas proved costly when fire broke out in the company’s storeroom and one of the company’s captains took his ship and joined the pirates. Other nations regularly seized Brandenburg ships, and a dispute with the Danish authorities at St. Thomas prompted the island’s governor to seize Brandenburg goods and ships. Poor business practices and disputes drained the confidence of potential investors, restricting the BAAC’s abilities not only to reequip its own ships but also to obtain credit in the Netherlands for doing so.63 Between 1699 and 1709, only a few ships fitted out by the BAAC went to sea, and some of these were lost. Unable to supply its own trading posts in Africa adequately, the company came to depend on Dutch interlopers to do so. Private traders based in Emden or Zeeland encroached on the BAAC charter, using Prussian passports and the Prussian flag as well as BAAC forts but trading on their own account. This continued until 1715, with some of the profits being distributed to the BAAC and later to Prussian King Friedrich I, who in 1711 seized the company’s assets without resistance from the shareholders. Friedrich’s successor, Friedrich Wilhelm I, effectively liquidated what was left of the company. In 1717 he sold Großfriedrichsburg to the Dutch West India Company in exchange for 6,000 thaler and twelve young Africans—for some years the Prussians had tried to get Africans to serve in their army as musicians.64 In 1725 the last effects of the company at Emden were auctioned, and six years later the Danes confiscated the company’s factory and lodges at St. Thomas to cover outstanding debts.65 The ultimate failure of the BAAC and Brandenburg’s intervention in the Atlantic slave trade should be considered within a broad context. The BAAC was not the only company involved in African trade to face severe problems and ultimately to collapse. The English Crown chartered several monopoly companies in the seventeenth century (the two largest in 1662 and 1672, respectively), before eventually opening the African trade to all its subjects in 1698.66 The French experimented with various companies but in the seventeenth century were unable to meet the labor demands of their Caribbean

The Slave Trade of Northern Germany

colonies.67 The Danes sent ships to Africa for slaves to supply their West Indian colonies at same time as the Brandenburgers. The difficulties of the Dutch West India Company are widely known, and it lost its monopoly after 1730. Brandenburg’s limited involvement in slave trading and the trade’s relatively early ending were only partly, therefore, institutional failures. The BAAC’s collapse owed something to internal problems, but it was linked also with wider infrastructural and political problems. In Europe, the BAAC had access to merchant capital through the Dutch but lacked the means to market and process colonial products. The Great Elector had wanted to participate in colonial trade, but his system was never fully completed. To realize profits in the Americas required secure bases and colonies that could be kept dependent and under control. The Brandenburg authorities failed to match their support for the BAAC by engaging in colonial ventures in the Americas. Small as their slave trade was, it was much larger than what St. Thomas or indeed all the Danish islands together could absorb, given the Danes’ parallel activity in the slave trade. Perhaps separate plantation colonies in the Americas were beyond the means of a sparsely populated duchy. Herein, perhaps, lies the underlying problem of Brandenburg’s Atlantic slave-trading venture. That said, we now know that at least fifty-six slaving voyages were dispatched to Africa under the flag of Brandenburg during its history of involvement in this traffic. These voyages carried an estimated 22,750 slaves from Africa, just 18,400 of whom arrived in the Americas alive. In view of the weaknesses in BAAC accounting, this should probably be considered a minimum figure. It does not include Brandenburg involvement in intra-Caribbean trafficking. Even with these caveats, Brandenburg’s slave trade was small by the international standards of the age and tiny relative to the overall volume of the Atlantic slave trade. It need scarcely be said that the human misery this trade represents is clearly beyond the quantifiable. In terms of its temporal distribution, the Brandenburg slave trade was not without significance. Figure 9.1 shows the arrivals of slaves in the Caribbean in the last fourteen years of the seventeenth century, distributed by the carriers’ national flags. Most of this period was characterized by war at sea, which certainly reduced the traffic of the major powers. Nevertheless, while the English were already the dominant slave traders in the region at the beginning of the period, Brandenburgers were the second largest carriers in 1692 and 1696 and the third largest, ahead of the French and the Danes, in every year except 1686, 1689, 1695, and 1697. Their trade reached its peak in 1692–93, when an estimated

261

Andrea Weindl

Imputed nation of slave registration French Dutch Danish English Brandenburg

10,000

8,000

Number of slaves disembarked

262

6,000

4,000

2,000

0 1699

1698

1697

1696

1695

1694

1693

1692

1691

1690

1689

1688

1687

1686

Years

Figure 9.1 Annual Arrivals of Slaves in Caribbean Territories Distributed by National Flag of Carrier, 1686–99 Source: TSTD2.

3,717 Africans arrived in the Americas in Brandenburg ships, more than was carried by the Royal African Company in those years and far more than the 900 disembarked by Dutch West India Company ships. That single year actually accounts for more than one in five of all the slaves carried to the Americas in Brandenburg slave vessels. It was not by chance that other nations then began to view German involvement in the trade seriously and to take measures against it. Neither before nor afterward were Brandenburg carriers ever able to match the slave deliveries of 1693, but in the 1690s they shipped on average 1,200 slaves a year in what was the most “successful” period of the slave trade from the Brandenburg point of view. Well over half of Brandenburg’s slave trade was concentrated in this one decade.

The Slave Trade of Northern Germany

Brandenburg activity at this time sheds light on the shifting balance of power and interest among the different European nations in the Caribbean in the late seventeenth century. By this time the economic position of the Dutch in the Americas was clearly in decline although, given the Dutch financing of Brandenburg slaving activities, that decline may not have been so pronounced as is sometimes assumed.68 The struggle for economic supremacy in the Caribbean by the 1690s, however, was clearly between the English and the French. Brandenburg participation in the slave trade occurred during the upheaval caused by the shift in the balance of power from the Dutch to others. Supported by Pedro van Belle, the Germans exploited Dutch connections with Spanish America and other parts of the Caribbean to gain entry into American slave markets. In retrospect, Van Belle’s association with the BAAC provides an index of its fortunes. He used the BAAC for his own benefit when the influence of the Dutch West India Company began declining. Then he retreated about 1702 to his own plantation on the British island of St. Christopher (later known as St. Kitts), when the BAAC no longer offered him significant commercial opportunities.69 More generally, the Brandenburg involvement in the slave trade was always conditional on broader political and economic contexts. Its peak years of involvement coincided with changes in international politics, war in the Caribbean, and the uncertainty this created in the 1690s.70 The ending of the Brandenburg slave trade came quite abruptly after 1700, not only because of financial constraints but also because the English and French chose to stop buying slaves from the Germans.

THE SLAVE TRADE OF THE HANSE TOWNS

In the age of mercantilism, the Hanse towns along or near the Baltic and North Sea coasts could participate in the slave trade only under the protection of other nation-states. One early involvement centered on the Danish African Company of Glückstadt, a port northwest of Hamburg. The company was founded in 1659, predominantly with resources from Hamburg, but it ran into difficulties and in 1671 became associated with the newly founded Danish West Indian Company based in Copenhagen. The new company traded with varying degrees of success between Guinea and the West Indies, often in competition with Brandenburg traders along the same routes, starting in the 1680s. At that time, the Danish company recruited many of its staff from Hamburg and Holstein. Further efforts by Hamburg merchants to enter the slave trade occurred after the War of the Spanish Succession, when they sought to establish a new

263

264

Andrea Weindl

company on the Elbe River. One vessel, the King of Prussia, reached St. Thomas with 212 slaves in 1715, but this effort was generally unsuccessful largely for political reasons, and from then until late in the eighteenth century there is only one further Hamburg voyage on record—in 1768—although there were certainly other slaving ventures to Africa conducted under foreign passports or as contraband trade.71 There are few sources about such activity, and we cannot easily distinguish between Danish and Hamburg trade. At the end of the eighteenth century, an interesting interlude of relatively free trading in slaves developed for the only time in the slave trade’s history. The British, the Dutch, and the Danes all allowed free entry into many of their ports for vessels carrying slaves.72 After 1789 the Spanish also gave free access to vessels carrying slaves to all their ports in the Americas. Thus, except for the French Caribbean—largely shut down during the 1790s by revolution and war—there was virtually no jurisdiction in the Caribbean that did not participate in a freer slave-trading environment. It did not last long, however. Beginning with the Danes in 1802, one nation after another instituted a different set of restrictions in the form of gradual abolition of the slave trade, until the last vessel crossed the Atlantic in 1867. In the preceding short period of freer trade, from about 1789 to the British and U.S. abolition of the slave trade commencing in 1808, Hanse town merchants organized a minirevival of the North German slave trade, dispatching at least eight slave voyages between 1798 and 1805. These vessels obtained slaves at Accra and Christiansborg on the Gold Coast or, in two cases, in Upper Guinea. They then proceeded to a range of ports in the New World where slaves could be sold without restriction. These included Havana, Montevideo, Suriname, and St. Thomas. As noted, Danish, British, and U.S. abolition efforts brought this interlude to a close, and it is clear that Hanse merchants were never able to capitalize on the new international environment in the 1790s the way the Brandenburgers had a century earlier. Thereafter, a different kind of Hanseatic involvement developed. After 1815 the British suspected Hamburg merchants of participating in what they now considered to be an illegal slave trade, with Hamburg merchants of Portuguese ancestry in particular being thought to use connections with Angola and Brazil to pursue it. In the 1830s Hanse merchants sent vessels carrying merchandise to be exchanged for slaves to the African coast, and in some cases they were suspected of transferring such vessels to Portuguese owners, with the result that the same vessels carried slaves to the Americas. The Hanse towns signed a treaty with Britain by which cruisers of the two powers had the

The Slave Trade of Northern Germany

right to arrest suspected ships, conduct them to their respective domestic tribunals, and attempt to get convictions. The British navy detained several vessels under the Hanseatic flag and accompanied them back to Bremen, where the vessels were invariably released without being convicted.73 These activities have not been researched exhaustively.74 Tables 9.1–9.3 summarize what is known about the size and geography of 160 years of slave trading from northern Germany. Table 9.1 shows that the minirevival of the traffic in the late eighteenth century, this time without the involvement of a monopoly company, meant that in total sixty-six ships carried an estimated 25,169 slaves from Africa and disembarked an estimated 21,209 in the Americas. (This excludes a voyage to Brazil in 1839, sailing under the flag of the Hanseatic League, which was owned by Portuguese slave traders who adopted that flag to escape the attentions of British antislave patrols.) Table 9.2 shows that the sample of voyages with information on African coastal origins of slaves is small. The preponderance of Gold Coast and the Bight of Benin origins—equivalent to 84 percent of all German-transported slaves for whom origins are known—that these data show is, however, likely to be representative of the traffic as a whole and is consistent with earlier discussions. Table 9.3 shows that in the Americas, where information on places of disembarkation is more complete, the Danish Virgin Islands played a similarly dominant role for the North German traders. It is also clear that the private traders during the minirevival sought a wider range of markets in the Americas than had their counterparts of a century earlier. Table 9.1 Estimated Number of Slaves Carried by North German Slave Vessels in Ten-Year Intervals Years 1641–50 1681–90 1691–1700 1701–10 1711–20 1761–70 1791–1800 1801–10 All years Source: TSTD2.

Voyages

Slaves Departing

Slaves Arriving

1 10 37 7 1 1 6 3 66

150 2,865 17,532 1,894 265 330 1,605 529 25,170

130 2,277 14,461 1,515 212 269 1,450 452 20,766

265

266

Andrea Weindl

Table 9.2 Estimated African Origins of Slaves Carried by North German Slave Vessels, 1641–1810 Senegambia 879

Gold Coast

Bight of Benin

Bight of Biafra

Angola

2,413

3,151

51

150

Source: TSTD2.

Table 9.3 Estimated Destinations in the Americas of Slaves Carried by North German Slave Vessels Years 1641–1720 1761–1810 All years

Dutch Caribbean

Dutch Guianas

Río de la Plata

Virgin Islands

Barbados

Cuba

1,404 0 1,404

254 266 520

0 232 232

13,912 618 14,530

200 0 200

0 220 220

Source: TSTD2.

CONCLUSION

German involvement in the Atlantic slave trade was always dependent on broader political and economic conjunctures, as the German states were unable to provide sufficient resources to promote trading companies on their own. Their involvement was always constrained as well by the lack of German plantation colonies in the Americas. German traders, whether from Courland, Brandenburg, or the Hanse ports, thus depended on groups from other nations to buy the slaves they delivered to the Americas. In this respect, German participation in the slave trade depended on the power and goodwill of other European nations. In Africa, Brandenburg managed for some time to acquire factories and forts as bases for trade. The BAAC was also for a time able to exploit political conflicts and shifting balances of power among the major European nations to develop markets for slaves among otherwise inadequately supplied non-German planters and other buyers. The African ventures of Brandenburg-Prussia were also made possible by Dutch capital and the ambitions of Great Elector Friedrich Wilhelm. They always lacked, however, a commercial infrastructure at home and thus remained dependent on external factors, including, in the age of mercantilism, the willingness and power of other nations to block trade. Although the Germans did not take a major role in the slave trade, the history of their

The Slave Trade of Northern Germany

involvement, especially that of Brandenburg-Prussia, provides an interesting example of how smaller states tried to share this trade in the seventeenth and eighteenth centuries. It also shows how merchants outside the Germanspeaking states of Europe used the ambitions of these states for their own commercial and financial gain. An overview of nearly 200 years of intermittent slave trade–related activity by German polities allows us to discern a pattern of sorts. The Brandenburg slave trade was most active when one of two conditions held. The first was war between the major slave-trading powers, which could disrupt the normal flow of slaves and provide an opportunity for a smaller neutral flag to take up some of the resulting slack. This was the situation in the 1690s, when the Brandenburg traffic peaked. The second condition was when one or more powers with a presence in the plantation Americas opened their markets in times of peace to particular German states, as Denmark did in the seventeenth century and as several major colonial powers did in the late eighteenth. Neither condition held for most of the era of the slave trade, and the involvement of the North German ports remained a peripheral activity except for a few years in the 1690s. NOTES

1. See Heinrich Volberg, Deutsche Kolonialbestrebungen in Südamerika nach dem Dreißigjährigen Krieg (Cologne, 1983). 2. The 1841 episode is included in TSTD2 as voyageid 11,653. 3. See “Vorgeschichte der deutschen Kolonien und Schutzgebiete,” http://www .deutscheschutzgebiete.de/vorgeschichte_der_deutschen_kolonien_und_schutzgebiete.htm. 4. See Otto Heinz Mattiesen, Die Kolonial- und Überseepolitik de kurländischen Herzöge im 17. und 18. Jahrhundert (Stuttgart, 1940). 5. See Richard Schück, Brandenburg Preußens Kolonialpolitik unter dem Großen Kurfürsten und seinen Nachfolgern (1647–1721) (Leipzig, 1889), 2 vols. Vol. 1 covers the history of the colonial politics of Brandenburg Prussia. Vol. 2 contains some important documents on the issue. This reference is to vol. 1, pp. 142–43. 6. “[E]in halb Dutzend junge Sklaven von 14, 15 und 16 Jahren, welche schön und wohlgestalt seien.” Schück, Brandenburg Preußens Kolonialpolitik, vol. 2, pp. 95–96. 7. Caboceer was derived from the Portuguese term caboçeira. Europeans used the word for noblemen or chieftains, especially of the Gold Coast. See Adam Jones, Brandenburg Sources for West African History, 1680–1700 (Stuttgart, 1985), 313. 8. See Schück, Brandenburg Preußens Kolonialpolitik, vol. 2, no. 63. 9. The company was also called “Guineische Compagnie” or “Africanische Compagnie.” See Eberhard Schmitt, “Die Brandenburgischen Überseehandelskompanien im XVII. Jahrhundert,” Schiff und Zeit 11 (1984): 6–20, esp. 9.

267

268

Andrea Weindl

10. See Peter Feddersen Stuhr, Die Geschichte der See- und Kolonialmacht des Großen Kurfürsten Friedrich Wilhelm von Brandenburg in der Ostsee, auf der Küste von Guinea und auf den Inseln Arguim und St. Thomas, aus archivalischen Quellen dargestellt (Berlin, 1839), 41. 11. See Schück, Brandenburg Preußens Kolonialpolitik, vol. 2, no. 123, par. 3. 12. See ibid., vol. 1, p. 236. 13. Construction lasted until 1708. Nearly all materials had to be brought from Europe so that the sum of 36,000 thaler spent by the Great Elector was insufficient to cover the costs. See Schück, Brandenburg Preußens Kolonialpolitik, vol. 1, p. 320. See also ibid., vol. 2, no. 74, Requisita zur Festung in Africa vom 20./10. Juni 1683. 14. See Schück, Brandenburg Preußens Kolonialpolitik, vol. 1, p. 322. 15. See Jones, Brandenburg Sources, 4. The post was called Tacrama or Taccerama and was located near what is today called Westpoint. 16. “Ein jeder weiß, dass der Sklavenhandel die Source des Reichthums ist, den die Spanier aus ihren Indien holen, und dass derselbe mit ihnen den Reichthum theilet, der die Sklaven anzuschaffen weiß. Wer weiß wieviel Millionen baar Geld die niederländische westindische Kompagnie aus dieser Sklavenlieferung an sich gebracht!” (Raule to Elector, Copenhagen, October 26, 1685, in Schück, Brandenburg Preußens Kolonialpolitik, vol. 1, p. 192). 17. See the treaty between Elector Friedrich Wilhelm of Brandenburg and King Christian V of Denmark concerning the island of St. Thomas, dated November 24, 1685, in Schück, Brandenburg Preußens Kolonialpolitik, vol. 2, no. 103, art. 20. 18. Ibid., art. 21. 19. See Nils Brübach, “ ‘Seefahrt und Handel sind die fürnembsten Säulen eines Estats’: Brandenburg-Preußen und der transatlantische Sklavenhandel im 17. und 18. Jahrhundert,” Lateinamerika-Studien 32 (1994): 30. 20. See GStA, I.HA, Rep. 65, No. 24, Blatt 23. These plans were probably never realized. 21. See GStA, I.HA, Rep. 65, No. 24a, Blatt 28f. 22. See Schück, Brandenburg Preußens Kolonialpolitik, vol. 2, no. 170. 23. See Jones, Brandenburg Sources, 6. 24. See GStA, I.HA, Rep. 65, No. 24, Blatt 24, and No. 40c, Blatt 32. 25. See Paul Oettinger, ed., Unter kurbrandenburgischer Flagge: Deutsche KolonialErfahrungen vor 200 Jahren; nach dem Tagebuche des Chirurgen Johann Peter Oettinger (Berlin, 1886), 51f. 26. See Hugh Thomas, The History of the Atlantic Slave Trade, 1440–1870 (London, 1997), 349. 27. See Patrick Manning, “The Slave Trade in the Bight of Benin, 1640–1890,” in Henry A. Gemery and Jan Hogendorn, eds., The Uncommon Market: Essays in the Economic History of the Atlantic Slave Trade (New York, 1979), 114. 28. See Robin Law, “Dahomey and the Slave Trade: Reflections on the Historiography of the Rise of Dahomey,” Journal of African History 27 (1986): 240. 29. See Johannes Postma, The Dutch in the Atlantic Slave Trade, 1600–1815 (Cambridge, 1990), 106–10. 30. See GStA, I.HA, Rep. 65, No. 24, Blatt 25, where it is said that “is geresolveert . . . dat geen Slaven meer uijt de Calbarij sullen werden gehaalt om d’andre daar meede niet

The Slave Trade of Northern Germany

instalug te maken.” In 1698, when the Charlotte Louize was taken by English pirates, the ship was cruising off the coast of Calabar. See Jones, Brandenburg Sources, doc. 89. 31. Ibid., 165. 32. See Schück, Brandenburg Preußens Kolonialpolitik, vol. 2, no. 144. 33. See Oettinger, Unter kurbrandenburgischer Flagge, 61. This is the only description of a Brandenburg slave transport, but we can suppose that most of the transports took a similar course. 34. Despite high prices, most slave ships used the island as the last stop before the Middle Passage. See Thomas, History of the Atlantic Slave Trade, 362. It is known that in 1698 the Kurprinzessin also stopped at São Tomé. See GStA, I.HA, Rep. 65, No. 28a, Blatt 306. 35. Ibid., No. 18b, Blatt 122. 36. See Brübach, “Seefahrt und Handel,” 34. 37. John J. McCusker, Rum and the American Revolution: The Rum Trade and the Balance of Payments of the Thirteen Continental Colonies, 1650–1775 (New York, 1989), 710. 38. See Waldemar Westergaard, The Danish West Indies under Company Rule (1671–1754) (New York, 1917), 148 and appendix J. 39. See Manuel Gutiérrez de Arce, La colonización danesa en las Islas Vírgenes: Estudio histórico-jurídico (Seville, 1945), 34–36. 40. See Schück, Brandenburg Preußens Kolonialpolitik, vol. 2, no. 113. 41. I. A. Wright, “The Coymans Asiento (1685–1689),” in Bijdragen voor Vaderlandsche Geschiedenis en Oudheidkunde, 6 (1924): 31. 42. See Brübach, “Seefahrt und Handel,” 32. The situation got so confused that nobody knew whether the treaties between Spain and Coymans or his successor, Carcau, were in force or not. In 1688 there were about 5,000 slaves at Curaçao who could not be sold within the treaties, and so the following year the island was declared an open market. See Postma, Dutch in the Atlantic Slave Trade, 43–46. 43. See Brübach, “Seefahrt und Handel,” 32. 44. A 1694 report from Danckelmann and Raule to the elector said that “in onderhandeling sijn getreden met een Coopman die van de Cooning van Hispanien geobtineert hebbende het assiento om negros Slaven in de spanshe Westindien tebrengen . . . bij succes van welck Accordt en Contract, om aen haer 2000–3000 Stucks jaerlijks televeren” (GStA, I.HA, Rep. 65, No. 18b, Blatt 13). On the asiento holders at this time, see Georges Scelle, La traite négrière aux Indes de Castile: Contrats et traités d’assiento (Paris, 1906), vol. 1, pp. 693–700. 45. See GStA, I.HA, Rep. 65, No. 24, Blatt 15. 46. Governor Codrington to Council of Trade and Plantation, Antigua, September 27, 1697, in J. W. Fortescue, ed., Calendar of State Papers, Colonial Series, America and West Indies (henceforth Cal. Col.) (London, 1904), vol. XIV, No. 1347. 47. See Cal. Col., vol. XIV, No. 997; vol. XVI, No. 431. 48. See Westergaard, Danish West Indies, 148, and Cal. Col., vol. XV, No. 864 III. 49. See Governor Sir Nathaniel Johnson to Lords of Trade and Plantations, June 2, 1688, Cal. Col., vol. XII, No. 1773. Johnson’s report notes, “There is now a Danish African Company established to carry on a trade with the Spaniards in negroes, which will be

269

270

Andrea Weindl

very detrimental to Jamaica and to the Royal African Company also, by sending slaves clandestinely into these islands which is constantly done.” According to Westergaard, no slave vessels registered in Denmark reached St. Thomas between 1688 and 1700. See Westergaard, Danish West Indies, appendix J. 50. This was Santiago Castillo, who worked as a factor for several asientistas. At this time he worked for the asiento holder, Nicolas Porcio. See Elizabeth Donnan, ed., Documents Illustrative of the History of the Slave Trade to America (New York, 1969), vol. 1, p. 327. 51. Lieutenant-Governor Sir William Beeston to the Earl of Nottingham, June 10, 1693, Jamaica, Cal. Col., vol. XIV, No. 392. At a price of about £24 per slave, with £300,000 they would have been able to buy 12,500 slaves. More probable is the sum of £30,000. 52. See Brübach, “Seefahrt und Handel,” 33. 53. See Clarence J. Munford, The Black Ordeal of Slavery and Slave Trading in the French West Indies, 1625–1715 (Lewiston, N.Y., 1991), vol. 2, p. 393. 54. See Hermann Kellenbenz, “Die Brandenburger auf St. Thomas,” Jahrbuch für Geschichte von Staat, Wirtschaft und Gesellschaft Lateinamerikas 2 (1965): 196–217, esp. 204. In 1693, for example, 300 slaves were brought from St. Thomas to Léogâne. A year later a French merchant shipped 400 slaves from St. Thomas to Saint-Domingue. See Munford, Black Ordeal, vol. 2, pp. 393–96. 55. See Jean Baptiste Labat, Pater Labats Sklavenbericht, 1690–1705 (Stuttgart, 1984), 181; Donnan, Documents, vol. 1, 96–99. 56. Cited in Munford, Black Ordeal, vol. 2, p. 394. 57. Ibid. 58. See Archibald Hutcheson to William Blathwayt, Antigua, April 3, 1691, Cal. Col., vol. XIII, No. 1302. 59. See GStA, I.HA, Rep. 65, No. 24, Blatt 14, and No. 40c, Blatt 6ff. In the papers of St. Thomas published by Kellenbenz we can find only cotton as cargo for Curaçao. Probably the Brandenburgers bought slaves at Curaçao when no ships came from Africa. See Kellenbenz, “Brandenburger auf St. Thomas,” 204–10. 60. See Brübach, “Seefahrt und Handel,” 33. 61. See GStA, I.HA, Rep. 65, No. 24, Blatt 25. 62. See Stuhr, Geschichte der See- und Kolonialmacht, 110. 63. Ibid., 107. 64. See Schück, Brandenburg Preußens Kolonialpolitik, vol. 2, no. 165. The king noted regarding an offer of merchants of Rotterdam to send ships to the gulf of Guinea: “Die Pesse will ich unterschreiben wen die Com. will 150 Mohren mitbringen, die 10 à 12 Jahre alt.” See also ibid., vol. 2, no. 186, Ramler’s Bericht über Anschaffung von 150 Mohren. 65. See Brübach, “Seefahrt und Handel,” 42. 66. See George L. Beer, The Old Colonial System, 1660–1754 (Gloucester, 1958), vol. 1, pp. 43–44. See also “Losses Reported by the Royal African Company,” in Donnan, Documents, no. 161. 67. See Munford, Black Ordeal, vol. 1, pp. 160–65. During the war of 1672–78 only eleven ships crossed the Atlantic on a slave journey. In 1679 only eight of twenty-two company ships traded in slaves. See also James Pritchard, David Eltis, and David Richardson,

The Slave Trade of Northern Germany

“The Significance of the French Slave Trade to the Evolution of the French Atlantic World before 1716,” this volume (chap. 7). 68. See Fernand Braudel, Sozialgeschichte des 15.–18. Jahrhunderts (Munich, 1990), vol. 3, p. 288. For the Dutch position in the slave trade, see Jelmer Vos, David Eltis, and David Richardson, “The Dutch in the Atlantic World: New Perspectives from the Slave Trade with Particular Reference to the African Origins of the Traffic,” this volume (chap. 8); and António de Almeida Mendes, “The Foundations of the System: A Reassessment of the Slave Trade to the Spanish Americas in the Sixteenth and Seventeenth Centuries, this volume (chap. 2). 69. See GStA, I.HA, Rep. 65, No. 40c, Blatt 15. 70. See figure 9.1. 71. See Heinrich Sieveking, “Die Glückstädter Guineafahrt im 17. Jahrhundert: Ein Stück deutscher Kolonialgeschichte,” Vierteljahrschrift für Sozial- und Wirtschaftsgeschichte 30 (1937): 19–71. The data set includes only two voyages by the Hanse towns in the eighteenth century: one in 1768, the other in 1798 (TSTD2, voyageids 79,001 and 97,001). 72. See TSTD2, voyageids 97,002 and 96,051. 73. David Eltis, Economic Growth and the Ending of the Transatlantic Slave Trade (New York, 1987), 87, 200. 74. See Percy Ernst Schramm, Deutschland und Übersee: Der deutsche Handel mit den anderen Kontinenten, insbesondere Afrika, von Karl V. bis zu Bismarck; ein Beitrag zur Geschichte der Rivalität im Wirtschaftsleben (Braunschweig, 1950).

271

This page intentionally left blank

Part III Some Wider Consequences and Implications of the New Data

This page intentionally left blank

Chapter 10 The Slave Trade,

Colonial Markets, and Slave Families in Rio de Janeiro, Brazil, ca. 1790–ca. 1830 Manolo Florentino

Four out of every ten African slaves who disembarked in the Americas were imported by Brazil, yet the history of the African slave trade in Brazil is perhaps the least known among all the large-scale trade routes of the modern era. For this reason, the production of The Transatlantic Slave Trade: An Expanded and Online Database (henceforth TSTD2) is auspicious, since it seeks to provide both to specialists and to the interested public research instruments that can illuminate the history of the African diaspora in the Americas. It is especially gratifying that much of the new material embodied in TSTD2 relates to the supply of slaves to Brazil. This chapter uses this database and other sources relating to the slave trade in Rio de Janeiro to redefine some classic positions in Brazilian historiography regarding the traffic in Africans, the colonial market, and slave-based family practices. According to some authors, the African slave trade was essentially a metropolitan endeavor, with the replenishment of captives taking place outside the circle constructed by colonial social relations. This was the case not only because the slaves themselves came from outside 275

276

Manolo Florentino

colonial society, but also because the mercantile agents who made the commerce in Africans economically viable were supposedly representatives of metropolitan capital.1 This theoretical approach is, in truth, the consequence of explanatory models which have assumed that colonial markets were completely atrophied and thus have situated planters at the top of the New World social hierarchies. Even when the authors who have promulgated these theories admit the possibility of small levels of endogenous accumulation, they still tend to exclude the mercantile sector. In the case of Brazil, at least, it is now possible to contest and even reverse the belief that the external reproduction of slavery served as a potent mechanism for migration of capital from the colony to the metropolis.2 We know, for example, that for almost a half century—from 1810 to 1850—the Brazilian elites were able to resist British pressures for an end to the slave trade. This fact can only be understood if we comprehend the commerce in Africans as an important internal circuit for capital accumulation, with a great degree of autonomy from international mercantile capital. The profits generated by this trade made slave merchants the most important faction of the Brazilian elite—a position that aided their attempts to influence state policies. It is not coincidental that a European passing through Brazil during the nineteenth century observed that free men there cultivated long fingernails to demonstrate their distance from the world of manual labor.3 The practice of buying and selling people contributed to the commonly held view in Brazilian society that slaves were nothing more than objects to be used at their owners’ convenience. It is impossible, however, to reduce any enslaved man to the condition of a mere instrumentum vocale. The captives always sought to (re)create means of cultural affirmation. To the two routes to freedom most commonly cited (self-purchase or escape), we must also add the family. It is our hypothesis that families played a key role in slaves’ struggle to be treated as, and to live like, human beings. This possibility justifies the hitherto overlooked study of the relationships between the slave trade and certain cultural practices that encouraged the familial socialization of the captives and affirmed their social identity.

THE SLAVE TRADE AND SLAVERY IN RIO DE JANEIRO, 1790–1830

We take Rio de Janeiro as the focus for our investigation since that region provides a classic example of colonial slavery. Half of the 120,000 people who

Colonial Markets and Slave Families in Rio de Janeiro

populated the captaincy in 1789 were slaves. In 1824 the captive population included some 150,000 souls—more than twice the slave population of 1789.4 The continual shipments of Africans sustained slave population growth during this period, with annual imports growing from around 10,000 in the late eighteenth century to some 40,000 during the 1820s.5 Moses Finley has made the intriguing observation that any colonial society whose elite reproduced itself through slavery must be classified as slavery-based and not merely slaveholding.6 Far from being tautological, this statement recognizes the fact that the ultimate end of slavery is to produce differences among free men. If slavery sustained the colonial elite, and this elite had to replenish their slave stock through the transatlantic slave trade, then we must conclude that the ultimate objective of forced migration of Africans was the reproduction of the social topos of Portuguese America. This is precisely what we find when we investigate the historical record in Rio de Janeiro: increases in the shipments of slaves corresponded with the creation of increasing socioeconomic differences among free men. In Rio de Janeiro at least 84 percent of those individuals who left probate inventories of property between 1789 and 1832 owned at least one slave, indicating that slave ownership was widespread in the province. This arguably undermines the validity of Finley’s maxim: if so many owned slaves, slaveholding was seemingly not, in and of itself, an element capable of creating distinctions between free men. Widespread ownership of slaves, however, was consistent with considerable inequality in property rights in slaves. Between 1789 and 1808, some 60 percent of Rio’s slave owners had fewer than ten captives. These were generally employed in growing foodstuffs for the plantations and the urban nuclei, but they also had a significant presence in craft manufacturing, in street sales, and in domestic service. By comparison, rural and urban establishments using the labor of twenty or more slaves were in the hands of just 15 percent of the slave owners, who between them owned 45 percent of all slaves in the province. Small-scale production units employing fewer than ten slaves generally accounted for less than one in four slaves, as shown in table 10.1. The growth of the Atlantic trade following the opening in 1808 of colonial ports to international commerce reinforced the tendency toward concentration of ownership, and the increase in shipments of Africans occurred simultaneously with this process. The colonial elite came, as table 10.1 illustrates, to own over 60 percent of the captives in 1810–32. At the same time, the percentage of slaves controlled by those proprietors with fewer than twenty captives decreased (dropping from slightly more than half to around 40 percent

277

278

Manolo Florentino

of the total). Likewise, the number of property owners without slaves increased from 11.2 percent in the wills registered between 1789 and 1808 to 16.4 percent in those registered between 1810 and 1830. Increased access to imported slaves seems to have strengthened the capacity of the richest free men to generate profits through the appropriation of slave labor. The slave trade thus emerges as a central mechanism in the reproduction of the social distance separating the slaveholding elite from the other free sectors of the socioeconomic hierarchy. Between 1790 and 1830, the thousands of Africans disembarked every year in the port of Rio de Janeiro supplied the captaincy of Minas Gerais (and its economy, directed toward the internal market) via overland routes. This internal commerce absorbed up to 40 percent of the slaves sold and resold in Rio. Via oceanic routes, Rio also supplied a significant portion of the demand for slaves in Rio Grande do Sul, Santa Catarina, São Paulo, and, on a somewhat smaller scale, Espírito Santo. Slaves in these regions were employed in producing agricultural goods for both the internal and external markets. They were also used in herding, craftwork, and domestic labor. Rio de Janeiro itself encompassed other economic sectors demanding African labor: these included sugarcane plantations (notably those in the region of Campos dos Goitacás); farms producing foodstuffs; the mercantile center of the capital and its surrounding periphery; and the coffee plantations of the Paraíba Valley. Exports from the latter rose from 160 tons in 1792 to 539,000 tons in 1820 and then peaked at 3,237,190 tons in 1835.7 Table 10.2 reveals the fluctuations in slave disembarkations. The numbers expressed here are taken from TSTD2 and relate to those ships entering Rio de Janeiro between July 5, 1795, and December 31, 1830. For an explanation of the procedures used to produce these data, see this volume’s chapter 1. There is a gap in data for the period before mid-1795, with no records of disembarkations registered for January 1, 1790, to July 5, 1795. To estimate total imports for 1795, we doubled the number of slaves recorded as embarking for or disembarking at Rio de Janeiro during the second half of the year. For the period before 1795, TSTD2 provides data for 1787–94 for some voyages departing from Luanda, Benguela, and “Angola” with the declared destination of Rio de Janeiro, but there are no records of them arriving in Rio. In the absence of registers of imports, we used totals of slave disembarkations provided by Rudolph W. Bauss, based upon the customs taxes paid on the slaves who entered the port of Rio between January 1790 and December 1794. From Bauss’s numbers for slave imports, we calculated estimates of total slaves

Table 10.1

Concentration of Slave Ownership across Rural and Urban Proprietors in Rio de Janeiro, 1789–1832 1789–1808

Number of Slaves per Proprietor 1–9 10–19 20+ Total

1810–1832

Percentage of Proprietors

Percentage of Slaves

Number of Proprietors

Number of Slaves

Percentage of Proprietors

Percentage of Slaves

Number of Proprietors

Number of Slaves

59.7 25.9 14.4 100

24.1 30.5 45.4 100

161 70 39 270

743 938 1,399 3,080

62.4 19.3 18.3 100

20.4 18.4 61.2 100

402 124 118 644

1,823 1,648 5,468 8,939

Sources: Inventários post-mortem (1789–1832), Arquivo Nacional (Rio de Janeiro); Inventários post-mortem (1825–32), Primeiro Ofício de Notas de Paraíba do Sul (data collected by João Fragoso); Inventários post-mortem (1820–32), Arquivo Público Judiciário de Itaguaí (data collected by Ricardo Muniz de Ruiz). Note: Only 34 wills between 1789 and 1808, and 126 from 1810 to 1832, make no mention of slaves.

280

Manolo Florentino

shipped to Rio de Janeiro from Africa by using slave mortality levels in transit for the period 1796–1800 (8.3 percent). Similar adjustments were made to the numbers generated from TSTD2, which provides the basis for the data presented in table 10.2.8 Table 10.2 identifies three phases in the Rio de Janeiro slave trade in 1790–1830. The first and longest phase, covering 1790–1810, was one of relative stability in slave shipments, with average disembarkations of 12,000 slaves per year. The second phase, from 1811 to 1825, includes the opening of colonial ports to international commerce and saw a rise in imports of 50 percent to just under 18,000 a year. The third and final phase, from 1826 to 1830, saw more than a doubling of annual imports to 40,000 a year, largely associated with British recognition of Brazilian independence after 1822 and the ratification of the treaty of March 13, 1827, which stipulated a legal end to the Brazilian slave trade within three years. Foreseeing the end of the transatlantic trade and demonstrating a great ability to marshal resources, the slaveholding elite of Southeast Brazil began to buy up all the Africans they could, even before formal ratification of the treaty, and continued to do so through 1830.9 Table 10.2 Slaves Exported from Africa and Imported to Rio de Janeiro over Five-Year Periods, 1790–1830 Years 1790 1791–95 1796–1800 1801–5 1806–10 1811–15 1816–20 1821–25 1826–30 Total

Number of Voyages (registered entries)

Number of Slaves Exported

Number of Slaves Imported

Index (importation)

0 0 138 142 182 264 299 327 504 1,856

13,029 69,672 60,990 70,993 86,323 116,871 129,426 152,872 220,010 920,186

11,999 64,050 55,915 65,540 75,319 107,389 115,871 135,126 201,171 832,380

— 100 87.2 102.3 117.6 167.7 180.9 210.9 314.1

Sources: Estimates page of http://www.slavevoyages.org and Rudolph W. Bauss, Rio de Janeiro: The Rise of Late Colonial Brazil’s Dominant Emporium, 1777–1808 (PhD diss.,Tulane University, 1977). Note: Rudolph Bauss worked with the “Resumo do Rendimento dos Direitos dos Escravos que Entrarão Neste Porto desde o Primeiro de Janeiro de 1790 até o fim de dezembro de 1794” (Correspondência do Vice-Rei para a Corte, códice 68, vol. 14, fl. 91, Arquivo Nacional). There he found annual taxes of 5:740$000 réis (1790), 7:478$000 réis (1791), 8:456$000 réis (1792), 11:096$000 réis (1793), and 10:225$000 réis (1794). Bauss assumed a tax of 1$000 réis per slave, however, which implies that he supposed that all the slaves disembarked in Rio de Janeiro during those years were adult Africans.

Colonial Markets and Slave Families in Rio de Janeiro

Whereas during the second part of the eighteenth century, half of the African slaves brought into Brazil entered through Rio de Janeiro, during the two decades after 1809, some 70–90 percent of them entered through this port.10 Taking a continental view, the number of African slaves brought into Rio between 1790 and 1810 was superior to the number of slaves arriving in the United States, Spanish America, and the Danish and Dutch colonies of the Caribbean combined. This volume was equal to about 70 percent of the total of all the Africans transported in British hulls and 85 percent of the size of the entire French slave trade in the same period. After 1810, when only Cuba could seriously compete with the Carioca-controlled trade in Africans, the number of slaves disembarked in the port of Rio was two to four times greater than the number reaching Cuba.11 When we ponder the fact that this trade was also a business, we begin to perceive that the volume of the Carioca importations must indicate the existence of a formidable endogenous circuit of wealth, the total accumulation of which set it apart, even when compared to the activities of agricultural production and exportation. For this reason, the question of whether this wealth accumulated in European circles or remained largely in Brazil assumes an immediate and obvious importance. The eighteenth century saw the Congo-Angolan region climb to dominance as the principal source of slaves for the Carioca market. Of the slaves reaching Rio in the eighteenth century, 88 percent boarded ship at Luanda and Benguela, compared to only 11 percent who boarded ship at the Central Atlantic and West African coasts and 1 percent at East Africa.12 The profile of African slave imports shown in table 10.3 reveals that in the period 1795–1830 trade varied from the earlier pattern. In this period, the opening of the colonial ports to international commerce was associated with the incorporation of East Africa into the trade and the diversification of supply sources within already established areas. The prohibition of the slave trade north of the equator by the Congress of Vienna caused a diminution in Western Africa’s participation in the Riobound trade during this period. However, this region never entirely ceased to supply Rio, even though the majority of the West Africans passing through the port after 1815 most probably were reimported internally from other areas in Brazil such as Bahia. Part of the remaining West African traffic came from the north sections of the coast, while the rest was from the Cameroon River region. An absolute and relative increase in exportations also occurred, starting around the turn of the century, from East Africa, which became a major supplier of slaves to Rio after 1811. This development was linked to a diversification of

281

282

Manolo Florentino

Table 10.3 African Embarkation Points of Slaves Disembarking in the Port of Rio de Janeiro, 1795–1830 1795–1810

Region

Port/Region of Embarkation

Western Africa

100 Sierra Leone Ajuda Lagos “Gold Coast” “Mina Coast” Bonny Old Calabar Cameroon River Gabon São Tomé Island Príncipe Island

Central Atlantic Africa

Eastern Africa “Mozambique” Quelimane Inhambane Lourenço Marques Madagascar Source: TSTD2.

Percentage of Region in Total Slave Exports 1.3

18.2 11.9 32.9 10.9 26.1

100 Loango Molembo Cabinda “North of the Congo” Congo Ambriz Luanda Benguela “Angola”

Total

Percentage of Regional Exports

96.6

Percentage of Regional Exports 100 9.9 6.6 5.0 10.1 4.5 4.1 11.1 31.6 1.4 13.6 2.1

100

Percentage of Region in Total Slave Exports 1.5

76.7

0.2 1.4 34.2 0.1 3.3 8.3 8.8 18.1 25.6

0.3 3.2

26.1 48.1 22.3

100 100

1811–1830

2.1

100

100 51.6 35.9 3.4 3.4 5.7

21.8

100

Colonial Markets and Slave Families in Rio de Janeiro

exports from East African sources. Originally derived largely from the island of Mozambique, East African slaves entering Rio increasingly came from the southern ports of Quelimane, Inhambane, and Lourenço Marques (later Maputo), with some even coming from the island of Madagascar. It is possible that the increase in East African exports reflected an inability of suppliers in the Congo-Angolan region to meet Southeast Brazil’s increased demand for slaves after 1808. Nevertheless, Central Atlantic Africa continued to dominate slave exports to Rio de Janeiro before 1830. Further confirmation can be found in the probate inventories of slaves, which indicate that, during the first three decades of the nineteenth century, eight out of ten Africans living in the captaincy of Rio de Janeiro were Congo-Angolans. From 1810 onward, however, important changes took place in this scenario—changes that involved the more intense incorporation of the northern region of Central Atlantic Africa. The hegemony of Luso-Brazilian traffickers intensified in the region north of Luanda after the withdrawal of the Dutch, English, and French (initially due to the Napoleonic Wars and later owing to the metropolitan restrictions on the slave trade). Luso-Brazilian interests thus began to gain ground in the states of Angoi, Congo, Kakongo, and Loango, and the port of Cabinda became involved in the large-scale exportation of Africans.13 Because Africans controlled the supply of slaves, I shall begin this analysis of the slave import trade into Rio de Janeiro by examining its origins in the Congo-Angola region. The focal points of this commerce were the Atlantic ports, which were centers of two flows: the export of slaves (generally captured in the interior and brought to the coast for sale) and the import of European and American goods (shipped to the interior savannas and forests). From an analytical viewpoint, these flows were complementary and inseparable. Their smooth functioning depended upon the existence in Africa of three different types of socioeconomic agents: the native merchants located in the ports with access to trade goods received from the captains of the slave ships; the backlands authorities who supplied these men with captives; and the small-scale native merchants who served as intermediaries between the first two groups.14 After weeks or months of traveling through savannas and forests, the hundreds of small-scale native slave dealers—pumbeiros, aviados, funantes, or sertanejos—and their employees contacted the authorities in the backlands who controlled the capture of slaves. The expense of keeping slaves alive prevented these authorities from holding stocks of them. Instead, they awaited the arrival of the traders from the coast and only then, when terms of trade were agreed upon and trade goods exchanged, would they produce prisoners

283

284

Manolo Florentino

from their border regions. Transferred to the traders, these groups of captives (libambos or quibucas) were then gathered until their number satisfied the requirements of the merchants of the Angolan coast.15 Three out of every four slaves brought out of the backlands were prisoners of war.16 Wars provided a massive and continual production of captives in those areas affected by the traffic, which created or accentuated already existing power imbalances among Africans. This, in turn, led to the reinforcement of the state as the only institution capable of producing wholesale lots of captives. It is unsurprising, therefore, to find that, during the highpoint of the slave trade in the eighteenth century, the majority of societies without formal state structures largely lay outside the principal routes of the slave trade. This was the first dimension of a deep sociological movement: the acquisition of goods on the coast and the exchange of these for prisoners in the backlands resulted in a corresponding political and economic reinforcement of dominant native groups. Warfare—especially state-directed warfare—thus accentuated social distinctions among classes and class factions, among ethnic groups, and—in the backlands—between 0lder and younger generations. As the foundation of slavery, wars intended to meet American demands for labor gave rise in Africa to greater social differentiation and to state expansion. The slave trade itself represented a dimension of this second process, linked with the first. While most captives were traded for European or American goods, a growing number were retained and employed in Africa itself. Scholars have discovered evidence of large percentages of slaves among the populations of the largest and most powerful African states.17 Violence also permitted the means by which the elite of the capturing societies accumulated ever greater wealth in the persons of enslaved men, women, and children.18 The violence that underpinned slavery disassociated commercial value from the social costs involved in producing a captive.19 In this sense, the initial exchange of people for goods implied the appropriation by merchants and the native elite of a large portion of the work produced by local communities. All the steps that followed, involving the circulation of the captives through the Atlantic trading circuits, revolved around the repartition of the slaves’ products. As in a transmission belt, the disassociation of production costs from commercial value continued through to Portuguese America, where the prices for captives did not reflect in the slightest their social value. This fact permitted even relatively poor free men to purchase at least one slave. In eighteenth-century Luanda, a healthy adult man in good physical condition would correspond in value to fifteen felt hats or fourteen pairs of silk

Colonial Markets and Slave Families in Rio de Janeiro

stockings. A slave might also be bought for three barrels of gunpowder or twenty-three blank books; a horse or two barrels of cachaça (sugarcane brandy) would pay for two more. For every $1,000 réis of imports into Angola, $600 comprised cloth, clothing, and objects of personal decoration; $200 comprised alcoholic beverages; and $110 comprised arms and other means of acquiring slaves. The remainder comprised construction materials, spices, cowrie shells, and tobacco.20 According to an anonymous author who wrote about the traffic in the middle of the eighteenth century, this had been the pattern since the installation of the traffic in Angola, given that “since the beginning” the trade for slaves had as its base “living supplies, liquor and hardware.”21 India was the source of about one-third of everything Angola imported. The rest came in more or less equal parts from Brazil, Portugal, and northern Europe, as shown in figure 10.1. Portugal supplied Angola cloth, alcohol, and general supplies, while textiles represented some 75 percent of the goods from other European sources. Sugarcane brandy constituted almost 60 percent of the goods originating in Brazil, with general foodstuffs amounting to a further 15 percent. The preferred “coin” of the slave trade, however, was Indian cloth, which made up 99 percent of all Asian sales to Angola, as shown in figure 10.2.

Figure 10.1 Origins of Merchandise Imported into Luanda, 1785–94 Source: “Balanço da importação e exportação deste Reino de Angola desde o anno e 1785 em que teve princípio o estabelecimento da alfândega até o anno de 1794 inclusivo,” Arquivo do Instituto Histórico e Geográfico Brasileiro (RJ), Seção dos Manuscritos, lata 77, documento 1.

285

286

Manolo Florentino

Other

25%

11%

64%

Ivory

100%

Spices

100%

Utensils

43%

48%

Construction Materials and Derivatives

89%

Cowries

11%

94%

Tobacco

6%

100%

Weapons

26%

74%

Food Supplies

49%

Alcohol Cloth and Derivatives

9%

42%

36%

64%

14%

28%

58% Portugal

9%

Brazil

Asia

Northern Europe

Figure 10.2 Origins of Merchandise Types Imported into Luanda, 1785–94 Source: Same as figure 10.1.

By the end of the eighteenth century, one-third of the cotton cloth sent to Brazil from the metropolis was white Asian cloth dyed and printed in Portugal.22 Assuming that the same percentage of Asian textiles was exported from Lisbon to Africa, it appears that some 60 percent of the cloth traded for captives had its origin in Asia. This means that goods made in India accounted directly and indirectly for some 40 percent of the slave purchases made in Luanda. Together, the products of India and of Portuguese manufactories in the Americas amounted to over 60 percent of the value of the materials traded for slaves in Angola during the second half of the eighteenth century. With the dawn of the nineteenth century, Angola’s sources of imports changed. During the first two years of the House of Bragança’s sojourn in Brazil, Portuguese America was the source of one-third of the goods sent to Africa, the same as the shares from Asia or Europe. The enlargement of Rio de Janeiro’s involvement in the slave trade in 1810–12, which saw the port’s share of the trade rise to between two-thirds and three-quarters of all slave purchases, brought about little, if any, increase in the share of Brazilian goods in purchases at Luanda.23 Admittedly, sugarcane brandy amounted to onequarter of all Angolan purchases in 1808–9, and in the same period supplies and hardware amounted to two-thirds of the goods involved in the trade. During the following three years, however, when Angola acquired most of its imports from Brazil, cloth represented some 70 percent of all African purchases

Colonial Markets and Slave Families in Rio de Janeiro

in exchange for slaves.24 It thus appears that the opening of the colonial ports and the centralization of the purchase of African slaves in Rio de Janeiro strengthened the position of cloth among goods tradable for slaves in Angola, while weakening somewhat the position of alcohol, food supplies, and weapons. Much of this cloth continued to come from India directly, but some Indian textiles arrived in Luanda via Rio de Janeiro and yet more were transshipped to Africa after being printed in Portugal. Indian goods therefore continued to dominate slave purchases at Angola during the 1810s, controlling roughly the same market share as they had in the previous century. A new factor in the nineteenth-century trade was nevertheless discernible by this time. This was an increasing share of the trade of European manufactured goods, particularly items from northern Europe. It is not difficult to trace the causes of this change, since European manufacturers could more easily satisfy African demands for goods that arose in response to the growth in slave exports. Overall, we can postulate that during the first decades of the nineteenth century, the southern sphere of the Portuguese Empire (Brazil and India) contributed half of the value of the goods used in the Congo-Angolan trade, with the metropolis and—above all—the rest of Europe providing the other half. From the eighteenth century onward, Rio’s own slave traffickers financed the port’s slave trade.25 The probate inventories of these merchants do not show any outstanding debts to India or to European mercantile houses. This suggests that these men were not simply the local representatives of foreign commercial concerns.26 In truth, as the owners of three-quarters of the slave ships docking in Rio and of the goods used in the African trade, these men relied on three mechanisms to acquire the European and Asian goods that, together with the colonial Brazilian products, made possible the purchase of African slaves. Many of these merchants bought goods in Africa itself, together with the European and Luso-African traders. Remittances from Brazil to Angola via letters of exchange or in specie covered the cost of such purchases. Others acquired goods in Europe or Asia, exporting these directly to Africa, which, in turn, required the forwarding of letters of credit and cash to enable purchases on the European and Indian markets. Though these mechanisms continued after 1808, it is possible that they— particularly the first mechanism—dominated the Angolan trade before the opening of the Brazilian ports. The third system used by Rio-based merchants, emerging as the dominant system after 1808, was to import Indian and European products into Brazil and subsequently to reexport them to Africa.

287

288

Manolo Florentino

For example, the company founded by Captain Manoel Gonçalves Moledo (of Rio de Janeiro) and Ignácio Correia Picanço (of Benguela) arranged for Portuguese goods to be sent directly to Africa, topping these off with purchases of foreign goods in Angola. Ninety percent of Gonçalves Moledo’s contributions to the partnership in 1804 amounted to money (plus 10 percent sugarcane brandy). In the following year, goods apparently sent from Portugal to Benguela represented one-third of the Brazilian partner’s expenditures, with letters of credit and cash accounting for another half. Only in 1808 did the value of textiles bought in Lisbon and sent directly to Benguela exceed the remittances of money.27 The arrangements of another Rio trader, Manoel Gonçalves de Carvalho, and his representative in Angola, Antônio Alves da Silva, illustrate the prevalent form of the trade after the arrival of the Braganças in Brazil. In October 1819 the Carioca merchant sent thirty bundles of cloth and twenty barrels of alcohol to Africa—“excellent goods for that country . . . where quality items are hard to come by,” according to Carvalho. In exchange, he asked for “young, good slaves even if these cost up to half again the normal price.” The textiles involved in this exchange had come from Malabar and reached Angola via Rio de Janeiro.28 In the case of another large slave trader, Elias Antônio Lopes, we know that in 1812 and 1814, two of his ships docked in Cabinda carrying imported merchandise. On his death in 1814, the customs warehouses of Rio were full of textiles belonging to Lopes. Those imported from Goa were valued at 46:653$806 réis—or at least one-fifth of his enormous fortune. Lopes had controlled an enormous mercantile network through which he acquired goods for the African trade (among other things) in transcontinental commercial centers such as Lisbon, Porto, London, Hamburg, Amsterdam, and Goa. He traded for slaves in Luanda, Benguela, and Mozambique. Lopes demonstrated that even before the Lisbon court migrated to Brazil, Carioca merchants already used Rio as a reexportation port for goods for the African trade. In 1803 the galley Resolução arrived in Rio outbound from Porto, carrying freight bought by Lopes and his partners. Upon arrival, it set off immediately on a slaving voyage to Angola.29 There are even registers attesting that, in the mideighteenth century, Rio already supplied Angola and Benguela with “goods from Europe, Asia, money in half dobles, liquor and other goods.”30 It is likely that the greater part of the non-Brazilian goods bought before 1808 by Rio transatlantic merchants went directly to Angola from Lisbon or Goa without stopovers in Brazil.31 A legal action in the Junta do Comércio by the Rio slave trader Bernardo Lourenço Vianna against a debtor, Antônio

Colonial Markets and Slave Families in Rio de Janeiro

Rodrigues de Moura, a merchant established in Luanda, offers some support for this. The relationship between the two litigants stretched back to 1800, when Antônio de Souza Portella, a Lisbon merchant operating in Vianna’s name, loaned large sums of money to Moura. During the first four years of the relationship, the Angolan managed to pay off some of his debt, but after 1804 all payment simply ceased. Portella then died, his commercial house went bankrupt, and the debts owed to the Carioca trader by the Angolan merchant rose to more than 20,000 cruzados (8:000$000 réis). Vianna’s Carioca backer was, in effect, purchasing goods in Asia, with the Lisbon-based Portella simply serving as a convenient intermediary.32 The entire mercantile process revolved around credit in the form of advancing merchandise: Angolan merchants advanced goods, liquor, tobacco, weapons, and gunpowder to the traders who journeyed into the interior in search of slaves. Before this could take place, however, they needed to receive merchandise from the captains of the slave ships and this, in turn, put them in debt to the slaving capitalists of Rio de Janeiro. The debts of the Angolans with the Crown, with the slave drivers, and with suppliers of goods were paid in letters of credit issued by Rio merchants. For these reasons, during the second half of the eighteenth century Carioca letters of credit circulated as means of exchange in Benguela, according to an anonymous author.33 Even before ships left Africa with slaves, many slaves died. Memória a respeito dos escravos e tráfico da escravatura entre a costa d’Africa e o Brasil, a paper presented by Luís António de Oliveira Mendes to the Real Academia de Ciências de Lisboa in 1793, estimated that between 30 percent and 50 percent of captives were lost between capture in the interior of Angola and their arrival in the port of Luanda.34 On the other side of the continent, according to Frei Bartolomeu dos Mártires, writing in 1819, almost 20 percent of the slaves who were acquired in Mozambique for sale in Rio de Janeiro, Bahia, and Pernambuco died between the time of purchase and their final embarkation—a statistic that basically referred to the time the slaves spent waiting while the purchase for each ship was completed.35 In Angola, one or two out of every ten slaves who arrived at the coast died in filthy port warehouses while awaiting transshipment overseas; one in every twenty survivors escaped from the warehouses or was abandoned at the embarkation port as too sick to continue the journey.36 Nothing, however, matched the Atlantic journey in terms of risk and damages to the traffickers’ human merchandise. Slaves were occasionally lost at sea because of the activities of corsairs, but the most serious sources of onboard

289

290

Manolo Florentino

slave deaths stemmed from inadequate food and water supplies, poor treatment, overloading, lack of sanitation, and the debilitating effects of sheer, sustained terror. Constant fear wore down and weakened many slaves and contributed significantly to the death tolls during the Middle Passage. These, in turn, constituted the traders’ primary losses. A few ships lost hardly any slaves, but others saw half their captives die during the transatlantic voyage. There were even cases in which slave ships docked in Brazil without a single living slave onboard. Fundamentally, the proportion of deaths was related to the duration of the voyage, which was largely determined by the distance between the African regions of embarkation and the Brazil ports of arrival. Table 10.4 compares death rates in ships arriving in Rio de Janeiro from several different regions in Africa from 1791 to 1830. It shows that, on average, losses on ships taking slaves in the Indian Ocean were two to three times those on ships embarking slaves at the Congo-Angolan region. This difference was obviously owing to the variations in durations of voyages: ships leaving the Central Atlantic coast of Africa could anticipate arriving in Rio in thirty-three to forty-three days, while those leaving Mozambique could take more than seventy-six days to reach their destination. As time passed, onboard mortality rates were reduced on both routes, a fact linked in part to declining crossing times. From the end of the eighteenth century to 1820, the average voyage length decreased by 15 percent for ships outbound from Congo-Angola and 20 percent for those sailing from Mozambique. It is possible that this decrease in journey length—and the consequent drop in onboard mortality among slaves—was related to the growing use of smaller, faster ships by nonspecialist traffickers. Whatever the cause, it is almost certain that losses of slaves in the crossing were lower after 1800 than during the sixteenth and seventeenth centuries.37 The risks to slaves did not cease with the arrival of ships in Brazil. Many arrived sick and died before they could be sold. The shock of a new environment full of unknown diseases, delays in selling slaves, and the subsequent journeys into the Brazilian interior were other experiences that routinely proved fatal to slaves. It is not surprising to learn that some 25 percent of the slaves who arrived in Brazil died before reaching their final destination.38 Of the 279 companies that engaged in the trade between Africa and Rio de Janeiro between 1811 and 1830, only one in twenty were active throughout the entire period, financing one or more trips a year. This small group of companies was responsible for 43 percent of all expeditions in these years and constituted the core of the specialized traffickers. Another group of companies— making up 9 percent of the total—equipped slaving expeditions every two or

Colonial Markets and Slave Families in Rio de Janeiro

Table 10.4 Voyage Mortality as Percentage of Slaves Embarked for Rio de Janeiro Years 1791–95 1796–1800 1801–5 1806–10 1811–15 1816–20 1821–25 1826–30 All years

West Africa

West Central Africa

Southeast Africa

— — — 8.9 (7) 10.6 (14) — 9.6 (2) 4.5 (3) 9.4 (26)

5.5 (13) 8.0 (103) 6.2 (103) 13.0 (111) 6.2 (146) 6.8 (126) 8.4 (149) 4.8 (365) 7.0 (1,116)

— 18.7 (2) 36.1 (2) 22.2 (9) 15.8 (15) 22.5 (23) 18.1 (81) 11.6 (99) 15.9 (231)

Source: TSTD2. Numbers in parentheses indicate numbers of voyages.

three years and was responsible for around 20 percent of the expeditions. The third and final group constituted the large majority (86 percent) of companies and entrepreneurs who dabbled in the Atlantic slave trade. These organized one expedition every four or more years and were responsible for about onethird of the voyages to Africa. They were adventurers who did not specialize in the traffic and entered into it as a strictly speculative endeavor. These numbers demonstrate how ownership in the business of selling slaves was highly concentrated. The preparation for a voyage included the purchase or rental of ships, their fitting out and staffing—including the employment of masters, surgeons, chaplains, and seamen (the last almost always slaves)—and their supplying with specialized equipment (such as chains), food, and, most important, trade goods. This was an expensive undertaking. When we take these costs into consideration and remember the risks involved in the commerce in people, it is easy to understand why so few companies had enough capital to engage in the traffic on a regular basis. The fact that the core group was able to operate on a continual basis, organizing one or more expeditions each year, reveals a great deal about their capacity for professional specialization. The slave trade was, nevertheless, a business in which speculators had a central role. Almost 90 percent of the Carioca traders were infrequent investors in the trade, accounting for a third of the trips to Africa from Rio in 1790–1830. Such speculators or occasional traffickers were urban merchants who sought to increase their already considerable wealth through slaving. They were also often former slave ship captains who, after several trips to

291

292

Manolo Florentino

Africa serving as the principal intermediaries with native slavers, decided to fit out ventures on their own account, using the experience and wealth gained in their earlier voyages. As was the case with piracy, the activities of these adventurers increased during moments of high demand—and consequently higher prices—for slaves in the colonial market. What profits could reasonably be expected from the commerce in people? The profitability of slaving ventures was tied to the difference between the purchase and sale price of the merchandise involved in the trade. In this activity, the Brazil-based trader had the advantage that the prices at which he could acquire slaves were often low, and typically only a fraction of the prices he received for slaves in Brazil. Moreover, even when slaves were sold in Brazil at prices that seemed low (and failed to reflect the social costs of their procurement), profits could still be achieved by merchants. Trafficking almost always resulted in significant gains. On a more general plane, however, the slave trader was a merchant just like any other. He sought to achieve not just “reasonable” profits, but the highest profits possible. Some estimates can be calculated of the gross profits made on some fourteen slaving voyages dispatched from Rio de Janeiro in 1812. These are summarized in table 10.5. We know that in this year the average price per slave on the Carioca market was 104$000 réis and that slaves cost 69$000 réis per head in Luanda, 50$000 réis in Benguela, and 40$000 réis on the island of Mozambique. Applying these prices to the data on slaves bought and sold per voyage in columns 3 and 6 and then deducting the estimates of investment in column 4 from the returns from sales in column 7 yields the percentage rate of profit or loss per voyage shown in column 8. Our estimates reveal substantial variations in profit margins per voyage, ranging from gross profits of almost 83 percent on investment outlays in the case of the voyage of the Feliz Dia to losses of almost 31 percent in the case of the Protector. This variability in profits per voyage is consistent with those in other branches of the slave trade. Superficially, what our data also reveal is that profits tended to be higher where the proportions of slave losses in transit were lower, and in voyages to Benguela and Mozambique over voyages to Luanda. These estimates, however, are based only on expenditures for trade goods, and these were not the only items on the cost side of slaving voyages. Food and supplies for the crew and slaves, voyage stores and equipment, salaries of crew, and customs charges, among other things, should also be factored into any calculation of profits. These may have been higher in the East African, if not Benguelan, trade compared with the Luandan trade. Taking the total investments involved

Colonial Markets and Slave Families in Rio de Janeiro

and setting these against the returns from sales of slaves in Rio suggests that average rates of voyage profits in the Luanda-Rio slave trade were 16 percent in 1810, 3 percent in 1812, 6 percent in 1815, 23 percent in 1817, and 44 percent in 1820. The average voyage profit for the whole decade was 18.4 percent on capital investment.39 These calculations are significant. The 18.4 percent return on investment is higher than that of any other slave trade before 1830 for which we have hard evidence. Though the bases for the estimates are different, returns in the Rio trade in the 1810s appear to have been double those in the late eighteenthcentury British and French trades and several times greater than those achieved by the Dutch Middelburg company between the 1730s and 1790. Returns exceeded those enjoyed during the ancien régime, which tended to fluctuate around 5 percent, and to be higher than the profits of the large slave plantations of Rio de Janeiro, Bahia, and parts of the Caribbean, which varied between 5 percent and 10 percent. By these standards, the Carioca slave trade appears to have been a veritable El Dorado for those who could find the means to participate in it. The trade was, however, also characterized by sharp fluctuations in profitability—a fact that may help explain the large but also highly variable presence of speculators. Such violent rhythms indicate a risky, structurally unstable, and atrophied market. These factors in turn encouraged the traditional trader, as a typical merchant of the late eighteenth and early nineteenth centuries, to diversify his investments.

THE SLAVE TRADE AND THE COLONIAL MARKET

From 1796 to 1812, and without taking into account the remittance of precious metals to the metropolis, the Carioca colonial market increasingly tended to lose in its dealings with Lisbon-based commerce.40 Between 1810 and 1812, for example, Rio’s negative trade balance reached 267:040$755 réis.41 Rio merchants also owed their counterparts in Goa, having incurred losses of 1:105:640$220 réis over the same period. It must be remembered that Asia supplied critical trade goods to the South Atlantic circuit and that these were generally acquired by Carioca merchants with hard cash.42 Possibly 10 percent of what Rio traders imported from Asia, as well as 20 percent of what they bought in Portugal and 30 percent of their purchases in the other countries of Europe, were destined to supply the slave trade—all of which contributed to the Carioca deficits. With regard to Africa, Carioca debts held

293

Table 10.5

Estimates of Profits of Selected Slaving Expeditions Originating in Rio de Janeiro in 1812

Ship Mercúrio Feliz Dia São José Americano Carolina Providente Nossa Senhora da Conceição Canoa Mariana Dáfne Fiança Flor do Mar Feliz Indiana Guadalupe Julia Protector

African Port of Purchase

Number of Slaves Bought

Total Investment (in réis)

Percentage of Dead Slaves

Number of Slaves Sold in Rio de Janeiro

Total Revenue from Sales (in réis)

Profits (percent)

Benguela Mozambique Benguela

404 235 547

27:270$000 12:690$000 36:922$000

5.0 5.1 5.7

384 223 516

39:936$000 23:192$000 53:664$000

+46.4 +82.8 +45.3

Benguela Mozambique Luanda

571 247 478

38:542$500 13:338$000 44:525$700

6.1 7.3 7.9

536 229 440

55:744$000 23:816$000 45:760$000

+44.6 +78.6 +2.8

Luanda Luanda Luanda Luanda Luanda Benguela Luanda Luanda

534 572 473 462 462 458 500 397

49:742$100 53:281$800 44:059$950 43:035$300 43:035$300 30:915$000 46:575$000 36:980$550

10.9 11.5 12.5 12.6 12.6 15.9 16.0 38.0

476 506 414 404 404 385 420 246

49:504$000 52:624$000 43:056$000 42:016$000 42:016$000 40:040$000 43:680$000 25:584$000

−0.5 −1.2 −2.3 −2.4 −2.4 +29.5 −6.2 −30.8

Source: Manolo Florentino, Em costas negras: Uma história do tráfico de escravos entre a Africa e o Rio de Janeiro (Séculos XVIII e XIX) (São Paulo, 1997), 171.

Colonial Markets and Slave Families in Rio de Janeiro

steady through the eighteenth century. Figure 10.3 shows, however, that by the early nineteenth century these debts were substantially lower, though they never disappeared completely. This change can be explained by Rio de Janeiro’s transformation into a major reexport center for Portuguese, other European, and Asian goods. The flows of merchandise that made the slave trade possible increasingly became centered on and passed through the mercantile community that funded the trade. Between 1810 and 1812 alone, Rio’s trade imbalance with Africa totaled 459:574$479 réis. This debt continued to be paid in gold coin in quantities that fail to show up in the African customs books, although they appear in the notes of the customs employees.43 If, however, we accept that the Carioca merchants took control of the slave trade, we need to redefine these losses as the diminution of Carioca debts with Africa that occurred after 1800. Yet the Rio-based merchants’ overseas debts likely grew in this period as manufactured goods—almost entirely imported from Asia and Europe—grew in importance in the slave trade. Such debts were partially offset by sales of slaves in

Figure 10.3 Movement of Commerce between Luanda-Benguela and Rio de Janeiro, 1785–1812 Sources: Graph 1 of “Resumo dos mapas de importação e exportação dos estados da India, Africa e Brasil: 1808–1814,” Arquivo Nacional, Junta do Comércio, caixa 448, pacote 1. Note: The movement of commerce between 1785 and 1794 is based on exports from all over Brazil (which in this case basically means Rio de Janeiro). “Profit” here means exports as a percentage of exports and imports combined.

295

296

Manolo Florentino

colonial markets throughout southern and southeastern Brazil. Revenues from these, some of it in specie, enabled the Carioca slavers to pay their chronic debts to Goa, Europe, and Africa with specie.44 Let us take a closer look at this process. The commercial balances and probate inventories from the decade beginning in 1810 provide some indication of the scale of the resources controlled by the networks that sold recently arrived slaves (boçais). Boçais were redistributed via the port of Rio de Janeiro to points throughout the captaincies of Rio de Janeiro, Minas Gerais, Santos, and Rio Grande do Sul. Table 10.6 provides data on the number of slaves sold in the ports of Santos and Rio Grande do Sul, taken from the “Resumo dos mapas de importação e exportação,” which belonged to the Junta do Comércio (Arquivo Nacional). We assume that Rio de Janeiro—the capital and its surrounding regions—absorbed about half the African slaves arriving from Luanda and Benguela. We do not know how many were reexported to Minas Gerais, but we do know that this captaincy, whose demographic estimates for 1819 reveal the largest concentration of slave labor in Brazil, received more than 40 percent of the Africans disembarked at Rio during the 1820s.45 We can safely assume that this level of reexport was probably occurring a decade earlier, which means that Minas would account for the balance of slaves arriving in Rio in the 1810s. Figures from the Resumos suggest that the price of slaves in Africa in 1810–12 was 63$938 réis on average, while prices in Santos and Rio Grande do Sul were 128$000 réis and 133$255 réis, respectively. This represents a median markup on African prices of at least 100 percent at Santos and Rio Grande do Sul. With regard to Minas Gerais, in Memórias do distrito diamantino Joaquim Felício dos Santos affirms that in the late eighteenth century, slaves were purchased in Rio de Janeiro for between 100$000 and 120$000 réis and then resold in Minas for up to 240$000 réis.46 If this picture persisted after 1800, it would again imply markups on slave prices between Africa and Minas Gerais of over 100 percent. Finally, in Rio de Janeiro itself, the average price for slaves between fourteen and forty years of age, calculated according to probate inventory lists, was 91$500 réis in 1810, 135$000 réis in 1811, and 93$400 réis in 1812—an average of 107$239 réis, almost 60 percent higher than in Angola. Clearly, reexporting slaves from Rio de Janeiro and selling slaves in other parts of southern Brazil incurred extra costs for merchants beyond those involved in selling slaves in the initial port of arrival. This may largely explain the different markups in slave prices between Africa and Rio de Janeiro on one hand and Africa and Santos, Rio Grande do Sul, and Minas Gerais on the

Table 10.6 Slave Purchases and Sales in the Port of Rio de Janeiro, Taking In Flows from Angola and Benguela, and to Rio de Janeiro, Minas Gerais, Santos, and Rio Grande do Sul, 1810–12 1810 Number Slaves imported from Angola and Benguela

13,954

Percentage 100

1811 Value

925:560$000

Percentage 100

Number 13,612

Percentage 100

Value 862:700$000

1812 Percentage 100

Number 11,238

Percentage 100

Value 692:789$000

Percentage 100

Slaves exported to Rio Grande do Sul

552

4.0

66:816$000

7.2

1,174

8.6

143:787$600

16.7

1,168

10.4

175:035$300

25.3

Slaves exported to Santos

320

2.3

40:960$000

4.4

248

1.8

31:744$000

3.7

299

2.7

38:272$000

5.5

Slaves exported to Rio de Janeiro state

6,977

50.0

638:695$511

69.0

6,806

50.0

917:033$634

106.3

5,619

50.0

524:926$980

75.8

Slaves exported to Minas Gerais

6,105

43.7

1:117:740$030

120.8

5,384

39.6

1:450:869$552

168.2

4,152

36.9

775:759$680

112.0

Sources: “Resumo dos mapas de importação e exportação . . .” (1808–14); Arquivo Nacional, Junta do Comércio, caixa 448, pacote 1; Inventários post-mortem (1789–1830), Arquivo Nacional.

298

Manolo Florentino

other. But it seems reasonable to assume that additional shipping and distribution costs did not account for all the price differential; the demand for slaves throughout southern Brazil affected the prices of slaves in Rio itself and facilitated sales to regional subports that offset the overseas debts incurred by Rio slave traffickers. Precisely how big that contribution was is impossible to say with the evidence currently available to us. The most extreme assumptions in calculating profits—which involve taking the price differentials between Africa and the various Brazilian ports and making no allowance for shipping and distribution costs—suggest that the contribution of reexporting of slaves from Rio de Janeiro toward clearing Rio slave traffickers’ debts was possibly huge. The gain from reexporting slaves to Santos and Rio Grande do Sul was, on these assumptions, greater than Rio de Janeiro’s accumulated deficit of 459:574$479 réis from trade with Africa in 1810–12. Similarly, the proceeds of sales to Minas Gerais were three times greater than the debt of 1:105:640$220 réis accumulated with India in the same period, while profits from the resale of slaves into Rio’s own hinterland conceivably amounted to close to ten times Brazil’s accumulated debt of 267:040$755 réis with Portugal. (The last calculation excludes remittances of gold and silver.) Overall, the 5:921:640$287 réis accumulated in sales to Santos, Rio Grande, Minas Gerais, and the rural zone of Rio de Janeiro may have been over three times greater than all of Rio de Janeiro’s deficits with Africa, Asia, and the metropolis combined, or close to three times the value of all the port’s sugar exports to Portugal between 1810 and 1812.47 Such estimates are based, however, on the most extreme assumptions. The contribution of slave reexport and sale outside Rio de Janeiro itself to covering the deficit with overseas markets was likely much smaller than these figures suggest. Nevertheless, returns from reexports made some level of contribution to keeping Rio’s commercial wheels turning in the early nineteenth century, helping to transform the port into a more complex mercantile center in which export of plantation products was matched by reexport of African slaves throughout the internal colonial market. These reexports were supported by a widespread network emanating from Carioca mercantile capital and embracing the pampas of the south and the Minas Gerais uplands, where purchasers of reexported slaves helped finance Rio de Janeiro’s trade deficits with Africa, Asia, and the metropolis. Buyers in the interior in turn depended for their markets not just on international trade but on regional markets for basic foodstuffs and other goods, most of them involving uncountable numbers of small, everyday transactions by a multitude of small- and medium-scale

Colonial Markets and Slave Families in Rio de Janeiro

producers. Taken as a whole, these transactions generated a revenue much larger than that created by agro-exporting monoculture, the most profitable and specialized sector of the colonial economy. Trends in Rio de Janeiro’s slave trade in 1790–1830 require us, therefore, to rethink the traditional view of colonial society that focuses on large-scale slave ownership and largely ignores smaller-scale free as well as slave-owning units that were clearly operating above subsistence levels.48

TRAFFICKING, SLAVERY, AND SLAVE FAMILIES

Although the port of Rio de Janeiro was an important center for the redistribution of Africans throughout south central Brazil, a significant proportion of arrivals remained within the captaincy. The proportion of African-born slaves among the captives in probate inventories in the Carioca agricultural districts jumped from nearly 50 percent in 1789 to 60 percent after the Lisbon court moved to Brazil in 1808. In urban inventories from the same period, the percentage of “African” slaves (as opposed to “crioulos,” or slaves born in Brazil) grew from 65 percent to 79 percent. Table 10.7 indicates that slave population growth in Rio de Janeiro in 1789–1832 depended on a constant flow of new captives from Africa. The trade ensured in turn that adults constituted a clear majority of the slave population; between 1789 and 1832, adults constituted more than half of the captives in all cases studied and were two-thirds of the population in the cities. Within the adult population males outnumbered females, in some cases by two to one. The children arising from adult slave unions were never sufficient to ensure natural reproduction of the slave population. We should not overlook, however, the significant population of crioulo children. These were in all cases more than half of the slave population native to Brazil.49 This was partly a result of slaves’ short life expectancies; no more than 10 percent of crioulos reached the age of forty. The number of crioulos also stemmed from slaves’ own efforts to overcome the rootlessness of a forcibly displaced population. In this respect, the number of crioulos reflected the creation and survival of family life among slaves. Until recently, there were few studies of slave families in Brazil, since it was commonly believed that the slave trade, with its marked sexual imbalance, inhibited the development of stable slave unions. Probate inventory lists from the rural zone of Rio de Janeiro, however, provide glimpses of slave families

299

Table 10.7

Distribution of Slaves in Rio de Janeiro by Age Groups and Sex, 1789–1832 1789–1808 Born in Brazil

Sectors Rural sectors

Urban sectors

1810–32 Africans

Born in Brazil

Age Groups

Age (%)

Male (%)

Number of Slaves in Sample

0–14 years 15–40 years +40 years

53.3

55.1

497

3.0

76.0

25

53.5

51.9

37.8

49.1

352

66.6

60.0

560

36.5

8.9

51.8

83

30.4

71.0

255

53.0

44.7

132

7.2

70.6

41.4

52.4

103

70.7

5.6

50.0

14

22.1

0–14 years 15–40 years +40 years

Sources: Same as table 10.1.

Age (%)

Africans

Male (%)

Number of Slaves in Sample

Age (%)

Male (%)

Number of Slaves in Sample

Age (%)

Male (%)

Number of Slaves in Sample

916

6.6

65.1

166

54.8

626

71.8

69.8

1,809

10.0

55.8

172

21.6

71.1

546

34

55.8

54.0

285

9.2

67.2

174

64.9

333

38.4

49.0

196

74.3

67.2

1,398

71.2

104

5.8

63.3

30

16.5

73.2

310

Colonial Markets and Slave Families in Rio de Janeiro

via lists of immediate relatives, notably parents and children and spouses without children, and one list includes details of parents and their respective children. This evidence, although rather limited, suggests that standard arguments about slave families are no longer sustainable. In particular, it suggests that the alleged incompatibility between the Atlantic traffic and slave family life is misleading. On the contrary, it reveals the presence of cultural patterns, including the reformulation of kinship, matrimonial endogamy according to “nation” of origin, and the predominance of older slaves in the matrimonial market, through which captives continually sought to reconstruct family life. It is reasonable to assume that family relations under slavery varied in accordance with shifts in the scale of the Atlantic traffic. The period from 1790 to 1830 divides, as we have seen, into three phases: a phase in which levels of disembarkations were stable (1790–1808), which we will call phase A; another, designated as phase B, in which disembarkations grew through time (1809–25); and a final phase, phase C, in which the supply of slaves in Africa was under threat (1826–30). These phases are captured in table 10.8, which also reveals the percentage of slaves with primary kin on different sizes of holdings. During phase A, when imports were relatively stable, captives with relatives made up about a third of all slaves in our sample, with some variation according to size of holding. This share fell to about a quarter during the growth phase of the traffic, before rising again to about a third in 1826–30. The decline in the percentage of slaves with primary relatives as import levels moved from phase A to phase B is unsurprising in view of the large numbers of new slaves arriving with no parental ties at all. The final phase of imports in 1826–30, when disembarkations reached exceptional levels, is something of a paradox, as the percentages of slaves with relatives returned to levels last seen in the period before 1809. While men remained a majority of the adult imports through 1830, it is possible that, faced with the official abolition of the slave trade, owners increasingly began to value female slaves and the reproductive capacity of African women. Some two-thirds to three-quarters of slaves in the Rio region failed to leave evidence of documented kin. Most documents consulted focus on primary relatives, though it is possible that other documents revealing other relationships, such as grandparents and children, uncles and aunts, nephews and nieces, or cousins and even godparents may raise the proportion of slaves with kinship ties closer to a half. One should not be surprised, however, that a high proportion of slaves were celibate, given the gender differences in slave imports. Kinship relationships existed, nevertheless, among

301

302

Manolo Florentino

Table 10.8 Slaves with Immediate Kinship Links on Plantations in Rural Rio de Janeiro by Size of Plantation and Phase of Slave Trade, 1790–1830 1790–1808

1809–25

1826–30

Size of the plantation’s holdings

1–9 slaves

10–19 slaves

+20 slaves

1–9 slaves

10–19 slaves

+20 slaves

1–9 slaves

10–19 slaves

+20 slaves

Percent with immediate kin

31.0

24.0

42.0

25.0

20.0

26.0

24.0

8.0

37.0

Sources: Same as table 10.1.

all age groups of slaves in the rural areas of Rio de Janeiro, a point illustrated by figure 10.4. The lower level of kin relationships among the very young largely reflected their forced separations from their parents, whether through sales or division of inheritances or from their parents’ emancipations or deaths. These tendencies were reinforced by increased African arrivals and the fact that many, if not most, recently disembarked slaves had no relatives in Brazil. Among slaves fifteen years old or more, however, there is evidence of emerging levels of primary kinship networks. This suggests that slaves who were arriving at their economically most valued age were able to begin to establish relationships. Familial socialization and the slave trade were not incompatible. Various cultural practices encouraged kinship not only to develop but to become the norm. Family ties and consensual unions were reflected in increasing birth rates. To be a slave child was almost synonymous with having been born in Brazil. The exception proves the rule. Between 1790 and 1830, we have been able to find only one family made up of an African mother and her African-born child in the probate inventories. They were Catarina Benguela, twenty-three years old, and her son Manoel, four, both of the same ethnic group. Generally, slaves younger than fourteen constituted a quarter to a third of the captive population. The evidence conveys a clear sense of procreative urgency among slaves. Captive women in the rural sector of Rio de Janeiro gave birth to their first child between fourteen and seventeen years if they were crioulas and between sixteen and nineteen if they were Africans. Slave women in rural Rio thus became pregnant earlier than free colonial women, who tended to have their first child when nineteen or twenty (earlier than free women in early modern Europe, whose equivalent age was around twenty-five). It appears that Brazilian slaves’ procreation patterns paralleled those of contemporary Africa, where women generally married and gave

Colonial Markets and Slave Families in Rio de Janeiro

Figure 10.4 Shares of Slaves Identified with Immediate Kin and Shares of Slaves from Africa by Age Group in the Rural Zone of Rio de Janeiro, 1790–1830 Sources: Same as table 10.1.

birth to their first child shortly after puberty. One might reasonably suppose, therefore, that the pattern of early childbirth among slave women in rural Rio de Janeiro reflected the transposition and adaptation of an African cultural pattern. Captives in rural Rio de Janeiro conceived, on average, at two- to three-year intervals (six months more if the mothers were African-born). This was a longer interval than the norm among free white Brazilian women, suggesting a further adaptation of yet another African cultural pattern, one in which a prolonged period of breastfeeding delayed renewed pregnancy for three years or so. This extended spacing of birth intervals, however, did not result in slave women who survived to menopause having fewer children, given that such women continued their childbearing well into middle age—indeed, almost up to the biological limits of fertility. The evidence suggests, in fact, that procreation among slave women typically ceased around the age of thirty-nine— a pattern unlike that found across much of Africa, where reproduction commonly ceased in women’s early thirties. This suggests that slave women, both crioulas and African, may have redefined in Brazil some aspects of African patterns of reproduction. In short, slave women began their reproductive lives shortly after puberty and had children once every two to three years—trends that apparently matched those found in parts of Africa. Unlike African women, however, slave women in rural Rio de Janeiro continued their reproductive lives until the onset of menopause. Procreation can thus be seen as a form of

303

304

Manolo Florentino

prolonged investment in creating family ties and fostering new kinship structures. These in turn were important factors in shaping the social lives of slaves. Another behavioral pattern that underlines the urgency with which slaves sought to establish kinship ties relates to marriage patterns and their adjustment in the face of variations in slave import levels. Figure 10.5 plots the percentages of nuclear, or religiously sanctioned, unions against matrifocal unions, defined here as families formed by a single African or crioula mother and her children. The two trends are graphed in terms of the phases of slave imports noted earlier, thereby enabling us to track the relationship over time between marriage patterns and the scale of slave imports. The figure and related data suggest that before 1809, when import levels were fairly constant, slave marriages were largely realized within Brazilian social norms, subjected to religious sanction, and generating nuclear families comprising married parents and their children or childless spouses. Nuclear families remained in the clear majority during 1790–1808 and included the greater part of the relatives listed. This situation changed, however, as the level of slave imports climbed from 1809 onward. Over the whole period, matrifocal family units grew in importance and even came to eclipse their nuclear and religiously sanctioned counterparts. Changes in the importance of nuclear and matrifocal family units between 1790 and 1830 provide insights into how, in the face of variations in slave imports, slave communities sought to stabilize and reproduce themselves. Marriage and family were key mechanisms in incorporating slaves into wider

Figure 10.5 Slave Families in Rural Rio de Janeiro by Nuclear or Matrifocal Status, 1790–1830 (in percentages; percentages sum to 100) Sources: Same as table 10.1.

Colonial Markets and Slave Families in Rio de Janeiro

Brazilian society and inculcating appropriate social attitudes. Legal matrimony served such a function at times when the inflow of new slaves was relatively stable. The much higher levels of slave imports from 1809 onward seems to have put pressures on legal marriage as a mechanism of slave socialization and prompted wider resort to matrifocal units. In short, the growth in nonreligious unions suggests that slave acculturation through matrimony had reached its limits as a social mechanism by 1808. Thereafter, increased numbers of imported and socially disorganized captives subverted earlier norms of slave socialization. During 1809–30, unions without religious blessing came to predominate as an adaptation by slaves to changing import patterns. The marriages referred to here formed the most commonly cited kinship linkage among slaves: 30 percent to 40 percent of all relatives registered in the probate inventories were spouses, a number that rose to 50 percent to 57 percent when only adult captives are taken into consideration. Considering only legal marriages, it appears that both Africans and crioulos had access to religiously sanctioned unions. Africans, however, always dominated in this form of matrimony, forming anywhere from 56 percent to 67 percent of all couples registered, as shown in figure 10.6. This higher incidence of sanctioned matrimony among Africans can be partly explained by the fact that the majority of these captives were older than fifteen on their arrival in Brazil. Given, however, the high percentage of adult men in this group, one might have expected a higher incidence

80

60

%

40

20

0 1790–1808 Africans x Africans

1809–1825 Years Crioulos x Crioulas

1826–1830 Mixed

Figure 10.6 Slave Marriages Sanctioned by the Church in Rural Rio de Janeiro, 1790–1830, by Origins of Spouses and Groupings of Years Sources: Same as table 10.1.

305

306

Manolo Florentino

of mixed African-crioulas marriages in periods of high slave imports, as African men sought mates among the crioulas. This, as figure 10.6 shows, plainly did not happen: on the contrary, during periods of increased slave importation, the proportion of mixed marriages tended, if anything, to decline. This finding highlights another element in the development of slave kinship relationships, namely that the choice of partners was influenced by ethnic considerations, including country of origin. As figure 10.6 shows, out of every ten couples, five to seven were typically formed by Africans, one to three were formed by crioulos, and two were mixed. Endogamy based upon origin was thus the norm. But a closer look at figure 10.6 reveals a declining percentage over time of African marriages, starting from a high of more than 60 percent and decreasing to under 60 percent. Mixed marriages also declined, from more than one in five to about one in ten. By contrast, the ratio of crioulo marriages rose from one in ten to almost one in three. This trend suggests that, as the scale of slave imports increased through Rio de Janeiro from 1808 onward, the crioulos population tended to close in upon itself. In periods of slave-trade expansion, then, demarcation in marriage between Africans and crioulos increased. The search for explanations of this finding leads us to another cultural practice that permitted the expression of slave kinship. Figures 10.7 and 10.8 present data for 1790–1808 on the average age differentials between slave couples at marriage by groupings of age for Africans (figure 10.7) and crioulos (figure 10.8). The data show that, in this period of stability in slave imports, both African men and crioulo men marrying after the age of twenty-four

Figure 10.7 Age Differentials of African Slave Couples Marrying in Rural Rio de Janeiro by Groupings of Ages of Spouses, 1790–1808 Source: Probate Inventories, 1790–1808 (Arquivo Nacional).

Colonial Markets and Slave Families in Rio de Janeiro

Figure 10.8 Age Differentials of Native-Born Slave Couples Marrying in Rural Rio de Janeiro by Groupings of Ages of Spouses, 1790–1808 Source: Probate Inventories, 1790–1808 (Arquivo Nacional).

tended to be older than their spouses. The older the male slave, the greater the age difference tended to be between him and his spouse, while the reverse applied to women. A pattern, therefore, seems to have been evident before 1809: as men grew older they married increasingly younger women, thus claiming a disproportionate share of younger and still fertile women entering the marriage market. This was true of all older men but especially of those born in Brazil. One consequence was that an unexpectedly high proportion of the youngest males seeking partners married women older than themselves, including women no longer capable of reproduction. This was especially so for male slaves born in Africa, who found themselves partnering women considerably older than themselves. Far removed from commonly accepted notions that conjugal unions formed more or less randomly, these patterns suggest that older males dominated the marriage market for slaves in rural Rio de Janeiro in times of modest and stable slave imports. Figures 10.9 and 10.10 plot, for the period after 1808, patterns of age differences for African and crioulo couples, using the same age bands as those for the period 1790–1808. The figures suggest that after 1808, when slave imports grew, the age differences between husbands and wives found in the earlier period tended to narrow, and in the case of the youngest group of Africans— those fifteen to twenty-four—a striking reversal of age profiles occurred, with males now being older than their partners. Increases in slave imports after

307

308

Manolo Florentino

Figure 10.9 Age Differentials of African Slave Couples Marrying in Rural Rio de Janeiro by Groupings of Ages of Spouses, 1809–30 Source: Probate Inventories, 1809–30 (Arquivo Nacional).

Figure 10.10 Age Differentials of Native-Born Slave Couples Marrying in Rural Rio de Janeiro by Groupings of Ages of Spouses, 1809–30 Source: Probate Inventories, 1809–30 (Arquivo Nacional).

1808 were accompanied, therefore, by changes in marriage patterns in rural Rio de Janeiro, perhaps linked in some way with a dilution of the control by older males in the matrimonial market. We can only speculate at this stage about the implications of these changes in marriage patterns in rural Rio de Janeiro during 1790–1830 and the impact upon them of higher levels of slave imports after 1808. It is possible that, given the apparent strong desire of slaves for unions, families, and kinship, the disproportionate influence of older males in the pre-1809 marriage market re-

Colonial Markets and Slave Families in Rio de Janeiro

flected their social influence within slave communities. If this was true, then their apparently diminished influence after 1808, when imports of young males increased sharply, may have signaled greater competition for partners and increased tension within slave communities, adding yet another factor to those that shaped slaves’ lives. Should we assume that increased slave imports intensified coercive unions or that slave communities experienced a period of internal social warfare in which women became trophies on the battlefield? Not necessarily. But one outcome of enlarged slave imports that seems to be corroborated by our sources was increased rivalry or tension between Africans and crioulos. Native-born males seem to have lost some control over fertile women after 1808, a not insignificant issue for a population group marked by large gender imbalances and situated at the bottom of the colony’s social pyramid. This perhaps helps to explain the increased tendency we noted earlier of the crioulo group to become introverted and to reject mixed unions. It may, in turn, have been just one of the mechanisms through which captive communities were able to adjust to exogenous factors such as changes in levels of slave imports, sustaining their populations and thus inadvertently extending the feasibility of the slave labor system. NOTES

1. Jacob Gorender, O escravismo colonial (São Paulo, 1978), 120, 208, 211, and 544; and Fernando A. Novais, Portugal e Brasil na crise do Antigo Sistema Colonial (1777–1808) (São Paulo, 1983), 104–5. 2. For a different point of view, see Manolo Florentino, Em costas negras: Uma história do tráfico de escravos entre a Africa e o Rio de Janeiro (Séculos XVIII e XIX) (São Paulo, 1997). Another contribution has been made by Luiz Felipe de Alencastro, O trato dos viventes: Formação do Brasil no Atlântico Sul (São Paulo, 2000). 3. João José Reis, A morte é uma festa: Ritos fúnebres e revolta popular no Brasil do século XIX (São Paulo, 1995), 38. 4. Antônio Duarte Nunes, “Memórias públicas e econômicas da cidade de São Sebastião do Rio de Janeiro para o uso do Vice-rei Luiz de Vasconcellos por observação curiosa dos annos de 1779 até o de 1789,” Revista do Instituto Histórico e Geográfico Brasileiro 48 (1884): 29; Fundação Instituto Brasileiro de Geografia e Estatística (IBGE), Estatísticas históricas do Brasil (Rio de Janeiro, 1986), 3:30. 5. This period (1790–1830) was largely situated during the time in which slaving was open to all of Brazil’s inhabitants and also completely legal. The Portuguese Crown recognized that metropolitan merchants were not able to supply the colony’s slave needs in 1758, when it declared slave trading as being a commerce open to all imperial subjects. See Maurício Goulart, A escravidão africana no Brasil (das origens à extinção do tráfico) (São Paulo, 1975), 192.

309

310

Manolo Florentino

6. Moses I. Finley, Escravidão antiga e ideologia moderna (Rio de Janeiro, 1991). 7. Stanley J. Stein, Vassouras: A Brazilian Coffee Country, 1850–1900 (Cambridge, Mass., 1957), 53. 8. The 832,380 Africans disembarked in Rio de Janeiro between 1790 and 1830 represent about 19.3 percent more than the 697,945 that I had earlier estimated in Em costas negras. 9. Though officially abolished in 1830, the slave trade to Brazil continued as contraband. Only in 1851 would the Brazilian overseas trade in African slaves be completely eliminated. See David Eltis, “The Nineteenth-Century Transatlantic Slave Trade: An Annual Time Series of Imports into the Americas Broken Down by Region,” Hispanic American Historical Review 67 (1987): 109–38 passim. 10. Florentino, Em costas negras, table 4. 11. Philip D. Curtin, The Atlantic Slave Trade: A Census (Madison, Wis., 1969), 216. 12. Nireu Oliveira Cavalcanti, “O comércio de escravos novos no Rio setecentista,” in Manolo Florentino, ed., Tráfico, escravidão e liberdade (Rio de Janeiro, 2005), 15–77. 13. See José C. Curto and Raymond R. Gervais, “A dinâmica demográfica de Luanda no contexto do tráfico de escravos do Atlântico Sul, 1781–1844,” Topoi 4 (2002): 85–138. 14. Joseph C. Miller, Way of Death: Merchant Capitalism and the Angolan Slave Trade, 1730–1830 (Madison, Wis., 1988). 15. See Alberto da Costa Silva, A manilha e o libambo: A Africa e a escravidão, de 1550 a 1700 (Rio de Janeiro, 2002), for the sixteenth and seventeenth centuries. 16. J. D. Fage, History of West Africa (Cambridge, 1959), 94. 17. See Richard Gray, “Introduction,” in Cambridge History of Africa, 8 vols. (Cambridge, 1975–86), 4:1–13; Patrick Manning, Slavery and African Life (Cambridge, 1990). 18. John Thornton, Africa and Africans in the Making of the Atlantic World, 1450–1680 (Cambridge, 1992). 19. Claude Meillassoux, L’anthropologie de l’esclavage (Paris, 1985). 20. See “Balanço da importação e exportação deste Reino de Angola desde o anno de 1785 em que teve princípio o estabelecimento da alfândega até o anno de 1794 inclusive,” Arquivo do Instituto Histórico e Geográfico Brasileiro, Rio de Janeiro, Seção dos Manuscritos, lata 77, documento 1. Many of these products met demands not directly related to the slave trade. The total value of this material, however, was quite small if compared with that required to purchase captives. It is thus quite possible that the imports registered by the Luanda customs house are a fairly accurate reflection of the goods used to form trade stocks for the purchase of slaves. Regarding the role of these goods in the traditional African economy, see Miller, Way of Death. 21. Anon., “Instruções em que se Mostra a Formalidade do Comércio do Reyno de Angola e Benguella, e o quanto tinha Florescido desde o seu Princípio athé o Anno de 1760 em que Principiou a sua Ruína . . . ,” manuscript section of Biblioteca Nacional, I:4455–94. 22. “The shipments of cotton cloth increased in size and value, making them Portugal’s most valuable export to Brazil toward the end of the eighteenth and the beginning of the nineteenth centuries.” From Valentim Alexandre, Os sentidos do Império: Questão nacional e questão colonial na crise do Antigo Regime português (Lisbon, 1993), 49. On the following page, Alexandre tells us that “an unfortunately unquantifiable portion of these textiles was made up of white cloth made in Asia and which, after dyeing and

Colonial Markets and Slave Families in Rio de Janeiro

printing in Portugal, was shipped on to Brazil.” See also Jorge Pedreira, Os homens de negócio da praça de Lisboa (Lisbon, 1996), chaps. 1 and 2. 23. Regarding Rio’s growing slave trade, see “Resumo dos mapas de importação e exportação dos estados da India, Africa e Brasil: 1808–1814,” Arquivo Nacional, Junta do Comércio, caixa 448, pacote 1. 24. Ibid.. The same documents indicate that, in 1813, Rio exported 428:623$932 réis of textiles to Angola out of a total of 600:218$602 réis purchased by that colony. 25. See José H. Rodrigues, Brasil e Africa: Outro horizonte (Rio de Janeiro, 1964), 27–30. Roy Glasgow states that the Angolan trade was controlled by Brazilian merchants starting in the second half of the seventeenth century. See Roy Arthur Glasgow, Nzinga: Resistência africana à investida do colonialismo português em Angola, 1582–1663 (São Paulo, 1982), 146–47. 26. Florentino, Em costas negras. 27. “Processo de Seqüestro de Bens de Micaela Joaquina Nobre Contra o Testamenteiro dos Bens de João Luciano Moura, Antônio Lopes Anjo,” 1809, Arquivo Nacional, Junta do Comércio, caixa 361, pacote 3. 28. See Arquivo Nacional, Junta do Comércio, caixa 398, pacote 1. 29. See part of the inventory in the Arquivo Nacional, Junta do Comércio, caixa 348, pacote 1. 30. “Descrição do estado do Brasil, suas capitanias, produções e comércio,” British Library additional ms 13981, cited in Rodrigues, Brasil e Africa, 26. 31. Corcino Medeiros dos Santos also considers the Indian-Angolan commerce to have been fundamental. Between 1785 and 1794, the documents of the Luanda customs house indicate that India was the city’s principal commercial partner (with textiles being the largest import from India). See Santos, O Rio de Janeiro e a conjuntura atlântica (São Paulo, 1993), 156. 32. Arquivo Nacional, Junta do Comércio, caixa 377, pacote 1. 33. “Instruções em que se Mostra a Formalidade do Comércio.” 34. Luís António de Oliveira Mendes, Memória a respeito dos escravos e tráfico da escravatura entre a costa d’Africa e o Brasil, apresentada à Real Academia de Ciências de Lisboa em (1793) (Porto, 1977), 48. 35. José Capela, O tráfico de escravos nos portos de Moçambique, 1733–1904 (Porto, 2002), 260–61. 36. Joseph C. Miller, “Mortality in the Atlantic Slave Trade: Statistical Evidence on Causality,” Journal of Interdisciplinary History 3 (1981): 385–523, esp. 198, 413–14; Antônio Carreira, A companhia geral do Grão-Pará e Maranhão (São Paulo, 1988), 136. 37. According to Glasgow, the loss in transit along the Angola-Brazil route in the seventeenth century was around 30 percent. See Glasgow, Nzinga, 140. 38. Pedro Carvalho de Mello, “Estimativa da longevidade de escravos no Brasil na segunda metade do século 19,” Estudos Econômicos 13 (1983): 172–73. 39. Working with a distinctly different methodology, David Eltis postulates similar profit levels for the Cuban trade in 1826–35 (19.6 percent) and 1836–45 (19.2 percent), and for the Brazilian trade to the south of Bahia in 1831–40 (17.8 percent) and 1841–50 (21.8 percent). See Eltis, Economic Growth and the Ending of the Transatlantic Slave Trade (New York, 1987), 281.

311

312

Manolo Florentino

40. João Fragoso and Manolo Florentino, “A comunidade de mercadores do Rio de Janeiro e o mercado atlântico português na passagem do século XVIII para o XIX,” in Cristiana Bastos, Miguel Vale de Almeida, and Bela Feldman-Bianco, eds., Trânsitos coloniais: Diálogos Luso-Brasileiros (Lisbon, 2002), 321–42; José J. Arruda, O Brasil no comércio colonial (São Paulo, 1980), tables 17 and 19. 41. “Resumo dos mapas.” 42. Between 1810 and 1812, cloth accounted for 98 percent of Rio’s imports from Goa, and gold coin accounted for 79 percent to 98 percent of the city’s exports to India. See “Resumo dos mapas” and João Fragoso and Manolo Florentino, “Negociantes, mercado atlântico e mercado regional: Estrutura e dinâmica da praça mercantil do Rio de Janeiro entre 1790 e 1812,” in Júnia Ferreira Furtado, ed., Diálogos oceânicos: Minas Gerais e as novas abordagens para uma história do Império Ultramarino português (Belo Horizonte, 2001), 155–79. 43. For example, the Luanda customs house registers declare that “Large portions of gold coins also come into this Kingdom, which some Commissioners bring from Brazil for the purchase of slaves.” See “Balanço da importação.” 44. Regarding the enormous quantities of precious metals exported by Rio de Janeiro between 1763 and 1808, Corcino Medeiros dos Santos affirms that “this hemorrhage [of coins leaving Rio for the kingdom] did not lead to the collapse of carioca commerce because the specie coined in other captaincies also freely circulated in the city; quite apart from this, one could find gold dust circulating in Minas Gerais, Mato Grosso and Goiás, and even on occasion in the city of Rio, in spite of the restrictions against its use.” Santos, Relações comerciais do Rio de Janeiro com Lisboa (Rio de Janeiro, 1980), 192–95. 45. João Fragoso, Homens de grossa aventura (Rio de Janeiro, 1998), 177; IBGE, Estatísticas históricas, 30. 46. Joaquim Felício dos Santos, Memórias do distrito diamantino (Petrópolis, 1978), 282–83. 47. “Quadros Financeiros, Balanças e Minutas, sobre a Administração do Ministro Thomás Antônio de Villa Nova Portugal, 1809–1830,” Biblioteca Nacional, Rio de Janeiro, Seção de Manuscritos, II, 30, 27, 27; “Resumo dos mapas.” 48. Fragoso and Florentino, “Negociantes, mercado atlântico.” 49. Cavalcanti’s “O comércio de escravos” shows that, between 1687 and 1809, children made up only 17 percent of the population of Rio de Janeiro. Although there was a slight imbalance among these in favor of males, among the elderly and adults the gender imbalance took on large proportions (64 percent were males). This sampling included 1,236 slaves mentioned in sixty probate inventories.

Chapter 11 The Suppression of

the Slave Trade and Slave Departures from Angola, 1830s–1860s Roquinaldo Ferreira

Approximately 1.3 million Africans were taken to the Americas between the 1830s and the 1860s—the last three decades of the slave trade. By then most of the slave trade focused on Central Africa, and the number of African regions shipping significant number of slaves to the Americas was fewer than it had been in the eighteenth century. Embarkations of slaves from Senegambia were already relatively small and decreased further in the last decades of the slave trade. On the Gold Coast, the British withdrawal from Atlantic slaving in 1807 had an immediate impact on slave exports. In the Bight of Benin and the Bight of Biafra, exports continued until the 1840s and 1850s but were significantly affected by the rise of attempts to suppress the trade. Unsurprisingly, by the mid-nineteenth century, Angola and Kongo accounted for approximately 50 percent of all enslaved Africans taken across the Atlantic.1 The changes sketched above were primarily related to influences external to Africa—particularly efforts to halt the trade.2 In Luanda suppressive measures clearly constituted a paramount force behind the drive to halt shipments of slaves in the 1840s. It was a key development 313

314

Roquinaldo Ferreira

in the struggle to end the Angolan slave trade, because of the pivotal role that Luanda played not only as a slave port but also as a logistical hub for shipments from other Angolan ports. David Eltis has pointed out that further understanding of the end of the slave trade necessitates a more comprehensive analysis of local factors, such as geography, internal African politics, and the particular conditions of commerce along coastal Africa.3 In northern Angola, for example, an unfriendly business environment and competition from African traders had a greater impact than British naval patrols on the demise of slave shipments.4 This chapter focuses on regional variations in Angola to explore how external and local forces shaped the shipments of slaves between the 1830s and the 1860s. It combines evidence from the new Transatlantic Slave Trade Database (henceforth TSTD2) and a wide variety of archival data to reconstruct local conditions in which the trade operated in different Angolan regions. I examine how differences in the dynamics of the coastal trade in the three Angolan slaving regions— northern Angola, Luanda, and Benguela—influenced the supply of slaves from the interior. Special attention is given to the understudied slave trade in Benguela in order to redress a Luanda-centered imbalance in the current literature. My analysis suggests that a combination of internal and external factors shaped the last decades of the Angolan trade and ultimately led to its demise.

BACKGROUND

Two cautionary notes are in order before delving into the details of the Angolan trade during the illegal period. First, the slave trade in Portuguese territories in Africa was formally outlawed only in 1836. However, the passing in Brazil of a law in 1831 that banned imports of slaves threw the Angolan trade into a state of effective illegality. The impact on Angola of the Brazilian law was a function of the close commercial ties between the two regions, since Angola was the primary supplier of slaves to Brazil. Scholars of Brazilian history have long downplayed the importance of the 1831 legislation by arguing that it was not fully enforced.5 Yet a closer look reveals that it led to a dramatic, albeit brief, reduction in imports of slaves, which in turn affected slave demography in Brazil. In the interior of São Paulo, for example, the number of Africanborn slaves dropped from 55 percent in 1829 to 21 percent in 1835.6 In Rio de Janeiro, “the slave market . . . changed substantially with the ‘legal’ prohibition on slave imports.”7 Furthermore, the law reverberated across the Atlantic and altered the ways in which the Angolan trade was conducted.8 First, it added further momentum to

Slave Departures from Angola

a process of decentralization of shipments of slaves away from Luanda and toward northern Angola, which had been under way since the early nineteenth century, when the British dropped out of slaving. Second, it marked the beginning of more intense, if initially utterly failed, anti-slave-trade measures geared toward the southern Atlantic slave trade.9 Although the reduction in shipments of slaves from Luanda and Benguela was brief, the law seems to have initiated a spurt of suppressive measures that forced slave dealers to devise new organizational strategies to sustain the trade. These were effective enough to enable the continuation of shipments of slaves for several decades and in effect to hold the rising tide of abolitionism at bay.10 In this chapter, therefore, the Brazilian law is considered the threshold to the illegal phase of the Angolan trade. Most of the scholarship has taken a Luanda-centric approach to the Angolan trade, despite the existence of an entrenched three-pronged commercial structure (northern Angola, Luanda, and Benguela) since the late seventeenth century.11 In the sixteenth century, the trade was concentrated in northern Angola, but it expanded southward in the final decades of the century due to diminishing Portuguese control and the rising importance of creolized groups from São Tomé and Kongo.12 The southward expansion grew out of Portuguese plans to strengthen their control and led to the foundation of Luanda in 1576.13 In Benguela, after an early attempt to develop slaving failed in the first half of the seventeenth century, exports of slaves increased toward the end of the century and became a full-fledged independent operation in the first decades of the eighteenth century.14 Northern Angola, Luanda, and Benguela each had distinct patterns of internal and coastal commerce. In northern Angola, for example, the internal trading networks were dominated by Vili merchants from Loango and extended into Kongo territory and southward to central Angola. Unlike the networks supplying slaves to coastal Luanda, they were not monopolized by states or bureaucratic structures, and they displayed a high degree of decentralization that made them better suited to respond to the rising demand for forced labor from the Atlantic world in the late seventeenth century. These networks contributed significantly to key transformations in the enslavement of Africans, which made Angolan slaving move away from military techniques toward a variety of nonmilitary methods to acquire slaves. As the decentralized Vili trading networks grew, the supply of slaves to northern Angola became more reliable than the supply to Luanda. The advantages of this structure became clear in the late seventeenth century, when ships that would otherwise have conducted slaving in Luanda began tapping into northern

315

316

Roquinaldo Ferreira

Angola’s slaving, to compensate for the increased difficulties in embarking slaves at Luanda.15 David Eltis and David Richardson have suggested that the existence of a European coastal stronghold at Luanda might have created a more reliable supply of slaves.16 In fact, Luanda’s commercial relationships with internal regions providing slaves was marked by strong dependency on the Kasanje and, secondarily, the Matamba kingdoms. These two kingdoms were located in the Luandan hinterland and played a pivotal role in the region’s internal slave trade. Their rulers prevented agents operating on behalf of Luandan merchants from tapping into slave markets east of the Kwango River, and they effectively held a grip over the internal trade. Since the Luandan government was largely dependent on Kasanje for soldiers and protection, the relationship was tinged by political and military considerations, limiting attempts to diversify internal sources of slaves to Luanda. Kasanje’s leverage and semimonopolistic control over the internal commerce seriously undermined Luanda’s capacity to meet the increasing demands for labor throughout the Atlantic. This was particularly clear at the end of the seventeenth century, when Luanda failed to respond adequately to the rising demand for slaves in the wake of the discovery of Brazil’s gold mines.17 The regulatory bureaucracy on the coast also affected shipment of slaves. While northern Angola was by far the least regulated region, merchants operating out of Luanda and, to a lesser extent, Benguela dealt with corrupt administrations that sometimes posed serious impediments to their capacity to conduct business. However, the lack of regulatory bureaucracy in northern Angola in no way meant full freedom to conduct trade. African authorities exerted tight control over the coastal trade and were highly conscious of the benefits of commercial independence. They firmly rejected European attempts to encroach on their territories. This was particularly clear when African forces frustrated a British attempt to build a fort in Cabinda in 1721. Africans also defeated an expedition sent by the Portuguese to cut off African commercial links with the French in Cabinda in the 1780s. Indeed, for most of the eighteenth century, northern Angola evaded outside Portuguese influence altogether. In the seventeenth century, the three Angolan regions providing slaves for the Atlantic trade were relatively integrated. For example, approximately 30 percent of the Africans exported from Luanda in the 1680s came from Benguela. In northern Angola, the Dutch held the upper hand in the coastal trade in the second half of the seventeenth century, but ships from Brazil and

Slave Departures from Angola

Luanda were still permitted to embark slaves. Trade in the three regions gradually diversified in the eighteenth century, signaling the growing commercialization of the Angolan trade. By the 1730s, instead of complying with the mandatory stopover in Luanda to pay duties on slaves, most of the slavers bound from Benguela sailed directly to Brazil.18 In northern Angola, as the Dutch and the Portuguese lost ground to the British and the French in the eighteenth century, the trade became even more international and links with Luanda faded. In the early nineteenth century, the introduction of measures to abolish the slave trade changed this dynamic significantly by triggering tighter integration between northern Angola and Luanda. Benguela, however, stood in sharp contrast to the other regions, as it maintained closer ties with Rio de Janeiro than with any other city in the Atlantic world.19

BENGUELA

Recent estimates indicate that Benguela was second only to Luanda in terms of shipments of slaves between the 1830s and 1860s.20 Reconstructing patterns of trade in Benguela remains a problem, however, due to lack of documentation and the shadowy nature of the slaving business in that region. Since the early eighteenth century, Benguela had served as a safe haven for outcasts seeking to conduct slaving away from Luandan officials. Largely independent from Luanda, the city gravitated toward Brazil, receiving capital and absorbing the economic influence of merchants from Rio de Janeiro.21 In the nineteenth century, several Benguelan merchants, many of whom had either been born or lived in Rio de Janeiro, still kept close ties with that Brazilian city. Unsurprisingly, the Benguelan economy was more closely linked with Brazil than with Portugal. As late as 1837, for example, Benguelan imports of Brazilian rum—one of the main commodities in the internal slave trade—stood at a volume of 1,000 pipas a year, half of the amount of Brazilian rum imported by the primary Angolan slave port at Luanda.22 The domination of slave dealers in local politics was a key dimension of the continuing slave shipments in Benguela, and several dealers even held the position of governor of Benguela. In 1836, for example, Justiniano José dos Réis, a Brazil-born individual with deep links to the slave trade since the 1820s, was the governor of Benguela.23 In 1841 another merchant with clear connections to the slave trade assumed the governorship.24 Two years later, the position was taken up by Manoel Joaquim Teixeira, then one of the main actors in the slave trade.25 Teixeira maintained close contacts with Brazil and left for Rio in 1844.

317

318

Roquinaldo Ferreira

He continued to exert influence in Benguela until his return in the early 1850s, when he was accused by the British of seeking to revive the city’s slave trade.26 As late as 1850, British authorities argued that their years of efforts to halt the trade were being undermined by a governor sympathetic to slave dealers.27 Furthermore, local politics were driven by attempts to ban the trade and the animosity that mulattoes—mixed-race individuals, blacks, and locally born white individuals—held toward Portugal. In 1835, for example, racially driven tensions surfaced during an aborted revolt against the Benguelan administration.28 Mulattoes played a pivotal role in the local bureaucracy. In 1842, for example, while the official in charge of administering the customs house in Benguela was a black man, his immediate subordinate was a mixedrace official.29 With their close connections to African chiefs in the interior, mulattoes framed the struggle against the slave trade as a European attempt to take away their power. Invariably, they used the influential positions they held in the local administration to undermine attempts to enforce anti-slave-trade measures. In 1850 regional authorities stated that “mulattoes from this district [Benguela] have become offensive and restive, being particularly bitter toward Portuguese individuals whom they pejoratively called galegos, in addition to stating publicly that Benguela will only calm down when [the] galegos are expelled from positions they have come to hold in the administration.”30 In anticipation of the expected end of the importation of slaves to Brazil in 1830, the Lisbon government ordered the Benguelan governor to seek to reorient the local economy to a post-slave-trade scenario in 1827.31 One year later, in an early sign of anti-slave-trade sentiment in Brazil, the Luandan administration reported that ships bound from Luanda were facing difficulties entering Brazil because of British pressure on the Brazilian authorities.32 Despite the pressures, shipments of slaves from Benguela were still significant in 1830.33 In that year ten ships recorded in TSTD2 picked up 3,926 Africans from Benguela and took them across the Atlantic to Rio de Janeiro. The end of the legal slave trade in Brazil dealt a significant blow to the Benguelan trade. While an average twelve annual voyages had originated in the city during the legal trade in the 1820s, TSTD2’s data set includes only one voyage registered between 1831 and 1835. Between 1836 and 1839, the number of voyages followed a decidedly upward spiral, and the Benguelan trade seems to have surpassed its 1820s levels. Although TSTD2 identifies only three ships leaving Benguela in 1836, the number of ships that actually sailed that year may have been higher. This possibility is suggested by the number of applications submitted by ship captains for

Slave Departures from Angola

licenses to leave Benguela. Fourteen ships received licenses that year, and although none of the ships declared that they were carrying slaves, the fact that all presented fabricated routes—usually including Mozambique and Uruguay—strongly implies a trade in slaves. In following years, both TSTD2 and the license applications suggest that slave shipments grew further. In 1837, for example, TSTD2 indicates that eight ships left Benguela, but since the number of licenses reached twenty-five, the number might have been higher. Documentation from Brazilian archives shows that nine ships from Benguela were apprehended in Rio de Janeiro in the first half of 1837 alone, which offers further evidence that the overall number of slavers boarding slaves in Benguela was higher than the cases included in TSTD2.34 In 1838, according to the German traveler George Tams, approximately 20,000 slaves were embarked from Benguela.35 Unprecedented British and Portuguese actions against the slave trade provide further, albeit indirect, evidence of the surge in slave shipments in the following years. In 1839, in order to discourage embarkations of slaves in the vicinity of Benguela, the Luandan government decreed that no ships could leave the city after five in the afternoon.36 As indicated in TSTD2, these measures were largely unsuccessful; fourteen ships left in that year alone. As British forces zeroed in on Benguela, however, abolitionist measures were bolstered and seem to have yielded more favorable results.37 Backed by a controversial and shortlived convention signed by the governor of Angola, Antônio Manoel de Noronha—an anti-slave-trade official who held a brief tenure in Luanda—the British received far more legal and practical freedom to interfere with Portuguese property and citizens than the Portuguese government was prepared to concede at the time.38 The British actions were short-lived but sufficient to reduce the number of ships sailing in 1840 to only seven.39 Thereafter, British attempts to suppress the trade lost considerable momentum as Luandan merchants—most of them deeply engaged in the slave trade—led a campaign to overturn the anti-slave-trade legislation.40 This development seems to have spurred an immediate growth in slave shipments in 1841 and 1842.41 As shown by TSTD2, fifteen ships left Benguela both years, with five of them clearing out in a single month of 1841.42 British demands for the cessation of slave shipments continued but were ignored by Luandan officials, who argued that shipments per se were not the most crucial element of the trade and that these took place outside the city. Two reports written in 1841 hint at the magnitude of slave shipments from Benguela. The first mentions 8,000 slaves, and the second refers to 5,000 Africans held for

319

320

Roquinaldo Ferreira

embarkation in Benguela.43 Although these numbers do not provide direct evidence of shipment, they strongly suggest that Benguela was, at the very least, a significant entrepôt for the region’s slaving operations. To cope with continued efforts to restrict the traffic, slave dealers relied on several strategies, including the extensive use of bribes for the few local officials who might have opposed the slave trade. This was reflected in a sequence of investigations. In 1840 British demands forced the Luandan government to launch an investigation into widespread corruption.44 One year later, a governor of Benguela fled to Pernambuco in northeastern Brazil after being charged with participating in the slave trade.45 In 1843 Lisbon ordered a full-scale investigation of Benguelan officials following the British detention of a Brazilian slave ship.46 As late as 1849, a governor of Benguela was accused of receiving bribes in exchange for turning a blind eye to slave shipments.47 In contrast with Luanda, where by the early 1840s Portuguese legislation had made a serious impact on the trade, the primary threat to the Benguelan slave dealers came from British cruisers. To circumvent them, slave traders declared their vessels to be sailing to ports other than Brazil. Thus, twenty-nine out of the forty-six ships that received licenses to embark from Benguela between 1835 and 1837 declared the Portuguese colony of Mozambique, not Brazil, as their final destination.48 By so doing, they took advantage of the clause of the Portuguese anti-slave-trade decree from 1836 that permitted the transport of slaves between different Portuguese colonies. In case ships were apprehended by the British, the loophole allowed them to provide justification for carrying enslaved Africans. This strategy was part of a deception that began in Brazil, where ships would obtain documents declaring Montevideo the final destination and places like Benguela and Ambriz as stopovers on the way to Uruguay.49 In fact the majority of the ships calling into Benguela in the 1830s were vessels that sailed to Angola to pick up slaves before returning to Brazil. Between May and July 1837, for example, nine Portuguese slave vessels seized in Rio de Janeiro were said to have sailed from Benguela.50 Another strategy that slave traders used to continue the transatlantic traffic was to move slave embarkations away from Benguela to nearby regions. Although not practiced on the same scale as the decentralization of the Luandan trade after 1831, this maneuver helped slave dealers circumvent the British cruisers in the mid-1840s. Slaves were sometimes loaded in regions such as Quikombo and Novo Redondo (a presídío created by the Luandan administration in the late eighteenth century to prevent smuggling), but slave dealers seem to have favored closer places south of Benguela, such as Egito and Baía

Slave Departures from Angola

Farta.51 In 1850 a British commander declared, “I have ascertained the existence of barracoons or shacks at Bahia Farta [six miles to the southwest of Benguela] and at point of Salinas, and I believe these barracoons may contain between 800 and 1,000 slaves each.”52 As late as the mid-1850s, slave dealers would circumvent efforts to suppress the trade by setting up farms where Africans were put to work in “licit” activities such as picking orchella gum. Once ships ventured into the region, these slaves were sold into Atlantic slaving.53 Despite these strategies, TSTD2 demonstrates that slave shipments from Benguela diminished to an average of just six a year between 1843 and 1849. This decline grew out of a Portuguese push to strengthen metropolitan sway over Benguela, which damaged local slaving networks, and particularly intense military operations carried out by British and Portuguese cruisers. The latter included an incursion into the port of Benguela by a British cruiser in 1842.54 More important, it led to the destruction by British and Portuguese cruisers of several entrepôts (barracons) used to hold slaves along the Benguelan coast.55 These measures were combined with a crackdown that led to the imprisonment of various individuals involved in shipping slaves.56 The local slaving network was particularly undermined by Portugal’s push to establish more rigid control over Benguela. Mulattoes and local merchants were those most affected by these measures. José Ferreira Gomes offers a case in point of the damage suffered by slave-trade networks managed by local elites. Gomes was the son of Francisco Ferreira Gomes, who was born in Rio de Janeiro and became a wealthy slave dealer in Benguela in the 1820s. In 1834 Francisco returned to Rio de Janeiro (where he continued to be involved in the slave trade) and assigned José the task of overseeing his business in Benguela. Married to the daughter of a traditional leader (soba) from nearby Catumbela, José had close links with African slave suppliers and was reputed to be one of the two primary merchants in Benguela in 1835. In the late 1840s, the arrest of José Ferreira Gomes and his associates from Benguela and Catumbela dealt a significant blow to the Benguelan trade.57

LUANDA

The case of the Luandan trade illustrates how another combination of external and local forces shaped the slave trade in Angola. The local forces are perhaps best represented by the changes in the supply of slaves to Luanda. Traditionally, Luanda relied on the Mbundu kingdoms of Kasanje and Matamba for slaves, but that changed significantly as internal trading became further

321

322

Roquinaldo Ferreira

fragmented in the nineteenth century. More than ever, pumbeiros (internal merchants) and sertanejos (also internal merchants, but with more resources) became key suppliers of slaves to Luanda and Benguela. In 1846, for example, three subjects of the Matamba kingdom—acting in their capacity as pumbeiros—arrived at Luanda with a number of slaves to sell.58 Meanwhile, the Luandan administration escalated pressures on the Kasanje ruler to comply with demands by Luandan merchants for better access to slave markets.59 While Kasanje and Matamba relied on the east Lunda Empire for slaves, the pumbeiros and sertanejos tapped into sources of slaves in regions that were closer to the coast and under more direct influence of the Luandan government. The decentralization of the internal trade made access to slaves easier than in the past, thus making it particularly difficult to bring an end to the coastal trade. Other local forces included the slave dealers’ hegemonic position in Luandan politics. As in Benguela, the Luandan elite comprised an amalgam of Brazilian and locally born merchants who had little, if any, commitment to Portuguese plans to turn the Angolan economy away from the slave trade. In 1840, for example, one of the key aides to the governor of Angola was José Botelho Sampaio, a Luanda-born official.60 By holding key government positions and exerting pressure on the Luandan administration, slave dealers such as Sampaio successfully undermined efforts to enforce abolitionist legislation. The backlash from 1839 onward, caused by an agreement with the British signed by Governor Antônio Manoel de Noronha, is particularly illuminating. The agreement resulted in the apprehension of several slave vessels and eventually led to Noronha’s resignation.61 Unsurprisingly, a subsequent Angolan governor argued that if he “enforced anti-slave-trade legislation, his closest aides would resign and he would be left with an unleashed unrest to deal with.”62 Furthermore, Luandan politics was also driven by racial tensions, though not as intensely as in Benguela.63 Brazilian nationals intimately involved in the slave trade played a significant role in Luandan politics. Francisco Teixeira Miranda was a case in point. One of the wealthiest slave dealers in Luanda in the 1840s, Miranda provided some of the soldiers and horses used by the Luandan administration to vanquish the Tala Mugongo outpost, thus effectively helping to expand the regions under legal Portuguese influence.64 Another example was Francisco Antônio Flores, who went to Luanda in the 1840s as a representative of a Rio de Janeiro commercial house and became the paramount Brazilian investor in Angola until the early 1860s. Despite being deported from Angola in the

Slave Departures from Angola

1850s, Flores maintained close links with several governors and even British officials serving on the Court of Mixed Commission established in Luanda. In the late 1840s, even while shipments of slaves were subsiding, the number of Brazilians—most of them involved in slaving in northern Angola—applying for licenses to live in Angola far surpassed the number of applications by individuals of any other nationality. In the first semester of 1848, for example, approximately twenty-five Brazilian nationals requested licenses to live in Luanda.65 Close links with Brazil and the advent of international efforts to suppress the slave trade are examples of external forces that affected the shipment of slaves from Angola. Even after serious Portuguese efforts to undercut Angolan ties with Brazil, most of the imports in Luanda still originated in Rio de Janeiro.66 In fact, according to a Portuguese consul in Rio, the volume of exports from Rio to Luanda and Benguela in 1843 was as large as Rio’s exports to Portugal.67 Unsurprisingly, the trade was seriously disrupted when British demands forced Brazil to ban slave imports in 1831. The decline is clearly reflected in TSTD2, which indicates that only five ships left Luanda in 1832, a sharp drop from the number of ships leaving in the previous years. Prices of slaves decreased dramatically, severely affecting Luanda’s economy.68 According to the Luandan government chamber, financial losses caused by the end of slave shipments were “incalculable.”69 Slave dealers responded to the prohibition by relocating to northern Angola, where they had been investing since the British withdrawal from Atlantic slaving in 1807.70 The revival of the Luandan trade paralleled that of slave imports in Brazil in the mid-1830s. In 1835, for example, TSTD2 indicates that fifteen ships left Luanda. Shipments of slaves may have been even greater, since Portuguese authorities indicate that forty-five ships might have left that year.71 In that same year, a British commander reported “thirty-two Portuguese vessels of all types and one Spaniard, trading for slaves.”72 In 1836, although TSTD2 indicates that the number of slave ships sailing reached twenty-seven, an Italian traveler and a newly arrived governor estimated thirty-five.73 In 1838 and 1839, Portuguese assessments put the annual number of ships leaving Luanda at around forty ships.74 TSTD2 indicates that the actual number reached forty-nine. As in Benguela, slave dealers had to develop new methods to conduct the trade in Luanda. An Italian traveler reported that governors charged a fee for each ship leaving Luanda.75 In addition, records of ship sales in Luanda reveal that Brazilian merchants would fictitiously sell their ships to Portuguese nationals to avoid British cruisers by sailing under the Portuguese flag. In 1832,

323

324

Roquinaldo Ferreira

for example, Remígio Luiz dos Santos, a Brazilian merchant based in Angola, sold a ship to Luiz Gomes Ribeiro, a Luanda-born Portuguese and one of the most prominent slave dealers in Luanda.76 Despite the strategy, authorities estimated that most of the ships sailing under the Portuguese flag were owned by Brazilian merchants.77 The practice of using the Portuguese flag became so common that a British commander reported that “although under Portuguese colors, I have every reason to suppose they were Brazilian property, as more than two-thirds of them belonged to ports in the Brazils [sic] and were bound for there.”78 In the early 1840s, as reported by the German traveler George Tams, slave shipments in Luanda were conducted at night, owing to the need to avoid heightened repression.79 In 1841 the number of ships leaving Luanda dropped by 40 percent, with only eighteen vessels clearing out of the harbor. According to TSTD2, shipments rose again during the following two years, but they never returned to the levels reached in the late 1830s. The downward trend was consolidated from 1845 to 1850, when the average number of ships leaving Luanda annually was less than one-tenth the number of the late 1830s, during the illegal trade’s peak. The downfall of the Luandan trade was a function of a variety of anti-slavetrade measures by Portuguese authorities and intensifying naval operations by British and Portuguese cruisers. Records show that, of the criminal cases prosecuted in Luanda between 1845 and 1848, approximately 20 percent were related to the slave trade.80 The crackdown on Luandan slave dealers bore a clear resemblance to what was also taking place in Benguela. Furthermore, special measures were taken to deal with the large number of Brazilian nationals who lived in Luanda while carrying out activities related to the slave trade. For example, Brazilians living in Angola were prohibited from freely circulating and were required to register with the administration.81 Naval patrolling along the Angolan coast grew more intense. The Portuguese squadron established a headquarters in Luanda, and French and U.S. warships joined British and Portuguese cruisers in efforts to control slaving activities. While neither the French nor the Americans had any treaty rights against Brazilian slave traders, the effect was to increase pressures against slave dealers. In 1846 British authorities stated, “The number of vessels engaged or supposed to be engaged in the slave traffic which have been captured or destroyed by these respective [British, French, Portuguese, and American] squadrons so far as our information demonstrates amounts to 61 and is as follows: captured by the British: 37; by the French: 11; by the Portuguese: 5; by the

Slave Departures from Angola

Americans: 2. destroyed by the British: 3; by the French: 1; by the Portuguese: 2.”82 In 1847 the British requested permission for their troops to land in territories under Portuguese control so that they could destroy barraco0ns where slaves were held for the slave trade.83 Between 1847 and 1852, a military ship entered or left the Luandan port every three days, and the number of cruisers far surpassed the number of ships seeking to conduct legal trade with Luanda.84

NORTHERN ANGOLA

In this period, the northern region, stretching from Ambriz to Cabinda, was the focal point of Angola’s illegal slave trade. In addition to benefiting from coastal features that facilitated secrecy and evasion, the dynamics of the trade in that region were distinctly different from those in Benguela and Luanda. By any standards, northern Angola was the most international of the three Angolan regions shipping slaves to the Atlantic. In Cabinda, the British had taken the lead in developing the trade, and the French were quick to join. David Eltis has pointed out that the Royal African Company rejected plans to establish a permanent outpost in Kongo in the 1680s. Between 1681 and 1720, however, as reflected in TSTD2, ten British ships were conducting business in Cabinda, as compared with only eight French ships. From the mid-seventeenth century onward, growing Dutch trade in northern Angola took a toll on Portuguese influence in the area. Despite their dominance, the Dutch allowed British and French ships to pick up slaves in Cabinda, dealing a further blow to the waning Portuguese influence. In addition to pushing prices upward, the French and British aggressively drove many Portuguese ships away.85 Writing in the early nineteenth century, Angolan Governor Miguel Antônio de Melo dated the beginning of the British and French slave trade to the 1760s, but by then they had been participating in the trade for some time.86 Although the British trade suffered a sharp decline after a military setback in 1721, when they sought to build a fortress in Cabinda, it soon resumed its expansion. From the 1720s to 1750s, as indicated by TSTD2, while forty-six British ships traded in Cabinda, forty-two French ships also embarked slaves in the region. In the second half of the eighteenth century, at least until 1792, French domination in Cabinda was almost incontestable; TSTD2 demonstrates that four times more French than English ships embarked slaves there. In the second half of the eighteenth century, only three Portuguese ships participated in slaving activities in northern Angola—after several failed

325

326

Roquinaldo Ferreira

attempts to break the British and French grip in Cabinda. In the 1760s Luandan authorities sought to provide incentives for the city’s merchants to engage in trade in the north.87 In Rio, authorities also sought to convince merchants that trade in northern Angola was worthwhile.88 In Luanda, however, the return of six ships that had not been able to purchase slaves in Cabinda made it clear that the foreign competition was not easy to overcome.89 In fact, Portuguese trade between Cabinda and Luanda was strengthened only after diplomatic overtures by the Luandan government to African rulers in the nineteenth century.90 In 1803, for example, approximately 1,000 Muxicongo slaves from Ambriz were sold in Luanda in a single month.91 The volume of the trade with northern Angola might even have increased when Mdembo Muene Damba Mangombe sent a large, but not specified, number of slaves to Luandan merchants in 1804. Authorities then sent him gifts and assured him that they would seek to guarantee favorable prices for slaves if he continued shipping slaves to Luanda.92 The British withdrawal from slaving in 1807 profoundly altered the dynamics of the northern Angolan trade by creating a vacuum that was quickly filled by Luandan and Brazilian merchants. TSTD2 shows that ships sailing under the Portuguese flag—most of them actually outfitted in Brazil—moved into Cabinda immediately after Britain’s ending of its slave trade in 1807. Although some funding for the trade came from Luanda, the center of gravity in the trade between Angola and Brazil shifted to Brazil. Brazilian participation soon led to unprecedented levels of trade. By 1814 ships would still stop in Luanda on their way to Cabinda, but they soon ceased doing so and the Brazilian-dominated northern Angola trade developed into an independent operation.93 Between 1808 and 1822, all but one of the approximately 255 ships leaving the area sailed under the Portuguese flag. Although Muxicongo Africans from northern Angola would still be sent to Luanda via land, they were by all standards a minority, and most of the Africans enslaved in northern Angola were shipped from loading points in the region itself.94 A more efficient route for slaves, compared with the Luandan route, lay at the heart of the growth of slaving in Cabinda. While the trade in Luanda was hindered by excessive reliance on the kingdom of Kasanje, the Cabindan trading networks were highly decentralized and reached far into Central Africa. In addition, while ships from Luanda had little chance of avoiding taxes on slaves being shipped to Brazil, those leaving Cabinda had a much better one of doing so.95 After independence from Portugal in 1822, Brazilian domination increased further, with 185 Brazilian ships embarking

Slave Departures from Angola

slaves in Cabinda from 1823 to 1831. The first ship under a Brazilian flag recorded by TSTD2 arrived in Cabinda in 1826, and it is certain that the subsequent trade in that region was controlled by merchants from Rio de Janeiro. Around the Loje River, just north of Luanda, the dynamic was not significantly different. The British first developed the trade, and the region became known by the name they used to refer to it: Ambriz. According to TSTD2, however, although a British ship was the first to embark slaves in Ambriz in 1786, the British lagged well behind the French until the early 1790s—when war and the Saint-Domingue rebellion dealt a huge blow to the French slave trade. From 1786 to 1792, there were eleven French ships, while the number of British ships was only seven. Soon thereafter, British ships dominated the trade until the British withdrew from slaving in 1807. Between 1793 and 1807, twenty-seven British ships picked up slaves in Ambriz, but the number may actually have been higher. In 1793, for example, TSTD2 documents only two slave voyages by British ships, yet authorities in Luanda noted that eleven British ships were slaving in Ambriz.96 In contrast with Cabinda, where TSTD2 shows that shipments of slaves continued seamlessly in the aftermath of the 1807 British abolition act, the trade in Ambriz seems to have been significantly disrupted by the end of the British trade. Slave embarkations in Ambriz only resumed in 1815, with nine ships recorded from 1815 to 1822, the year of Brazilian independence. Between 1823 and 1832, however, the number rose dramatically, with ninety-one ships embarking slaves in Ambriz. Brazilian investors were responsible for this rise of shipments from Ambriz, as acknowledged by the Luandan administration in 1826.97 In contrast with the situation in Cabinda, however, several nations were involved in Ambriz. TSTD2 shows that the Spanish were operating in Ambriz by 1809, and they seem to have intensified their trade in that region by the early 1830s. A British commander calculated that 4,000 slaves were boarded on Spanish ships annually.98 The fact that the French, the Americans, and the British all owned factories in Ambriz in 1845 provides further evidence of the international nature of the Ambriz trade.99 Both Cabinda and Ambriz offered extremely propitious conditions for the slave trade.100 In the 1840s, for example, several barracoons to hold enslaved Africans existed five miles inland from the coast of Ambriz.101 According to Portuguese and British authorities, “there were at that time several hundred slaves in barracoons near Ambriz ready for shipment.”102 Furthermore, ships operating out of Ambriz and Cabinda could count on a local network of

327

328

Roquinaldo Ferreira

seasoned Cabindan sailors to help with the trade. Cabindans were particularly involved in the trade, either through slave trading or by controlling coastal trade operations in northern Angola.103 Powerful Cabindan families maintained tight links not only with Luanda but also with Brazil, where they often sent their children to be educated.104

CONCLUSION

Although shipments of slaves in nineteenth-century Angola were directly affected by external forces such as the international efforts to suppress the slave trade and the close economic links with Brazil, local politics and the supply of slaves from the hinterlands were equally key factors in the continuation of the trade during illegality. The partial abolition of the slave trade affected first northern Angola, and only much later took a toll on the trade in Luanda and Benguela. In Luanda and Benguela, Brazilian and local investors held a tight grip over slave shipments, in addition to dominating local politics and crippling early attempts to ban the trade. It is not surprising, therefore, that slave trading continued until the mid-1840s. The geography of the northern Angolan coast and the Congo River protected a daunting number of potential slave embarkation points, and there was no local Portuguese administration present to either help or hinder shipments of slaves. The region thus provided a safe haven for slave dealers as pressure on the traffic built up in Luanda and Benguela. NOTES

1. David Eltis, “African and European Relations in the Last Century of the Transatlantic Slave Trade,” in Olivier Pétré-Grenouilleau, ed., From Slave Trade to Empire: Europe and the Colonisation of Black Africa, 1780s–1880s (London, 2004), 24. David Eltis and David Richardson, “Prices of African Slaves Newly Arrived in the Americas, 1673–1865: New Evidence on Long-Run Trends and Regional Differentials,” in David Eltis, Frank Lewis, and Kenneth Sokoloff, eds., Slavery and the Development of the Americas (Cambridge, 2004), 187. 2. Eltis, “African and European Relations,” 24; Eltis and Richardson, “Prices of African Slaves Newly Arrived in the Americas,” 187. 3. Eltis, “African and European Relations,” 21–46; idem, The Rise of African Slavery in the Americas (Cambridge, 2000). 4. Roquinaldo Ferreira, “Dos sertões ao Atlântico: Tráfico ilegal de escravos e comércio lícito em Angola, 1830–1860” (MA thesis, Universidade Federal do Rio de Janeiro, 1996), chap. 5.

Slave Departures from Angola

5. Leslie Bethell, A abolição do tráfico de escravos no Brasil: A Grã-Bretanha, o Brasil e a questão do tráfico de escravos, 1807–1869 (São Paulo, 1976); Luis Henrique Dias Tavares, Comércio proibido de escravos (São Paulo, 1988); Robert Conrad, Tumbeiros: O tráfico de escravos para o Brasil (São Paulo, 1985); Jaime Rodrigues, O infame comércio: Propostas e experiências no final do tráfico de africanos para o Brasil (1800–1850) (Campinas, 2000); Jeffrey Needell, “The Abolition of the Brazilian Slave Trade in 1850: Historiography, Slave Agency and Statesmanship,” Journal of Latin American Studies 33 (2001): 681–711. 6. Francisco Vidal Luna and Herbert Klein, “African Slavery in the Production of Subsistence Crops: The Case of Minas Gerais,” in Eltis, Lewis, and Sokoloff, Slavery and the Development of the Americas, 131. 7. Zephyr Frank, Dutra’s World: Wealth and Family in Nineteenth-Century Rio de Janeiro (Albuquerque, 2004), 20–21. 8. Ferreira, “Dos sertões ao Atlântico,” 2. 9. Ibid., 3. 10. Ibid., chap. 4. 11. Joseph C. Miller, Way of Death: Merchant Capitalism and the Angolan Slave Trade, 1730–1830 (Madison, Wis., 1988); José Curto, Enslaving Spirits: The Portuguese-Brazilian Alcohol Trade at Luanda and Its Hinterland, c. 1550–1830 (Leiden, 2004); Jaime Rodrigues, De costa a costa: Escravos, marinheiros e intermediários do tráfico negreiro de Angola ao Rio de Janeiro (1780–1860) (São Paulo, 2005). 12. John Thornton, Africa and Africans in the Making of the Atlantic World, 1400–1800 (Cambridge, 1998); Herbert Klein, “The Atlantic Slave Trade to 1650,” in Stuart Schwartz, ed., Tropical Babylons: Sugar and the Making of the Atlantic World, 1450–1680 (Chapel Hill, 2004). 13. Joseph C. Miller, “The Slave Trade in Congo and Angola,” in Martin L. Kilson and Robert I. Rotberg, eds., The African Diaspora: Interpretive Essays (Cambridge, Mass., 1976); Joseph C. Miller, “Central Africa during the Era of the Slave Trade, c. 1490s–1850s,” in Linda Heywood, ed., Central Africans and Cultural Transformations in the American Diaspora (Cambridge, 2002). 14. Roquinaldo Ferreira, “Transforming Atlantic Slaving: Trade, Warfare, and Territorial Control in Angola, 1650–1800” (PhD diss., University of California, Los Angeles, 2003), 72. 15. Ibid., 84. 16. Eltis and Richardson, “Prices of African Slaves Newly Arrived in the Americas,” 210. 17. Ferreira, “Transforming Atlantic Slaving,” chap. 1. 18. Ibid., 97–101. 19. Ibid., 138. 20. Eltis, “African and European Relations,” 25. 21. Ferreira, “Transforming Atlantic Slaving,” 138–42. 22. “Carta de Manoel da Cruz,” October, 20, 1837, BNL, códice 600. 23. “Ordem do Governador de Benguela,” June 12, 1836, AHNA, códice 521, ff. 34–34v. 24. “Ofício do Governador de Angola,” July 20, 1841, AHNA, códice 15, ff. 60–60v. 25. “Ofício do Governador de Benguela,” March 26, 1844, AHNA, códice 455, f. 51.

329

330

Roquinaldo Ferreira

26. In 1846, three years after leaving Benguela on the grounds of seeking medical care in Brazil, Teixeira was still in Rio. See “Ofício do Secretário de Governo de Benguela,” July 24, 1846, AHNA, códice 461. For Teixeira’s return to Benguela, see “Ofício do Governador de Benguela,” December 5, 1853, AHNA, códice 459, ff. 22v–25v. 27. “Portaria do Ministério da Marinha e Ultramar,” September 27, 1850, AHNA, códice 263, f. 177. See “Draft of Dispatch to the British Consul in Lisbon,” October 30, 1850, BNA, FO 84, 799, ff. 33–33v. The British successfully pushed for the dismissal of the official. See “Draft of Dispatch to the British Consul in Lisbon,” December 14, 1850, BNA, FO 84, 799, ff. 77–77v. 28. “Carta para o Juiz do Crime Interino e Relator da Alçada,” August 11, 1835, AHNA, códice 163, ff. 10–10v. 29. “Portaria da Secretária de Governo da Marinha e Ultramar,” September 6, 1843, AHNA, códice 261, ff. 38v–39. 30. “Ofício do Governador de Benguela,” September 9, 1850, AHNA, códice 459, ff. 16v–17. 31. “Ofício de Antônio Manoel de Noronha,” April 7, 1827, AHNA, códice 7183, f. 58v. 32. “Ofício do Governador de Angola,” March 31, 1828, AHU, Angola, caixa 158, doc. 58. 33. “Ofício do Governador de Angola,” December 14, 1830, AHNA, códice 12, ff. 51v–55. 34. “Ofício de Sá da Bandeira,” 1837, AHU, papéis de Sá da Bandeira, maço 824. 35. George Tams, Visita às possessões portuguesas na costa ocidental da Africa (Porto, 1850), 110. 36. “Bando do Governador de Benguela,” May 11, 1839, AHNA, códice 522, ff. 7v. 37. At least one ship was apprehended in 1840. See “Ofício do Governador de Angola,’ April 22, 1840, AHNA, códice 15, f. 1. 38. “Carta do comandante do Navio Urania,” October 10, 1839, BNA, FO 84, 322, ff. 135–136v. 39. Documentation indicating that thirty ships left Benguela with slaves in 1841 seems to be a clear exaggeration. See “Apontamento sobre Angola,” May 27, 1841, AHU, papéis de Sá da Bandeira, maço 827. 40. José Joaquim Lopes de Lima, Ensaios sobre a statistica das possessões portuguezas na Africa occidental e oriental: Na Asia occidental, na China, e na Oceania, 5 vols. (Lisbon, 1844–62), 1:133; Almanak statistico da província d’Angola e suas dependências para o ano de 1852 (Luanda, 1851), xxi; Maria do Rosário Pimentel, Viagem ao fundo das consciências: A escravatura na época moderna (Lisbon, 1995), 344. See also Ferreira, “Dos sertões ao Atlântico,” chap. 1. 41. The British did apprehend two ships loaded with slaves in 1841 and 1842, but that did little to rein in slave dealers. See “Portaria do Governador de Benguela,” July 1, 1841, AHNA, códice 522, f. 114; “Edital do Governador Interino de Benguela,” May 25, 1843, AHNA, códice 522, ff. 172v–174. 42. “Portaria da Secretária de Estado da Marinha do Ultramar,” May 26, 1841, AHNA, códice 260, ff. 53v.–54v. 43. “Apontamento sobre Angola,” May 27, 1841, AHU, papéis de Sá da Bandeira, maço 827; “Apontamento de Pedro Alexandrino da Cunha,” August 6, 1841, AHU, papéis de Sá da Bandeira, maço 827. 44. “Portaria do Governador de Benguela,” July 7, 1840, AHNA, códice 522, ff. 63–63v.

Slave Departures from Angola

45. “Decreto Real,” March 2, 1841, AHNA, códice 260, ff. 17v–18; “Ofício do Governador de Angola,” June 12, 1841, AHNA, códice 15, ff. 55–55v. 46. “Portaria da Secretária de Governo da Marinha e Ultramar,” May 20, 1843, AHNA, códice 261, ff. 12–12v. 47. “Relatório do Juiz de Direito da Comarca de Luanda,” December 18, 1849, AHU, segunda seção de Angola, pasta 16. 48. AHNA, códice 521. 49. AHU, Coleções Especiais, lata 25, maço 4, pasta 1, Paquete do Sul, 1833–34. 50. “Ofício de Sá da Bandeira,” 1837, AHU, papéis de Sá da Bandeira, maço 824. 51. “Ofício para o Comandante do Brigue Audaz,” August 4, 1842, AHNA, códice 452, f. 165v; “Ofício do Governador de Benguela,” October 26, 1846, AHNA, códice 7183, ff. 109v–110v; “Ordens e Instruções,” December 24, 1846, AHNA, códice 1178, ff. 1–1v.; “Ofício do Governador de Benguela,” November 27, 1848, AHNA, códice 459, ff. 2v–3v. See also E. Bouët-Willaumez, Commerce et traite des noires aux côtes occidentales d’Afrique (Paris, 1848), 177–78; Tams, Visita às possessões portuguesas, 110. 52. “Carta de Hastings,” March 4, 1850, AHU, segunda seção de Angola, pasta 16A. 53. “Circular aos Negociantes de Benguela,” March 19, 1850, AHNA, códice 461, ff. 22v–23v.; “Ofício do Secretário do Governo de Angola,” April 24, 1850, AHNA, códice 511, ff. 218–218v; “Dispatch by Gabriel and Jackson,” February 20, 1854, BNA, FO 84, 932, ff. 42–43v; “Ofício do Governador de Benguela,” March 11, 1854, AHNA, códice 467, ff. 153–55; “Ofício da Estação Naval,” April 17, 1854, Arquivo Geral da Marinha (hereafter AGM), caixa 319. 54. “Edital do Governador Interino de Benguela,” May 25, 1843, AHNA, códice 522, ff. 172 v–174; “Ofício do Secretário de Governo de Benguela,” October 18, 1843, AHNA, códice 454, ff. 76–76v. 55. “Ofício do Governador de Angola,” February 17, 1846, AHU, segunda seção de Angola, pasta 10A; “Ofício do Governador de Benguela,” October 26, 1846, AHNA, códice 7183, ff. 109v–110v; “Edital do Governo de Benguela,” October 31, 1846, AHNA, códice 523, ff. 16–16v. 56. “Registros da Cadeia Pública de Benguela,” March 12, 1846, AHNA, caixa 145. 57. “Ofício do Governador de Benguela,” November 27, 1846, AHNA, códice 455, ff. 251v; “Ofício do Governador de Benguela,” December 25, 1848, AHNA, caixa 5426. 58. “Ofício do Secretário do Governo de Angola,” August 5, 1846, AHNA, códice 105, ff. 87–87v. 59. “Portaria do Governador de Angola,” March 2, 1838, AHNA, códice 101, f. 19v. 60. “Mapa Demonstrativo dos Oficiais do Estado Maior da Província em junho de 1840,” AHU, segunda seção de Angola, pasta 3C. 61. “Representação da Câmara Municipal de Luanda,” October 9, 1839, AHU, segunda seção de Angola, pasta 5; “Representação da Câmara de Luanda,” December 2, 1839, AHU, segunda seção de Angola, pasta 5; “Protesto de José da Lomba, Capitão do Patacho Portuguêse Treze de Junho,” June 30, 1840, BNA, FO 84, 322; “Representação do Tenente Ananias Xavier,” August 7, 1840, AHU, segunda seção de Angola, pasta 4B; “Portaria do Governador de Benguela,” July 1, 1841, AHNA, códice 522, f. 114; “Ofício do Conselho de Governo de Angola,” September 4, 1842, AHNA, códice 15, ff. 157v–158;

331

332

Roquinaldo Ferreira

“Ofício do Secretário de Governo,” May 20, 1843, AHNA, códice 102, f. 98; “Edital do Governador Interino de Benguela,” May 25, 1843, AHNA, códice 522, ff. 172v–174; “Ofício do Secretário de Governo,” October 18, 1843, AHNA, códice 454, ff. 76–76v. See also Anne Stamm, “L’Angola à un tournant de son histoire, 1838–1848” (PhD diss., Université de Paris I, Panthéon-Sorbonne, 1971), 159. 62. “Ofício do Governador de Angola,” May 17, 1843, AHU, segunda seção de Angola, pasta 6A. 63. “Ofício do Governador de Angola,” December 6, 1837, AHU, segunda seção de Angola, pasta 2; “Ofício do Governador de Angola,” June 5, 1840, AHU, segunda seção de Angola, pasta 3C. 64. “Ofício do Governador de Angola,” January 19, 1841, AHNA, códice 15, ff. 36–36v. Miranda had been living in Luanda for seven years and dealing in slaves on behalf of a commercial house in Rio de Janeiro. See “Ofício confidencial do Juiz de Direito de Luanda,” January 20, 1842, AHU, segunda seção de Angola, pastas 5A and 5B. 65. AHNA, códices 167, 168, and 171. 66. “Ofício do Governador de Angola,” April 8, 1836, AHNA, códice 13, ff. 1–3. 67. “Ofício do Vice-Cônsul Português no Rio de Janeiro,” April 24, 1843, AHNA, caixa 1479. 68. “Ofício do Governador de Angola,” April 8, 1836, AHNA, códice 13, ff. 10v–12v. 69. “Representação da Câmara de Luanda,” October 5, 1833, AHU, caixa 177, doc. 15. 70. As Trevor Getz, Robin Law, and Silke Strickrodt demonstrate, the relationship between heightened abolitionism and the decentralization of slave embarkations also took place elsewhere in Africa. See Trevor Getz, Slavery and Reform in West Africa: Toward Emancipation in Nineteenth-Century Senegal and the Gold Coast (Athens, Ohio, 2004), 38; Robin Law, Ouidah: The Social History of a West African Slaving “Port,” 1727–1892 (Athens, Ohio, 2004), 157–58; Silke Strickrodt, “Afro-European Trade Relations on the Western Slave Coast, 16th to 19th Centuries” (PhD diss., University of Stirling, 2002), 238. 71. “Carta de Luanda,” May 24, 1835, AHU, papéis de Sá da Bandeira, maço 2922. 72. “Relatório do Comandante do Brigue Charybolis,” September 6, 1835, AGM, caixa 311. 73. Tito Omboni, Viaggi nell’Africa occidentale: Gia medico di consiglio nel regno d’Angola e sue dipendenze (Milan, 1845), 83; “Ofício do Governador de Angola,” April 8, 1836, AHNA, códice 13, ff. 5–10v. 74. “Apontamento do Tenente Lima,” 1838, AHU, papéis de Sá da Bandeira, maço 827; Douglas L. Wheeler, “The Portuguese in Angola, 1836–1891: A Study in Expansion and Administration” (PhD diss., Boston University, 1963), 80. 75. Omboni, Viaggi nell’Africa occidentale, 107; “Ofício de Feliciano António Pereira,” June 18, 1838, AHU, papéis de Sá da Bandeira, maço 824. 76. Biblioteca Municipal de Luanda, hereafter BML, códice 37, f. 124v. 77. “Ofício do Cônsul Português no Rio de Janeiro,” June 16, 1834, AHNA, Luanda, avulsos, caixa 1465. 78. “Relatório do Comandante do Brigue Charybolis,” September 6, 1835, AGM, caixa 311. 79. Tams, Visita às possessões portuguesas, 212. 80. “Boletim oficial do Governo Geral da Província de Angola,” October 3, 1846; “Boletim oficial do Governo Geral da Província de Angola,” July 31, 1847; “Boletim oficial do

Slave Departures from Angola

Governo Geral da Província de Angola,” January 1, 1848; “Boletim oficial do Governo Geral da Província de Angola,” October 1, 1848. See also Ferreira, “Dos sertões ao Atlântico,” 20. 81. “Ofício do Secretário de Governo de Angola,” November 14, 1845, AHNA, códice 104, f. 65; “Ofício do Secretário de Governo,” January 17, 1846, AHNA, códice 104, ff. 132–132v. 82. Edmond Gabriel and George Jackson, “Report on the Slave Trade,” February 18, 1847, BNA, FO 84, 671, ff. 99–110. 83. “Dispatch by the British Consul in Lisbon,” May 17, 1847, BNA, FO 84, 676, ff. 30–33; “Ofício do Ministério da Marinha e Ultramar,” May 22, 1847, AGM, caixa 311. 84. Ferreira, “Dos sertões ao Atlântico,” 22. 85. Ferreira, “Transforming Atlantic Slaving,” 86. 86. “Carta do Governador de Angola,” January 3, 1801, AHU, Angola, caixa 98, doc. 6. 87. “Ofício do Governador de Angola,” November 25, 1768, AHNA, códice 3, ff. 261–261v. 88. “Ofício do Luiz de Vasconcellos e Silva,” June 30, 1783, AHU, Rio de Janeiro, caixa 131, doc. 85. 89. “Ofício do Governador de Angola,” December 31, 1794, AHNA, códice 246, ff. 53–55. 90. António Carreira’s records of Muxicongo Africans in Pará in the 1760s suggest that sporadic shipments of slaves via land from northern Angola to Luanda occurred in the second half of the eighteenth century. See António Carreira, As companhias Pombalinas de navegação, comércio e tráfico de escravos entre a costa africana e o nordeste brasileiro (Porto, 1969), 274. 91. This occasional trade was via land and failed to evolve into long-term commercial relations. See “Ofício do Governador de Angola,” July 30, 1803, AHNA, códice 9, ff. 37v–38. 92. See “Carta do Governador de Angola,” August 14, 1804, AHNA, códice 240, ff. 29–29v. 93. “Portaria do Governo de Angola,” September 6, 1814, AHNA, códice 277, ff. 148–148v. 94. “Registro de Ofício do Governador de Angola,” June 18, 1825, BML, códice 44, ff. 99v–100v. 95. Indeed, seemingly aware of looming fiscal losses, the Luandan administration unsuccessfully attempted to establish a fixed location for trade transactions, so that it could levy taxes on the business. See “Carta do Governador de Angola,” November 14, 1819, AHNA, códice 240, f. 128; “Ofício do Governador de Angola,” September 4, 1824, AHU, Angola, caixa 145. 96. AHU, códice 1633. The arrival of enslaved Africans in the Americas provides further evidence of what seems to have been a more intense British trade. While TSTD2 indicates that 1,596 slaves from Ambriz arrived in the Americas in 1799, Luandan authorities claimed that British ships carried off 6,000 slaves that year. See “Ofício para o Governador de Angola,” March 8, 1800, AHNA, códice 254, ff. 144–144v. 97. “Relatório do Governador de Angola,” March 5, 1826, AHU, Angola, caixa 151. 98. “Relatório do Comandante do Brigue Charybolis,” September 6, 1835, AGM, caixa 311. 99. “Ofício de Joaquim Xavier Bressane Leite,” August 31, 1845, AHNA, códice 16, ff. 125v–126. 100. Ferreira, “Dos sertões ao Atlântico,” chap. 4; Susan Herlin, “Brazil and the Commercialization of Kongo, 1840–1870,” in José Curto and Paul Lovejoy, eds., Enslaving

333

334

Roquinaldo Ferreira

Connections: Changing Cultures of Africa and Brazil during the Era of Slavery (New York, 2004), 261–85. 101. “Portaria da Secretária de Governo da Marinha e Ultramar,” April 27, 1844, AHNA, códice 261, ff. 145–145v. 102. “Draft of Letter by Lord Palmerston,” November 11, 1846, BNA, FO 84, 630, ff. 99–101. See also “Ofício do Governador de Angola,” February 17, 1846, AHU, segunda seção de Angola, pasta 10A. 103. “Dispatch by the British Representative in the Mixed Commission in Luanda,” March 30, 1844, FO 84, 517, ff. 40–41; Phyllis Martin, “The Cabinda Connection: An Historical Perspective,” African Affairs 76 (1977): 47–59; idem, “Family Strategies in Nineteenth-Century Cabinda,” Journal of African History 28 (1987): 65–86; idem, “Cabinda and Cabindans: Some Aspects of an African Maritime Society,” in Jeffrey Stone, ed., Africa and the Sea: Proceedings of a Colloquium at the University of Aberdeen, March 1984 (Aberdeen, 1985), 80–96. 104. “Documento de 23 de agosto de 1840,” AHU, papéis de Sá da Bandeira, maço 827; Anon., Quarenta e cinco dias em Angola: Apontamentos de viagem (Porto, 1862), 10. See also Ferreira, “Dos sertões ao Atlântico,” 88.

Chapter 12 The Demographic Decline of

Caribbean Slave Populations: New Evidence from the Transatlantic and Intra-American Slave Trades David Eltis and Paul Lachance

The inability of slave populations to sustain their numbers without constant replenishment from outside sources is one of the most widely accepted ideas among scholars of slavery, as well as being of great underlying significance. Parts of the North American mainland after 1720, and the whole of mainland North America by 1800, are often cited as having the only societies in the history of slavery where the enslaved maintained their numbers without augmentation from new coerced immigrants. Slave registration data in the British Caribbean and more problematic census information from mid-nineteenth-century Cuba have enabled scholars to add Barbados, Antigua, and Cuban slave populations late in the slavery era to this short list, as well as to hypothesize a broad trend over time for the whole era of African slavery in the Americas from strongly negative to mildly positive natural population growth. The significance of this hypothesis can hardly be overstated. “Depletion rates,” as they have been termed, shape our view of the viability of slavery as an institution and put the traffic that allowed slave societies to maintain their numbers into central focus. Indeed, estimates of the size of the slave trade itself have often hinged 335

336

David Eltis and Paul Lachance

on calculations of the natural increase of slave populations. More important, the date at which Creoles—that is, in the case of the Atlantic world, people born in the Americas rather than in Africa—became dominant has a rather large cultural significance. New data on the arrival of slaves in the Americas together with new information on what has come to be called the intra-American slave trade provide an opportunity to reassess this critical issue. The natural rate of Caribbean slaves’ population growth has been a continual subject of debate from the abolitionist era down to the present, and no doubt was a topic of private discussion among slave owners. Two methods have dominated attempts to elicit the facts—comparison of probable volumes of the slave trade with population counts for various islands on one hand, and direct calculations from the records of the estates on the other. Surviving records of individual Caribbean plantations are rare before 1750. Indeed, the only major example of documents that allow such compilation is for the Codrington plantations of Barbados after 1712.1 Population counts and estimates of slave arrivals from Africa, by contrast, are available for all but the earliest years of the Caribbean plantation complex. Much of the discussion, especially comparisons of slave inflows with slave populations, has been dominated by data from the English Caribbean. This last fact, as we shall emphasize, has led to some dubious inferences. The picture that emerges from these two approaches is broadly consistent. Rates of natural population change of −5 percent to −6 percent per year in the late seventeenth century gradually improved to about −1 percent in the early nineteenth century, but modern scholars have still been able to argue for Barbados, for example, that in the half century from 1712 to 1762 it took the arrival of 150,000 slaves to increase the population of the island by 28,000.2 In the later period, the range of demographic regimes was still wide. If the Barbados and Antigua slave populations were already experiencing positive rates of growth by the late eighteenth century, Tobago was still exceeding −3 percent in the early 1800s, and Grenada, Trinidad, and Demerara-Essequibo also had strongly negative rates of natural increase at this time.3 The broad explanation for the pattern is a variation of one that demographers have commonly used to account for change in immigrant-dominated societies. Migrants tend to be male and adult, as well as to experience higher mortality rates when moving long distances. Population pyramids in newly settled societies are therefore skewed and top-heavy, with relatively few women and children and lower life expectancies. Thus, quite apart from the exigencies of slavery, fertility will be lower and mortality higher than in established

The Demographic Decline of Caribbean Slave Populations

societies where migrants comprise a small share of the population, age–sex structures approach normality, and rates of natural population growth are positive. In Caribbean slave societies the emergence of such steady-state or growth conditions was severely delayed by a combination of the rigors of sugar production to which the slave labor force was subjected; the unhealthy, disease-ridden environments in which many sugar plantations were located; and the massive, three-century-long expansion of the sugar sector, with its huge labor demands that continued long after the ending of the slave trade from Africa. The best data for analysis of slave demography are the slave registration records for the British West Indies generated between 1813 and 1834, which have been exhaustively studied by Barry Higman.4 John Ward has assembled a set of plantation records to extend the period for which there are reliable data back for a few decades into the later eighteenth century.5 Gabriel Debien has examined mortality statistics drawn from reports of the managers of eight plantations in Saint-Domingue.6 But plantation records are much more numerous for the later eighteenth and the nineteenth centuries than for pre1760. For the seventeenth and most of the eighteenth century, estimates of depletion rates of Caribbean slave populations depend heavily on population counts and estimates of slave imports. These have been most comprehensively analyzed in Philip Curtin’s census of the Atlantic slave trade.7 Curtin used censuses of Barbados and Jamaica and reports on the number of slaves imported into the two colonies to calculate annual rates of growth and immigration between the censuses. Applying the standard formula for components of population change, r=i+m he derived rates of natural increase (i )—or natural decrease, in these two cases—by subtracting migration rates (m), calculated by dividing average annual imports over a period by the mid-interval population, from growth rates (r).8 These rates of natural decrease were applied in turn as “precedents” to other plantation societies at comparable stages of development. Subtraction of projected rates of natural increase from annual growth rates based on population counts provided estimates of migration rates and the corresponding number of migrants where records of imports were lacking or incomplete.9 Figure 12.1, drawn from John McCusker’s rounded population estimates at ten-year intervals, compares the growth of the black populations in the British and French Caribbean and in British mainland North America from

337

338

David Eltis and Paul Lachance

1650 to 1780. The second and third regions caught up with the first by 1760. In 1780 the black population was 575,000 in the British mainland colonies, soon to obtain recognition of independence, 489,000 in the British Caribbean, and 438,000 in the French Caribbean.10 Combining McCusker’s statistics with Curtin’s estimates of slave imports over twenty-five- and twenty-year periods in the seventeenth and eighteenth centuries, table 12.1 summarizes the components of population change for blacks in these three jurisdictions. The series for the British mainland begins in 1700 because Curtin does not provide statistics for the small number of slaves imported into this area in the seventeenth century. Natural increase there from the beginning of the eighteenth century contrasts with rates of natural decrease in the two Caribbean areas. Although the latter lessen over time, they do not yet approach natural increase by the end of the period under observation. The precocious natural increase of blacks in the thirteen colonies on the Atlantic seaboard is as often remarked by historians today as their natural decrease in the Caribbean up to

Figure 12.1 Black Population Growth in British and French Jurisdictions, 1650–1780 Source: John J. McCusker, “The Rum Trade and the Balance of Payments of the Thirteen Continental Colonies, 1650–1775” (PhD diss., University of Pittsburgh, 1970), tables B-26 and B-104, pp. 584 and 712.

The Demographic Decline of Caribbean Slave Populations

339

the abolition of slavery, but that was not yet the case in 1969 when Curtin’s study was published.11 The difference between rates of migration and depletion in the French and British Caribbean in table 12.1 is surprising in view of Curtin’s use of precedents of natural decrease in Barbados and Jamaica to predict vital rates in French colonies. Until 1760 the rate of growth of the slave population was higher in the French Caribbean. Its migration rate was consistently higher than the British, more than compensating for larger rates of depletion in every period except 1721–40. Migration rates and rates of natural decrease vary in tandem. The higher the rate of migration, the greater was the excess of deaths over births. This can be considered indicative of the disparity between African and Creole vital rates.12 Nevertheless, in both jurisdictions, growth rates and migration rates fell as the slave population grew, and rates of natural decrease converged after 1720, ranging from −3.3 percent to −2.3 percent per year over the next six decades. While the population figures used to calculate the annual rates of population growth in table 12.1 can be accepted as reasonably accurate (regional summaries and trends over time rectify aberrant counts in particular censuses), the estimates of slave imports used to calculate migration rates are more problematic. Curtin’s assumptions in estimating the number of slaves arriving in the Caribbean in the “long” eighteenth century (1700–1810)—the period when well over half of all slaves carried across the Atlantic made their journeys—led

Table 12.1 Components of Population Change in the British Caribbean, French Caribbean, and British Mainland, 1650–1780 British Caribbean Years 1650–75 1676–1700 1701–20 1721–40 1741–60 1761–80

French Caribbean

British Mainland

r

m

i

r

m

i

r

m

i

5.9% 2.4% 2.1% 1.8% 1.6% 1.8%

8.5% 8.1% 5.7% 4.8% 4.6% 4.1%

−2.6% −5.7% −3.5% −3.0% −3.0% −2.3%

6.2% 4.5% 4.1% 3.3% 2.0% 1.3%

12.8% 15.5% 10.3% 5.8% 5.4% 4.4%

−6.6% −11.0% −6.2% −2.5% −3.3% −3.1%

— — 4.6% 4.0% 4.0% 2.9%

— — 2.2% 2.5% 2.3% 1.0%

— — 2.4% 1.5% 1.7% 1.9%

Sources: Population estimates in John J. McCusker, “The Rum Trade and the Balance of Payments of the Thirteen Continental Colonies, 1650–1775” (PhD diss., University of Pittsburgh, 1970), tables B-26 and B-104, pp. 584 and 712; estimates of slave imports in Philip D. Curtin, The Atlantic Slave Trade: A Census (Madison, Wis., 1969), tables 34 and 65, pp. 119 and 216. Note: r = annual growth rate; m = annual migration rate; i = annual rate of natural decrease/increase.

340

David Eltis and Paul Lachance

to a major discrepancy when he turned to calculating number of departures from Africa. From shipping records, Curtin estimated separately 5,461,400 slaves exported from Africa by the three major carriers between 1701 and 1810: 2,466,800 by the British, 969,100 by the French, and 2,025,500 by the Portuguese.13 Most of the Portuguese slaves were shipped to Brazil, where estimated imports of 1,891,400 slaves correspond approximately with the number who would have survived the Middle Passage. Britain, France, and various minor carriers supplied the Caribbean, Spanish America, and British North America. Allowing for loss in transit of 15 percent of the slaves, British ships transported an estimated 2,108,000 slaves to the non-Brazilian New World in the eighteenth century. Since only 1,704,000 were imported into their own colonies, 404,000 were left over for reexport into other jurisdictions. The French, on the other hand, with the same losses in transit of 15 percent of exports from Africa, supplied only 838,000 of the estimated 1,348,400 slaves imported into French colonies between 1700 and 1810, a shortfall of 511,000. An additional 1,109,000 slaves were imported into Spanish America, the Danish and Dutch colonies, and the United States after independence during this period, for a total of 1,619,000 slave imports, or—with French imports— 1,215,000 more than could have been supplied by the British surplus. These 1,215,000 “missing slaves” represent 20 percent of the total imports Curtin previously estimated as arriving in the Americas in the eighteenth century, far more than could have been supplied by the British and minor carriers.14 Available evidence from the shipping side of the computation suggests that the Danes, Dutch, and Americans had a combined share of only 11.9 percent of this period’s slave trade.15 Curtin proposed several possible explanations for the discrepancy between the data on exports from Africa and the estimates of imports into different regions of the Americas. The Portuguese, he argued, may have reexported some slaves from Brazil to the Caribbean and exported others directly to Caribbean ports. Or the sources may have underestimated the volume of slaves carried by minor slave-trading nations. The Dutch in particular may have supplied a larger proportion of slaves before 1760 than afterward, when four separate estimates of its share of the Atlantic slave trade average 5.2 percent.16 An updated version of the Transatlantic Slave Trade Database (TSTD2), which combines records of embarked and disembarked slaves for the same voyage, provides the data needed to evaluate the accuracy of Curtin’s estimates of exports from Africa and imports in the Americas, and to test the merits of possible explanations for discrepancies in his data.17

The Demographic Decline of Caribbean Slave Populations

In the last two decades more precise information on the populations of all Caribbean territories has become available, and the broad patterns of slave arrivals from Africa are also much clearer. The shifts of slaves from the Caribbean to the mainland colonies of North and South America are more difficult to estimate for the seventeenth century than for the eighteenth. At the other end of the period of the slave trade, the Saint-Domingue rebellion and the attendant end of slavery there makes both population counts and movements of the largest black population in the Caribbean unusually speculative. As a consequence, the analysis here is limited to two periods in the eighteenth century: 1716–50 and 1751–90. Eventually it will be possible to push the analysis back before 1716 and forward past 1790, but analysis of the components of population change between 1716 and 1790 enable us to address the data discrepancy for exports from Africa versus imports in the Americas that Curtin has identified. Furthermore, comparing rates of change before and after 1750 enables us to reexamine the current consensus concerning a general ameliorative trend in the vital rates of Caribbean slave populations over the course of the eighteenth century. Table 12.2 summarizes revisions of Curtin’s estimates of slave exports and imports for the period 1716 to 179018 in light of the new data in TSTD2, adjusted for missing voyages according to the methodology explained in 2001.19 After 15 percent mortality on the Middle Passage, Curtin’s estimate of 4,393,000 exports from Africa implies 3,734,000 arrivals in the Americas, 657,000 less than his independent estimate of 4,391,000 imports from records of imports and colonial censuses.20 Embarkations in Africa in TSTD2 total just over 5,000,000, over 615,000 more than Curtin’s estimates of slaves exported from Africa between 1716 and 1790.21 Arrivals amount to 4,269,000 slaves, implying losses of 14.7 percent during passage. Although exports are higher and shipboard mortality is the same as Curtin’s estimates, imports in TSTD2 are 121,491 less than his total for the same period, roughly the amount by which he may have exaggerated rates of natural decrease in the eighteenth-century Caribbean. Whereas Curtin had to estimate slave imports to most regions of the Americas indirectly, owing to a lack of data for doing otherwise, TSTD2 provides direct evidence of imports into every jurisdiction of the Caribbean. Table 12.2 shows that 164,432 slaves disembarked in a colony whose national affiliation differed from that of the ship that carried them. Such was the case for almost 15 percent of the slaves transported on Danish ships and 8 percent of those on British ships, but only 2 percent of slaves on French ships and 1.5 percent of

341

Table 12.2

Comparison of Estimates of Exports from Africa and Imports in the Americas, 1716–1790 Philip Curtin

Transatlantic Slave Trade Data Set Disembarked in Americas

Nationality of Carrier at Embarkation in Africa British French Dutch Danish Spanish Portuguese Total

Exports from Africa

Imports in Americas

Embarked in Africa

In Another Jurisdiction

In Same Jurisdiction

In Same Jurisdiction by Foreign Carrier

1,904,600 857,750 228,430 74,680 0 1,327,400 4,393,000

1,244,470 1,159,600 323,060 17,740 401,960 1,244,000 4,390,830

2,062,132 1,009,864 274,465 46,717 4,671 1,610,131 5,007,980

133,992 18,949 3,502 5,675 425 1,889 164,432

1,551,102 836,301 235,840 32,987 3,634 1,445,042 4,104,906

15,295 53,793 6,641 15,402 65,152 8,149 164,432

Total Disembarked in Jurisdiction 1,566,397 890,095 242,480 48,389 68,786 1,453,191 4,269,338

Sources: Curtin, Atlantic Slave Trade, tables 63 and 64, pp. 211–12 (exports from Africa); table 65, p. 216 (imports into the Americas); TSTD2. Note: British figures combine statistics for British North America and the British Caribbean.

The Demographic Decline of Caribbean Slave Populations

those carried by the Dutch. The largest number of slaves arriving on foreign carriers ended up in Spanish and French colonies. Most disembarkations in foreign ports were legal, either because they occurred when a colony was temporarily occupied during war or because they were authorized by the asiento or administrative exceptions to mercantilist restrictions. In addition to imports directly from Africa on foreign ships as well as national carriers, we shall observe that other slaves were reexported from the port where they originally disembarked. Table 12.3 incorporates the latest estimates on both the volume of the transatlantic slave trade to the Caribbean and the black populations of the various Caribbean regions in the eighteenth century. In 1750—except in the Spanish West Indies, where two-thirds of blacks were free—98 percent of the blacks in the Caribbean were enslaved.22 Thus we use the terms “black” and “slave” interchangeably. Columns 1 and 2 show the black populations at the beginning and end of the selected periods for each Caribbean region. Column 3 lists the annual average rates of population change calculated from columns 1 and 2. Column 4 shows the gross number of arrivals from Africa via the transatlantic slave trade, whether on a foreign ship or national carrier, according to TSTD2. Column 5 presents the annual migration rate m, computed from the following equation: m = __Tr ____ Nt − N0 where T equals the number of immigrants, r the annual growth rate, Nt the population at the end of a time period, and N0 the population at the start of the period.23 Column 6 shows the rate of natural decrease or increase, obtained by subtracting the migration rate from the annual growth rate. The table is divided into two panels, one for each of the two designated periods. Gross imports of slaves in the new estimates derived from TSTD2 imply an average rate of natural decrease of −2 percent per year in the entire Caribbean between 1716 and 1790, both before and after 1750. This rate is a percentage point less than the rates of natural decrease of around −3 percent per year implied by Curtin’s estimates for the British and French Caribbean in the eighteenth century, shown in table 12.1. It is well below the rates of −5 and −6 percent attributed to slave populations in the seventeenth century, and just above a rate of −1 percent for the beginning of the nineteenth century. If rates of natural decrease were as high as −5 percent per year in the seventeenth century, table 12.3 describes a sharp fall in depletion rates after 1700. Rather than a

343

344

David Eltis and Paul Lachance

Table 12.3 Components of Change in Caribbean Slave Populations by Period, 1715–90 Black Population Jurisdictions in 1715

Black Population in 1750

Annual Rate of Increase

Slaves Imported from Africa

Migration Rate

Natural Increase

1715–50 British French Dutch Danish Spanish Total

166,820 84,951 34,330 3,750 30,000 319,851

278,509 325,118 74,494 12,554 120,731 811,406

1.5% 3.9% 2.2% 3.5% 4.1% 2.7%

481,061 261,727 93,022 7,543 3,166 846,519

6.4% 4.3% 5.2% 3.0% 0.1% 4.6%

−4.9% −0.4% −3.0% 0.5% −4.0% −1.9%

1751–90 British French Dutch Danish Spanish Total

278,509 325,118 74,494 12,554 120,731 811,406

472,558 692,871 127,743 30,178 270,196 1,593,546

1.3% 1.9% 1.4% 2.2% 2.0% 1.7%

861,118 619,670 149,605 40,846 31,600 1,702,840

5.9% 3.2% 3.8% 5.1% 0.4% 3.7%

−4.6% −1.3% −2.4% −2.9% 1.6% −2.0%

Sources: TSTD2. In 1715, data for French colonies compiled by James S. Pritchard, In Search of Empire: The French in the Americas, 1670–1730 (Cambridge, 2004), app. 1, 423–27; for other colonial jurisdictions from McCusker, “Rum Trade,” app. B, 601–711, for the most part by interpolation between his estimates for 1710 and 1720; for British Virgin Islands from Robert Wells, The Population of the British Colonies in America before 1776: A Survey of Census Data (Princeton, 1975), 209 and 212; and for Curaçao from David Cohen and Jack Greene, Neither Slave nor Free: The Freedmen of African Descent in the Slave Societies of the New World (Baltimore, 1974), app., 335. In 1750, estimates from Stanley L. Engerman and B.W. Higman, “The Demographic Structure of the Caribbean Slave Populations in the Eighteenth and Nineteenth Centuries,” in Franklin W. Knight, ed., General History of the Caribbean, vol. 3, The Slave Societies of the Caribbean (London, 1997), with assignment of population of islands not yet British to French, Dutch, and Spanish as required; population for Bermuda from McCusker, “Rum Trade,” 704, and figure for Curaçao interpolated between figure for 1715 and 1790 (Engerman and Higman’s 1750 estimate is from 1789 census of Curaçao, cited in our table as a 1790 estimate). Black population in the Dutch and Spanish Caribbean in 1715 and 1750 may be incompletely covered in available sources. In 1790, data for British from John J. McCusker, “Growth, Stagnation, or Decline? The Economy of the British West Indies, 1763–1790,” in Ronald Hoffman, John J. McCusker, Russell R. Menard, and Peter J. Albert, eds., The Economy of Early America: The Revolutionary Period, 1763–1790 (Charlottesville, 1988), 275–80, as well as from “Rum Trade”; French data from Alexandre Moreau de Jonnès, Recherches statistiques sur l’esclavage coloniale (Paris, 1842), 27; Léo Elisabeth, “La société martiniquaise aux XVIIème et XVIIIème siècles, 1664–1789” (thèse d’Etat, Université de Paris I, 1988), 39–42; estimates for Barbados and Puerto Rico from Cohen and Greene, Neither Slave nor Free, 338 and 335; Curaçao from census of 1789 in table 1 of section on “Nineteenth-Century Willemstad” in Linda Rupert, Roots of Our Future (Curaçao Chamber of Commerce and Industry, 2002), also available online at http:// books.caribseek.com/Curacao/Commercial_History_of_Curacao/nineteenth-century-willemstad.shtml. Note: This table differs from table 12.2 in being limited to Caribbean colonies. Imports include slaves arriving in each jurisdiction on national and foreign carriers, but not slaves on national carriers whose arrival port was in another jurisdiction.

The Demographic Decline of Caribbean Slave Populations

gradual trend of improvement in vital rates, the pattern seems instead to be one of very high rates of natural decrease in the first thirty to fifty years following settlement (when Africans made up most of the slave population and before a full generation of Creole women reached the age of reproduction), followed by a substantial drop in rates of natural decrease as the proportion of Creoles increased in the eighteenth century. In addition, the persistence of the same overall rate of natural decrease for the entire Caribbean from 1751 to 1790 as from 1715 to 1750 appears somewhat at odds with the assumption of a long-term ameliorative trend. Differences between the Caribbean colonial jurisdictions in table 12.3 are striking. Whereas Curtin’s data implied a higher rate of natural decrease in French than in British colonies, the rates of natural decrease implied by TSTD2 are much higher in British than in French colonies. Rates are also negative in Dutch colonies in both periods and in Danish colonies after 1750. The rates of depletion in all these jurisdictions contrast with positive rates of natural increase in the Spanish Caribbean. The rate of natural increase of 4 percent per year before 1750 is too high to be taken seriously, but it is quite possible that births did exceed deaths in the populations of African descent in Cuba, Santo Domingo, and Puerto Rico up to the moment they embraced plantation production and began to import large numbers of slaves. Rates of natural decrease are higher after rather than before 1750 in the French and Danish jurisdictions, and the positive rate of natural increase in Spanish colonies declines after 1750. A slight improvement in vital rates is observed in British and Dutch colonies. Is it plausible that rates of natural increase/decrease of Caribbean slave populations in the eighteenth century varied so much from one colonial jurisdiction to another? The divergence of Spanish islands from the Caribbean norm comes as no surprise.24 According to Neville Hall, excess mortality in the Danish colonies of St. Croix, St. John, and St. Thomas persisted into the 1840s. The crude death rate that he cites for the 1780s, 35.7 per thousand, combined with a crude birth rate of 20 per thousand reported in 1791, is only half as high as the depletion rate of −2.9 percent per year in table 12.3 for the Danish colonies from 1751 to 1790.25 Holland, Britain, and France each held several colonies at various stages of development, some of which belonged to one or the other of the imperial powers at different times. The slave demography of the Dutch entrepôt colonies Curaçao and St. Eustatius was quite different from that of the plantation colony Suriname.26 In the case of the French colonies, the depletion rate was less in the Lesser Antilles than in

345

346

David Eltis and Paul Lachance

Saint-Domingue, founded later and characterized by much higher migration and stronger growth.27 A similar variation is likely to have characterized the British colonies. Certainly Higman considers the diversity of demographic experience in the British Caribbean in the nineteenth century to have been as important as the general failure of slave population growth.28 One would be less surprised by divergence in vital rates within jurisdictions than between the averages for all colonies in them. It remains to determine how much of the variation in rates of natural decrease in table 12.3 might be reduced by one important factor potentially affecting migration rates: movement of slaves between colonies and jurisdictions subsequent to their arrival in the Americas. Large numbers of slaves moved around the Americas after they had crossed the Atlantic and after they had had their arrivals recorded. There was considerable displacement, first between islands within a given jurisdiction (for example, from Martinique to Guadeloupe), second between one Caribbean jurisdiction and another, and third between the Caribbean and the non-Caribbean Americas. If these were great enough to significantly alter the gross migration rates in table 12.3, then they would also alter rates of natural decrease obtained by subtracting migration from annual growth rates. In this chapter we offer new evidence on and estimates of the second and third of these movements, use them to calculate rates of net migration, and then recalculate the final column of table 12.3 on the basis of this new information. “Net” in this context means slave arrivals less slave departures for other parts of the Americas. All nationalities involved in the transatlantic slave trade sought markets beyond their own territories, and many of these were best served not by vessels sailing directly from Africa but by smaller ships based in the colonies. The Portuguese and Dutch supply of slaves to the Spanish Americas in the seventeenth century is well documented. In the Dutch case, the great majority of these slaves were first disembarked in Curaçao.29 The French supplied slaves to Guadeloupe through Martinique rather than directly from Africa, especially before 1760.30 When Saint-Domingue collapsed, many slave owners moved with their human chattels to eastern Cuba, several other Caribbean locations, or Louisiana.31 The small Danish islands both imported a large proportion of their slaves from the foreign islands of the Leeward/Windward chain in the second half of the eighteenth century and reexported many of the slaves arriving from Africa. They became major suppliers of slaves to Cuba in the 1790–1810 period.32 In the British Americas, plantation owners moved their coerced labor forces en masse to South Carolina in the seventeenth century, to

The Demographic Decline of Caribbean Slave Populations

the islands ceded by the French in 1763, and later to Trinidad, Demerara, and Cuba. Barbados supplied much of the eastern Caribbean, including the French islands, with slaves in the seventeenth and eighteenth centuries. And the eastern seaboard of North America received a small but steady flow of slaves from the English Caribbean until 1807.33 The largest slave export market for the British Caribbean was, however, Spanish America. While scholars have long assumed that Spanish markets were the main destinations of the departures of slaves from Jamaica in the series compiled by Stephen Fuller in the late eighteenth century from returns of export duties, they have ignored the other regions of the British Caribbean involved in this trade and made insufficient allowance for the Jamaican trade itself.34 Estimates of population decline that do not take into account this intraAmerican movement of slaves are likely to be unreliable. Because the British were such efficient slave traders and supplied so many slaves from their Caribbean possessions to non-British parts of the Americas, the unreliability is likely to be greatest for estimates based on the British experience. Unfortunately, this is precisely the source of many estimates in the literature, if not Curtin’s.35 Not only do earlier estimates of the volume of the slave trade inflate the number of arrivals in the British Caribbean, the role of the British colonies as slave entrepôts for much of the rest of the Caribbean and the English and Spanish mainland Americas has not hitherto received its due. There are two main sources of information on the movements of Caribbean slaves subsequent to their arrival from Africa that historians have exploited. One is Fuller’s series of departures from Jamaica already mentioned, and the other is arrivals in Curaçao, almost all of whom are known to have been carried subsequently to the Spanish mainland Americas.36 The problem is that Curaçao’s entrepôt role declined by the end of the 1720s, while Jamaica was only one source of slaves on which the Spanish and others drew in the eighteenth century. In addition, the Jamaican series does not reveal whether the slaves went to other parts of the Caribbean or to the Spanish mainland. As important as Jamaica and Curaçao were, a great part of the movement of slaves around and out of the Caribbean is not reflected in these series. From the 1660s to the early eighteenth century, the Spanish bought far more slaves than anyone else in the secondary or intra-American market for slaves in the Caribbean, and no reasonable assessment of the scale and direction of trade in this market can be made without drawing on Spanish records. A new data set of 786 vessels carrying slaves into Spanish American ports from other parts of the Americas between 1642 and 1811 provides a basis for

347

348

David Eltis and Paul Lachance

improving our understanding of the intra-American traffic.37 This data set includes vessels of all nationalities, and while it is not complete, it was collected from several sources within the AGI in Seville and there is no particular reason to expect bias in the inclusion of certain national flags, origins, or destinations over others. The early years have scattered entries, and the last year in our period was highly atypical and must be treated to some extent separately.38 For the complete period that interests us here, the database contains 479 voyages carrying 64,135 slaves from 1716 to 1789. If we incorporate the data for 1790 (104 voyages and 906 slaves), we have a total of 583 vessels carrying 65,041 slaves. While thirty-five years have no entries, there are enough in others to form an idea of trends in war and peace years, and new estimates of the volume and direction of the traffic are possible if we compare the new data with the series of departures from Jamaica and Curaçao. Table 12.4 summarizes what the new database reveals about the structure of the eighteenth-century Table 12.4 Origins and Destinations of Slaves Arriving in Spanish Ports from Other Parts of the Caribbean, 1716–1790 Origins in the Caribbean

Destinations in Spanish America

Number of Slaves

Percentage

Jamaica Barbados St. Kitts Curaçao Danish territories Subtotal**

42,886 1,405 816 253 59 45,419

94.4 3.1 1.8 0.6 0.1 100.0

1751–90 Barbados Jamaica Puerto Rico Curaçao Danish territories Subtotal** Total

9,530 3,789 3,510 343 413 17,585 63,004

54.2 21.5 20.0 2.0 2.3 100.0

1716–50

1716–50 Portobelo Cartagena Cuba Mainland* Other West Indies

1751–90 Puerto Rico Cuba Portobelo Mainland* Other West Indies

Number of Slaves

Percentage

22,494 9,774 8,102 6,191 372 46,933

47.9 20.8 17.3 13.2 0.8 100.0

8,699 7,654 1,560 194 3 18,110 65,043

48.0 42.3 8.6 1.1 0.0 100.0

Sources: AGI, Contaduría, 266–68; Indiferente, 2769, 2779, 2807, 2808, 2813; Contratación 2896; Panama, 364; Santa Fe, 1167; Santo Domingo, 2167, 2207, 2483, 2516. *Spanish mainland may include additional slaves carried to Cartagena and Portobelo. **Some slaves included in the table had their destinations, but not their origins, recorded (and some their origins, but not their destinations). This helps explain why the subtotals and totals in the table do not match.

The Demographic Decline of Caribbean Slave Populations

trade to Spanish ports and gives a broad idea of changes in that trade over time. For 1716 to 1789, almost four out of five vessels were English, and most of the rest were Spanish ships sailing from English ports. Such ratios underline the dominance of English slave traders in the intra-Caribbean traffic at this time. We also know that just under 4,000 slaves were taken from Curaçao to the Spanish mainland between 1716 and 1730.39 Only one of the voyages carrying these slaves shows up in the Spanish records, but the absence of other national flags probably reflects the reality that the major non-English traders—the French, Portuguese, and North Americans—were not much involved in supplying eighteenth-century Spanish slave markets, at least before 1790. The left-hand panel of table 12.4 shows 99 percent of the slaves in the first period and 76 percent in the second originating in the British Caribbean. While Puerto Rico had an apparently prominent role as a source after 1750, between 1766 and 1770 the island was the only designated point of entry into the Spanish Americas for slaves.40 Both transatlantic and interisland slave traders sailed into San Juan, but the ultimate destination for most of the slaves they carried was another port in the Spanish Americas, to which they were transported on Spanish vessels. Slaves reexported from the British Caribbean were not likely to have been there for long before their intra-American voyages began—most were no doubt transferred more or less directly from a transatlantic slaver to a colonial sloop. Within the British Caribbean, Jamaica was clearly the dominant provenance before 1750, but thereafter Barbados became the leading port for reexport of slaves to Spanish America, as reflected in the AGI database. A major shift over time is also apparent in the Spanish American ports to which the slaves were carried, shown in the right-hand panel of table 12.4. More than four out of five slaves in the intra-American traffic to Spanish ports went to the Spanish mainland between 1716 and 1750 (and indeed for many years prior to 1716). By contrast, the Spanish mainland accounted for less than 10 percent of imports after 1750, with Cuba and Puerto Rico the destinations for the remainder. We know that both precious metal and plantation produce sectors expanded through most of the second half of the eighteenth century, but the growth of the precious metal sector, located entirely in Central and South America, was clearly not predicated on expanding supplies of African labor.41 A central question for the present analysis is the volume of slaves flowing into Spanish ports from other parts of the Americas. Table 12.4 does not provide the

349

350

David Eltis and Paul Lachance

answer to this directly, but it provides enough information for a minimum estimate that we propose here as a point of departure for future research, fully expecting it to be revised upward as the intercolonial trade is more adequately documented. The main problem with the AGI data set is the large gaps in the years that it covers. From 1716 to 1738, it contains reports of slave shipments arriving from other colonies in the Caribbean in every year but two, but then it does not provide information on another voyage until 1760, despite the continuation of the British asiento to 1748 and the Spanish adoption of a policy of free trade in slaves thereafter. Over the three decades from 1760 to 1790, records of the intra-American slave trade to Spanish colonies are extant for eighteen years, but only one voyage during the 1780s is documented. Our minimum estimate is based on the assumption that AGI coverage is fairly complete in the years for which records have survived. We have reviewed the records of voyages departing from Jamaica in this data set against Stephen Fuller’s series, which has been regarded as being complete apart from smuggling.42 The comparison revealed that the AGI sources document more slaves disembarking in Spanish ports alone in five of nineteen years between 1716 and 1738 than Fuller did theoretically for all destinations of slave ships leaving Jamaica in the same years. Only one year in the AGI data is obviously incomplete. The methodology of our estimate is straightforward. We computed for years with more than 100 slaves embarked in the major ports of departure in the AGI data set—Jamaica, Barbados, the Dutch Caribbean (Curaçao and St. Eustatius), and the Danish Caribbean (St. Croix and St. Thomas)—the proportion of arrivals from Africa that these reexports represented. Then we multiplied annual imports from Africa, as documented in the TSTD, by the average proportion of reexports from each port over each period (1716–50 and 1751–90) to obtain separate estimates of reexports from each port of departure. To the extent that coverage in the AGI records is incomplete, this method underestimates the true number of reexports; but it has the advantage of producing an estimate of the intercolonial trade to the Spanish Americas that is consistent with yearly fluctuations in the number of imports from Africa into the Caribbean. Moreover, it allows us to empirically distribute intra-American voyages originating in the English, Dutch, and Danish colonies, the three exporting jurisdictions in the Caribbean, and to apportion in each the percentage of slaves transshipped to mainland and Spanish Caribbean colonies. Table 12.5 summarizes the British-Spanish slave trade and other movements of slaves from their jurisdictions of entry to other jurisdictions in the

The Demographic Decline of Caribbean Slave Populations

Caribbean and on the mainland. Positive numbers indicate imports, and the negative numbers indicate the colonies receiving slaves. Though much less active than in the seventeenth century, the Dutch sent slaves to the Spanish Americas after 1715 and provided markets for Spanish buyers—in Curaçao in particular. Because it seems obvious that our methodology underestimates this part of the intra-American slave trade—reexports from this port in the six years before 1750 represent only 6 percent of transatlantic imports—we have adjusted upward the number of slaves from the Dutch Caribbean destined for Spanish mainland ports in the first period: from 1,272 to 7,500. For the second period, the procedure used for estimates yields a realistic 8,791 Dutch reexports to Spanish colonies, 7,126 to Puerto Rico and Cuba, and most of the remaining 1,665 to Caracas and Portobelo.43 The estimated 14,000 slaves moving from the Danish islands to the Spanish Americas at the end of the second period is perhaps on the high side, but they are little more than a third of the over 40,000 Africans who disembarked in St. Croix and St. Thomas between 1750 and 1790, according to the TSTD, and they compensate for the absence in AGI records of Danish reexports in the period 1716–50, when 7,500 Africans arrived.44 In total, we estimate 96,062 slaves moving from the English and Dutch Caribbean colonies into Spanish jurisdictions between 1716 and 1750, and 64,981 moving from the English, Dutch, and Danish colonies between 1751 and 1790. As already noted, the AGI series also provides the destinations of slaves within the Spanish Americas, which allows us to distribute these estimates between the Spanish mainland and the Spanish Antilles as shown in table 12.5. The movement of slaves to Spanish jurisdictions was easily the biggest of the intra-American slave trades, but it was scarcely unique. The Dutch supplied many slaves to the French Windward Islands in the seventeenth century. Between 1719 and 1727, the Dutch West India Company carried nearly 10,000 slaves to St. Eustatius and sold them on an open market. Given the large supply of slaves at this time to the English islands, it seems reasonable to assume that all these slaves were bought for the French islands.45 The experiment was not a financial success, and St. Eustatius received no further slaves directly from Africa until 1750. After 1750 a small direct trade to St. Eustatius reappeared—amounting to 6,650 arrivals—and again our assumption is that all these slaves were moved to the French islands.46 Of greater importance throughout the eighteenth century was a steady trickle of slaves from the English to the French islands in the eastern Caribbean. We have no database to support estimates of the size of this traffic, but

351

Table 12.5

Estimates of Reexports of Slaves Subsequent to Their Arrival in the Caribbean from Africa, 1716–90 Reexports

Caribbean Jurisdictions 1716–50 British French Dutch Danish Spanish West Indies Subtotal 1751–90 British French Dutch Danish Spanish West Indies Subtotal

Imports from Africa

British (West Indies)

British (Mainland)

French

Dutch

Danish

Spanish ( West Indies)

Spanish (Mainland)

Total Movement

Net Migration

481,061 261,727 93,022 7,543 3,166

— 31,000 — — 16,259

−10,216 — — — —

−31,000 — −9,897 — —

— 9,897 — — —

— — — — —

−16,259 — — — —

−72,303 — −7,500 — —

−129,778 40,897 −17,397 — 16,259

351,282 302,624 75,625 7,543 19,425

846,519

47,259

−10,216

−40,897

9,897

—

−16,259

−79,803

−90,019

756,499

861,118 619,670 149,605 40,846 31,600

— 10,000 — — 37,415

−11,299 — — — —

−10,000 — −6,644 — —

— 6,650 — — 1,665

— — — — 14,254

−37,415 — −1,665 −14,254 —

−4,521 — −7,126 — —

−63,236 16,650 −15,435 −14,254 53,334

797,883 636,320 134,171 26,593 84,934

1,702,840

47,415

−11,299

−16,644

8,315

14,254

−53,334

−11,647

−22,940

1,679,900

Sources: Same as table 12.4. Subtotals may not add up due to rounding. Note: Positive numbers indicate the colony from which slaves were reexported, negative numbers the colony receiving slaves. Where numbers are the same, they refer to the same movement of slaves; for example, 31,000 reexported from the British to the French Caribbean between 1716 and 1750. All such cases reflect movements between Caribbean islands. Reexports to British and Spanish mainland colonies appear without a corresponding positive number.

The Demographic Decline of Caribbean Slave Populations

an agent of the Royal African Company in Barbados, John Ashley, prepared a detailed breakdown of its size and composition in 1725: Many negroes are exported from Barbados to different places and tho’ the Numbers are very different some Years to what are exported others, yet in a general average I take the following to be a moderate calculation: Martinique Mary galant Guadaloop Leeward Islands Grenada Coast of Caracas New England & Rhode Island New York & the Jerseys Virginia and Maryland Carolina

800 100 150 200 150 uncertain 150 150 500 100

[Subtotal] Sold in the Island for its own use

2300 3200

[Total]

5500

Besides supplying the Spaniards this is a very moderate Calculate.47

Ashley ignored the other British islands in the Eastern Caribbean, perhaps justifiably given that Barbados was by far the region’s most important English entrepôt. Based on Ashley’s assessment, we have assumed 1,000 slaves a year moved from Barbados to the French islands during each of the thirty-one peacetime years between 1716 and 1750 inclusive. After 1750 the slave price differential between the English and French islands narrowed. Although we have no direct evidence of significant movements of slaves from English to French jurisdictions in the east after this date, we have allowed for a further 10,000 coerced migrants between 1751 and 1790. Even this figure may underestimate English reexports to Saint-Domingue, especially in the first period. At the beginning of the eighteenth century, according to Charles Frostin, a high proportion—often more than a third—of slaves brought by North American and British slavers to Jamaica were reexported to Saint-Domingue, whence some were again transshipped to Santo Domingo, Puerto Rico, and Cuba, but most were retained to expand the new French colony’s own sugar production.48 Besides reexports to the Spanish mainland, another “leakage” of slaves from the Caribbean system was the relatively small number of slaves

353

354

David Eltis and Paul Lachance

transported—usually in small batches—from the Caribbean to the thirteen continental colonies. Lorena Walsh has estimated these to have been no more than 7 percent of total arrivals in the Chesapeake on the basis of her own database of slave vessels coming into Virginia and Maryland.49 Sampling of newspapers and the Naval Office shipping lists for South Carolina (henceforth NOSL) suggests that the same ratio is appropriate for this other major North American mainland market for slaves prior 1790. The prevalence of shipboard mortality and shipwreck requires allowance of a further percentage point to arrive at an estimate of departures from the Caribbean. Because slaves arriving in the Chesapeake and South Carolina made up 90 percent of slaves arriving in the entire North American mainland in this period, we can apply the 8 percent ratio to estimates of arrivals in all mainland regions, and the resulting sum can be taken to represent the total of the intra-American slave trade between the Caribbean and what had become the United States by 1790. At 8 percent of total imports, 10,216 slaves arrived by way of the Caribbean between 1715 and 1750, followed by 11,299 between 1751 and 1790, for a total of 21,515 over the two periods.50 Our new estimates of arrivals in the thirteen colonies between 1715 and 1790 by way of the transatlantic trade and of the intercolonial trade add up to 268,937 slaves: 127,702 before 1750 and 141,236 after 1750. The number who arrived between 1720 and 1780 (244,658) confirms Curtin’s estimate for these decades (it is just 3 percent less than his), and suggests that Robert Fogel’s revised estimate of North American imports may have been too high by around 15 percent.51 The predominance of reexports from the British West Indies and of imports into the Spanish colonies is clearly evident in table 12.5. In addition to these population movements were others between Caribbean colonies belonging to the same imperial power. For example, 3,729 slaves were transported from Puerto Rico to Cuba between 1751 and 1790, and reexports were made from Martinique to Guadeloupe and Cayenne throughout the eighteenth century, but these are not indicated in table 12.5 because they would not affect estimates of net migration into each jurisdiction. The difference between total imports from Africa and net migration into the Caribbean—over 90,000 from 1716 to 1750 and almost 23,000 from 1751 to 1790—is the movement out of the region: in each period around 10,000 slaves to the British mainland, the remainder to the Spanish mainland. The estimated number of slaves moving out of the Caribbean would be even higher if they included reexports to Cayenne and Suriname and as many as 17,500 slaves arriving in Spanish Louisiana on British and French ships.52

The Demographic Decline of Caribbean Slave Populations

Using the estimates of net migration in table 12.5, table 12.6 adjusts the migration and depletion rates for each jurisdiction in table 12.3. Reexports to the mainland and other jurisdictions lower the migration rates and lessen the depletion rates in the British and Dutch Caribbean in both periods. Conversely, imports into the French and Spanish Caribbean increase migration rates and either raise the rate of depletion or, in the case of the Spanish West Indies after 1750, lower the rate of natural increase. The overall effect is to diminish somewhat the difference between rates of natural decrease observed in different jurisdictions. Average rates of natural decrease for the entire Caribbean remained as high as −1.5 percent per year before 1750 and worsened slightly in the second half of the eighteenth century, contrary to the general trend of improvement assumed from comparing the high rates of natural decrease at the end of the seventeenth century with much lower rates of natural decrease at the beginning of the nineteenth century. In both periods, though, the rate of depletion in the Caribbean region as a whole now ranks as significantly lower than ordinarily assumed. The British West Indies are the major exception. Even after adjustments for reexports, the rates of natural decrease there remain very high relative to the French Caribbean and to rates of natural decrease revealed by British registration data in the nineteenth century. In table 12.5 we estimated the movement of 40,897 slaves from British and Dutch to French colonies in the period 1716–50 and 16,650 in the period 1751–90. The combined rates of natural increase of British and Dutch colonies are −2.9 percent in the first period and −3.7 percent in the second, compared with −1.0 percent and −1.4 percent, respectively, in the French colonies. To equalize their rates of natural increase at −2.2 percent in the first period and −2.5 percent in the second, an additional 70,000 slaves would have had to have been reexported to French colonies before 1750, plus 210,000 in the four decades after 1750. Movement on this scale is not inconceivable. It would have increased the proportion of arrivals from Africa in British and Dutch colonies who underwent reshipment from 26 percent to 41 percent before 1750 and from 8 percent to 29 percent after 1750, while the percentage of slaves in French colonies supplied by foreign ships would have increased from 14 percent to 33 percent in the first period and from 3 percent to 27 percent in the second. Large as these primarily illegal transfers might seem, they are within the scale at which contemporaries estimated intercolonial movements of slaves. Accused of smuggling, the French governor Joseph d’Honon de Galiffet wrote in 1702 to Louis Phélypeaux, the Count of Pontchartrain and the chancellor

355

Table 12.6

Migration and Depletion Rates Based on Net Rather Than Gross Imports Rates Based on Gross Imports

Jurisdictions

Annual Rate of Increase

Gross Imports

Rates Based on Net Migration

Net Migration

Migration Rate

Implied Natural Increase

Net Migration Rate

Implied Natural Increase

1715–50 British French Dutch Danish Spanish West Indies Subtotal

1.5% 3.9% 2.2% 3.5% 4.1% 2.7%

481,061 261,727 93,022 7,543 3,166 846,519

351,282 302,624 75,625 7,543 19,425 756,499

6.4% 4.3% 5.2% 3.0% 0.1% 4.6%

−4.9% −0.4% −2.9% 0.5% 3.9% −1.9%

4.6% 4.9% 4.2% 3.0% 0.9% 4.1%

−3.2% −1.0% −2.0% 0.5% 3.2% −1.5%

1751–90 British French Dutch Danish Spanish West Indies Subtotal

1.3% 1.9% 1.4% 2.2% 2.0% 1.7%

861,118 619,670 149,605 40,846 31,600 1,702,840

797,883 636,320 134,171 26,593 84,934 1,679,900

5.9% 3.2% 3.8% 5.1% 0.4% 3.7%

−4.6% −1.3% −2.5% −2.9% 1.6% −2.0%

5.5% 3.3% 3.4% 3.3% 1.2% 3.7%

−4.1% −1.4% −2.1% −1.1% 0.9% −2.0%

Sources: Same as tables 12.3 and 12.5. Subtotals may not add up due to rounding. Note: Gross imports are all slaves arriving directly from Africa in the jurisdiction, as reported in table 12.3. Net imports are those remaining after reexports to a colony in the Caribbean outside the jurisdiction or to a British or Spanish mainland colony, as reported in table 12.5. The difference between gross imports and net migration represents movement outside the Caribbean to the mainland.

The Demographic Decline of Caribbean Slave Populations

of France, from Léogane in the south of Saint-Domingue: “Je n’ai jamais eu d’autre vue dans l’introduction des nègres étrangers que le bien du service qui s’y trouvait si considérablement qu’il serait à souhaiter qu’il en fût entré 50.000.”53 Nor do they exceed one informed evaluation of the extent to which the French slave colonies depended upon foreign suppliers: comparing Curtin’s estimates of total imports into French colonies with his own careful count of the French slave trade, Robert Stein remarked that France supplied 87 percent of the slaves required in its colonies between 1721 and 1740 and only 45 percent from 1740 to 1760.54 Moreover, greater reexports to British mainland colonies than we have estimated could partly account for the additional movement out of the British Caribbean needed to reduce its rates of natural decrease to the regional norm. Reexports to British mainland colonies were too small relative to the size of slave populations in the British West Indies and African imports into this jurisdiction to affect either migration rates or rates of natural decrease there, and they were too small a proportion of imports into the mainland colonies to have had much of an effect there on the components of population change. Nevertheless, TSTD2’s distribution by twenty-year periods of African imports into the mainland colonies does alter Curtin’s rather summary estimates for British North America referenced in table 12.1. Whereas Curtin estimated 19,800 arrivals in 1701–20 and 50,400 in 1721–40, TSTD2 includes 25,000 and 87,700, respectively—over 5,000 more in the first two decades and over 27,000 more in the second two decades. From 1741 to 1760, on the other hand, voyages in TSTD2 disembarked only 62,000 slaves in the thirteen colonies, barely three-fifths the 100,400 estimated by Curtin. Estimates of imports based on TSTD2 are also less for 1761–80: 75,500 compared to 85,800 according to Curtin.55 Substituting the new estimates of the slave trade’s volume for Curtin’s estimations in table 12.1 lowers the rate of natural increase of the North American black population from 2.4 percent to 1.8 percent in 1701–20, changes the positive rate of 1.5 percent in 1721–40 to a rate of natural decrease of −0.3 percent, and raises the rate of natural increase from 1.7 percent to 2.6 percent in 1741–60 and from 1.9 percent to 2.0 percent in 1761–80. Additional imports from the West Indies, assuming the ratio to African imports was the same in each period, further reduces the rate of natural increase to 1.6 percent in 1701–20 and −0.7 percent in 1721–40, and to 2.4 percent and 1.9 percent in the next two twenty-year periods, respectively. The improvement in the last two decades fits current notions of an improvement in living conditions of slaves in

357

358

David Eltis and Paul Lachance

the thirteen colonies in the eighteenth century, but the alternating pattern of increase and decrease from decade to decade is unexpected.56 These anomalies could result from problematic population counts rather than from gaps in the slave-trade data. The decade-by-decade figures on the colonial black population in Historical Statistics of the United States, McCusker’s source, are based on a handful of censuses.57 Overall, the adjustments that the new slave-trade data and more refined population data now make possible for the eighteenth century, both transatlantic and intra-American, go far toward resolving a number of problems that Curtin encountered when he attempted to reconcile his own estimates of the slave trade with the population counts then available. In general, adjustments in migration rates and the rates of natural decrease/increase of slaves using TSTD2’s data narrow the differences between different regions in the Caribbean and in the British mainland colonies in the eighteenth century, but hardly by enough to call into question the contrast between failure of the slave population to grow naturally in the Caribbean and its positive natural increase in North America. It still seems likely that lower migration rates to the mainland were associated with more rapid creolization of its slave population, enabling it to grow from an early date by the higher birth rate and lower death rate of Creole compared with African slaves. To determine whether or not blacks in North America also experienced natural decrease in the first decades following conversion from indentured to slave labor, it will be necessary to extend the analysis backward into the seventeenth century and reexamine population counts as well as slave-trade data. Linkage of rates in the Caribbean from 1715 to 1790 with periods before and after this time span is another obvious next step in analyzing the demography of New World populations of African descent. Anomalies in rates obtained using data from TSTD2, such as lower rates of depletion in French than in British Caribbean colonies, also call for further investigation. We have suggested that they could result from underestimation of clandestine Dutch and English reexports to Saint-Domingue, but we are obviously not claiming that these were exactly the numbers needed to equalize rates of natural increase among all northern European jurisdictions in the Caribbean. We do expect, however, that future research will adjust upward our minimum estimates in table 12.5 of the intercolonial movement of slaves. Although not the final accounting, the revisions for the 1715–90 period described in this chapter do demonstrate that the Transatlantic Slave Trade Database offers an essential resource for analysis of the components of population change in this and other periods of Caribbean and North American history.

The Demographic Decline of Caribbean Slave Populations

NOTES

1. J. Harry Bennett, Bondsmen and Bishops: Slavery and Apprenticeship on the Codrington Plantations of Barbados (Berkeley, 1958), 53–62. 2. David Lowenthal, “The Population of Barbados,” Social and Economic Studies 6 (1957): 445–501; Michael Craton, “Jamaican Slave Mortality: Fresh Light from Worthy Park, Longvale and the Tharp Estates,” Journal of Caribbean History 3 (1971): 1–27. 3. Barry Higman, Slave Populations of the British Caribbean, 1807–1834 (Baltimore, 1984), 307–14. 4. Ibid., 6ff. 5. J. R. Ward, British West Indian Slavery, 1750–1834: The Process of Amelioration (Oxford, 1988). 6. Gabriel Debien, Les esclaves aux Antilles françaises (XVIIe–XVIIIe siècles) (Basse-Terre, Guadeloupe, 1974), 343–47. 7. Philip D. Curtin, The Atlantic Slave Trade: A Census (Madison, Wis., 1969). 8. Ibid., 32, 58–59. 9. Ibid., 62 (Leeward Islands), 79 (Saint-Domingue), and 80 (Martinique and Guadeloupe). 10. John McCusker, “The Rum Trade and the Balance of Payments of the Thirteen Continental Colonies, 1650–1775” (PhD diss., University of Pittsburgh, 1970), 584 (table B26) and 712 (table B-104). 11. Curtin, Atlantic Slave Trade, 73: “By the early nineteenth century, Humboldt noticed this contrast to the Caribbean pattern [i.e., natural growth among the slaves of the North American colonies], and other demographers have occasionally remarked on it. It is therefore all the more curious that historians have neglected it almost completely.” 12. Observed by Higman in the British West Indies in the early nineteenth century (Slave Populations, 322, 358), a marked difference between African and Creole natality and mortality also characterized the region in the seventeenth and eighteenth centuries. See Herbert S. Klein and Stanley L. Engerman, “The Demographic Study of the American Slave Population” (paper presented at the International Colloquium on Historical Demography, 1975). 13. Curtin, Atlantic Slave Trade, 211, table 63. 14. Ibid., 219–20. 15. Ibid., 212, table 64, col. 6. 16. Ibid., 220, referring back to table 64 on p. 212 for Dutch exports. 17. TSTD2 distinguishes the version of the data set used for this chapter from the original version in SPSS format (COLE.sav) in the subdirectory “SPSSDATA” in David Eltis, Stephen D. Behrendt, David Richardson, and Herbert S. Klein, eds., The Transatlantic Slave Trade: A Database on CD-ROM (New York, 1999). Further modifications in the database may slightly alter statistics computed from it. The filename of the SPSS version used for this paper is TSTD2.sav and is available from the authors on request, as are the Excel spreadsheets used to calculate statistics from SPSS output. 18. Curtin, Atlantic Slave Trade, 211–12, tables 63 and 64, and 216, table 65, extracting data for the period 1716–90 from Curtin’s statistics for the longer period from 1700 to 1810.

359

360

David Eltis and Paul Lachance

The major carriers are the Portuguese, English, and French. The North American British, Dutch, Danish, and Spanish are considered minor carriers. One-half of the exports of the three major carriers in the decade 1711–20 are attributed to the years 1716–20. Using Curtin’s estimate in table 64 that the three major carriers accounted for 88.1 percent of exports, their 3,870,200 over the entire period of 1715–90 implies 4,393,000 by all carriers. The proportion of this amount attributed to each of the minor carriers is calculated from the average proportions in col. 6 of table 64: 5.2 percent for the Dutch, 1.7 percent for the Danish, and 5.0 percent for the British mainland colonies and the United States. Imports from 1716–20 are computed as one-fourth of those for 1701–20 in table 65, and imports from 1781–90 by the ratio of exports in this decade to exports from 1781 to 1810 as reported in table 63: 35 percent for England, 82 percent for France, 29 percent for Portugal, and an average of 41 percent for minor carriers. 19. Our estimates were calculated in spreadsheets from five-year totals of the variables Slaximp (slave departures from Africa) and Slamimp (slave arrivals in the Americas) for each national carrier in TSTD2. The procedures are described in David Eltis, “The Volume and Structure of the Transatlantic Slave Trade: A Reassessment,” William and Mary Quarterly 58 (2001): 17–46. 20. The 15 percent difference for the period 1716–90 is slightly less than the difference between the imports implied by exports in tables 63 and 64 and the imports in table 65 in Curtin’s Atlantic Slave Trade for the entire period of 1700–1810. 21. TSTD2 indicates that 4,269,145 Africans were embarked on British, French, and Portuguese ships between 1715 and 1790. An additional 527,940 were embarked on ships of minor slave-trading nations. 22. Stanley L. Engerman and B. W. Higman, “The Demographic Structure of the Caribbean Slave Societies in the Eighteenth and Nineteenth Centuries,” in Franklin W. Knight, ed., General History of the Caribbean, vol. 3, The Slave Societies of the Caribbean (London, 1997), 48–49, table 2.1. 23. Curtin, Atlantic Slave Trade, 32. 24. Robin Blackburn, The Making of New World Slavery from the Baroque to the Modern, 1492–1800 (London, 1997), 127–60. 25. Neville Hall, Slave Society in the Danish West Indies: St. Thomas, St. John, and St. Croix (Baltimore, 1992), 85. The crude birth rate is equivalent to the annual birth rate of about 2 percent in the report of the Grand Commission on the Slave Trade in 1791, cited in Svend E. Green-Pedersen, “The Scope and Structure of the Danish Slave Trade,” Scandinavian Economic History Review 19 (1971): 156. The annual rates of natural decrease in St. Croix implied by statistics in this article on slave imports into Christiansted and Frederiksted, however, are higher: −2.6 percent between 1766 and 1775, and −4.4 percent between 1775 and 1786 (−4.1 percent when free persons of color are included in the population counts). 26. Johannes Postma, The Dutch in the Atlantic Slave Trade, 1600–1815 (Cambridge, 1990), 174–200. 27. Paul Lachance, “The Demography of French Slave Colonies, Part I (1700–1760)” (paper presented at the annual meeting of the Social Science History Association, Chicago, 1998), 17. Between 1700 and 1760, after readjustment for reexports from Martinique to

The Demographic Decline of Caribbean Slave Populations

Guadeloupe, the rates of depletion in these two colonies and in Saint-Domingue were −1.6, −0.7, and −2.9 percent per year, respectively. 28. Higman, Slave Populations, 5. 29. For the Portuguese, see António de Almeida Mendes, “The Foundations of the System: A Reassessment of the Slave Trade to the Spanish Americas in the Sixteenth and Seventeenth Centuries,” this volume (chap 2). For the Dutch, see Johannes Postma, “A Reassessment of the Dutch Atlantic Slave Trade,” and Wim Klooster, “An Overview of Dutch Trade with the Americas, 1600–1800,” in Johannes Postma and Victor Enthoven, eds., Riches from Atlantic Commerce: Dutch Transatlantic Trade and Shipping, 1585–1817 (Leiden, 2003), 115–38 and 365–84. 30. The commercial network in which Saint-Pierre served as the main nexus of French colonies in the southern Caribbean with the larger Atlantic world is described in Anne Pérotin-Dumont, “Cabotage, Contraband, and Corsairs: The Port Cities of Guadeloupe and Their Inhabitants,” in Franklin Knight and Peggy Liss, eds., Atlantic Port Cities: Economy, Culture, and Society in the Atlantic World (Knoxville, 1991), 58–86. From 1661 to 1700, according to TSTD2, 18,757 Africans arrived in Martinique and 1,455 in Guadeloupe, yet Martinique’s black population numbered 15,073 and Guadeloupe’s 7,050 in 1700. From 1701 to 1760, 109,775 Africans were imported into Martinique, and only 2,347 into Guadeloupe; yet Martinique had 69,375 blacks and Guadeloupe 55,204 in 1760. The disparity between imports and population implies reexport to Guadeloupe of many slaves who arrived in Martinique. After 1760, direct imports to Guadeloupe slightly exceeded imports to Martinique. See Anne PérotinDumont, La ville aux îles, la ville dans l’île: Basse-Terre et Pointe-à-Pitre, Guadeloupe, 1650–1820 (Paris, 2000), 191, table 4.8. 31. Olga Portuondo, “La inmigración negra de Saint-Domingue en la jurisdicción de Cuba (1791–1809),” Espace Caraibe 2 (1994): 169–98; Paul Lachance, “Repercussions of the Haitian Revolution in Louisiana,” in David Geggus, ed., The Impact of the Haitian Revolution in the Atlantic World (Columbia, S.C., 2001), 209–30. 32. Hall, Slave Society, 23; Green-Pedersen, “Scope and Structure,” 159, 164–65, 170; Svend E. Green-Pedersen, “Colonial Trade under the Danish Flag: A Case Study of the Danish Slave Trade in Cuba, 1790–1807,” Scandinavian Journal of History 5 (1980): 93–120. 33. Greg O’Malley, “The Intra-American Slave Trade: Forced African Migrations within the Caribbean and from Islands to the Mainland” (paper presented at the annual meeting of the American Historical Association, Philadelphia, 2006). 34. Fuller’s series is summarized in five-year periods from 1701 to 1808 in Richard B. Sheridan, “Slave Demography in the British Caribbean and the Abolition of the Slave Trade,” in David Eltis and James Walvin, eds., The Abolition of the Atlantic Slave Trade: Origins and Effects in Europe, Africa, and the America (Madison, Wis., 1981), 274. 35. Curtin (Atlantic Slave Trade, 57–59) takes interisland migration into account in estimating natural decrease in Barbados and Jamaica. Because Fuller’s series of slave exports from Jamaica has been well known for some time, studies of Jamaica have usually taken slave departures into account, but the eastern Caribbean has not always been assessed in the same light—see note 2 above.

361

362

David Eltis and Paul Lachance

36. Han Jordaan, “The Curaçao Slave Market: From Asiento Trade to Free Trade, 1700–1730,” in Postma and Enthoven, eds., Riches from Atlantic Commerce, 219–57. 37. Collected by Oscar Grandío Moráguez for the authors from AGI: Contaduría, 266–68; Indiferente, 2769, 2779, 2807, 2808, 2813; Contratación, 2896; Panama, 364; Santa Fe, 1167; and Santo Domingo, 2167, 2207, 2483, 2516. 38. In terms of voyages rather than slaves, 1790—the first full year of unrestricted slave entries into the Spanish Caribbean—contains no less than 104 or 17.8 of all voyages in the data set for the years 1716–90. On the other hand, these voyages carried only 906 slaves or just over 8.7 slaves per vessel, compared to an average in preceding years of 134 slaves per ship. It is likely that many small vessels took advantage of the new regulation; but the average for the next year with data on voyages—1792, and thus not of direct concern in the present discussion—was 120. 39. Calculated from Jordaan, “Curaçao Slave Market,” 248–49. Similarly, we know from TSTD2 that 20,458 slaves arrived in Curaçao between 1716 and 1780. Only one of these voyages can be found in the AGI records for these years. 40. For the role of Puerto Rico at this time, see Bibiana Torres Ramírez, La Compañía Gaditana de negros (Seville, 1973). 41. For the expansion of silver output and its implication for African labor, see Dennis Nodin Valdes, “The Decline of the Sociedad de Costas in Mexico City” (PhD diss., University of Michigan, 1978), 139–74. 42. The Fuller series was based on returns of duties paid on slaves carried out of the island, and Sheridan suggested that smuggling or avoidance of duty was significant. McCusker’s general strictures against making excessive allowance for smuggling apply here; but to the extent that it was a major problem, our estimates of slave departures will have a downward bias, and the population growth rates calculated using these estimates will have an upward bias. 43. Jordaan, “Curaçao Slave Market,” 219–57. 44. Calculated from the data set of slaves arriving in the Spanish Americas from other parts of the Americas. See table 12.4. 45. For the St. Eustatius experiment, see Henk den Heijer, “The West African Trade of the Dutch West India Company, 1674–1740,” in Postma and Enthoven, eds., Riches from Atlantic Commerce, 161. 46. Calculated from TSTD2. 47. Huntington Library, San Marino, California, Stowe Ms., ST 9, pp. 50–51. Nine hundred of this estimated number (slaves to New England, New York, the Jerseys, Virginia, Maryland, and Carolina) are included in the estimate of arrivals of slaves from the Caribbean in the continental colonies in table 12.5. 48. See Charles Frostin, “Histoire de l’autonomisme colon de la partie française de St. Domingue aux XVIIe et XVIIIe siècles: Contribution à l’étude du sentiment américain d’indépendance,” 2 vols. (Thèse, Doctorat d’Etat, Université de Paris 1, 1972), 1:350, and the discussion later in this chapter that note 53 references. 49. Lorena S. Walsh, “The Chesapeake Slave Trade: Regional Patterns, African Origins, and Some Implications,” William and Mary Quarterly 58 (2001): 139–70.

The Demographic Decline of Caribbean Slave Populations

50. These should be considered as minimum estimates. Colonial census statistics show 13 to 15 percent of slaves were imported to particular colonies from the Caribbean, and in the years 1768–72 the proportion was 35 percent. See U.S. Bureau of the Census, Historical Statistics of the United States, Colonial Times to 1970, bicentennial ed., part 2 (Washington, D.C., 1975), 1172–74: “Slave Trade, by Origin and Destination: 1768 to 1772” [Z 133–45], “Slave Trade in Virginia: 1619–1767” [Z 146–49], “Slave Trade in New York: 1701 to 1764, except 1701–18, unknown” [Z 150–54], and “Slaves Imported into Charleston, S.C., by Origin, 1706–1775” [Z 155–64]. In a paper using a database of the British intercolonial slave trade constructed for his dissertation, Greg O’Malley has presented preliminary findings that the cumulative total of imports from the British Caribbean to North America surpassed 60,000 by 1808—in effect, tripling our estimate, which is deliberately conservative. See O’Malley, “Intra-American Slave Trade.” 51. In Curtin, Atlantic Slave Trade, 140, table 40, imports from 1720 to 1780 total 236,000, while Robert Fogel has calculated 289,630 imports in the same decades. See Robert W. Fogel, “Revised Estimates of the U.S. Slave Trade and of the Native-Born Share of the Black Population,” in Robert W. Fogel, Ralph A. Galantine, and Richard L Manning, eds., Without Consent or Contract: The Rise and Fall of American Slavery: Evidence and Methods (New York, 1991), 55. 52. The number of imports to Louisiana is estimated from population growth between 1766 and 1788 and the rate of natural decrease of −2.2 percent per year estimated by Curtin. See Paul Lachance, “The Politics of Fear: French Louisianians and the Slave Trade, 1786–1809,” Plantation Society in the Americas 1 (1979): 196–97. 53. Frostin, “Histoire de l’autonomisme,” 1:350. 54. Robert Stein, The French Slave Trade in the Eighteenth Century: An Old Regime Business (Madison, Wis., 1979), 26. 55. Curtin, Atlantic Slave Trade, 140, table 40. Total imports in TSTD2 are closer to the estimates of African immigration to Virginia and South Carolina from 1700 to 1790 in Philip D. Morgan, Slave Counterpoint: Black Culture in the Eighteenth-Century Chesapeake and Lowcountry (Chapel Hill, 1998), 59, table 9. 56. For a representative overview of U.S. slavery history, see Peter Kolchin, American Slavery, 1619–1877 (New York, 1993); see 38–39 regarding the eighteenth century. 57. Henry Gemery, “The White Population of the Colonial United States,” in Michael Haines and Richard Steckel, eds., A Population History of North America (Cambridge, 2000), 144–45.

363

This page intentionally left blank

Contributors

Daniel Barros Domingues da Silva is a graduate student at Emory University. He is the author of “The Atlantic Slave Trade to Maranhão, 1680–1846: Volume, Routes and Organization,” Slavery and Abolition (forthcoming). David Eltis is the Robert W. Woodruff Professor of History at Emory University. He is the author of Rise of African Slavery in the Americas (Cambridge, 2000). Roquinaldo Ferreira is assistant professor of African and African-American history at the University of Virginia, Charlottesville. His book, Angola and Brazil in the Atlantic World: Trade and Social and Cultural Landscapes of Slaving, 1680–1830, will be published by Cambridge University Press. Manolo Florentino is a professor in the graduate program in history at the Universidade Federal do Rio de Janeiro. He is the author of Em costas negras: Uma história do tráfico atlântico de escravos entre a Africa e o Rio de Janeiro (São Paulo, 1997). Paul Lachance is visiting professor of history at Emory University and a member of the project development team for the online version of the Transatlantic Slave Trade Database. 365

366

Contributors

António de Almeida Mendes received his PhD from the Ecole des Hautes Etudes en Sciences Sociales in 2007. He has published several articles on the early Portuguese slave trade and is currently affiliated with the Centre International de Recherches sur les Esclavages in Paris. Philip Misevich is a graduate student at Emory University. He is the author of “In Pursuit of Human Cargo: Philip Livingston and the Voyage of the Sloop ‘Rhode Island,’ ” New York History 86 (2005): 185–204. Oscar Grandío Moráguez is a PhD candidate in history at York University, Ontario. He is the director of the project “Collections on African Slaves and Their Descendants at Cuban Provincial Archives,” cosponsored by the Harriet Tubman Institute on the Global Migration of African People and the Instituto de Historia de Cuba. James Pritchard is emeritus professor of history at Queen’s University, Kingston, Ontario, and the author most recently of In Search of Empire: The French in the Americas, 1670–1730 (Cambridge, 2004). Alexandre Vieira Ribeiro is a doctoral student at the Universidade Federal do Rio de Janeiro. He is the author of “O comércio de escravos e a elite baiana no período colonial,” in João Fragoso, Antônio Carlos de Jucá Sampaio, and Carla Maria Carvalho de Almeida, eds., Conquistadores e negociantes: Histórias de elites no Antigo Regime nos trópicos. América Lusa, séculos XVI a XVIII (Rio de Janeiro, 2007). David Richardson is professor of economic history at Hull University, Yorkshire, UK, and director of the Wilberforce Institute for the study of Slavery and Emancipation. He is the author of numerous articles on the slave trade and editor of several collections of essays. Jelmer Vos is a postdoctoral research fellow at the Instituto de Investigação Científica Tropical in Lisbon and is affiliated with the National Institute for the Study of Dutch Slavery and Its Legacy in Amsterdam. He is the author of “Slavery in Southern Kongo in the Late Nineteenth Century,” in CEAUP, ed., Trabalho forçado africano: Experiências coloniais comparadas (Porto, 2006). Andrea Weindl is an affiliate of the Institut für Europäische Geschichte in Mainz, where she has been working on European peace treaties of the early modern era. She is the author of Wer kleidet die Welt? Globale Märkte und merkantile Kräfte in der europäischen Politik der Frühen Neuzeit (Mainz, 2007).

Index

Abolition of slave trade, 138, 140, 160, 264, 328. See also Laws, anti-slave-trade; Suppression of slave trade, British Abréu, Nery Gómez, 180 Accra, 255, 264 African languages. See Linguistic identification and African languages AGI. See Archivo General de Indias (Seville) AHMS, 131 AHU. See Arquivo Histórico Ultramarino (Lisbon) Alden, Dauril, 109 Alencastro, Luiz Felipe de, 18, 110 Allada, 255–256 Alvarás para navegar, 131 Alves da Silva, Antônio, 288 Amazonia, 16–17, 19–20, 95, 110 Ambriz, 192, 194, 326, 327 Amistad (schooner), 159, 168

Angola, slave trade from, 313–334; and asientos, 76; background, 314–317; to Bahia, 18; and Benguela, 317–321; and Dutch slave trade, 120; and imported merchandise, 285–287; and Luanda, 20, 314, 315–317, 320, 321–325, 326, 328; and northern Angola, 325–328; to Pernambuco, 97, 98–99, 100, 101, 107, 110, 113, 115–116; to Rio de Janeiro, 20–21, 323; to São Tomé, 84; and Spanish Americas, 66, 67, 85; and suppression, 313–334; and TSTD2, 314, 318–319, 321, 323, 324, 325–327 ANRJ. See Arquivo Nacional do Rio de Janeiro Antigua, 335, 336 Antislavery squadrons, 157, 160, 184, 186–187, 324–325 Antonil, André João, 20, 108, 110 367

368

Index

APEB. See Arquivo Público do Estado da Bahia Archivo General de Indias (Seville), 68, 69, 80, 183, 348, 349, 350, 351 Archivo Histórico Nacional (Madrid), 183 Ardra, 223, 240, 242, 243 Arguim, 65–66, 70–72, 93–94, 255 Arquivo Histórico Municipal de Salvador (AHMS), 131 Arquivo Histórico Ultramarino (Lisbon), 18, 20, 101, 110 Arquivo Nacional do Rio de Janeiro (ANRJ), 131 Arquivo Público do Estado da Bahia (APEB), 131 Arrival estimates (slaves). See Departurearrival estimates (slaves) Ashley, John, 353 Asientistas, 73, 74, 88, 89, 233, 258. See also Asientos Asientos: and British slave trade, 350; and Dutch slave trade, 22, 254, 258; and French slave trade, 206, 210, 215; Grillo and Lomelin asiento, 35, 68, 77, 88; and northern German slave trade, 258, 259; and Portuguese asiento period (1595–1640), 68, 70, 75–79, 81–82, 85, 97, 256; and São Tomé, 76. See also Asientistas; Contracts (slave) Atlantic markets, arrivals of slaves in, 48–51 Atlantic Slave Trade: A Census, The (Curtin), 3, 4 Atlas of the Slave Trade (Yale University Press), 4 BAAC. See Brandenburg African American Company BAC. See Brandenburg African Company Badagry, 143 Bahia, slave trade to, 130–154; and African origins, 18–19, 140–142, 144, 149; attacks on slave shipments,

145–146; importance to transatlantic trade, 148–149; length of voyage, 148, 149; literature review, 130–131, 132; and political instability, 142–144; and redistribution of slaves, 136–138; and Rio de Janeiro, 137, 138; shipboard mortality, 144–145, 146–148, 149; and sugar production, 133, 135, 136, 138, 149; and tobacco, 142; in TSTD2, 15, 18–19, 133, 140; and volume of slave arrivals, 130–135, 137, 139–140, 141, 144, 148–149, 153–154. See also Brazilian slave trade Baia Farta, 320–321 Bance Island, 160, 161 Barbados, 24, 25, 244, 335, 336, 339, 349, 353 Barbosa, Anacleto José, 139 Barnet, Miguel, 179 Bauss, Rudolph W., 278 Beato, Manuel Pérez, 178 Behrendt, Stephen, 11 Below-deck conditions, 2, 146, 147, 149, 170. See also Shipboard mortality Bengo River, 107–108 Benguela, 20, 99, 189, 314, 315, 316–321, 328 Benin, 66, 70, 75, 84. See also Bight of Benin Biblioteca Nacional do Rio de Janeiro (BNRJ), 131 Bight of Benin: and Bahian slave trade, 18; and Cuban slave trade, 186, 188, 190, 194, 195; and Dutch slave trade, 120, 239, 243; and French slave trade, 222; and tobacco trade, 142. See also Benin Bight of Biafra: and Cuban slave trade, 182, 184, 186, 187, 191, 194, 195; and Dutch slave trade, 120, 239, 243 Binder, Franz, 22, 213, 229–230, 231, 232 Birmingham, David, 20 Blanco, Pedro, 189–190

Index

BNRJ (Biblioteca Nacional do Rio de Janeiro), 131 Boer, Cornelis Jansz, 232 Bonny, 188–189, 191, 194, 195 Boston, 28, 30 Bourbon reforms, 34, 36 Bowser, Frederick, 68, 118 Brandenburg, 31, 34, 263–267 Brandenburg African American Company (BAAC), 253, 254, 255–256, 257, 258, 260–261, 263, 266 Brandenburg African Company (BAC), 252–253, 254, 256, 257 Brásio, António, 67 Brazilian slave trade: to Amazonia, 19–20; early decades of, 13–14; and slave estimates by decade, 16–17. See also Bahia, slave trade to; Pernambuco, slave trade to; Rio de Janeiro, slave trade in Bristol, 26, 122 British Caribbean, 25, 26, 339, 346, 347, 349, 357. See also British slave trade British Parliamentary Papers, 105, 115 British slave trade, 23–26, 39, 347. See also British Caribbean; Suppression of slave trade, British British West Indies. See West Indies Brito, Domingos de Abreu de, 107 Brübach, Nils, 251 Caballos, Esteban Mira, 68 Cabinda, 120, 139, 192, 316, 325–328 Cabrera, Lydia, 179 Cádiz, 74, 83, 117, 252 Calabar, 121, 191 Caldeira, Manuel, 74, 79 Caminho Novo, 137 Campbell, James, 170 Canot, Theodore, 190 Cape Verde Islands, 65–66, 67, 69, 70, 118 Caribbean slave population, demographic decline of, 335–363

Carioca, 281, 288, 289, 291, 293, 295, 296, 298 Carreira, António, 113, 115 Cartagena, 70, 76, 77, 85, 86, 88 Carvalho, Marcus de, 105, 115 Casa da Mina e da India, 68, 74 Casa de Contratación, 69, 76, 254 Casanova, Manuel Martínez, 180 Castile, James, 259 Castro, André de Melo e, 144 Cavalcanti, Nireu, 21 Charles I (later Charles V), 64 Chaunu, Huguette, 68 Chaunu, Pierre, 68 Child slaves, 52, 72, 135, 284, 299, 301, 302, 304 Cibao, 65, 66, 74 Codrington, Christopher, 258 Colbert, Jean-Baptiste, 22, 218, 220, 224, 235 Colón, Diego, 64 Companhia de Cach˜eu, 256, 258 Companhia Gaditana, 36 Companhia Geral do Grão-Pará e Maranhão, 19, 20 Congo. See Kongo Congo River, 188–189, 192–194, 195. See also Kongo Contracts (slave), 75, 76, 77, 81–82. See also Asientos Contraduría series, 68 Contratación series, 68 Coopstad and Rochussen, 230, 231 Correia Picanço, Ignácio, 288 Costa da Mina, 18, 21, 100, 101, 110, 120–122, 139 Costa Ferreira, Felipe Justiniano, 146 Cotton, 111, 286 Coughtry, Jay, 28, 29 Courland, Duke of, 31, 34, 221, 251–263 Courts of mixed commissions, 157, 162 Coymans, Balthasar, 254, 258 Coymans, Caspar, 254

369

370

Index

Creoles, 339, 345, 358. See also Crioulos Crioulos, 135, 299, 302, 303, 304, 305–306, 307–308, 309. See also Creoles Cuba, origin of slaves arriving in, 176–201; literature review, 178–181; and North American slave trade, 28; and slave ethnicities, 180–181, 189, 193; and slave registers, 158, 162; and Spanish slave trade, 36; from specific embarkation points, 188–193, 194; and volume of slave arrivals, 182, 183–185; from West Central Africa, 183–184, 186, 187, 191–192, 194 Cuban National Archives, 180 Curaçao, 35, 213, 235, 347, 348–349 Curtin, Philip: and Brazilian/Portuguese slave trade, 14, 15; and census of Atlantic slave trade, 337, 338–339; on coastal slave transport, 171; and departure-arrival estimates, 39, 42, 53, 340–342, 357–358; and early slavetrade studies, 3–4; and North American slave trade, 354; and shipboard mortality, 4; and size of slave trade, 67; and slave demography, 5, 345; on slave records, 177 Curto, José, 20–21

Os descobrimentos e a economia mundial (Godinho), 67 Diaz, João, 84 Dolben’s Act of 1788, 88 Domingues da Silva, Daniel Barros, 14, 45, 95–129 Dorotheenschanze, 253 Duarte de Silva, Manuel, 145 Ducasse, Jean-Baptiste, 259 Dumont, Henri, 178, 179 Dutch Guianas, 243 Dutch slave trade, 228–249; and Bight of Benin, 120, 239, 243; and Bight of Biafra, 120, 239, 243; decline of, 237, 238, 240, 244; destinations of voyages, 232, 233–235; five periods of, 232–233, 235–236; and Gold Coast, 239, 242; and involvement in Brazil, 97–98, 142; literature review, 229–231; and Luanda, 240, 242; overview of, 22–23; and Pernambuco, 233, 237; rivals of, 237–238; and sugar production, 244–245; in TSTD1, 231; in TSTD2, 22–23, 231, 233, 242; and volume of slave arrivals, 232, 237, 238, 249 Dutch West India Company, 97, 140. See also West India Company Dyewood, 106

Daget, Serge, 27 Dahomey kingdom, 142–143, 144 Danish African Company, 263 Danish slave trade, 31, 34, 261, 264 Danish West India Company, 31, 263 De la Fuente, Alejandro, 179–180 De Laet, Johannes, 97, 106, 108 Debien, Gabriel, 337 Den Heijer, Henk, 230–231 Departure-arrival estimates (slaves), 39, 42, 44–51, 53, 340–343, 357–358. See also Size of slave trade, estimated Depletion rates (slaves), 335, 339, 345–346, 355, 356

Ehingher, Enrique, 73 Elmina, 98, 99, 101, 118, 240, 243 Eltis, David: on Caribbean slave populations, 335–363; on Dutch slave trade, 228–249; on end of slave trade, 314; on estimated slave arrivals, 131; on European stronghold at Luanda, 316; on French slave trade, 205–227; on slave trade in northern Angola, 325; on slave trade to Pernambuco, 95–129; on transatlantic slave trade, 1–60 Emden, 252, 260 Emmer, Pieter C., 98, 99 Engenhos, 97, 106, 109, 134

Index

English Navigation Acts, 22 Escribania series, 80 Ethnic groups, African, 162–163, 165, 177–178, 180–181, 189–190, 193 Eusébio de Queiroz law, 140 Exclusif policies, 22, 211, 214, 218, 219, 233 Female slaves, 52, 72, 73, 135, 299, 301, 302–304, 307 Ferdinand II of Aragon, 73 Ferreira, André, 74 Ferreira, Pedro Gomes, 146 Ferreira, Roquinaldo, 52, 313–334 Filho, Luís Viana, 130 Finley, Moses, 277 Florentino, Manolo, 11, 52, 275–312 Flores, Francisco Antonio, 322–323 Fogel, Robert, 354 Folmer-van Prooijen, Reinders, 230 Franco, Marisa Vega, 68, 77 Freetown, Sierra Leone, 161 Freire, Braamcamp, 67 French Caribbean, 27, 207, 212, 214, 218, 223, 338–339, 355. See also French slave trade French slave trade, 205–227; African origins of, 221–223; and African produce trade, 217; and Bight of Benin, 222; defined, 206–207; destinations of voyages, 215–216; formulas for estimating, 207, 208, 209, 210; overview of, 26–28; problems encountered by, 217–221, 223; shipboard mortality, 217–218; and sugar production, 218, 219; in TSTD1, 207; in TSTD2, 27, 214, 226–227; and volume of slave arrivals, 205, 207–210, 211–215, 216, 220, 222, 223. See also French Caribbean Friedrich Wilhelm (Elector of Brandenburg), 252, 253, 261, 266 Frostin, Charles, 353

Fuentes, Lutgardo García, 68, 74, 77, 79, 80, 83 Fuller, Stephen, 347, 350 Furé, Rogelio Martínez, 179 Galegos, 318 Galiffet, Joseph d’Honon de, 355 Gallinas, 161, 163, 167, 168, 171, 188, 189, 190 Gambia, 251–252 Ganga, 162, 163 Geggus, David, 206 Geheimes Staatsarchiv Preußischer Kulturbesitz (GStA), 251 Gibbs, Josiah W., 159–160 Gildem, Cornel, 251 Goa, 288, 293, 296 Godinho, Vitorino Magalhães, 67 Gold, 65, 109, 120, 136, 137, 138 Gold Coast: and Cuban slave trade, 184, 186, 190; and Dutch slave trade, 239, 242; and North American slave trade, 29; and Northern German slave trade, 253 Gomes, Francisco Ferreira, 321 Gomes, José Ferreira, 321 Gonçalves de Carvalho, Manoel, 288 Gonçalves Moledo, Manoel, 288 Goslinga, Cornelis, 230 Goulart, Maurício, 14, 20, 115, 131 Gouvenod, Lorenzo de, 64 Graça, Raimundo Justino da, 145 Grand Cape Mount, 167, 169, 171 Great Elector. See Friedrich Wilhelm (Elector of Brandenburg) Grillo, Domingo, 35, 68, 77, 82, 88 Gross Friedrichsburg. See Großfriedrichsburg Großfriedrichsburg, 31, 253, 255, 260 GstA. See Geheimes Staatsarchiv Preußischer Kulturbesitz Guanche Pérez, Jesús, 180–181 Guinea Company, 26

371

372

Index

Hair, Paul, 159, 205 Hall, Neville, 345 Hamburg, 263–264 Hanse towns, 251, 263–266 Havana Court of Mixed Commission, 157, 183 Havana Registers, 157–158, 161, 162, 163, 164 Hawkins, John, 24 Hernaes, Per, 5, 31, 34, 53 Higman, Barry, 170, 337, 346 Hispaniola, 63, 64, 65, 86 Hispanoamérica y el comércio de esclavos (Vilar), 67 Holsoe, Svend, 31 House of the Contratación. See Casa de Contratación Humboldt, Alexander von, 182 Iberian Peninsula, 70, 72, 79, 89, 118 Idade d’Ouro do Brasil, A, 131 Iles de Los, 160, 161, 189 Illegal slave trading: from Angola, 314, 315, 325, 328; on Benin River, 75; and courts of mixed commissions, 157; and Cuban slave trade, 192; Hanseatic involvement in, 264–265; and national flags, 12; in North America, 30; to Pernambuco, 104; in Rio de Janeiro, 20; in Spanish Americas, 79–80, 81, 83, 86 India, 285, 286, 287 Indigenous slaves, 65, 134, 135 Inikori, Joseph E., 53 Intra-American slave trade, 19, 42, 85, 336, 347–351, 354, 355, 358 Isabella I of Castile, 72–73 Jamaica, 86, 339, 347, 348, 349, 350 Jewish merchants, 233 João III (Portuguese king), 66 Jones, Adam, 251 Junta do Commercio series, 101

Kasanje, 316, 321–322, 326 Kellenbenz, Hermann, 251 Kettler, Jacob (James), 251 Kissi, 166, 167, 168 Klein, Herbert S., 11, 20–21, 182, 183 Koelle, Sigismund Wilhelm, 177 Kongo, 66, 75, 84, 120, 313, 315, 325. See also Congo River Koopman, Ruud, 230 La Rochelle, 220 Lachance, Paul, 5, 39, 52, 335–363 Lachatañeré, Rómulo, 179 Lagos, 65, 70, 143, 190, 194, 195 Laing, Alexander Gordon, 166 Lamb, George, 187 Lançados, 83, 84 Laws, anti-slave-trade, 30, 134, 140, 315, 319, 320. See also Abolition of slave trade Licenses, slave: and Angolan slave trade, 319; and Bahia, 131, 139; and slave trade to Spanish Americas, 70–75, 76, 77–79, 82, 85 Lima, Domingos José, 146 Lima, José de Cerqueira, 140 Linguistic identification and African languages, 164, 165–168, 172, 177, 181–182, 189, 190–191, 193–195. See also Mande languages; Mel languages Lisboa, Félix da Costa, 146 Lisbon, 70, 72, 83, 88, 93–94, 117 Liverpool, 24 Livros de Visitas da Saúde, 131 Loango, 192, 194, 240, 242, 243 Lomelin, Ambrosio, 35, 68, 77, 82, 88 London, 11–12, 24 Lopes, Edmundo Correia, 14, 115 Lopes, Elias Antônio, 288 Luanda: and Angolan slave trade, 20, 314, 315–317, 320, 321–325, 326, 328; and Cuban slave trade, 184, 189, 192; and Dutch slave trade, 240, 242; merchan-

Index

dise imported into, 285–286; ships from Pernambuco to, 99, 107, 118, 120 Lucumí, 178, 182, 194 Luso-Brazilians, 136, 143, 283 Ly, Abdoulaye, 205–206 Macías, José Miguel, 177 Malebo Pool, 192, 193, 194 Malembo, 240, 242, 243 Mande languages, 164, 194 Mandinga, 162–163, 165, 166 Mandingo, 162–163, 165, 166 Mangombe, Mdembo Muene Damba, 326 Maranhão, 19, 20, 111 Marinho, Joaquim Pereira, 140 Marriage, slave, 304–309 Márties, Frei Bartolomeu dos, 289 Martinique, 213, 214–215, 218, 259, 346 Matamba, 120, 316, 321–322 Mboma, 193, 194 MCC. See Middelburgse Commercie Compagnie McCusker, John, 337–338, 358 McMillin, James, 12, 29 Mel languages, 164, 194 Melo, Miguel Antônio de, 325 Memórias do distrito diamantino (Joaquim Felício dos Santos), 296 Mende names, 158–159, 160, 165, 167–168, 171. See also Names, African Mendes, António de Almeida, 13, 45, 63–94, 107 Menezes, César de, 101 Menezes, José César de, 100 Mettas-Daget catalogue, 10–11, 27 Meyer, Jean, 206 Middelburgse Commercie Compagnie, 230, 231 Middle Passage, 88, 133, 146, 147, 290, 340, 341 Migration rates, 42, 339, 343–344, 346, 352, 355, 356, 358 Miller, Joseph, 113, 115

Minas Gerais, 108, 109, 136–138, 298 Miranda, Francisco Teixeira, 322 Misevich, Philip, 45, 155–175 Molembo, 120, 139 Montserrat, 212, 213 Moráguez, Oscar Grandío, 45, 176–201 Mortality rates. See Vital rates Moura, Antônio Rodrigues de, 288–289 Mozambique, 188–189, 193 Mulattoes, 318, 321 Muslim slaves, 64–65 Names, African, 155, 157–160, 162. See also Mende names Los ñáñigos (Sosa), 179 Nantes, 122, 220–221 National participation in slave trade: comparison of, 40–41, 44, 82; and national flags, 5, 8, 12–13, 183, 215, 348, 349. See also Brazilian slave trade; British slave trade; Dutch slave trade; French slave trade; North American slave trade; Northern German slave trade; Portuguese slave trade; Spanish slave trade Navigation Act (1660), 218, 219 New Calabar, 189, 191 New Holland, 97–98, 108 Norman, Try, 168 Noronha, Antônio Manoel de, 319, 322 North American slave trade, 28–31, 32–33, 354 Northern German slave trade, 250–271; African origins of, 265, 266; Brandenburg-Prussia enterprise, 251–263, 266–267; Courland initiatives, 251–263; destinations of voyages, 265, 266; and Hanse towns, 251, 263–266; shipboard mortality, 256; with Spanish America, 257–258; and volume of slave arrivals, 261–262, 265 Oliveira, João de, 143 Oliveira Mendes, Luís Antônio de, 289

373

374

Index

Ortiz, Fernando, 178–179, 180 Ouidah, 190, 195, 222, 240, 242, 243, 256 Ovando, Nicolás de, 64, 72 Oyo Empire, 143, 190, 191 Pará, 19, 20 Pernambuco, slave trade to, 95–129; and African origins, 117–122; from Angola, 97, 98–99, 100, 101, 107, 110, 113, 115–116; and Dutch involvement, 112–113, 233, 237; estimates in TSTD2, 14–15, 96, 98, 100, 103–105, 108, 110–112, 127–129; and Luanda, 99, 107, 118, 120; and sugar production, 97–98, 106, 108, 109, 110, 111; and volume of slave arrivals, 96–117, 127–129. See also Brazilian slave trade Peru, 76, 86, 118 Phélypeaux, Louis, 355, 357 Pichardo, Esteban, 177–178 Pinelo, Antonio, 73 Pombal, Marquis of, 138 Pombalina Company, 19, 100, 101, 115 Ponta do Monserrat, 147, 148 Population growth, slave, 337–339, 343, 344, 358 Portobelo, 77, 86, 351 Portocarrero, Antonio, 73 Portugal, Dom Jorge de, 64 Portuguese slave trade: compared to British slave trade, 39; destinations of, 21; early decades of, 13–14; new information since 1999, 7–8, 11; and slave estimates by decade, 16–17 Postma, Johannes, 23, 229, 230, 231, 232 Prices, slave, 73, 292, 296, 298. See also Slaves Principe Island, 191 Pritchard, James, 27, 52, 205–227 Profits, slave trade, 292, 294 Providence Island, 23–24 Puerto Rico, 36, 349

Ratelband, Klaas, 230 Raule, Benjamin, 252, 253–254, 260 Rebelo, Manuel dos Anjos da Silva, 115 Recife, 97, 101, 115, 117, 120, 121, 122 Reers, Cornelis, 253 Reexports of slaves, 352–353, 354–355, 357, 358 Registers of Liberated Africans, 155 Reinel, Pedro Gomez, 76 Réis, Justiniano José dos, 317 Revised Transatlantic Slave Voyage Data Set. See TSTD2 Rhode Island, 28–30 Ribeiro, Alexandre Vieira, 15, 45, 130–154 Ribeiro, Luiz Gomes, 324 Richardson, David, 1–60, 205–227, 228–249, 316 Rio de Janeiro, slave trade in, 275–312; and African origins, 20–21, 281, 282, 283; from Angola, 20–21, 323; and Bahia, 137, 138; and colonial market, 293–299; and profitability, 291–293, 294, 298–299; and Rio-based traders, 287–288; from 1790–1830, 276–293; shipboard mortality, 290, 291; and slave families, 276, 299–309; and status of free men, 277–278, 279, 280; and trade imbalance, 293, 295–296, 297, 298; and trafficking, 299–309; in TSTD2, 20, 21, 275, 278; and volume of slave arrivals, 278, 280–281. See also Brazilian slave trade Rio Pongo, 161, 163, 165, 166, 167, 169, 170, 171 Royal African Company, 11, 24, 25, 219, 220, 262, 325, 353 Royal Asiento Company, 26 Rufisque, 83–84 Sá, Salvador Correia de, 140 Saco, José Antonio, 182 Saint-Domingue, 11, 138, 208, 213–214, 215, 259, 341, 353

Index

St. Eustatius, 213–214, 235, 236, 238, 351 St. Thomas, 31, 214, 254, 255, 256–257, 258–259 Salvador, 130, 131, 136, 139–140, 147, 148 Sampaio, José Botelho, 322 Sandoval, Alonso de, 107, 108 Santiago, 66, 70, 83, 84 Santo Domingo, 64, 65, 73, 74, 345, 353 Santos, Corcino Medeiros dos, 11 Santos, Joaquim Felício dos, 296 Santos, Remígio Luiz dos, 324 São Jorge da Mina, 66, 74, 140 São Tomé: asientos for, 76; and direct exportation of slaves, 66, 191; and first transatlantic slave voyage, 64; and mortality rates, 88; as pre-1570 slave center, 69; and Rio de Janeiro, 84; slave voyages to West Indies from, 71, 73 Saugera, Eric, 27 Sayller, Jeronimo, 73 Scarano, Francisco, 36 Scelle, Georges, 68 Schön, James Frederick, 167–168 Schück, Richard, 251 Schwartz, Stuart, 15, 106, 108, 134, 138 Senegambia, 84, 118, 171, 217, 220, 221, 222, 223 Sergipe, 134–135 Seville, 63, 72, 74, 83 Sherbro Island, 163, 167, 169, 171 Shipboard mortality: average rate, 10; and Bahian slave trade, 144–145, 146–148, 149; Curtin on, 4; and departure-arrival estimates, 341; and French slave trade, 217–218; and North American slave trade, 354; and northern German slave trade, 256; and Rio de Janeiro, 290, 291; and Spanish slave trade, 88–89. See also Below-deck conditions Sierra Leone: and African languages, 164; and African names, 158, 159; and British suppression of slave trade,

160–161; and courts of mixed commission, 157; and Cuban slave trade, 186, 187, 194; and nineteenth-century slave trade, 160–162 Silva, Alfonso Franco, 72 Silva, Daniel Barros Domingues da, 95–129 Size of slave trade, estimated, 37–38, 39, 40–41, 42, 43, 45, 53, 67. See also Departure-arrival estimates (slaves); specific locations for volume of slave arrivals Slave Coast, 221–222, 239, 240, 243, 254, 255 Slave rebellions. See Slave revolts Slave revolts, 72, 138, 140, 318, 327, 341 Slaves. See Below-deck conditions; Child slaves; Female slaves; Indigenous slaves; Marriage, slave; Migration rates; Muslim slaves; Names, African; Population growth, slave; Reexports of slaves; Shipboard mortality; Slave revolts; Vital rates Smuggling. See Illegal slave trading Soares, Manuel, 134 Sosa, Enrique, 179 South Carolina, 28, 31, 354 Souza Portella, Antônio de, 289 Spanish Americas, slave trade to, 63–94; and African origins, 83, 84–85; and Angola, 66, 67, 85; destinations of voyages, 71, 73, 86–87, 93–94; early decades, 13–14; and government regulation, 86, 88; and intra-American slave trade, 347–349; literature review, 67–68; and new data, 68–70; overview, 63–67; and period of individual licenses (to 1595), 70–75; and Portuguese asiento period (1595–1640), 68, 70, 75–79, 80–81, 85; and post-1640 period, 82–83; and sources of demand for African slaves, 64–65; in TSTD2, 69, 77, 81, 82; and volume of slave arrivals, 74, 77, 80, 81–83, 86, 89. See also Spanish slave trade

375

376

Index

Spanish slave trade, 34–37, 38, 88–89. See also Spanish Americas, slave trade to Stein, Robert, 357 Sugar mills. See Engenhos Sugar production: in Bahia, 133, 135, 136, 138, 149; and Dutch slave trade, 244–245; and French slave trade, 218, 219; and Pernambuco, 97–98, 106, 108, 109, 110, 111; and Portuguese, 134; and slave demographics, 337; as source of demand for slaves, 65, 66 Suppression of slave trade, British, 138–139, 140, 144–145, 160–161, 167, 186–187, 319–320, 325. See also Abolition of slave trade; British slave trade Suriname, 213, 229, 235, 236, 242–243, 345 Swedish slave trade, 31, 34 Tams, George, 319, 324 Teixeira, Manoel Joaquim, 317–318 Tobacco, 142, 285 Tobago, 251, 336 Toledo, María de, 64 Torre, José María de la, 177 Torre do Tombo, 20, 101, 105 Trading Company of the Coast of Guinea. See Brandenburg African Company Transatlantic Slave Trade: A Database on CD-ROM (1999). See TSTD1 The Transatlantic Slave Trade: An Expanded and Online Database. See TSTD2 Treaties, anti-slave-trade, 121, 139, 157, 160, 184, 187, 192, 280 TSTD1: and Bahia, 133; and Cuban slave trade, 182, 183; and Dutch slave trade, 231; and French slave trade, 207; and overall assessment of slave trade, 5–6; scope since 1999, 6–8; and Spanish Americas, 69 TSTD2: and Amazonia, 19–20; and Angola, 314, 318–319, 321, 323, 324, 325–327; and Bahia, 15, 18–19, 133,

140; and British slave trade, 24–26; completeness of data in, 8–9, 11, 43, 44; and Danish slave trade, 34; and departure-arrival estimates, 340, 341, 343, 357; and Dutch slave trade, 22–23, 231, 233, 242; and French slave trade, 27, 214, 226–227; inferences and estimations assumed in, 9–10; and North American slave trade, 28–29, 31; and Pernambuco, 14–15, 96, 98, 100, 103–105, 108, 110–112, 127–129; and Portuguese/Brazilian slave trade, 16–17; and Rio de Janeiro, 20, 21, 275, 278; and Saint-Domingue, 11; and size of slave trade, 37, 40–41; and Spanish Americas, 69, 77, 81, 82; and Spanish slave trade, 35, 36–37 United States. See North American slave trade Upper Guinea Coast, 155–175; and country of origin, 157, 162–164, 171; and distances between enslavement and embarkation, 169–170, 171; geographic origins of recaptives from, 164–170; and importance to slave trade, 76; map of, 156; and Sierra Leonean slave trade in nineteenth century, 160–162; and slave registers, 155, 157–158 Valdés, Rafael López, 179, 180 Van Belle, Josua, 254 Van Belle, Pedro, 254, 258, 259–260, 263 Van Dantzig, Albert, 230 Van den Boogaart, Ernst, 98, 99, 229, 231 Van Pere, Abraham, 254 Ventura, Maria da Graça Mateus, 68 Vera Cruz, 70, 76, 77, 85, 86 Verger, Pierre, 18, 131, 139, 140 Vianna, Bernardo Lourenço, 288–289 Vila Vilar, Enriqueta, 13, 67–68, 69, 70, 76–77, 80–82, 83 Vili, 192, 242, 315

Index

Vital rates, 216, 339, 341, 345, 346 Vivaldo, Lorenzo, 73 Vos, Jelmer, 23, 52, 228–249 Walsh, Lorena, 354 Ward, John, 337 Wätjen, Hermann, 98 Watson, John, 145 Weindl, Andrea, 34, 52, 214, 250–271 Welsers, 73

West India Company, first, 232 West India Company, second, 22, 23, 230–231, 235–236, 253, 254, 262 West Indies, 71, 355, 357 West-Indische Compagnie. See West India Company, second WIC. See West India Company Windward Coast, 189, 240, 242, 243 Yoruba, 182, 189, 190–191, 194

377