Digitized by the Internet Archive

in 2017 with funding from

Kahle/Austin Foundation

https://archive.org/details/isbn_9780714643908
ROUTES TO SLAVERY
Studies in Slave and Post-Slave Societies and Cultures

Co-Editors: GAD HEUMAN

JAMES WALVIN

THE SLAVES’ ECONOMY

Independent Production b\ Slaves in the Americas
Edited by IRA BERLIN and PHILIP D. MORGAN
THE ECONOMICS OF THE INDIAN OCEAN SLAVE TRADE IN THE
NINETEENTH CENTURY
Edited by WILLIAM GERVASE CLARENCE-SMITH

OUT OF THE HOUSE OF BONDAGE

Runaways, Resistance and Marronage in Africa and the New World
Edited by GAD HEUMAN

THE BUSINESS OF ABOLISHING THE BRITISH SLAVE TRADE, 1783-1807

JUDITH JENNINGS

AGAINST THE ODDS

Free Blacks in the Slave Societies of the Americas
Edited by JANE LANDERS

UNFREE LABOUR IN THE DEVELOPMENT OF THE ATLANTIC WORLD

Edited by PAUL E. LOVEJOY and NICHOLAS ROGERS

RECONSTRUCTING THE BLACK PAST

Blacks in Britain, 1780-1830
NORMA MYERS
SMALL ISLANDS, LARGE QUESTIONS
Society, Culture and Resistance in the Post- Emancipation Caribbean
Edited by KAREN FOG OLWIG

THE HUMAN COMMODITY

Perspectives on the Trans-Saharan Slave Trade
Edited by ELIZABETH SAVAGE

THE WAGES OF SLAVERY

From Chattel Slavery' to Wage Labour in Africa, the Caribbean and England
Edited by MICHAEL TWADDLE
 ROUTES TO SLAVERY
Direction, Ethnicity and Mortality
the Transatlantic Slave Trade

Edited by
David Eltis and David Richardson

FRANK CASS
LONDON PORTLAND, OR
•
 First published in Great Britain by
FRANK CASS & CO LTD
Newbury House, 900 Eastern Avenue
London IG2 7HH, England

and in the United States by

FRANK CASS
do ISBS
5804 N.E. Hassalo Street, Portland, Oregon 97213-3644

Copyright © 1997 Frank Cass & Co. Ltd.

Library ot Congress Cataloging-in-Publication Data

A catalog record of this book is available from the

Library of Congress.

British Library Cataloguing in Publication Data

A catalogue record of this book is available from the

British Library.

ISBN 0-7146-4820-5 (hb)

0-7146-4390-4 (pb)

This group of studies first appeared in a

Special Issue on ‘Direction, Ethnicity and
Mortality in the Transatlantic Slave Trade’ in Slavery
and Abolition, Vol. 8, No. 1 1

published by Frank Cass & Co. Ltd.

All lights reserved. No part of this publication

may be reproduced, stored in a retrieval svstem
o> transmitted in any form, or by any means, electronic, mechanical,
photocopving. recording,
or othenvise. w ithout the prior permission
of Frank Cass and Company Limited.

Printed in Great Britain by

Antony Rowe Ltd.
 Contents

The ‘Numbers Game’ and Routes to Slavery David Eltis and

David Richardson 1

West Africa and the Transatlantic Slave Trade: David Eltis and
New Evidence of Long-Run Trends David Richardson 16

Long-Term Trends in African Mortality in Herbert S. Klein and

the Transatlantic Slave Trade Stanley L. Engerman 36

Crew Mortality in the Transatlantic Slave

Trade in the Eighteenth Century Stephen D. Behrendt 49

‘My own nation’: Igbo Exiles in the

Diaspora Douglas B. Chambers 72

‘Of a nation which others do not understand’

Bambara Slaves and African Ethnicity in
Colonial Louisiana, 1718-60 Peter Caron 98

The Cultural Implications of the Atlantic Slave

Trade: African Regional Origins, American
Destinations and New World Developments Philip D. Morgan 122

Notes on Contributors 146

Index 147

nil43006150651
380. 144 Rou
Routes to slavery :

direction, ethnicity, and

mortality in the
transatlantic slave trade

0
V

%
The ‘Numbers Game’ and Routes to Slavery

DAVID ELTIS and DAVID RICHARDSON

Despite a major research effort in the lastfew decades, less is known about
themovement of African peoples to the New World than the much smaller
movement of their European counterparts before the mid-nineteenth
century. Given that the record keepers were Europeans who regarded
Africans as outsiders, it is likely that we shall never have as much
information on the personal lives of individual Africans making the Atlantic
crossing as we do of Europeans. But on the identities of large groups
entering the African stream as well as the size and demographic
characteristics of these groups, the picture is much less discouraging.
Indeed, in a few years it may well be the case that in these areas, and in the
early modern period at least, we will actually know more about these
aspects of African than of European transatlantic migration. As knowledge
of the patterns of the trade is basic to evaluations of the cultural implications
of long-distance movements of people, this is an exciting prospect. One of
the developments that has made it possible is, of course, the computer
revolution and the related, but ultimately more important, explosion in

archival research that has occurred since the late 1960s.

Historians are sometimes prone to exaggerate the significance of
published works, but the largest single influence over the exploitation of the
archives was arguably the publication in 1969 of Philip Curtin’s Census of
the Atlantic slave trade.' It was a landmark in the historiography not Just of
and migration. Drawing
the slave trade but in the larger fields of slavery
almost exclusively on previously published work, Curtin provided the first
detailed assessment of the overall volume of the transatlantic traffic in
enslaved Africans between 1500 and 1867. His estimates of the trade - up
to embarked at the coast of Africa and 9.4 million
11.8 million slaves
arrivals in the Americas - was substantially lower than most of the figures
previously assumed by historians, some of which were several times greater
than those calculated by Curtin.- Curtin’s book provided, however, more
than a reassessment of the overall dimensions of the Atlantic slave trade,
valuable though that was. In the course of producing his census, he also
2 ROUTES TO SLAVERY
generated data on temporal changes in the scale of the trade in slaves; on
mortality levels of slaves in the Atlantic crossing or middle passage; on the
numbers of slaves carried by different national carriers; and on shifts in the
geographical distribution of slave departures from Africa and of slave
arrivals in the Americas. In each ot these areas, Curtin’s findings
represented a major advance on existing knowledge of the transatlantic
slave trade.
The radical nature of Curtin’s revision of the most frequently cited of the
earlier estimates of the magnitude of the Atlantic slave trade provoked a
lively and, at times, heated debate.’ Most discussion centred on the last
two
centuries ot the trade, for which records are most abundant and when the
movement ot African slaves across the Atlantic was unquestionably at its
height. Disagreements continue over estimates of the scale of slave
shipments by some countries.’ Consequently, the ‘numbers game’ relating
to the volume of the Atlantic slave trade is likely to remain a significant
historical industry for some time to come. The latest estimates tend,
nevertheless, to corroborate Curtin’s overall assessment of the trade, at least
for the period from 1650 to 1870, though they also suggest that he probably
overestimated slave shipments before 1700 and underestimated them in the
nineteenth century.^ On the basis of the most recent surveys, it appears that
some 10.1 million people
Africa for America in 1660-1867, most of
left

them carried in British, Portuguese, and French ships. This is close to

Curtin’s assessment which suggested that between 1650 and 1870 some
10.5 million entered the transatlantic traffic, with some 8.9 million
surviving the Atlantic crossing. Assuming that, at most, one million slaves
were shipped from Africa before 1650, then the most recent evidence
suggests that perhaps 11 million Africans were forced to leave their
homeland for America between 1500 and 1870.
Further refinements of estimates of the magnitude of the Atlantic slave
trade will doubtless occur. But an important by-product of these
efforts to
quantify the trade has been the discovery of new records in Europe and
America relating to the shipping
and sale of African slaves. Such records
have generally been regarded as the most reliable sources for gauging the
dimensions of the trade in slaves. The discovery and analysis of such
records has, therefore, been a major feature of debates since 1969
over the
volume of the slave trade.
A review of the shipping records unearthed since Curtin produced his
census is outside the scope of this paper, but it is important to note that
the
discovery of such records has resulted in the publication of several
compilations of voyage histories as well as the creation of various
unpublished data sets of voyages. Some of the latter have been lodged in
archives in Britain and the United States. Among the larger published
THE ‘NUMBERS GAME’ AND ROUTES TO SLAVERY 3

collections are Johannes Postma’s study of the Dutch slave trade; Jean
Mettas’s two-volume compilation of eighteenth-century French slaving
voyages; Serge Daget’s compilation of illegal French voyages after 1813;

Jay Coughtry’s study of Rhode Island slavers; and David Richardson’s four-
volume examination of the eighteenth-century Bristol slave trade.* In
addition, there are now available unpublished data sets relating to slave
voyages to colonial Virginia; voyages by Liverpool ships in 1744-86;
Portuguese, British, and other ships in 1790-1815; and ships of various
flags (or, in many cases, no known flag) in 1815-67. Other data sets have
been created for the British slave trade before 1714 and again from 1780 to
1807 as well as for the eighteenth-century slave trade of London and smaller
British ports such as Lancaster and Whitehaven.’ Much work still remains
to be done to trace voyages undertaken before 1660 and to fill gaps in our
knowledge of voyages in later years. But, compared with the time that
Curtin’s census appeared, there is now a wealth of information about
transatlantic slaving voyages in the period from 1660 to 1867. '

It sometimes suggested that recent quantitative approaches tend to

is

sanitize the slave trade and need to be balanced by placing a greater

emphasis on the personal experiences of the slaves themselves.* There are,
of course, autobiographies of African slaves, such as that of Olaudah
Equiano (or Gustavus Vassa) to whom one of the contributors to this volume
refers.^ Moreover, some historians have sought to draw individual and

collective portraits of the lives of slaves, whether during the course of their
enslavement or on American plantations.'® This will be greatly augmented
by the work of Paul Lovejoy." We welcome such research. But we also
believe that it is difficult to assess the significance or representativeness of
personal narratives or collective biographies, however detailed, without an
understanding of the overall movement of slaves of which these
individuals’ lives were a part. The reconstruction of slaving voyages and,
even more importantly, the creation of a single, consolidated data base of

voyages offers the best means available for charting the routes to slavery of
Africans forced into exile from their homelands. In other words, voyage
histories, when combined, represent a powerful tool for understanding the
African diaspora and the contribution of Africans to the creation and
development of the transatlantic world between 1500 and 1850.
Since 1992 several scholars have collaborated in seeking to build an
integrated and consolidated data set of transatlantic slaving voyages. Hosted
by the W.E.B. Du Bois Institute of Harvard University, the project has
brought together a number of published and unpublished data sets and,
through consultation of previously unused records, has also enhanced
existing sets and created additional ones. Details of the various data sets
included in the consolidated Du Bois Institute set are provided in Table 1.
 4 ROUTES TO SLAVERY
TABLE 1

STRUCTURE AND SOURCES OF DATA INTEGRATED INTO THE W E B DU BOIS INSTITUTE’S

TRANSATLANTIC SLAVE SHIP DATABASE

Source by Nation No. of Voyages Years Covered Collected by

Portuguese
(Arrivals in Spanish America) 571 1595-1640 Enriqueta Vila
Vilar
British 1,626 1644-1714 David Eltis
British/US C .870 1644-1808 David Richardson
(Bntish Stephen Behrendt
Amencan ports) Eltis
Dutch 1,211 1675-1792 Johannes Postma
Danish 169 1683-1792 S.Green-Pedersen
W. Westergaard
British 2,337 1698-1779 Richardson
(Liverpool) M M. Schofield
Katherine Beedham
Behrendt
British 2,200 1698-1807 Richardson
(Bristol)
British c 1,600 1700-1779
.
James A. Rawley
(London) Behrendt
Richardson

French 3,298 1707-1792 Jean Mettas

Serge Daget
Robert L. Stein
Various C .650 1707-1807 Elena de Studer
(Arrivals in
Colin Palmer
Spanish America)

British/US 945 1709-1808 Jay Coughtry

(Rhode Island Behrendt
American ports)
British C.450 1714-1777 Schofield
(English
Richardson
outports)
Portuguese
(Angola) 550 1723-1772 Herbert S. Klein
British 3,132 1780-1807 Behrendt
Various C.800 1790-1820 Klein
(Arrivals
in Havana)
Portuguese 374 1795-1811 Klein
(Arrivals in
Rio de Janeiro)
Various
(traffic to 5,378 1806-1867 Philip D. Curtin
Brazil,
Eltis
Cuba)
Klein
Joseph C. Miller
French 640 1814-1845 Daget
THE ‘NUMBERS GAME’ AND ROUTES TO SLAVERY 5

The shows that, with the exception of some 571 voyages which were
table
undertaken between 1595 and 1640, almost all the voyages included in the
Du Bois Institute set were undertaken between the mid-seventeenth century
and 1870. As yet, therefore, this data set contains little information about
the Atlantic slave trade during its first century, but it does cover the years
when the transatlantic traffic in slaves was at its peak. Overall, the set
currently includes records of over 26,000 slaving voyages or probably more
than two-thirds of those undertaken after 1650. This represents the largest
data set for any area of transatlantic migration or trade currently available
or, indeed, of any pre-nineteenth century migration anywhere. It is not
overstating its significance, therefore, to suggest that, when more widely
accessible, it will be relevant to the work of all scholars interested in the
African diaspora and the African-American heritage. It will, moreover, help
to illuminate the human experience of the victims of this tragic chapter in
modem history.
The Du Bois Institute data set comprises over 160 fields of information

about slaving voyages. These cover the geography of voyages, including

ports of provenance of ships, trading places in Africa and the Americas, and
ports of return; dates of departure of ships and their time schedules for
different stages of their voyage; details of the constmction, registration,
owners, size, rig, crew and armaments of ships; and the number, age and
gender of slaves loaded in Africa, of those lost during the voyage, and of
those delivered or sold in the Americas. In most -cases, the range of data
available for individual voyages remains incomplete, with information on
ships, their ports ofprovenance and trading places tending to be rather more
abundant than that on the numbers of slaves carried and lost. Even so, the
data set currently contains information about slaves shipped in over 17,000
voyages and slave mortality in some 6,100 voyages. These are very
substantial foundations upon which to project long-run trends in the

movement and mortality of slaves in the transatlantic traffic.

Several studies of the slave trade using data from the Du Bois Institute
sethave already appeared. One is an analysis of shipping productivity
between 1673 and the mid-nineteenth century which reveals that long
swings occurred in the productivity of shipping in the trade and argues that
‘these can only be explained by African forces operating through loading
times’.'- It also suggests that there was no significant improvement in

shipping productivity before the last quarter of the eighteenth century but
that ‘efficiency gains’ in the nineteenth century ‘were very strong’;'^ this
was presumably reflected in a relative decline of the transoceanic costs of
transporting slaves. The impact of these productivity trends on trading
patterns remains to be fully explored, but as children tended to fetch lower
prices than adult slaves in the Americas, any significant and sustained
6 ROUTES TO SLAVERY
decline shipping costs must have encouraged carriers to ship more
in

children. This finding is consistent with the increase in proportions of

children entering the transatlantic traffic after ISOO.'"’
A second and, in the context of this volume, even more important paper
that uses the Du
Bois Institute set explored the geography of transatlantic
slaving between 1662 and 1867.'^ This provides breakdowns of ports of
provenance of ships and distributions of slaving activity in Africa and the
Americas. Drawing on evidence of African departure points for over 12,500
voyages, of American destinations for over 15,000 voyages, and of
combined African and American data for over 11,500 voyages, the paper
suggested regional patterns of slave embarkations in Africa and arrivals in
the Americas that are not radically different from those ot Curtin, although
he confined his analysis of patterns of slaves departures to the period after
1700.
From the Du Bois Institute data,
appears that between 1662 and 1867
it

over four out of five slaves left Africa from just four regions - the Gold
Coast, the Bights of Benin and Biafra, and West-Central Africa. The data
also permit even more disaggregated analysis of departures by providing
details of shipments by ports. This shows that. Just as slave ship departures
from Europe and the Americas were concentrated at relatively few ports, so
large proportions of the slaves exported
were shipped at a small number of
outlets in Africa. Prominent among these were Cape Coast Castle
and
Anomabu on the Gold Coastj Whydah in the Bight of Benin; Bonny and
Calabar Bight of Biafra; and Cabinda, Benguela, and Luanda in West-
in the

Central Africa. The history of those ports in this period was clearly
dominated by slave trafficking. On the American side, arrivals of slaves
were more dispersed geographically; nevertheless, some 40 per cent of
slaves landed in Brazil, over 20 per cent in the British Caribbean,
about 17
per cent in the French West Indies, and over 10 per cent in the
Spanish
islands. Overall, this preliminary analysis suggested that over
per cent
90 of
the slaves arriving in the Americas between 1662 and 1867 disembarked at
places in the Caribbean islands and Brazil. Even though the United States
had a substantial slave population by 1810, arrivals in mainland North
America constituted probably no more than 7 per cent of the total. The main
findings of this paper related, however, not to Africa and
America
separately, but rather to the intensity of the links between particular regions
in the two continents.
Some attempt to trace connections between regions of departure and
arrival of slaves was made by Curtin,'^ but the Du
Bois Institute data allow
one to trace in detail the principal routes to New World slavery followed by
Africans. Some further refinement of the findings reported in the
1995
paper is inevitable as more data are added to the Du Bois Institute set.
THE ‘NUMBERS GAME’ AND ROUTES TO SLAVERY 7

Nevertheless, important links between Africa and the Americas were

uncovered by the information then available. For instance, with the
exception of Bahia, which largely drew on slaves from the Bight of Benin,
slaves from West-Central Africa dominated arrivals in South America. In
the West Indies, St Domingue also depended heavily on slaves from south
of the equator, but elsewhere in the Caribbean slaves from West Africa
dominated Moreover, within the Caribbean, the proportions of
arrivals.

slaves entering particular colonies or groups of colonies from individual

sub-regions of West Africa were uneven. Thus, the Gold Coast seems to
have supplied a disproportionate share of slaves arriving at Barbados,
Surinam, and the Guyanas, while the Bight of Benin played a similar role in

the case of the French islands outside St Domingue. In the other major
Caribbean islands - Jamaica and the British Leeward Islands - the Bight of

Biafra was easily the largest source of supply of slaves. Significantly, it

appears that slaves from Senegambia and Sierra Leone represented a

relatively small share of arrivals at the major American destinations. As the
middle passage from these regions to the West Indies was shorter than that
from the four regions to the south,’’ this is striking and, together with the
Bight of Benin’s dominance of slave arrivals at Bahia, should caution one
against assuming that geography and transport costs were of overwhelming
importance in shaping the patterns of transatlantic slave routes.
Like the papers just outlined, the essays volume rely in varying
in this

degrees on recent research on slave ship records and quantitative analyses

of the slave trade. Four of the essays, directly or indirectly, use data from
the Du Bois Institute set. Of the other two, one examines evidence on
French slaving voyages collected by Mettas which is included in the Du
Bois Institute set, while the other uses recently published quantitative
studies of the British slave trade to estimate the number of Igbo forced into
the transatlantic traffic after 1700. The essays here build, therefore, on
modem quantitative approaches to the study of the transatlantic slave trade,
though they also seek to relate more qualitative evidence to quantitative

findings. More specifically, they underline the potential of the Du Bois

and fluctuations in slave shipments
Institute data set for investigating trends
and the patterns of connections between Africa and the Americas in the two
centuries after 1660. They also highlight the human costs of sustaining the
slave traffic and expose some problems in tracing the ethnicity of slaves and
evaluating its role in the creation of slave identities and communities in the

Americas.
The essay by Eltis and Richardson extends the analysis of regional
breakdowns and intercontinental links in slave labour Hows which, together
with Behrendt, they first developed in 1995."’ Their paper concentrates on
the Gold Coast, Bight of Benin, and Bight of Biafra, which dominated the
8 ROUTES TO SLAVERY
West Atrican trade, and explores long-run trends in slave shipments by
regions, variations in influence over regional trade by national carriers, and
patterns ot slave arrivals in the Americas. The evidence suggests that before
1830 slave shipments from the Bight of Benin were closely linked to
demand tor labour in Bahia andjnvariably exceeded those from the Gold
Coast, while shipments trom the Bight of Biafra rose in tandem with
increasing British involvement in the trade and, at the height of the British
trade in 1780-1807, outstripped shipments from its neighbour. As British
traders also dominated the Gold Coast, it is perhaps not surprising to find
that, while most slaves from the Bight of Benin travelled in Luso-Brazilian
ships to Bahia, the route to slavery for those leaving the Gold Coast and
Bight of Biafra more often than not led to the British Caribbean before
British abolition in 1807. Thereafter, shipments of slaves from the Gold
Coast collapsed, but Biatran slaves began to be shipped in greater numbers
to the French islands and Cuba. In the latter case, they were joined by slaves
trom the Bight of Benin after 1830. Even though links between West
African regions and American markets for slaves varied through time, Eltis
and Richardson conclude that the distribution of peoples from West Africa
in the Americas was far from random. Given the mercantilist framework
within which Atlantic trade developed before 1815, it was, in fact, largely
shaped by patterns of European colonial control in the Americas and by the
influence of national carriers over slave shipments from regions in West
Africa. What determined the latter remains an issue for further
investigation.
The Du Bois Institute data set not only allows us to begin to identify the
principal transatlantic slave routes, but also offers closer investigation
of the
experiences ot slaves in the Atlantic crossing or so-called ‘middle passage’.
This aspect of the slave trade has, quite rightly, consistently attracted
attention since the late eighteenth century when the British Parliament
first
debated the appalling conditions endured by enslaved Africans in the
Atlantic crossing.For many, the term middle passage has become
synonymous with the cruelty and inhumanity of the traffic. Modern
quantitative study of the middle passage was initiated by Curtin who
relied
heavily on data for eighteenth-century Nantes ships and figures
generated
by the British Foreign Office for 1817-4^3 in order to explore trends in
shipboard mortality of slaves.*'’ Since Curtin’s work, large amounts of
initial
new data have been unearthed, with the result that the Du Bois Institute data
set now contains vastly more evidence on slave
mortality than was available
to Curtin.^'

The Du Bois Institute data

underpin the essay on slave mortality trends
by Herbert Klein and Stanley Engerman in this volume. The
authors are,
quite appropriately, cautious about trying to quantify
overall mortality in the
THE ‘NUMBERS GAME’ AND ROUTES TO SLAVERY 9

transatlantic slave trade. In part, this reflects the continuing incompleteness

of data on shipboard slave mortality as a result of gaps in the records and
inadequate knowledge of ships lost at sea, but it also reflects problems in

defining the middle passage and in uncovering data on those parts of the
enslavement experience of Africans that might legitimately be embraced by
the concept. In particular, as Klein and Engerman remind us, we know very
little about mortality at the point of enslavement or in transit to the African
coast, but there are suggestions that those who boarded ship may have been
only half of those originally enslaved. This broader interpretation of the
middle passage and associated slave mortality has implications for
comparisons sometimes made between slave and crew mortality in
that are

the slave trade. Even if we ignore this, however, and simply focus on slave
mortality on board ship, it appears that over the three and a half centuries of
the transatlantic slave trade, perhaps 15 per cent (or over 1.5 million) of
those who embarked at the African coast died during the Atlantic crossing.
At the peak of the trade in 176()-1810 losses of slaves on the Atlantic
voyage perhaps averaged 6,000-8,000 a year.^^ Clearly, for a large number
of those bound for sale in the Americas - the great majority, it should be
noted, aged under twenty-five - the route to slavery ended either before
leaving the African coast or in mid-ocean.
As far as trends in slave mortality on board ship are concerned, Klein
and Engerman emphasize that, regardless of trading places, voyage lengths
and other dominant feature of the middle passage was a wide
factors, the
distribution of mortality rates per voyage as measured by the number of
slaves who
died in the crossing as a proportion of those loaded at the
African coast. Mortality rates on slave ships were highly unpredictable
between one voyage and the next and added to the uncertainties surrounded
to the trade. They also observe, however, that slave mortality rates tended
to decline in the long run, that mortality varied by region of trade in Africa
- ships from the Bight of Biafra having the worst record - and that, in the
late eighteenth century, the British appear to have been the most efficient in

keeping slaves alive. The last finding may partly be ascribed to

Parliamentary measures after 1788 to regulate slave-carrying and improve
conditions on board ship. But the general trend in mortality seems to have
reflected, in part at least, a more widespread capacity among carriers to shift
the overall distribution of mortality rates by reducing the incidence of
shipboard epidemics through technical and other changes. similar pictureA
emerges from Stephen Behrendt’s study in this volume of the pattern of
crew mortality on board eighteenth-century British slave ships, though, as
he observes, losses of crew on these ships tended to be higher than those on
their French counterparts. At this stage of his research, Behrendt is unable
to offer a considered explanation of this, but it is possible that it was
 10 ROUTES TO SLAVERY
connected with higher concentrations of British trade at the Bight of Biafra.
Despite the general tendency of mortality to fall, comparisons with other
contemporary oceanic migrations suggest that mortality rates on board slave
ships - both Behrendt so graphically shows, for crews -
for slaves and, as
remained very high indeed throughout the history of the transatlantic slave
trade. In this respect, the
journey to the Americas was always a perilous one,
whether tor the Atrican-born victims of the trade or the European and
American crews who manned the ships that carried them. The contribution
that increased contemporary knowledge of the extraordinary waste of life
associated with the slave trade made to its eventual abolition in Britain in
1807 remains a source of debate.
Another issue that excites controversy is how the slaves who survived
the middle passage and their subsequent ‘seasoning’ in the Americas
managed to rebuild their lives. It is widely recognized that, under plantation
regimes, the obstacles that slaves faced in seeking to establish their
own
identity or to shape the communities in which they worked were
formidable.
Some have noted that within most plantation systems opportunities existed
tor slaves to develop independent economic activities,-^
while others,
including John Thornton and Gwendolyn Hall, have argued that slaves
drew
on the language and culture ot their homelands in seeking to adjust to and
intluence their new social environments.--' In the latter case, it is argued,
ethnicity was especially important as slave populations in many New World
societies seem to have been drawn disproportionately from a few regions or
ethnic groups in Africa. In this scenario, therefore, the middle
passage was
more than a route to slavery that separated Africans from their past; it was
a series of cultural highways that linked the history of transatlantic slave
societies to the history of specific peoples of West and West-Central Africa.
Someevidence on the ethnicity of slaves has been published.-- But we
have, as yet, no overall picture of the movements of ethnic
groups from
Africa to the Americas. However, the patterns of transatlantic links
between
African and American regions revealed by Eltis and Richardson
appear to
be consistent with claims that there may have been concentrations of people
of particular African ethnic groups in the Americas. Despite such evidence,
the essays in this volume by Douglas Chambers, Peter Caron and Philip
Morgan, all of which are concerned with issues relating to slave
ethnicity,
offer, on balance, no more than qualified support,
at best, to the claims made
by Thornton and others.
On the surface. Chambers appears provide strong support for the case
to
that ethnicity had an important influence on the life
of enslaved Africans in
the Americas. Concentrating on the Bight of Biafra,
Chambers argues that
most of the slaves shipped from the region were Igbo-speaking
and^because
they were largely taken by the British, they tended to
be landed in British
THE ‘NUMBERS GAME’ AND ROUTES TO SLAVERY 11

American colonies. Within the British Americas, he suggests, particularly

large numbers of Igbo-speaking slaves were to be found in Jamaica, the
Leeward Islands and Virginia, though he also notes that, especially in the
late eighteenth century, Igbo arrived in significant numbers in the French

island of St Domingue. Seeking to investigate Igbo influences on the

cultural life of slave societies. Chambers then examines evidence of
‘Igboisms’, including customs, diet, language and masquerades, on which,
he argues, Biafran slaves ‘drew in order to adapt to slavery’. In the course
of his analysis, he challenges claims that slaves from the Gold Coast,
notably Akan groups, had a dominant influence on Jamaican slave
communities, arguing that important local institutions and practices such as
jonkonii and obeah, which are sometimes attributed to Akan connections,
were probably of Igbo origin. Although his argument seems broadly in
sympathy with recent claims about the influence of ethnicity and ancestral
traditions on slave life in the Americas, Chambers acknowledges that
evidence of Igboisms is widely dispersed throughout British America and is
to be found in places where, under slavery, concentrations of Igbo-speakers
were probably low. This lack of symmetry between the intensity of flows of
slaves between regions in Africa and the Americas and cultural carryovers
is also highlighted by Morgan. Perhaps even more significantly. Chambers

argues that ethno-genesis or the creation of ethnic identities, Igbo or

otherwise, may have been largely a product of the trade in slaves itself, as
confinement on board ship and the harshness of life on plantations
encouraged slaves to group together to confront the problems arising from
their enslavement. If this was so - and, as Morgan reminds us, recent work
suggests that in Africa ethnic affiliations may have had only marginal
significance - it appears that the concept of ‘ethnicity’ was rather more
pliable than has sometimes been assumed.
The last point is addressed more fully by Caron in his essay on slaves in

Louisiana in the first half of the eighteenth century. Caron focuses on slaves
from Senegambia who, it seems, constituted about a half of all arrivals in
the colony in 1718-31 and on the use by Louisiana officials of the term
‘Bambara’ to describe slaves from the region. Some scholars have assumed
that such official labels were accurate and that the slaves so described were
‘true’ Bambara in the sense of being people from the Niger bend.-* Caron

reminds us, however, European concepts of ethnicity or nationality did

that

not correspond with those of Africans. Moreover, building on the work of

Curtin,-^ he notes that, in addition to describing the ethnic-linguistic group
of that name, ‘Bambara’ was a term used in Senegambia to describe, among
other things, captives from east of the River Senegal, slave soldiers, and
pagans or non-Muslims. It appears, then, that apparently precise ethnic
labels were, in Senegambia at least, highly flexible. At the same time, a
 12 ROUTES TO SLAVERY
close Study of the origins of slaves shipped from the region by the French
in the 1720s strongly suggests that the great majority probably came from
near the coast rather than from the deep interior or Niger bend. While some
of the slaves in Louisiana may indeed have been ethnic Bambara, it appears
that many of those described as Bambara were almost certainly not.
According to Caron, this does not mean that all evidence relating to slave
ethnicities should be disregarded but, rather, that historians should be more
fully aware than they sometimes have been of the context in which labels
are used or applied.
Caron’s reservations about the way evidence on ethnicity has been used
are endorsed by Morgan in the final essay in this volume. But whereas
Caron and Chambers focus tightly on particular segments of the
transatlantic slave trade, Morgan adopts a wide-angled lens and ranges
across the whole Atlantic basin, from Angola to Senegambia and from Rio
de Janeiro to Virginia. The picture he creates involves sharp contrasts. For
instance, he applauds Thornton for his vision of Atlantic history and follows
him in adopting a similar approach to the slave trade, but challenges his
depiction ot ethnic identities in Africa and dismisses his use of concepts of
nation and national loyalty as anachronistic. Similarly, he draws on the work
of Eltis and Richardson on transatlantic movements of slaves and accepts
that a majority of slaves from most regions were ‘funnelled ... to one region
in the Americas’, but goes on to contest their suggestion that the picture of
a confusing mix of African cultures with all the attendant barriers to
establishing African carryovers in the New World needs revising. Fie argues
that in Africa there was
mixing of slaves as the hinterlands of the leading
a
ports and the trading networks that served them expanded. Further mixing
is said to have occurred in the Americas as slaves were often moved
on from
their original landing places. Despite continuities in Bows of slaves between
specific African departure and American arrival points, therefore, Morgan
questions the view that the slave trade was non-random and rejects the
image of ‘homogeneous peoples being swept up on one side of the ocean
and set down en masse on the other’. As depicted by Morgan, the journey
to slavery for most Africans assumes a very different image, more
reminiscent in some respects of the ‘melting pot’ imagery of white
nineteenth-century migration to the United States, only in this case it was
heterogeneous groups of African peoples being thrown together in the
Americas and drawing on and modifying often very diverse individual and
collective experiences and skills in order to build new identities and
communities. In this process, African ethnicity is seen to have been, at best,
no more than one variable among many that helped to shape slave
communities, while the cultures that underpinned them are depicted as
having been fluid, permeable, pluralistic and syncretic. In the final analysis.
 .

THE ‘NUMBERS GAME’ AND ROUTES TO SLAVERY 13

therefore, we are left by Morgan with the impression that the African

heritage of slaves cannot be ignored but that the cultures of slave-based

communities were essentially new and owed more to the conditions that
Africans encountered in America than to their upbringing in Africa.
The essays in this volume do not provide an agreed set of answers to
questions relating to the movement of enslaved Africans to the Americas.
What they do suggest, however, is that the Du Bois Institute data set

promises to be a vital source for historians seeking to investigate the

structure, direction, and regional and temporal distributions of transatlantic
movements of enslaved Africans. The papers offer some preliminary
discussion of some of the patterns that are beginning to emerge from the
reconstruction of slaving voyages. The capacity to document trading links
between specific African and American regions in far more detail than
hitherto represents one of the more important contributions of the Du Bois
Institute data to historical research on transatlantic slaving. But as several
papers in this volume attest, interpreting the impact of these linkages on the
evolution of the social history of the Atlantic world remains a matter for
debate. Resolving issues arising from patterns of slave movements will, of
course, inevitably generate new data and raise new questions relating to
transatlantic slavery.
Although we now know far more about the Atlantic slave trade than was
imaginable less than thirty years ago, it is important to remind ourselves that
routes to slavery for Africans led to places other than the Americas.
Knowledge of the flows of slaves across the Sahara, the Red Sea and Indian
Ocean has grown in tandem with that across the Atlantic.’’’ But it is unlikely
that we shall ever be able to document these slave trades - much less the
doubtless large-scale movements of slaves within Africa - as fully as their
Atlantic counterpart. In this respect, understanding of the African diaspora
is bound to remain very incomplete. Nevertheless, the studies contained in
this volume indicate that, by co-ordinating and integrating the data on slave

movements that are available, it is possible to shed new light on important

elements of the African diaspora and, more specifically, on the contribution
of Africans to the making of modern Atlantic history.

NOTES

1. Philip D. Curtin, The Atlantic Slave Trade: A Census (Madison, Wisconsin, 1969).
2. For a review of earlier estimates, see ibid., ch. 1

3. For reviews of the debates see Paul E. Lovejoy, ‘The Volume of the Atlantic Slave Trade; A
Synthesis’, Journal of African History, 23 (1982), pp. 473-501; idem, ‘The Impact of the
Atlantic Slave Trade on Africa; A Review of the Literature’, ibid., 30 (1989), pp. 365-94.
4. The volume of the British slave trade between 1655 and 1807 remains the source of much
debate; see Joseph E. Inikori, ‘The Volume of the Bntish Slave Trade, 1655-1807’, Cahiers
 14 ROUTES TO SLAVERY
d Etudes africaines, XXXII (4), 128 (1992), pp. 643-88; David Richardson and Stephen D.
Behrendt, ‘Inikori’s Odyssey; Measuring the British Slave Trade’, ibid., (2-3), 138-9 XXXV
(1995), pp.599-615; David Eltis, ‘The Volume and African Origins of
the British Slave
Trade before 1714’, ibid., XXXV (2-3), 138-9 (1995), pp.616-27.
5. David Richardson, ‘The Volume of the Slave Trade’, in Seymour Drescher and Stanley L.
Engerman (eds.). Encyclopedia of Slavery (New York, forthcoming).
6. Johannes M. Postma, The Dutch in the Atlantic Slave Trade 1600-1815 (Cambridge, 1990);
Jean Mettas, Repertoire des Expeditions Negrieres Frani^aises au XVlIl Siecle. eds. Serge
and Michele Daget (Paris, 1978-84); Serge Daget, Repertoire des Expeditions Negrieres
Francoises a la Trade lllegale (1814-1850) (Nantes, 1988); Jay Coughtry, The Notorious
Triangle: Rhode Island and the African Slave Trade, 7 700-/^07 (Philadelphia, 1981); David
Richardson (ed.), Bristol, Africa and the Eighteenth-Century Slave Trade to America, 4 vols.,
Bristol Record Society Publications, 1986-96.
7. Details of these data sets are given in Table 1 below.
8. James Walvin, Black Ivory: a History of British Slavery (London, 1992), pp.57-8, 317.
9. Olaudah Equiano, The Interesting Narrative of the Life of Olaudah Equiano, Written by
Himself ed. Robert J. Allison (1995 edition).
10. See, for example, Paul E. Lovejoy, ‘Background to Rebellion: The Ongins of Muslim Slaves
in Bahia’, in Paul E. Lovejoy and Nicholas Rogers (eds.). Unfree Labour in the Development

of the Atlantic World, special issue of Slavery and Abolition, 15, no.2 (1994), pp. 151-80;
Richard S. Dunn, ‘Sugar Production and Slave Women in Jamaica’, in Ira Berlin and Philip
D. Morgan (eds.). Cultivation and Culture: Labor and
the Shaping of Slave Life in the
Americas (Charlottesville, Virginia, 1993), pp. 49-72.
11. Lovejoy is currently seeking to create a biographical dictionary of slaves shipped from
Africa, as part of the UNESCO-sponsored Slave Route project, ‘The Development of an
African Diaspora: the Slave Trade of the “Nigerian” Hinterland, 1650-1900’, which Lovejoy
co-directs with Robin Law (University of Stirling) and Elisee Soumoni (National University
of Benin).
12. David Eltis and David Richardson, ‘Productivity in the Transatlantic Slave Trade’,
Explorations in Economic History, 32 (1995), pp. 465-84 (quotation, p.480).
13. Ibid., p. 480.
14. For a recent review of trends in the age and sex structure of the Atlantic slave trade see David
Eltis and Stanley L. Engerman, ‘Fluctuations in Sex and Age Ratios in the
Transatlantic
Slave Trade, 1663-1864’, Economic History Review, 46 (1993), pp. 308-23.
15. David Eltis, David Richardson and Stephen D. Behrendt, ‘The Structure of the Transatlantic
Slave Trade, 1595-1867’, in Henry Louis Gates Jr., Carl Pederson and Maria Diedrich (eds.).
Transatlantic Passages (forthcoming).
16. Curtin, Atlantic Slave Trade, pp.205-10.
17. Herbert Klein and Stanley L. Engerman, ‘Slave Mortality on British Ships 1791-1797’, in
S.
Roger Anstey and P.E.H. Hair (eds.), Liverpool, the African Slave Trade, and Abolition
(Liverpool, 1976), p. 116.
18. Richardson and Behrendt, ‘Structure’.
Eltis,
19. The term was, of course, used by Klein as the
title for one of his books on the slave trade;
Herbert S. Klein, The Middle Passage: Comparative Studies in the Atlantic
Slave Trade
(Princeton, 1978).
20. Curtin, Atlantic Slave Trade, pp. 275-82. Curtin used data for Nantes
carriers published by
Gaston Martin, Nantes au XVlll Siecle: L'Ere des Negriers (1714-1774) (Paris,
1931) and
Dieudonne Rinchon, Le Trafic Negrier, d’apres les Livres de Commerce du
Capitaine
Gantois Pierre-lgnace-Lievin van Alstein (Paris, 1938).
21 . Itshould be noted that major gaps still exist in the evidence on slave mortality, especially for
pre-1790 Portuguese ships.
22. The figure of over 1,5 million deaths assumes that about 11 million left Africa, as noted
earlier. The figure of 6,000-8,000 deaths a year in 1760-1810 assumes annual average
shipments of 60,000-80,000 of slaves and losses of 10 per cent during the
crossing. For
estimates of shipmentsin this period, see David Richardson, ‘Slave
Exports from West and
West-Central Afnca, 1700-1810’, Journal of African History, 30
(1989), p.lO.
THE ‘NUMBERS GAME’ AND ROUTES TO SLAVERY 15

23. Ira Berlin and Philip D. Morgan (eds.), The Slaves’ Economy: Independent Production by
Slaves in the Americas (London, 1991).
24. John Thornton, Africa and Africans in the Making of the Atlantic World, 1400-1680
(Cambridge, 1992); Gwendolyn Midlo Hall, Africans in Colonial Louisiana: the
Development of Afro-Creole Culture in the Eighteenth Century (Baton Rouge, Louisiana,
1992).
25. See, for example, Curtin, Atlantic Slave Trade, pp. 184-99; David Geggus, ‘Sex Ratio, Age
and Ethnicity in the Atlantic Slave Trade: Data from French Shipping and Plantation
Records’, Journal of African History, 30 (1989), pp. 23-44.
26. Hall, Africans in Colonial Louisiana, p.43.
27. Philip D. Curtin, Economic Change Precolonial Africa: Senegambia in the Era of the
in

Slave Trade (Madison, Wisconsin, 1975), pp. 178-79.

28. See, for example, Ralph A. Austen, ‘The Trans-Saharan Slave Trade: A Tentative Census’,
in Henry A. Gemery and Jan S. Hogendom (eds ). The Uncommon Market: Essays in the
Economic History of the Atlantic Slave Trade (New York, 1979), pp. 23-76; William Gervase
Clarence-Smith (ed.). The Economics of the Indian Ocean Slave Trade in the Nineteenth
Century (London, 1989); Elizabeth Savage (ed ). The Human Commodity': Perspectives on
the Trans-Saharan Slave Trade (London, 1992).
 West Africa and the Transatlantic Slave
Trade: New Evidence of Long-Run Trends

DAVID ELTIS and DAVID RICHARDSON

The transatlantic slave trade was the largest long-distance coerced migration
in history. On the African side, three regions -
Gold Coast, the Bight of
the
Benin and Bight ot Bialra - dominate the historiography. These areas tend
to be seen as the centre of gravity of the traffic not just from West Africa but
from the whole of sub-Saharan Africa, a situation captured by the
description of a section of the Bight of Benin as ‘the Slave Coast’ on most
maps printed before 1820. With the possible exception of Senegambia, the
history of these regions is better known than the rest of sub-Saharan Africa.
These regions also contained the largest population densities on the sub-
continent, and, consistent with this, contained
the greatest urban
development and, in the cases of the Gold Coast and Bight of Benin, the
most sophisticated state structures. All three regions tended to draw on
largely exclusive provenance zones, and indeed there is a reasonably
exclusive ethno-linguistic homogeneity within their hinterlands. Yet it is
still useful to treat these areas together. African slave traders were always
likely to view embarkation points in the east of the Bight of Benin and the
west ot the Bight of Biafra as alternate routes - particularly after Lagos
began its rise to major outlet status in the late eighteenth century - and
European traders often ensured that ties existed between ports in the eastern
Gold Coast and western Slave Coast.'
both the popular and scholarly worlds these heartland regions of West
In

Africa - defined here as the west coast of Africa from Cape Apollonia
to
Cape Lopez inclusive - have a higher profile than
do other regions. They
are seen as the major source of slaves entering the Americas;
they contain
most of the sites and monuments to the slave trade on the western African
littoral; and at the same time they are the subject
of the majority of the
monographs on the trade. The Du Bois Institute’s transatlantic slave trade
database permuts the reassessment ot the role of this region in the slave
trade
and uncovers many new detailed patterns in the movement of its
peoples
WEST AFRICA AND THE TRANSATLANTIC SLAVE TRADE 17

into the Atlantic trade. No long-run overview of slaving activity in any of

these regions has been undertaken since the late 1970s.^ It should be noted
that the data currently available are still incomplete as far as distinguishing
actual places of trade in Africa and America are concerned. Where such
data are not available, we have used intended ports of embarkation and
disembarkation of slaves instead. Further, for about 15 per cent of the
voyages, we lack data on numbers of slaves carried. In these cases we have
estimated the numbers of slaves on board from means computed from very
largenumbers of observations.
The database suggests that, overall. West Africa probably played a larger
role in the transatlantic traffic than non-specialists have appreciated.
Curtin’swork and the research that his 1969 book triggered suggested that
about two out of five slaves entering the transatlantic traffic did so in one of
the two Bights between 1700 and 1867 (no breakdown of African origins
before 1700 existing in the current literature). The new data set suggests a
somewhat higher ratio - perhaps 48 per cent of all slaves - for the period
1595-1867, though much of the higher ratio comes at the expense of
Senegambia, rather than regions south of the equator. West-Central Africa -
in other words sites near the mouth of the Zaire river and from Luanda,
Benguela and the adjacent regions - is still the dominant regional supplier
of slaves in the eighteenth and nineteenth centuries as a result of this

reassessment. This large region south of the equator supplied 40 per cent of
the slaves entering the Atlantic slave trade. However, while West-Central
Africa was far more important than any of the three individual West African
regions examined here, the size of the three regions together is comparable
to the West-Central African slave provenance zone as a whole. A West
AfricanAVest-Central African comparison may thus be more appropriate
than a strictly regional comparison.
Within West Africa as defined here, the new data suggest an increased
importance for the Gold Coast and the Bight of Benin relative to other West
African regions. Whereas earlier estimates gave the Gold Coast six per cent
of the total trade and the Bight of Benin about 17 per cent, the distribution
made possible by the new data set almost doubles the importance of the
Gold Coast and increases the Bight of Benin share to 22 per cent. The Bight
of Biafra’s share, by contrast, falls slightly to 15 per cent."* In sum, the data
set currently contains records of 7,085 voyages to these three regions and
implies that Just under 2.5 million left Africa for the Americas from points
between Cape Apollonia and Cape Lopez. All but one of these voyages
began between 1662 and 1863. There is no way of estimating with any
precision the proportion of all voyages to these regions that the current set
represents, but given the range of sources and the coverage of the major
national traders, we would be surprised if the current estimates comprised
 18 ROUTES TO SLAVERY
less than halt the Africans entering the traffic after the mid-seventeenth
century from the area between Cape Apollonia and Cape Lopez. With few
signs that gaps in the present data are ‘clumped’ by region or period, the
trends suggested here are likely to be reliable, whatever opinions might be
on the volume of the traffic — and therefore the share of the total represented
by the present data.

FIGURE 1

SLAVES LEAVING WEST AFRICA. 1595-1866

Somte. All diagrams and tables in this study have been compiled from information
in the Du
Bois Institute slave ship data set.

The temporal distribution of these departures is shown in Figure 1,

which presents estimates of slave departures grouped in broad 25-year
bands. Keeping in mind that the Du Bois Institute data set is a sample only,
and not a population, it is likely that departures from the Bight of Benin
exceeded 10,000 per year over a period of 125 years from 1687 to 1811.
Indeed, if we allow for bias in the declarations of Bahian slave ship captains
in theperiod after the British took suppressive action against the trade after
1815, this high volume era probably extended over a century and a half to
1830. The two peaks in the traffic from this region came at the very
beginning - 1687-1711 - and near the end - 178^1811 - of these 125
years, with a rather lower, but still high and steady volume of
departures in
between. More refined analysis, not shown here, does little to change
this
picture of a large and sustained trade in people in the sense that
annual
fluctuations in the volume of departures were relatively small.
European
wars and conflicts within the African supply networks might interrupt
business at particular ports and for short periods, but not until the
1830s
were these disruptions anything more than temporal^'.
WEST AFRICA AND THE TRANSATLANTIC SLAVE TRADE 19

A quite different pattern emerges from the Bight of Biafra, the second
most important West African region for slave departures. Here, ships and
numbers carried per ship were somewhat smaller than in the Bight of Benin,
313 slaves per ship compared to 387, and ship size averaging some 122 tons
in the Bight of Biafra and 192 in the Bight of Benin."* While slave trading
began early in the Niger delta, the traffic approached Slave Coast levels
only toward the mid-eighteenth century. Moreover, for a time in the late

seventeenth century, slave departures from the Bight of Biafra appear to

have lagged behind those from the Gold Coast. For fifty years after the
1760s, however, slave departures from the Bight of Biafra exceeded those
from both the Gold Coast and the Bight of Benin, though they also
subsequently fell off far more rapidly than in the Bight of Benin in the era
of suppression after 1815. This rapid decline of slave shipments from Niger
delta ports is well documented, with most of the few thousand leaving after
the early 1840s passing through points closer to Cape Lopez.^ As the British
were always by far the most important slave traders in Biafran ports, and as
British records are particularly well-represented in the data, it seems
unlikely that further improvements in the data set will change the shape of
the time profile for the region shown in Figure 1.

The Gold Coast was the least important of the three West African
regions examined here, with total departures running at half or less of those
in the other two regions. This finding is entirely consistent with recent
estimates of regional departures from Africa.* However, the temporal
distribution of this Gold Coast total is at odds with some recent estimates.
Figure 1 shows departures gradually doubling in the course of the
eighteenth century, as opposed to the currently accepted profile of a strong
increase into 1730s followed by a gradual decrease through to the 1780s.
Further contributions to the data set may yet clarify this discrepancy.
Slave traders from all parts of the Atlantic West Africa. Yet
world traded in

there is little doubt that the trading community was much more cosmopolitan
on the Slave Coast than on the Gold Coast or the Bight of Biafra. In brief,
Portuguese based in Brazil dominated trade in the Bight of Benin, with 55 per
cent of all voyages recorded there setting out from Bahia. On the Gold Coast
and in the Bight of Biafra, however, the British were even more dominant,
sending out 81 per cent of the voyages recorded as trading (or intending to
trade) on the Gold Coast and 78 per cent of those trading or intending to trade
between the main outlets of the Niger River and Cape Lopez. It might be
noted that one quarter of the British voyages to the Gold Coast is accounted
for by voyages from the British Americas, the Gold Coast being the African
mecca for the small rum-carrying ships from New England and the
Caribbean. Within these British-dominated regions, an interesting
specialization by port emerges, with the vast majority of slavers in the Bight
20 ROUTES TO SLAVERY
of Biafra at the peak of the region’s slave trade being from Liverpool, while
ships from Bristol and London dominated trade with the Gold Coast. After the
British, Dutch ports were responsible for the second largest number of
voyages to the Gold Coast, accounting for 10 per cent of voyages to the
region, while French ports played a similar role in the Bight of Biafra,
accounting for 12 per cent of voyages there. After Liverpool merchants
switched from slaves to palm oil in 1808, Cuban-based Spanish slave traders
became the largest group in the Bight of Biafra. On the Slave Coast, French
ships from a wide range of French ports were the second largest group of
traders after the Portuguese, accounting for 20 per cent of voyages, but Dutch,
English, and, later, Spanish (mainly Cuban) ships were also well represented.
Where did the slaves carried on these ships go after leaving West Africa?
The first point to note is that many did not reach the Americas. Mortality data
exist for 2,3 1 7 of the slaving voyages leaving West Africas. Of those leaving
the Gold Coast (n= 1,008), on board. The equivalent figure
14.3 percent died
for Slave Coast ships (n=760) was 14.6 per cent and for Bight of Biafra ships
(n=496) 1 9.2 per cent. Mortality on ships leaving all three regions was above
the average for all regions in the data set and was well above the 8.4 per cent
observed for the 1,916 voyages recorded as leaving West-Central Africa.
However, as always with slave ship mortality, the high variances observed -
typically approaching the size of the mean itself - makes firm conclusions
on interregional mortality variations difficult.’
The pattern of West African Americas was far from
arrivals in the
random. The major single destination of Gold Coast slaves was Jamaica
which accounted for 36 per cent of arrivals, but as many again went to other
parts of the British Americas. Thus, well over two-thirds of all slaves
leaving the Gold Coast went to the English-speaking New World, Barbados
being the major seventeenth-century destination and Jamaica dominating
the eighteenth century. Akan cultural prominence in Jamaica (including
Ahanta, Fanti, Akim and Asante peoples among others) is well rooted in the
slave trade according to this data Spanish America, the second most
set.
important destination for Gold Coast slaves after Jamaica, accounted for 15
per cent of departures.
By comparison
with the British American dominance of slave shipments
from the Gold Coast, six out of every ten slaves leaving the Bight of Benin
went to Bahia two out ot ten went to the French Americas, notably
in Brazil,

St Domingue, and one out of ten went to the British Caribbean. This broad
breakdown disguises, however, major shifts in time in the pattern of
destinations in the Americas of slaves shipped from the Bight of Benin. In
the late seventeenth century, the English Americas took the majority of
slaves leaving the Slave Coast. After 1700 the Reconcavo of Bahia in Brazil
and the French Caribbean replaced the English, with the former
WEST AFRICA AND THE TRANSATLANTIC SLAVE TRADE 21

predominating until the quarter century before 1791 when the French share
came closer to matching that of Bahia. From 1791 to 1830, however, Bahia
took 75 per cent of deportees from the Bight of Benin. Thereafter, Cuba
became more important as a market for slaves from the region, and for
fifteen years after 1851 was, indeed, the only market available to shippers
of slaves from the Bight of Benin. Compared to Bahia, Cuba’s links with the
Bight of Benin were relatively short-lived. It is, nevertheless, striking that
the strong Yoruba presence Cuba noted by some historians should be
in

based on such limited exposure to the region - at least compared to Cuban

ties to other African regions.
In the Bight of Biafra, a much simpler pattern emerges. In the Niger
delta ports, which supplied over 90 per cent of the slaves leaving the Bight
of Biafra, the first English ship is recorded as shipping slaves from Calabar
in 1662 and from this point to 1807, the British took perhaps 80 per cent of
all slaves from the region. Igbo and Ibibio peoples, who dominated slave
shipments from the region, went to all parts of the British Americas, with
Jamaica and the Leewards taking no less than three out of every five, and
Barbados and the British American mainlands playing a smaller role. This
region also supplied those non-British regions with which the British traded,
notably the Spanish Americas. The French Americas received about 10 per
cent of Biafran slaves. After the British ended their direct involvement in
the traffic, the Spanish Americas, particularly Cuba, continued to draw on
the Niger delta. For a decade and a half between 1814 and 1 830 the revived
and largely illicit French slave trade also drew on these outlets, delivering
the slaves mostly to Martinique and Guadeloupe. The absence of
Portuguese traders and the slight presence of the Dutch and the pre-1800
French in this area is striking.*

The major contribution of the Du Bois Institute data set, however, is to

provide more refined geographic and temporal analyses of the slave trade.
At this point in the project it is possible to identify the specific points of
embarkation (as opposed to regions of departure) of 28 per cent of those
leaving the Gold Coast, 37 per cent of those leaving the Bight of Benin, and
69 per cent of those leaving the Bight of Biafra. The relatively low
identification rate for the Gold Coast is accounted for by the limited range
of the coastline that comprises the region, or from which slaves
at least

could be expected. The same phenomenon in the Bight of Benin has a

different origin. A large part of this region’s data is drawn from passes for
ships leaving Bahia de Todos os Santos for Africa. These are held at the
Arquivo Publico da Bahia. Most of these ships simply declared their
destinations to be the ‘Mina Coast’, which, following an agreement with the
Dutch - the effective controllers of Portuguese access to this part of Africa
- meant the four ports of Grand Popo, Whydah, Jaquin and Apa. Unhappily,
22 ROUTES TO SLAVERY
the Brazilian passes do not allow breakdowns among these ports, although
Whydah was certainly the most important of the four ports to the Portuguese
from the 1680s, when the records begin, until the second quarter of the
nineteenth century, when Lagos emerged From the last
as a serious rival.
third of the eighteenth century, the designation ‘Mina’ came to include
additional ports, all east of Whydah - namely, Porto Novo, Badagry and
Lagos (also commonly known as Onim).^ After 1810, British records of the
Portuguese slave trade allow a more precise identification of African
embarkation points. A crude assumption that 75 per cent of those ships
designated for Mina in fact took in their slaves at Whydah would perhaps
not do too much violence to the historical reality, although it should be
noted that in some years departures from Whydah were very low.
Table 1 breakdown of departures from individual sites as far
presents a
as these can be identified in the Du Bois Institute set. Most of these are
shown on the accompanying map which dates from the last years of the
transatlantic traffic. The concentration of departures from a handful of sites
is particularly marked in the two Bight regions, but even on the Gold Coast

two embarkation points - Cape Coast Castle and Anomabu - account for 76
per cent of departures.'" On the Slave Coast, if we allow that at least 75 per
cent of slaves leaving from ‘Costa da Mina’ were taken from Whydah, then
the latter port by itself accounts for no less 695,000 slaves, or a little over
two-thirds of recorded departures from the Bight of Benin. Indeed, as the
current data set is clearly less than the total trade, probably well over one
million slaves left from Whydah, making it perhaps the single most
important oceanic outlet for slaves sub-Saharan Africa. Lagos, the second
in
most important Slave Coast port, lagged far behind, though in the closing
years of the trade it was probably the most important embarkation point in
either of the Bights, a fact accounting for the intense attention it was by then
receiving from the British - at this time seeking to end the traffic rather than
encourage it."

In the Bight of Biafra, the concentration of departures was almost as

severe, with almost 80 per cent ot all slaves leaving from just two outlets,
Bonny and Calabar, and a further 8 per cent leaving from a third port. New
Calabar. The Cameroons and, further south, Gabon, were by comparison of
only minor importance, accounting together for only one in ten of
departures from the region. There is no doubt that Whydah, Bonny and
Calabar were key trading sites not only in Africa, but in the whole Atlantic
slave trading system. By this we mean that overall each perhaps imported
more goods be exchanged for slaves than, with the possible exception of
to
Liverpool, any single port in Europe exported into the traffic, and Whydah
may have sent more slaves into the trade than any single port in the
Americas attracted in the way of arrivals.
WEST AFRICA AND THE TRANSATLANTIC SLAVE TRADE 23

TABLE I

SLAVE DEPARTURES (THOUSANDS) FROM OUTLETS IN THE GOLD COAST. BIGHT OF BENIN
AND BIGHT OF BIAFRA, 1662-1863 (LISTED FROM WEST TO EAST)

Slaves Ships
Gold Coast

Assinie 0.3 1

Axim 0.1 1

Quaqua 6.7 17
Kormantine 3.0 11

Elmina 2.4 9
Cape Coast Castle 38.9 191
Anomabu 53.9 196
Apam 1.2 5
Tantumquerry 0.4 1

Wiamba 0.2 1

Accra 7.1 24
Christiansborg 4.7 14
Alampo 2.0 6

Total identified 120.9 477

Bight of Benin

Keta 4.0 14
Popo
Little 6.1 18
Grand Popo 0.4 1

Popo (Unspecified) 11.3 36

Whydah 260.2 696
Jaquin (Offra) (Ardrah) 38.8 106
Apa (Epe) 9.3 25
Porto Novo 31.8 77
Badagry 21.0 57
Lagos 63.5 162
Costa da Mina 579.5 1,497
Benin 29.5 84

Total identified 1,055.5 2,773

Bight of Biafra

Rio Nun 2.2 9

Formosa 0.2 1

Rio Brass 5.0 16

New Calabar 46.3 150
Bonny 240.4 678
Andony 0.8 3
Calabar (or Old Calabar) 196.5 633
Bimbia 1.1 4
Cameroons 22.7 63
Cameroons River 4.8 16
Corisco 0.3 2
Gabon 25.3 84
Cape Lopez 6.6 26

Total identified 552.2 1,686

24 ROUTES TO SLAVERY
MAP 1

PORTS OF EMBARKATION OF SLAVES AT THE GOLD COAST, BIGHT OF BENIN

AND BIGHT OF BIAFRA, 1662-1863

Lopf/rii

Cap*
WEST AFRICA AND THE TRANSATLANTIC SLAVE TRADE 25

1843

1827

1791

1680-1863

1772

WHYDAH,

2
1754

FIGURE

DEPARTURES,

1736

SLAVE

ANNUAL

1720

1704

1688

9
26 ROUTES TO SLAVERY
TABLE 2
SLAVE DEPARTURES (THOUSANDS) FROM THE BIGHTS OF BENIN AND BIAFRA: MAJOR
OUTLETS FOR SLAVES BY TWENTY-FIVE YEAR INTERVALS, 1662-1863

Slaves (thousands) Ships

Whydah
1662-1686 7.9 23
1687-1711 44.2 102
1712-1736 75.6 223
1737-1761 40.8 108
1762-1786 41.7 107
1787-1811 5.8 21
1812-1836 24.2 70
1837-1866 19.9 42
Eastern Bight of Benin^
1712-1736 0.4 1

1737-1761 5.3 14
1762-1786 28.8 65
1787-1811 16.5 39
1812-1836 26.7 83
1837-1866 39.3 94
Bonny
171221736 7.3 27
1737-1761 63.6 176
1762-1786 96.8 253
1787-1811 22.4 66
1812-1836 47.8 149
1837-1866 2.5 7
Calabar
1662-1686 14.0 55
1687-1711 12.3 47
1712-1736 15.3 66
1737-1761 34.2 98
1762-1786 77.2 218
1787-1811 21.9 74
1812-1836 20.4 71
1837-1866 1.1 3
South-East Bight of Biafra^
1712-1736 0.5 1

1737-1762 6.5 18
1762-1786 30.9 90
1787-1811 4.7 24
1812-1836 11.6 42
1837-1863 6.6 20

a. includes Porto Novo, Badagry and Lagos combined.

b. includes Gabon, Bimbia, Cameroons, Corisco, Cape Lopez.

Anexamination ot the temporal distribution of departures at these major

ports is presented in Table 2. Gold Coast ports are excluded because none
of them rivalled the Bights ports listed in this table. Whydah’s greatest
relative importance came in the quarter century from 1712 to 1736. As a
result, peak departures from Whydah fall outside the periods for peak
WEST AFRICA AND THE TRANSATLANTIC SLAVE TRADE 27

departures from the Bight of Benin as a whole shown in Figure 1. Given the
attention Dahomey has received in the historiography, Figure 2 provides
annual totals for Whydah as well. Interestingly, the peak of departures
corresponds with the era of internal instability in Whydah and its conquest
by Dahomey in 1727. It most of the 1730s. Conquest by
also includes
Dahomey did bring about an immediate decline in departures, but this was
very short-lived, and while departures in the 1730s were below those of
earlier decades, they were still running at over 1,000 per year - higher than
the 1740s and much higher than, say, the twenty years from 1791 to 1810.
Cumulatively, more slaves left Whydah after the conquest than before, and
close inspection suggests that the decline in the 1730s may have been due
more to weakness in the Brazilian markets than to supply constraints. One
possible explanation for the quarter-century peak in the Whydah trade was
the fact that the period from 1712 to 1736 saw the least disruption from
European wars of any of the quarter centuries listed. The troughs in
departures in 1756-62, 1779-84 and 1790-1814 are clearly associated with
European conflict. Moreover, the decline of the slave trade of Dahomey in
the 1760s and 1770s suggested by Akinjogbin turns out scarcely to have
existed - except in comparison with the pre-1727 era.'*
Apart from European wars, fluctuations in departures from Whydah at
the end of the eighteenth century and thereafter are best explained by
increased competition from other outlets in the Bight of Benin. The shift
eastwards main flow of departures is very marked in the annual
in the

breakdowns of departures. Badagry was most important in the late 1770s,

and Porto Novo in the following decade. Lagos became of major
importance at the end of the Napoleonic Wars and, as already noted, from
1837 down to the British take-over, sent more slaves each year into the
Atlantic trade than did Whydah. It is important to note here that the key to
understanding this situation lies in Africa rather than Europe. The collapse
of the Oyo Empire and the flow of slaves associated with this event was
probably the major impulse behind the emergence of these eastern outlets.’’
Table 2 shows departures from the three ports grouped together. Yet the
most easterly of all, Benin, appears to have followed a path independent of
these events. It is at least clear that with nearly 30,000 departures in the
seventy years after 1721, Benin was not sealed off from the slave trade to
would have us believe. '* Outlets
quite the extent that earlier interpretations
between Whydah and Porto Novo - Jaquin and Apa - supplied slaves only
in the middle quarters of the eighteenth century, according to these data,
which capture nothing of slaving activity there before 1700. Likewise, the
western ports are underrepresented here. These were clearly more important
before 1750 than after, but so far the DuBois Institute set has provided a
disappointingly small amount of detail on Keta, the Popos and Awey.
28 ROUTES TO SLAVERY

1662-1838

CALABAR.

FIGURE

DEPARTURES,

SLAVE

ANNUAL
WEST AFRICA AND THE TRANSATLANTIC SLAVE TRADE 29

In the Bight of Biafra, it was the ports due east of the Niger delta which
entered the trade first. The first record of a departure at Calabar in the
present data set is for 1625, and from the early 1660s down to the late 1830s
(with the exception of the 1690s) departures were rarely far short of a
thousand a year. Figure 3 gives the annual trend line for this embarkation
point. By the 1780s and 1790s shipments from Calabar were three or four
times earlier levels, yet they still lagged behind the flow of slaves from the
River Bonny Bonny
at this time. entered the trade as early as the 1660s, one
of the Efik chiefs being named after an English slave captain of the period.
However, Bonny does not show up in the present data set until 1712, but by
the 1770s and 1780s was sending more people into the trade than Whydah
peak. This time profile also holds for the other major Niger delta outlet
at its

ot New Calabar as well as the less important outlets of Bimbia and

Cameroon to the east of Calabar and Gabon to the south. The most southerly
outlets, Corisco and Cape Lopez were significant only in the nineteenth
century. Unlike the Slave Coast ports which tended to take turns in stepping
up trading volumes, swings in departures at Biafran ports north of Corisco
tended to coincide with each other. During the years from the 1760s to the
1 790s, all these ports contributed to the record stream of departures from the
Bight of Biafra. It is much more difficult to link these trends with African
political and economic developments than it is in the Bight of Benin. But
perhaps there is The massive number of departures
a significant point here.
through the eastern rivers of the Niger delta and the Cross River was neither
the product nor the cause of the rise and fall of major new political
structures. The Kings of Bonny and Dukes of Calabar remained rulers of
what were essentially still trading enclaves at the end of the eighteenth
century and continued to be enclaves throughout, first, the slave trading and
palm oil era, and then after 1839, the palm oil era proper. Indeed, beyond
the growth of Duke Town at the expense of Old Town in the Cross River,
the rise of Ekpe, and the consolidation of the Aro trading network, it is
difficult to see how political structures and the distribution of power at
Bonny and Calabar differed in 1830 from what they had been in the 1660s
when slave trade volumes were very low or non-existent.'^
Some additional insights are possible if we take into account the age and
sex characteristics of the coerced migrants. It is now clear that over the last
two centuries of the slave trade the proportion of males and especially the
proportion of children in the traffic both increased. In addition, it is also
now clear that substantial variations existed in the sex and age profiles of
slaves from the different African coastal regions supplying the trade.
Generally, the Gold Coast, the Bight of Benin, and the Bight of Biafra
shared in the general increase in male and child ratios over two centuries,
but all three regions typically sent fewer males and fewer children into the
30 ROUTES TO SLAVERY

traffic than did other areas at any given period. Indeed, a larger proportion
of women was always to be observed among those leaving the Bight of
Biafra than from any other major embarkation zone in sub-Saharan Africa.'^
Table 3 provides additional detail to this broad picture by presenting age and
sex ratios among those leaving the major ports within the Bights broken
down by century - the data not yet being sufficiently abundant to allow
working with the quarter-century intervals used The table focuses
earlier.

once more on the three leading outlets in the regions - Whydah, Bonny, and
Calabar - and two composite categories comprising the three most
important embarkation points in eastern Bight of Benin - Badagry, Porto
Novo and Lagos - and the outlets south of Calabar - Gabon, Bimbia,
Cameroon, Corisco and Cape Lopez. Data for the Gold Coast as a whole are
added for comparative purposes. Recorded observations are rather small for
the grouped ports, but we expect to be able to increase these in the near
future.
Generally the extra detail provided by these more localised breakdowns
simply confirms the larger picture. All outlets shared in the trend toward
more males, fewer women, and more children through time. These broad
similarities across embarkation points should not be allowed to obscure
significant differences, however. In the Bight of Benin, a smaller proportion
of deportees from the eastern ports than from Whydah were boys and girls,

although the gap did narrow in the nineteenth century and for boys, at least,

becomes statistically insignificant. In addition, while shares of women

leaving Whydah on the one hand, and ports further east on the other, do not
seem markedly 40 per cent fall in the
different after controlling for time, the
proportion of women leaving the eastern Slave Coast between the
eighteenth and nineteenth centuries is rather larger than the 27 per cent fall

among those leaving Whydah. The more than threefold increase in the share
of children among slave departures from Whydah, albeit over two centuries,
is also worthy of note. The major political and military event in the
hinterland of the Bight of Benin outlets in the late eighteenth and early
nineteenth centuries was the collapse of Oyo. The resulting disturbances can
plausibly explain why relatively more slaves left the eastern ports over time,
but it is harder to make a link between the collapse of the Oyo empire and
a surge in the number of children entering the slave trade. Interestingly, the
pattern on the Gold Coast is closer to that displayed at Whydah - the point
on the Slave Coast in Table 3 most adjacent to the Gold Coast - than to that
at any other West African port. Nevertheless, the high ratio of women in the
seventeenth century, the subject of frequent comment by contemporaries, is

reminiscent of the Bight of Biafra.'’

In the Bight of Biafra, the minor ports south and east of Calabar had
smaller shares of women and more children among their deportees than did
 WEST AFRICA AND THE TRANSATLANTIC SLAVE TRADE 31

either of the major ports, though the differences are less marked in the
nineteenth century than earlier. The principal points of note in the Bight of
Biafra, however, relate to changes at the major ports through time. At
Bonny, the decline in the proportion of women entering the trade - from 40
to 14 per cent - is the largest observed for any outlet. The fourfold increase
in the ratio of children
both Calabar and Bonny, apparently confined to
at

the period from the eighteenth to the nineteenth century, is also dramatic.
Both the decline in the share of women and the rise in proportion of children
carried from east and south of the Bight of Biafra were much less than
among main ports of Bonny and Calabar, though we
the slaves leaving the
should take note of what is at present a relatively small number of
observations here.

TABLE 3
MEAN PERCENTAGES OF AFRICAN MALES, WOMEN, MEN, GIRLS AND BOYS CARRIED TO
THE AMERICAS FROM MAJOR OUTLETS IN THE BIGHTS, 1662-1863 (NUMBER OF
OBSERVATIONS CM THOUSANDS DM PARENTHESES)

Males Women Men Girls Boys

Gold Coast
1662-1700 56.0 41.0 49.4 3.0 6.5 (12.8)
1701-1809 66.6 28.3 54.6 5.5 11.7 (57.3)
1810-1863 - - - - -

Whydah
1662-1700 58.1 37.1 49.3 4.8 8.8 (11.5)
1701-1809 62.0 30.0 48.6 7.9 13.5 (42.6)
1810-1863 63.2 21.3 37.5 16.2 25.0 (7.4)

Eastern B. of Benin^
1662-1700 — — — —
1701-1809 63.4 33.8 57.5 2.6 6.0 (6.6)
1810-1863 67.4 20.2 44.0 12.4 23.3 (13.6)

Bonny
1662-1700 — — — —
1701-1809 56.2 40.3 52.7 3.5 3.4 (20.0)
1810-1863 68.2 14.3 51.3 17.1 17.3 (12.2)

Calabar
1662-1700 49.5 47.3 39.0 3.4 10.3 (5.5)
1701-1809 57.4 38.6 50.9 5.9 4.6 (12.2)
1810-1863 64.7 16.6 42.8 18.8 21.9 (7.3)

South-East Bight of Biafraf

1662-1700
1701-1809 62.0 29.8 48.6 8.7 12.4 (2.0)
1810-1863 62.5 18.3 38.3 18.9 24.5 (4.9)

a. includes Porto Novo, Badagry and Lagos combined.

b. includes Gabon, Bimbia, Cameroons, Corisco, Cape Lopez.
32 ROUTES TO SLAVERY

To what extent can the regional trends be explained by shifts in the flow
of slaves from outlets located within the regions themselves - in other words
by shifts in the relative importance of individual embarkation points? The
short answer supplied by Table 3 is very little. In the Bight of Benin, the
minor ports increased significantly in importance so that a major impact on
the overall ratios is However, some elementary interactive analysis
possible.
suggests that this was not the case. This involves two steps. First, we
recalculate Table 2 so that the time periods correspond with those in Table
3 (century-long instead of quarter-century), and in addition compute the
proportion of the total departures from the region accounted for by each
outlet (or grouping of outlets) in each century. We then select the age/sex
ratios for which change over time appears to be greatest - the proportion of
women and the proportion of children - and ask the question how would the
ratios in the nineteenth century have been different from what they were if

the distribution of departures among ports had changed, but the

demographic profile from each outlet had remained constant between the
eighteenth and nineteenth centuries. For the Bight of Benin, the women
ratio would have increased slightly rather than declined, and the child ratio

would have fallen by 17 per cent rather than increased. The same question
posed in the Bight of Biafra results in about 10 per cent of the observed rise

in child ratios being explained by shifts in the distribution of departures

between ports, but the women’s ratio stays constant. It is thus clear that the
key factor influencing trends demographic structure of the slave trade
in the

was not shifts between one strongly differentiated African region and
another. Differences between ports certainly existed, but these do not
explain the major changes through time - shared by all ports - that emerge
from the historical record.
It is obviously easier to highlight these large changes than account for
them - whether we look within or without Africa. Why, over time, would
fewer women, more men, and more children of both sexes leave all ports?
These trends were as strong when the trade was expanding between the
seventeenth and eighteenth centuries as they were between the eighteenth
and nineteenth centuries, when the volume of departures declined.
European wars, indeed events in the Atlantic World generally, might have
had temporary effects, but it is difficult to pinpoint any developments in
Europe or the Americas that could account for these secular trends.
Plantation crop type, for example - a factor that might be construed,
potentially, as a major influence on the demographic profile of peoples
entering the slave trade - remained largely constant throughout most of the
two centuries of the trade covered here. Erom the seventeenth century to the
1820s, it was sugar production, and sugar production alone, that absorbed
90 per cent of those carried across the Atlantic. Thereafter, coffee became
WEST AFRICA AND THE TRANSATLANTIC SLAVE TRADE 33

important, but the pattern of more children, fewer women, and more men
held for all branches of the traffic - the one that led to Cuban and Bahian
sugar plantations as well as the one that fed their coffee counterparts in
southern Brazil. In effect, the demographic characteristics of the coerced
migrant flow from Africa changed from one of rough balance between
males and females and the presence of some children in the seventeenth
century, to one in which males and children predominated by the nineteenth
century. Such a pattern, observed across a wide range of African cultures,
political and social structures, and economic activities, should pose
something of a challenge for specialists in African history.
To the extent that the age and sex of the forced migrants reflect African
patterns of slave supply as well as the requirements of European planters in
the Americas (and behind them European consumers of plantation produce),
the West African regions examined here exhibit interesting
three
similarities. They provide evidence which, taken together with other

language structure, religion, and even musical

cultural traits as diverse as
instruments, provides grounds for treating the three regions as a unit when
pursuing the topic of cross-Atlantic influences - at least for some
purposes.'* For example, in parts of Jamaica today, Tgbo’ is a term for a
nation or sub-group within a religion that is called ‘Akan’ by all adherents
and has clear Akan origins. We are on the threshold of acquiring much more
precise knowledge of the direction and composition of the slave trade from
West Africa to the Americas. Full recognition of the traits shared by West
African peoples will permit more effective use of the new Du Bois Institute
data set in nailing down what it was Africans carried with them and what it

was that they developed after their arrival. Clearly, the forced migration
from West Africa to the Americas was no more random and chaotic than
was its free European counterpart.

NOTES
Research for this paper was supported by funding from the National Endowment of the
Humanities. For comments on an earlier draft of this study, we wish to thank Stanley L.
Engerman and participants at the York University (Canada) conference. The African Diaspora
and the Nigerian Hinterland’, held in February 1996.

1 . Aboh, a major slave market at the head of Niger delta supplied slaves to both Bight of Benin
and Bight of Biafra outlets. Also see the comments in ‘Extracts from Mr Lyall’s Journal’, in
Macgregor Laird to Malmesbury, 14 March 1859, Public Record Office FO 84/1095, on
slaves destined for the Atlantic slave trade being carried down the Niger River. These slaves
could only have been embarked via the lagoons of the Slave Coast. More generally see
Mahdi Adamu, ‘The Delivery of Slaves from the Central Sudan to the Bight of Benin in the
Eighteenth and Nineteenth Centuries’, in Henry A. Gemery and Jan S. Hogendom (eds ). The
Uncommon Market: Essays in the Economic History of the Atlantic Slave Trade (New York,
1979), pp. 172-8.
34 ROUTES TO SLAVERY
2. Apart from the syntheses on the Atlantic slave trade published by Paul Lovejoy (The
Volume of the Atlantic Slave Trade: A Synthesis’, Journal of African History, 23 [1982],
pp.473-501 and The Impact of the Atlantic Slave Trade on Africa: A Review of the
30 [1989], pp. 365-94), which evaluate the relative importance of African
Literature’, ibid.,
regions, the last full assessment of either regionwas Patrick Manning, The Slave Trade in
the Bight of Benin, 1640-1890’, in Gemery and Hogendom (eds.). The Uncommon Market,
pp. 107-40. For studies on parts of the period covered here see Robin Law, The Slave Coast
of West Africa. 1550-1750: The Impact of the Atlantic Slave Trade on an African Society
(Oxford, 1991), and David Northrop, Trade Without Rulers: Pre-Colonial Economic
Development in South-Eastern Nigeria (Oxford, 1978).
3. For assessments of earlier quantitative work on the slave trade and comparisons between this
and findings based on the Du Bois Institute set, see David Eltis, David Richardson and
Stephen D. Behrendt, The Structure of the Transatlantic Slave Trade, 1595-1867’, in Henry
Louis Gates Jr, Carl Pedersen and Maria Diedrich (eds.). Transatlantic Passages
(forthcoming). For the modem starting point of this work see Philip D. Curtin, The Atlantic
Slave Trade: A Census (Madison, Wisconsin, 1969).
4. For tonnage, numbers are 181 ships with a standard deviation of 59.5 in the Bight of Biafra,
and for the Bight of Benin 332 ships were used to calculate the mean, the standard deviation
being 91.3. Definitions of tonnage varied over time and between countries. These data were
calculated after attempts to standardize the data; see the appendix to David Eltis and David
Richardson, ‘Productivity in the Transatlantic Slave Trade’, Explorations in Economic
History, 32 (1995), pp. 465-84 for the procedures used.
5. For the ending of the traffic in the Bights see David Eltis, Economic Growth and the Ending
of Trade (New York, 1987), pp. 168-73, and more recently, Robin
the Transatlantic Slave
Law, ‘An African Response to Abolition: Anglo-Dahomian Negotiations on the Ending of
the Slave Trade’, Slavery and Abolition, 16 (1995), pp. 281-31 1.
6. David Richardson, ‘Slave Exports from West and West-Central Africa, 1700-1810: New
Estimates of Volume and Distribution’, Journal of African History', 30 (1989), pp.1-22.
7. For a fuller discussion of the inter-regional differentials see the article by Herbert S. Klein
and Stanley L Engerman in this volume.
8. See Eltis, Richardson and Behrendt, ‘The Structure of the Transatlantic Slave Trade,
1595-1867’, Table 7 for a more precise breakdown of destinations in the Americas within
the major African regions of provenance.
9. Much of this is based on Pierre Verger, Flicx et Reflux de la Trade des Negres entre Le Golfe
de Benin et Bahia de Todos os Santos du XVlle au XIXe siecle (Paris, 1968), pp.3-5. For the
relative importance of Whydah see Law, Slave Coast of West Africa, pp. 8-48.
1 1

10. Before 1740, Dutch slave ships from the Gold Coast obtained almost all their slaves at the
Dutch castle of Elmina, even though they enter the Postma data set with designation ‘Gold
Coast’. A recoding of this embarkation point yields the result that these three ports supplied
82 per cent of Gold Coast slaves from known destinations (36 per cent from Anomabu, 26
per cent from Cape Coast Castle, and 21 per cent from Elmina).
11. Lagos, of course, was also far more accessible to the mid-nineteenth century British navy
than was Whydah. The town of Whydah was located beyond the range of British guns and
the Admiralty recognized that it could not be taken without a massive commitment of
resources. See Admiralty to Malmesbury, 9 February 1858, enclosure. Commander Wise, 21
December 1857, Public Record Office FO 84/1068.
12. I. Dahomey and its Neighbours, 1708-1818 (Cambridge, 1967), pp.76-7,
A. Akinjogbin,
141-51; Law, Slave Coast of West Africa, pp. 103-4, 284-6.
cf.

13. Akinjogbin, Dahomey and its Neighbours, pp. 69-71; Robin Law, The Oyo Empire c.
1

1600-c. 1836: A West African Imperialism in the Era of the Atlantic Slave Trade (Oxford,
1977); idem, ‘Dahomey and the Slave Trade: Reflections on the Historiography of the Rise
of Dahomey’, Journal of African History, 21 (1986), pp. 237-67.
14. Ryder’s account of slaving activity in Benin is broadly consistent with the pattern displayed
by the data set. Alan Ryder, Benin and the Europeans, 1485-1897 (New York, 1969),
pp. 196-238.
15. For a convenient description of the ‘ritual trading’ networks supplying Bonny, the Cross
WEST AFRICA AND THE TRANSATLANTIC SLAVE TRADE 35

River and Cameroon (the Aro, Ekpe and Bilaba), see Ralph Austen, African Economic
History (London, 1987), pp.94-5. For Gabon and its slave supply networks, see Henry
Bucher, ‘The Atlantic Slave Trade and the Gabon Estuary’, in Paul E. Lovejoy (ed.), Africans
in Bondage: Studies in Slavery and the Slave Trade (Madison, Wisconsin, 1986),
pp. 136-54.
16. Eltis,Richardson and Behrendt, ‘Structure of the Transatlantic Slave Trade’; David Eltis and
Stanley L. Engerman, ‘Fluctuations in Sex and Age Ratios in the Transatlantic Slave Trade,
1663-1864’, Economic History Review, 46 (1993), pp. 308-23.
17. See, for example, Stede & Gascoigne, Barbados, 4 April 1683, Public Record Office, T
70/16, f.50; Stede & Company, 22 August 1688, T 70/12,
Skutt, Barbados, to Royal African
p.31; Royal African Company to Petley Wybome and Henry Stronghill, 8 August 1688, T
70/50, f.70; Royal African Company to Browne, Peck, and Hicks at Cape Coast Castle, 23
July 1702, T 70/51, f.l31.
18. See John Thornton, Africa and Africans in the Making of the Atlantic World, 1400-1680
(Cambridge, 1992), pp.l84— 92, 235-71 for a discussion of the language and religious
commonalities.
 Long-Term Trends in African Mortality in
the Transatlantic Slave Trade

HERBERT S. KLEIN and STANLEY L, ENGERMAN

The transatlantic slave trade from Africa to the European settlements in the

New World has been only one of many large-scale international movements
of free and enslaved peoples in recorded history, but it has clearly been one
of the most frequently studied and discussed. The slave trade was entered
into by merchants and vessels from all the western European powers; it
involved trade with most areas of the West African coast as well as some of
the East African coast, and it led to sales of slaves throughout the regions of
the western hemisphere. The overall slave trade persisted for more than
three centuries, was quite large, involving at least 10-12 million slaves, and
since the carrying capacities of slave-trading vessels was small - averaging
about 300 slaves per vessel' - the trade entailed a large number of ships and
consisted of many voyages from several different nations. Because of the
role of slavery in New World settlement, and given the general range of
mercantilist policies by European nations, the trade and its vessels were
often regulated by government policies. These policies included the
granting of monopolies to slave trading companies, controls regarding
provisioning of ships, and also regulations affecting the numbers of slaves
to be carried on ships of different sizes and tonnages.
Debates among contemporaries and, subsequently, among scholars,
cover many issues. Particularly noteworthy have been those debates
concerning the consequences of the granting of monopoly privileges to
favoured companies, such as the English Royal African Company,- and
those estimating the extent to which mortality on the ‘middle passage’ could
be reduced from high levels by the introduction of appropriate government
regulations, as well as the effects of changes in mortality rates on profit
rates. The death on slave ships, of slaves and also of the crews, had
rates
attracted attention as one part of the attack on the slave trade (if not directly
on slavery and both the high overall mortality rates as well as some
itselO,
particularly bad cases of mortality on individual ships provided the
MORTALITY IN THE ATLANTIC SLAVE TRADE 37

argument that the system as practised was inhumane and inefficient, and
should be either regulated (at a minimum) or abolished. In Britain, Dolben’s
Act of 1788 was the major legislated attack by the British on slave ship
first

mortality, following Parliamentary debate and discussion as to the expected

effects of such regulation on mortality and profits.^ The Act linked the
maximum number of slaves to be carried to the tonnage of ships. Such a
relation had also been presented in a Portuguese act of 1684, which also
regulated the food and water to be carried on board, and the intent to
improve slave accommodations in transit led to similar Dutch regulations in
1789. Dolben’s Act also provided bonuses to those ship captains and
surgeons achieving relatively low death rates. The British regulations did
change over the next decade, the number of slaves carried being based upon
the dimensions of the vessel not tonnage, but this concern with mortality
continued through 1808, as long as the British slave trade remained legal.'*

About one century after the transatlantic slave trade ended, the analysis
of mortality in the slave trade again became an issue, this time among
scholars who have concentrated more on explaining patterns and rates of
mortality and the determinants of mortality, including the examination of
the nature of the specific diseases that caused the deaths of slaves and the
crew.^ Of particular interest was the trend in mortality over time, of obvious
importance in itself, but also to permit better calculations of the numbers of
slaves transported and the numbers arriving from the available but
incomplete data. Variables used to explain these trends have included the
number of slaves carried per vessel, the ship’s tonnage, the number of slaves
carried per ton (or other ship dimensions), and the number of days at sea
between Africa and the Americas, as well as various other aspects of the
slave-trading process within and on the coast of Africa. As the studies have
become more systematic, distinctions have been recognized between the
mortality rates for departures from different parts of Africa, differences by
the nationality of the ships carrying slaves, and also mortality differences
based on New World ports of arrival, differences not easily explained by
ship dimensions, sailing times, or other variables.
The new data set prepared by the Du Bois Institute represents a major
step forward in the attempt to deal with the many issues and problems of
analysing the slave trade, including the study of shipboard mortality. It is

drawing together many different data sets of slave trade data, for several
different flag carriers, and in making them as consistent as is possible,
permitting examination for comparative purposes. This is, of course, a most
difficult task to undertake, given that few of these sources were based on a
systematic and standardized form, for any country, let alone across
countries. Some of the data sets have been collected by other scholars, and
used by them in published works, before being made available to the
38 ROUTES TO SLAVERY
Institute. Others are newly processed and have not been previously
examined. The expansion in available information has increased the number
of voyages for which we now have information, and will therefore permit
more comparisons across time and place than has heretofore been possible.
It obvious that use of these sources, based on listings prepared by
is

contemporaries for several different countries over several different

centuries, is not problem free, since recording methods differed, and the
presence (or absence) of certain key pieces of information varied over time.
There are also problems of interpretation arising from the fact that the
distribution of surviving records which form the data set is no doubt
different from that which would have occurred if we had all the data from
the entire slave trade. This might be due either to systematic factors
influencing data collection and presentation, or to the accidents of data
survival and retrieval. Nevertheless, the care in preparation of what is made
available should make this a better data set to use to study the slave trade
than such a systematic presentation was not undertaken or if the analysis
if

were limited to only one body of data, for one nation or region.
Before seeing where the analysis of this data set takes us in the study of
slave trade mortality, it will be useful to see what questions the data set does
not permit us to answer. A most general piece of information on the slave
on only a limited number of records, concerns the length of
trade, available
time between the departure from Africa and the date of arrival in the
Americas - the sailing time or voyage length. This is for those ships that
have completed the voyage. If,however, the voyage was not complete and
the vessel lost at sea, due to wartime destruction, a wreckage, or rebellion,
then the losses of slaves are more difficult to determine.^ There is also some
information on time on the coast spent loading, but this is so for only a
limited number of voyages.
For several reasons we cannot answer questions about the overall
mortality in the transatlantic slave trade.’ We most frequently know what
were the deaths from the time of leaving the coast to the time of arrival in
the Americas — the so-called middle passage. We do not know, however,
after the number of deaths, the original enslavement in Africa - including
the losses of life in the capture of slaves, the deaths during the march to the
coast, and the mortality during the time spent on the coast before sailing,
either when held by African traders or European shippers. Nor can we
estimate those deaths after arrival in the Americas that may have been
attributable to prior conditions, including both the voyage and the delayed
effects attributable to African enslavement. Further losses occurred after
arrival in the New
World, both during the time awhiting sale, as well by the
increased rates during the initial period of adjusting to the local environment
and working conditions.
MORTALITY IN THE ATLANTIC SLAVE TRADE 39

This allocation of the losses of African lives between capture within

Africa and settled location within the New World was examined by Thomas
Fowell Buxton in 1840. Buxton described the mortality losses at five
different stages of the enslavement process: ‘The original seizure of the
slaves; the march to the coast, and detention there; the Middle Passage; the
sufferings after capture [of slave vessels by the anti-slave trade fleet] and
after landing; and the initiation into slavery, or the “seasoning”, as it is

termed by the planters.’ Buxton argues that the total loss of life in this
process is about 70 per cent of the ‘victims of the slave trade’. He estimates
that the losses within Africa between enslavement activities and departure
from the coast were 50 per cent; mortality in the middle passage accounted
for 25 per cent of the slaves embarked; and of those landed, 20 per cent died
in seasoning. By his estimates, the middle passage accounted for 17.9 per
cent of deaths in the slave trade, with 71.4 per cent of deaths occurring
within Africa. There have been several recent attempts to update Buxton’s
estimates of mortality prior to the middle passage. These are for different
periods and regions, and also for varying time in transit, but while the range
is wide, and the estimates varied, they generally exceed the mortality rates
on the transatlantic crossing, with works by Hogendom and by Miller
placing the mortality rate from point of captivity to the coast at about 50 per
cent.*
The issues can pose a problem in the attribution of deaths to specific
stages of the shipment of slaves. Diseases that were initially generated in
Africa will be mixed in with the diseases caused by conditions on the slave
ship, while those diseases that originated on board ship might have led to
deaths not at sea but only after arrival in the Americas. These factors will
influence any attempt to explain mortality at sea based upon ship-specific or
region-specific variables. But there no major difference in the use of these
is

ship data to study slave trade mortality from the data used in the study of
other historical problems, which also work from limited documentation and
the use of supplementary data to answer historical questions.
Why is the middle passage such a widely studied event? First, it is the
part of the slave trade system that provides the best recording of information
dealing with the numbers of slaves involved and their mortality in transit.
Second, there is the importance of the European involvement in the middle
passage, and, unlike the internal movement of slaves within Africa, the
European practice of detailed record-keeping and of government regulation
of the transatlantic traffic in slaves. Thus shipboard mortality became the
focus of many of the debates in western Europe regarding the slave trade,
and there was concern both with the mortality of slaves during the
transatlantic passage and the deaths of the crew in the slave trade. While the
shipboard mortality of slaves was used to demonstrate the harshness of the
40 ROUTES TO SLAVERY
trade, it was the high mortality among the crew that was used by the British
abolitionist, Thomas Clarkson, in his arguments to end the slave trade. To
those who supported the slave trade as a ‘nursery for seamen’, Clarkson
argued that, to the contrary, the rates of mortality of the crew on slave ships
were considerably higher than for crews on other routes, including the
commodity trade with Africa.'^
Our discussion of results here will be somewhat tentative, since we have
utilized only a sub-sample of the data set, which is itself a sub-sample of the
slave trade, with, in both cases, the precise shares of different trades not
certain. For the present analysis we have utilized only those ships for which
complete data are available, and for which the patterns of slaves
relatively
loaded and landed, and mortality, appear consistent.'" This does leave
enough observations, however, to reach conclusions on certain major issues
and to raise some questions for further analysis.
The basic findings in regard to long-term trends in slave ship mortality,
are somewhat familiar from the many earlier studies." The basic measure of
interest, reflecting the concerns of contemporaries, is the mortality rate per
voyage - the number of slaves who were recorded as died in the middle
passage (or else the difference between slaves loaded and slaves landed)
divided by the number of slaves loaded on the African coast. The most
significant pattern for understanding mortality in the middle passage is the
very wide distribution of mortality rates by voyage. This is found even
when holding constant other features, including sailing times, ship sizes,
African embarkation areas, and the age-sex composition of slaves carried.
There was a broad range of outcomes, with very many quite different
experiences, even for the same captains or the same nationality of shippers.
Very high mortality rates tend to be associated with unexpectedly long
voyages, or with unusual outbreaks of disease, but, in general, it is the very
broad range of outcomes rather than any bunching at specific mortality rates
that has been the main characteristic of the transatlantic slave trade.
While we are concerned with factors that influence the mortality in the
transatlantic slave trade, it has been demonstrated that the mortality per
voyage on slave ships exceeds those of other transoceanic voyages at the
same time, although those voyages also have a wide range of outcomes.
Studies of other voyages and their comparisons with the slave trade indicate
lower mortality rates for ships carrying convicts, contract labourers,
military and free immigrants." As with the slave ships, these
troops,
voyages experienced declines in their mortality rates over time, with, in
some cases, the magnitude of the decline exceeding that of the slave ships.
From the initial period, 1 590-1 7(X), slave shipboard mortality declined
to less than half its initial level in the late eighteenth century. While the
decline was relatively monotonic over time, there was a large decline in the
MORTALITY IN THE ATLANTIC SLAVE TRADE 41

PERIODS)

TIME

(TWO

PASSAGE

MIDDLE

THE

1
Embarkeij

DURING
FIGURE

LOSSES

Slaves

SLAVE of

OF

Percentage

DISTRIBUTION

as

Deaths

RELATIVE

81.s06bAoa 10 iu0Oj0c|

4
 PERIODS)

TIME

(TWO

TIME

VOYAGE

2
PASSAGE

FIGURE
(jays

in
MIDDLE

passage
OF

DISTRIBUTION
miiddle

on

spent
RELATIVE

Time

S06 bAoa |o ju0Oj0d

MORTALITY IN THE ATLANTIC SLAVE TRADE 43

last quarter of the eighteenth century, a period in which British ships

achieved an unusually low slave mortality. As sharp as was the decline in
the mean of slave mortality rates, the median of slave ship mortality
declined even more rapidly, with the entire distribution of mortality rates
shifting down (see Figure 1). In the last forty years of the slave trade, when
it was restricted to mainly Spanish and Portuguese vessels, mortality rates
rose again, but they did not reach pre-1800 levels. Although there were
declines in time spent in the middle passage, particularly in the late
eighteenth and early nineteenth century, it is doubtful that speedier voyages
alone were sufficient to explain the observed declines in mortality (see
Figure 2).

TABLE 1

MORTALITY BY FLAG CARRIERS (%)

(n) Portugal (n) Spain (n) Dutch (n) French (n) British (n) All
1590-1700 (67) 29.8^ (139) 14.3 (195) 21.3 (401) 20.3
1701-1750 (427) 15.8 (488) 15.6 (176) 15.4 (1091) 15.6
1751-1800 (93) 8.6 (607) 14.0 (698) 12.5 (846) 10.6 (2287) 11.8
1801-1820 (657) 8.2 (12) 4.2 (6) 9.5 (8) 6.3 (112) 15.0 (807) 9.1

1821-1867 (962) 9.1 (136) 118 (16) 15.0 (124) 9.4^^ (1380) 10.4

Total (1712) 8.7 (215) 17.2 (1180) 14.7 (1210) 13.8 (1453) 12.9 (5966) 12.4

a. Portuguese vessels carrying slaves for Spain.

b. Not British Flag.

This decline over time was seen for each of the major slave trading
nations in the eighteenth century, but none was as steep as was that of the
British (see Table I).'"* The ratio of mortality rates, 1751-1800 to 1701-50,
was 0.80 for the French, 0.89 for the Dutch, and 0.69 for the British. This
indication of relative British efficiency is consistent with the growing
importance of the British in the international slave trade, and of the
expanding British economy internationally.'^To some extent the decline
was influenced by the legislation regarding the numbers of slaves to be
carried, which also included provisions and bonuses
for shipboard surgeons,
paid to captains and surgeons on ships with low mortality. While there were
regulations on numbers of slaves carried for other nations, none required
surgeons or paid bonuses for holding down the rate of mortality. In the
nineteenth century the mortality rates of the relatively new and important
Spanish- American and Brazilian flag carriers were at levels above those that

the British had achieved in the last decades of its slave trade, but below
those of French and Dutch shippers.'® From his studies of numerous
migrations in the nineteenth century, Ralph Shlomowitz has pointed to the
44 ROUTES TO SLAVERY
important role played in the absence of correct medical knowledge by
administrative reforms and various empirical approaches in lowering the
rates of mortality over time.'’ Thus whatever forces of institutional,
technical, and organizational change influenced mortality in the slave trade,
they apparently became widely diffused among all trading nations. Some of
the key changes in ship construction may have reflected technological
developments, were also changes in vessel design and
but there
organization that permitted a change to more optimal speed or size for the
shipowners, and there were also some organizational and medical changes
without any new technology. The rising prices of slaves would have meant
that there would have been an increased return to improved care of slaves

in the middle passage. It is probable that the overall reductions in mortality

rates reflected some combination of these various types of change.

An important factor explaining the pattern of declining mortality rates is

that, given the large number of ships involved in the trade, and the shifting
overall distribution of mortality rates, some widespread changes had
occurred. These changes affected not just a limited number of shippers but
a great majority of them. There was a large increase in the share of voyages
coming in at relatively low mortalities (however low is defined).
Correspondingly, the percentage of ships with mortality rates above a
selected threshold fell, meaning that over time there were relatively fewer
ships with very high mortality rates. While the explanations for the changes
at these two ends of the distribution may differ, it is clear that the process of
diffusion of information and techniques among shipowners and captains
played a crucial role in this measured mortality decline.

TABLE 2
MORTALITY BY AFRICAN REGION OF DEPARTURE (Vo)

(n) All In) 1590- in) 1701- In) 1751- In) 1801- In) 1821-
1700 1750 IHOO 1820 1867

Senegambia (247) 12.8 (33) 13.6 (94) 10.6 (105) 15.6 (8) 9.0 (7) 6.4
SierraLeone (286) 8.6 (12) 11.0 (17) 12.0 (174) 94 (28) 6.7 (55) 6.4
Gold Coast (1,041) 13.2 (69) 21.6 (243) 15.5 (636) 12.4 (83) 5.9 (10) 7.7
Benin (758) 13.8 (83) 17.5 (301) 17.1 (185) 13.6 (34) 5.1 (155) 7.4
Biafra (480) 17.4 (30) 29.5 (24) 43.2 (280) 15.1 (18) 9.5 (128) 15.6
West-Central
Africa (2.117) 9.2 (103) 22.9 (169) 12.1 (634) 9.4 (503) 7.7 (708) 7.1
South-East (248) 18.3 (1) 29.5 (8) 19.3 (46) 19.9 (193) 17.6
Misc. (673) 15.5 (71) 19.2 (242) 15.7 (249) 13.9 (79) 16.7 (36) 15.9

Total (5.854) 12.5 (401) 20.3 (1,091) 15.6 (2,271) 12.1 (799) 9.0 (1 ,292) 9.8

and 2 show significant differences in mortality rates, among

Tables 1

European flag carriers, and also persisting differences in mortality from

various African ports of departure. Table 2 indicates that the mortality rate
MORTALITY IN THE ATLANTIC SLAVE TRADE 45

differences by African region of departure remain rather systematic during

the period studied, with a general persistence in the nature of the
differentials over time.'* A rough measure of this pattern can be seen by
examining the seven regions for five time periods - three regions are either
always above or always below the average mortality rate for that time
period, and two differ in direction in only one period. While there are those
persistent differences by port of departure. Table 2 also indicates that all

regions had mortality rate declines over time and generally had their lowest
rates in the nineteenth century. While the lowest rate among the major
regions classified, for West-Central Africa, is for the region with the
shortest average sailing time, in general the times at sea cannot alone
explain the magnitude of differences in mortality rates by African port. It

might be expected that, in general, these persistent differences in the

transatlantic mortality rates based on African port of departure reflect as

much the conditions in Africa as the characteristics of the ships and the
voyages. They were influenced by changes in catchment areas within
Africa, shifts that occurred with political and economic changes, and that
influenced the age-sex composition of the slaves sent as well their overall
physical condition. Therefore, more attention should be given to the slave
losses inland, to the costs of transporting slaves to the coast, and to the time
spent on the coast prior to transatlantic sailing. In regards to length of
voyage, there was some increase in mortality with numbers of days at sea,

particularly for those unexpectedly long voyages on which water and

provisions ran short, and accelerated the spread of disease. For the great
number of voyages, however, there was little variation in mortality that
could be directly explained by differences in the number of days at sea.

There was some decline in mortality rates with tonnage of ship, except
for a limited number of very large vessels. This probably reflected the nature
of ship construction and measurement, with the area used for slaves
increasing more rapidly than did overall measured tonnage. There is a small
positive correlation of mortality with slaves carried per ton, but there is no
indication of a sharp distinction between so-called ‘tight packers’ and ‘loose
packers’. The number of slaves per ton was always high on slave ships,
higher than on other voyages. This means that while the numbers carried per
ton might help explain why slave ships had higher mortality than did other
transport vessels, within the general range of slaves per ton carried there was
no additional effect, despite the arguments made in the ‘packing’ debate.

As we have indicated, there remain a number of puzzling aspects regarding

the transatlantic slave trade. The large dispersion of mortality experiences,
even within specific national carriers and for departures from specific
46 ROUTES TO SLAVERY
African ports, demonstrates the role of the particular factors influencing
individual voyages. Similarly the higher mortality rates on slave vessels, for
crew as well as slaves, compared with all other transoceanic carrying trades,
further suggests the unusual conditions determining the supply of slaves and
influencing the middle passage experience on slave-carrying voyages. The
persistent differentials in mortality rates by African port of departure points
to the importance of further studies of the initial process of enslavement
within Africa and in the march to the coast. Since the middle passage
accounts for a relatively limited number of all deaths in the overall
transatlantic slave trade, it would obviously be useful to extend the analysis
to these other steps within the enslavement process.

NOTES
1. The average number of slaves per vessel was computed from the Du Bois Institute data set.
See also Herbert S. Klein, The Middle Passage: Comparative Studies in the Atlantic Slave
Trade (Princeton, 1978). There was some upward trend over time, offset by the onset of
regulation beginning in the late eighteenth century. These estimates imply that there were
over 30,000 separate voyages in the transatlantic slave trade.
2. On the slave trade of the Royal African Company, see, in particular, K.G. Davies, The Royal
African Company (London, 1957); David W. Galenson, Traders, Planters, and Slaves:
Market Behavior in early English America (Cambridge, 1986).
3. See Klein, Middle Passage, pp. 141-75, and Herbert S. Klein and Stanley L Engerman,
‘Slave Mortality on Bntish Ships, 1791-1797’, in Roger Anstey and P.E.H. Hair (eds.),
Liverpool, the African Slave Trade, and Abolition (Liverpool, 1976), pp. 113-25, for a
discussion of the impact of Dolben’s Act on the number of slaves carried.
4. See Charles Garland and Herbert S. Klein, ‘The Allotment of Space for African Slaves
aboard Eighteenth-Century British Slave Ships’, William and Mary Quarterly, 3rd senes, 42
(1985), pp. 238-48, for an examination of ship design and the relation between various
physical dimensions, tonnage, and carrying capacity. For the Portuguese decree, see Klein,
Middle Passage, pp. 29-30. The Dutch regulation is noted in Johannes Menne Postma, The
Dutch in the Atlantic Slave Trade, 1600-1815 (Cambridge, 1990), pp. 287-8.
5. For recent studies of the mortality rates on slave vessels and their causes, see, for example,
Richard H. Steckel and Richard A. Jensen, ‘New Evidence on the Causes of Slave and Crew
Mortality in the Atlantic Slave Trade’, Journal of Economic History, 46 (1986), pp. 57-77;
Raymond L. Cohn, ‘Deaths of Slaves in the Middle Passage’, Journal of Economic History,
45 (1985), pp. 685-92; David Eltis, ‘Mortality and Voyage Length in the Middle Passage;
New Evidence from the Nineteenth Century’, Journal of Economic History, 44 (1984),
pp. 301-18; Raymond L. Cohn and Richard A. Jensen, ‘The Determinants of Slave Mortality
Rates on the Middle Passage’, Explorations in Economic History, 19 (1982), pp. 269-82;
Galenson, Traders, Planters, and Slaves, pp. 29-52; Joseph C. Miller, ‘MortJility in the
Atlantic Slave Trade: Statistical Evidence on Causality’, Journal of Interdisciplinary'
History, 11 (1981), pp. 385-423; David Northrup, ‘African Mortality in the Suppression of
the Slave Trade; The Case of the Bight of Benin’, Journal of Interdisciplinary History, 9
(1978), pp. 47-64; and David Richardson, ‘The Costs of Survival; The Transport of Slaves in
the Middle Passage and the Profitability of the 18th-Century British Slave Trade’,
Explorations in Economic History, 24 (1987), pp. 178-96.
6. See, however, the studies reconstructing such lost vessels by Stephen D. Behrendt, ‘The
British Slave Trade, 1785-1807: Volume, Profitability and Mortality’ (unpublished Ph D.
of Wisconsin, 1993) and by Joseph E. Inikori, ‘Measuring the Unmeasured
thesis. University
Hazards of the Atlantic Slave Trade: Documents Relating to the Bntish Trade’, Revue
MORTALITY IN THE ATLANTIC SLAVE TRADE 47

fran^aise d'histoire d’outre-mer, 83 (1996), pp. 53-92. Inikori places estimated losses at
about 10 per cent of all voyages, but notes that the effect was to add altogether no more than
one or two percentage points to the overall mortality rate.
7. See Thomas Fowell Buxton, The African Slave Trade and its Remedy (London, 1840; 1967
edition), Ch.2. For a participant’s description of the stages of the enslavement experience
within Africa, the middle passage, and New World arrival, as well as his experiences as a
slave and in the antislavery movement, see The Interesting Narrative of the Life of Olaudah
Equiano, Written by Himself (1191). Clearly the debates on mortality due to enslavement and
mortality in the middle passage refer to two quite different sets of issues.
8. See, for example, Jan S. Hogendom, ‘Economic Modelling of Price Differences in the Slave
Trade between the Central Sudan and the Coast’, Slavery and Abolition, 17 (1996),
pp. 209-22; Joseph C. Miller, Way of Death: Merchant Capitalism and the Angolan Slave
Trade, 1730-1830 (Madison, Wisconsin, 1988), pp.l53, 441. For other discussions of this
problem, see Patrick Manning, Slavery and African Life: Occidental, Oriental, and African
Slave Trades (Cambridge, 1990), p.58 and Paul E. Lovejoy, Transformations in Slavery: A
History of Slavery in Africa (Cambridge, 1983), pp.60-3.
9. See Thomas Clarkson in Sheila Lambert (ed.). House of Commons, Sessional Papers of the
Eighteenth Century (Wilmington, Delaware, 1975), vol.69, and his subsequent publications
of these data. Clarkson used the muster rolls and related records of slave vessels from
Liverpool, London, and Bristol to estimate the overall loss of crew, of about 20 per cent. He
then compared crew losses with that for other trades, including Newfoundland, Greenland,
the East Indies, Petersburg, and also the trade with Africa which did not involve slaves, and
made allowance for differences in sailing times. The losses in the African commodity trading
vessels were only one half of those on slavers, in part because of less time spent loading on
the coast. For a study of crew mortality at various stages of the slaving voyage, see Stephen
D. Behrendt’s essay in this volume.
10. Thus some of the patterns shown in Tables and 2 may not accurately reflect the mortality
1

experiences of different national flag carriers or of slaves of different African origins. For the
discussion in this paper, particularly noteworthy is the absence of Portuguese data on
mortality prior to the 1790s.
11. See, for example, the sources cited in note 5 above. Since the Du Bois Institute data set
includes most of the materials used in earlier studies of individual nations, some similarity
of resultsis to be anticipated. The ability to make comparisons among nations represents a

major contribution of this data set.

12. The reason for our using this measure (mortality rates per voyage) is to allow separate
discussion of the effects of length of voyage on mortality and to highlight the impact of
unusually long voyages on slave deaths in transit. Since earlier studies indicate that the
correlation between mortality rates and sailing times was low, except for the very long
voyages, division of mortality rates by sailing times gives an artificial pattern regarding the
timing of deaths during the voyage. See Steckel and Jensen, ‘New Evidence’; Klein, Middle
Passage', Galenson, Traders, Planters, and Slaves, pp.48-51.
13. See the summary studies by Raymond L. Cohn, ‘Maritime Mortality in the Eighteenth and
Nineteenth Centuries: a Survey’, International Journal of Maritime History, (1989),
1

pp. 159-91; Robin Haines, Ralph Shlomowitz and Lance Brennan, ‘Maritime Mortality
Revisited’, International Journal of Maritime History, 13 (1996), pp. 133-72; Ralph
Shlomowitz (with co-authors Lance Brennan and John McDonald), Mortality and Migration
in the Modern (Aldershot, 1996).
14. The British mortality decline was highly concentrated in the last two decades of its slave
trade, falling to less than one-half the 1776-89 level in the years 1790-1807. In no year after
1790 was the mortality rate above 8 per cent, and in no year after 1794 was it above 4 per
cent. To indicate the uniqueness of the British mortality experience, the decline in mortality
between the 1776-89 average and 1790-1807 for other carriers was only about 16 per cent.
And while these other nations had, in the post- 1790 period, frequent rates below 10 per cent,
the lowest annual mortality rate was above 6 per cent, higher than the British average for the
eighteen years between 1790 and 1807. The British mortality rate per day at sea declined
sharply after 1790, meaning that the pattern cannot be explained by changing time in the
48 ROUTES TO SLAVERY
middle passage. The non-British ships, however, had a relatively constant mortality per day
at sea alter 1776. For a detailed breakdown of slave mortality on British vessels before 1780

and 1799, pointing to a very strong post- 1792 decline, see Stephen D. Behrendt, The Annual
Volume and Regional Distribution of the British Slave Trade \l'&0-\%Qn' Journal of African
,

History (forthcoming).
15. For more direct examination of total factor productivity in the slave trade, see David Eltis
and David Richardson, ‘Productivity in the Transatlantic Slave Trade’, Explorations in
Economic History, 32 (1995), pp. 465-84.
16. There were numerous Portuguese slave ships in the seventeenth and eighteenth centuries, but
the information did not contain mortality data. The Portuguese (and Brazilian) data in Table
1 start we cannot determine the overall trend in Portuguese mortality.
with 1795, so
17. Shlomowitz, Mortality and Migration, pp.lO-l 1, and also the various case studies presented.
18. The same systematic differences by African regions are seen in the mortality rates per day at
sea, and their declines over time, indicating the importance of factors other than days at sea
in explaining these mortality differentials.
 Crew Mortality in the Transatlantic Slave
Trade in the Eighteenth Century

STEPHEN D. BEHRENDT

During the past ten years historians have written several new studies on the
mortality of non-slave shipboard populations in the eighteenth and
nineteenth centuries. The passenger groups examined include European
emigrants to colonies and to the United States, British convicts to North
America and Australia, and African, Indian, Chinese, and Pacific islander
indentured labour to the Americas, South Africa, Mauritius, Australia, and
Fiji.' By comparison, there have been only two recent studies which present
new information on crew mortality in the transatlantic slave trade. Using
voyage from surgeons’ logs, Steckel and Jensen analysed crew
details
mortality rates on the African coast and middle passage for 92 British slave
voyages in 1792-96.- From examination of muster rolls, Behrendt
calculated the mortality losses of captains in the Bristol and Liverpool slave
trades between 1785 and 1807.^ Several earlier studies of crew mortality on
slave ships relied on Rinchon’s pioneering research of the major French
slaving port of Nantes more than fifty years ago"* and Unger’s examination
of selected slave voyages organized by the Dutch Middelburg Commercial
Company.^ Other works include Mettas’ examination of crew mortality on
24 slave voyages from Honfleur in 1763-88 and 452 French slave voyages
to the Angolan coast in 1714-91 and Stein’s analysis of crew mortality on
130 slave voyages from Nantes in 1715-78.*
As mentioned in the introductory paper by Eltis and Richardson in this
volume, in the 1970s Jean Mettas collected and organized slave voyage data
for Nantes and other French ports; in 1978 and 1984 the Dagets published
the data in the Repertoire des Expeditions Negrieres Frangaises an XVIIle
Siecle. These French slave voyage data, which have been scanned into the
Du Bois Institute data set, contain information on the number of crew at the
outset of the voyage and the deaths of sailors on each of the five stages of
slaving voyages - namely, the outward passage, the stay at the African coast,
the middle passage, the time in the Americas, and the homeward passage.’
There is crew mortality information for 1,792 of the 3,343 French slave
50 ROUTES TO SLAVERY
voyages between 1691 and 1793 contained in the Mettas-Daget catalogue.
In this large sample, there are also dates of departure and arrival on select
voyage legs which permit analysis of crew mortality by time and place.
Detailed information on crew mortality in the British slave trade is

contained in ledgers of Liverpool and Bristol muster rolls.* During the

period of the British slave trade, there are musters which record information
on the crew sailing on four out of five Liverpool slave voyages between
1770 and 1807.’ Bristol muster rolls survive from voyages which sailed
from 1748 to 1794.'" The musters report the names of the crew and the dates
of their discharge, desertion, impressment, or death. Some musters list crew
rank or cause of death. Departure and return dates are given, and other
departures and arrivals throughout the voyage can be inferred from dates of
crew discharge or desertion." We
have presently integrated into the Du Bois
Institute data set slave voyage information from 2,120 Liverpool muster
rolls in 1780-1807. Of this total, musters report reliable data on crew deaths

for 1,963 voyages, of which 1,709 successfully completed the triangular

voyage and returned to Liverpool.'*
By examining this large run of Liverpool musters, one can determine the
number of crew on the outward passage to Africa and the number of
‘original’ crew deaths during the triangular voyage.'^ These data are
contained in the Du Bois Institute set in two of the twelve variables
concerning crew mortality. These twelve variables are:

1 . Crew at the outset of the voyage

2. Crew deaths prior to the African coast (voyage leg one)
3. Crew deaths on the African coast (voyage leg two)
4. Crew who sailed from the African coast (crew at outset of middle
passage)
5. Crew deaths in the middle passage (voyage leg three)
6. Crew who arrived in the Americas
7. Crew deaths in the Americas (voyage leg four)
8. Crew who sailed from the Americas
9. Crew deaths on the homeward passage (voyage leg five)
10. Crew who arrived at the port of return
1 1 . Crew deaths prior to Americas (inferred from 2, 3, 5 above)
12. Crew deaths on the voyage
In selecting these variables,was decided to include the basic crew
it

mortality information which would be reported in most maritime history

sources. Sub-sets of data which report additional crew mortality
information, however, may be linked easily to the Du Bois Institute data set.
This paper will analyse the new crew mortality data for the French and
British slave trades contained in the Du Bois Institute set. It will also
CREW MORTALITY IN THE ATLANTIC SLAVE TRADE 51

examine two sub-sets of data compiled from a selection of Liverpool slave

voyages in the periods 1770-75 and 1801-7. It will analyse aggregate crew
mortality on the triangular voyage, crew mortality by region of trade, crew
mortality by voyage leg, and crew mortality by rank. Mortality on the
maritime legs of the voyage (legs one, three and five) will be compared to
the mortality experience of other non-slave shipboard populations. French
and British crew mortality data on the African coast and in the Americas
(legs two and four) will be compared to that of other European migrant
groups. The essay will conclude with some brief comments about the
relationship between crew and slave mortality.
Previous studies have shown that between 1748 and 1792 about 17 per
cent of the outward bound crew of slave ships from Nantes died during the
voyage and most deaths occurred on the African coast and middle
that
passage. Based on a sample of 598 voyages, these Nantes data suggest that
crew mortality declined from 18.7 per cent in 1748-72 to 14.7 per cent in
1773-92.'^ The mortality data from our larger sample of 1,570 completed
voyages from all French slaving ports corroborate these findings. As Table
1 shows, about 17 per cent of the outward bound crew died in the
eighteenth-century French slave trade.'*

TABLE 1

VOYAGE CREW MORTALITY IN THE FRENCH SLAVE TRADE, 1711-95 (by five-year periods)

Five-year period Voyage sample^ Mortality' loss (%) Voyage Mortality-

duration (days) rate^

1711-1715 38 23.1 498.5 13.9

1716-1720 61 21.0 475.1 13.3
1721-1725 64 22.8 512.2 13.4
1726-1730 68 16.5 525.3 9.4
1731-1735 54 20.8 537.2 11.6
1736-1740 94 19.8 549.9 10.8
1741-1745 97 18.7 539.9 10.4
1746-1750 46 21.1 555.4 11.9
1751-1755 138 17.3 518.9 10.0
1761-1765 70 20.9 517.5 12.1
1766-1770 172 20.0 540.7 11.1
1771-1775 145 12.2 488.8 7.5
1776-1780 88 12.4 496.3 7.5
1781-1785 91 17.7 488.3 10.9
1786-1790 263 14.1 472.9 8.9
1791-1793 81 12.3 440.1 8.4

Totals 1,570 17.2 506.3 10.2

Source: Atlantic slave trade database, Du Bois Institute, Harvard University (see introduction).
Data for Tables 1-8 are computed from this consolidated slave voyage set.

Note: Table 1 excludes data from the five-year periods, 1706-10, and 1756-60 because of small
sample sizes The eighteenth century French slave trade ended in 1793.
(less than ten voyages).

a. Sample includes data from French slave voyages which completed the triangular journey.
b. Crew deaths per 1,000 crew per 30-day period.
52 ROUTES TO SLAVERY

Most crew deaths were on the African coast and middle passage, though
about one in four deaths were in the Americas. ’’ The data show a reduction
in crew mortality loss over the course of the century, from 20 per cent in
171 1-70 to 14 per cent in 1771-93. This reduction was not associated with
faster triangular voyages: crew mortality rates declined from 11.3 to 8.6
crew deaths per thousand crew per month (of 30 days)'** during the same two
periods.
For comparative purposes, we can evaluate French crew mortality by
African region based on triangular voyage data. The relevant data are
presented in Table 2.

TABLE 2
VOYAGE CREW MORTALITY IN THE FRENCH SLAVE TRADE
(by African region of trade)

Crew mortalit}' loss Crew mortality' rate

African region Voyage sample^ Mortality loss {%) Voyage sample^ Mortality
rate^

Senegambia 150 14.8 149 9.6

Sierra Leone-
Wmdward Coast 123 15.3 1 22 9.1
Gold Coast 116 15.9 115 8.7
Bight of Benin 404 18.3 401 10.6
Bight of Biafra 107 17.1 106 11.0
West-Central Africa 628 16.0 628 9.3
South-East Africa 11 8.3 11 5.0

Totals 1,539 16.5 1,532 9.7

a. Sample includes data from French slave voyages which completed the triangular journey
b. Crew deaths per 1,000 crew per 30-day period.

Note: African regions arranged from north to south.

Though this is a crude measurement, because of varying coastal slave-

loading rates and different voyage lengths on the outward and middle
passages, the breakdowns suggest that crew mortality rates were highest on
French slave voyages to the Bight of Benin and Bight of Biafra and lowest
on voyages to South-East Africa (Madagascar and Mozambique) and the
Gold Coast. On the middle passage alone,
from a sample of 1,535 French
slave voyages we find that crew died at a rate of 15.1 per thousand crew per
month. The highest mortality rates occurred in the middle passage on
voyages which had slaved at West-Central Africa (19.4 per thousand per
month) and the Bight of Biafra (17.3 per thousand per month). For
European shipboard populations in the eighteenth century, greater mortality
rates of about 50-60 crew deaths per thousand crew per month have been
CREW MORTALITY IN THE ATLANTIC SLAVE TRADE 53

calculated from small samples of British convict voyages to North America

in 1719-36 and British and German troop shipments to the West Indies and
North America during the American Revolution. German adult male
emigrants to Pennsylvania on 14 voyages between 1727 and 1805 died at
mortality rates of 15.8 per thousand per month, a rate very similar to that of
French crew in the middle passage.-" On ships sailing from the two principal
regions of French slaving activity in the eighteenth century - the Bight of
Benin and West-Central Africa - the highest crew mortality rates in the
middle passage occurred, as Table 3 shows, when vessels sailed to the
Americas during months of comparatively high levels of rainfall in these
regions. The lowest crew mortality in the Atlantic crossing occurred on
ships sailing during the dry season.-' This finding is consistent with other
studies which found evidence of increased European mortality during
African rainy seasons.-- Greater amounts of stagnant water (in pools or in
open casks) fostered larger populations of Aedes aegypti. Anopheles
gambiae, and Anopheles melas, the principal mosquito vectors of the yellow
fever virus and malarial parasites on the West African coast.-’

TABLE 3
MIDDLE PASSAGE CREW MORTALITY IN THE FRENCH SLAVE TRADE
(by African region and month)

Bight of Benin West-Central Africa

Month of Voyage Mortality Month of sail Voyage Mortal

sail from Benin sample rate^’ from Angola sample rate^

January 28 18.7 January® 59 22.2

February 18 1 0.0 February® 55 22.9
March 39 15.1 March® 55 19.6
April 34 19.0 April® 37 11.7
May^ 25 18.8 May 43 18.8
June“ 35 15.9 June 45 19.6
July® 25 29.5 July 38 9.9
August® 18 17.9 August 37 20.0
September® 26 14.4 September 44 14.4
October 36 13.3 October 42 10.5
November 28 12.0 November® 45 15.5
December 40 18.0 December® 57 22.6

Totals 352 16.9 Totals 557 17.8

a. Months of greatest
rainfall (wet season).
b. deaths per I,(X)0 crew per 30-day period. We approximate the number of crew on the
Crew
middle passage as the original crew less deaths on the outward passage and coast of Africa. We
assume that no additional crew embarked on the African coast.

Turning to the British slave trade, the calculation of crew mortality data
from muster rolls continues the work begun in the 1780s by the British
abolitionist Thomas Clarkson. For the Liverpool trade, Clarkson calculated
54 ROUTES TO SLAVERY
the crew mortality of the ‘last eighty-eight vessels in this trade, that had
returned to Liverpool from their respective voyages in the September of the
year 1787’. Clarkson calculated that 20.5 per cent or 631 crew died out of
-*
the 3,082 crew taken on during the triangular voyages of these ships.
Regarding the Bristol and London slave trades, he reported that ‘Every
vessel from the port of Bristol loses ... almost a fourth of the whole crew’,
while ‘Every vessel from the port of London loses ... between a fourth and
a fifth of thewhole complement of her men.’-^ For the year 1786 Clarkson
estimated that the British slave trade employed 5,000 seamen and that 1,130
were ‘upon the dead list’.’^ In Spring 1791 a House of Commons committee
examined muster rolls for 350 Liverpool and Bristol slave voyages in
1784-90. The musters indicated that 2,643 ‘original’ crew died out of
12,263 enlisted, a mortality loss of 21.6 per cent.^^ The committee report
also noted the average voyageand one can calculate from this
length,
information mortality rates of 16.7 crew deaths per thousand crew per
month for the Liverpool and Bristol slave trades.-* Furthermore, the
parliamentary abstract separated the deaths of crew who drowned, were
killed or were lost. Of the 2,400 crew deaths on 298 Liverpool slave
voyages from 1784 to 1790, 159 crew were reported drowned, 27 were
killed, and six were lost. Of the 285 deaths on 52 Bristol slave voyages in

the same period, six crew drowned, one was killed, one was lost, and one
died ‘falling from the mast’. Thus, on these voyages 93 per cent of the crew
‘died’, six per cent drowned, and one per cent were ‘killed’.
Analysis of Liverpool slave voyage muster rolls confirms the accuracy
of these contemporary findings and extends the period of examination to the
years from 1780 to 1807. For these years there are muster rolls for 1,709
Liverpool slave voyages which completed the triangular journey. Evidence
for these voyages is presented in Table 4. As the table shows, 17.8 per cent
or 10,439 out of the 58,778 crew who entered pay in Liverpool died during
these voyages. The average duration of these voyages was about a year;
mortality rates, as shown in Table 4, were 14.7 crew deaths per thousand
crew per month. In the Liverpool slave trade, crew mortality declined over
the two decades prior to abolition, from 15.7 deaths per thousand per month
in 1780-93 to 13.8 deaths in 1794-1807. There are additional crew mortality
data for 254 Liverpool slave voyages which did not return to England.
Including these with those for completed voyages, it appears that, in all,
there were 12,234 deaths among the 66,323 crew reported sailing from
Liverpool on ,963 Liverpool voyages, a mortality loss of 1 8.4 per cent.-*^ Of
1

the total of 12,234 deaths, 10,639 crew were reported to have ‘died’, 1,314
to have drowned, and 281 to have been killed. A smaller sample of Bristol

and Liverpool musters indicates that of the crew who ‘died’, about four in
five deaths were from ‘fevers’ (mainly yellow fever or malaria), one in ten
CREW MORTALITY IN THE ATLANTIC SLAVE TRADE 55

were from ‘fluxes’ (mainly dysentery), and a small proportion arose from a
variety of respiratory, urological, or diet-deficiency diseases.^" Also, in the
period from 1780 to 1807, 20 Liverpool vessels were lost at sea with the loss
of 349 crew by drowning.^' Most other drownings occurred while vessels
traded at the African coast. About half of the crew reported killed were
killed by slaves. From the sample of 1,709 Liverpool slave voyages, crew
mortality data for 1,651 voyages where the African location of trade could
be determined were examined.” Shown in Table 5, the results reveal large
regional variations in crew mortality rates. Crew members embarking at
Liverpool, for example, were three times more likely to die on slaving
voyages to the Gambia River than on voyages to the Gold Coast. As in the
French trade, trew mortality levels on ships trading at the Gold Coast were
comparatively low.

TABLE 4
VOYAGE CREW MORTALITY LOSS ON LIVERPOOL SLAVE VOYAGES, 1780-1807

Year Voyage sample Mortality loss (%) Triangular Mortality

voyage (days) rate^

1780 13 23.7 442.9 16.1

1781 21 15.0 385.2 11.7
1782 21 9.9 305.8 9.7
1783 64 19.9 388.4 15.4
1784 48 21.4 350.0 18.3
1785 53 21.0 367.5 17.1
1786 59 24.1 447.1 16.2
1787 56 22.3 404.1 16.6
1788 60 19.4 386.8 15.0
1789 55 14.3 367.7 11.7
1790 80 17.4 360.2 14.5
1791 77 18.2 348.0 15.7
1792 97 22.4 388.6 17.3
1793 29 23.6 360.8 19.6
1794 46 19.4 365.9 15.9
1795 33 14.9 354.8 12.6
1796 52 14.5 357.7 12.2
1797 56 10.6 322.7 9^9
1798 104 17.1 347.4 14.8
1799 83 20.0 378.2 15.9
1800 87 18.2 371.3 14.7
1801 92 16.9 333.0 15.2
1802 98 14.0 324.8 12.9
1803 52 12.9 354.5 10.9
1804 78 17.3 353.3 14.7
1805 64 16.8 339.2 14.9
1806 83 15.1 352.3 12.9
1807 48 17.0 424.5 12.0

Totals 1,709 17.8 364.2 14.7

a. Crew deaths per 1,000 crew per 30-day period.

Sample: 10,439 deaths of 58,778 crew who sailed from Liverpool on 1,709 slave voyages,
1780-1807, which returned to Liverpool after landing slaves in theAmericas.
56 ROUTES TO SLAVERY
TABLE 5

VOYAGE CREW MORTALITY RATES IN THE LIVERPOOL SLAVE TRADE, 1780-1807

(by African region)

African region Voyaf’e sample Marta lit}' loss (%) Triangular Mortality
voyage (days) rate^

Gambia River 19 30.3 330.9 27.5

Sierra Leone 89 25.5 342.6 22.3

Windward Coast 193 21.3 368.3 17.3

Gold Coast 129 10.7 407.8 7.9

Bight of Benin 82 21.6 433.9 14.9

Bight of Biafra 729 18.4 349.1 15.8

West-Central
Africa 410 14.9 361.4 12.4

Totals 1,651 17.7 362.7 14.6

a. Crew deaths per 1,000 crew per 30-day period.

Note: African regions arranged from north to south.

Sample: 10,109 deaths of 56,968 crew who sailed from Liverpool on 1,651 slave voyages,
1780-1807, which returned to Liverpool after landing slaves in the Americas.

To examine crew mortality during the triangular voyage with more

precision, a sub-set of 158 Liverpool slave voyages was created for the

period 1770-75, with five variables to record the number of crew at risk per

voyage leg. A sixth variable was added, the voyage identification number,
to link this sub-set of 158 records to the Du Bois Institute set which includes
variables for crew deaths per voyage leg, dates of sail between legs, and
locations of trade. We can thus calculate death rates throughout the voyage
and by African region. On the outward, middle and homeward passages, the
number of crew at risk takes into account the timing of sailor deaths. On the
African coast and in the Americas, the number of crew at risk takes into
account crew deaths, desertions, discharges, and the dates when ‘new’ crew
enlisted. Most deaths occurred during months of trade on the
the first

African coast. In the Americas, most crew deserted ship or were discharged
within a few days or weeks of arrival. New crew usually entered pay close
to the dates of sail from England, Africa or the Americas. The number of
crew-days per voyage leg were calculated, because deaths, desertions and
discharges do not occur at regular intervals.”
Table 6 presents the new data on crew mortality by voyage leg from the
sub-set of Liverpool voyages in 1770-75. Mortality losses and rates are
calculated from the original crew on each voyage leg; death losses and rates
are calculated from the crew at risk totals. Regarding the percentage
distribution of crew deaths by voyage leg, this Liverpool sample is
CREW MORTALITY IN THE ATLANTIC SLAVE TRADE 57

consistent with the French slave trade data presented earlier, as a majority
of deaths occurred at the African coast and
middle passage. Death in the
rates were highest on the African coast, averaging 45.8 crew deaths per
thousand crew per month and, at these levels, were three times the French
coastal mortality in1715-78 calculated by Stein and twice the British rates
in 1792-96 calculated by Steckel and Jensen. Middle passage death rates
of 28 crew per thousand crew per month in the 1770s were also greater than
these French and British samples, but less than the eighteenth-century
voyage mortality of convicts and soldiers cited earlier.’^

TABLE 6
CREW MORTALITY IN THE LIVERPOOL SLAVE TRADE, 1770-75
(by voyage leg)

Voyage leg Beginning At risk Crew Mortality Death Voyage Mortality Death
crew crew deaths loss % loss % duration (days) rate‘s rate^

Outward passage 4,660 4,645 30 0.64 0.65 67.5 2.8 2.9

African coast 4,655 3,969 1,300 27.93 32.75 214.5 39.1 45.8
Middle passage^ 3,205 3,139 151 4.71 4.81 51.7 27.3 27.9
Americas 3,055 2,381 91 2.98 3.82 45.4 19.7 25.2
Homeward passage 2,439 2,420 44 1.80 1.82 53.3 lO.I 10.2

a. Rates per 1,000 crew per 30-day period.

b. Middle passage defined as the voyage from the last African location to the first port of arrival
in the Americas.

Note: Outward passage and Middle Passage voyage duration (by African region) calculated
from Liverpool and British slave voyage data, 1770-1807. Mortality losses and rates
calculated from beginning crew totals; death losses and rates calculated from crew at
risk totals.

Sample: Crew mortality data from 158 Liverpool slave voyages, 1770-75, which completed the
triangular journey. Totals include all crew who entered pay at any point of the voyage.

The monthly crew death rates in the Americas were, at 2.5 per cent,
greater than those for most other migrant groups in the late eighteenth and
early nineteenth centuries. This rate is, for example, two to three times
higher than that of British troops in the West Indies in 1819-36, before the
‘mortality revolution’ later in that century.^^ The crew mortality in the
Americas in 1770-75 was not exceptional, however, when measured against
the catastrophic mortality of British soldiers during the yellow fever
epidemic in St. Domingue in the mid- 790s. 1 Indeed, we will probably find
a sharp rise in crew mortality for all voyages trading in the Caribbean during
wartime as frequent troop arrivals maintained human population densities
necessary to sustain the yellow fever virus. Furthermore, the threat of
impressment into the British navy increased the movement of sailors and
58 ROUTES TO SLAVERY
contagious disease West Indian harbours and on land.^” Death rates on the
in

homeward passage would also presumably increase during war years. In the
peace years of 1770 to 1775, the monthly death rate on the homeward
passage of ten crew deaths per thousand crew is comparable to the mortality
level on twelve British convict voyages 1768-75 when
to North America in

the death rate averaged 12.5 deaths per thousand per month. As can be
seen in Table 6, the crew death rates on the homeward passage were three
times greater than the rates on the outward passage. Most crew deaths on the
final voyage leg were from yellow fever or malaria acquired in Africa, on

the middle passage, or in the Americas.

In the sample of 158 British slave voyages in 1770-75, all but five
voyages slaved within one African region, allowing analysis of regional
variation in crew mortality.'^ Data on crew mortality by African trading
region are provided in Table 7. These indicate that death rates varied along
the African coast from a low of 1 1 .8 deaths per thousand crew per month on
the Gold Coast to 84-85 deaths per thousand in the Gambia River and at the
Bight of Benin.^' Though based on a small sample of 17 voyages, the death
rates in the last two regions are, outside the catastrophic troop mortality
during the yellow fever epidemic in St. Domingue, the highest recorded for
any seagoing population or migrant group.'^-

TABLE 7

COASTAL CREW MORTALITY IN THE LIVERPOOL SLAVE TRADE, 1770-75

(by African region)

African ref^ion Voyage Beginning At Crew Mortality Death African Mortality Death
sample crew risk deaths loss (9c) loss (9c) coast rate^ rate‘s

crew voxage
duration
( days)

Gambia River 6 160 125 67 41.9 53.6 189.2 66.4 85.0

SierraLeone 7 116 105 17 14.7 16.2 146.0 30.2 33.3
Windward Coast 28 1,071 853 All 44.5 55.9 292.5 45.6 57.3
Gold Coast 15 412 390 30 7.3 7.7 196.0 11.2 11.8
Bight of Benin 11 354 255 165 46.6 64.7 230.4 60.7 84.2
Bight of Biafra^ 78 2,281 2 ,016 509 22.3 25.2 192.2 34.8 39.3
West-Central Africa 8 202 179 33 16.3 18.4 230.8 21.2 23.9

a. Mortality and death rates per 1,000 crew per 30-day period.
b.Death rates for select Biafran ports are as follows: Old Calabar, 26 voyages (47.5); Bonny, 22
voyages (33.1); Cameroons, 18 voyages (33.5).

Note-. African regions arranged from north to south. Mortality losses and rates calculated from
beginning crew totals; death losses and rates calculated from crew at risk totals.

Sample-. Crew mortality data from 153 Liverpool slave voyages, 1770-75. Totals include all
crew who entered pay at any point of the voyage.
CREW MORTALITY IN THE ATLANTIC SLAVE TRADE 59

As most crew died from yellow fever and malaria, it is reasonable to

assume that large variations in regional mortality, for example, between the
Gambia River and the Gold Coast, relate to the frequency of contact
between non-immune sailors and Aedes and Anopheles mosquito vectors.
Mosquito population densities, moreover, relate directly to precipitation
levels (or flooding).-*^ In the late eighteenth century the British occupied
James Fort at the mouth of the Gambia River as well as a ‘considerable
number of establishments’ on both banks. British vessels slaved along the
Gambia as far as Barokunda 300 miles upriver."^ The first 100 miles
Falls,
of the lower Gambia are bordered by mangrove forests, saline swamps, and
and the riverside lands thereafter are characterized for 200
alluvial flats,
miles by open fresh-water swamps and grassy flats. From recent surveys,
average annual precipitation at the mouth of the Gambia River was 53
inches; and there are major spring tides in the lower Gambia which dictate
The most important malaria vectors in the Gambia are the
flood levels.^^
freshwater-breeding Anopheles gambiae and the salt water breeding
Anopheles melas. Studies have shown that A. gambiae live during the rainy
season from June to October and have a higher infection rate because they
live longer and are less zoophilic. A. melas mosquitoes can live for a few

months into the November-May dry season and are the main malaria vector
in the western Gambia outside the usual summer transmission period.
Comparatively few anopheline mosquitoes, however, live during the six or
seven months of the dry season. In the Gambia forests live large numbers of
monkeys and baboons which act as the reservoir hosts of the yellow fever
virus. Aedes aegypti thrive during the rainy season though inter-human
transmission by A. aegypti been documented in the
also has
November-May dry season. Malaria and yellow fever still pose major
health problems in the Gambia today.-^ Given this disease environment, it is
not surprising that in the late eighteenth century many non-immune British
crew trading up the Gambia River died at epidemic rates.
By contrast, the climate and physical geography of the British slave-
trading regions along the Gold Coast (or modern Ghana) may have
supported a less dense mosquito population. Furthermore, mariners had
limited on-shore contact. At Accra precipitation levels measured about 30
inches a year on average in 1888-1960, and rain falls more evenly
throughout the year than in the Gambia, allowing year-round transmission
of yellow fever and malaria in Ghana.-** The British organized the slave trade
on the Gold Coast through forts and factories along a 200-mile coastline
from Apollonia in the west to Accra in the east. By the late eighteenth
century, trade centred around Anomabu fort. Slaves confined at these
locations were sold to coastal traders, fort officials and British merchants.
As the Gold Coast lacks large harbours, British and other European vessels
60 ROUTES TO SLAVERY

anchored well away from the pounding surf. Forts at Dixcove and Sekondi
were accessible, but only to small coastal vessels. Cape Coast Castle was
the administrative centre of British trade; there resided an African Company
surgeon who managed the ‘hospital’ and store rooms. Crew on board
British (and French) slavers on the Gold Coast, therefore, died at lower rates
because they spent less time on shore in, probably, a comparatively less
deadly disease environment. The coastal forts and factories also could
supply captains with provisions and medical supplies.
We have seen that there crew mortality by year,
were variations in

voyage leg, season of sail from Africa, and region of trade on the African
coast. A sub-set of 313 Liverpool slave voyages from the early 1800s was
also created to assess whether crew mortality varied by rank. In this file, I
recorded the voyage histories of each crew member in six variables: a rank
variable; the date of discharge, death, or desertion; and boolean variables
assessing whether there was a crew death, and whether the crew member
‘died’, drowned, or was killed.^" A seventh variable was added, the voyage
identification number, to link this sub-set Institute of 11,399 records to the
Du Bois set, upon the dates of departure of vessels from
thus drawing
Liverpool to calculate voyage time per crew. The results are reported in
Table 8. As the table shows, captains and first mates died at the lowest rate
in the late Liverpool slave trade. Surgeons had the highest mortality; one in

four died on the voyage, a mortality rate double that of first mates. These
differences in mortality are probably due to the degree of immunity each
mariner had to yellow and malarial fevers. Mariners who survived fevers on
the African coast gained immunities to future attacks and were promoted to
higher rank on subsequent slave voyages.^'
Crew making their first voyage in the slave trade - usually surgeons, and
some sailors, coopers and carpenters - had higher mortality. The fact that
surgeons had the highest mortality may not be surprising. These men had
the most contact with slaves and crew who were sick; they entered the slave
trade without previous maritime experience; and they frequently left the
vessel to examine slaves for sale on shore. First mates may have had the
lowest mortality because they gained immunities on previous voyages and
frequently remained offshore in command while the captain organized the
purchases of slaves.
There are two important questions yet to consider in this essay: first,

why did crew mortality decline in the late eighteenth century; and, second,
why did crews on ships in the British slave trade die at greater rates than
those in the French trade?
The decline in British crew mortality more evident when we include
is

mortality data from the sub-set of 158 Liverpool voyages for 1770-75 with
our figures for 1780-1807. For the period 1770-75, the aggregate (original)
 CREW MORTALITY IN THE ATLANTIC SLAVE TRADE 61

crew data reveal that, of the 4,660 crew who

1,493 crew died,
enlisted,
suggesting a loss by mortality of 32 per cent. Given an average voyage
duration of 434.5 days in 1770-75, this suggests a crew mortality rate of
22.1 crew deaths per thousand crew per month. These mortality levels are
higher than for all five-year periods in the French slave trade or for the
British trade between 1780 and 1807.

TABLE 8
VOYAGE CREW MORTALITY IN THE LIVERPOOL SLAVE TRADE
(by rank)

Rank or Station Number Deaths Mortality Days on Mortality-

loss {%) board (ave.) rate^

Captains, supercargoes 325 47 14.5 303.2 14.3

1st mates 318 38 11.9 293.0 12.2
2nd, 3rd, 4th, 5th mates 742 114 15.4 258.1 17.9
Surgeons, surgeons’ mates 333 76 22.8 279.9 24.5
Carpenters, coopers, joiners 669 114 17.0 259.6 19.7
Other crew 9,012 1,430 15.9 240.1 19.8

Totals 11,399 1,819 16.0 246.9 19.4

a. Crew deaths per 1,000 crew per 30-day period.

Sample: Crew mortality data from 313 Liverpool slave voyages, 1801-7.

The declines aggregate crew mortality in the late eighteenth century

in
were due mostly to shifts in the region of trade on the African coast. In both
the French and British slave trades, there were shifts away from the higher-
mortality Senegambia region and Bight of Benin to the lower-mortality
Angolan coast over the course of the eighteenth century. Two-thirds of the
French trade was to the Benin Coast in the early eighteenth century; more
than half of French slavers traded on the Angolan coast in 1760-90.” The
British trade to Angola, which comprised only 5-6 per cent of the total
British trade in 1780-91, increased to about a third of the total trade in
1792-1807.” We also have evidence, however, that mortality rates declined
through time on British ships trading at specific African markets. For
example, closer analysis of the large sample of 729 Liverpool slave voyages
to the Bight of Biafra reported in Table 5 reveals that monthly crew
mortality rates declined from 1.70 per cent in 1780-93 to 1.48 per cent in
1794-1807. Taking only the crew who ‘died’ (that is, excluding crew who
drowned or were killed), these monthly mortality rates are .63 per cent and 1
62 ROUTES TO SLAVERY

1.36 per cent, respectively. Muster roll information suggests that this
decline in aggregate crew mortality was the result of fewer ship-board
epidemics at the Bight of Biafra and all West African regions. For example,
at least half the ‘original’ crew ‘died’ on 26 or 16.5 per cent of the 158

Liverpool slave voyages 1770-75, whereas half the outbound crew died
in

on 37 (4.6 per cent) of 809 Liverpool slave voyages in 1780-93 and on 32

(2.8 percent) of 1,154 Liverpool slave voyages in 1794-1807.'^
What explains the decline in crew epidemics on board British slave
vessels? It is useful, first, to look at specific trading practices on the coast
during the 1770s and early 1780s before addressing this question. James
Stanfield, a mariner in the Liverpool slave trade, believed that crew
mortality was caused primarily by the cruelty of officers, contagion on
board ship because of confined air, the provisioning duties of crew on the
African coast, and the cleaning of ship holds. Stanfield wrote in 1788:

Among the many causes of destruction, which originate from the

trade,and not from the climate, the bulk-heads between the decks,
excluding a salutary circulation of air, have been insisted upon as
producing their effects. But there is another, which has not claimed
such notice, and which yet is a terrible assistant to African mortality.
This is the fabricating of an house over the vessel for the security of
slaves, while on the coast.
This enclosure helps the stagnation of air, and is, in that point of
view, dreadful: but it is more fatal in the act of its preparation. know
I

nothing more destructive than the business of cutting wood and

bamboe, for the purpose of erecting and thatching this structure. The
process is generally by the river-side. The faces and bodies of the poor
seamen are exposed to the fervour of a burning sun, for a covering
would be insupportable. They are immersed up to the waist in mud
and slime; pestered by snakes, worms and venomous reptiles;
tormented by muskitoes, and a thousand assailing insects; their feet

slip from under them at every stroke, and their relentless officers do
not allow a moment’s intermissionfrom the painful task. This
employment, the cruelty of the officers, and the inconceivably
shocking task of scraping the contagious blood and filth, at every
opportunity, from the places where the slaves lie, are, in my opinion,
the three greatest (though by no means the sole) causes of the
destruction of seamen, which this country experiences by the
prosecution of the trade in slaves.

Stanfield also attributed crew mortality to poor nourishment, a scarcity of

water, crowded living conditions, and the fact that most of the crew slept
upon deck.”
CREW MORTALITY IN THE ATLANTIC SLAVE TRADE 63

Stanfield’s
pamphlet, published by the abolitionist James Phillips,
perhaps influenced Members of Parliament to amend the Dolben Act of
1788.^^ The Dolben Act had on the number of slaves British
set limits
vessels could transport from the coast, effectively reducing the number of
slaves loaded per ton by 13 per cent. Consequently, the number of crew
shipped per ton declined by 17 per cent. Furthermore, the Act required that
surgeons certify for the slave trade. In 1789 an amendment stipulated the
terms of contract between captain and crew. Among the clauses, sailors
were required to receive a regular diet of daily provisions and sleeping
accommodations for at least half the crew below deck."^ The captain also
agreed to ‘hire and employ the Natives in their Craft, to wood and water the
said Ship, during her Continuance on the Coast of Africa, if such Natives
can be procured’ and ‘whenever the Officers and Seamen are employed
trading in Craft up the Rivers’ the captain will ‘furnish the Parties so
employed with a sufficient Quantity of painted Canvas, or Tarpawling, for
an Awning, and Provisions for the Time’. In 1799 Parliament passed a
further regulation act which increased aerial space below deck for slaves
and set a maximum slave per crew ratio of ten to one as well as a maximum
loading level of 400 slaves. The new space requirements reduced loading
levels to aboutone slave per ton; this amounted to a 47 per cent reduction
on pre-1799 levels. With fewer slaves to control, crew size declined by 13
per cent.^**

Parliament passed these three acts to reduce crowding, the amount of

‘contagion’ generated below deck, and the cycle of disease transmission
between slaves and crew. Did they improve the health and working
conditions of sailors? Several studies have shown that there is no
relationship between ship-board mortality and crowding.^’ The minimum
diets required by Dolben Act amendment were still
the nutritionally
inadequate and would not have increased sailors’ ability to withstand
yellow fever or malaria.^® The medical qualifications of surgeons certainly
improved after the Dolben Act. Some surgeons entered the slave trade after
a period of hospital study where they learned about the relationship between
hygiene, sanitation and epidemic diseases. Despite this, doctors would not
have been able crew with yellow fever, though tonics of Peruvian,
to treat
or cinchona, bark, which we know now contains quinine, may have proved
effective against malaria.*'
Perhaps the improved health of slaves, suggested by the lower slave
mortality on the middle passage in the 1790s, reduced transmission of
fevers and gastro-intestinal diseases between slaves and crew during and
after slave loading on the African coast.** But why did slave mortality
decline in the 1790s? The 1788 Dolben Act raised shipping costs*’ and may
have prompted British slaving merchants to pay greater attention to hygiene
64 ROUTES TO SLAVERY

and sanitation on board ship in an attempt to lower slave mortality.

Specifically, coal tars may have been used in the last two decades of the
British slave trade as disinfectants. Archibald Cochrane developed a new
process of manufacturing tar from coal in 1781 and he believed that tar

would protect hulls from worms that lived in warm waters.^ In the late
1780s British manufacturers offered to supply the African Company
storeships with various coal tars and varnishes. The British Coal Tar Office
claimed that soaking planks in coal tars ‘destroys insects and their eggs, and
by empoisoning Wood to a certain depth, prevents the entrance of those
worms which feed on, and lodge in wood’. They recommended that other

coal tars could be used ‘for Ships masts, sides, and decks which, in warm
Climates, are apt to rent and open, by the heat of the sun’ and to blacken
yards and mast-heads. Painting and fumigating vessels with coal tar may
have killed bacteria and mosquitoes on board ship, thus reducing chances of
an outbreak of amoebic dysentery, malaria, or yellow fever. Recent work on
English voyages to Australia has suggested that improved shipboard
hygiene and sanitation in the mid-nineteenth century, promoted by
government agencies, helped lower passenger mortality.^ Fifty years earlier,
British merchants also may have employed new methods to disinfect slave
vessels effectively. When William Young inspected the Bristol slave ship
Pilgrim, anchored in Kingstown Bay, St. Vincent in December 1791, he
found the interior of the ship ‘as clean as a Dutch cabinet’.^’ Though the
Pilgrim slaved for eighty-two days at Bonny in the Bight of Biafra during
the rainy season, the vessel lost only five of 371 slaves on her sixty-four day
passage to America. Moreover, only one sailor died (on the homeward
passage) of the 35 crew who shipped out from Bristol.^* By the last years of
the trade, medical authorities, in fact, contrasted the cleanliness of British
slavers withgovernment transports.
The clause in the Dolben Act amendment requiring captains to hire
Africans to provision slave vessels on the coast would have reduced the
contact of non-immune crew with African disease environments. Captain
Robert Hume stated that during his voyages in the 1790s he employed sixty
to eighty African sailors to trade for provisions along the Windward Coast.
This practice he deemed necessary to avoid ‘burying a great Number of
White Men’.^" Hume’s commercial transactions with African sailors may
not have been prompted by British legislation, however. Rather, the
development of African coastal economies may have led to a reorganization
of trading practices. As Klein has suggested in regard to the long-term
decline in slave mortality,^' epidemics may have been less frequent in the
late British slave trade because of efficiencies gained in marketing slaves on
the African coast. In Table 4 we see that during the years of lowest crew
mortality - 1782 and 1797 - British slavers made the fastest triangular
 CREW MORTALITY IN THE ATLANTIC SLAVE TRADE 65

voyages. This suggests that there were minimal delays in loading slaves.'- If
African traders could increase productivity in provisioning British vessels
and supplying slaves, British merchants could organize voyages to
minimize shipping costs and perhaps avoid African rainy seasons.'^
Moreover, in regions such as the Gambia River and the Windward Coast,
there would be less need for crew to travel up rivers to seek out slave
suppliers. If African merchants increased the regularity of slave supplies
from the upper to the lower Gambia, for example, fewer British vessels
would have to sail along the fresh water swamplands where A. gambiae
thrived.
Why was crew mortality in the French slave trade lower than that in the
Liverpool trade in 1771-75 and 1781-93? In the first period, mortality rates
in the Liverpool trade were three times greater than in the French trade,

while in the later period, they were twice as great. In part, these mortality
differences are explained by varying voyage leg duration. Triangular French
slave voyages took, on average, 80-100 more days to complete than
Liverpool voyages, primarily because French vessels remained in the West
Indies for many months. On the West African coast, French captains
sometimes completed trade more quickly than their Liverpool counterparts.
For example, the ships involved in 143 French slave voyages in 1771-75
remained on the African coast for 170 days on average, whereas, as Table 6
shows, 158 Liverpool slavers took 215 days on average to complete their
trade on the coast during more or less the same period. French slavers thus
spent a greater part of their triangular voyages in less deadly disease
environments. French slave ships also may have been manned by greater
proportions of ‘seasoned’ crew who had
gained immunities to yellow or
malarial fevers on previous voyages to Africa or the Americas. In the
British trade, captains frequently hired ‘landmen’ who were not apprenticed
to sea and may have died at high rates.'" There may have been
comparatively few landmen on French vessels. It is also possible that
French merchants paid greater attention to ship-board medical care, hygiene
and sanitation during the period from 1771 to 1793.
Even though there was a crew mortality in the late
reduction in
eighteenth-century French and British slave trades, merchants organized
slave voyages to maximize profitability regardless of potential crew loss.
European vessels continued to slave in regions of Africa such as the Gambia
River and Bight of Benin which were comparatively unhealthy for sailors.
Merchants also organized some voyages to arrive in Africa during the rainy
season. For example, they tried to send slave vessels to Bonny in June or
July so that they would arrive there alter the harvest when yam stocks were
plentiful. By arriving at Bonny
August or September, slave vessels would
in

be expected to reach the Americas during the December-January sugar

66 ROUTES TO SLAVERY

harvest when demand for slaves was relatively high. The primary aim of
merchants in the late eighteenth century was
minimize slave deaths in the
to

middle passage to ensure a profitable voyage. Minimizing crew mortality

was a secondary consideration.

NOTES
1. Raymond Cohn, ‘Maritime Mortality in the Eighteenth and Nineteenth Centuries: A
L.
Survey’, International Journal of Maritime History, (1989), pp. 159-91; Robin Haines,
1

Ralph Shlomowitz, and Lance Brennan, ‘Maritime Mortality Revisited’, International

Journal of Maritime History (1996), pp. 133-72.
2. Richard H. Steckel and Richard A. Jensen, ‘New Evidence on the Causes of Slave and Crew
Mortality in the Atlantic Slave Trade’, Journal of Economic History, 46 (1986), pp. 57-77.
3. Stephen D. Behrendt, ‘The Captains in the British Slave Trade from 1785 to 1807’,
Transactions of the Historic Society of Lancashire and Cheshire, 140 (1991), pp.79-140.
More than one in four captains died in the trade during these years.
4. Dieudonne Rinchon, Le Trafic Negrier, d’apres les Livres de Commerce du Capitaine
Gantois Pierre-Ignace-Lievin van Alstein, (Paris, 1938), pp.l46, 247-96; Philip D. Curtin,
1

The Atlantic Slave Trade: A Census (Madison, Wisconsin, 1969), pp. 282-6; Herbert S.
Klein, The Middle Passage: Comparative Studies in the Atlantic Slave Trade (Princeton,
1978), pp.l80, 190-207; Robert L. Stein, The French Slave Trade in the Eighteenth Century:
An Old Regime Business (Madison, Wisconsin, 1979), pp. 98-100; Herbert S. Klein and
Stanley L. Engerman, ‘A Note on Mortality in the French Slave Trade in the Eighteenth
Century’, in Henry A. Gemery and Jan S. Hogendom (eds.), I'he Uncommon Market: Essays
in the Economic History of the Atlantic Slave Trade (New York, 1979), pp. 266-8.

5. W.S. Unger, ‘Bijdragen tot de Geschiedenis van de Nederlandse slavenhandel 11’,

Economisch-Historisch Jaarboek, 28 (1958-60), pp.26-7; Johannes Postma, ‘Mortality in
the Dutch Slave Trade, 1675-1795’, in Gemery and Hogendom (eds.). Uncommon Market,
p.260. Postma published slightly revised captains’ mortality figures in a later work {The
Dutch in the Atlantic Slave Trade, I600-I8I5 (Cambridge, 1990), pp. 156-7, 377-89). As
Postma notes, there are few surviving records which document crew mortality in the Dutch
slave trade. Many records, such as logbooks of the Dutch West India Company, have been
lost almost entirely. Sources which detail crew mortality for other transatlantic slave trades

also are lacking.

6. Jean Mettas, ‘Honfleur et la Traite des Noirs au XVllle Siecle’, Revue fran^aise d’histoire
d’outre-mer, 60 (1973), pp.20-2; idem, ‘Pour une Histoire de la Traite des Noirs Fran^aise;
10.
Sources et Problemes’, Revue frangaise d'histoire d’outre-mer, 62 (1975), pp.38-9; Robert
Stein, ‘Mortality in the Eighteenth-Century French Slave Trade’, Journal of African History,
21 (1980), pp. 35^1.
7. In his notes, Mettas recorded the dates and locations of crew deaths in the French slave trade.
The Dagets, however, grouped crew deaths by voyage leg in the two-volume catalogue. Jean
Mettas in Serge and Michele Daget (eds.). Repertoire des Expeditions Negrieres Frangaises
au XVIIIe Siecle, 2 vols. (Paris, 1978, 1984), 1, pp.xv-xvi.
8. In 1747 Parliament ordered that customs houses record ‘ship’s agreement and crew lists’ in
‘An Act for the Relief and Support of maimed and disabled Seamen and the Widows and
Children of such as shall be killed, slain or drowned in the Merchants Service’ (20 Geo. II
C.38, XX). For British voyages in the eighteenth century, there are also surviving muster rolls
from the ports of Dartmouth, Plymouth, and Shields. Public Record Office (hereafter PRO)
information leaflet no. 5, p.3.
9. PRO, BT 98/33-69; Behrendt, ‘Captains in the British Slave Trade’, p.81.
Merchant Venturers Hall, Bristol. Kenneth Morgan,
Bristol musters are kept at the Society of
‘Shipping and the Atlantic Trade of Bristol, 1749-1770’, William & Mary
Patterns
Quarterly, 46 (1989), pp.532^; David Richardson (ed.), Bristol, Africa and the Eighteenth-
CREW MORTALITY IN THE ATLANTIC SLAVE TRADE 67
Century Slave Trade to America, Bristol Record Society, vol.3 (1991), p.x.
1 1 For example, during war years sailors frequently deserted ship upon arrival in the Americas.
12. Crew deaths were often not recorded on the Liverpool muster roll when the vessel was
captured on the outward passage or middle passage. For some musters from the 1780s, it is
not possible to differentiate between crew discharges or deaths.
13. I have not yet determined crew deaths by voyage leg from all Liverpool slave voyage muster
rolls.

14. For example, Trade, Mediterranean Passes record the number of crew at the
in the British
outset of the voyage. Naval Office shipping lists report the number of crew who arrive in the
Americas, and Seamen’s Sixpence ledgers enumerate the crew who returned.
15. Curtin, Atlantic Slave Trade, p.283. See also Klein and Engerman, ‘Note on Mortality’
p.263.
16. 1 would like to thank David Eltis for calculating the crew mortality data for the eighteenth-
century French slave trade.
17. Klein’s analysis of the Nantes data, 1711-77^ indicated that more than half the crew deaths
occurred on the African coast and about a third took place on the middle passage (Klein,
Middle Passage, p.l97 n.44). From new French slave trade data on 1,730 voyages, we
calculate that 3 per cent of crew deaths occurred on the outward passage; 42 per cent on the
African coast; 22 per cent in the middle passage; 26 per cent in the Americas; and 7 per cent
on the homeward passage.
18. Crew deaths per thousand crew per thirty-day period (approximating a month) is the standard
measurement of maritime mortality. Cohn, ‘Maritime Mortality’, p.l87. Regarding the
importance of distinguishing between mortality losses and rates in the transatlantic slave
trade, see Joseph C. Miller, ‘Mortality in the Atlantic Slave Trade: Statistical Evidence on
Causality’, Journal of Interdisciplinary History, 11 (1981), pp. 385-409.
19. This contrasts with a middle passage mortality rate of 18.5 crew deaths per thousand crew
per month on 130 Nantes voyages in 1715-78. Stein, ‘Mortality in the French Slave Trade’,
pp.37, 40.
20. Klein, Middle Passage, p.71; Cohn, ‘Maritime Mortality’, pp. 188-9; Haines, Shlomowitz,
and Brennan, ‘Maritime Mortality Revisited’, p.l35.
21. Monthly rainfall data for the mid-twentieth century are found in Great Britain
Meteorological Office, Tables of Temperature, Relative Humidity and Precipitation for the
World (London, 1958), Part IV, pp.13-15, 120-5. For Benin, the sample of 352 French slave
voyages yielded crew death rates on the middle passage of 16.9 and a standard deviation of
20.7. For West-Central Africa, the sample of 557 voyages yielded death rates of 17.8 and a
standard deviation of 29.4. We excluded voyages which were lost on the passage. In the
Bight of Benin sample, standard deviations ranged from 10.8 (in November, during the dry
season) to 41.7 (in July, the height of the rainy season). In West-Central Africa, standard
deviations ranged from 1 1 .7 (in July, during the dry season) to 37.2 (in December, during the
rainy season). There was thus greater variance around mean crew mortality rates during the
rainy season, suggestive of epidemic mortality on some voyages.
22. Mettas, ‘Honfleur et la Traite des Noirs’, p.22; Mettas, ‘La Traite des Noirs Frangaise’, p.38;
K.G. Davies, ‘The Living and the Dead; White Mortality in West Africa, 1684-1732’, in
Stanley L. Engerman and Eugene D. Genovese (eds ). Race and Slavery in the Western
Hemisphere: Quantitative Studies (Princeton, 1975), pp.94-6. As Davies notes,
contemporaries frequently commented on the fatal rainy seasons in Africa. Steckel and
Jensen, however, suggest that rain did not influence the chances of crew deaths from fevers
or gastrointestinal diseases, the two most common causes of death. Steckel and Jensen,
‘Slave and Crew Mortality’, p.67.
23. R. Mansell Prothero, Migrants and Malaria in Africa (Pittsburgh, 1965), p.lOO; Walther H.
Wemsdorfer and Ian McGregor (eds.). Malaria. Principles and Practice of Malariology, 2
vols. (London, 1988), II, pp. 1384-5.
24. Thomas Clarkson, An Essay on
the Impolicy of the African Slave Trade (Freeport, New York,
1971 edition), pp.5()-2. These 88 Liverpool voyages sailed to Africa from 7 December 1784
to 23 December 1786 (PRO, BT 98/46, No.8; PRO, BT 98/47, No.340). I re-examined the
88 musters and found that Clarkson made only a few errors in counting. He also mistakenly
 .

68 ROUTES TO SLAVERY
included data from two African produce vessels in his totals.

25. Clarkson, Impolicy of the African Slave Trade, pp.53-5.

26. Clarkson estimated that more than 1,500 crew left their slave vessels in Africa and the West
Indies (they either deserted or were discharged) and that most of these sailors died or ‘were
rendered unserviceable at home’ through illness. In all, he stated that ,950 out of 5,000 slave
1

trade crew ‘were lost to the service of [England] by the prosecution of the slave trade in the
year 1786’. Clarkson also calculated that more than twice as many sailors died in the slave
trade than in the East India, West India, Petersburg (timber), Newfoundland (fishing) or
Greenland (whaling) trades combined. Ibid., pp.54— 75).
27. Roger Anstey, The Atlantic Slave Trade and British Abolition (Atlantic Highlands, New
Jersey, 1975), p.32, n.95.
28. Journal of the House of Commons, 46 (1790-91), pp.337, 397, 433. In the debate on the
abolition of the slave trade on 18-19 April 1791, speakers in the Commons and Lords cited
these aggregate crew mortality data. Parliamentary History of England, 29 (1791-92),
p.270; The Senator; Clarendon s Parliamentary Chronicle, 2 (1791), pp. 569-70, 604-5.
Or,
29. Including mortality in the Bristol and London slave trades, estimate that there were about
I

20,000 crew deaths in the British slave trade, 1780-1807.

30. Stephen D. Behrendt, ‘The British Slave Trade, 1785-1807: Volume, Profitability, and
Mortality’ (unpublished Ph.D. thesis. University of Wisconsin-Madison, 1993), pp.330, 341.
See also Steckel and Jensen, ‘Slave and Crew Mortality’, p.62.
31. The greatest loss of life occurred on the slave ship (and letter of marque) Saint Ann (249
tons). All 61 crew on board drowned when the ship was lost on the homeward passage in
1798 (PRO, BT 98/59, No. 342; William Fergusson Irvine (ed ). An Index to the Wills and
Inventories now preserved in the Probate Registry at Chester, 1791-1800 (Liverpool, 1902),
will of Captain Robert Jones).
32. Sources report the locations of African trade for four in five British slave voyages,
1780-1807. For most other voyages, I based my estimates on the location of slaving on
reported African destinations from England and/or slave voyage histories of captains,
shipowners, merchants, and vessels (see Stephen D. Behrendt, ‘The Annual Volume and
Regional Distribution of the British Slave Trade, 1780-1807: ‘The Annual Volume and
Regional Distribution of the British Slave Trade, 1780-1807’, Journal of African History
forthcoming).
33. For example, a slave vessel arrived on the African coast with fifteen crew and remained there
for six months. The original crew on voyage leg 2 (African coast) is fifteen. On the Coast,
four sailors deserted ship after two months and two sailors died after two months. On the day
of sail from Africa, three crew were hired to sail on the middle passage. Thus, nine crew
remained on the coast for six months ( 80 days) and six crew remained for only two months
1

(60 days). The total number of crew-days is 1,980. As

remained on the coast for
the vessel
180 days, the crew at risk on the coast is eleven (1,980 crew-days/180 days). It is standard
procedure to define the average shipboard population at risk as the number of passengers
embarked less half the voyage deaths. This method examines the mid-voyage population,
and assumes that deaths occurred at regular intervals during the voyage leg (Cohn, ‘Maritime
Mortality’, pp. 162-3; Haines, Shlomowitz, and Brennan, ‘Maritime Mortality Revisited’,
p.l36). As I calculated the number of crew-days per outward, middle, and homeward
passages, my crew at risk totals do not always equal the mid-voyage population.
34. Stein, ‘Mortality in the French Slave Trade’, pp.37, 40; Steckel and Jensen, ‘Slave and Crew
Mortality’, p.61
35. See above note 20.
36. Philip D. Curtin, ‘Epidemiology and the Slave Trade’, Political Science Quarterly, 83
(1968), p.203.
37. David Geggus, ‘Yellow Fever in the 1790s; The British Army in Occupied Saint Domingue’,
Medical History, 23 (1979), pp.45-6; idem, ‘The Cost of Pitt’s Caribbean Campaigns,
1793-1798’, Historical Journal, 26 (1983), pp. 699-706.
38. Slave trade surgeon, Elliot Arthy, listed a variety of factors which he believed contributed to
high crew mortality in the Americas: ‘the desertion of seamen from one merchantman to
another, in order to obtain large sums of money for the run-home from the West-Indies; their
CREW MORTALITY IN THE ATLANTIC SLAVE TRADE 69

leaving their ships, and going on-shore, to avoid being impressed; their exposure to the
causes of the Yellow Fever in the boating duty; their disorderly conduct when, on those
several occasions, they are on-shore; the badness of their accommodations both in health and
sickness; [and] their want of proper medical assistance and attendance’. Elliot Arthy, The
Seamens Medical Advocate (London, 1798), pp. 146-9. Perhaps with these mortality
concerns mind. Parliament passed an act in 1797 to try to prohibit the desertion of seamen
in
from Bntish merchant vessels in the West Indies (37 Geo. Ill c.73). This law fined captains
who employed ‘deserted’ sailors and regulated the wages paid to crew for the ‘run home’.
39. Haines, Shlomowitz, and Brennan, ‘Maritime Mortality Revisited’, p.l35.
40. Locations of slaving are listed in Lloyd's List and in the Liverpool plantation registers. An
important government document reports the intended slaving location for Liverpool slave
voyages, 1750-76 (PRO, BT 6/3, ff 100-29). Almost all Liverpool slave voyages slaved at
their intended destinations.
41. Davies also notes that
in the early eighteenth century the Royal African Company station at

Gold Coast was healthier for Europeans than bases at the Gambia, Sierra Leone, or
the
Whydah. Davies, ‘Living and the Dead’, p.93.
42. Tom W. Shick, ‘A Quantitative Analysis of Liberian Colonization from 1820 to 1843 with
Special Reference to Mortality’, Journal of African History, 12 (1971), pp.5I-7; Davies,
‘Living and the Dead’, pp. 83-98; Philip D. Curtin, Death by Migration: Europe’s Encounter
with the Tropical World in the Nineteenth Century (Cambridge, 1989); Antonio McDaniel,
Swing Low, Sweet Chariot: The Mortality Cost of Colonizing Liberia in the Nineteenth
C^n/wry (Chicago, 1995), pp.91-135.
43. M.W.Service, ‘The Effect of Weather on Mosquito Biology’, in TE. Gibson (ed ). Weather
and Parasitic Animal Disease, World Meteorological Organization, Technical Note No. 159
(Geneva, 1978), p.l53.
44. Philip D. Curtin, Economic Change in Pre-Colonial Africa: Senegambia in the Era of the
Slave Trade, 2 vols. (Madison, Wisconsin, 1975), II, pp.84, 296-7; S.M.V. Golbery, Travels
in Africa. Performed during the years 1785, 1786, and 1787, 2 vols. (London, 1803), II,

pp. 112-13. British vessels anchored near James Island and paid an anchorage fee to Nomi
rulers who controlled territories there. Nomi sailors handled the river trade and they assisted
in navigating vessels upstream. British merchant, William Lyttleton, owned property 450
miles up the Gambia, and may have been the only British slave merchant
on the upper
to live
Gambia. F. William Torrington (ed.). House of Lords Sessional Papers, 1798-99 (Rahway,
New Jersey, 1975), III, pp.l8I-2, evidence of William Lyttleton; Robin Hallett (ed.).
Records of the African Association, 1788-1831 (London, 1964), pp.l32, 146.
45. Great Britain Meteorological Office, Tables of Temperature, p.74; U.S. Department of
Commerce, World Weather Records, 1951-60 (Washington, D.C., 1967), V, p.l74.
46. T.P Monath et al, ‘Yellow Fever in the Gambia, 1978-1979: Epidemiologic Aspects with
Observations on the Occurrence of Orungo Virus Infections’, American Journal of Tropical
Medicine & Hygiene, 29 (1980), pp. 912-5; J.H. Bryan, "Anopheles gambiae and A. melas at
Brefet, The Gambia, and their Role in Malana Transmission’, Annals of Tropical Medicine
and Parasitology, 11 (1983), pp.1-10; B.M. Greenwood and H. Pickering, ‘A Review of the
Epidemiology and Control of Malaria in The Gambia, West Africa’, Transactions of the
Royal Society of Tropical Medicine and Hygiene, 87, Supplement 2 (1993), pp.3-1 1.
47. For example, the brig Maria arrived at Jamaica from the Gambia, ‘with only one white Man
on board, the Captain & the rest of the People having died on the Passage’ (Lloyd’s List, 30
November 1773). In 1805 the Liverpool slave ship Mars was captured by the Spanish: ‘the
Mars had lost 25 of her people by sickness, which is said to have been the cause of her
capture. She was bound from the river Gambia with 200 slaves for Charleston’. Twenty-three
of twenty-nine crew died in August-October 1805 during the rainy season. Three other sick
sailors were sent on shore (Roval Gazette and Bahama Advertiser, 17 January 1806; PRO,
BT 98/66, No.245).
48. Rainfall levels varied along theGold Coast (Ghana). To the west, average precipitation at
Axim, 1943-52, was 83 inches, with half of the rain falling in May-June. Great Britain
Meteorological Office, Tables of Temperature, p.75; U.S. Department of Commerce, World
Weather Records, V, p. 179. Axim was the location of a Dutch fort. Bntish slavers traded
70 ROUTES TO SLAVERY
farther east, closer to Accra. Perhaps the reddish loam soils of the Gold Coast allow greater
drainage than, for example, the clay soils of the Gambia. R.C. Muirhead-Thomson, ‘Recent
Knowledge about Malaria Vectors in West Africa and their Control’, Transactions
of the
Royal Society of Tropical Medicine and Hygiene, 40 (1947), p.512; M.E.C. Giglioli and I.
Thornton, ‘The Mangrove Swamps of Keneba, Lower Gambia River Basin. 1. Descriptive
Notes on the Climate, the Mangrove Swamps, and the Physical Composition of their Soils’,
Journal of Applied Ecology, 2 (1965), pp. 8 1-103. Most documented outbreaks of yellow
fever in Ghana have occurred on the coast and in the region around Accra (David Scott,
Epidemic Disease in Ghana, 1901-1960 (London, 1965), p.49).
49. K.G. Davies, The Royal African Company (London, 1957), pp. 240-52; John Adams,
Remarks on the Country Extending from Cape Palmas to the River Congo (London, 1966),
pp.7, 159.
50. For these 313 Liverpool slave voyages, which sailed from late 1801 through early 1807, the
rank (or station) of each crew member is recorded in the muster rolls. Ranks are listed
infrequently in the Liverpool musters before the year 1801. As shown in Table 8, for
convenience various ranks are grouped into six categories.
51. For example, in the 1790s mariner Hugh Crow survived several fevers and was promoted to
mate of larger Liverpool slave vessels and then to captain in 1798. Memoirs of the Late
Captain Hugh Crow of Liverpool (London, 1970 edition), pp.34, 64-6.
52. David Richardson, ‘Slave Exports from West and West-Central Africa, 1700-1810: New
Estimates of Volume and Distribution’, Journal of African History, 30 (1989), pp.l2, 14.
53. Behrendt, ‘Annual Volume’.
54. Studies on both slave and non-slave ship-board populations have numbers shown that small
of voyages usually experienced epidemic mortality rates and thus mortality on most voyages
was below average. Cohn, ‘Maritime Mortality’, p.l75; Klein and Engerman, ‘Note on
Mortality’, pp. 263-5.
55. James Field Stanfield, Observations on a Guinea Voyage. In a Series of Letters Addressed to
the Rev. Thomas Clarkson (London, 1788), pp. 14-16. On a Liverpool slave voyage in 1786,
another sailor recalled how captains re-allocated provisions to healthy crew and slaves: ‘No
sooner was a wretched sailor’s name entered on the sick list, than the pitiful allowance of a
quarter of a pound of beef, and the small glass of brandy, were denied him, without any thing
being given in lieu thereof. A little bad bread, with a proportionate quantity of water, was
nearly the whole of what the patients had to subsist on’. William Butterworth, Three Years
Adventures, of a Minor, in England, Africa, the West Indies, South-Carolina and Georgia
(Leeds, 1822), p.41.
56. For a sympathetic appraisal of Stanfield’s tract, see the London Monthly Review, 79
(July-December 1788), pp.70-1.
57. The articles of agreement stipulated that crew were to receive ‘a quarter of a pint of spirits,
or half a pint of wine’ per day. Dajly water allowances also were legislated, but the amount
was not specified.
58. Behrendt, ‘British Slave Trade’, pp. 161-3, 342; 28 Geo. Ill c.54; 29 Geo. Ill c.66; 39 Geo.
Ill C.80.
59. Charles Garland and Herbert S. Klein, ‘The Allotment of Space for Slaves aboard
Eighteenth-Century British Slave Ships’, William & Mary Quarterly, 42 (1985), pp. 238^8;
David Eltis, Economic Growth and the Ending of the Transatlantic Slave Trade (New York,
1987), p.l37.
60. J. Freeman and W.F Bynum (eds.). Starsing Sailors. The Influence of Nutrition upon
Watt, J.

Naval and Maritime History (Greenwich, 1981). Undernourished sailors would have a
lowered antibody response to yellow fever. G. Thomas Strickland (ed.). Hunter’s Tropical
Medicine (Philadelphia, 1991), p.948. Nutrition levels do not play a role in malaria infection.
Wemsdorfer and McGregor (eds.). Malaria, I, pp. 759-63.
61. Behrendt, ‘British Slave Trade’, pp. 177-87.
62. On the African coast and middle passage, slaves were at greater risk of dying from
gastrointestinal diseases; crew were at greater risk of dying from fevers. Steckel and Jensen,
‘Slave and Crew Mortality’, pp.60-2. Stein argues that slave and crew mortality levels in the
Nantes trade rose and fell together. Stein, ‘Mortality in the French Slave Trade’, pp. 38-41.
CREW MORTALITY IN THE ATLANTIC SLAVE TRADE 71

63. David Eltis and Stanley L. Engerman, ‘Fluctuations Sex and Age Ratios
in in the
Transatlantic Slave Trade, 1663-1864’, Economic History Review, 46 (1993), p.3I4.
64. Research suggests that coal tar derivatives (phenols) are superior wood disinfectants. Susan
C. Morgan-Jones, ‘Cleansing and Disinfection of Farm Buildings’, in C.H. Collins et al.

Their Use and Evaluation of Effectiveness (London, 1981), pp. 205-9.

(eds.). Disinfectants:
65. PRO, T 70/1556, George Glenny to African Committee, British Coal Tar Office, 7 July 1788;
Description of and Directions for using the Different Kinds of Coal Tar and Varnish,
prepared by the British Tar Company, at their Works, in Shropshire, Staffordshire, and in
Scotland (undated pamphlet, c. 1780s), pp.1-15.
66. Haines, Shlomowitz, and Brennan, ‘Maritime Mortality Revisited’, pp. 142-63.
67. Cited by Richard B. Sheridan, ‘Sir William Young (1749-1815): Planter and Politician, with
Special Reference to Slavery in the British West Indies’ (unpublished paper, 1996). 1 am
grateful to Professor Sheridan for providing a copy of this paper.
68. House of Lords Record
Office, Main Papers, House of Lords, 28 July 1800, f2; Society of
Merchant Venturers, Bristol Muster Rolls, 1791-92, No.lll.
69. James Veitch, A Letter to the Commissioners for Transports, and Sick and Wounded Seamen,
on the Non-Contagious Nature of the Yellow Fever; and Containing Hints to Officers, for the
Prevention of this Disease Among Seamen (London, 1818), pp. 59-64.
70. Torrington (ed ). Lords Sessional Papers, 1798-99, III, p.l02, evidence of Robert Hume.
7 1 . Klein, Middle Passage, p. 160.

72. I estimated a crude voyage efficiency index based on triangular voyage duration and the
approximate number of slaves that Liverpool vessels loaded in 1780-1807. Based on the
sample of Liverpool voyages in Table 4, the years of greatest ‘efficiency’ were 1782, 1781,
and 1797, with ratios of 0.64, 0.85, and 0.88 voyage days/slaves loaded, respectively.
73. In a sample of British slave voyages sailing to the Gambia, only 22 of 103 voyages sailed
from England in May-September, suggesting that captains tried to avoid arriving at the
Gambia during the June-October rainy season. Efficiencies in slave marketing would
determine whether the vessels could sail from the Gambia before the next seasonal rainfall.
74. Sheffield noted that the crew of a slave vessel set out to carry 500 slaves ‘generally consists
of 20 real seamen, and 30 or 40 landsmen, the very dregs and outcasts of the community; the
real sailors are necessary for the navigation of the ship, the others are employed to attend the
negroes on the coast and middle passage’. Lord John (Holroyd) Sheffield, Observations on
the Project for Abolishing the Slave Trade (London, 1791), p.l8.
‘My own nation’: Igbo Exiles in the Diaspora

DOUGLAS B. CHAMBERS

When Olaudah Equiano (c. 1745-97) was taken on board the slave ship that
carried him out of the Bight of Biafra he feared for his life. Everything he
saw that day seemed to confirm these initial fears that the whites (whom he
suspected were evil spirits) had acquired him in order to eat him or perhaps
to sacrifice him to their gods. Equiano also saw, however, other people ‘of
my own nation’ on board which, as he later remembered, ‘in a small degree
gave ease to my mind’. These people, whom he recognized as fellow
‘natives of Eboe’, or ‘Eboan Africans’, or simply ‘my countrymen’, told the
young boy what little they knew: that ‘we were to be carried to these white
people’s country to work for them’; that this country was a far distant one;
that they had their own women there; and that the whites used ‘some spell
or magic they put in the water when they liked, in order to stop the vessel’.'
At the end of Barbados in 1756, Equiano and his
his Atlantic crossing, at
fellow slaves again feared that the whites were planning to eat them. These
fears, Equiano reported, caused ‘much dread and trembling among us, and
nothing but bitter cries to be heard all the night’. Finally, the whites brought
some ‘old slaves from the land to pacify us’; these ‘old slaves’ had
presumably not forgotten their natal language. Equiano wrote: ‘They told us
we were not to be eaten, but to work, and were soon to go on land, where
we should see many of our country people. This report eased us much. And
sure enough, soon after we landed, there came to us Africans of all
languages.’^
Years later, in 1772, when he was
grown man, Equiano went to
a
Kingston, Jamaica, where he was impressed by the numerous Africans he
saw there, and especially how those people grouped themselves by their
particular ethnicities during their free time. On
Sundays, he noted, African
slaves gathered in large numbers at a general meeting place outside of town.
‘Here’, wrote Equiano, ‘each different nation of Africa meet and dance after
the manner of their own country. They still retain most of their native
customs; they bury their dead, and put victuals, pipes, and tobacco, and
other things, in the grave with the corpse, in the same manner as in Africa.’^
IGBO EXILES IN THE DIASPORA 73

Although Equiano did not mention seeing other Igbo among ‘each
different nation of Africa’ in Jamaica, other near contemporaries did. In
1788 Captain Hugh Crow, who became a major trader at Bonny in
1791-1810, witnessed the public execution of an Igbo man in Jamaica. He
reported that the man’s ‘Eboe friends continued to cheer him with the hope
that he would return to his own country [after his death] until he was turned
off the scaffold.’'* Over twenty-five years later, Matthew Lewis, who kept a
journal of his time in western Jamaica in 1815-17, observed how his slaves
grouped themselves according to their ethnicity; he noted, for example, that
one day he ‘went down to the negro-houses to hear the whole body of Eboes
lodge a complaint against one of the book-keepers’.^ Even recent arrivals
such as Lewis quickly learned to identify individual slaves by their African
ethnicities. In one passage Lewis described a conflict between two of his
adult male slaves named Pickle and Edward, both of whom were said to be
‘Eboes’. Pickle, Lewis reported, had accused Edward of breaking into his
house and stealing some of his goods. Upon being asked by Pickle to use
‘obeah’ to find where the stolen goods were, Edward was said to have gone
out at midnight and to have collected the requisite herb and prepared it the
proper way, but then, according to Pickle, to have used obeah to poison him
rather than using it to find the stolen property as promised. As proof of
being ‘obeahed’ by his fellow Igbo slave. Pickle claimed that ‘he had a pain
in his side, and, therefore, Edward must have given it to him’.**

The Argument
The Igbo peoples were a distinct ethno-historical group who shared a
distinctive set of ancestral traditions and drew on the same or very similar
material, social and ideological resources in order to adapt to the situations
in which they commonly found themselves as slaves. They were a people
whom modern scholars can study as a separate ‘nation’ in the transatlantic
diaspora. The Igbo diaspora originated in the Nigerian hinterland of the
Calabar coast and was immense, amounting to perhaps some 1.4 million
people in the era of the Atlantic slave trade. About half, or some 750,000,
of those forcibly shipped to the Americas left in the period from 1750 to
1807. For the most part, Igbo were shipped by the British and were taken to
British America. The movement of Igbo had, therefore, important
consequences for the historical development of a number of British
American slave societies.
Although fiercely localistic in their home areas, Igbo-speaking peoples,

once thrown into the diaspora, embraced a collective identity derived from
being a member of ‘my own nation’. As the English explorer, W.B. Baikie,
put it in the 1850s, ‘In I’gbo each person hails ... from the particular district
74 ROUTES TO SLAVERY
where he was born, but when away from home all are I’gbos.’’ When S.W.
Koelle solicited information around 1850 from Igbo-speakers who had been
‘recaptured’ and sent by the British to Sierra Leone in the 1820s and 1830s,
they told him that ‘certain natives who have come from the Bight are called
‘Ibos’. Koelle went on to note, however, that, in ‘speaking to some of them
respecting this name, I learned that they never had heard it till they came to

Sierra Leone’. Before being thrown into the slave trade, he wrote, such
people knew ‘only the names of their respective districts or countries’.* Like
Olaudah Equiano, once taken from their homeland, they came to see each
other as fellow ‘natives of Eboe’ and members of ‘my own nation’.
These various Igbo-speaking peoples in the Calabar backcountry were
the main sources for the transatlantic slave trade from the Bight of Biafra.
It is likely, or at least possible, that they comprised 80 per cent or more of
the Africans loaded on European slave ships there. The structure of the
export trade in slaves from Calabar, in which Igboized coastal bigmen
brokered exchanges between newly arrived European captains and a series
of Igbo-speaking headmen in the interior villages, combined with certain
other social and political changes to make Igbo-speaking peoples the
region’s principal source population for slaves.
The rise of a new meritocratic elite in the densely populated interior
regions of Nri-Awka and Isuama, whose wealth was fuelled by the
importation of commodity-currencies like iron bars and copper rods brought
by new traders such as the Aro, encouraged the export of people as slaves.'”
The community-culture of Eboan society, with its mix of gerontocracy and
meritocracy, of fatalism and localism, of obia (or ‘doctoring’) and aja
(‘sacrifice’), lent ideological support to a system of exchange that
transferred large numbers of people as slaves to entrepots on the coast.
Until recently, anthropologists and historians tended to emphasize the
heterogeneity of social, cultural, and economic patterns of life in Igboland.
The basic ethnographic sources, produced as adjuncts to the administration
of twentieth-century British colonialism, stressed the multiplicity of sub-
ethnic groups among the Igbo-speaking peoples. The most comprehensive
ethnographer counted some 29 ‘sub-tribes’ and 59 ‘clans’; another
estimated that ‘there must be 2,000 [distinctive] Ibo clans’." The
at least
classic summary ethnological report of Forde and Jones documented five
basic Igbo groups, with a total of 13 separate sub-groups.'-
Earlier generations of historians tended to focus on the various
mercantile city-states or village-groups that ringed Igboland, from the Niger
mini-monarchies of Onitsha and Aboh, to the Delta polities (Brass-Nembe,
New Kalabar, Bonny, Okrika, and later Opobo), and the trading towns of the
Calabar (Cross) River, especially the group of settlements that made up Old
Calabar.'’ More recently, however, and perhaps in response to a revival of
 IGBO EXILES IN THE DIASPORA 75

‘traditional’ Igbo culture in the wake of the 1967-1970 Nigerian Civil War,
there has been a tendency to argue in favour of the homogeneous nature of
a loosely definable pan-Igbo culture, and of placing the evolution of
regional socio-cultural and political forms within the context of a broader
Igbo history.'^

The Numbers
In the whole era of the transatlantic slave trade, about one in seven Africans
transported to the New
World originated from the Bight of Biafra.'^ A
review of recent secondary literature and classic primary sources suggests
that, of the 11.6 million people estimated to have been shipped from Africa
between 1470 and 1860, some 1.7 million were transported from the Bight
of Biafra. Of these 1 .7 million, about 1 .2 million, or some 70 per cent, were
taken during the period from 1700 to 1809. Taking the rather longer period
from 1700 to 1840, it appears that 1.5 million people were loaded as slaves
onto European ships Calabar coast. Assuming that only 50 per cent of
at the

those originally enslaved were actually shipped to America, this suggests

that around 3 million people were enslaved and uprooted in the hinterland
of the Bight of Biafra in the century and a half after 1700.'* The contribution
of the Bight of Biafra to the transatlantic slave trade rose sharply in this
period, from approximately one-tenth of the traffic before 1750 to one-fifth
of it following quarter century and one-fourth in the years from 1780
in the

to 1809. The data upon which these calculations are made are presented in
Tables 1 and 2.

TABLE 1

ESTIMATES OF THE NUMBER OF SLAVES EXPORTED. 1470-1867'’

(FIGURES ROUNDED)

Years Bight 1
')f Biafra Africa

1470-1600 20 ,000 367,000

1601-1699 125 ,000 1,868,000
1700-1809 1,212 ,800 6,672,000
1811-1867 353 ,100 2,738,000

Total 1,710 ,900 11,645,000

Overall, estimate that about 80 per cent of the people shipped from the
I

Bight of Biafra as slaves were Igbo-speaking. Many of the others may well
have been Igboized through long-term contact. The first slaves shipped
from the region before 1650 were mostly Ijo and coastal non-Ijo such as
Delta Edo, Ogoni, Andony, and Ibibio.'" In the sixteenth and early
seventeenth centuries when the Biafran export slave trade was very limited
76 ROUTES TO SLAVERY

and largely based on the plundering of near delta villages by local warlords,
non-Igbo were captured and sold in relatively high proportions. Pereira in
1508 identified the peoples of the rivers along the coast as ‘Jos’ (Ijos)

without mentioning Igbo, and of the 17 communities of ‘caravalies

particulares’ that in the early seventeenth century Sandoval mentioned by
name, at least 1 1 can be identified as non-Igbo. ’’ During the major period of
slave exportation between the 1680s and 1830s, however, when the source
of captives shifted steadily inland, the proportion of Igbo rose dramatically,
perhaps reaching four out of five of the people shipped overseas as slaves.
This suggests, as Table 3 reveals, that in total up to 1.4 million Igbo were
put on European ships as slaves during the whole slave trade era.

TABLE 2
ESTIMATES OF SLAVES EXPORTED, BY PERIOD, 147CL1870
(ALL CARRIERS; FIGURES ROUNDED)

Years Bight of Biafra Average per Year Africa Percentage

from Biafra

1470s- 1600 20,000 200 367,000 5.5

1600- 1650s 25,000 500 631,000 4.0
1660s- 1699 100,000 2,500 1,237,000 8.0
1700-1749 273,600 5,500 2,430,000 11.3
1750-1779 409,100 13,600 2,083,000 19.6
1780-1809 530,100 17,700 2,159,000 24.6
1811-1840 333,700 11,100 1,962,000 17.0
1841-1870 19,400 970 776,000 2.5

Total 1,711,000 11,645,000 15.0

TABLE 3
IGBO EXPORTED IN CALABAR TRADE. I470-1867'''

Years % Number of Igbo

1470-1600 33 6,660
1600- 1650s 50 12,500
1660s- 1699 75 75,000
1700-1809 80 970,240
1811-1867 80 282,480

Total 1,346,880

As Table 4 shows, were loaded on British

the vast majority of these Igbo
ships and were thus bound for the British Americas. Of the estimated 1.2
million people exported from the Bight of Biafra in the period from 1700 to
1809, about 95 per cent were probably carried away on British ships. Put
 IGBO EXILES IN THE DIASPORA 77

another way, appears that during the eighteenth century the Bight of
it

Biafra supplied over one-third of the Africans carried in British ships and
that during the peak years of the Bight trade in 1750-1807, the region
supplied over 40 per cent of the slaves of British ships (see Table 5).
Assuming that Igbo constituted 80 per cent of those shipped, and that 20 per
cent died in the Atlantic crossing, then it would seem, on the basis of the
figures given in Table 4, that nearly 750,000 Igbo-speaking people reached
the Americas in British ships between 1700 and 1807. The vast majority of
these probably disembarked in the British Caribbean and North American
mainland, thus accounting for perhaps a third of all slave arrivals in these
colonies in this period. Such concentrations of Igbo arrivals had important
consequences for the development of slavery and society in the era of
revolution in British America.-'

TABLE 4
BRITISH SLAVE TRADE FROM CALABAR 1700-1807

Years Biafra Igbo

1700-9 19,350 15,480

1710-19 44,820 35,856
1720-29 59,990 47,992
1730-39 61,330 49,064
1740-49 75,880 60,704
1750-59 104,050 83,240
1760-69 134,980 107,984
1770-79 151,120 120,896
1780-89 211,000 168,800
1790-99 170,070 136,056
1800-07 123,000 98,400

Total 1,155,590 924,472

Source: Richardson, ‘Slave Exports’, p.l3.

TABLE 5
REGIONAL EXPORTS OF ENSLAVED ATLANTIC AFRICANS, 1700-1807

Sene- Sierra Gold Coast Bight of Bight of West

gambia Leone Benin Biafra Central

1700-1749 144,860 60,470 201,190 110,210 261,370 3 1

1 ,460
1750-1779 43,210 276,420 88,610 88,250 390,150 94,200
1780-1807 5,520 146,920 118,660 34,840 504,070 233,860

Total 193,590 483,810 408,460 233,300 1,155,590 639,520

1750-1807 48,730 423,340 207,270 123,090 894,220 328,060

Source: Richardson, ‘Slave Exports’.

78 ROUTES TO SLAVERY

Slaving in the Hinterland

The export slave trade of the Bight of Biafra reached further inland over the
course of two centuries. In the sixteenth century, individual Kalahari and
associated villages raided each other, with one village group, Agbaniye
Ejika (Bile), remembered as the most disruptive and warlike. According to
one historical anthropologist, ‘Bile got its slaves by sacking neighbouring
Delta villages, rather than by trade with the hinterland’." The oral traditions

mark the disruption of this early predatory period, and they contrast with the
later expansion of New Calabar under King Amakiri I. Whereas the early

Bile attacks had scattered villages, perhaps creating refugee settlements, in

the seventeenth century New Calabar colonized other village groups and
offered protection from the ravages of slave-raiding in return for

acknowledgement of Elem Kalahari’s suzerainty.-’ This small regional shift

in power is remembered in the following extract from a twentieth-century

popular song: ‘Agbaniye Ejika ama fama te /Amachree ama paka mam’
[‘Agbaniye Ejika destroyed towns/Amachree founded towns’].’’
The rise of Bonny in the late seventeenth and early eighteenth centuries
was based on creating ties with interior Igbo peoples and cornering the
market on trade with Europeans. Ibani principal men, who headed their own
‘canoe-houses’ under the paramountcy of one (or often two) local kings,
apparently built their trade on their connections with Ndoki and Ngwa Igbo
of eastern Isuama on the headwaters of the Imo River, initially for yams,
later for slaves, and finally for palm-oil. The early Bonny king-lists
emphasize that the immediate descendants of the culture-hero king,
Asimini, who gained office by persuading the Europeans to trade at the
town, were Ngwa Igbo. The oral histories relate that Asimini gave his
daughter in marriage to one ‘Opoli of Azuogu
Ndoki country’, and
in the

that she returned to rule Bonny as Queen Kambasa. Her son, Kumalu,
whose name later was one commonly given to children in Ngwa-Igboland,
succeeded her as king. The name ‘Kumalo’ suggests a ruthless, perhaps
‘evil’ ruler, and seems to evoke slave trading. All of this must have taken

place in the seventeenth century, as these changes predated the Pepple (or
Perekule) dynasties of the pre-colonial era. In the mid-nineteenth century,
the Ibani still Ndoki as
referred to the ‘brothers’.”
Before it could benefit from the ties with Ngwa-Igboland, however.
Bonny and its allies had to fight a series of wars to defeat the Kalahari,

Okrika, Ogoni, andAndony peoples who occupied land between the sea and
Isuama.” These conflicts are remembered in Bonny oral histories, which
explain the rise of the littoral village-group in mythic terms. In order to
‘widen the river’ at Bonny point, the culture-hero Asimini was made to

sacrifice a daughter, and thus started the rite of sacrificing a virgin woman
IGBO EXILES IN THE DIASPORA 79

every seven years order to propitiate the god of the sea wind. The new
in

cult of the war-god Ikuba, whose totem was the monitor lizard or iguana (as
well as the ‘house of skulls’) which so fascinated European visitors in the
eighteenth and nineteenth centuries, was also a function of Bonny’s new
reach into the interior.^’ Elem Kalahari traditions state that Bonny stole the
European trade from them, and describe the process in terms of a conflict in
spiritland between the tutelary gods, that is, Ikuba of Bonny and
Owoamekaso of New Calabar:

After Oruyingi (Mother of the gods) had given birth to all the gods of
all ibe [named peoples] in the delta, she asked them to make requests
for the benefit of their people. Owoamekaso asked for a book that
would attract European ships
Elem Kalahari. After they left the
to
presence of Oruyingi, Ikuba became jealous and tried to seize the book.
In the ensuing struggle, Ikuba was able to make off with the larger
fragment of the book, and so got the bigger ships calling at the port of
Bonny, only smaller ships being able to go upstream to Elem Kalahari.

As Map 1 illustrates, pre-colonial Igboland had three overlapping trade

systems, each of which served to link people in the densely populated
central plateaux with sources of outside goods that were controlled by
various other peoples on the peripheries of the region. In the Niger system
of cowrie-based canoe-trading, the mini-monarchies of Aboh, Asaba, and
Ossomari rose out of trading villages and incorporated the cultural
influences of Benin, Nri-Awka, and the non-Igbo Igalla, respectively.’^ In
central and eastern Igboland, African brokers demanded payment in a
distinctive assortment of trade-goods based on manillas (iron and/or copper
ingots), iron barsand guns.’” The early leading market-towns in central
Igboland included Azumini, Obegu, and Ohuhu, all on the middle or upper
reaches ol the Imo whence merchants brokered trade with Igboized
River,
rulers at Kalahari, Bonny, and, by the 1820s, Opobo. To the east, the fierce
itinerant Aro merchants, who were protected in their travels by the divine
sanction of Chukwu
(hence calling themselves umucluikwu or ‘god boys’),”
conducted slaves to the group of towns on the lower Cross River known
collectively as ‘Old Calabar’. Hailing from the oracle of Arochukwu, and
relying on the muscle of their armed porters (and their relations) as much as
on their reputation for retribution, the Aro ran major and minor fairs at
villages all across Igboland in the eighteenth and nineteenth centuries.
The largest fairs, such as that at Bende in the mid-nineteenth century,
lasted four days every twenty-fourth day (that is, one izu every three izu
Likwii),and were scheduled so that traders making rounds could attend
several of them in succession. Bende was said to be ‘two to four days’
journey north from Bonny’, and in the mid-nineteenth century was.
80 ROUTES TO SLAVERY
MAP 1

EXTENT OF IGBO CULTURE

IGBO EXILES IN THE DIASPORA 81

according to Baikie, grand depot tor slaves, as well as for palm-oil and
‘a

provisions, and supplies with the former New Kalabar, Bonny, and Andony,
as well as other neighbouring countries’. Baikie also noted that when the
‘foreign slave trade was being actively carried on, this town was in the
zenith of itswealth and importance, and even since has declined but little,
as it still remains the centre of the home slave mart for the coast, and the
south of rgbo’."*-
This Eboan world was one in which nothing happened by chance; leaving
anything to chance was, therefore, dangerous.” Sacrifice was the principal
way that people in historical Igboland attempted to create order out of the
events of their individual and collective lives. Sacrifice, the ritual killing of
consecrated victims, structured much of Eboan African ritual life. Igbo
people understood that they sacrificed living things because the deities
demanded it. The sacrifice was a prerequisite for gaining the benefits that
came with ‘feeding’ a particular spirit; or it could be the means of placating
a powerful spiritual being, a prerequisite for avoiding the wrath of capricious
deities who constantly demanded attention. The more powerful the deity, or
the more powerful the need for ‘asking’ (a-jiiju), the more valuable must be

the sacrificial victim. Some sacrifices, called ichu aja (or ‘joyless’ sacrifice),
were of worthless, disfigured or ugly things, such as rotten eggs, sick
chickens, aborted lambs, or lizards. Such sacrifices were made specifically to
distract otherwise malevolent spirits, ‘much as a dangerous dog is given a
bone to keep
busy’.” The most specialized sacrifices were to remove the
it

‘abomination’ of having violated a taboo {nso), usually thought of as a

transgression against A/?/ (the ground or Earth Deity), and therefore requiring
the services of an Nri-man, because the ndi Nri were the only people with the
power to remove such pollution. More generally, Igbo sacrificed regularly to
the ancestors (ndi ichie) and the spirits {ndi miio), offering not Just libations
at every meal but at periodic personal and collective rites, at times of
misfortune, or at funerals, festivals, harvests, and other calendrical events. At
Bonny in the 1 820s, for example, the people performed sacrifices ‘at different
periods, which are governed by the Moon’. At such events they would ‘offer
up Goats & Fowls as a sacrifice to their departed Progenitors, & the more
especially if they had performed any signal achievements’. At one large
festival at Bonny in 1826, when the king distributed rum, cloth, beads and
brass manillas to the general population, Fowls were ‘Goats, Dogs &
sacrificed in immense numbers - long poles on which were suspended dead
carcuses [sic] of the canine race met the eye at every turn.’”
In his narrative, Equiano recounted that sacrifices were made all the
time. He remembered that they marked special events, especially for
thanksgiving and were most often were done by the lineage heads {okpala)
and other leading men:
82 ROUTES TO SLAVERY

They have many moons; generally two at

offerings, particularly at full
harvest before the fruits are taken out of the ground: and when any

young animals are killed, sometimes they offer up part of them as a

sacrifice. These offerings, when made by one of the heads of a family,

serve for the whole. I remember we often had them at my father’s and
my uncle’s, and their families have been present. Some of our
offerings are eaten with bitter herbs. We had a saying among us to any
one of a cross temper, ‘That if they were to be eaten, they should be
eaten with bitter herbs’.

The dibias control over ‘medicine’ (ogwu) and his ability to communicate
with the spirits to find out what to ‘feed’ them may have made them
increasingly important in Igbo life as contact with Aro traders intensified in
the eighteenth and nineteenth centuries. In fact, dibia may have played a
major role in the increasing use of oracles and petitions to Chukwii and in
relations with the Arochukwu, and therefore in the displacement of the older
Nri hegemony with a new one based on petitioning (a-jiiju) and sacrifice
(aja). This is evident especially in the contrast between the pacifist role of
the Nri,one of whose main functions was to remove the abomination caused
by the shedding of human blood, and a secondary function of dibia which
was to provide the special ogwu that facilitated warfare.
The rise of dibia to great power throughout Igboland in the era of the
slave trade, however, may have come in part from arrogating the ability to
‘sacrifice’humans, and then having them sent away instead of having them
killed (and thus avoiding the required intervention by an Nri-man). Just as
a successful sacrifice to the spirits was signalled by the appearance of
vultures who carried the sacrifice to the muo world, so may have the
appearance of Aro and other interior traders demanding people in order to
please and placate them seem like the vultures sent by the spirits for a
sacrifice.^’

It should be noted, moreover, that dibia and juju were closely connected
with the secret-societies and masquerades, whose members generally were
the most powerful in the village and village-group. The only other figure
that would have had the kind of authority to send people away would have
been each compound’s headman or the okpara, or the ndi Nri, but it is likely
that the secret-societies and Mmuo (or masquerades) were instrumental in

such seizures. Not only could the powerful local men act with impunity and
anonymity in Mmuo, and with the sanction of the local dibia, but they were
precisely the ones who had the personal connections with outside long-
distance traders.
Walter Rodney may have been right, therefore, to argue that the export
of slaves tended to increase social exploitation within source areas, but, as
IGBO EXILES IN THE DIASPORA 83

far as historical Igboland isconcerned, he would be surprised to find that the

form such exploitation took was ‘traditional’/’* The question arises,
however, did the people who were seized or sent away see themselves as
outcasts or exiles? Did they turn their collective backs on things and ways
Eboan, or did they strive to re-create as much of their Eboan African world
as they could in the new worlds that they entered after the Atlantic crossing?
But first, where did these diasporic Igbo go?

Destinations

Up 20 per cent of the slaves embarked in the Bight of Biafra, as the paper
to
by Klein and Engerman in this volume shows, died before reaching
America. Moreover, in the nineteenth century, significant numbers of Igbo,
amounting 30,000 between 1821 and 1839 alone, were landed at
to at least
Sierra Leone. Retaken by anti-slave trade patrols, these ‘recaptured’ Igbo
have been little studied even though European explorers and Christian
missionaries of the 1 840s and
850s relied on Sierra Leonean Igbo as guides
1

and translators during their initial ventures up the Niger and Cross rivers.’^
As noted earlier, however, probably about 750,000 Igbo reached the
Americas between 1700 and 1809. As the British dominated slave
shipments from the Bight of Biafra in this period, it seems reasonable to
assume that most landed in the Anglophone Caribbean. Since Igbo were
widely regarded as unsatisfactory slaves by their European masters, it is
possible that disproportionate numbers of them were sold in the more
marginal regions of plantation agriculture in the New World. This seems to
have been the case in the French colonies, where, according to David
Geggus, ‘Igbo slaves, in large measure females, show up disproportionately
in Guadeloupe and the least-developed parts of Saint Domingue.’"^ By
contrast, slaves from the Slave Coast (Ewe, Eon and others), who were
usually regarded with greater favour by planters, tended to be shipped to the
prime areas of French Caribbean sugar-production. Whether this pattern
was replicated in the British colonies is uncertain, but, if it was, then one
might anticipate that, in the first hall of the eighteenth century, relatively
few Igbo went to Jamaica, Antigua, and St Kitts, where sugar production
was expanding vigorously, and relatively large numbers went to Barbados,
where sugar production was stagnating, the Bahamas, and the mainland
North America, especially the Chesapeake. After 1760, when the tide of
Igbo exiles was at its height, it is likely that Igbo were shipped in relatively
largenumbers to the islands of the Lesser Antilles and to areas around
Montego Bay and Savannah la Mar in western Jamaica.’”
While the precise pattern of Igbo arrivals in the Americas has yet to be
determined, scholars are beginning to identify those parts of Anglophone
84 ROUTES TO SLAVERY

and Francophone America where one might anticipate discovering a major

Igbo presence in the diaspora. It is believed that in the eighteenth century,
the Chesapeake, the Bahamas and Leeward Islands, and Jamaica were
probably the most common destinations of diasporic Igbo. Scholars have
noted the striking similarities in Atlantic creole languages, especially in the
Anglophone Americas, as well as the wide distribution of certain elements
of slave culture, including oheah, jonkomi (especially the so-called root
version with its associated cow-horn and ragman maskers, and ‘gambys’ or
box-drums), and the temi hiickra. Given that in 1750-1807 British planters
received probably twice as many Igbo as any other African ethno-historical
group (and perhaps four times the number of people from Gold Coast),"*- it

is important to investigate how the Igbo ethnicity of those arriving from the
Bight of Biafra contributed to the ‘bricolage’ and historical development of
the various Afro-Caribbean slave societies in the British Americas.

Igboization

In some colonies such Chesapeake piedmont in the second quarter of

as the
the eighteenth century Igbo forced migrants were Tirst-comers’. In such
cases, they set the basic patterns of material, social and ideological culture
of enslaved communities to which succeeding waves of saltwater (for
example. Western Bantu and Mande) and creole (or tidewater) slaves
acculturated. In other colonies such as post- 1750 Jamaica Igbo were
‘second-comers’ who ‘Igboized’ existing institutions and cultural patterns
as people drew on ancestral material, social and ideological resources in

order to adapt both to slavery and to the culture of slaves already there. In
both cases, Igbo made use of what they had at hand what they
to fashion

needed in order to sustain themselves, to forge connections amongst and

between each other, and to make sense of their new worlds. This African-
derived process of bricolage was one of mixing and matching, of adapting
and adopting a combination of new and old ways of doing and of being, and
often resulted in Igboesque regional ‘common traditions’.

Examples of Igboisms

We can see evidence of Igboization in the food ways, power ways, and
magic ways of slaves in Anglophone American slaveThese societies.^
various Igboesque artifacts suggest the range of material, social and
ideological resources on which enslaved Eboan Africans drew in order to
adapt to slavery.
Masters generally provided their slaves with only the barest essentials
needed to shelter, feed and clothe the people. Many slaves, therefore, had
 IGBO EXILES IN THE DIASPORA 85

their own
provisioning grounds, whether kitchen gardens, house-yards or
outlying plots, where they grew their own fruits and vegetables. One of the
most common slave-grown vegetables was Hibiscus esciilentiis, whose
podlike fruit was a staple of slaves’ ‘one-pot’ cooking practices. Slaves used
okra to thicken stews and soups, and the mucilaginous vegetable practically
represented slave (and, in the United States, southern) food ways.**^
was an Igbo word (okro).^^ People in other parts of western
‘Okra’
Africa knew and grew the vegetable, called ‘nkru-ma’ in Twi and variations
of ‘-ngombo’ in Western Bantu languages. But
was Igbo people who
it

brought the most common name of ‘okra’ into English. The word appeared
in English dictionaries in the eighteenth century, notably the Oxford English

Dictionary (1707, 1756 editions), although the plant remained largely

unknown to whites until the nineteenth century. Thomas Jefferson, for
example, did not have okra grown in his garden until 1809.'*’
The vegetal part of the basket of African-American ‘soul food’ indicates
a strong Igbo presence in Anglophone slave food ways. Most notable were
‘yams’ (both Dioscorea and Ipomoea batatas); black-eyed peas (or
cowpeas) Vigna unguicuilata); ‘greens’ {Brassica sp.); watermelon
(Citrullus vulgaris); and eggplant (Solanum melongena). The last was also
known as ‘Guinea squash’ in Virginia around 1800, but whites did not eat
it."** and early twentieth centuries in Igboland, the basic
In the nineteenth
kitchen crops were yams, black-eyed peas, greens, squashes, pumpkins, and
okra and watermelons, as well as gourds and groundnuts and Guinea
squash, and coco-yams, plantains and bananas (but not cassava).**^ In the
diaspora, Igbo continued to grow yams and maintained nearly all the
secondary subsistence crops of their ancestral village agriculture except for
coco-yam, plantains, bananas, and papaws (or papayas). The loss of
plantains and bananas seems to have been made up with maize and meal,
while butter and lard replaced palm-oil and cayenne replaced melegueta
pepper. Okra, associated with fertility as well as with proverbial knowledge
in some parts of pre-colonial Igboland, black-eyed peas (which in the south
of the United States are still reputed to bring good luck if eaten on New
Year’s Day), and squashes, watermelons, gourds, ‘greens’, and other forms
of material culture quickly reappeared and remained as staples of African-
American slave food ways.-'”

There is also clear evidence of Igboization in slave power ways. Fischer

has defined power ways as ‘attitudes toward authority and power’ and
‘patterns of political participation’." Slaves in Anglophone America drew
on similar social resources to order their individual and collective lives.
Although the influence of Akan on the Maroons of eighteenth-century
Jamaica has attracted much attention from historians, there is evidence of a
clear Igbo influence on those who stayed on the local plantations. The world
86 ROUTES TO SLAVERY

of the plantation was essentially divided into two largely separate spheres;
that of the whites and that of the slaves.^^ In the British Atlantic, whites
learned to call blacks ‘niggers’ and slaves learned to call their masters (and
other whites) ‘buckra’. The widespread use of the latter term rellected the
ubiquity of Igbo peoples in the diaspora or, for slaves of other ethnicity, the
utility of Igboesque ways of defining whites. Although derived from the
Ibibio mhakara {nih- plural; -karci, to encircle, rule, abuse, master,
understand), slaves everywhere in Anglophone America used the term to
denote ‘white folk’."^ It was Igbo people who brought the term into English
as buckra, and perhaps also supplied the social subtext of seeing the white
man as ‘he who surrounds or governs’ or as a ‘demon, powerful and
superior being’.'"' Or, as a slave saying from South Carolina in the 1770s put
it, ‘Da buccary no be good fatru’ (‘That white man is not good, to be
).-'•'
sure’
How did Igboized slaves perceive the buckral Monica Schuler has
provocatively suggested that, in Jamaica at least, many slaves ‘defined
slaveowners as sorcerers’ because of the slaves’ common conviction that
sorcery had played a prominent role in their original enslavement.^^ There is

good evidence to support such a supposition, at least in terms of the initial

interpretation that many West Africans had of Europeans as spirits and

jujus. For example John Jea, who had been born in 773 in Old Calabar and 1

then taken to New York as a slave remembered that he and his parents ‘were
often led away with were our gods; and at other
the idea that our masters
times we placed our ideas on the sun, moon, and stars, looking unto them,
as if they could save us’.^^ In general, however, one may suggest that many
of the slaves saw the buckra in Igboesque terms, that is, as eze (masters) or
as little kings. As a missionary resident in Onitsha in the 1850s explained,
'Eze literally means “Master”, and is applied to kings and to those who are
in an important office’."" Just as eze were spiritually and materially powerful
beings who could lord it over the people and yet were subject to the
omenani or customs of the place, so slaves saw their masters as cruel and
yet bound by customary law. Furthermore, slaves everywhere routinely
addressed not Just their owners but all buckra with the ritual salutation of
some variation of ‘master’, whether ‘massa’, ‘mossa’, ‘marster’, or ‘marse’.
Itmay well be that Igboized slaves appropriated familiar political terms to
make sense of their masters’ formal powers over them, even as they were
coerced into signifying their own subordination.
One may also see how an Igboesque political consciousness contributed
to the importance of resistance rather than rebellion in the slave worlds of
the British Americas. It is perhaps not without significance that successful
slave revolts - whether in seventeenth-century Palmares, among the
Maroons of eighteenth-century Jamaica, in St. Domingue in the 1790s, or in
IGBO EXILES IN THE DIASPORA 87

Bahia in the 1830s - tended to occur in areas with relatively few Igbo. By
contrast, in places such as the Chesapeake and western Jamaica, where there
was a strong Igbo presence, slaves tended to resort to resistance within
small-scale communities to force the buckra to abide by unwritten but well-
known plantation customs. Such customs (Igbo omenani) varied from place
to place but shared many core elements; these included two days off, abroad
spouses, ‘negro daytime’, basic rations, and individual compounds. The
growth of plantation customs also contributed to the shift from corporal
punishment to systems of discipline and incentives. In the more marginal
areas where planters routinely subdivided their slave holdings into
plantations, farms and quarters, Igbo concepts of localism, dual division and
segmentary social relations would have been adaptive for Igboesque
communities, with the incidental consequence of discouraging large-scale
revolutionary movements.
Igboized slaves drew on other Eboan African institutions to forge their
own power ways and regulate their own
The best known (and perhaps
lives.
least understood) was the common Anglophone slave masquerade,

Generally given a Gold Coast provenance, this slave social fact was more
likely an artifact of Igbo people in the diaspora.^ This is suggested by the
association of jonkomi (the ‘root’ version, that is) with ‘gambys’, or box-
drums, instruments that Matthew Lewis
1815-17 called ‘Eboe drums’.
in
Jonkomi is also associated with cow-horn and other animal masks as well as
distinctive peaked-hat masks, many of which are nearly identical to mid-
twentieth century Kalahari ones. Moreover, the development of jonkomi in
its Afro-Caribbean form seems to have occurred between 1750 and 1830

when arrivals of Igbos were at their greatest. Such evidence points,

therefore, towards Igboland as the primary source of jonkomi in
Anglophone America.^'
Two important institutions in historical Igboland were njokkii (or
ifejiokii), the ‘Yam spirit cult’, and the yam-associated okonko (men’s
clubs). The latter was the southern Igbo secret society equivalent of the Efik
ekpe society. Nineteenth-century sources attest to animal- and ragman-
maskers in Bonny and Elem Kalahari, and these appear very similar to ‘root
jonkomi' and not unlike the twentieth-century northern Igbo Oniabe and
Odo masquerades.^- It is likely that diasporic Igbo combined these
essentially shared traditions into a creole institution. Whites came to call the
Christmas-time masquerade ‘John Konnu’ because that is what the slaves
chanted or yelled during the visitation of njokkir, the rest of the year it may
have served as a kind of men’s club. So-called ‘root’ jonkomi was a male
preserve whose leader wielded a whip or stick in association with fierce
animal-masked others. From an Igboesque perspective, we can imagine that
what slaves got was an annual visitation by njokkii in mniiio (fearsome
88 ROUTES TO SLAVERY

masking); those who participated as maskers gained honour and prestige

within the slave community. To those enslaved, therefore, jonkonii may
have been an internal means of establishing status and hierarchy among the
slave men independently of any of their relations, occupational or
otherwise, with whites, or even with women. The slaves then masked its

importance by turning it into a Christmas-time buffoonery for the biickras

viewing. During the rest of the year, though, it may have had a much more
serious purpose.
Another important Afro-Caribbean slave artifact, the system of
‘doctoring’ called ‘obeah’, points finally towards an Igbo presence in the
ideological domain of slave community-culture. Igbo slaves drew on
ancestral ideological resources to make sense of their new world and in the

process ‘Igboized’ slave religious traditions throughout British America.

Because of its association with Jamaican slave communities, Afro-
Caribbean obeah, like jonkonii, has tended to be given an Akan
provenance.^’ This may, however, be misleading, for the functions of obeah
men and obeah women in the third quarter of the eighteenth century were
similar to those of the ndi obea (or dibia) of pre-colonial Igboland. In the
West Indies, as Edward Long explained in 1774, obeah men were ‘consulted
upon all occasions in order to revenge injuries and insults, discover and
punish thieves and adulterers; to predict the future, and for the conciliation
of favour’.^ As in Igboland, therefore, obeah men (and sometimes women)
in the Caribbean were diviners, doctors, and petitioners who specialized in
finding out why things happened in daily life and in determining what
needed to be done to placate the gods in given situations. In short, they were
JiiJii-mcn par excellence.
In historical Igboland, the dibia or obea was the person, usually a man,
who could communicate directly with the spirits. Known across Igboland
and the heavily Igboesque coastal settlements as powerful and dangerous,
and thus both feared and respected everywhere, such ‘doctors’ provide the
most common link between the visible and the invisible worlds. Privy to
secret information, purportedly including a separate ritual language, and
often idiosyncratic in their own lives (and thus thought to be gifted or
‘touched’), dibia combined knowledge of the spirit-world with
their sacred
a practical pharmacological knowledge. The latter presumably required
them to spend time in the forest collecting herbs, to divine what ailed
individuals, to determine the necessary remedies, and then to apply them.
Not only were ‘Oboe doctors, or Dibbeah’ able to ‘cure diseases by
charms’; they could ‘foretell things to come, and discover secrets’ as well.*’
The various magical and religious characteristics of dibia seem to have
changed very little in Igboland between the 1750s and the 1920s. In
southern Igboland, an ‘Ibo chief’ told P.A. Talbot in the 1920s that ‘With
IGBO EXILES IN THE DIASPORA 89

our people a native doctor is called Onye Dibia. All know witchcraft, but
some good and only make medicine to help men. Others can make both
are
bad and good medicine and yet others only busy themselves with bad
ones. ..Every Onye Dibia has great power, because everyone fears to offend
him on account of his medicines.’*^ Seventy years earlier, at Onitsha, it was
said that ‘the doctor, or priest, called Dibia, is another person of
consequence, and is very much feared by the people. He has a great sway
over the people, from his pretension to be able to foretell things to come,
and discover secrets’. At Bonny in the early ISOOs, dibia were said to
combine the sacred and the profane and through their powers held ‘the
populace most absolute awe and subjection’. The author of such
in the

remarks, Hugh Crow, also emphasized the curative medical abilities of

dibia, and wrote that although ‘they apply certain remedies, chiefly
decoctions of herbs and cupping, which they perform with a small calabash,
after having made incisions, they depend upon charms, in a great measure,
for relief’.^* In his vivid description of dibia medicine. Crow went on to
reveal the role of sacrifice in their curing rites. Thus he noted that, after
killing a male fowl by slitting its throat, the dibia ‘then threw himself into
many strange postures, and while muttering some incantations over the sick
men, he sprinkled the blood on Presumably the dibia also
their heads’.
applied some physical medicine, for Crow noted that in general dibia ‘make
much use of pod pepper, palm oil and various kinds of herbs for the cure of
diseases’.
Remembering his childhood in the 1750s near present-day Orlu in
Isuama and within the Nri/Awka Olaudah Equiano in 1789
cultural ambit,’"
provided a description of dibia similar to Crow’s, although Equiano seemed
to conflate the work of dibia with that of another major group of ritual
specialists, the atama or Nri-men. Equiano wrote:

Though we had no places of public worship, we had priests and

magicians, or wise men. do not remember whether they had different
I

offices, or whether they were united in the same persons, but they
were held in great reverence by the people. They calculated our time,
and foretold events ... These magicians were also our doctors or
physicians. They practised bleeding by cupping, and were very
successful wounds and expelling poisons. They had
in healing
likewise some extraordinary method of discovering jealousy, theft,
and poisoning ...[which] is still used by the negroes in the West
Indies.’”

obea in historical Igboland and the

In short, therefore, the dibia or ndi
‘obeahmen’ of pre-modern Alro-Caribbean societies were responsible for
ascertaining why things happened, remedying or inHuencing them, and
90 ROUTES TO SLAVERY

punishing transgressors. They also provided other things such as war-

medicine to protect those who would shed human blood.The world of the
invisibles was as real to diasporic Igbo as it had been in Eboan Africa,
because those hosts of invisibles had also survived the Atlantic crossing.
Eboan peoples drew on this ancestral tradition of the dibia to make sense of
their new world, calling it by the Igbo variant of obea, and in the process
adding a major Igboesque artifact to the religious bricolage of slaves in

British America.

Ethnogenesis

The Igboization of enslaved African-American communities in the era of

the transatlantic slave trade, and the essential reality of Igbo as a nation in
the diaspora, begs the question of ethnogenesis. In particular, to what extent
can we see the construction of ‘Igbo’ peoples as a function of slave trading?
After all, Olaudah Equiano, like hundreds of thousands of others from the
hinterland of the Bight of Biafra, first learned to call himself Igbo only after
he had been taken out of Eboan Africa. Equiano recognized his countrymen
in places as diverse as Barbados, Jamaica and London. In the nineteenth
century Koelle’s Sierra Leone informants were quite explicit in observing
that they never heard the term were sent or taken away. Scholars
until they

of African history have emphasized the colonial era and the role of
European administration in the construction of pan-ethnic or meta-ethnic
identities. If Equiano and Koelle’s informants are to be believed, however,
we perhaps need to look to an earlier era to see the historical creation of
‘nations’, Igbo included, as an integral part of the process of enslavement
and entry into the diaspora.

Conclusion

It is the contention of this paper that the ethnicity of Africans thrown by

force or by force of circumstance into the diaspora affected (and effected)
the historical development of regional slave cultures in the Americas. The
ancestral traditions they brought with them across the Atlantic continued to
inform their ‘secret lives’ in the early slave communities. Assuming that the
experience of the Atlantic crossing and adjustment to slavery, painful and
traumatic as it undoubtedly was, did not obliterate the identities, memory,
beliefs and customs of the Africans forced into the diaspora, it is important
to examine how those who survived the crossing drew on ethnic-African
material, social and ideological resources to adapt to slavery in their new
worlds.
As Monica Schuler has noted for the Caribbean, the natural response to
 IGBO EXILES IN THE DIASPORA 91

this uprooting of people was for them to associate with others of their own
specific ethnicity or nation, that is, for people to ‘re-group’ in
order to
confront the challenges of being slaves. It now appears that the transatlantic
slave tradewas not so random and randomizing as was once thought.^- As
John TTiornton has shown lor the seventeenth century and this author has
suggested elsewhere for later periods,'^ the organization of Atlantic slave
trading perhaps tended to concentrate rather than disperse African ethnic
groups. The
vast outpouring of literature in the last twenty years on the
numbers, origins and destinations of Africans shipped to the Americas
promises to provide powerful new tools to identify and describe the
historical significance of particular African ethnicities in the various
regional cultures of slaves in the Americas.’" Moreover, one could argue
that, in certain respects, these African ethnicities or ‘nations’ were
a product
of the trade in slaves.

When Olaudah Equiano published his memoirs in 1789, he included his

Igbo name in the manuscript’s title and referred to himself as ‘the African’.
And when he summarized the first chapter of his life’s story, as ‘some
account of the manners and customs of my country’, he specifically wanted
his audience to know that.

They had been implanted in me with great care, and made an

impression on my mind, which time could not erase, and which all the
adversity of fortune I have since experienced served to rivet and
record...! still look back with pleasure on the first scenes of my life,
though that pleasure has been for the most part mingled with sorrow.’-'

Equiano’s own personal diaspora transformed his identity from that of a

member of a particular kindred in a local village-group in ‘Eboan Africa’ to
being one of many ‘natives of Eboe’. His forced migration did not make
him an outcast, but an exile from his ‘own nation’.

ACKNOWLEDGEMENTS
The author wishes to thank James Walvin and Maureen Wamer-Lewis for their encouragement
and Joseph C. Miller, Paul Lovejoy, and David Richardson for their suggestions and comments
on earlier drafts of this article.

NOTES
1. Olaudah Equiano, The Interesting Narrative of the Life of Olaudah Equiano, Written by
Himself ed. Robert J. Allison (original edition, 1789; Boston, 1995 edition), pp.54-5. For
other examples of Equiano’s use of ‘my own nation’ or similar phrases, see ibid
pp 39 45
50.
2. Ibid., p.57.

3. Ibid., p. 145.

4. Captain Hugh Crow, Memoirs of the late Captain Hugh Crow, of Liverpool (1830; London,
1970 edition), p.26.
 .

92 ROUTES TO SLAVERY
5. Matthew G. Lewis, Journal of a West India Proprietor (London, 1834), p. 129.
6. Ibid., pp. 133-4 (emphasis in original).

7. William Balfour Baikie, Narrative of an Exploring Voyage up the Rivers Kwora and Binue
(1856; London, 1966 edition), p.307.
8. Reverend Sigismund W. Koelle, Polyglotta Africana, ed. RE.H. Hair (1854; Graz, Austria,
1963 edition), pp.7-8. See also Margaret M. Green, Tgbo Dialects in the Polyglotta
Africana’, African Language Review, 6 (1967), pp.l 1-19. 1

9. Cf. Northrup, who suggests that 60 per cent were Igbo and Inikori who suggests that only
one-third were Igbo, David Northrup, Trade without Rulers: Pre-colonial Economic
Development in South-eastern Nigeria (Oxford, 1978), pp.60-2; J.E. Inikori, ‘The Sources
of Supply for the Atlantic Slave Exports from the Bight of Benin and the Bight of Bonny
(Biafra)’, in Serge Daget (ed.), De la Trade a TEsclavage: Acts du Colloque International
sur la Traite des Noirs, 3 vols. (Nantes, 1988), II, Obichere has suggested that
p.35. Recently
Igbo were not the actual source population, and argues that Idoma and Tiv were, but he offers
no real evidence to support this; Boniface Obichere, ‘Slavery and the Slave Trade in the
Niger Delta Cross River Basin’, in Daget (ed ), De la Traite, II, p.50. Obichere’s argument
is also supported by A.E. Afigbo; see Inikori, ‘Sources of Supply’, p.35n.
10. For discussion of the role of commodity-currencies in the transatlantic slave trade from West-
Central Africa, see J.C. Miller, Way of Death: Merchant Capitalism and the Angolan Slave
Trade, /7iO-/^iO (Madison, Wisconsin, 1988), pp. 71-104.
11. C.K. Meek, Law and Authority in a Nigerian Tribe: a Study in Indirect Rule (1937; New
York, 1970 edition), pp.3-4.
12. Daryll Forde and G.l. Jones, The Ibo and Ibibio-speaking Peoples of South-eastern Nigeria
(London, 1950), passim; cf. Baikie, who in the 1850s recognized four major groups and six
or seven other major sub-groups, with a number of lesser districts or places; Baikie,
Narrative, pp. 308-1 1

13. Ebiegberi Alagoa and Adadonye Fombo, A Chronicle of Grand Bonny (Ibadan, Nigeria,
J.

1972), pp.3-16; Ebiegberi J. Alagoa, A History of the Niger Delta: an Historical

Interpretation of Ijo Oral Tradition (Ibadan, Nigeria, 1972), pp. 123-71; G.l. Jones, The
Trading States of the Oil Rivers: a Study of Political Development in Eastern Nigeria
(London, 1963), pp.9-48.
14. Elizabeth Isichei, A History of the Igbo Peoples (London, 1976); idem, A History of Nigeria
(London, 1983); A.E. Afigbo, Ropes of Sand: Studies in Igbo History and Culture (Ibadan,
Nigeria, 1981); M. Angulu Onwuejeogwu, An Igbo Civilization: Nri Kingdom and
Hegemony (London, 1981); John N. Oriji, Traditions of Igbo Origin: a Study of Pre-colonial
Population Movement in Africa (New York, 1990); Don C. Ohadike, Anioma: A Social
History of the Western Igbo People (Athens, Ohio, 1994).
15. See David Eltis and David Richardson, ‘West Africa and the Transatlantic Slave Trade: New
Evidence on Long Run Trends’, in this volume.
16. On losses before shipment from the coast, see Miller, Way of Death, pp. 379-442; Inikori,
‘Sources of Supply’, pp.31-2. Northrup suggested that about million slaves were shipped
1

from the Bight; Northrup, Trade without Rulers, pp.54-7.

17. The total African export figures for 1470s- 1699 are based on Paul E. Lovejoy, ‘The Volume
of the Atlantic Slave Trade: A Synthesis’, Journal of African History, 23 (1982), pp.478,
480-1; for 1700-1809 based on Richardson, ‘Slave Exports’, p.l7; and for 1811-1870, on
David Eltis, Economic Growth and the Ending of the Transatlantic Slave Trade (New York,
1987), pp.249, 250-2. The Bight of Biafra estimates for 1470s-1600 are arbitrary (simply an
average of 200 per year for 100 years), but reflect Pereira (c.l508); Alonso Sandoval, De
Instauranda Aethiopum Salute (1627; Begota, 1956 ed.); Northrup, Trade without Rulers,
pp.50-1. For the 1600-1659, it again is arbitrary, simply an average of 500 per year for 60
years, but reflects Johannes M. Postma, The Dutch in the Atlantic Slave Trade 1600-1815
(Cambridge, 1990), pp.1-25, 56-83, 106-25. For 1660-1699 the number assumes increases
from ,000 to 4,000 a year over four decades and is based on Northrup, Trade without Rulers,
1

pp.52-4; Postma, Dutch in the Atlantic Slave Trade, pp. 106-25. The figures for 1700-1809
are from Richardson, ‘Slave Exports’; for 181 1-1870 from Eltis, Economic Growth.
18. Alagoa, Niger Delta, pp. 10-16; Jones, Oil Rivers, pp. 12-13. The fact that the twentieth-
 .

IGBO EXILES IN THE DIASPORA 93

century ‘complexity of the language map of the Niger Delta’ (Alagoa, Niger Delta, p.I6)
largely mirrors a similar complexity in the seventeenth century powerfully supports
Hair’s
thesis of the ‘ethnolinguistic continuity’ of the Guinea Coast; see P.E.H. Hair,
‘An
Ethnolinguistic Inventory of theLower Guinea Coast before 1700: part I’, African Language
Review, 1 (1968), pp .47-73 and part II, African Language Review, 8 (1969), pp.225-56.
19. Pereira wrote: ‘ha outros negros que ham nome Jos e possuem grande terra e sam jente
belicosa e comem os homees...Rio dos Ramos.. .A jente desta terra sam chamados Jos. ..Rio
de Sam Bento. ..Rio de Sant Ilefonso...Rio de Santa Barbara. ..Rio Pequeno...estes quatros
Rios. ..sam abitados d’aquelles povoos a que chaman sam
Jos. ..Rio Real... A jente d’este Rio
chamados Jos; estes e os de que atras falamos, todos
e sam huus todos comem came
humana...’, cited in Hair, ‘Ethnolinguistic Inventory, part IT, p.250n. See also Alonso
Sandoval, De Instauranda, p.94; and also the listing of Ijo and non-Ijo groups in Jones, Oil
Rivers, pp. 12-13.
20. Numbers from Richardson, ‘Slave Exports’, p.l7. The Bight of Biafra numbers that he gives,
however, are only the Bntish and French trades. I would guess that the smaller European

national slave trades, including the Brandenburgers, Danes, Swedes and, after 1750, the
Dutch, as well as the ‘Portuguese’ (Brazilian) and Spanish (Cuban, Puerto Rican, Louisiana)
trades all would have drawn slaves from Calabar. The point
is that the numbers should be
considered a conservative estimate, and perhaps 10 to 25 per cent too low.
21. The era was best defined by David Brion Davis, The Problem
of Slavery in the Age of
Revolution 1770-1823 (Ithaca, 1975). The general Igbo reputation for being ‘bad’ slaves
who
tended to run away or shirk work must have played a role in the social relations of
when so many Igbo were flowing into British colonies.
slavery in that era,
22. This was based on oral traditions and corroborative material evidence (captured sacred
tusks)’ collected by Robin Horton, presumably in the 1950s; cited in Hair, ‘Ethnolinguistic
Inventory, part IT, p.252n.
23. Perhaps modem distinction between the ‘tme’ and ‘affiliated’ Kalahari. The
this led to the
acknowledge that Elem Kalahari also subjugated peoples like the Engenni by force
traditions
of arms; see Alagoa, Niger Delta, pp. 139^0.
24. Alagoa, Niger Delta, pp. 38-9. In fact, as the slave trade expanded into the interior, the slave
1

trade places shifted coastwise, from Bile and New Kalabar eastward to Bonny, and later,
Opobo.
25. In the twentieth century,Ndoki and Ibani used the phrase ‘Amina mina’ (‘brothers’) to refer
to each other; Alagoa and Fombo, Grand Bonny, pp.7— 16, 74; Baikie, Narrative, p.438;
Onji, Traditions of Igbo Origins, pp.157-9. Baikie {Narrative, pp.314-15) related that in
Igboland ‘the greatest or worst of evil spirits, is named Kamallo, possibly equivalent with
Satan. His name is frequently bestowed on children, and in some parts of I’gbo, especially
in Isuama, Kamallo is worshipped. No images are made, but a hut
which are
is set apart, in
kept bones, pieces of iron, &c., as sacred. Persons make inquiries of this spint, if they wish
to commit any wicked action, such as murder, when they bring presents of cowries and cloth
to propitiate this evil being and render him favourable to their designs’. He also wrote that
^Kamallo means “one going about evei^where and in all directions’”.
26. Alagoa, Mger De//a, pp. 1 53-7.
27. Ibid.,pp.154, 156-7; see also Baikie, Narrative, p.337; Crow, Memoirs, p.212; Richard M.
Jackson, Journal of a Voyage to Bonny River on the West Coast of Africa, ed. Roland Jackson
(Letchworth, 1934), p.68; Alexander Falconbridge, An Account of the Slave Trade on the
Coast of Africa (1788; New York, 1973 edition), p.51. Drawing on classical Greek
mythology to interpret Bonny practices. Crow described the seven-year sacrifice as ‘a
propitiatory offering to Boreas, the god of the north wind’ {Memoirs, p.83).
28. Alagoa, Mger De/ra, p. 141
29. Richard N. Henderson, The King in Every Man: Evolutionary' Trends in Onitsha Ibo Society
and Culture (New Haven, 1972).
30. For an explanation of ‘iron bars’ as a money of account, see Elizabeth Donnan (ed ).
Documents of the History of the Slave Trade to America, 4 vols. (Washington,
Illustrative
D C., 1935; reprinted New York, 1965), IV, pp. 69-84.
31.. See Major Arthur G. Leonard, ‘Notes of a Journey to Bende’, Journal of the Manchester
 .

94 ROUTES TO SLAVERY
Geographical Society', 14 (1 898), p. 19 1

32. Baikie, Narrative, pp.309-10.

33. Kenneth O. Dike and Felicia Ekejiuba, The Aro of South-eastern Nigeria, 1650-1980
(Ibadan, Nigena, 1990), p.l32; Francis Arinze, Sacrifice in Ibo Religion (Ibadan, Nigeria,
1970), p.37. The closest thing to ‘chance’ in ‘traditional’ Igbo ontology (at least in the mid-
twentieth century) was accidentally bumping into particular types of spirit, ado, which are
said to be ‘wicked’ and ‘like electric wire’ and unable to ‘consider bona fide transgression
[of any kind]’. Arinze, Sacrifice, pp.13-14.
34. For ichu aja, see Annze, Sacrifice, pp.37-8, 57, 83.
35. Jackson, Voyage to Bonny River, pp.83, 89.
36. Equiano, Narrative, p.l 1 ;
see also Arinze, Sacrifice, pp.42^.
37. For the importance of vultures, see the proverbs Northcote W. Thomas, Anthropological
in

Report on the Ibo-speaking Peoples of Nigeria, Vol.ll, English-Ibo and Ibo- English
Dictionary (1913; New York, 1969 edition), pp.l57, 382; Arinze, Sacrifice, p.44; John E.E.
Njoku, A Dictionary of Igbo Names, Culture and Proverbs (Washington, D.C., 1978), p.78.
For the historiography of the slave acquisition mechanism(s) in Igboland, see Jones, Oil
Rivers', Northrup, Trade without Rulers', John N. Oriji, ‘The Slave Trade, Warfare and Aro
Expansion in the Igbo Hinterland’, Transafrican Journal of History, 16 (1987), pp. 15 1-66;
Obichere, ‘Slavery and the Slave Trade’.
38. Walter Rodney, History of the Upper Guinea Coast 1545-1800 (Oxford, 1970).
A
39. See, for example, Simon Jonas in the various accounts connected with Baikie’s expedition in
the 1850s. For the presence of Igbo in early nineteenth-century Sierra Leone, see Northrup,
Trade without Rulers, pp.231, 235-9.
40. David Geggus, ‘Sex Ratio, Age and Ethnicity in the Atlantic Slave Trade: Data from French
Shipping and Plantation Records’, Journal of African History, 30 (1989), p.42.
41. For general patterns in sugar production, see David Watts, The West Indies: Patterns oj
Development, Culture, and Environmental Change since 1492 (Cambridge, 1987), pp. 232-3.
My hypothesis has to be evaluated ultimately by reference to distributions of shipments of
slaves from Biafran ports to specific destinations in the Americas. The Du Bois data set will
provide information on this in due course, but, unfortunately, it was not available at the time
of writing this paper.
42. See Richardson, ‘Slave Exports’, p.l 3, for coastal variations in slave shipments.
Richardson’s numbers show that from 1750 to 1807 the Bight of Biafra supplied some
894,000 slaves, or 44 per cent of the 2 million slaves shipped from Africa in British vessels
in this period. In the 1780s, the region accounted for over 60 per cent of the British trade.
43. For a theoretical discussion of ancestral and common traditions in the context of small-scale
societies in equatorial Africa, see Jan Vansina, Paths in the Rainforests: Toward a History of
Political Tradition in Equatorial Africa (Madison, Wisconsin, 1990), pp. 7 1-100, 249-63.
For bricolage, see Claude Levi-Strauss, The Savage Mind (1962; Chicago, 1966 edition),
pp. 16-22.
44. David Hackett Fischer, Albion’s Seed: Four British Folkways in America (New York,
Cf.
1989), pp.7-1 1, and his discussion of the domains of vernacular culture and his typology of
folk ways.
45. See Stacy G. Moore, “‘Established and Well Cultivated’’; Afro-American Foodways in Early
Virginia’, Virginia Cavalcade, 39 (1989), pp. 70-83; Robert L. Hall, ‘Savoring Africa in the
New World’, in Herman J. Viola and Carolyn Margolis (eds.). Seeds of Change: a
Quincentennial Commemoration (Washington, D C., 1991), pp. 161-9.
46. Isichei, History' of Nigeria, p.28. Cf. Western Bantu ‘gumbo’ for H. esculentus; Lorenzo Dow
Turner, Africanisms in the Gullah Dialect (Chicago, 1949), p.l94; Joseph E. Holloway and
Winifred K. Vass, The African Heritage of American English (Bloomington, Indiana, 1990),
p.29.
47. Moore, ‘Afro-American Foodways’, pp. 79-80.
48. On the last point see Mark Wagner, ‘The Introduction and Early Use of African Plants in the

New World’, Tennessee Anthropologist, 6 (1981), pp.l 12-23.

49. See Equiano, Narrative, pp.36-7; V.A. Oyenuga, Agriculture in Nigeria: an Introduction
(Rome, 1967), pp 134-9; Barry Floyd, Eastern Nigeria: a Geographical Review (New York,
IGBO EXILES IN THE DIASPORA 95

1969), pp. 174-7; L.C. Uzozie, ‘Agricultural Land Use

in the Nsukka Area’, in G.E.K.
Ofomata The Nsukka Environment (Enugu, Nigeria, 1978), pp.l55, 170-1. For
(ed.).
historical sources, see Jean Barbot, A Description of the Coasts of North and South-Guinea
[1678-1682], in Awnshawn Churchill and John Churchill (eds.), A Collection of Voyages and
Travels (London, 1732), p.379; Captain John Adams, Sketches Taken during Ten Voyages to
Africa, Between the Years 1786 and 1800 (1823; New York, 1970 edition), p.53; Crow,
Memoirs, pp.l46, 252-3, 258; Richard Lander and John Lander, Journal of an Expedition to
Explore the Course and Termination of the Niger, 2 vols. (New York, 1832), II, pp.l74, 201,
246; R.A.K. Oldfield, ‘Mr. Oldfield’s Journal’, in Macgregor Laird and R.A.K. Oldfield,
Narrative of an Expedition into the Interior of Africa, by the River Niger, 2 vols. (London,
1837), I, pp.374, 386-7; Captain William Allen, A Narrative of the Expedition Sent by Her
Majesty's Government to the River Niger in 1841, 2 vols. (London, 848), 1, p.25 Reverend1
1 ;

John C. Taylor, ‘Journal of the Rev. J.C. Taylor at Onitsha’, in Reverend Samuel Crowther
and Reverend John C. Taylor, The Gospel on the Banks of the (1859; London, 1968
edition), pp. 367-8.
50. It is unlikely that Thomas Jefferson, who startedgrowing okra in his garden in 1809, was, as
Moore suggests, ‘among the first in Virginia to grow this foreign plant’; Moore, ‘Afro-
American Foodways’, pp. 79-80. In fact, in the 1790s and 1800s Jefferson suddenly became
interested in the African plants that his slaves knew about; see Edwin M. Betts (ed.), Thomas
Jefferson’s Garden Book. 1766-1824 (Philadelphia, 1944), passim.
5 1 . Fischer, Albion 's Seed, p.9.
52. For an analysis of the implications of the separate realities for the continuing relevance of
things African, that is, for the importance of the ‘substrate’ influence in Historical
Creolization in the Eighteenth-century Chesapeake, see my ‘
He Is an African But Speaks
Plain”: historical creolization in eighteenth-century Virginia’, in Joseph E. Harris, Alusine
Jalloh, Joseph E. Inikori, Colina A. Palmer, Douglas B. Chambers and Dale T. Graden, The
African Diaspora, Walter Prescott Webb Memorial
Lectures, No. 30, (eds.) Alusine Jalloh
and Stephen E. Maizlish (College Station, Texas, 1996), pp. 100-33.
53. Elaine M. Kaufman, Ibibio Dictionary' (California, 1972), p.223; Turner, Gullah Dialect,
p.l91. This was true for places as diverse as eighteenth-century Philadelphia, the
Chesapeake, the Carolina lowcountry, Jamaica, Barbados, and the Leeward Islands.
54. Turner, Gullah Dialect, p. 191; Oxford English Dictionary (compact edition), p.288; cf.
Captain Becroft and J.B. King, ‘Details of explorations of the Old Calabar River, in 1841 and
1842’, Journal of the Royal Geographical Society, 14 (1844), p.261, on the Cross River
above Old Calabar. Interior Igbo did not have a common term for ‘white folks’ until after the
1850s (or even later); see Taylor, ‘Journal at Onitsha’, pp.251, 261, 281; Thomas,
Anthropological Report, II, pp.l04, 299; Njoku, Dictionary of Igbo Names, pp.3-4. The
various terms included oibo (stranger), onye oicha (white-coloured person), and beke (light-
skinned) or nwambeke (children of Baikie, the 1850s Niger explorer).
55. James Barclay, The Voyages and Travels of James Barclay, Containing Many Surprising
Adventures and Interesting Narratives (n.p., 1777), p.26. It is curious that Mande toubab,
Akan abroni, or Western Bantu mundele did not seem to resonate with slaves in the
Americas. This is especially so of mundele, since Kongo and Angolan peoples were widely
distributed. By comparison, Igboesque buckra was fairly commonly adopted, even where, as
in South Carolina, there were few Igbo.
56. Monica Schuler, ‘Afro-American Slave Culture’, Historical Reflections, 6 (1979), pp.l32,
124, 131.
57. Monica Schuler, ‘The Life, History, and Unparalleled Sufferings of John Jea, the African
Preacher [1815]’, in Graham R. Hodges (ed.). Black Itinerants of the Gospel (Madison,
Wisconsin, 1993), p.90.
58. Taylor, ‘Journal at Onitsha’, p.264. The locals routinely called Taylor eze or master.
59. Cf. Genovese’s thesis of ‘paternalism’; Eugene D.Genovese, Roll Jordan Roll: the World the
Slaves Made (New York, 1974), passim.
60. Edward Long originally attributed the custom to a Fante caboceer named John Conny in the
1720s, but the widespread distribution of Jonkonu, especially after 1750, suggests that such
an explanation was, at best, unlikely and, at worst, fanciful; see Judith Bettelheim, ‘The
96 ROUTES TO SLAVERY
Afro-Jamaica Jonkonnu Festival: Playing the Forces and Operating the Cloth’ (unpublished
PhD. thesis, Yale University, 1979), pp.7-20; Orlando Patterson, The Sociology of Slavery:
an Analysis of the Origins, Development and Structure of Negro Slave Society in Jamaica
(1967; Rutherford, New Jersey, 1969, edition), pp. 238-9, 243-4; Mullin, Africa in America,
pp.70, 326n; Michael Craton, ‘Decoding Pitchy-patchy: the Roots, Branches and Essence of
Junkanoo’, Slavery' and Abolition, 16 (1995), pp. 15-3 Sterling Stuckey suggests a Yoruban
1 .

or Fon influence, from the Egun ma.squerade, overlooking the fact that neither ethnic group
was numerically important to the British trade; see Sterling Stuckey, Slave Culture:
Nationalist Theory and the Foundations of Black America (Oxford, 1987), pp. 68-73. Others,
such as Frederick Cassidy, argue for a possible Bambara or Ewe derivation; see Craton,
‘Decoding Pitchy-patchy’, pp.38-9. Patterson {Sociology of Slavery, p.245) recognized that
Igbo Mmo, Yoruba Egugun, and Ga homowo were most like Jamaican jonkonu, but again,
the proportion of Yoruba and eighteenth-century Ga (who if enslaved would have been sent
from the Bight of Benin) in the British trade was small.
6 1 . Bettelheim, Afro-Jamaican Jonkonnu’, pp. 1 2- 3, 20. Bettelheim goes on to write (pp.45-6)
‘ 1

that ‘This Kalahari peaked hat is so close in style to those worn

Jamaica that one is in

tempted to make certain assumptions. Yet, the region of Kalahari demonstrates as varied a
culture history as does Jamaica and any firm conclusions are impossible’. For various
historical descriptions, see Roger D. Abrahams and John F. Szwed (eds.). After Africa:
Extracts from British Travel Accounts and Journals of the Seventeenth, Eighteenth, and
Nineteenth Centuries Concerning the Slaves, their Manners, and Customs in the British West
Indies (New Haven, 1983), passim. See also Judith Bettelheim, ‘Jamaican Jonkonnu and
Related Caribbean Festivals’, in Margaret E. Crahan and Franklin W. Knight (eds.), Africa
and the Caribbean: the Legacies of a Link (Baltimore, 1979), pp.8()-l()0.
62. See Philip A. Oguagha, ‘Historical and traditional evidence’, in Philip A. Oguagha and Alex
1. Okpoko, History and Ethnoarchaeology in Eastern Nigeria: a Study of Igbo-Igala
Relations with Special Reference to the Anambra Valley (Cambridge, 1984), pp. 268-9; Uche
Okeke, ‘The Art Culture of the Nsukka Igbo’ and Chike Aniakor, ‘The Omabe Cult and
Masking Tradition’, both in Ofomata (ed ), Nsukka environment, pp.271-2, 286-306. See
also Augustine Onyeneke, The Dead among the Living: Masquerades in Igbo Society' (Nimo,
Nigeria, 1987).
63. Patterson (Sociology of Slavery, pp. 1 85-6) attributes it to Twi and Ga obeye. See also Mullin,
Africa in America, pp. 175-86.
64. Long’s account was from his experiences in the West Indies in the 1750s and 1760s, cited in
Mullin, Africa in America, p.l75. See also Albert J. Raboteau, Slave Religion: the “nvisible
Institution” in the Antebellum South (Oxford, 1978), p.34.
65. The from Crow, Memoirs, p.226; Reverend Samuel Crowther, ‘Appendix 11: a
citations are
Few Notices of Onitsha, Idda, and Gbegbe, and of the Overland Route to Abbeokuta’, in
Crowther and Taylor, Gospel on the Banks of the Niger, p.435.
66. P. Amaury Talbot, Tribes of the Niger Delta: their Religions and Customs (London, 1932),
p.l31. For how dibia were consulted for any misfortune, theft, illness, as well as for every
birth and death in central Igboland in the 1930s, see Margaret M. Green, Ibo Village Affairs
(1947; New York, 1964 edition), p.54. Just as importantly, dibia also made ‘war-medicine’.
67. Crowther, ‘Appendix 11’, p.435.
68. Crow, Memoirs, pp.21 1, 225.
69. Ibid., p.227.

70. For recent and convincing evidence on Equiano’s origins, and the location of his natal village
Acholonu, ‘The Home of Olaudah Equiano - A Linguistic
east of the Niger, see Catherine O.
and Anthropological Search’, Journal of Commonwealth Literature, 22 (1987), pp.5-16;
idem. The Igbo Roots of Olaudah Equiano - an Anthropological Research (Owerri, Nigeria,
1989), especially her map p.9. Cf. the early guess, based on less convincing evidence, that
Equiano’s home was west of the Niger, among the ‘Ika’ Igbo (now generally referred to as
‘Anioma’). G.I. Jones, ‘Olaudah Equiano of the Niger Igbo’, in Philip D. Curtin (ed.), Africa
Remembered: Narratives of West Africans from
the Era of the Slave Trade (Madison,
Wisconsin, 1967), pp.60-9.
71. Equiano, Narrative, p.l2. The ellipsis omits a couple of sentences about how they called
IGBO EXILES IN THE DIASPORA 97
these specialists Ah-affoe-way-cah, which signifies calculators or yearly men, our year
being called Ah-affoe’. This is problematical, but may be related to the annual New Year
festival {ife-njokku) as the root-word afo is the Igbo word for ‘year’. The term dibia
or a close
variant of obia apparently was universal in and around Igboland; see Koelle, Polyglotta
Africana, p.28; Thomas, Anthropological Report, II, pp.27, 303.
72. Cf Sidney Mintz and Richard Price, The Birth of African-American Culture: an
Anthropological Perspective (Boston, 1992), originally published as An Anthropological
Approach to the Afro-American Past (Philadelphia, 1976) and the essay by Morgan in this
volume.
73. John Thornton, Africa and Africans in the Making of the Atlantic World. 1400-1680
(Cambridge, 1992); Douglas B. Chambers, “‘He Gwine Sing He Country”: Africans, Afro-
Virginians and the Development of Slave Culture in Virginia, 1690 to 1810’ (unpublished
Ph D. of Virginia, 1996).
dissertation. University
74. Schuler, ‘Slave Culture’, pp. 123-4, 126-8; Thornton, A/r/ca, pp.98-125, 183-205.
75. Equiano, Narrative, p.l5.
 ‘Of a nation which the others do not
Understand’: Bambara Slaves and African
Ethnicity in Colonial Louisiana, 1718-60

PETER CARON'

The French North American colony of Louisiana, parts of which were first
settled in 1698, is an interesting and unique example of early eighteenth-
century life in a non-English North American colony, especially as it was
lived by the several thousand African slaves imported in the 1720s.
Comprising fewer than a dozen isolated settlements along the Gulf Coast
and on the banks of the Mississippi River, Louisiana experienced a sharp
increase in immigration of West Africans which only lasted thirteen years
from 1718 yet had a profound affect on the social and economic
development of the colony.* By the end of the African slave trade to the
colony in the early 1730s, Africans made up over half of the total population
of the colony’s largest settlement and capital. New Orleans.^ Between 1719
and 1731, the Louisiana colony was the final destination for just over 5,000
Africans. Captives arrived from various points along the West Africa coast
including Cabinda, Whydah, Cape Lahou, Cape Apollonia, Bissau,
Albreda, Goree and Saint Louis.^ Many were sold to plantations along the
Mississippi River from New Orleans to Natchez. Thorough and detailed
examinations of contemporary events at and near these ports may
eventually make it possible to offer more specific conclusions about the
geographic origins and cultural backgrounds of Louisiana’s African slaves
than has been possible in other New World slave societies. As a result,
Louisiana may provide for historians a rare opportunity to explore the
complexities of African political and social structures in the New
World as
well as the African contribution to the social and economic development of
a part of North America that is often neglected. As an example of an
approach which stresses the value of wedding local histories in Africa and
in the New World, this essay will focus upon Senegambia since that region

accounted for roughly one-half of all the African captives brought to French
Louisiana between 1719 and 1731.^
BAMBARA SLAVES IN COLONIAL LOUISIANA, 1718-60 99

The
Atlantic slave trade brought together Africans of distinct and varied
religious, social and linguistic backgrounds in the hostile and oppressive
environment of New World slavery. Louisiana was no exception and the
colony had an ethnically, linguistically and religiously diverse African
population. Numerous ethnic groups lived in areas that fed the transatlantic
slave trade, and individuals from many of these ethnic groups were brought
to Louisiana. Approximately 2,000 were shipped from ports along the West
Africa coast from Portentic in northern Senegal to Cape Appolonia several
hundred kilometres to the south. An additional 2,000 came via the port of
Whydah on the Slave Coast, and approximately 300 more departed from
Cabinda. We
have only a small number of observations relating to the
ethnicity of slaves in Louisiana in' 1720—50. Despite the geographic
diversity of Africans arriving in the colony, a disproportionate number of
those to whom an ethnic label is attached were described as Bambara or
belonging to the Bambara nation.*
Between 1720 and 1750, few African slaves in Louisiana were identified
by nation. In fact, prior to the late 1730s most estate inventories did not even
identify individuals by name.’ Though there were more than 3500 Africans
living on plantations along the Mississippi river in the 1720s and 1730s, few
were ever identified by either name or nation before the late 1730s.
Thereafter, inventory records began consistently to note both.* Despite the
anonymity of individual Africans in early French colonial Louisiana, the
ethnonym ‘Bambara’ appears much more frequently than other
designations. of course, possible that large numbers of ethnic Bambara
It is,

from the region of the Niger bend were captured and transported to the
Senegambian whence, having been bought by Europeans, they were
coast,
shipped to Louisiana. It is possible, too, that Bambara were especially prone
to trouble in Louisiana and thus came to the attention of colonial officials.
That said, it is also important to treat with caution descriptions of slaves’
ethnicity given by Louisiana officials such as Michel Rossard, registrar and
royal notary of the colony in this period. It is, for example, not at all clear
that when Rossard recorded a slave as being Bambara that this meant the
same to Frenchmen as it did to the individual whose national or ethnic
identity it was intended to describe. In other words, eighteenth-century
meanings of the term Bambara are neither obvious nor certain.^
In searching for the origins of Senegambian captives brought to
Louisiana, iteasy to confuse the labels which contemporary Europeans
is

used to describe Africans with the ethnic origins of those individuals. How
both Europeans and Africans adopted, used and changed ethnic terminology
in the New World depended in large measure on how ethnic labels were
applied in Africa. We must begin, therefore, by exploring the complex and
often deceptive phenomenon of ethnic identification before we can attempt
100 ROUTES TO SLAVERY

to reach back into their African past to uncover the true geographical or

ethnic origins of Louisiana slaves.

Perhaps not surprisingly, Europeans tended to understand Africans in

European political terms. This has to be taken into account when examining
the terminology employed by eighteenth-century European observers to
describe Africans.'" Contemporary European descriptions of African
political groups and alliances were often misleading because European
conceptions of nation and race were not fully shared by the African groups
to whom they were applied." Using contemporary European labels of
ethnicity can, therefore, lead to confusion in historical studies of slavery and
the African diaspora because such labels do not always accurately describe
the ethnic origins of African captives. Indeed, uncritical use of definitions
of African ethnicity may create confusion for scholars of New World
slavery and may even result in erroneous conclusions being drawn about the
ethnic and cultural heritage of American slaves.
Though it may seem at first a trivial point, the implications of variations
and and meanings are potentially enormous. Some
shifts in ethnic labels
ethnic labels have been applied to African captives which did not accurately
describe any particular ethnic or language group. This is true of the
designation ‘Bambara’ and exemplifies the problems of using eighteenth-
century European classifications of Africans. Understanding how ethnic
labelswere applied and, more particularly, how they relate to the origins of
African slaves must be central, therefore, to the study of community
development and African identity in New World societies.
Despite the problems of identifying slave ethnicity and nation, such
terms have provided the framework within which some historians of slavery
have sought to suggest that common ethnic heritage between African slaves
necessarily implied cultural homogeneity in the New World.'- However, the
framework becomes imprecise, unstable and often simply unusable once
one begins to explore eighteenth-century definitions of ethnicity and nation.
‘Bambara’, for instance, is an example of a generic label sometimes
employed exclusively as an ethnic designation in modern slave
historiography in the United States, even though, in Senegambia in the
1720s, was a generic term
it for slave and was used to describe non-Muslim
captives more often than to identify precisely the ethnic heritage of
individuals. Similarly, ‘Guinea’, ‘Senegalais’, ‘Mina’ and other generic
even though the meanings of such words varied
labels cast large nets,
between regions and over time. For example, ‘Guinea’ could mean a slave
from almost anywhere on the West African coast; ‘Senegalais’ was a catch-
all geographical phrase; and ‘Mina’ was used to describe any slave from the
region extending from Elmina on the Gold Coast to the Bight of Benin.
Other examples include ‘Igbo’, which referred to slaves from the Niger
BAMBARA SLAVES IN COLONIAL LOUISIANA, 1718-60 101

Delta, and ‘Congo’, which often meant any person from the Kongo empire
or any captive purchased between ‘Cap Lopes [and] le Cap Negre’.'^
A word used to identify a captive’s ‘nation’ (or ‘pays’) did not always
correspond to his or her ethnicity. Moreover, even shared ethnicity between
captives did not necessarily imply a shared culture. '* Ethno-labels were often
littlemore than rough geographical pointers. Colonial Louisiana notary
records and inventories contain a number of references to obviously generic
or geographical African origins such as ‘nation Congo’, ‘nation Senegal’, or
‘nation Guinee’.'" Furthermore, it should be noted that such labels were
invariably determined by the captors rather than the slaves, and they should,
therefore,be used primarily as a guide to the geographical point of
embarkation or, in some cases, the point of capture of slaves rather than as a
description of their ethnic origin. Nevertheless, French colonial officials
insistedon re-interpreting African communal associations in terms of national
allegiance and tended to identify groups of Africans by their ‘nationality’. For
example, in 1728 a fifteen or sixteen-year-old youth brought aboard the ship.
La Renee Frangoise, was identified as ‘de nation Cap Apollonie’.'^ In cases
such as these, ethno-labels were obviously improvised by Europeans with
apparently little or no knowledge of African geography. A remarkable
example of European misunderstanding of African ethnic groups is provided
by the case of a slave named Benjamin, who was shipped to the French
Caribbean in the late eighteenth century and was identified as ‘un Congo de
nation Poulard’. ’’ ‘Congo’ referred to the large geographical area along the
central African coast while was a far more specific reference,
‘Poulard’
usually to someone from Senegambia.'* Though there is no equally absurd
example for Louisiana, Benjamin’s case nevertheless highlights European
ignorance of African ethnicity and geography.
Ethnic labels varied according to national or geographical perspectives.
As the involvement of Europeans in the slave trade increased in the
eighteenth century, so too did their knowledge of Africa. And as Africans
from hitherto unaffected regions fed the insatiable appetite of European
slavers, new peoples - and with them new ethnic tenns - entered the
vocabularies of Europeans. A French merchant trading in Senegal in 1723
meant something when he used the term ‘Bambara’ than a Muslim
different
slave trader from, say, Futa Toro. To the French trader, Bambara was an
abstraction. It could mean a ethno-linguistic group, or not. What mattered

most, however, was not the accuracy of the label, but whether it influenced
the price that might be obtained for the captive. To a Muslim slave trader,
on the other hand, identifying a captive as Bambara - meaning, in this case,
non-Muslim - may have been a way of avoiding, if not exonerating, his
complicity in the enslavement of a fellow Muslim, an activity clearly
forbidden by Islamic law. In addition, both of these definitions or
102 ROUTES TO SLAVERY
interpretations of Bambara probably differed from that used by an enslaved
African who may have used the term to identify himself or herself with a
group for whom the definitions ascribed by their enslavers held little or no
relevance. In the New World, as in Africa, what was perceived by Europeans
as an ethnic or national label could, on occasions, be a reference to a
religious designation. At other times, it may have referred to a slave’s
geographical origin or his or her social condition in Africa (i.e. whether slave
or free).’’’ In whatever context the ethnonym was used, as James Newman has
noted, ‘virtually everywhere on the continent, group boundaries, even
religious ones, were flexible and permeable. The notion of rigid and
unchanging “tribes” belongs to the minds of others; it is not evident in the
actions of Africans’.-” Thus, in addition to referring to the ethnic-linguistic
group Bambara (or Banmana, Bamana, and Bamanakan), the term
‘Bambara’ was variously used in the eighteenth century to refer to a slave,
especially at Galam and on the island of Goree; to a captive from east of the
Senegal river; to a slave soldier; to the ethnic group of an individual’s captor;
or to a pagan or a non-Muslim.-' The variety of meanings was further
complicated by the nature of ethnic Bambara society in the eighteenth
century. Like other expansionist military states, the Segu Bambara kingdom
depended on captives to augment and replace its members. According to
Richard Roberts, the ton, or military and social unit that formed from the
bachelor’s age set, ‘reproduced laterally, by incorporating young men taken
as captives into the ranks of the ton itself’.-- Young men of ethnic groups
other than Bambara were therefore forcibly assimilated into Bambara society
with the result that, in a sense, not even all ‘Bambara’ were ethnic Bambara.
Though the young men who constituted the warrior class spoke or learned to
speak Bamanakan,-^ it is not certain that they would have constituted a
homogenous cultural or linguistic group even if transported together as
slaves to the Americas. In any case, since the Segu Bambara state was
expanding from 1712, it is far from certain that these soldiers of the Segu
army would have been enslaved in large numbers. On the contrary, in fact, it

seems more likely that they would have been enslaving others.
Bambara was perhaps most often used as a generic term to describe
people of an, at best, vaguely defined group or geographic region. In a
monograph published in the late eighteenth century, Moreau de Saint-Mery
described Africans known as Bambara as coming from points to the east of
Senegal. In his description of African ethnic groups, he characterized
Bambara slaves as ‘amenes de plusiers centains de lieus a I’Est de I’Afrique
et vendus avec Bambara de I’Est], sous la denomination generique
lui [les

de Bambara, que ce mot sert a indiquer un grand corps sans graces. Le

sobriquet qu’ il a aux lies, est celui de voleiir de dindes et voleiir de moiitons,
dont il est tres-friand’.-''
BAMBARA SLAVES IN COLONIAL LOUISIANA, 1718-60 103

Saint-Mery’s characterization of Bambara as thieves is coincident with

the appearance of Bambara in a number of criminal cases in Louisiana.
More recently, Philip Curtin has suggested that the term ‘Bambara’ could be
interpreted as ‘a general designation for all Malinke-speaking peoples, or
even of all people from east of the rivers’.*^ Curtin claims that ethnic
Bambara which he means captives from the Niger bend, did
slaves, by
indeed constitute a large percentage of exports from the Senegambia after
the 1680s. Significantly, Curtin goes on to note that an exception to this
pattern occurred after 1721when warring factions in the Kingdom of Segu
united to defeat a common enemy, the Ormankoobe. The result was that ‘the
flow of slaves from the east was sharply cut’. Curtin claims that the next
evidence that trade from the east resumed dates from 1733.-’ It is, however,
precisely in this period when was disrupted that slaves
trade with the east
were shipped from Senegambia to Louisiana. This would seem to suggest
that most of the Africans sent to Louisiana from Senegambian ports did not
originate in the far interior.
Jean-Baptiste Labat, an eighteenth-century travel writer, described
slaves exported from the Senegambia in the first half of the eighteenth
century as Bambara. Labat’s use of the term, however, is not necessarily
indicative of the actual ethnic origins of the slaves departing from the region
and must be treated with caution. Labat’s Noiivelle Relation de I'Afriqiie
Occidentale is often cited by modem historians. Published in 1728, much of
the information contained in it from accounts of visitors
derives, however,
to West Africa in the previous century or was borrowed from other
accounts. In Noiivelle Relation, Labat identified what he described as a
‘very large kingdom’ situated between Timbuktu and Khasso called
Bambara Cana.’* Historians have sometimes assumed this to be a reference
to the kingdom of Marmari Kulibali (1712-25), even though it is more
likely that it refers kingdom of Kaladian Kulibali in the late
to the
seventeenth century. The latter interpretation would be more consistent with
the events during the 1680s, the era from which Labat drew much of his
information. In a similar account of the region of Lower Guinea, Labat
relied heavily on accounts from Willem Bosman and Nicolas Villault de
Bellefond, both of whom
had visited the coast of Africa between the 1667
and 1702. This period corresponds with the rise of the Bambara state of
Kaladian Kulibali, a kingdom described as ‘extensive but ephemeral’ by
Nehemia Levtzion.’*^

Among its various other meanings, Bambara referred to any enslaved

individual. The term was used in this fashion by French officials to mean
slave on the island of Goree. There, enslaved blacks were categorized as
‘bambaras’, to distinguish them from free blacks and mulattoes.’” This
application of the term in Africa dates from at least the 1730s. At the French
104 ROUTES TO SLAVERY
Fort Saint-Joseph at Galarn, too, Bambara referred to slaves. In 1731, and
again in 1734, in addition to the European residents, each rolle general at
the tort listed 80 and 40 Bambara, respectively.^' According to Abdoulaye
Bathily,Bambara was a term by which one designated domestic slaves
permanently employed by the Company at Fort Saint-Joseph who served as
guards in the captiveries or as cooks, labourers, or perhaps as translators.
Furthermore, Bathily claims, all slaves coming from Galam were presented
as ‘of the race Bambara or of the nation Bambara’. In truth, he argues, ‘these
slaves were captured in all the lands of the upper Senegal-Niger’,^- or, in
other words, from the region near Futa Djallon. Bathily believe that
throughout the early eighteenth century the slave traffic through Galam had
three principal sources - the Niger valley, the regions immediately
neighbouring Galam, and Galam - but that between 1720 and 1730
itself

slaves came primarily from the upper Falame and Futa Djallon. Caravans
brought captives taken from the Tenda region (Konyagi and Basari) and
from Jallonke in the Fulbe wars of conquest.
Bambara apparently also held a generic meaning for Parisian officials,
which may not have been the same as in Senegal. In 1723 Saint Robert, the
director of the Senegal concession, was ordered by the company to ‘retain
50 Bambara and to leave 20 other Bambara at Goree, 20 at Arguin, and 5 at
Bissau’.’^ It is hard to imagine that a company official in Paris cared whether
fifty ethnic Bambara were retained, or whether they were fifty Fulbe or fifty

Malinke. It is unlikely in any case that he understood the difference. In this

case, the official, writing from was not concerned with the ethnicity
Paris,
of the slaves left at those depots; by ‘bambaras’ he simply meant slaves.
Slaves or captives may also have been the intended meaning of a
reference to the cargo of Le Due dii Noailles in 1727. The ship’s log notes
that ‘Les Noirs charges au Senegal sont des Bambaras’.’^ Perhaps these
captives were, as the phrase suggests, ethnic Bambara. It is equally possible,
however, meant simply black Africans, perhaps as opposed
that the phrase
to Moors. The term was also used generically to mean captured slave
soldiers whose use was common in early eighteenth-century Senegal.^*
There is reason to believe that some of the above meanings associated
with the term Bambara, which enjoyed wide currency in Africa and in
Europe, were also used in French colonial Louisiana. Indeed, French
officials in the New
World sometimes added to the confusion of ethnic
identification. In a letter in 1731 to Paris in which he expressed his concern
over the possibility of an alliance between local Indians and Africans, the
Governor ot Louisiana, Etienne de Perier, described a conspiracy in which
an African came as a secret emissary from the Chickasaws to convince other
Africans to join the Anglo-Indian alliance against the Erench. ‘The Indians
[Erench: sauvages]', Perier wrote, ‘sent one of the negres who had joined
BAMBARA SLAVES IN COLONIAL LOUISIANA, 1718-60 105

them here to tell our negres that they would have their liberty and would
want for nothing among the English. This negre,' Perier continued, ‘is a
Bambara of a nation that the others do not understand’. Perier clearly did not
intend to imply that Bambara was the ethnic origin of this black Chickasaw
agent; rather he used the term generically, to imply that the agent was black,
Senegambian, or even a black slave soldier of the Chickasaw.^’ As one who
had served the Company of the Indies for several years in Senegambia,
Perier would have been familiar with the term as it was used in the West
Africa.^* Because he served in both Africa and Louisiana, Perier’s voice is

especially and cannot be easily ignored. His specific

authoritative
identification of the agent as a Bambara who was not understood by other
Africans argues against the existence of a homogenous linguistic group and,
even more importantly, indicates that it was normal to group Africans of
more than one nation under the umbrella term Bambara.
Bambara also carried religious connotations. Fulbe and other Muslim
traders or translators often identified non-Muslim captives as Bambara, by
which they meant non-believers or animists.^^ In his early twentieth-century
translation of oral histories of the region, Maurice Delafosse wrote that
while the Arabic word written by Sire-Abbas, an historian who recorded
Senegambian oral histories in the late nineteenth century, should be
translated literally as ‘les Berberes’ (or barbarians), he also noted that both
oral tradition and Sire-Abbas intended that the word be understood as
Bambara. This word, Delafosse explained,

non pas exactement la valeur d’un nom de tribu a proprement parler,

mais I’acception de ‘barbares, paiens sauvages’. C’est d’ailleurs
I’acception que revet communement au Soudan le mot Bambara, dans
la bouche des Musulmans; ceux-ci I’appliquent, non-seulment a la
tribu des ‘Bambara’ proprement dits (tribu apparentee de pres aux
Mandingues et repandue au Sahel, au Kaarta, a Bamako, a Segu, etc),
mais a quantite d’autres peuplades tres differents ethniquement des
Bambara propres.'**’

It is accepted that in the eighteenth century the term Bambara could refer to
a slave’s ethnic identity in the sense of his or her ethnic origin. Just as often,
however, it could be used for other reasons by slave traders. European and
African slavers eager to secure the highest possible price for slaves might
apply the label Bambara to a slave or group of slaves without detailed
knowledge of their geographic or ethnic origin in the belief that, since
Bambara were sometimes characterized as passive, they might earn a higher
sale price than others."*' Whether or not Bambaras actually fetched higher
prices is uncertain, but it underlines, nevertheless, the ethnic stereotyping
which occurred in both Africa and the New World.
 s

106 ROUTES TO SLAVERY

one of the most famous accounts of early Louisiana, Jean Baptiste Le
In

Page du Pratz, a relatively minor company employee, referred to an alleged

slave conspiracy in 1731/- Writing several decades after the incident and his
own departure from Louisiana, du Pratz alleged that the conspiracy was led
by Samba, an African commandeiir of the concession for which du Pratz
was responsible, and that all the Africans of his country would wrest control
ot the colony from the French and assume power/^ Samba, a name which
means ‘second may have been the same
son’ in the language of the Fulbe,
individual who had acted as a translator for the Company and had converted
to Christianity.^ Du Pratz claimed that Samba had not only led a revolt

against the French at Fort Arguin in Senegal in 1726, but, deported for that
crime, had also plotted to seize the ship L'Annibal which in 1727 was said
to have carried him to Louisiana. The probability that these incidents were
related or that the leader in each case was the same is small. There is no
documentary evidence other than du Pratz’s own self-serving version to
support this contention. Further, it would have been highly unusual in the
early eighteenth century to first exile and then spare the life of an African
who had any revolt against the French. This would be especially true of
led
revolts aboard ship. These were not uncommon during this period and in
almost every case one or several of the leaders was executed or thrown to
the sharks as an example to the rest of the slaves.'*^

On its 1727 voyage, many of L'AnnibaV slaves were sold at Cap Saint-
Louis in Saint Domingue. One hundred and fifty ‘Noirs invendus’ were,
however, brought to New Orleans by Perier who, coincidentally, was in
SaintDomingue on his way to Louisiana to assume his new post as
Commandant. Had such a notorious figure as Samba been one of the 150
aboard the L’Annibal, Perier would doubtless have been alerted to the fact.^’
Yet, several years later. Governor Perier, writing of the same incident in
which du Pratz supposedly played such a critical role, mentions neither du
Pratz nor Samba, nor, for that matter, any Bambara. Perier, in fact, seems to
have doubted the veracity of stories about the conspiracy when in a letter to
the Company in July 1731 he wrote that it ‘is not at all sure that these blacks
[accused of the conspiracy] actually plotted this dark deed, though all the
city was alarmed’. Combined with the characterization of the incidents
leading up to the execution of several African slaves in 1730 as an
ethnically-based conspiracy, the almost mythical, non-historical, treatment
of Samba and imposition of imagined cultural characteristics has
the
clouded our understanding of the historical record.
The persistent identification of Africans by nation in inventories and
legal proceedings reinforced for contemporaries and for many modern
historians the idea that African slaves recreated New World African national
alliances based primarily on ethnicity and, by extension, language. One
 BAMBARA SLAVES IN COLONIAL LOUISIANA, 1718-60 107

should not underestimate the influence of ethnicity on the lives of Africans

in the Americas. But historians cannot invent ethnicities where records
of
them do not exist nor should they allow contemporary European
assumptions about African ethnicities to distort the lens through which we
must necessarily view early African captives in New World societies.
One recent scholar has argued that Bambara slaves ‘arrived in Louisiana
in large numbers’ and that ‘they were truly Bambara’. Although a number
of slaves in Louisiana were identified as Bambara, it is not at all certain that
they were, in fact, ‘truly Bambara’.The source of some of the confusion is
based on slave testimonies. Under questioning a number of slaves did
indeed identify themselves as Bambara and this has led historians to
conclude that if they themselves say so, then they must be ethnic Bambara.
Yet there is reason to believe that they were not from the region of the Niger
bend. Why, then, would some identify themselves as Bambara?
There are several explanations to account for a slave’s alleged self-
identification as Bambara, even if the slave in question were of another
ethnicity. First, as has already been suggested, the names and nations
assigned to slaves were often arbitrarily assigned by captors or buyers.
Second, the method by which European slave traders identified particular
ethnic groups was based on information supplied by African traders and
often supplemented by only the vaguest notion of African politics. Third,
events in the New World could encourage a slave to identify with one
African community over another and to assume the ethnic identity of that
group. This is suggested by Barth, by Rodney (in the case of the Mandinga),
and by Reis (in the case of nineteenth-century Male or Muslims in Brazil).^”
For Louisiana’s Africans, Bambara may not have referred to an ethnicty per
se but instead to a group identification of another sort. Lastly, Africans often
spoke through translators whose own interpretations could easily mix with,
and conceivably contradict, the words of the slaves themselves.
An example of translations failing to express accurately the sentiments
of individuals is provided in September 1729 when several Africans
identified as Bambara were questioned before the Louisiana Superior
Council. For three of the interrogations the Council used a Frenchman
named Jean Pinet to translate.'' Pinet was a gunsmith who had lived in
Senegal and was married to a woman who may
have been a Senegambian
mulatto.” By November 1729, however, the Council had begun to employ
an African named Malene to translate for a slave who was also identified as
Bambara. Why was not used in this last case is unclear. He was still
Pinet
alive in 1731 and appears from census records to have owned six slaves.”
What is important, however, is that Pinet appears to have spoken a language
which was understood by the first three Africans brought before the Council
in September 1729. These ‘Bambaras’ were probably from the Galam area.
108 ROUTES TO SLAVERY

even if they had not had direct contact with the French there or served as
slaves, and probably spoke a language of the coastal region rather than
Banmanakan, a language of the far interior with which Pinet was unlikely
to have been familiar. Nevertheless, Pinet’s identification of the Africans for

whom he interpreted as Bambara could easily have been based on his own
knowledge of Senegal, though his conceptualization of the individual’s
nation would have differed from the African, Malene, who acted as
translator for another Bambara in November 1729. Again, it is important to
remember the context and meaning of the term ‘nation’ to eighteenth-
century Frenchmen such as Pinet and to remind ourselves that the word did
not necessarily carry any explicit territorial meaning and was not
synonymous with ‘ethnicity’ which is a modem constmct. The term ‘nation’
could be, and was, used in a variety of ways depending on the background
and purposes of individual writers or speakers.
Europeans ascribed a ‘national identity’ to some slaves, so slaves
Just as
disembarking in the New World were often given new names by ship
captains, slave traders, or owners. In many cases, these help to obscure the
ethnic, religious or geographic origins of slaves. Interestingly, modem
historians have accepted that names such as ‘Sans Souci’ given to slaves by
their owners were not their original ones, yet have also gone on to assume
that when a slave is identified as ‘de nation Bambara’ this is probably
accurate. If, however, one of the most important representations of an
individual’s identity - his or her name - could be changed, so too could his
or her ethnicity.
Modern historians have attempted to determine the African origins of
New World slaves from inventory records and from interrogations of slaves
accused of running away, conspiracy, and petty crimes, whether real or
imagined. A transcript of a slave interrogation could typically begin, as it

did for a slave accused of marronage in 1741, ‘Interroge de son nom, age,
qualite et demeure’. As recorded, the slave ‘a dit le nomme Pierrot negre
Esclave de Chaperon de nation Nago’."^ Of course, some of the information
gathered during such interrogations is suspect. In the majority of cases the
slave’s identity was probably already known and itwas therefore likely that
the preliminary questions as to name and ‘nation’ of origin were not asked.
Instead, information of this sort was probably simply entered by the
recorder on the basis of evidence given by the slave’s owner. Often in
interrogations of slaves, the circumstances of the case are outlined in the
opening statements. The questioning then follows. There is little, however,
to indicate that this was the actual order in which ‘interrogators’ obtained
information or even if questions relating to the nationality of slaves were
asked. A slave incorrectly identified as Bambara, Mina or Ibo upon arrival
in Louisiana would not have had an opportunity to correct the error. This is
BAMBARA SLAVES IN COLONIAL LOUISIANA, 1718-60 109

particularly so during interrogations when any attempt by a slave to

challenge the nationality ascribed to him or her by European or African
traders sometime could be interpreted as a form of deception and
earlier
carry severe, even mortal, consequences. As a result, it is likely that
information concerning a slave’s ethnicity can be as misinterpreted in
Louisiana to colonial officials as it was misrepresented by traders in Africa.
Despite the caution with which one must approach the term Bambara, it
nevertheless offers a vital clue as to the actual geographical origins of
Senegambians sent to Louisiana. While we clearly cannot take French use
of the term at face value, neither can we discard the term entirely. When
understood within the context of Senegambian history, Bambara is still a
valuable key to unlocking the origins of Senegambian captives brought to
the Americas in the 1720s. It is not my intention, therefore, to dismiss or
disregard the ethnic labels attached to individuals or groups of Africans but,
rather, to re-assess the significance of those labels in a manner
consistent
with their eighteenth-century usage, meanings and implications.
Simply put, the problem is as follows: if Bambara did not necessarily
refer to an ethnic group, then how do we explain the disproportionately
large number of Africans identified as Bambara
eighteenth-century
in
Louisiana? In addition, what did the term mean
to Senegambians in
Louisiana? It is likely that individuals identified as Bambara were of
Senegambian origin. Moreover, it is probable that most were non-Muslims
from the littoral rather than the far interior and may have been slaves in
Africa of the Fulbe or even the French prior to their sale and shipment to
Louisiana. For further evidence to support these suggestions, however, we
must look more closely at the history of Senegambia, particularly in the
period of the late 1720s when five ships carried some 2,000 slaves to
Louisiana.
For many decades prior to the 1720s, religion in Senegambia was both a
force which divided many people and an umbrella under which they united.
Religious affiliation, most notably Islam, united peoples of different
regions, villages, and lineages in the face of common enemies. But, if Islam
was an umbrella, it was a porous one for despite their profession of faith in
Islam and their public expressions of unity, the peoples of Senegambia
remained divided throughout the eighteenth century. Serious divisions
existed between the emerging theocratic states of Futa Djallon, Futa Bundu
and Futa Toro, and even between individual villages.^^ Religion could unify
people across regions, class, language and geographical origin, but it could
also provoke deep divisions along the same or similar lines. It could be, and
was, a divisive which segregated peoples and communities,
force
sometimes causing conflict between members of the same ethnic or
linguistic group. The degree to which Senegambia was disrupted by
110 ROUTES TO SLAVERY

political, economic and ecological upheaval is important to the discussion

of Louisiana slaves tor three reasons. First, it calls into question the

assumption most of the slaves exported from Senegambia during this

that

period came from the far interior or, more specifically, from near the Niger
bend. Second, the level of commitment to Islam, which was forced upon
many, may not have been so great even while hostility toward Islam,
especially for those captured and enslaved in the jihads, may have been
significant both in Africa and in the Americas. Finally, the enslaved
Senegambian population was ethnically heterogeneous, and so were those
exported as captives. Taken together, these three factors are important for
our understanding of the early Louisiana slave population because even
though the largest numbers of captives brought to the colony came from the
Senegambia their common geographical origins, religious unity and shared
languages did not necessarily imply that in Louisiana they developed any
political alliances or even cultural affinities.

Within the crescent formed by the Senegal river as it winds through the
region from the Futa Djallon highlands to the coast there are more than one
hundred ethnic groups and sub-groups. Most belonged to a large yet vague
cultural zone known as Mande. Yet, according to Wondji, ‘each ethnic
group, speaking a language distinct from those of its neighbors, [was] aware
of its individuality’.^* Among
major groups involved in the
the
Senegambian struggles of the 1720s were the Jallonke Fulbe, Sissibe,
Mande Juula (Jaxanke), Timbo, Soso, Sereer, Malinke, Bobo and Fulbe in
the Gambia River basin as well as related peoples from south of the
Casamance river such as the Kasanga, Papel, Beafada, Bijago, Nala and
These representative examples are sufficient to show the great
Balante.^'^

variety of ethnic and linguistic differences in eighteenth-century

Senegambia. Even more importantly in the present context, these various
groups were surely represented among the slaves arriving in Louisiana.
In the first decades of the eighteenth century non-Muslim ceddo
(warlord states) of Waalo, Kayor and Baol controlled the coastal region of
the Upper Guinea coast between Senegal and Guinea-Bissau. In the mid-
1720s, both Kayor and Baol, whose rulers and tyeddo^ armies were
especially hostile to Islam because Islam offered ‘an alternative source of
political power’, were themselves embroiled in civil wars as competing
factions vied for political control of the region.^'
Futa Toro, like the northern Wolof state of Waalo and the Soninke
Kingdom of Gajaaga, was Moroccans on an
the object of invasions by
almost annual basis beginning around 1720.*- Adding to the political
disruption caused by these invasions and severe grain shortages between
1723 and 1725, the leadership of Futa Toro was in dispute until Samba
Gelaago Jegi assumed the position of Satigi in 1725. He and his mercenary
BAMBARA SLAVES IN COLONIAL LOUISIANA, 1718-60 111

allies captured and enslaved their rivals, selling some to European factors;
many were probably Fulbe and some may have been Muslims. Some of
these individuals could easily have been identified as non-Muslims or
Bambara. Since Islamic law proscribed the sale of Muslims by fellow
Muslims, the network of slave traders extending to the coast may have
labelled Muslim captives non-Muslim (or Bambara) for fear that their
captives might not otherwise be purchased.*^
The Senegambia in 1720-35 was marked by periodic civil and
history of
ecological disruption. The latter included drought, infestations of locusts,
and famines and served to limit severely the numbers of slaves transported
through the region from the interior in the 1720s and 1730s. The same
events also led, however, to the displacement of people within the region,
leaving them vulnerable to kidnapping, organized slave raids and sale.^ In
was a steady flow of captives from the region
the seventeenth century there
beyond the confluence of the Senegal and Falame rivers, but in the
eighteenth century this flow of captives was periodically disrupted,
especially during the 1720s when slave shipments to Louisiana were at their
height.
The French fort at Galam was the principal nexus of slaves transported
both from the interior and from the region immediately surrounding Galam.
Beginning in late 1728, Senegambian exports began to rise, though there is

Galam itself was the primary source for the increase.

reason to doubt that
Throughout the summer of 1729, at least three French ships, Le Saint-Louis,
Le Saint-Michel and La Nereide, were moored near Goree hoping to
complete purchase of slaves. There were, however, delays in
their
shipments expected from Galam. The captain of La Nereide, Contault
Dentuly, wrote in June that his ship and Le Saint-Michel were hoping to
receive slaves from Galam because at that time ‘there were only blacks for
Le Saint-Louis' On 23 October, however, the captain of Le Saint-Louis,
Breban, wrote that he was still awaiting his captives and went on to note that
‘we have learned from a courier from Senegal that the barques from Galam
have descended without captives’. Despite this all three ships eventually
managed to sail forAmerica with apparently full loads of slaves, Le Saint-
Louis sailing with 380 slaves. La Nereide with 203, and Le Saint-Michel
with at least 350. It is unlikely that these slaves came from the interior since
expectations of slave deliveries from Galam were low by 1729; as Breban
wrote in his log in 1729, ‘the 6-800 captives from Galam are lost this year’.
Moreover, shipments of slaves from Galam were small at other times in the
1720s; in 1720, for example, only 260 slaves had been dispatched from the
port. Overall, therefore, it seems highly unlikely that slaves forwarded from
Galam constituted more than a small fraction of the 1,400 or so slaves
exported each year in 1729-31 from Senegambia on French ships. It seems
1 12 ROUTES TO SLAVERY

that these captives came for the most part from sources much closer to the

coast.
The rise in French slave shipments from Senegambia after 1724 may
have been the result of increased religious warfare in the region. The
drought which struck the region in 1723 appears to have hindered the supply
of slaves reaching markets Goree and other places along the Senegal
at

coast in the following year; fewer than 500 were shipped from Senegambia
in French ships in 1724. The drought conditions which again struck the

region in 1729 and 1730 did not, however, have the same effect as in
1723-24, because in the later years the captives were of more local origin
and the impact of drought on movements of slaves from more distant
sources was therefore less severe. On balance, the slaves shipped by the
French after 1724 were probably primarily of Senegambian origin (i.e. the
littoral) rather than from the Niger bend.“ More specifically, it is likely that

a large proportion came from areas near or bordering Futa Toro, Futa Bundu
and Futa Djallon where the majority of the fighting occurred. This suggests
that captives would largely have been non-Muslim Mandinka from
Bambuk; partially Islamicized Soninke from Gajaaga; Muslim and non-
Muslim Fulbe residing throughout the entire region; and persons from
smaller ethnic groups.
Amid the ecological disruptions of the 1720s, Muslim Fulbe and their
allies launched from about 1725 a series of jihads against their neighbours.
They captured and enslaved non-Muslims, including non-Muslim Fulbe, as
well as rivals who were Muslims or Muslims whose faith did not measure
up to the standards of the Jihadists. At the same time, non-Muslims captured
and enslaved Muslims and non-Muslims alike. Many of these individuals
would have been men and either full or part-time soldiers. In addition, bands
of slave raiders generated further supplies of slaves throughout
Senegambia. European merchants purchased the captives arising in these
various ways, shipping them to the Americas without regard to ethnicity or
religion. Added were slaves brought from much further
to these captives
inland, traditionally from the Niger bend region. These captives were
brought by Senegambian middlemen where they were then
to the coast

purchased by Europeans factors for shipment overseas. The trade from the
Niger bend, however, was irregular during the 1720s and the number of
captives originating from this source, as we have seen, was probably
relatively small.
A
combination of economic and political conditions made slaves
coming from the Niger bend to Senegambia less attractive to both European
and African traders than individuals captured more locally. Drought and
famine, exacerbated by war, had a negative impact on the scale of slave
shipments from the interior in the 1720s.^** In an important analysis of the
 BAMBARA SLAVES IN COLONIAL LOUISIANA, 1718-60 113

grain trade, Searing has argued that a critical link existed between the
harvest of grain and the commerce in slaves. The harvest of slaves and gum
arable, sought by Atlantic merchants,’ Searing argues, ‘was directly related
to the harvest of grain. In the simplest sense, this relationship existed
because the ability of Atlantic merchants to export slaves depended directly
on their ability to feed them from the moment of purchase until they
departed for the Americas. This argument is, of course, consistent with
the suggestions made here that in 1723-24 and again in 1729-31 political
and ecological impediments militated against the shipment of large numbers
of slaves from the Niger bend and reinforces the claim that the captives
exported by the French in these years came from areas closer to the coast.
Even though Senegambia accounted for almost half of all the African
slaves brought to Louisiana in the 1720s and early 1730s, it is not
necessarily the case that this common
would translate into the
origin
foundation of an ethnic, linguistic, or other community in the French
colony. While it is tempting to assume that slaves from the same
geographical region of Africa would develop mutual affinities once
transported and re-settled in the New World, it would be premature to make
this assumption in the case of slaves shipped to Louisiana. Separated from

their families and living in a strange environment, slaves would have

organized their communities and sub-communities according to a variety of
factors. These include, among other things, shared (or similar) culture,
language, religion, personal preferences and residence. Even Africans of the
same ethnic group may not have found common ground in the New World
if there was something in their past or their enslavement which mitigated
against such bonding. For example, Fulbe captives arriving in Louisiana
were not necessarily bonded by their common condition or even as
members of the same linguistic group. Non-Muslim Fulbe who had been
victims of jihad may not have associated with Muslims, Fulbe or otherwise,
regardless of their mutual condition. The same reasoning would apply to
other slaves from Senegambia thrown together on Louisiana plantations.
In an interesting and revealing passage in his Histoire de la Loiiisianne,
du Pratz advised that for domestics and for comrnandeiirs he chose only
Senegals (‘who are called amongst themselves, Djolaiifs') because of their
‘fidelity and gratitude’.’" He also characterized Senegals as having ‘plus de
fidelite & I’esprit plus penetrant que les autres, & sont par consequent plus
propres a apprendre un metier ou a servir; il est vrai qu’ils ne sont pas si

robustes que les autres pour les travaux de la terre, & pour resister a la
grande chaleur’.”
Du Pratz also wrote that seemed as if Senegals, whom he also
it

described as the ‘most black’, were ‘bom leaders’ (‘nes pour commander’).
Though speculative, it is interesting to compare these characterizations with
1 14 ROUTES TO SLAVERY

those attributed to Muslim slaves in English North America.^- It is possible,

given his name, du Pratz’s own preference for Senegals (or Djolaufs), and
his cultural characterizations, that du Pratz’s own commandeur. Samba,
whom we encountered earlier, may have been a Muslim, perhaps Wolof,
Fulbe, or Tukulour.^^ Du Pratz also claimed that, in the case of Senegals,
once they were made servants ‘one sees them sacrifice their own friends to
serve their masters’. If true, this raises the intriguing possibility that the
‘own friends’ of these Senegals, as du Pratz imagined them, could have
been non-Muslims in disharmony with Muslim Senegals. It is certainly
possible that some of those shipped from Senegambia may have retained an
antipathy towards fellow captives forged during political upheavals in

Africa. There is more than vague suggestion that

as yet, however, little

Muslims and non-Muslims from Senegambia continued their conflicts in

Louisiana.
The death in 1751 of a slave named Marboux illustrates the importance
of appreciating how Senegambian political and religious divisions affected
New World African communities and how those basic divisions may have
changed over time. Marboux’s death arose out of a confrontation between
himself and another slave, Francois, a confrontation that may have had a
meaning for Africans beyond what French officials were willing to
understand or were even capable of understanding. In April 1751, Marboux,
whose name was probably derived from marabout (a spiritual leader and
religious adviser in Senegambia), was fatally stabbed by Francois with a
small knife which Marboux wore around his neck. The judge in the case,
Jean Baptiste Raguet, decided that the stabbing was accidental and that
Francois was intoxicated at the time, but there may have been more to the
incident than this. As with many cases dealing with conflicts between
individual slaves, whites rarely cared to investigate too deeply. But at a
hearing held several days after the event, four slaves named Pierrot, Jean,
Pierre Birame, and Pierre testified that they witnessed the stabbing. All, like
Francois, were baptized Christians.’^
Marboux’s origins are unknown, though his name strongly suggests that
he was a Muslim. The term ‘marabout’ or maRabu - from the Arabic
morabit - referred to Islamic scholars or saints who were religious and
political teachers Senegambia. Marabouts were also
and leaders in

consulted by the local population and often provided them with amulets, or
small leather pouches in which verses from the Qur’an were placed. These
were designed to protect the wearer from harm.’^ The death of Marboux
may have been an accidental stabbing as the Superior Council ultimately
ruled, but it is possible that religious differences may have created
antagonism between the men involved in the incident. The questions raised
by the incident resonate loudly when seen in the context of Senegambian
BAMBARA SLAVES IN COLONIAL LOUISIANA, 1718-60 115

and underline the fact that relationships between Bambara,

religious strife
Muslims, Christians, and captives brought together from geographically and
culturally dissimilar regions contributed to world of
the cultural
transplanted Africans. Much still remains hidden, however, behind
ethnonyms, Christian names and nicknames which have little bearing on or
resemblance to the original identities of the individuals who carried them as
reminders of their enslavement.
This essay has sought to suggest that the national or ethnic labels of
slaves in Louisiana were flexible and subject to different meanings and
interpretationsdepending on the speaker, the language and the context of
usage. Ethno-labels alone cannot be used as a precise indication of African
origin, and in any case terms such as ‘Bambara’ can be useful only if
understood within the contexts of West African history. Slaves from the
same ethnic group were often identified by different labels such as Tula,
Poulard, and Fulbe, while others with little or no relation to each other could
fall under the umbrella of labels such as Bambara, Mina, or Congo. As
Africans consolidated their communal affiliations within the larger French
and Indian communities of Louisiana, the significance of group
identifications changed still further. Ethnicity is a complicated phenomenon
and one understood, especially in the context of a slave system that
little

often purposely attempted to sever individuals’ links with their past.^^

Ethno-labels remain important pieces of evidence relating to African
slaves’ origins, but they cannot be accepted without question. For
problem with terms such as Senegal and Bambara, Igbo and
historians, the
Mina, or Congo and Angola is that one can rarely be certain what the words
meant in the eighteenth century and how their meanings differed through
time and between individuals. Clearly, ethnic labels did not always mean the
same thing to all people. It is likely that in many cases the word ‘Bambara’,
as used by Africans in Louisiana, meant a group of Louisiana slaves who
identified themselves as from Senegambia and who were initially distinct
from Muslims and later perhaps from Christians.
Bearing in mind the permeability of ethnic groups and boundaries in the
western Sudan and the inaccuracy of Senegambian ethnic labels as used by
European observers, accepting definitions of Bambara - and, by extension,
other ethnonyms - that are based upon assumptions of colonial officials is
highly problematical. What at first glance seems to have been a community
of ethnic Bambara slaves among the Louisiana population
was in all
probability something different. In fact, it is unlikely that a community of
ethnic Bambara formed the core of the African creole community in
eighteenth-century Louisiana. Rather, many Senegambians - ethnically,
religiously, and, perhaps, linguistically heterogeneous - either identified
themselves or were identified by others as belonging to a group collectively
116 ROUTES TO SLAVERY

known as Bambaras. For the individuals included, however, the term

referred to something other than an explicit reference to their geographical
origins or to their African social or language group. Perhaps, as Bathily
suggested, many Bambaras were slaves in Senegambia before being
captured and sent to Louisiana. Equally possibly, they may have been non-
Muslims of many different ethnic groups captured in jihads and sold to

European slavers.
Bambara in colonial Louisiana comprised all of the
In all probability, the
above groups as well as some ethnic Bambara (or Banmana) from the Niger
bend. Whether they ‘became’ Bambaras while in Africa or purposely chose
to identify themselves with others who were called Bambaras after arriving

in the New World, the African identity known as ‘Bambara’ was probably

far more complex than most historians of the Americas have been willing to
acknowledge. The presence of a group of individuals who identify
themselves as Bambaras is perhaps most interesting in what it suggests
about the presence and significance of Senegambian Muslims in colonial
Louisiana. The term Bambara - comprising a group of peoples from
Senegambia distinct from, and probably antagonistic toward, Muslims -
poses important historical questions with profound implications for the
cultural formation of creole communities in French North America. These
questions will remain unanswered, however, if historians continue to accept
uncritically the lexicon of eighteenth-century French colonial officials.
The Bambara phenomenon - when examined alongside other African
ethno-national groups such as Igbo, Mina, and Congo - can extend our
understanding of African ethnic and cultural re-grouping in the New World.
It is a key to helping us unlock the identities and communal development of
New World Africans, an enterprise which rests on a clear and precise
understanding of local histories in both Africa and the New World
interpreted within the broader framework of an Atlantic perspective. In
examining other ethno-labels such as Congo and Mina, historians must
determine who applied the label, if possible why, and then decide what
significance the application of a common label has for group identity. It is a
tall order, but one which is essential if we are accurately to identify the

origins of Africans in the Americas.

NOTES

1. Versions of this paper were presented at Tulane University in 1994, the Louisiana Historical
Association Conference in Houma, Louisiana, in March 1995, and at a symposium on ^he
African Diaspora and the Nigerian Hinterland|held York University, Ontario, Canada, in
at

February, 1996. would like to

1 thank Carl Brasseaux, Angela A. Caron, Emily J. Clark, Katy
Coyle and Gwendolyn Midlo Hall for reading early versions of the paper, and Paul E.
 BAMBARA SLAVES IN COLONIAL LOUISIANA, 1718-60 117
Lovejoy, Philip D. Morgan, Patrick Manning, and especially Sylvia R. Frey for
their helpful
comments on later versions.
2. See Daniel H. Usner, Indians, Settlers, and Slaves in a Frontier Exchange Economy: The
Lower Mississippi Valley before 1783 (Chapel Hill, North Carolina, 1992); Gwendolyn
Midio Hall, Africans in Colonial Louisiana:
The Development of Afro-Creole Culture in the
Eighteenth Century (Baton Rouge, Louisiana, 1992).
3. ‘Recensement general des habitations le long du fleuve - 1731’, Louisiana, Historical
Collection, Records of the Superior Council 1718-1769 (hereafter LHC, RSC).
After 1731,
only one ship - Le Saint Ursin 1743 - was recorded as having brought slaves directly from
in
Africa, but slaves may have been smuggled into the colony from the Caribbean
in the 1750s
in small numbers.

4. Jean Mettas, Repertoire des ExpMtions Negrieres Frant^aises au XVIlIe Siecle, Vol.I,
Nantes
and Vol. Autres Ports, eds. Serge and Michele Daget (Paris, 1978-1984).
II,

5. Senegambian ports accounted for some 45 per cent of the Africans brought to Louisiana
between 1718 and 1731, with Whydah accounting for 40 per cent and other points along the
coast including Cabinda 15 per cent. Unfortunately, data relating to ethnic composition
of
Louisiana slaves prior to the 1750s do not exist. The relatively small sample of existing
references prior to 1750 makes impossible any estimate comparable to that made by David
Geggus for Saint Domingue. Even for Saint Domingue, Geggus cautions that definitive
conclusions about the ethnic composition of slaves on eighteenth-century sugar plantations
cannot be drawn from surviving records because the available sample of records was heavily
weighted toward the late eighteenth century. After 1769, Louisiana was no longer a French
colony. Consequently there is no easy way to compare slave ethnicity in the two colonies.

David Geggus, ‘Sex Ratio, Age and Ethnicity in the Atlantic Slave Trade; Data from French
Shipping and Plantation Records’, Journal of African History, 30 (1989), pp. 23-44.
6. LHC, RSC records. Baptismal records of the Cathedral of Saint-Louis, ANC C13A. See also
Hall, Africans in Colonial Louisiana, esp. ch. 1, 2. The number is large only in comparison
with other ethnicities. From recorded references to African nations (including ethnically-
all

based names, i.e. Louis Congo) during the period 1720-50, the number of individuals
identified as Bambara is less than twenty-five persons. No other ethnic label appears more
than three times.
7. See, for example, the estate inventories for Pierre Manade, Bechemin Corbin, and Joseph
Paris-Duvemay (LHC, RSC 1728100601, 1736012301, LHQ 21 (October 1938), p.987), as
well as the estate sale for RSC 1728091701). Of the eleven slaves sold’ by
M. Brusle (LHC,
Brusle only three were identified by name. Exceptions were for Africans baptized with
Christian names or for individuals being tried for crimes. In the Sacramental records of the
Saint Louis Cathedral at New Orleans more than 300 Africans were baptized in the 1730s.
The vast majority of these, however, were infants and they almost invariably took the name
of one or the other of their godparents, often white French.
8. See, for example, the estate inventory of le Marechal d’Asfeld, LHC, RSC 1738012401.

9. See Stephan Palmie’s review of Gwendolyn Hall’s book in Africa, 64 (1994), pp. 168-71.
10. See, for instance, Geggus, ‘Sex Ratio’, pp.34-9.
11. For a discussion of the meanings associated with the word ‘nation’ in eighteenth-century
Europe, see E.J.Hobsbawm, Nations and Nationalism Since 1780: Programme, Myth,
Reality (Cambndge, 1990, 2nd edition), pp. 14-45.
12. Stephan Palmie, ‘Ethnogenetic Process and Cultural Transfer’, in Wolfgang Binder (ed
).
Slavery in the Americas (WUrzburg, 1993). Palmie discusses the relationship of ‘tribal
affiliations’ with naciones in Cuba and cautions against a direct transfer of ethnic labels.
13. Roseline Siguret, ‘Esclaves d’indigoteries et de cafeieres au quartier de Jacmel (Saint-
Domingue), 1757-1791’, Revue Fran(;aise d'histoire d'Outre Mer, 55, no. (1968), 2
pp. 224-5. For Igbo involvement in the transatlantic slave trade, see the essay by Chambers
in this volume.
14. Frederick Barth cnticized the once popular notion that a race equals a culture equals a
language, and even that a society equals a unit which rejects or discriminates against others;
 ,

1 18 ROUTES TO SLAVERY
see Le Page and Andre Tabouret-Keller (eds.), Acts of Identity: Creole Based
R B.
Approaches to Language and Ethnicity (Cambridge, 1985), p.208.
15. As an e.xample of a generic application of the term Congo, see ANC C13A 9, fo. 267-8,
Louis Congo.
16. Mettas, Autres Ports, p.698.
17. Gabriel Debien et ai, Les Origins des Esclaves des Antilles (Extraits du Bulletin de
r Institute fran(;ais d’Afrique Noire - B, 1961-1967), pp.241, 243. Debien cautions, ‘On ne
sera done jamais trop prudent dans les identifications ethniques, et ce ne sera qu’a des
conclusions tres generales qu’on pourra aboutir’.
18. Benjamin conceivably could have been a Peul from northern Cameroon.
19. For example, in Brazil Male referred to Muslims while Mina meant any captive taken from
the Bight of Benin; Joao Jose Reis, Slave Rebellion in Brazil: The Muslim Uprising of 1835
in Bahia, translator, Arthur Brakel (Baltimore, 1993).
20. James L. Newman, The Peopling of Africa: A Geographic Interpretation (New Haven,
1995), p.6.
21. See Philip D. Curtin, Economic Change in Precolonial Africa: Senegambia in the Era of the
Slave Trade (Madison, Wisconsin, 1975), pp. 178-9; James F. Searing, West African Slavery
and Commerce: The Senegal River Valley 1700-1860 (Cambridge, 1993), p.l07; Maurice
Delafosse, Sire-Abbds-Soh, Chroniques du Fouta Senegalais: Traduits de Deux Manuscrits
Arabs Inedits (Paris, 1913 edition).
22. Richard Roberts, Warriors, Merchants, and Slaves: the State and the Economy in the Middle
Niger Valley, 1 700-19 14 (Stanford, 1987), p.34.
23. The Banmana people spoke a Mankdekan language called Bamanakan or Bamana which is

more recent times this language is

closely related to Malinke and Maninka. In also called
Bambara. See Barbara F. Grimes (ed.), Ethnologue: Languages of the World, Electronic
Ethnologue Database, 12th edition.
24. M.L.E. Moreau de Saint-Mery, Description Topographic, Physique, Civile. Politique et
Historique de la Partie Fran^aise de Lisle de Saint-Domingue, 2 vols. (Philadelphia, 1797),
1, p.27.
25. A case in which several slaves identified asBambaras in 1729 did indeed involve the theft
of livestock; in this case a heifer and some chickens. LHC, RSC 1729090503, 05,06;
1729111601.
26. The rivers would presumably be the Senegal and Falame; Curtin, Economic Change,
pp. 178-9.
27. See also Daniel H. Usner, ‘From African Captivity American Slavery; the Introduction of
to
Black Laborers to Colonial Louisiana’, Louisiana History, 20 (1979), pp.25^8.
28. Jean-Baptiste Labat, Nouvelle Relation de I’Afrique Occidentale, 3 vols. (Paris, 1728), 111,

p.334.
29. Adam Jones, ‘Semper Aliquid Veteris: Printed Sources for the History of the Ivory and Gold
Coasts, 1550-1750’, Journal of African History, 27 (1986), pp.2 15-35; Nehemia Levtzion,
‘The Bambara States’, Cambridge History of Africa, from c.1600 to c.1790, ed. Richard Gray
(Cambridge, 1975), IV, p.l75.
30. Searing, West African Slavery and Commerce, p.l07.
31. Abdoulaye Bathily, Les Portes de I’Or: Le Royaume de Galarn (Senegal) de I'Ere
Musulmane au Temps des Negners (Vllle-XVlIe Siecle) (Paris, 1989), pp. 259-60, 284.
32. Ibid., p.267.
33. Ibid., p.264.
34. Hall, Africans in Colonial Louisiana, p.68.
35. Mettas, Autres Ports.
36. Martin A. Klein, ‘Servitude among theWolof and Sereer of Senegambia’, in Suzanne Miers
and Igor Kopytoff (eds.), Slavery in Africa: Historical and Anthropological Perspectives
(Madison, Wisconsin, 1977), pp. 335-63.
37. 10 December 1731, New Orleans, Pener Company of the Indies, C13A, 13, 63-4.
to the
38. Marcel Giraud, A History of French Louisiana: The Company of the Indies, 1723-1731
BAMBARA SLAVES IN COLONIAL LOUISIANA, 1718-60 119

vols., translator, Brian Pearce (Baton Rouge, Louisiana, 1987), V, p.54.

39. Levtzion, ‘Bambara States’, p.I7I.
40. The Arabic word is translated by Delafosse as ‘paien’ or pagan; Delafosse, Sire-Abbas-Soh
p.200.
41. Bathily, Les Ports de I’ Or, p.265.
42. Le Page du Pratz actually dates the conspiracy to 1730; Jean-Baptiste
Le Page du Pratz,
Histoire de la Louisianne, 3 vols. (Paris, 1758),
III, pp.304-17. The incident which led to the

execution of several Africans actually occurred in 1731. ANC C13A, letter from Perier, f. 85,
21-28 Juillet, 1731.
43. Le Page du Pratz does not indicate from what country Samba came. Hall speculates that
Samba was Bambara and that the conspirators spoke that language, but cites no documentary
evidence of this; Hall, Africans in Colonial Louisiana, pp. 107-8.
44. Michael Gomez, ‘Muslims in Early America’, Journal of Southern History, LX (1994), no. 4,
p.685; LHC, RSC 1728071001. There were many Africans in the colony named Samba and
it is not at all certain which one was the Samba to whom du Pratz referred.
Samba was a
fairly common name in Senegambia, but it is also a Kongo name; John Thornton, ‘Central

African Names and African American Naming Patterns’, William and Mary Quarterly, 3rd
series, 50(1993), pp.736-7.
45. Du Pratz, Histoire, III, p.3 15. For an account of du Pratz’s version of the conspiracy, see Hall,

Africans in Colonial Louisiana, pp. 107-11. There are many inconsistencies in du Pratz’s
version, but easily the most glaring is his assertion that Samba was involved in the revolt
aboard the ship L’Annibal. Hall suggests that Samba came to Louisiana in 1726 aboard the
L’Annibal. A revolt did, indeed, occur aboard the ship, but not until its 1729 voyage and on
this voyage the ship did not go to Louisiana. Records of the 1726 voyage make no mention
of a revolt or conspiracy. In the unlikely event that Samba was detained for almost three
years in Senegal before being exiled, then he was not the Samba who acted as interpreter for
the Company in 1728. Neither Samba nor any Bambaras are mentioned in logs for either
voyage. See Mettas, Autres Ports', LHC, RSC 1728071001.
46. See, for example, Mettas, Nantes and Autres Ports. After a revolt aboard the ship L’Affriquin,
the two leaders were to be made an example to the others. They were executed and their
corpses were hung from the mast. Other examples include Le Courrier de Bourbon bound
for Grenada and Louisiana in 1723. A plot was uncovered after the crew threatened to torture
two female captives, one from
Goree and the other from Senegal. The leader of the plot, a
forty-five-year-old ‘sorcier’, was hanged from the mast and shot. Dunng another especially
grisly journey aboard Le Dauphin, the crew put down two revolts. After overcoming the
slaves involved in the first, the leader was hanged from the mast as an example, shot, then
thrown to the sharks. Other references are made in ships’ logs to slave revolts. Often slaves
were killed in revolts, as in the cases of L’Aimable Renotte in 1730, when 33 died and Le
Neptune in 1729 when three died. On occasion slaves were thrown overboard alive as befell
a slave involved the second revolt aboard Le Dauphin on its fateful 1723 voyage.
47. Mettas, Autres Ports, lettre de Perier, du Cap, 8 fevrier 1727.
48. Perier to Company, Cl 3A 1 3, f 85, 2 1-28 juillet 1731.
49. Hall, Africans in Colonial Louisiana, p.43.
50. Walter Rodney, ‘Upper Guinea and the Significance of the Origins of Afncans Enslaved in
the New World’, Journal of Negro History, 54 (1969), p.335. For the relationship of Male to
Muslim, see Reis, Slave Rebellion in Brazil, Frederick Barth, Ethnic Groups and
Boundaries: The Social Organization of Culture Difference (Boston, 1969), p.l4.
51. LHC, RSC 1729090503, 05, 06.
52. ANC C13A II, f349.
53. Recensement-1731 LHC.
54. Hobsbawm, Nations and Nationalism, p. 17.
55. LHC, RSC, 74 101 1602.
1

56. LHC, RSC, 174101 101 (translated; ‘Asked

1 his name, age, quality, and residence. Answered
“Pierrot, black slave of Chaperon, of the nation Nago’’’). See also slave interrogations LHC,
120 ROUTES TO SLAVERY
RSC 1729111601, 1729090503.
57. Michael A. Gomez, Pragmatism in the Age of Jihad: the Precolonial State of Bundu
(Cambridge, 1992), pp. 52-73.
58. C. Wondji, ‘The States and Cultures of the Upper Guinea Coast’, in B .A. Ogot (ed ), UNESCO
General History' of Africa, 5 vols. (Paris, 1992), V, p.368. Most Africans captured in this area
would have spoken a language in the Western Atlantic branch of the Niger-Congo family. This
family includes, for example, the languages Diola, Serer and Wolof, Coniagui and Baga (in
modem Guinea), Temne, and Foulfoulde (Fula, Peul) (Green, Languages of Africa).
59. Wondji, ‘States and Cultures’, pp. 368-97. Wondji discusses at length the difficulty of
producing an historical synthesis when describing peoples and societies that did not belong
to large political states.
60. Tyeddo refers to non-lslamic rulers and the paid or slave soldiers in their armies (Martin A.
Klein, ‘The Impact of the Atlantic Slave Trade on the Societies of the Western Sudan’, 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 (Durham, North
Carolina, 1992), p.35).
61 . Searing, West African Slavery and Commerce, p 80.
62. Ibid. Moroccan forces also attacked northern Senegalese kingdoms in the 1670s and, for a
brief time, controlled Waalo, Futa Toro, Jolof, and Kajoor (Klein, ‘Servitude among the
Wolof and Sereer of Senegambia’, pp. 344-5).
63. In addition to the Bambara label, a number of Muslim names appear in Louisiana
significant
colonial records. In the 1720s and 1730s, these Muslims could have only come from the
Senegambia. For more on the presence of Muslims in Louisiana, see Peter Caron, ‘Problems
and Approaches to the Study of the African-American Slave Community; Louisiana, A Case
Study’ (unpublished MA thesis, Tulane University, New Orleans, 1996).
64. Klein, ‘Servitude among the Wolof and Sereer of Senegambia’, pp. 343-9.
65. Mettas, Nantes and Autres Ports.
66. Bathily, Les Ports de I’Or, p.282.
67. Not all would have been taken from among refugee populations, nor were all
captives
commoners. There are several examples of educated or wealthy Senegambian Muslims
captured and sold into slavery in the eighteenth century. The earliest known, Ayub b.
Sulyman (Job Ben Solomon), a native of Bundu, was kidnapped in 1731 by Mandingos
dunng a voyage to the coast to sell slaves of his own. He was brought to Maryland though
he eventually returned to Africa several years later. Allan Austin, African Muslims in
Antebellum America: A Sourcebook (New York, 1984), ch. 2.
68. Ibid., pp. 278-82. The region referred to as the interior has a vague meaning in the eighteenth

century. Here the expression refers to the area around the Niger bend approximately
1000-1500 kilometres from modem Dakar.
69. Searing, West African Slavery' and Commerce, p.46.
70. Djolaof may be a corruption of Djola or Wolof both of whom came from the Senegambia,
Djola from the southern or Casamance region and Wolof from northern Senegal (Greenberg,
Languages of Africa).
71. Du Pratz, Histoire, III, p.344.
72. See, for example, Ayub b. Sulyman and other Muslims in Austin, African Muslims, p.80.
73. Du Pratz’s characterization of Senegals is echoed in Saint-Mery’s account published forty
years later in which he descnbes Senegalais as having ‘des marques d’une espece de
superiorite’; ‘intelligent, bon, fidele, ... reconnaissant, excellent domestique’; ‘...tres-sobre,
tres propre a la garde des animaux, discret & sur-tout silencieux’. Saint-Mery also described
the Yoloffes as ‘voisins des Senegalais’, and ‘leur couleur noire, est plus foncee que celle du
Senegalais’; Saint-Mery, Description de la Partie Fran^aise de Sainte-Dorningue, p.27.
74 Du Pratz, Histoire, III, pp. 344-5.
75. LHC, RSC 1751041401; 1751041601; 1751042301; 1751050101
76. The Male slaves who rebelled in Brazil in 1835 carried amulets to protect themselves from
injury. Amulets and talismans were quite popular in both Africa and among the black
BAMBARA SLAVES IN COLONIAL LOUISIANA, 1718-60 121

Brazilian population in the first third of the nineteenth century; Reis, Slave Rebellion in
Brazil, p.98.
77. Perhaps the best example of this effort to disassociate Africans from their past is to be found
in the naming processes practised by many New World owners. See Ira Berlin, ‘From Creole
to African: Atlantic Creoles and the Ongins of Afncan-American Society in Mainland North
America’, William and Mary Quarterly, 3rd Series, 53 (1996) no.2, pp. 251-88; Allan
Kulikoff, Tobacco and Slaves: The Development of Southern Cultures in the Chesapeake,
1680-1800 (Chapel Hill, North Carolina, 1986); Thornton, ‘Central African Names and
African American Naming Patterns pp.727—42; John C. Inscoe, ‘Carolina Slave Names; An
,

Index to Acculturation’, Journal of Southern History, 49 (1983), pp.527-54; Cheryll Ann

Cody, There was no “Absolom” on the Ball Plantations; Slave-naming Practices in the
South Carolina Low Country, 1720-1865’, American Historical Review, 92 (1987)
pp.563-96.
 The Cultural Implications of the Atlantic
Slave Trade: African Regional Origins,
American Destinations and New World
Developments

PHILIP D. MORGAN

In the early modern era, an increasingly integrated and cohesive Atlantic

world began to emerge. The Atlantic was the first ocean in the history of the
world to be regularly crossed, and the lands that bordered it came to have a
common history. Over time, a variety of links, bonds and connections drew
the territories around the Atlantic - that vast ‘inland sea’ - more closely
together. People,goods and ideas circulated in ever wider and deeper flows
between the pan-Atlantic continents. Changes in one corner of the Atlantic
world had repercussions in others; even seemingly local and provincial
developments invariably had Atlantic dimensions. Diverse and
heterogeneous, this Atlantic world became one - a unitary whole, a single
system.'
Slavery was a central feature of this emergent Atlantic system. It was the
cornerstone of a vast Atlantic labour market which, though inelastic and
inefficient, nevertheless functioned as one. Of course, the institution of
slavery varied enormously from one locale to another, but it was no curious
abnormality, no aberration, no marginal feature of this world. Few Atlantic
peoples before the late eighteenth found servile labour
century
embarrassing or evil; rather, slavery was fundamental and acceptable,
bearing an ancient pedigree to be sure, but readily adaptable to a variety of
needs and circumstances. Prior to 1820 two to three times as many Africans
as Europeans crossed the Atlantic to the New World. Much of the wealth of
the Atlantic economy derived from slave-produced commodities in what
was the world’s first system of multinational production for a mass market.
Slavery defined the structure of many Atlantic societies, underpinning not
just their economies but their social, political, cultural and ideological
systems. If slavery then must be situated squarely at the centre of the
 THE CULTURAL IMPLICATIONS OF THE ATLANTIC
SLAVE TRADE 123
Atlantic world, must be considered as a single sphere of inquiry,
it also
encompassing Europe, Africa and the Americas. Slavery
must be viewed in
its full Atlantic context."
At the heart of Atlantic slavery was the slave trade,
a vast co-ordinated
system for the forced migration of Africans often
from hundreds of miles in
theirhomeland interiors to virtually every corner of the Americas. Both
Europeans and Africans participated in the trade, and
four continents were
deeply influenced by it. The best studies of the
trade, beginning with Philip
D. Curtin s seminal The Atlantic Slave Trade, have
sought to explore ‘whole
institutions and whole processes, seen in
the large and separate from the
mere national subdivisions’. Since, as Curtin put it, ‘the
institutions of the
slave trade were common to the Atlantic
community’, an Atlantic
perspective is way to understand fully what was the largest inter-
the only
continental migration then known to the world. Curtin’s book is
still the best
place to start for an understanding of the Atlantic
slave trade. Indeed, in
many ways, it has still not been superseded.’
The book perhaps has done most to build on Curtin’s insights
that
is
John Thornton’s lavishly praised Africa and Africans
in the Making of the
Atlantic World. Thornton describes his work as
an attempt to assess the
‘migration of Africans to the Americas and to place
this assessment in the
growing field of Atlantic history’. He argues, among
other things, that
randomization was not a function of the middle passage;
rather, slave ships
drew their entire
cargo from only one or two African ports, and
their
catchment areas were homogeneous. Thus, ‘an entire ship
might be filled,
not just with people possessing the same culture,
but with people who grew
up together Once in the Americas, most slaves ‘on any
.
sizeable estate
were probably from only a few national groupings’. Therefore,
Thornton
continues, ‘most slaves would have no shortage of
people from their own
nation with whom to communicate’. In Thornton’s view,
particular African
national groups tended to dominate particular slave
societies in the
Americas, Africans in the New World often shared common
languages and
cultures that helped them survive in a hostile setting.
In most parts of the
Americas, it is now contended, slaves perceived themselves
as part of
communities had distinct ethnic or national roots.-*
that
Thornton’s book ostensibly ends in 1680, but he and others
are willing
to argue that ethnicity or nationality was central to slave
life beyond the
seventeenth century. In a general text designed by Thornton
and others for
the college student and informed reader, the concept of
nation as an ethno-
linguistic entity serves as the key social force driving
the development of
slave life well beyond 1680. One
or two African nations in most New World
settings, it is argued, dominated most slave societies.
Gwendolyn Hall
credits transplanted Bambara^as the central players in
Afro-American
 ^

124 ROUTES TO SLAVERY

culture in Louisiana. ‘The Louisiana experience,’ she observes, ‘calls into

question the common assumption that African slaves could not regroup
themselves language and social communities derived from the sending
in

cultures.’ Mervyn Alleyne believes that ‘one African ethnic group (the Twi)
provided political and cultural leadership’ among Jamaican slaves; he also
thinks that ‘entire functioning languages’ and ‘entire religions’, not just
general cultural orientations or religious beliefs, were carried to Jamaica.
Michael Mullin has argued which he sees as a euphemism
that ‘ethnicity’,

for tribalism, was particularly important among Anglo-American slaves,

especially in the West Indies. Thus, for Mullin, Coromantee was ‘the most
conspicuous and important nationality in Anglo- America’. In short, an
orthodoxy seems to have emerged that sees slaves as forming identifiable
communities based on their ethnic or national pasts.

This essay will explore emerging paradigm in two ways. First, it will
this

examine evidence from the latest and most comprehensive analyses of the
Atlantic slave trade, especially in so far as these bear on the question ot
African ethnicity and nationality. Second, it will explore three key issues
raised by the slave trade material and Thornton’s (and others’) arguments.
Throughout, this study will aim for the widest angle of vision, the broadest
transoceanic framework, seeking to see the Atlantic as a single, complex
unit of analysis, and trying to break out of the national boundaries
traditionally set for the study of slavery, whether African or American.
Exciting new material is beginning to emerge from an extraordinarily
important project sponsored by the W.E.B. Du Bois Institute for
Afro-American Research at Harvard University. With David Eltis, David
Richardson and Stephen D. Behrendt at the helm, this project is compiling
information on all known individual voyages drawn from the records of all
the major European and American slaving powers. To date, records exist on
almost 27,000 voyages, extending from the late sixteenth to the late

nineteenth centuries. When complete, the project will have information on

well over half of all the ships that made a transatlantic slave voyage. This
project has already compiled the largest data set for the study of the
long-distance movement of peoples before the twentieth century. As a result
of this project, and the work on which it builds, we now know more about
the forced migration of Africans than the voluntary migration of Europeans
in the early modem era.'’

The brief of this study is limited: to think about the cultural implications
of the project’s preliminary findings. Clearly, these thoughts are provisional,
because the analyses are available in largely aggregate terms. More refined
analyses of smaller temporal and spatial units will add much to the general

picture and lead to much more sophisticated conclusions than are presently
possible. As much as this author recognizes the necessity of Atlantic history.
 THE CULTURAL IMPLICATIONS OF THE ATLANTIC SLAVE
TRADE 125
his expertise is confined to theBritish-American world, followed by some
passing acquaintance with the French and Dutch
sectors, an even less
nodding awareness of the Iberian zones, and least
understanding of the
African dimensions of the story.
As David Eltisand David Richardson have argued, the key findings of
their consolidated and comprehensive set of data concern
neither Africa nor
the Americas treated alone, but
rather the connections between the
continents. In short, Eltis and Richardson and Behrendt
are engaged in true
Atlantic history. They are able to view the
intercontinental flow of people
from both sending and receiving poles. Erom the vantage
point of Africa, it
is nowpossible to look outward from each coastal region
and trace where
the forced migrants went. Most African regions
funnelled a majority of their
forced emigrants to one region in the Americas.
Thus, three-quarters of
those leaving South-East Africa went to South-Central
Brazil; two of three
Africans from the Bight of Biafra left for the British
Caribbean; 60 per cent
of the Bight of Benin’s emigrants went to Bahia; a half
of those leaving
Senegambia went to the Erench Caribbean; a half of West-Central
Africa’s
emigrants went to South Brazil; and a half of the
Gold Coast’s and
Windward Coast’s emigrants went to the British Caribbean. To be
sure, all
the regions of Africa sent slaves to almost all the regions of the Americas,
but people tended to flow in one dominant channel. In
some cases, there was
a subsidiary stream, thus a quarter of the Gold
Coast’s slaves went to
Surinam and the Guyanas; a quarter of the Windward Coast’s slaves
went to
St. Domingue; and a fifth of West-Central Africa’s slaves went to the French
Caribbean. Nevertheless, the regional African perspective on
slave
destinations reveals a distinct geographic concentration, or
in a few cases
two concentrations, in where the slaves went.
Equally striking patterns emerge when the transatlantic links are
examined from the more usual perspective of the American regions of
disembarkation. What stands out — and these observations are only a
variation on theemphases of Eltis and Richardson — are two extremes. First
the two main regions of Brazil - Bahia and the South-Central
area - drew
heavily on a single region of Africa. In Bahia’s case
about nine in ten
Africans came from the Bight of Benin; in South-Central Brazil
about eight
in ten came trom West-Central Africa. Second, at the
other extreme, true for
much of the Caribbean and North America, is the absence of a
dominant
single African provenance zone. No region of Africa, for example, supplied
more than about 30 per cent of arrivals to either Cuba, Barbados,
Martinique, Guadeloupe, or the Danish islands. Between the two
extremes
were some major destinations that received about half of their arrivals
from
a particular African coastal region: St. Domingue from
West-Central Africa,
the British Leeward Islands from the Bight of Biafra, and the Guyanas and
126 ROUTES TO SLAVERY
-
Surinam from the Gold Coast. In each of these American destinations
from St. Domingue to Surinam — the other halt ot their African influx came
from a number of regions. Brazil, then, was exceptional in drawing slaves
heavily from one region, while most other parts of the Americas drew on a
wider mix of African peoples, even if in some cases about halt of slaves
came from one region.^
These broad summaries of aggregate patterns disguise marked shifts
over time. David Eltis has provided a detailed chronological analysis of the
British trade before 1714. A close, decade-by-decade examination of this
trade reveals that the leading African provenance zones that supplied
Africans to Barbados and Jamaica were constantly changing. In the 1660s
the Bight of Biafra was the leading supplier; in the 1670s the Gold Coast,
Bight of Benin and Bight of Biafra were roughly equal providers; in the
1680s the Bight of Benin was the leader; in the 1690s again the Bight of
Benin dominated the trade, but with strong infusions from West-Central
Africa into Jamaica and from the Gold Coast into Barbados; in the 1700s
and early 1710s, the Bight of Benin and increasingly the Gold Coast were
pre-eminent. David Richardson has explored the aggregate
eighteenth-century British trade. He reveals, for example, that from the
1710s through 1730s British shipments of slaves from Senegambia reached
an all-time high; during the 1760s and 1770s about a third of British
Africans came from the Windward Coast; and from the 1780s through 1807,
the Bight of Biafraand West-Central Africa accounted for about 70 per cent
of British slave exports. A dynamic, diasporic approach indicates how
slaves came from a changing series of African coastal regions. The
aggregate picture masks a fluid, evanescent reality.*
The age and sex structure of a migration, just as much as its size and
regional origin, also had a differential impact on both sending and receiving
societies. While long-distance migrations are typically dominated by young
men, variations occurred. One significant finding is that the slave migration,

stereotypically portrayed as heavily male, was not in fact so, when put in its

full Atlantic context. Compared to the trade in indentured servants, the slave

trade comprised a remarkably large number of women and children. As

Eltis and Engerman note, ‘a higher proportion of children left Africa than
left Europe’. Indeed, overall, women and children outnumbered men in the
slave trade. The sex and age ratios of the Atlantic slave trade were most
comparable to free, not contractual, migrant flows.
Furthermore, the proportion of women varied quite markedly, both
between African regions of embarkation and American regions of arrival
and over time. Looked at from the perspective of African regions, the ratio
of male to female slaves varied from about 75:25 in upper Guinea to about
55:45 at the Bight of Biafra. From the perspective of American regions.
 THE CULTURAL IMPLICATIONS OF THE ATLANTIC SLAVE TRADE 127

much larger shares of women were carried to the British areas than
elsewhere, with women and men arriving in almost equal numbers in early
Barbados and Jamaica, whereas between two and three times
more men
than women arrived in Cuba and in Brazil. Over time, the
share of women
among African arrivals uniformly across African regions. From one
fell

region to the next, the proportion of women dropped by well

over 50 per
cent from the seventeenth to the nineteenth centuries. After
1810 women
constituted a quite small proportion of Africans from every
coastal region.
Lastly, the proportion of children also fluctuated widely.
West-Central
Africa, and to a lesser extent Upper Guinea and the Bight of
Benin exported
greater shares of children than other African
regions. Even more
dramatically, the proportion of children entering the transatlantic
trade more
than tripled from the seventeenth to the nineteenth centuries.
After 1810
over a half of those leaving South-East Africa and West-Central Africa,
and
just under a half of those from Upper Guinea, were children. The rise in the
number of children carried in the Atlantic slave trade occurred in all
regions, but was most pronounced most northerly and southerly
at the
extremes of the African coast. From the perspective of receiving regions,
South-Central Brazil imported the most children, accounting for a half of
all
African arrivals, Cuba received 38 per cent, Bahia
35 per cent, and the
French Caribbean and North America about 25 per cent. At the other
extreme, only about 10 per cent of the slaves who arrived in the British
Caribbean were children.^
Aggregate, sequential and structural analyses therefore emphasize the
complexity of the slave trade. As a way of summarizing their data, Eltis and
Richardson single out the regional composition of an African migration, its
duration, and its demographic character, which leads them to posit their
own
dual pattern. Cuba represents one extreme. Its newcomers were drawn from
a wide array of African coastal regions; its African influx was fairly

short-lived, lasting about eighty years (from 1790

and the African
to 1867);
arrivals comprised many children and few women. For all these reasons -
the relative absence of a shared background, the short span of the
slave-
trading connection, and the youthful and predominantly male character
of
its immigrants, which militated against family life and
the transmission of
culture - ‘Cuban African population had the potential for the greatest
the
loss of culture and language specific to particular African regions’.
At the
other extreme, Eltis and Richardson argue, was the British Caribbean,
where most immigrants came from Just two coastal regions (three out of five
newcomers came from the Bight of Biafra and the Gold Coast), where
African immigration lasted about twice as long as Cuba’s, and where
women were twice as numerous as Cuba’s arrivals. Eltis and Richardson
therefore conclude that ‘the opportunities for family fonnation and the
128 ROUTES TO SLAVERY

perpetuation of language and culture were likely stronger [in Barbados

specifically, and the British Caribbean more generally] than anywhere else
in the Americas’.'”
In several ways the preliminary findings of the Atlantic slave trade
project seem to lend support to the emerging paradigm propounded by
Thornton and others. Eltis and Richardson emphasize that ‘the distribution

of Africans in the New World was no more randomized than was its

European counterpart’. With the exception of Bahia and Minas Gerais, they
conclude that ‘the African part in the re-peopling of South and Central
America was as dominated by West-Central Africa, as was its European
counterpart by Iberians’. In the Caribbean, they continue, ‘West Africa was
as dominant as was West-Central Africa in South America’. Even where the
mixture of African peoples was greatest, African regions tended to supply
slaves in sequence, therefore minimizing the mixture at any one time. In
short, they conclude, ‘the picture of a confusing mix of African cultures
with all the attendant barriers to establishing African carryovers to the New
World needs revising’. Revisionism then is widespread. My question is

simple: is it justified? Was the slave trade markedly less random than we
once thought?"
To answer these questions, three central issues must be explored. First,
how homogenous or heterogenous was the Atlantic slave trade seen from
the vantage points of African coastal regions and American destinations?
Second, is it best to focus attention on ports, seeing Atlantic slaving vessels
largely visiting one or at most two African ports and then delivering their
forced migrants to one American port ? Finally, and most importantly, what
do we mean by ethnic and national identity in the early modern era and what
implications has this for New World cultural development? These questions
will be addressed from an Atlantic perspective, thinking not just of Africa
or America separately; but rather viewing them as linked or interconnected
continents. The Atlantic was a bridge as well as a barrier; the lands ringing
this ocean were joined as well as sundered by the sea.

The Europeans on the African coast were engaged in a highly

competitive trade. To be sure, certain parts of the coast tended to be
dominated by one power, as and Richardson and others have pointed
Eltis
out. In the eighteenth century the British dominated slave exports from the
Bight of Biafra; the French held the upper hand in trade with Senegambia;
and the Portuguese controlled shipments from most of the region south of
Zaire. But domination was never absolute and was always under challenge.
Thus, although the Portuguese based in Brazil certainly dominated trade
with the Bight of Benin - just over half of all slave voyages that arrived in
Benin set out from Bahia - yet French ships accounted for a fifth of all
slavers, and the Dutch, English and later Spanish were at various times
 THE CULTURAL IMPLICATIONS OF THE ATLANTIC SLAVE TRADE
129

significant players on the Slave Coast’. As much as one might say that
Bahia dominated the Bight of Benin trade, nevertheless almost a half of the
slavers arriving there were non-Bahian. Furthermore European slavers
always encountered shifting fortunes along the coast. In the early eighteenth
century the British acquired about a sixth of their Africans
from
Senegambia but by the 1750s less than half that proportion; in the
1720s the
Dutch West India Company dramatically turned from the Bight of
Benin to
the Gold Coast for its main source of slaves, and, with
the era of free trade,
the Dutch again shifted their centre of gravity westward
to the Windward
Coast, as the Dutch moved westward, the French moved eastward, and by
the middle of the eighteenth century they had relocated
their slave-trading
energies from the Bight of Benin to Central Africa.'-
Even within a single African
marked shifts often
coastal region,
occurred in the peoples forcibly expelled. The complex
competition for
trade was as much among Africans as among Europeans.
The supply of
slaves to the Bight ot Benin, for example, changed
drastically from the
eighteenth to the nineteenth centuries. Down to the late eighteenth century,
when the Oyo were a principal supplier of slaves to the Slave Coast, peoples
from the north and west of Oyo - Nupe, Borgu, Hausa, and
various
Ewe-speaking peoples — were readily available. In the nineteenth century,
after the collapse of the Oyo
empire, Yoruba-speaking peoples dominated
the flow leaving Bight of Benin ports, while the emergence
of the Sokoto
Caliphate in the Central Sudan generated a growing secondary
stream of
Hausa slaves. By the early nineteenth century ‘there were at least two
demographically distinct components of the trade at the Bight of Benin’,
notes Paul Lovejoy, ‘one that brought males from the distant interior
to the
coast and another that siphoned off slaves (men, women and
children) from
the coast itself . Thus it is somewhat misleading to say that Bahia received
virtually all its slaves from the Bight of Benin, by
if this is meant to imply
some uniformity over time. The ethnicity of those leaving the Slave Coast
and arriving in Bahia changed drastically over time.'^
The relationship of coastal ports to hinterlands grew more
complicated
over time, which again enhanced the increasing diversity of peoples
shipped
across the Over time,
Atlantic. example, the region known as
for
West-Central Africa came to cover a wider range of coastline and drew on
an increasingly expanding hinterland, extending hundreds of
kilometres
from the coast. At least three, sometimes four, distinct commercial networks
drew slaves from the interior toward the Atlantic shores. The mix of peoples
flowing from that region grew more, not less, heterogeneous. Different
ports within a single coastal region might draw upon different and
fluctuating streams of peoples. Thus, along the early nineteenth-century
Bight of Biafra coast, Igbo-speakers dominated slaves shipped from Bonny,
130 ROUTES TO SLAVERY

Ibibio-speakers comprised 40 per cent of slaves shipped trom Old Calabar,

and slaves from much further inland - Nupe, Kakanda, and Hausa, tor
example - formed between 5 and 25 per cent of the exported slaves from
various Biafran ports.
If the lens focuses primarily on the American side ot the Atlantic, the
emphasis again ought to be on heterogeneity. This is so because, as Eltis and
Richardson have noted, the ‘geographic concentration of arrivals in the
Americas is much less than that of African departures’. Even if attention is

directed to the places that received most Africans band stretching trom- a
about Cuba in the north to Central Brazil in the south - such American
destinations encompass a much wider span of landscapes, climates, and
environments than the coastline ot West and West-Central Africa. It
attention broadens to all the places in the Americas that received Africans
-
from New York in the north to Buenos Aires in the south - then the
geographical diversity is staggering. New World slavery knew no limits; it
penetrated every economic activity, every type of settlement, every setting.
American slaves lived in temperate highlands as well as in tropical

lowlands, on large continental plains and on small mountainous islands, on

farms as well as on plantations, in cities as well as in the countryside; they
worked in fields and in shops, in manual and skilled occupations, in civilian
and in military life, up trees and down mines, on land and at sea."’
Even when the lens zooms to a single region or city, such as Rio de
Janeiro centred in South-Central Brazil, which, as noted, drew heavily on
West-Central Africa for its slaves, heterogeneity still seems the most
accurate description of its African residents. This may seem a ridiculous
proposition, for on the face of it Rio seems the perfect place to find
homogeneity. At least two-thirds of Africans living in nineteenth-century
Rio traced their homelands to West-Central Africa. As Mary Karasch puts
it, the ‘Central Africanness’ of the city’s slaves is fundamental to an
understanding of their culture. Yet Karasch stresses Rio’s ‘extraordinary
ethnic diversity’. Slaves from West-Central Africa were from three distinct
sub-regions. The first, Congo North, supplied ‘thousands of ethnic groups’
to Rio, and in certain decades - the 1830s and 1840s, in particular - the

ethnic mix from Cabinda, the central port of Congo North, was especially
notable. Second, although slaves from Angola came from a more restricted
area than the Congos, they still comprised numerous ethnic groups and at

least two major Kimbundu-speaking populations of

linguistic groups: the

Luanda and its hinterland; and the Lunda-Tchokwe of eastern Angola.

Finally, the third important port and sub-region was Benguela in southern
Angola, whence came Ovimbundu and Ngangela peoples among others. So,
the West-Central origins of most Rio slaves was in fact a congeries of
peoples, languages and cultures. Further, important as Central Africa was to
 THE CULTURAL IMPLICATIONS OF THE ATLANTIC SLAVE TRADE 131

Rio s slaves, as many

as a fifth to a quarter of the city’s Africans could trace
their heritage to East Atrica. The so-called Mocambique
nation became one
of the largest in the city. Finally, although West Africa was the least
important source of Rio s slaves, so-called Minas and Calabars were
prominent in the city. Rio, then, was truly a babel of African peoples.'*
Just as thenew paradigm tightly links one African region to a single
American society, so the new orthodoxy sees slave ships connecting one
African port to a single American destination. In some ways,
this is a
valuable perspective. One of the most exciting features of the
Du Bois
Institute project is its capacity to follow a ship (and
perhaps, in the future,
individuals) from a port in Africa to a port in the Americas.
Preliminary
analysis from the project reveals the importance of a small
number of
African ports. Two-thirds of African slaves shipped from known
points,
rather than from broad coastal regions, left from just twenty
ports. Ranking
the ports that shipped the most slaves, three of the top five -
Cabinda,
Benguela and Luanda - were in West-Central Africa. About two-thirds of
departures from the Bight of Benin were from Whydah. Indeed, as Eltis
and
Richardson note, ‘probably well over one million slaves left from Whydah,
making it most important oceanic outlet for slaves in sub-Saharan
the single
Africa’. Similarly, almost 80 per cent of all slaves in the Bight
of Biafra
region left from just two outlets. Bonny and Calabar. The concentration
of
African departures from just a few sites is notable.'^
Nevertheless, some caution is in order. One cannot accept uncritically
the stated African destinations of ships clearing European ports. Further,
ships designated as having boarded their slaves at a particular port did
not
always obtain all their slaves even from the coastal region of that port. The
African point of embarkation may have been simply the last port of call. In
1744 a Dutchman at Elmina on the Gold Coast reported that most of his
countrymen purchased their slaves on the Windward Coast, sailing on to
Elmina only when their slave cargo was still deficient. From detailed
records of over fifty free-trade Dutch ships from the 1740s through
the
1780s, about three-quarters of all slaves were acquired before the ship
reached the Gold Coast. Even on the Gold Coast itself, the Dutch drew
fairly diversely. Elmina was their major port of call, but slaves from various
Dutch trading stations were often taken by small boats to Elmina. One
statistical record from the mid- 1720s to mid- 1750s indicates that
Elmina
provided only one-third of the Gold Coast slaves; ports from Axim in the
west to Accra in the east supplied the other two-thirds. Similarly, in the sub-
region of West-Central Africa known as Congo North, slave traders
sometimes picked up slaves along the entire coast even if they then called
them Cabindas after the central port of the region.'"
These cautionary notes should not be exaggerated. It was unusual for a
132 ROUTES TO SLAVERY

ship to purchase slaves in more than one coastal region. The major
exceptions were ships that bought in the Gold Coast and Bight ot Benin
chietly before 1720, and ships that touched at various points ol the
Windward Coast on way
Gold and Slave Coasts. Probably no
their to the

more than a fifth of the Gold Coast ships went on to the Bight of Benin,
even before 1720; and Gold and Slave Coast ships normally (with perhaps
the exception of the Dutch) bought only a few slaves before reaching their
major markets. Coastal trading was common, so slaves would come from a
range of fairly proximate places rather than from a single port. Nevertheless,
there were exceptions. For example, Whydah seemed to be able to supply
all its prospective purchasers, without ships having to trade elsewhere, and
the same appears to have been the case at various ports on the Gambia.'^
In another sense, however, whether a ship landed at one or two or more
African ports is somewhat beside the point; rather, the real issue is the
complexity of networks deep within Africa that funnelled slaves into nodal
points on the coast. Joseph Miller has written a brilliant study of just such
networks for West-Central Africa. As he points out, the whole region
consisted of over 1,000 kilometres of coastline and a slaving hinterland that
by the early nineteenth century extended 2,500 kilometres inland. Overall,
slaves were drawn from locales within a region of 2.5 million square
kilometres, an area larger than the United States east of the Mississippi
River. In most of the central market places in the interior of West-Central
Africa, traders dispatched slaves in sizeable caravans that marched at best

150 kilometres a month. As the moving frontier zone of slaving violence

advanced eastwards, a complex fan of trade routes, ever more extensive and
convoluted, radiated out into the interior. The sequence of multiple sales
that attended transfers of slaves between their place of seizure and the coast
could divert the tlow in almost any direction. As slaves plodded westward,
many died and others were added, so that by the time they reached the coast
the caravans were indeed a motley crew. The process was, in Miller’s
words, an ‘agonizing progress toward the coast that lasted months if not
The actual port of embarkation was therefore just one link in a highly
years’.
complex chain. No other slaving hinterland was large as West-Central
Africa’s, but then again no other coastal region supplied as many slaves to
New World."’
The Du Bois Institute project is able to rank not just African ports of
embarkation but American ports of disembarkation; once again, the
concentration is notable. Eltis and Richardson rank the top seventeen
American ports, descending to St. George’s in Grenada, the lowest ranked
on their list with a known importation of less than 50,000 slaves. The top
two ports are Brazilian - Rio de Janeiro on Guanabara Bay which received
over 800,000 slaves, and Bahia (presumably primarily Salvador) which
 THE CULTURAL IMPLICATIONS OF THE ATLANTIC SLAVE TRADE 133

received almost 700,000. Directly below these two are

many Caribbean
ports, the most important of which in descending
rank order are Cap
Francais in St. Domingue with 350,000 known arrivals,
Bridgetown in
Barbados with 340,000, Havana in Cuba with 210,000, and Kingston in
Jamaica with 190,000.*'
Most slave ships probably had a single destination, but whether most
slaves were sold and remained in the vicinity of their
point of
disembarkation seems somewhat more problematic. The whole question
of
Africans subsequent movements within the Americas is complicated.
Much more is known about the first place of landing than the Africans’ final
destinations. Just as the African port of embarkation is easier
to document
than the complex chain that led from interior to coast, so it is infinitely
easier to record slaves landing at an American port than it is to trace their
later movements.
At certain times, a lively re-export trade in slaves unquestionably arose
in various American ports. The sometimes landed slaves in their
British
Caribbean islands before taking them to the North American mainland.
After 1763, slaves were commonly re-exported from most British islands
in
the eastern Caribbean. Equally well known is that Jamaica
was a major
re-exporter of slaves. Over the course of the eighteenth and early nineteenth
century, Jamaica imported over 800,000 Africans and transshipped
about
200,000. A
British slave trade of sorts persisted for a quarter of a century
after abolition when over 20,000 thousand slaves were shipped from the
older islands to the newer colonies, especially Trinidad and Demerara.
Smaller European slave-trading nations engaged in much re-exporting.
Many slaves imported by theDanes were transshipped. The Dutch were
well known for landing slaves at islands like Curasao and afterwards
reshipping them to the Spanish-American mainland ports. Some slaves
were transshipped from Curasao to other Dutch colonies in the Caribbean
or Guiana. War or market conditions sometimes forced a ship to alter
its
course and make more than one landing. With the growth of the Dutch free
trade, Johannes Postma notes, the restless search for the most
profitable
markets often led to more than one landing. Before the direct trade to Cuba
developed in the nineteenth century, Cuban and other Spanish Caribbean
planters regularly purchased slaves in the well-established markets of
Jamaica and Dominica. A robust inter-island and island-mainland trade
existed in slaves as in much else.-*

But the forced migration of Africans in the Americas was not just
confined to transshipment; far more consequential were the long marches
on American soil, sometimes in stages, far into the interior. In many ways,
then, America may be conceived as a mirror image of Africa: ports on
either
side of the Atlantic were funnels for large slaving hinterlands that fanned
134 ROUTES TO SLAVERY

out across the land. As Miller points out, ‘Even the ships headed to a single
Brazilian captaincy must, finally, be understood as moving through no more
than an intermediate stage in a complex redistribution to turther
destinations.’ Rio received more slaves than any other New World port
because it supplied Minas Gerais, was a route of access to Sao Paulo, and
constituted a smuggling station on the way to the estuary of the Plate.
Salvador was the second ranked largest port for slaves because many of
those Africans continued south to Minas Gerais and north to Pernambuco
and the Amazon. Slaves who had already been transshipped from various
Caribbean islands to Cartagena and Portobello then faced further journeys
to Colombia or Peru. Slaves landed in the Rio de la Plata went overland to
Tucuman and then onto the silver mines at Potosi. Africans who arrived in
Virginia after the early eighteenth century or South Carolina after the late
eighteenth century were usually destined for the piedmont or even further
inland. The American trek from port to places of residence was often as long
and agonizing as the African march from point of capture to the coast.-^
Movement both within Africa and the Americas complicates not Just the
notion of port to port correspondences but also the conception of
homogenous peoples being swept up on one side of the ocean and set down
en masse on the other. Because many African slaves came in tortuous and
convoluted ways from the interior to the coast, whatever ethnic identity they
originally had was undoubtedly in flux. Furthermore, it is often impossible
from a late twentieth-century vantage point to reconstruct what, if anything,
might have been. Miller’s description of the functioning
that ethnic identity
of the Angolan slave trade - ‘individuals being kidnapped, sold, resold, and
captured again in the course of repeatedly disrupted lifetimes’ - leads him
to conclude that their so-called ethnic origins probably meant ‘very little’.

In addition, when Philip Curtin assembled a number of different samples of

contemporaneous on
opinionthe ethnic distribution of the
eighteenth-century Senegambian slave trade, what is most impressive is that
three-quarters of the exported slaves occupy the ‘non-ethnic’ category. Only
about 17 per cent of the slave exports from Senegambia were Wolof, 5 per
cent Fulbe, and 3 per cent Sereer. Most slaves seem to have come from east
of the heads of navigation - by way of Gajaaga and the Gambia - and
cannot be assigned to a specific ethnic group.-^

Even more fundamental, how aware were people of belonging to an

imputed ethnic and cultural tradition? Whether the search is for a
pan-African culture, broad regional cultures - as in Thornton’s tripartite
division of West and Central Africa into something akin to Caesar’s Gaul -
or more localized ethnic cultures, the same problem inheres: are we not in
danger of adopting the hermeneutics of the observer? Do we not fall into the
trap of denying the social and cultural worlds created by local actors, of
THE CULTURAL IMPLICATIONS OF THE ATLANTIC SLAVE TRADE 135

seeing similarities where the actors were aware only of difference?

Consider, for example, the eighteenth-century use of the term Yoruba. As
Robin Law has pointed out, originally the name designated only the Oyo,
being the term by which the Hausa of northern Nigeria referred to the Oyo
kingdom. Before the nineteenth century, he continues, ‘the name Yoruba
was not used group of which the Oyo form part ... It
to designate the larger
must, indeed, be doubtful whether the various “Yoruba” groups were
conscious ot forming, on linguistic or other grounds, any sort of unity or
community.’ Similarly, David Northrup, writing of pre-colonial
South-Eastern Nigeria, observes that ‘the largest unit of identity for most
inhabitants does not appear to have been the primary ethnic unit such as
Igbo or Ibibio, but rather the smaller dialect or cultural group’. Indeed,
Igbo-speakers enslaved in the early nineteenth century had apparently never
heard the name Igbo homelands. Or consider the farmers of the
in their

central highlands of West-Central Africa, many of whom were shipped to

Brazil; they became known as the Ovimbundu because they shared similar
linguistic traits, but, notes Miller, ‘none of them in the eighteenth century
would have claimed much unity’. Ethnicity, in so far as it existed, was
clearly very localized in precolonial Africa.’^
In fact, a distinct danger exists inapplying terms such as ethnic group
and nation indiscriminately in African and African-American studies.
Thornton adopts rather uncritically early modem
European usages by
talking of ‘countries’ or ‘nations’ and even of ‘national loyalty’, which are
not just imposed taxonomies but anachronistic ones. As Karen Fog Olwig
notes of Danish West Indian slaves, they ‘did not seem to identify strongly
with nations, so when asked the name of his nation, a slave often responded
“with the name of the place where he lived in Guinea’”. Similarly, while
ethnicity can be used to stand for some kind of group {ethnos), it is often a
residual term applied when too little is known about some group to be able

to label more precisely. The ethnic lexicon of New World planters and
it

slave traders - and they must be distinguished - is often mysterious. As

David Geggus has pointed out in the Francophone context, the labels ‘Mine’
and ‘Caramenty’ obviously derived from the ports of Elmina and
Kormantin ‘situated close together on the Gold Coast, but the sex ratios and
morbidity levels of these two sets of slaves suggest that they were drawn
from quite different, perhaps distinct’ locales. Many ethnic labels were
affixed inaccurately. Primarily on the basis of scattered references to large
numbers of Bambara in early Louisiana, Gwendolyn Hall argues that they
served as a charter group, but, as Philip Curtin had earlier noted, Bambara
was a catch-all term. Some early Louisiana slaves doubtless were Bambara,
an ethnic group, generally non-Muslim, who comprised the dominant
people of the new kingdoms of Segu and Kaarta in the eighteenth century.
136 ROUTES TO SLAVERY

But the word also meant, in Senegambian French, any slave soldier serving
in Senegal, and it could be taken as a very general designation for all

Malinke-speaking peoples, or even of all people from east of the Senegal

and Gambia rivers. Curtin authoritatively declares that. The “Bambara”
slaves shipped west as a result of eighteenth-century warfare or political
consolidation could be dissident people who were ethnically Bambara, or
they could just as well be non-Bambara victims of Bambara raiders.’ The
terni was more geographical than cultural.’^

In the New World, so-called African ethnic or national identities were

often convenient reconstitutions or inventions. Nor could these identities
easily remain pristine in the pluralistic Americas. Analysis of the national
origins and mating patterns of the Trinidadian slave population in the early
nineteenth century - one of the most African slave populations at that time
- reveals that the population of about 14,000 was drawn from an wide range
of territories, from Senegambia to Mozambique. Early
extending
nineteenth-century Trinidad was also unusually urban and had relatively
small slave-holdings. In this newly settled society, most African slaves -
and thus about half of all slaves - did not live in families. The fortunate
Africans who found partners generally found other Africans, but not often
from their own ethnic group or even region. Ethnic identity, Barry Higman
concluded, dissolved rapidly as a result of extensive inter-ethnic marriage.-^
and the Americas should be
Just as ethnic identities both within Africa
viewed as fluid and permeable, so cultural development in the New World
involved borrowing and adaptation, modification and invention. Slaves
functioned as bricoleiirs, to borrow Claude Levi-Strauss’s term, picking and
choosing from a variety of cultural strains to create something new. This
plasticitywas evident in all aspects of slave culture - from the way the
slaves wore their clothes, the way they combed their hair, to the way they
organized their yards.’**

The complexity of ethnic cultural development among New World

slaves is well captured in a transcription of three African songs - ostensibly
from ‘Angola’, ‘Papa’ and ‘Koromanti’ respectively - heard in Jamaica in
1688. In the 1680s Africans from West-Central Africa (including Angola)
and the Bight of Benin (where Popo referred to a people as well as to two
ports) comprised about three-quarters of allincoming slaves to Jamaica, so
it is not surprising that white visitors heard a Papa and an Angolan song. A
little more Koromanti song, because only 7 per cent of
surprising is the
known Africans imported in the 1680s were from the Gold Coast (where the
port of Kormantine was located). Koromantis are often associated with
early Jamaican slavery, but in the 1660s there were no Africans imported
from the Gold Coast and in the 1690s less than one in ten arrivals were from
that region. The only decade in the second half of the seventeenth century
THE CULTURAL IMPLICATIONS OF THE ATLANTIC SLAVE TRADE 137

when the Gold Coast was a major supplier

Jamaica was the 1670s: it
to
contributed a quarter of Jamaica’s Africans, but was easily dwarfed by the
60 per cent who came from the Bights of Benin and Biafra. Nevertheless,
so-called Koromantis apparently led two Jamaican slave rebellions in 1673
and 1686, which seems to indicate they exercised a power out of all
proportion to their numbers.
A
musical analysis of the three songs reveals that they were far from
ethnically distinctive. Rather, even at this early date in Jamaica’s history
when Africans were numerous and might be thought to retain a measure of
ethnic integrity, the songs showed influences from other African regions. In
particular, the ‘Angola’
song incorporated musical elements from the Akan
peoples of the Gold Coast, even though retaining many Bantu features.
Again, then, there seems evidence that Gold Coast peoples had undue
influence. But the ‘Koromanti’ song had little Western Kwa (or Gold Coast)
features but was rather ‘a loosely bundled set of associations centered on
West Africa’. In other words, perhaps the Koromanti, clearly a minority
among late seventeenth-century Jamaica’s were especially
slaves,
influential because of their adaptability, their pan-West African outlook. At
any rate, as early as the late
seventeenth century, syncretism and ‘a process
of interchange and experimentation’ had clearly occurred among African
musical cultures in Jamaica.-*^
Another conundrum concerning the syncretism of African traditions and
the relative contributions of various African peoples is evident in the early
formation of Haitian vodun or voodoo. The Dahomean or Aja-Fon influence
in Haitian voodoo is paramount: it contributed the major deities,

ceremonies, and most of the African vocabulary to this New World religion.
But from extant records, which begin in about the 1720s, the Aja-Fon seem
to have been a minority of Africans in St. Domingue, comprising at most
about a of the island’s black immigrants. Like the Koramanti in
fifth

Jamaica, what has to be explained is the undue influence of the Aja-Fon

minority in Domingue. One possible explanation is that perhaps before
St.

1720 the Aja-Fon were not a minority. Perhaps in the earliest years of St.
Domingue’s history they were numerous, and constituted a charter group,
creating many of the cultural norms for later newcomers. Or perhaps
Aja-Fon languages were particularly easy to learn or their pantheon of gods
were especially structured, which gave them influence disproportionate to
their numbers throughout the eighteenth century.
While the Dahomean influence in voodoo was dominant, other African
sources were important. Perhaps most notable were the Kongo strains in
voodoo, which is not surprising for at the height of St. Domingue’s power
Congos formed the largest single ethnic group in the colony’s slave
population. John Jenzen has noted the presence in Haitian voodoo of
138 ROUTES TO SLAVERY
various Kongo rituals, associated with Lemba, a cult of healing, trade, and
marriage relations in seventeenth-century West-Central Africa. Luc de
Huesch secs a syncretism of Dahomey’s rada and Kongo’s petro divinities
in vodiin. David Geggus has highlighted two religious chants dating from

the late eighteenthand early nineteenth centuries which were Kongo in

origin and which were sung during an initiation ceremony into a snake cult.
Quite why the Kongo intluence was secondary and not primary may be
explained variously: the low status of Congo slaves based perhaps on their
size, the ability of Congos to assimilate, the Congos’ relative lack of women
which may have hampered cultural transmission, and their prevalence in the
mountains rather than on the plains. Secondary or not, what is most notable
here, as in Jamaican music, is the apparent joining together of different
Atrican cultures in an early New World cultural form.’*^
An ‘Igboized’ slave Douglas Chambers argues, arose in
culture,
eighteenth-century Virginia. He bases his argument on the large number of
slaves drawn from the Bight of Biafra who entered the Chesapeake Bay in
the 1720s and 1730s. These ‘first comers’, he contends, shaped the material
culture of the region; the cultivation of okra, itself an Igbo word; the
reliance on sweet potatoes, which were the nearest equivalent to yams, the
staple of the Bight of Biafra region; the prevalence of root cellars in
Chesapeake quarters, where the slaves stored their sweet potatoes and other
goods. Similarly, the Igbo connection is said to explain much of the musical,
magical, and ceremonial culture of Virginia slaves: the incorporation of
Igbo instruments such as the gamby (Eboe drum) and the balafo
(xylophone); the practice of conjuring and root-doctoring, which allegedly
had precise Igboland analogues; and jonkonu, a masquerade involving
cow-horn and other animal masks, which can be likened to spirit cults and
secret societies in the Bight of Biafra region.^'
Although the Bight of Biafra region was an important source of slaves
for the Chesapeake region, it seems rather far-fetched to claim that Virginia
slave culture was predominantly Igbo. First, Africans from Upper Guinea
constituted 44 per cent of arrivals in Chesapeake between 1662 and
the
1713, outstripping those from the Bight of Biafra. Although the Bight of
Biafra was a considerable supplier in these early years - providing just over
a third ot the Chesapeake’s slaves - the numerically predominant ‘first
comers’ came from Senegambia and Sierra Leone. The Gold Coast was a
significant supplier too, particularly decade of eighteenth
in the first

century. The Bight of Biafra was the dominant provenance zone in the
1720s and 1730s, but this influx represented a second, not the first, wave.
Other influxes, from the 1740s to 1760s, were quite heterogeneous, with
Senegambia the dominant zone in the 1740s and Angola in the 1760s.
Second, many of the so-called ‘Igboisms’ in Chesapeake slave life could
 THE CULTURAL IMPLICATIONS OF THE ATLANTIC SLAVE TRADE 139
just as easily be explained as general West and Central African traits; and,
as already noted, Igbo identity is itself problematic. Rather than posit a
single ethnic influence, syncretism is the real story of Virginia slave culture.
Finally, the reason for the notably syncretic character
of Chesapeake black
life must be sought not just in the scale and timing of the immigration and
the enforced mingling of heterogeneous Africans, but in
the local context.
Africans in Virginia found themselves purchased in tiny lots, dispersed
onto
widely scattered estates, resident on small plantations, soon surrounded
by
a majority of native-born slaves, and brought into close contact
with whites.
All of this was not conducive to the reconstitution of an
African ethnic
identity.

The foundation of Louisiana’s Afro-Creole culture, Gwendolyn Hall

claims, rests on the numerical predominance of Senegambia
slaves
imported to the colony in the early decades of the eighteenth century.
Bambaras, she maintains, ‘played a preponderant role’ in the formation of
the colony s slave culture. They ‘constituted a language
community’,
mounted rebellions, were accused of a disproportionate number of crimes,
and influenced other slaves with their magical beliefs, evident in the
widespread resort an amulet, or grisgris, a harmful charm, both
to zinzin,
Mande terms. Aside from whether Bambaras were a true ethnic group, the
alleged paramount influence ol Mande on Louisiana culture needs to be
questioned. First, in the early 1720s, as Peter Caron has emphasized,
Africans from the Bight of Benin dominated the colony’s African
population. The contribution of enslaved Aja peoples from the Slave Coast,
Caron observes, may have been especially significant given the large
numbers of children born to Africans between 1721 and 1726. Further,
Caron demonstrates that most slaves from Senegambia came from the
coastal areas, not the Niger bend, which is the area of Bambara and Mande
influence. In addition, even by the middle of the eighteenth century most
Louisiana slaves lived on units of ten or fewer Africans, which inhibited the
domination of one ethnic group. For all these reasons and others, such as the
much more heterogeneous and larger influx of Africans when Louisiana
became a Spanish colony —
seems sensible to emphasize the pluralistic
it

quality of Afro-Creole culture. The Congo inlluence in folklore and magic,

the Fon role in voodoo, the Yoruba origins of shotgun houses, and the many
African religious traditions (from Islam to Congo-Christianity to
non-universal variants) that infused Louisiana religion must all be
recognized.^’
Cuban Santeria can serve as one last example of a New World
neo-African cultural form. Eltis and Richardson, it will be recalled, argued
that the heterogeneity of the island’s Africans, the lack of women and
abundance of children, and the short period of African importation should
140 ROUTES TO SLAVERY

have undermined the ability of Cuban slaves to maintain regional African

cultures. Cuban Santeria seems an unlikely model for such an argument,
because one of its major components was a traditional African religion, the

orisha worship as practised by the Yoruba. Eltis and Richardson

demonstrate that a quarter of Cuba’s Africans came from the Bight of
Benin, making it most important provenance zone for Cuba’s
the second
Africans (behind West-Central Africa, which supplied about 28 per cent of
the island’s Africans). Furthermore, at certain times, the Bight of Benin was
a dominant supplier; indeed, from 1851 to 1866, Cuba was the only market
available for Bight of Benin slaves. Thus, a powerful Yoruba influence in
Cuba should not be all that surprising, although perhaps the question should
again be, as in the Haitian case, why West-Central African influences were
not more dominant than they were.
But a number of other facilitating forces were also at work, which will
demonstrate that the process of New World slave cultural development was
far from straight-forward. First, there is the issue of ethno-genesis. As
already noted, in eighteenth-century Africa there was no overall term for all

the heterogeneous Yoruba subgroups, Cuba, descendants of

but in

Yoruba-speakers and some of their neighbours became Lucumi. ‘Just as

Apulians, Sicilians, and Calabians all became Italians in the United States,’
George Brandon analogizes, ‘Oyos, Fgbas, IJebus, and IJeshas all became
Lucumis in Cuba.’ Who was primarily responsible for generating this ethnic
label - the slaves, the slave traders, or most likely a combination of the two
- is unclear, but this emergent ethnic identity was a New World
development. Furthermore, what is evident is that while Lucumi culture had
a Yoruba focus, it incorporated traits from far afield. People sold by the
Yoruba became Lucumi; people of Allada and the Ibo were incorporated
within the so-called Lucumi nation. Some Lucumi words and phrases are
not Yoruba in origin, and seem to be Ewe or Fon in derivation. Once again,
a minority exercised dominion in excess of its numbers.
In the Cuban case, an important - perhaps the most important - reason
why Africans were able to retain so much of their regional cultures should
be sought in towns. Cuba was a quite urbanized society by New World
standards, and urban slaves experienced a less regimented existence than
their rural counterparts. They had greater latitude for cultural development.
Slaves and freed people created Afro-Cuban cabildos or lodges, clubs,
fraternities, and dance groups. They drank in taverns with their carabelas or

shipmates. By the end of the eighteenth century, twenty-one cabildos

existed in Havana alone. Each had its own ritual centre, its African
language, its distinctive drums and drumming styles and its dances. Just as
with Jamaican music, an exchange of musical and other cultural styles no
doubt occurred across cabildo and ethnic lines. These clubs were, in
 THE CULTURAL IMPLICATIONS OF THE ATLANTIC SLAVE TRADE
141

Brandon s words, important centers for the preservation of African

religion
in Cuba s cities One or more Lucumi cabildos provided the crucible in
.

which Santeria evolved.^'*

The towns in facilitating ethnic identity in the New World points
role of
to the importance of context in the formation
of slave culture in the
Americas. Perhaps neo-African cultural forms in the New World
appeared
most readily in urban settings, as the example of Cuban Santeria
(and
Bahian Candomble) suggests. Conversely, the most generic
African cultural
forms may have been most conspicuous in heavily rural places, such as
Virginia. Yet, important as the urban/rural contrast was in moulding ethnic
affiliations in the New World, it was from being the only contextual
far
variable in shaping black cultures in the Americas. The scale of immigration
from a particular African coastal region or regions was fundamental
to the
process ot New World ethno-genesis. Equally important was
timing:
first-comers in some situations, as in the rapid formation ol a new language
and religion in Surinam, were extremely influential;
in other settings,
late-comers, as in the Yoruba influx to Cuba, played a vital
role. The
demographic structure of the immigrant influx — the ratio of men to women,
adults to children - was an important key to cultural transmission. Whether
planters bought Africans singly or in large groups was another crucial factor
m shaping slave life. The size of plantations, the proportion of black to
white, and the nature of the economy were yet other features of a New
World context determined the character of a slave culture.'"
that
Finally, even after making allowances for all the demographic,
economic
and social variables that fostered or inhibited ethnic identity among slaves,
many unpredictabilities remain. In some cases, particular African minorities
were influential in shaping certain slave societies,
perhaps because these
African minorities were especially adaptable and could incorporate others,
or perhaps because some features of their homeland culture were
especially
attractive to others, or perhaps because of their high status among
Europeans. Such might explain the influence of ‘Koromantis’ in Jamaica.
Conversely, in other cases, particular African majorities were surprisingly
uninfluential, perhaps because they were too adaptable and were
therefore
readily assimilated by others, or perhaps because they lacked
enough
women to transmit effectively their culture, or perhaps because they were
located on the margins rather than in the heartlands of their new society, or
perhaps because they were viewed negatively by Europeans. Such might
explain the relative lack of influence of ‘Congos’ in St. Domingue.
Overall, Africans in the Americas had to adapt to survive. They had no
time for debates about cultural purity or precise roots; they had no necessary
continuing commitment to the societies from which they came. They were
denied much of their previous social and cultural heritages: the personnel
142 ROUTES TO SLAVERY

who maintained their homeland institutions, the complex social structures

of their ancestral societies, their kings and courts, their guilds and
cult-groups, their markets and armies. Even what they brought they
ruthlessly jettisoned because was no longer applicable or relevant to their
it

new situations. No wonder, as Mintz puts it, when we think of the history of
African-American slaves, ‘we are speaking of mangled pasts’. For that
reason, he continues, ‘It is not the precise historical origins of a word, a
phrase, a musical instrument or a rhythm that matters, so much as the
creative genius of the users, molding older cultural substances into new and
unfamiliar patterns, without regard to purity or pedigree’.’^
Whether the focus is on African regional origins, American destinations,
or New World cultural developments, the emphasis should be on
heterogeneity, on Ouid boundaries, on precarious and permeable zones of
interaction, on hybrid societies, on mosaics of borderlands where cultures
jostled and converged in combinations and permutations of dizzying
complexity. A key way in which the many and disparate parts of the Atlantic
world were coming together - albeit at unequal speeds - was in the creation

of ever more mixed, heterogeneous cultures. The homogenizing tendency of

stressing cultural unity in Africa, of emphasizing the non-random character
of the slave trade, and of seeing the dominance of particular African coastal
regions or ethnicities in most American settings, is at variance with the
central forces shaping the early modern Atlantic world. This tendency
should be resisted.

NOTES
1. Bernard Bailyn, ‘The Idea of Atlantic History’, Itinerario, 20, no.l (1996), pp.l9—T4; Daniel
W Howe, ‘American History in an Atlantic Context: An Inaugural Lecture delivered before
the University of Oxford on 3 June 1993’ (Oxford, 1993).
2. David Eltis, ‘Free and Coerced Transatlantic Migrations: Some Comparisons’, American
Historical Review, 88 (1983), pp. 25 1-80.
3. Philip D. Curtin, The Atlantic Slave Trade: A Census (Madison, Wisconsin, 1969), p.xvi.
4. John Thornton, Africa and Africans in the Making of the Atlantic World, 1400-1680 (New
York, 1992), pp.l ,
195-7.
5. Michael L. Conniff and Thomas J. Davis (eds.), with various contributing authors, Africans
in the Americas: A History of the Black Diaspora (New York, 1994), pp.53-8; Gwendolyn
Midlo Hall, Africans in Colonial Louisiana: the Development of Afro-Creole Culture in the
Eighteenth Century (Baton Rouge, Louisiana, 1992), p.l59; Mervyn C. Alleyne, Roots of
Jamaican Culture (London, 1988), pp. x, Michael Mullin, Africa in America: Slave
18;
Acculturation and Resistance in the American South and the British Caribbean, 1736-1831
(Urbana, Illinois, 1992), p.l60.
6. In addition to the three principal organizers are three associates: Herbert S. Klein, Joseph C.
Miller, and Barbara L. Solow. The core data consist of 177 fields of information, including
thenames of vessels, captains and shipowners, regions and dates of trade in Europe, Africa,
and the Americas, and the number, age, and gender of slaves confined on the middle passage.
The directors eventually hope to link related information, such as African climatic patterns,
slave phenotypes, slave rebellions, and slave prices to the main data-set.
 THE CULTURAL IMPLICATIONS OF THE ATLANTIC SLAVE TRADE 143
7. David Eltis and David Richardson. ‘The Structure of the Transatlantic
Slave Trade,
1595-1867’ (unpublished paper presented to the Social Science
History Meeting, 1995).
This an important paper. Much of my analysis, which sometimes varies from
is
that of the
authors, is based on their tables. owe them a great debt for sharing their information with
1

me.
8. David ‘The Volume and African Origins of the British Slave Trade before 1714’,
Eltis,
Cahiers d’ Etudes africaines, 138 (1995), pp.6 17-27; David Richardson, ‘Slave
Exports from
West and West-Central Africa, 17()()-1810; New Estimates of Volume and Distribution’,
Journal of African History, 30 (1989), pp.1-22; David Richardson, ‘The Bntish Empire
and
the Atlantic Slave Trade 1660-1807’, in Peter J. Marshall (ed
). The Oxford History of the
British Empire: The Eighteenth Century (forthcoming). Table 3.
9. David Eltis and Stanley L. Engerman, ‘Was the Slave Trade Dominated by Men?’, Journal
of Interdisciplinary History, 23 (1992), pp.237-57 (quote p.245); David Eltis and Stanley L.
Engerman, ‘Fluctuations in Sex and Age Ratios in the Transatlantic Slave Trade,
1663-1864’, Economic History' Review, 2nd series, 46
(1993), pp. 308-23. For a less
persuasive view on this subject, see Joseph E. Inikori, ‘Export versus Domestic Demand;
The
Determinants of Sex Ratios in the Transatlantic Slave Trade’, Research in Economic Historv
‘

14 (1992), pp. 11 7-66.

10. Eltis and Richardson, ‘The Structure of the Transatlantic Slave Trade’
11. Ibid.
12. Ibid.; Richardson, ‘Slave Exports’, pp. 13-14; Johannes Menne Postma, The Dutch in the
Atlantic Slave Trade J 600-1 81 (New York, 1990), pp.l 14-15, 122-3.
13. Robin Law, The Oyo Empire c. 1600-c. 1836: A West African Imperialism in the Era
of the
Atlantic Slave Trade (Oxford, 1977), pp.206, 219-29. 274, 281-2, 303-8; Paul E.
Lovejoy,
The Central Sudan and the Atlantic Slave Trade’, in Robert W. Harms, Joseph C. Miller,
David S. Newbury, and Michele D. Wagner (eds.). Paths Toward the Past: African Historical
Essays in Honor of Jan Vansina (Atlanta, 1994), pp. 345-70 (quote p.359).
14. 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 (Durham, North Carolina, 1992), pp.77-115; David Northrop, Trade without
Rulers: Pre-Colonial Economic Development in South-Eastern Nigeria (Oxford, 1978),
pp.60-4.
15. Eltis and Richardson, ‘The Structure of the Transatlantic Slave Trade’.
16. Mary
C. Karasch, Slave Life in Rio de Janeiro 1808-1850 (Princeton, 1987), pp.xxiv, 8, 18.
17. Eltis and Richardson, ‘The Structure of the Transatlantic Slave Trade’; David Eltis
and David
Richardson, ‘West Africa and the Transatlantic Slave Trade: New Evidence of Long-Run
Trends’, this volume.
18. Adam Jones and Marion Johnson, ‘Slaves from the Windward Coast’, Journal of African
History, 21 (1980), pp.l 7-34, for arguments for not accepting destinations; Postma, The
Dutch in the Atlantic Slave Trade, pp. 120-4; Karasch, Slave Life in Rio,
p. 16. When detailed
records on individual voyages are available, the ship often stops at a number of ports; see,
for example, 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 1 698-1 725
(London, 1991), and Suzanne Schwarz, Slave Captain: The Career of James Irving in the
Liverpool Atlantic Slave Trade (Wrexham, Clwyd, 1995).
19. Personal communication from David Eltis, 23 October 1996.
20. Joseph C. Miller, Way of Death (Madison, Wisconsin, 1988), pp.7-8, 141-53, 189-203,
223,
224 (quote), 226.
21 . and Richardson, ‘The Structure of the Transatlantic Slave Trade’.
Eltis
22. Richardson, ‘The Bntish Empire and the Atlantic Slave Trade’, in Marshall (ed
), O.xford
Histor\' of the British Empire', David Eltis, ‘The Traffic in Slaves between the British West
Indian Colonies, 1807-1833’, Economic History Review, 2nd series, 25 (1972),
pp. 55-64;
Svend E. Green-Pedersen, ‘The
Scope and Structure of the Danish Negro Slave Trade’,
Scandinavian Economic History Review, 19 (1971), pp.l 49-97; Postma. The Dutch in the
Atlantic Slave Trade, pp.l 68-9, 226; Franklin W. Knight, Slave Societ}' in Cuba during the
144 ROUTES TO SLAVERY
Nineteenth Century (Madison, Wisconsin, 1970), p.8.
23. Miller, ‘Number, Origins’, in Inikori and Engerman (eds.), Atlantic Slave Trade, p.89;
Patnck Manning, ‘Migrations of Africans to the Americas; The Impact on Africans, Africa,
and the New World’, The History Teacher, 26 (1993), pp. 279-96, especially p.281 Philip D. ;

Morgan and Michael Piedmont Virginia, 1720-1790’, William and

L. Nicholls, ‘Slaves in
Mary Quarterly, 3rd series, 46 (1989), pp.21 1-51; Philip D. Morgan, ‘Black Society in the
Lowcountry, 1760-1810’, in Ira Berlin and Ronald Hoffman (eds ). Slavery and Freedom in
the Afte of the American Revolution (Charlottesville, Virginia, 1983), pp.83-141.
24. Miller, Way of Death, p.225; Philip D. Curtin, Economic Change in Precolonial Africa:
Senegamhia Era of the Slave Trade (Madison, Wisconsin, 1975), pp. 187-8. See also
in the

Adam Jones, ‘Receptive Nations; Evidence Concerning the Demographic Impact of the
Atlantic Slave Trade in the Early Nineteenth Century’, Slavers' and Abolition. 11 (1990),
pp.42-57.
25. Richard Rathbone, Review of Thornton, Africa and Africans in Journal of African History,
34 (1993), 495-6; Law, Oyo Empire, pp.5-7; Northrup, Trade Without Rulers, p.l5; Miller,
Way of Death, p.28. This subject is treated at much Hawkins and Philip
greater length in Sean
Morgan, ‘Patterns of Cultural Transmission; Diffusion, Destruction, and Development in the
African Diaspora’ (paper delivered at York University Workshop, ‘The African Diaspora and
the Nigerian Hinterland’, 2-4 February 1996, which the authors hope to publish in a revised
form soon).
26. Karen Fog Olwig, ‘African Cultural Pnnciples in Caribbean Slave Societies; A View from
the Danish West Indies’, in Stephan Palmie (ed). Slave Cultures and the Cultures of Slavery
(Knoxville, Tennessee, 1995), p.29; Sidney W. Mintz, ‘More on the Peculiar Institution’,
New West Indian Guide, 58 (1984), pp. 185-99, especially p.l89; David Geggus, ‘Sex Ratio,
Age and Ethnicity in the Atlantic Slave Trade; Data from French Shipping and Plantation
Records’, Journal of African History, 30 (1989), pp. 23-44, especially p.35; Curtin, Atlantic
Slave Trade, p.l84, and Curtin, Economic Change in Precolonial Africa, p.l79 (quote).
27. Barry W. Higman, ‘African and Creole Slave Family Patterns in Trinidad’, Journal of Family
History, 3 (1978), pp. 163-80. For a contrasting argument in a different context, see Colin A.
Palmer, ‘From Africa to the Americas; Ethnicity in the Early Black Communities of the
Amencas’, Journal of World History, 6 (1995), pp. 223-36.
28. Claude Levi-Strauss, The Savage Mind (London, 1966); Shane White and Graham White,
‘Slave Hair and African-American Culture in the Eighteenth and Nineteenth Centuries’,
Journal of Southern History, 61 (1995), pp. 45-76, and ‘Slave Clothing and
African-American Culture in the Eighteenth and Nineteenth Centuries’, Past and Present,
148 (August 1995), pp. 149-86; Grey Gundaker, ‘Tradition and Innovation in
African-American Yards’, African Arts (April 1993), pp.58-7I, 94-96; Richard Westmacott,
African-American Gardens and Yards in the Rural South (Knoxville, 1992).
29. rely heavily on an excellent article by Richard Cullen Rath, ‘African Music in Seventeenth
I

Century Jamaica; Cultural Transit and Transition’, William and Mary Quarterly, 3rd series,
50 (1993), pp. 700-26, although my emphases are a little different, and Rath did not have the
advantage of the most recent analysis of the slave trade to Jamaica, which can be found in
Eltis, ‘Volume and African Origins’, p.619.
30. Serge Larose, ‘The Meaning of Africa in Haitian Vodu’, in loan M. Lewis (ed.). Symbols and
Sentiments (London, 1977), pp. 85-1 16; John M. Janzen, Leniba. 1650-1930: A Drum of
Affliction in Africa and the New World (New York, 1982), especially pp. 273-92; Luc de
Heusch, ‘Kongo in Haiti; A New Approach to Religious Syncretism’, Man, XXIV (1989),
pp. 290-303; David Geggus, ‘Haitian Voodoo in the Eighteenth Century; Language, Culture,
Resistance’, Jahrbuch Fur Geschichte Von Staat, Wirtschaft und Gesellschaft Latein
Amerikas, 28 (1991), pp.21-51; David Geggus, ‘The Bois Caiman Ceremony’, Journal of
Caribbean History, 25 (1991), pp. 41-57, especially p.50; David Geggus, ‘Sugar and Coffee
Cultivation in Saint Domingue and the Shaping of the Slave
Labor Force’, in Ira Berlin and
Philip D. Morgan (eds.). Cultivation and Culture: Labor and
the Shaping of Slave Life in the
Americas (Charlottesville, Virginia, 1993), pp. 73-98, especially p.80. See also John K.
Thornton, ‘On the Trail of Voodoo; African Christianity in Africa and the Americas’, The
Americas, 44 ( 1987-1988), pp.261-78.
THE CULTURAL IMPLICATIONS OF THE ATLANTIC SLAVE TRADE 145
31. Douglas B. Chambers, He Gwine Sing He Country”: Africans, Afro- Virginians, and the
Development of Slave Culture in Virginia, 1690— 1810’, unpub. Ph.D. thesis. University
of
Virginia, 1996; Douglas B. Chambers, '“He is an African But Speaks
Plain”; Historical
Creolization in Eighteenth-Century Virginia’, Alusine Jalloh and Stephen Maiziish (eds.),
in
Africa and
the African Diaspora (College Station, Texas, 1996),
pp. 100-133; and Douglas
B. Chambers, ‘“My own nation”; Igbo Exiles in the Diaspora’, this volume.
Chambers has
broadened his argument to two other regions in his ‘Eboe, Kongo, Mandingo: African Ethnic
Groups and the Development of Regional Slave Societies in Mainland North America, 1700
1820’, Working Paper No. 96-14, International Seminar on the History
of the Atlantic World
1500-1800, Harvard University, September 1996.
32. Eltis, Volume and African Origins p.619; Allan Kulikoff, Tobacco and Slaves: The
,

Development of Southern Cultures in the Chesapeake, 1680-1800 (Chapel Hill, North

Carolina, 1986); Philip D. Morgan, Slave Counterpoint: Black Culture in the
Eighteenth-Century Chesapeake and Lowcountry (Chapel Hill, North Carolina, 1997).
33. Hall, Africans in Colonial Louisiana, pp.41, 51, 111-13, 164,
293, 302; Peter Caron, ‘The
Peopling of French Colonial Louisiana; The Origins and Demographic Distributions
of
African Slaves, 1718-1735’ (paper delivered at Tulane-Cambridge Atlantic World Studies
Group, 21-23 November 1996); Peter Caron, “‘Of a nation which the others do not
understand”; Bambara Slaves and African Ethnicity in Colonial Louisiana, 1718-1760’, this
volume; John Michael Vlach, ‘The Shotgun House: An African Architectural Legacy’, in
John Michael Vlach (ed.). By the Work of Their Hands: Studies in Afro-American Folklife
(Charlottesville, Virginia, 1991), pp. 185-213.
34. This account rests on Knight, Slave Society' in Cuba, pp.60-1; Manuel
Moreno Fraginals,
‘Africa in Cuba: A Quantitative Analysis of the African Population in the Island of Cuba’,
in Vera Rubin and Arthur Tuden (eds.). Comparative Perspectives on Slavery’
in New World
Plantation Societies (New York, 1977), pp. 187-201; Herbert S. Klein, S/aven- in the
Americas (Chicago, 1967); Stephan Palmie, ‘Ethnogenetic Processes and Cultural Transfer
in Afro-American Slave Populations in Wolfgang Binder (ed.). Slavery in the Americas
,

(Wurzburg, 1993), pp. 337-33; and most usefully, George Brandon, Santeria from Africa to
the New World: The Dead Sell Memories (Bloomington, 1993), pp.55, 69 (quotes).
35. Sidney W. Mintz and Richard Price, The Birth of African-American Culture: An
Anthropological Perspective (Boston, 1992), pp.48-50; for size of purchases, see Stephanie
Smallwood, ‘After the Atlantic Crossing: The Arrival and Sale of African Migrants in the
Bntish Americas, 1672-1693’, Working Paper No. 96-13, International Seminar on the
History of the Atlantic World 1500-1800, Harvard University, September 1996, and Trevor
Bumard, ‘Who Bought Slaves in Early America? Purchasers of Slaves from the Royal
African CompanyJamaica, 1674-1708,’ Slavery and Abolition, 17 (1996), pp. 68-92.
in
36. Sidney W. Mintz, ‘Foreword’, in Norman E. Whitten, Jr. and John F. Szwed (eds.),
Afro-
American Anthropology: Contemporary’ Perspectives (New York, 1970), p.9, and Sidney W.
Mintz, ‘Creating Culture in the Americas’, Columbia Forum, 13 (1970), p.8.
 Notes on Contributors

David Eltis is Professor of History at Queen’s University, Kingston,

Ontario; Research Lecturer at the University of Hull; and Visiting Fellow at

the W.E.B. Du Bois Institute for Afro-American Research, Harvard.

David Richardson is Reader in Economic History, University of Hull.

Herbert S. Klein is Professor of History at Columbia University.

Stanley L. Engerman is John H. Munro Professor of Economics and

Professor of History at the University of Rochester.

Stephen D. Behrendt has taught at Drake University and Northern Iowa

University and is an Honorary Fellow of the W.E.B. Du Bois Institute,
Harvard.

Douglas B. Chambers teaches at the Corcoran Department of History,

University of Virginia, Charlottesville.

Peter Caron is a graduate student at the Department of History, Tulane

University, New Orleans.

Philip D. Morgan is Professor of History at Florida State University.

 Index

Aboh, 33n, 74, 79 Benin, 23-4, 27, 79

Accra, 23-4, 59, 69-70n, 1 3 Bettelheim, Judith, 96n
aja (sacrifice), 74, 81-2 Bight of Benin (Slave Coast), 6-8, 16-33,
Akim, 20 52-3, 56, 58, 61, 65, 67n, 77, 83, 100, 118n,
Akinjogbin, I.A., 27 125-9, 131-2, 136-7, 140
Alampo, 23-4 Bight of Biafra, 6-8, 10, 16-33, 52, 56, 58,
Albreda, 98 61-2,64, 72-97, 125-9, 131, 137-8
Allada, 140 Bile (Agbaniye Ejika), 78
Alleyne, Mervyn, 124 Bimbia, 23^, 26, 29-31
Amakiri (King of New Calabar), 78 Bissau, 98, 104
Amazon river, 134 black-eyed peas, 85
Andony, 23-4, 75, 78, 81 Bonny, 6, 22-4, 26, 29, 31, 58, 64-5, 73-4,
Angola, 4, 12, 17, 61, 130, 136, 138 also 78-9, 81, 87, 89, 93n, 129, 131
West-Central Africa Bosman, Willem, 103
Ani (god), 81 Brandon, George, 140-1
Anomabu, 6, 22-4, 34n, 59 Brass-Nembe, 74
Antigua, 83 Brazil, 4, 6, 107, 120n, 125-8, 130, 135
Apa (Epe), 21, 23^, 27 Bridgetown, see Barbados
Apam, 23^ Bristol, 3-4, 20, 54, 64
Ardrah, see Jaquin Coal Tar Office, 64
British
Arguin, 104, 106 Buenos Aires, 130
Aro network, 29, 74, 79 Bunda, 120n
Arthy, Elliot, 68n Buxton, Thomas Fowell, 39
Asaba, 79
Asimimi (King of Bonny), 78 cabildos {clubs), 140-1
Assinie, 23-4 Cabinda, 98-9, 117n, 130-1
6,
Australia, 49, 64 Calabar, 6, 21-4, 26, 28-31, 58, 74, 79, 86,
Awey, 27 130-1
Axim, 23-4, 69n, 131 Cameroons (Cameroons river), 22-4, 26,
Ayub b. Sulyman (Job Ben Solomon), 120n 29-31,58
Azumini, 79-80 Candomble (Bahia), 141
Cap Francois (St. Domingue), 133
Badagry, 22-^, 26-7, 30-1 Cap Negre, 101
Bahamas, 83-4 Cap Saint-Louis (St. Domingue), 106
Bahia, 7-8, 19-21, 87, 125, 127-9, 132, 134 Cape Apollonia, 16-18, 59, 98-9
Baikie, William B., 73, 81, 92-3n Cape Coast Castle, 6, 22-3, 34n, 60
Bambuk, 1 12 Cape Lahou, 98
Barbados, 7, 20, 72, 83, 90, 95n, 125-6, 128, Cape Lopez, 16-18, 23-4, 26, 29-31, 101
133 Caron, Peter, 10-12, 98-121, 139
Barokunda Falls, 59 Cartagena, 134
Barth, Frederick, 107, 117n Casamance river, 1 10, 120n
Abdoulaye, 104
Bathily, Cassidy, Frederick, 96n
Beedham, Katherine, 4 Chambers, Douglas, 10-12, 72-97, 138
Behrendt, Stephen, 4, 7, 9-10, 46-7n, 49-71, Charleston, 69n
124 Chesapeake, 83-4, 87, 95n, 1 38
Belleford, Nicolas Villault de, 103 Chickasaw Indians, 104-5
Bende, 79 Christiansborg,23-4
Benguela, 6, 17, 130-1 Clarkson, Thomas, 40, 47n, 53^, 67-8n
 1

148 ROUTES TO SLAVERY

coal tar, 64, 7 In Coromantee (Koromanti), 124, 135-7, 141
Cochrane, Archibald, 64 Djolauf (Senegal), 113-14, 120n
coffee, 33 Edo. 75,
Colombia, 134 Ewe-Fon, see Aja-Fon
Congo nver, see Zaire river Fanti, 20
Conny, John, 95n Fulbe, 104-5, 110, 112, 134
convict trade, 53, 58 Hausa, 129-30, 135
Consco, 23-4, 26, 29-3 Ibani, 78, 93n
Coughtry, Jay, 3-4 Ibibio, 21, 75. 135
crew, impressment, 57-8 Igalla, 79
mortality of, 9-10, 39^0, 49-71 Igbo (Ibo), 7, 10-11, 21, 33, 72-97, 100,
Crow, Hugh, 70n, 73, 89, 93n 116, 135, 138^0 a/jc; Calabar
Cuba, 4, 8, 21, 117n, 125, 127, 130, 133, Ijo, 75
140-1 Jallonke, 104
Cura(^ao, 133 Jallonke Fulbe, 1 10
Curtin, Philip, 1-2, 4, 6, 8, 11, 17, 103, 123, Kakanda, 130
134-6 Kasanga, 10 1

Kwa, 137
Daget, Serge, 3—4, 49-50 Lucumi, 140-1
Dahomey, 27, 1 38 Lunda-Tchokwe, 130
Dakar, 120n Male, 107, 118-20n
Dartmouth, 66n Malinke, 110, 136
Davies, K.G., 67n, 69n Mande, 110, 139
Debien, Gabriel, 118n Mande Juula (Jaxante), 1 10
Delafosse, Maurice, 105, 119n Mandinga, 107
Demerara, 133 Mandinka, 12 1

dibia, 82, 88-9, 96-7n see also obeah Mina, 118n, 131, 135
Dixcove, 60 Nala, 110
Dolben Act, 37, 63^ Ndoki, 93
Dominica, 133 Ngangela, 130
Du Bois Institute (Harvard), 3-5, 124, 131-2 Ngwa, 78
Duke Town (Calabar), 29 Nomi, 69n
dysentery, 55 Nupe, 129-30
Ogoni, 75, 78
ekpe society, 29, 87 Ovimbundu, 130, 135
Elmina, 23-4, 34n, 100, 131, 135 Papel, 1 10
Eltis, David, 1-35, 49, 67n, 124-8, 130-2, Sereer, 110, 134
139^0 Sissibe, 1 10
Engerman, Stanley L., 8-9, 36-48, 83, 126 Soso, 1 10
Epe, see Apa Timbo, 1 10
Equiano, Olaudah, 3, 72—4, 81, 89-91, 96n Twi, 124
ethnicity, Wolof, no, 134
groups, names Yoruba, 21, 96n, 135, 139-41
Ahanta, 20 see also slaves, ethnicity
Aja-Fon (Ewe-Fon), 83, 137, 139—40
Akan, 11, 20, 33, 85, 88, 137 Falame, 104
Asante, 20 Falame river. 111, 1 1 8n
Balante, 1 10 Fiji, 49

Bambara, 11-12, 99-121, 123, 135-6, 139 Forde, Daryll, 74

Beafada, 1 10 Formosa, 23^
Bantu, 137 Fort Saint-Joseph (Galam), 104
Bijago, 1 10
Bobo, 110 Gabon, 22-4, 26, 29-31
Borgu, 129 Galam, 102, 104, 107, 111
Calabar, 131 see also Igbo Gambia. 55-6, 58-9, 65, 69-7 In, 134, 136
Congo, 137-9, 141 gamby (Eboe drum), 84, 138
 INDEX 149
Geggus, David, 83, 117n, 135, 138 Labat, Jean-Baptiste, 103
Gold Coast, 6-8, 11, 16-33, 52, 55-6, 58-60, Lagos (Onim), 16, 22^, 26-7, 30-1, 34n
69n, 77, 100, 125-7, 131-2, 135-8 Lancaster, 3
Goree, 98, 102-3, 111-12, 119n Law, Robin, 135
Grand Popo, 2 23-4, 27 1 , Leeward Islands, 7, 1, 84, 95n, 125 see also
1

Green- Pederson, Sven, 4 Antigua and St. Kitts

Grenada. I19n, 132 lemba (cult), 138
grisgris (charm), 139 Levi-Strauss, Claude, 136
Guadeloupe, 21, 83, 125 Levtzion, Nehemia, 103
Guinea-Bissau, 110 Lewis, Matthew (‘Monk’), 73, 87
Guyanas, 7, 125 Little Popo, 23-4, 27
Liverpool, 3-4, 20, 22, 54-7, 61-2
Hair, P.E.H., 93n London, 3-4, 20, 54, 90
Hall, Gwendolyn Midlo, 10, 123-4, 135, 139 Long, Edward, 88, 95-6n
Havana, 4, 140 see also Cuba Louisiana, 11-12, 98-121, 124, 135, 139
Higman, Barry, 136 Lovejoy, Paul E., 129
3,
Hogendom, Jan, 39 Luanda, 6, 17, 130-1
Honfleur, 49 Lyttleton, William, 69n
Horton, Robin, 93n
Hume, Captain Robert, 64 Madagascar, 52
malaria, 53, 55, 58-60, 6.3-5, 70n
ife-njokku (New Year festival), 97n Maroons, 85-7
Igboland, see ethnicity, Igbo Martinique, 21, 125
Ikuka (god), 79 Maryland, 120n
Imo river, 78-80 masquerades (mnnio), 82, 87-8
Inikori, Joseph, 46n Mauritius, 49
Islam, see slavery Mettas, Jean, 3-4,49-50 7,
Isuama region, 74, 78, 89 Middelburg Commercial Company, 49
migration, white, 53, 64
Jamaica. 7, 11, 20-1, 33, 69n, 72-3, 83-4, Miller, Joseph C., 4, 39, 132, 1.34-5, 142n
86-7, 90, 95n, 126, 133, 136-7, 141 Mina Coast, 21-4
James Fort (Gambia), 59 Minas Gerais, 128, 134
James Island, 69n Mintz, Sidney, 142
Jaquin, 21, 23-4, 27 Mississippi river, 99, 132
Jea, John, 86 Montego Bay, see Jamaica
Jefferson, Thomas, 95n Morgan, Philip D., 10-13, 97n, 122-45
Jensen, Richard, 49, 57, 67n mortality, see crew; slaves; troops
Jenzen, John, 137 Mozambique, 52, 131, 136
Jonas, Simon, 94n Mullin, Michael, 124
Jones, G.I., 74
jonkonii, 1 1, 84, 87-8, 96n, 138 Nantes, 8, 49
juju, 81-2, 86, 88 Natchez, 98
New Calabar (Elem Kalahari), 22-4, 29, 74
Kamallo 93n
(spirit), 78-9, 87, 93n, 117n
Kambasa (Queen of Bonny), 78 New Orleans, 98, 106
Karasch, Mary, 30 1 New York, 86, 130
Keta, 23-4, 27 Niger river, 19, 79
Khasso, 103 njokku (yam spirit), 87-8
Kingston, see Jamaica Northrup, David, 92n, 135
kitchen crops, 85 Nn-Awka region, 74, 79, 89
Klein, Herbert S., 4, 8-9, 36-48, 64, 67n, 83, Nn-man (atama), 81, 89
142n
Koelle, Sigismund W., 73-4, 90 obeah {obia, obea), 11, 73-4, 84, 88-90 see
Koromantine, 23-4, 135-6 also dibia
Kumalu (King of Bonny), 78 Obegu, 79-80
 1 3 1 1

150 ROUTES TO SLAVERY

Obichere, Boniface, 92n Richardson, David, 1-35, 49, 124-8, 130-2,
Offra, see Jaquin 139-40
Ohuhu, 79-80 Rinchon, Dieudonne, 49
okonko society, 87 Rio Brass, 23-4
okpala (lineage head), 81 Rio de Janeiro, 4, 12, 130-2, 134
okpora (headman), 82 Rio de la Plata, 134
okra (okro), 85, 95n, 138 Rio Nun, 23-4
Okrika. 74, 78 Robert, Saint (Director, Senegal concession),
Old Calabar, see Calabar 104
Old Town (Calabar), 29 Roberts, Richard, 102
Olwig, Karen Fog, 135 Rodney, Walter, 82, 107
Onim, see Lagos Rossard, Michel, 99
Onitsha, 74, 86, 89 Royal African Company, 36, 69n
Opobo, 79-80
74,
Opoli of Azuogu, 78 Sahara, 13
orisha (worship), 140 St. Domingue, 7, 1
1 , 20, 57, 59, 83, 86, 1 17n,
Orlu, 89 125-6, 133, 137, 141
Ormankoobe, 103 St. Kitts, 83
Oruyingi (goddess), 79 St. Louis, 98
Ossomari, 79 Saint-Mery, Moreau de, 102-3, 120n
Owoamekaso (god), 79 St. Vincent, Kingstown Bay, 64
Oyo Empire, 27, 30, 129, 135 Salvador, see Bahia
Samba Gelaago Jegi, leader of Futa Toro, 1 10
palm oil, 20 Sandoval, Alonso, 76
Palmares, 86 Santeria, 139-41
Palmer, Colin, 4 Sao Paulo, 134
Palmie, Stephan, 117n Savannah la Mar, see Jamaica
Paris, 104 Schofield, M.M., 4
Pereira, 76, 93n Schuler, Monica, 86, 90
Perier, Etienne de (Governor of Louisiana), Searing, James, 1 1

104-6 Sekondi, 60
Pernambuco, 134 Senegal, see Senegambia
Peru, 134 Senegambia, 7, 11-12, 1^17, 52, 61, 77,
petro divinity (Kongo), 138 98-100, 103-21, 125-6, 128-9, 134, 136,
Philadelphia, 95n 138-9
Phillips, James, 63 Sheffield, Lord John, 71n
Pinet, Jean (translator), 107-8 Shendan, Richard B., 7 In
Plymouth, 66n Shields, 66n
Popo, 23^, 27, 36 1 shipmasters,
Portentic, 99 Breban, Captain, 1 1

Portobello, 134 Dentuly, Contault, 1 1

Porto Novo, 22-4, 26-7, 30-1 shipping, productivity of, 5-6, 64-5, 7 In
Postma, Johannes, 3-4, 133 ships, names,
Potosi, 134 L’Ajfriquin, 119n
Pratz, Jean Baptiste Le Page du, 106, 113-14, L'Airnable Renotte, 119n
119-20n L'Annibal, 106, 119n
Le Coitrrier de Bourbon, 1 19n
Quaqua, 23 Le Dauphin, 119n
Le Due du Noailles, 104
(Dahomey), 138
divinity Maria, 69n
Raguet, Judge Jean Baptiste, 14 1 Mars, 69n
Rath, Richard Cullen, 144n La Neptune, 9n 1 1

Rawley, James A., 4 La Nereide, 1 1

Red Sea, 13 Pilgrim, 64

Reis, Jaoa Jose, 107 La Renee Frangoise, 101
Rhode Island, 3-4 Saint Ann, 68n
 INDEX 151

Le Saint-Louis, 1 1
states, African,
Le Saint-Michel, 1 1
Baol, 110
Le Saint Ursin, 17n 1 Futa Bundu, 109, 112
ships, tonnages, 19, 34n FutaDjallon, 104, 109-10, 112
Shlomowitz, Ralph, 43 Futa Toro, 109-10, 112, 120n
shotgun houses, 139 Gajaaga, 110, 112, 134
Sierra Leone, 7, 52, 56, 58, 69n, 74, 77, 83, Jolof,120n
90, 138 Kaarta, 135
Sire-Abbas (historian), 105 Kaladian Kulibali, 103
Slave Coast, see Bight of Benin Kayor (Kajoor), 110, 120n
slave, voyage dataset, 4-5, 7-8, 16-33, 37-8, Kongo, 137-8
124, 131, 142n Segu, 135
slave-carrying, regulation of, 9, 37, 43, 63^ Segu Bambara, 102-3
slavery, and Atlantic development, 122-3 Sokoto Caliphate, 129
and drought, 111-13 Waalo, 110, 120n
and Islam, 101-2, 107, 109-14, 118n, 120n Steckel, Richard, 49, 57, 67n
and sugar, 32 Stein, Robert L., 4, 49, 57, 70n
slaves, carriers of, 2-4, 8, 19-21, 43, 111, Stuckey, Sterling, 96n
128-9, 131-2 Studer, Elena de, 4
composition, 29-33, 126-7 Surinam, 7, 125-6, 141
ethnicity, 10-12, 73-4, 90-1, 98-121,
1 23^, 29-30, 34-42
1 1 see also ethnicity, Talbot, P.A., 88
groups Tantumquerry, 23-4
executions, 119n Tenda region, 104
family life, 127-8 Thornton, John, 10, 123, 128, 134-5
mortality, 2, 5, 8-10, 20, 36-49 Timbuktu, 103
named, Trinidad, 133, 136
Benjamin, 101, 1 18n troops, mortality, 57
Edward, 73 Tucuman, 134
Fran9ois, 114 ts eddo army, 110, 1 2 On
Jean, 1 14
Malene, 107-8 Unger, W.S., 49
Marboux, 1 14 Upper Guinea, 129, 138
Pickle, 73
Pierre, 1 14 Vilar, Enriqueta Vila, 4
Pierre Birame, 1 14 Virginia, 3, 11-12, 95n, 134, 138
Pierrot, 1 14 voodoo, 137, 139
Samba, 106, 119n
narratives of, 3, 47n, 72-3, 86, 89 West-Central Africa, 6-7, 52-3, 56, 58, 67n,
numbers shipped, 1-2, 23, 27-8, 36, 73, 125-30, 132, 140 see also Angola
75-7, 83, 92n, 99, 111, 11 7n Westergaard, W., 4
per ship, 36, 111 Whitehaven, 3
recaptured, 83 Whydah, 6, 21-7, 29-31, 34n, 69n, 98-9,
re-export of, 133-4 117n, 131-2
revolts, 119n, 137 Wiamba, 23^
slaving, Windward Coast, 52, 56, 58, 65, 126, 129,
geography of, 6, 8, 12, 16-33, 44, 75-7, 131-2
94n, 117n, 125-6, 131-4, 136-8 Wondji, C., 1 10, 120n
networks, 132 see also Aro
temporal patterns, 18-20, 25-9, 129 yams, 85, 138
Solow, Barbara L., 142n yellow fever, 53, 55, 57-60, 63-5, 69-70n
South Africa, 49 Young, William, 64
South Carolina, 86, 95n, 34 1

South-East Africa, 52, 125, 127 Zaire river, 17, 128, 130-1
Stanfield, James, 62-3 zinzin (amulets), 139
V
I

!
V
 0651

00615

1143

3
 nr he scale ot the Atlantic slave trade
.1 is central to the debate
over transatlantic slavery from the sixteenth to the mid-
•
^

nineteenth centuries. Res CHMOND PUBLIC L BRAR CA 94804- 165

"1st amount of

data on slaving voya] I has been

collected by the W.E.B. 3 1143 00615 0651
ro-Ameriean
Research Harvard University. Containing records of some
at

25,000 slaving voyages between 1595 and 1867, this data set
forms the basis of most of the studies in this collection. The
other essays supplement the quantitative analysis by examining
issues relating to the ethnicity of slaves. In addition to
presenting new evidence on mortality trends in the slave trade
and on African influences on the history of American slave
societies, the volume raises important questions about how
slaves reconstructed their identities outside of their homeland.

TTV'
’rofessor of History at Queen’s University, Kingston,
DATE DUE thor of Economic Growth and Ending of the
the
{ Trade (1987) and edited, with James Walvin, The
Hon of the Atlantic Slave Trade (1981 ).

ARDSON is Reader in Economic History at the

•1 He has edited Bristol, Africa and the Eighteenth-
de America, 4 volumes (Bristol Record Society,
to

vlition and its A ftermath: The Historical Context

1790-1816 {¥v 2iX\k Cass, 1985).

ISBN a7mbq3'=io-q

je, 900 Eastern Avenue

5H, UK

Return Material Promptly '

llE Hassalo Street
-
J 13-3644, USA 9 780714 643908
Printed in Great Britain ;