The Atlantic Slave Trade: a Census


The Atlantic Slave Trade: a Census

For more than a hundred years, even serious students of the history of slavery believed the number of African slaves brought to the Americas was between 15 to 20 million. Then, in 1969, Philip D. Curtin questioned those numbers in his book, The Atlantic Slave Trade: a Census.

Until Curtin’s book, discussions regarding the scale of the slave trade had largely been based on speculative estimates without ever really being subjected to critical scrutiny.   The author, who completed his graduate studies in 1953 at Harvard University with special interests in the fields of African and Caribbean history, argues that the estimates are fallacious. He contends these estimates were mostly based on nineteenth century information, when Latin American history was in its infancy. Since then, new historical knowledge has been available, but the old synthesis has not been corrected.   Curtin seeks to provide us with a study which is more closely and honestly attuned to the data. Furthermore, he has stated that the purpose of this book is to provide a departure point for new research, rather than a definitive study.

Slave Ship Fredensborg II, 1788_jpg

To carry out the task he assigned himself, Curtin chose not to perform monographic research. Instead, he examined the massive amounts of published literature on the slave trade as a point of departure. He conceived the task “as that of building with the bricks that exist, not in making new ones” (Curtin, 1969: xviii).

The author shows how major scholars used estimates regarding the number of African slaves shipped without verifying their sources. The lower estimation of 15 million begins in 1861, with the journalist Edward Dunbar’s calculations. Curtin declares, “The estimates were guesses, but they were guesses educated by knowledge and the historical literature. They earned the approval of later generations who were in a position to be still better informed” (1969: 8). He proceeds to explain how W.E.B. Dubois quotes Dunbar’s estimates in a speech to London’s Congress in 1911. The 15 million Dubois estimate is then picked up and cited by R.R. Kuczynski in 1936. Thereafter the estimate is cited and used by numerous scholars, including Basil Davidson in Black Mother (1961), the best general history of the slave trade at the time, according to Curtin. The same type of thing is shown with the higher estimate of 20 million: it to was merely compounded from old, flimsy estimates. The methodological flaw that these sources shared in common was that none were based on historical research. In Curtin’s words, “The literature of the slave trade, then, carries a broadly held impression that the total slave imports into the Americans came to something like 15 to 20 million. But….the vast consensus turns out to be nothing but a fast inertia, as historians have copied over and over again the flimsy results of unsubstantial guesswork” (1969: 11).

cans Thrown Overboard from a Slave Ship, Brazil, ca_ 1830s_jpg

Curtin goes on, using a variety of both direct and indirect evidence, to formulate new quantitative estimates. The direct evidence consisted of slave import figures found in records of particular ports or colonies over time, and shipping records, which included carrying capacity of slave ships and importation contracts like Spain’s asientos. The asiento was a system of monopoly licenses issued to foreign firms that gave them the right to ship a stipulated number of African slaves to the Spanish colonies over a period of time. The patchwork of existing data, along with the adjustment problems regarding smuggling, required some fancy footwork which he justifies carefully. The indirect evidence includes population estimates and censuses, where they existed. The author also devised a method using slave population counts adjusted for mortality and natural increase to guesstimate the number of imports necessary to account for increases in the population. In addition, to calculate the number of Africans imported in relation to economic productivity, he used annual time-series of sugar production. Some estimates of slave exports from Africa were based on goods exported to Africa. And finally, Curtin uses analogy to help develop a new estimate of slaves brought to the Americas, using comparable cities, where records were kept in one but not the other.

Ultimately, the author comes up with tentative totals for various territorial, regional and continental areas culminating with a total estimation of 9.566 million African slaves imported via the Atlantic slave trade: 175,000 going to Europe and the islands off the African coast and the rest going to the Americas. This number is significantly less than the minimum estimate of 15 million regurgitated by historians in the past and his estimation, unlike the previous, is based on a synthesis of historical research gathered from the existing literature on the slave trade. Curtin stresses that his estimates should be regarded as calculations within a range of possibilities and not set figures. He repeatedly cautions throughout the book that his estimates could be as much as 20% off, either up or down.


One reviewer of The Atlantic Slave Trade: a Census, Roger Anstey, agrees that Curtin has convincingly shown the old consensus of 15 to 20 million, although used by distinguished scholars, is only an estimate which is not based on scholarship. He goes on to test Curtin’s estimates of the British trade between 1791 and 1800, which were based on the study of existing published work, to those made by Anstey using a wider range of sources. Anstey found that Curtin’s estimate of 325,500 compared to his findings of 360,000 is clearly within the 20% range Curtin regarded as acceptable (Anstey, 1970: 457). Anstey concludes that not only are Curtin’s methods and conclusions plausible but his goal of providing a point of departure for other researchers has been achieved.   He goes on to state, “Curtin has worked through a mountain of very varied and sometimes difficult material, and has applied it to the skills, not only of a historian but of a demographer and a statistician. In doing so he has given us an assessment, partly in the form of numerous tables and charts, which has long been lacking, and a base from which much further work can be undertaken. One might hazard that The Atlantic Slave Trade: a Census will remain unrivaled for many years and that fuller study will ratify many, at any rate, of Curtin’s conclusions” (Anstey, 1970: 459).

Slave Ship Henrietta Marie, 1699 1700_jpg

This book is a notable addition to the body of literature written on the colonial history of Latin America. The colonies of Latin America were created to provide individuals with opportunities for upward mobility by exploiting the land and labor of the New World. This wealth was also meant to be shared with the Royals of the Old World. With the decline of native populations and the introduction of sugar crops to the New World, a fresh source of labor was needed.   African slaves filled this void and their distribution undoubtedly influenced the social and political history of Latin America. It is therefore relevant for those students to have approximations of how many African slaves were imported across the various territories of Latin America.

From reading the book, I became aware of some interesting conclusions: between 1492 and 1770, Africans out numbered Europeans in the Americas. From 1701 to 1810, England, France and Portugal were the major players in the Atlantic slave trade. During that time, the three might have exported between five and six million Africans into the New World. Portuguese trade to Brazil, the largest African slave importer into the 19th century, accounted for nearly one third of the whole 9.566 million estimate.   Curtin’s figure of African slave imports into Spanish America for the whole period of the Atlantic slave trade is 1.552 million. Of this total 700,000 were shipped from 1521 to 1773 and 853,000 were shipped from 1774 to 1867. After 1807, Spain’s large slave exports were going only to Cuba and Puerto Rico.

Of the 9.566 million African slaves Curtin estimates came to the new world, three out of every five would have arrived between 1721 in 1820. At the height of the Atlantic slave trade between 1700 and 1840 the number of Africans forced to the new world was between 40,000 and 60,000 per year. After reaching this plateau, the numbers started to fall during the 1790’s, coinciding with the Napoleonic wars. Curtin thinks the decline was due to several factors, including the St. Dominique slave revolt and the British abolition in 1808.


Curtin’s pattern of distribution reveals that the United States received less than 5% of Africans brought to the New World. In contrast, Jamaica had 7.8%, Haiti 9%, Brazil 38.1%. “Contrary to the parochial view of history that most North Americans pick up in school,” he explains, “the United States was only a marginal recipient of slaves from Africa. The real center of the trade was tropical America, with almost 90% going to the Atlantic fringe from Brazil through the Guianas to the Caribbean coast and islands” (Curtin, 1969: 89).

However, the African slaves who suffered on plantations in the United States seem to have fared much better than their counter parts in Latin America. Determinations of survival and increase show Caribbean plantations, more so than others, worked African slaves to death. The Brazilian plantations were not far behind in life taking brutality. Biologically, the United States was the most favorable toward African slaves. Because reproduction rates can be considered an indication of treatment, challenges arise to historian Frank Tanenbaum’s hypothesis, popular at the time, which claims that slave conditions were harsher in the United States compared to the Latin American colonies. Curtin does not explore this or other observations of this sort in detail, but rather opens the door for further discussion.

In conclusion, Curtin’s book challenged the long held notion that 15 to 20 million Africans were brought to the Americas as slaves. In fact, he believes further study may well reduce that number to his own estimate of 9.566 million, give or take 20%. The Atlantic Slave Trade: a Census by Philip Curtin is easy to read, despite its complicated content. It is logically organized and well written. The tables and graphs are conveniently placed. His explanations are clear enough that those without statistical training can understand them. This book will certainly provide a suitable departure point for future researchers, because it opens the door to other observations, contains extensive notes, explanations and a detailed bibliography to aid that research.

References |  The Atlantic Slave Trade: a Census.

Curtin, Philip D. 1969. The Atlantic Slave Trade: a Census. Madison, Wisconsin: University of Wisconsin press.

Anstey, R. 1970. Numbers and the Slave Trade. The Journal of African history, 11 (3): 457-459.

Leave a Reply

Your email address will not be published. Required fields are marked *