The Intra-American Slave Trade Database

This essay was originally published in 2018.

Introduction

The text that follows in each of the introductory sections is intended to be read in combination with the overview of the Transatlantic Slave Trade Database. Having emulated the format and methodology of the transatlantic database wherever possible, here we only examine the sources, methods, and construction that are specific to the Intra-American Slave Trade Database. In most cases, the methodology, variable names, data codes, and user interfaces of each database are nearly identical and described thoroughly in the overview of the transatlantic database. What remains is to describe the origins of the intra-American database and explain its few distinctions in method and practice from the transatlantic precursor.

Whereas most twentieth-century slave trade scholarship focused on the Atlantic crossing, in the first decades of the twentieth-first century, historians working on disparate parts of the Americas increasingly document intra-American networks of forced migration, asserting their significance for both the African diaspora and the geopolitics of the Atlantic World. Scholars working on the era of the transatlantic slave trade have demonstrated that many African people who survived the Middle Passage across the Atlantic were forced to board subsequent vessels soon after arriving at a port in the Americas. These additional passages within the Americas expanded the African diaspora to far more destinations than were targeted by vessels coming directly from Africa, often moving enslaved people across imperial lines.¹ Likewise, in the era after the Atlantic slave trade closed, robust domestic trafficking networks emerged to forcibly move people from older slaveholding regions to burgeoning new ones, disrupting enslaved people's communities in the Americas and severing family ties.² The Intra-American Slave Trade Database seeks to pull data on all such forced migrations together. The consolidation of such information in a single database reveals patterns, connections, and entanglements previously hidden by language barriers and geographic specializations.

The Intra-American Slave Trade Database reveals the overwhelming importance and ubiquity of slave trading in the New World. It also expands on the geographic coverage offered by the Transatlantic Slave Trade Database, adding more than 27,000 slave-trading voyages within the Americas stretching from Newfoundland to Buenos Aires, and both the Atlantic and the Pacific littorals. This database offers basic demographic evidence to examine when, where, and how African captives and enslaved people born in the Americas endured forced journeys within and across the European empires of the Americas and newly formed nations such as the United States and Brazil. Such information has the potential to enrich cultural studies of the African diaspora with a clearer sense of the routes African women and men traveled to life in enslavement in various American locales. It also sheds light on lived experiences of captives, by adding complexity to our understanding of the duration of journeys in the slave trade, shipboard conditions, patterns of age and sex, and of course the horrific mortality of the traffic.

The new database also redraws the map of American slavery. If the launch of the online version of the Transatlantic Slave Trade Database in 2008 spotlighted the significance of Brazil as the most important broad region of slave disembarkation in the Atlantic slave trade (a fact previously underappreciated), the dominant pattern to emerge from this new intra-American database is the importance of the Spanish Americas as, probably, the second largest broad region of slave disembarkation in the Americas. Looking at the transatlantic traffic alone vastly understates the importance of Spanish territories to the history of slavery in the Americas, but the data contained here adds thousands of slave voyages from (mainly) Portuguese, British, and Dutch colonies to Spanish dominions, mostly in Cuba, Puerto Rico, Mexico, Colombia, Venezuela, and the Río de la Plata region. For those interested in the slave trade to the many Spanish colonies across the New World, it is imperative to use both the transatlantic and the intra-American databases, in order to examine the volume and routes of this traffic as well as the origins of African and African-descended captives.

The same is true for many colonies in other empires that were not the primary centers of slave exploitation. In the British case, for example, the transatlantic slave trade supplied most of the captives to major plantation zones, such as Barbados, Jamaica, South Carolina, and Virginia. But smaller colonies, and those less dependent on slave labor, tended to draw significant numbers of enslaved children, women, and men from intra-American routes of forced migration. As a result, those studying New York, the Bahamas, Massachusetts, or Belize will find as much (or more) relevant information in the intra-American as in the transatlantic database.

Coverage of the Intra-American Slave Trade

Internal American slave trades existed alongside the transatlantic traffic and continued after its prohibition, only coming to an end with the abolition of slavery itself. This database gathers a systematic documentation of the most important American seaborne slave trading routes in terms of duration and volume within and across imperial boundaries. The earliest voyages in the dataset conducted captives from the Caribbean islands (all under nominal Spanish dominion) to the American mainland in the sixteenth century.

The data covering English and Dutch intra-American trafficking (both within their empires and across colonial borders) seems largely comprehensive from the mid-seventeenth to the late-eighteenth century thanks to the work of O’Malley and Jelmer Vos, with the notable exception of the traffic from English and Dutch colonies into French territories and from the latter to the Spanish. (Numerous anecdotal accounts refer to the importance of intercolonial voyages for delivering enslaved people to the French Caribbean, but French officials prohibited such foreign commerce. As a consequence, the trade operated clandestinely, making data on individual voyages scarce.) Data on eighteenth century arrivals in the Spanish Americas are strong for the Río de la Plata region (with the exception of the Portuguese enclave of Colonia), Colombia, and Venezuela. The Danish and Swedish slave voyages internal to the Americas were significant for the late-eighteenth century traffic in the Caribbean, as well, particularly to Cuba and Puerto Rico. Jorge Felipe parsed out intra-American and transatlantic voyages for Cuba between 1790 and 1820, when both intra-American and transatlantic slave arrivals to Cuba significantly increased before the larger peak of direct transatlantic arrivals from 1820 to 1860.

New data added to the Intra-American Slave Trade Database in late 2021 and early 2022 added substantial coverage of trafficking within the Americas after abolition movements curtailed transatlantic slaving in the nineteenth century. Jennie K. Williams contributed data on more than three thousand maritime voyages that forcibly moved enslaved individuals from other U.S. ports, especially on the eastern seaboard, to New Orleans between 1820 and 1860. Research is ongoing to document more of this U.S. domestic slave trade, and these entries complement existing data on nineteenth-century trafficking between Spanish Caribbean colonies. Meanwhile, Daniel Domingues and researchers working with him at Rice University used Brazilian newspapers to document more than twelve thousand voyages that delivered enslaved people to Rio de Janeiro between 1831 and 1860, mostly from elsewhere in Brazil. It is noteworthy that each of these massive datasets focuses only on a single port of arrival, albeit a vitally important one. This suggests that the maritime portions of both the U.S. and Brazilian domestic slave trades were enormous and not yet completely documented.

The unit of record of this dataset (like the transatlantic database) is the shipment—a single maritime journey of one vessel. This organization of the database around shipments generally works well for documenting the slave trade since numerous kinds of records—port logs, newspapers, merchants’ papers, logs of import duties, insurance records—recorded the comings and goings of ships. The focus on individual shipments is not without shortcomings, however. First, this methodology offers no good way to capture statements or reports from colonial officials that described broader patterns in the slave trade, such as declarations that in a certain decade some total number of enslaved people arrived in a particular place. A database of discreet voyages cannot accommodate such information.

Whereas that limitation is a challenge for both the transatlantic and intra-American databases, another limitation of the shipment model is more specific to documenting intra-American movements of enslaved people. Namely, the focus on maritime voyages limits coverage of overland slave trading, a common feature of the intra-American traffic. Overland forced migrations occurred routinely in the mainland Spanish Americas from the sixteenth century onward, from Veracruz to Mexico City, or from Cartagena and the Río de la Plata to Peru. Overland trades also emerged in Brazil and British North America in the eighteenth century as settlement expanded westward away from coastal areas. After the abolition of the transatlantic slave trade, such overland trafficking peaked as domestic slave trades replaced the transatlantic traffic, especially in the nineteenth-century United States and Brazil.

Such overland movements are crucial to understanding captive experiences and the formation and dislocation of enslaved populations, but they are not easily tracked using a shipment-based model because governments rarely tracked and logged overland movements of people and goods as systematically as they monitored the departures and arrivals of ships. The 2021 expansion of the Intra-American Database includes substantial coverage of We certainly intend to incorporate maritime slave voyages within the United States and Brazil, and we anticipate additional research on maritime trafficking in these countries, whether riverine or coastwise, enhancing the database in coming years. We also intend to add data on overland migrations where records exist to document “slave coffles” or other overland migrations in a format compatible with the “voyage” structure of the database. For Brazil this seems feasible, because the law required slave traders to register with authorities in Rio de Janeiro before marching inland (creating documents akin to port records), leaders of the caravans were appointed (akin to ship captains), and these forced marches tended to follow repeated routes and pass through checkpoints making them traceable. For the domestic traffic in the nineteenth-century United States, it is less clear that systematic records will allow for the creation of such an overland voyage-based dataset. Scholars such as Michael Tadman, who have worked to quantify the U.S. domestic slave trade, have relied on different methodologies to estimate the scale of the forced migration—such as tracking changes in census data. That methodology has provided crucial insight on the U.S. traffic, but is not well suited to a database of voyages.³

In addition to recognizing gaps in the data that this database can provide, it is also crucial for users to recognize the limited perspective that any database can provide. As Hans Rosling notes, “the world cannot be understood without numbers, and it also cannot be understood with numbers alone.”⁴ Here, we intend to advance the quantitative and qualitative examination of the intra-American slave trade without objectifying all over again the people forcibly trafficked as commodities. We operate from a belief that compiling quantitative information in such a database leads to discoveries about the direction, routes, and volume of various branches of the traffic; aids the development of studies of culture and language in the African diaspora; and reveals the overwhelming importance and ubiquity of slave trading in the New World. For those reasons, we think compiling such a database is a worthwhile undertaking, but we also remain painfully aware that the sources that make such quantitative analysis possible are dehumanizing, tallying people in ledgers as trade goods alongside barrels of sugar and crates of textiles. We encourage all users of the database to also consult other types of information about the slave trade—such as survivors’ accounts, cultural studies, and more qualitative analyses—and we welcome suggestions for how to give this database project a more human touch that highlights enslaved people’s perspectives despite the source-based challenges.

Classification as an Intra-American Slaving Voyage

For the purposes of this database, an intra-American slave voyage is defined as a shipment that embarked enslaved people anywhere within the Americas and disembarked them at another site in the New World. Slave voyages that initially embarked captives in Africa are located in the Transatlantic Slave Trade Database instead, even if those transatlantic voyages disembarked captives at more than one port in the Americas (or stopped at multiple American ports of call before disembarking captives). Such continuations of transatlantic voyages to multiple American ports are not included in the intra-American data to avoid overlap with the transatlantic database.

The distinction between intra-American and transatlantic voyages is not always simple. For instance, some records documenting arrivals of enslaved people at a given American port fail to indicate ports of origin for the voyages, leaving researchers to guess whether a voyage originated in Africa or in the Americas. This was a particular problem for the Cuban records from archives in Spain that Herbert Klein consulted to compile his set of voyages to Havana, 1790 to 1820. The editors of the original Transatlantic Slave Trade Database (1999) decided to use a benchmark of 100 captives to predict whether a voyage was coming from Africa or elsewhere in the Americas, since transatlantic slavers rarely found it profitable to cross the ocean with small groups of captives. Hence, the original transatlantic dataset included all of Klein’s documented voyages that carried 100 or more captives, and when O’Malley began his intra-American dataset he adopted the same methodology, incorporating all of Klein’s voyages with fewer than 100 captives. Recently, however, and in combination with the making of this Intra-American database, Jorge Felipe has parsed out intra-American and transatlantic voyages for Cuba between 1790 and 1820 more definitively, by using shipping data from Havana’s newspapers, which include the previously missing information on ports of origin. This more definitive differentiation between intra-American and transatlantic voyages shows that the method of using the number of captives to predict whether voyages were transatlantic or intra-American was imperfect.

Thankfully, Felipe’s research resolved the big problem of disentangling intra-American and transatlantic voyages for Cuban arrivals, but other grey areas persist. For the early seventeenth century traffic to Buenos Aires, for example, it is difficult to differentiate transatlantic from intra-American slave voyages, given that it is sometimes impossible to know if slave voyages departing Rio de Janeiro or Salvador to the Río de la Plata were continuations of transatlantic voyages or voyages that embarked captives in the Americas. Some voyages were indeed continuations of transatlantic ventures initially departing from West-Central Africa (and thus they are not included in the intra-American data but rather in the transatlantic dataset). Meanwhile, other ships that connected Rio de Janeiro and Buenos Aires had not previously crossed the Atlantic (and thus are included in the intra-American data).

Even more confounding, the South Sea Company often operated voyages from Jamaica to Spanish America that mixed both transatlantic and intra-American slave trading. Company vessels sometimes ferried people across the Atlantic from Africa to Jamaica, offloading a portion of the captives in Kingston and embarking others from the island in their place. Thus, for some captives on board, the voyage from Jamaica to Spanish America was a continuation of their transatlantic journey, but other captives experienced the same shipment as a discrete intra-American journey. In such cases, we logged the voyage in the intra-American database, but only counted the people who embarked in Jamaica to avoid double-counting with the record of the same vessel in the transatlantic database.

To the extent possible, we have avoided duplications between the intra-American the transatlantic databases, but given the incomplete nature of the records and the entangled nature of the transatlantic and intra-American slave trades, some voyages listed in the intra-American dataset may have been continuations of Atlantic crossings.

Nature of the Sources

This database, though newly consolidated, reflects the work of dozens of scholars working independently over the past several decades. Only their generosity in sharing data—either by sending it directly to the project team, publishing datasets, or donating them to archives—makes such an accumulation of information on a hemispheric scale possible.

The foundation for the current intra-American database was a dataset of 7,685 voyages compiled by O’Malley. The bulk of that information derived from O’Malley’s research in the Naval Office Shipping Lists—port records from nearly all of Britain’s colonies in North America and the Caribbean, which are housed in the British National Archives and available on microfilm at several American research libraries. Such systematic records were supplemented by references in newspapers, merchants’ papers, imperial correspondence, and secondary sources. In addition to O’Malley’s own research, his initial database incorporated several datasets compiled by other scholars. This included published shipment lists by Walter Minchinton (on North Carolina), Darold Wax (on Pennsylvania), and Elizabeth Donnan (on a variety of locations); datasets donated to public repositories by Lorena Walsh (on the Chesapeake) and Herbert Klein (on Cuba); and datasets and research notes shared directly by David Eltis (assorted) and Philip Morgan (on South Carolina). Funding support from the National Endowment for the Humanities facilitated the reformatting of O’Malley’s dataset for integration with the slavevoyages.org website, and paved the way for the incorporation of additional datasets.

The first of these additions came from the research of Jorge Felipe on the slave trade to Cuba. Drawing on Havana’s newspapers, Felipe’s work added more than 500 new voyages to the database, and improved the information on 1,500 voyages already documented from Klein’s dataset on Cuba and O’Malley’s work in British sources. (Felipe’s work also improved the information on direct African arrivals to Cuba in the transatlantic database.) Felipe also supplemented his work on newspapers with colonial reports on Santiago de Cuba housed in the AGI in Seville, which were provided by José Belmonte. Another dataset of the slave trade to the Spanish Caribbean, based on AGI documents, was gathered by Oscar Grandío for the work of Eltis and Paul Lachance, and was also merged into this database.

Further additions came from Borucki, who compiled data on slave arrivals in the Río de la Plata region—some 550 voyages for the period 1777-1812, from port records; and some 250 seventeenth-century voyages, mostly from secondary sources. (Since the seventeenth-century traffic in this region remains under research by other scholars, additional voyages will surely surface for inclusion in future iterations of the database.)⁵ Borucki also employed Spanish records both from Venezuela and Seville, to create a dataset of some 260 intra-American voyages to Venezuela, which of course only partially documents the internal traffic to this region.

Dutch data gathered by a team of researchers led by Jelmer Vos is key to understanding the slave trade to the Spanish colonies ranging from Venezuela to Mexico from the late-seventeenth century to the mid-eighteenth century. Drawing on records housed in Amsterdam and The Hague, Vos’s team provided entries for some 1,400 slave voyages departing Curaçao for various Caribbean islands and the Spanish American mainland, as well as smaller datasets of about 240 voyages from St. Eustatius and St. Maarten.

Another addition to the initial version of the intra-American database on the SlaveVoyages website in 2019 came from Daniel Domingues, who documented 121 voyages to Maranhão and Pará from elsewhere in Brazil, using the Portuguese Overseas Historical Archive, covering the period from 1778 to 1797.

Thanks to a Digital Extension Grant from the American Council of Learned Societies, a major expansion of the database was completed in late 2021 and early 2022. One major piece of this expansion was Jennie K. Williams’s contribution of data on more than 3,200 voyages compiled from shipping manifests for her dissertation, “Oceans of Kinfolk: The Coastwise Traffic of Enslaved People to New Orleans, 1820-1860.” These voyages moved tens of thousands of enslaved people, almost all of them American-born, from elsewhere in the United States to New Orleans in the antebellum era. One exciting aspect of this research is that the manifests examined by Williams include the names and other basic biographical information for the enslaved people transported. (Shipping data from prior centuries almost always omitted such identifying information about captives.) Names and other information about these enslaved individuals is available in the Oceans of Kinfolk Database, which is part of the new People of the Atlantic Slave Trade database.

A second outcome of the Digital Extension Grant was the addition of new slave voyages to the Spanish Americas in the eighteenth century, and the revision of many existing entries for Spanish America with more complete information. The work, spearheaded by Alex Borucki began (and will continue) with research on digitized sources in the collection “Blacks and Slaves” from the AGN Colombia (the Colombian National Archives) with one hundred voyages carrying captives mostly from the British Caribbean to Cartagena and Portobelo between 1740 and 1780. In addition, another set of nearly one hundred voyages were provided by José Belmonte on slave arrivals in Cartagena and Portobelo as well, between 1791 and 1797. This list was produced by colonial authorities in Cartagena to be sent to Spain, and that is why this list is located in Seville.

The third major expansion supported by the ACLS Digital Extension Grant was the addition of a dataset of more than 12,000 voyages in the Brazilian domestic slave trade that delivered enslaved people to Rio de Janeiro between 1831 and 1860. This research initiative in Rio’s newspapers was led by Daniel Domingues at Rice University and documents the massive scale of coastwise trafficking in Brazil in the period after transatlantic deliveries of enslaved Africans were gradually suppressed. This dataset should continue to grow with research on newspapers from the 1860s, 1870s, and 1880s.

Whereas many port records, newspapers, and merchants’ accounts were mined effectively to create entries in the database, some other types of sources, such as Spanish notary records, present problems for the organization of the database around individual voyages. Some official reports and notary records offer a precise total figure for people arriving during a given period (such as a year), rather than data on arrivals of individual vessels carrying enslaved people. Similarly, documents from Venezuela that record slave owners’ purchases of captives from Dutch traders are specific at the level of the individual about African arrivals, but do not include shipping data. Likewise, in the case of early-seventeenth century Buenos Aires, some lists from the Royal Treasury enumerate captives by the person responsible for their introduction, rather than the ship that delivered them. All such records contain useful information, but not in a format that allows for the entry of a voyage in the current database structure.

Contraband trade poses another major problem for the voyage-based model. In the case of Spanish America, Borucki, Eltis, and David Wheat provide an estimate of the overall internal traffic that also accounts for contraband to the Spanish Americas, particularly during the pre-1640 era.⁶ In addition to this article, please take a look at these two documents:

Notes on the Estimates of the Intra-American Slave trade to the Spanish Americas

Origins of slaves entering Spanish Americas via Intra-American Slave trade

Serial records of individual shipments are also rare for the slave trade to French colonies—especially Martinique and Guadeloupe, which saw less frequent arrivals than Saint-Domingue of slave ships directly from Africa. French imperial officials barred non-French traders from their colonies until the late-eighteenth century, but anecdotal references from both the British and French Caribbean suggest that illegal foreign trafficking accounted for perhaps half of the enslaved African people arriving in Martinique and Guadeloupe. Yet the clandestine nature of this trade means that systematic records of individual shipments do not survive. As a result, the intra-American traffic to the French Caribbean is vastly underrepresented in this database and smuggling along other routes is poorly documented as well.⁷

As in the transatlantic database, each voyage has at least one source of documentation, and many have two, or more. Occasionally, these multiple sources contradict each other on exact dates, names, places, or numbers of captives. This is unsurprising, given that some sources record departures of ships, while others document arrivals. We have not eliminated all inconsistencies in the records, for which curious users are invited to examine the archives and evidence themselves. To guard against double counting of voyages found in multiple sources, all new entries in this database were cross-referenced against the other intra-American entries and against the transatlantic database (by vessel name, captain’s name, and voyage itinerary).

See the entire list of sources of both intra-American and transatlantic slave voyages.

Data Variables

The intra-American database uses mostly the same data variables as the transatlantic dataset, but includes a few additional ones. These additional intra-American variables are not visible through the online user-interface, however. Rather, they are only accessible when a user downloads the complete database here.

The first of these variables, “VoyagelD2,” is used when captives in the intra-American traffic can be traced to a particular transatlantic voyage (i.e., the preceding stage in their forced migration). In such cases, we enter the ID number of the transatlantic voyage from the transatlantic database in the record of the captives’ intra-American voyage to allow users to trace the multi-stage journey of a group of captives. Another variable specific to the intra-American database is “AfrInfo,” which records information about the origins of the captives—for instance if they were noted as coming from a particular African coastal region, such as the Gold Coast or from a specific cultural group, such as Igbo, or if they were born in the Americas. Finally, the variable “other_Cargo” is used to record available information on any goods transported alongside the captives.

Scholars interested in such African connections, or the ties between the slave trade and other commerce, are thus encouraged to download the database to examine these additional variables offline.

Age Categories

The demographics of captives selected for transshipment within the Americas did not necessarily mirror the larger population of transatlantic captives from which they were chosen. At times these populations were quite similar, but in many cases, the people transshipped within an empire were those deemed least desirable by the planters and speculators buying slaves in an entrepôt. Thus, descriptive records describe more children, older men, women, and so-called “refuse” slaves (the sick or disabled) in the internal trade, for example, within or between British colonies.⁸

Conversely, trade across imperial boundaries, particularly to Spanish colonies, reveals that a combination of high prices and available specie (silver pesos) could reverse the dynamic. Non-Spanish merchants in Jamaica and Curaçao were willing to speculate on relatively large groups of captives whom they deemed especially valuable for transshipment to Spanish colonies, especially on the mainland. As a result, adult males of prime working age were more likely on such routes.

Unfortunately, many port records did not differentiate captives by age or sex, so such patterns cannot be analyzed with this database in all cases. When such data are available, it is recorded in the database in five age and sex categories: women, men, girls, boys, and infants (undifferentiated by sex because that was typically true of the available sources). Such age and sex data was typically lacking in British records; it is most complete for the slave trade to Cuba.

For domestic trafficking within the United States, demographic data was handled differently because records of that traffic were unique in providing the names of enslaved individuals moved, along with information about the age and sex of such individuals. Instead of grouping such individuals into arbitrary subjective categories that distinguished men, women, boys, and girls, researchers captured the more individual data on exact age and sex for each person. That demographic and personal information for the U.S. domestic traffic can be explored through theOceans of Kinfolk Database.

Dates

Intra-American slave voyages were typically shorter than transatlantic crossings. To give the most extreme example, a sail of only a day or two separated some Caribbean islands. On the other end of the spectrum, however, some intra-American voyages were comparable to Atlantic crossings in length. Voyages from the Río de la Plata to Valparaíso (Chile), for example, could take an entire month through the Magellan straits.

In cases of trans-imperial voyages, records most commonly survive to document either the departure or arrival end of a slave voyage, but not both. This partial data makes it difficult to ascertain voyage durations in some cases. As such, when individual vessels appear multiple times in a set of records, it can be difficult to determine whether records contain multiple references to a single voyage or whether a single vessel conducted multiple slave-trading voyages to or from the same port, with separate groups of captives, within a relatively short span of time. In such cases, we used average voyage durations and other circumstantial evidence to make the best determination possible about whether a record indicated a new voyage for the database or just an additional reference to a previously documented voyage.

Names

Names appear in their original language, where possible, and we have made efforts to systematize names in English, Spanish, Dutch, French, and Portuguese, where multiple records rendered those names differently. This includes changing English renditions of Spanish names and vice versa for captains, and also, finding the complete names of ships when some sources recorded only partial names (e.g. “Brothers” versus “Three Brothers”). In some instances, such standardization is easy as the “Favorite” becomes “Favorita,” but, in other cases, sources that rendered names in languages other than the person or ship’s original seem to reflect attempts at phonetic transcriptions and in some cases the names given for the same ship or person in different sources have no apparent connection to each other.

Given the dubious legality of some trans-imperial trade, one explanation for the variance of names across sources is that ships often used aliases in foreign ports. Other name discrepancies stemmed from simple misunderstandings or errors of transcription. Smugglers (which is not a clear-cut category) often changed the names of both captains and vessels as they crossed imperial borders, so we have tried to identify vessels that appear different by the names to avoid double-counting of voyages, but if contraband traders could fool their contemporary port authorities, surely they sometimes fool us.

In the case of Spanish ships sailing within the Caribbean, and between Brazil and the Río de la Plata, the usage of two or three Catholic saints and figures of devotion in the ship’s name complicates the process of identification if other variables, such as captain's name, are missing. For example, “Nuestra Señora del Carmen y Buen Jesús,” means that this ship could have been recorded as “Nuestra Señora del Carmen,” and in other times, simply as “Buen Jesús.” Likewise, certain names appeared as ship names (and captain’s names) repeatedly, leading to some confusion, but we made our best effort to use rig type, tonnage, the number of captives on board and other circumstantial evidence to clarify between multiple references to single voyages and references to distinct ventures featuring similar names.

Where multiple names (that were not mere translations or abbreviations) appeared in different sources to refer to the same vessel or person, we employed a convention of listing both names in the appropriate variable field, separated by “(a)” meaning alternatively or AKA. For example, voyage #100272 shows the captain as “Fleming (a) Heming, James” because two different sources rendered the last name differently but all other evidence points to these being two records of the same voyage.

Imputed Variables: Numbers of Slaves and Mortality

Most imputed variables in the Intra-American Slave Trade Database use the same assumptions as the Transatlantic Slave Trade Database. However, not all assumptions from the transatlantic database are applicable to the intra-American traffic given that intra-American slave voyages were usually shorter in distance and duration and were conducted in ships that often carried captives and merchandise together. As with the transatlantic database, users can download the intra-American database and create their own imputed variables from the data variables displayed.

The different organization of the intra-American slave trade required two main departures from the methodology of the transatlantic database with regards to imputed variables. First, because intra-American traders targeted markets of widely varying size (and often mixed slave trading with other commerce), imputing the numbers of captives aboard documented voyages requires greater attention to variations by trade route than is necessary for the transatlantic trade. Full cargoes of one or two hundred enslaved women and men were common in the intra-American traffic, but so were shipments with just a few enslaved people on board.

We have followed the methodology of O’Malley for estimating the numbers of captives aboard slave ships for which the precise number of enslaved people went unrecorded: 

The second main departure from the transatlantic database’s methodology for imputed variables is required for assessing mortality. Estimating mortality aboard voyages for which the number of captives is documented only at embarkation or disembarkation (but not both) is complicated by numerous factors. For one, the data for more fully documented voyages suggests that there was no “typical” experience for intra-American slave ships with regard to mortality; experiences ranged widely, with some groups of captives suffering devastating outbreaks of contagious disease or deadly maritime accidents, while many voyages saw no deaths among the captives at all.

Nonetheless, to allow users to develop an aggregate picture of the slave trade along various intra-American routes, it is useful to estimate both a number of captives embarked and disembarked for each voyage in the database. These imputed numbers of captives appear in the variables SLAXIMP (embarked) and SLAMIMP (disembarked). The data are derived from the documented numbers of captives embarked or disembarked on a vessel whenever possible, but typically our sources only reveal the number of people embarked or disembarked, not both. To approximate the number of people on board at the undocumented end of such voyages, we ran a linear regression analysis on the 655 cases in the intra-American database for which the number of deaths on board was either reported, or could be estimated with some confidence from cases in which our records document both the number of captives embarked and disembarked. From that analysis, the following formula was derived to impute the number of captives disembarked from voyages for which only the number of captives embarked had been determined:

SLAMIMP = 0.316 + .953(SLAXIMP)

And the following formula was derived to impute the number of captives embarked on voyages for which only the number of captives disembarked had been determined:

SLAXIMP = 0.757+1.015(SLAMIMP)

In the rare cases in which we found records indicating a larger number of people disembarking from a vessel than other records indicated had embarked on the vessel, we also used the methodology described above to impute a plausible number of people embarked on the voyage. The data variables for reporting numbers of captives (e.g. tslavesd, slaarriv, slas32) present the data exactly as we found it—discrepancies included. But in the imputed variables (slaximp and slamimp) it seemed appropriate to reconcile the data. Because under-reporting seems more likely then over-reporting in the surviving records (due to both tax evasion and the possibility of unreported stops at additional ports) we opted to elevate the number of captives embarking on a vessel to account for the number of people reported disembarking from that voyage at the other end.

Geographic Data

One big advantage to adding intra-American coverage to slavevoyages.org is that the combined intra-American and transatlantic datasets offer a picture of the African diaspora more inclusive of all of the Americas. The intra-American database adds much more information about movements to regions of the Americas that saw few enslaved people arrive directly from Africa. Thus, the Intra-American database adds places previously not listed by the Transatlantic Slave Trade Database in North, Central, and South America. In adding these new places, the Intra-American Slave Trade Database reveals the overwhelming importance and ubiquity of slave trading throughout the New World.

This expanded geographical coverage does pose challenges for the website’s existing structure for grouping ports into regions, however. The transatlantic database offers great detail for parts of the United States and the British Caribbean, but one finds less detail (and less geographic specificity) for non-British areas such as the Dutch Caribbean, and particularly Spanish America, where the single regional category “Spanish Circum-Caribbean” lumps together the Spanish territories from North America (Mexico), Central America (from Guatemala to Panama), and South America (Colombia and Venezuela). (Meanwhile, regional variables for North America are much more specific, corresponding to individual states, such as Maryland or Georgia.) Note that these geographical divisions within the Spanish Americas do not correspond to the district divisions under Spanish colonialism but rather lump Spanish American territories together regardless of administrative divisions. Those wishing to use alternative groupings of ports and places into regions and broad regions may download the database and use the geographic appendix to the SPSS codebook to create their own regional categories.

Geopolitical groupings should be cautiously examined for some Caribbean and Circum-Caribbean places such as Jamaica and Trinidad, as well as Honduras, among other places, as these islands and mainland regions passed from one European empire to another over time.

In virtually all cases, the Intra-American Slave Trade Database provides data on where slaves embarked or disembarked, and in most cases such geographic information is available for both embarkation and disembarkation. Where a specific location is only known for one end of the voyage, imputed data on embarkation or disembarkation of captives depends, frequently, on what was the flag of the ship (English, Dutch, etc.). For instance, English and Dutch records of departure provide precise information of slave embarkation in the British and Dutch Caribbean, but often less precise data on where these vessels planned to disembark captives (e.g., the records noted a vessel was bound for Spanish America, but did not specify which port or colony). Spanish records on disembarkation of captives usually suffer the opposite problem, providing less precise data on where captives had embarked. In most cases, we have not attempted to guess where vessels were headed, but rather to populate data fields with information of greater or lesser specificity, as the sources allow. For example, a vessel known to have disembarked captives in Havana, will list “Havana” as the port of disembarkation, but if a vessel was only known to have delivered enslaved people to Cuba, we do not assume Havana (or any other specific port) as the place of landing. Instead, we use more generic codes, such as “Cuba, port unspecified.”

Dutch records of the intra-American trade often give vague descriptions of the destinations of slave ships sailing from Curaçao and other Dutch islands. When the source indicates “Upper Coast” or “Coast” as intended destination, we infer the place of disembarkation to be in the “Spanish Circum-Caribbean, unspecified,” given that all mainland colonies where the Dutch commonly disembarked slaves were Spanish. There might be some Dutch voyages going to North America, but relative to the more common voyages to the Spanish Main, such voyages were negligible. When Dutch sources refer to “West” as intended destination, we impute “Spanish Americas, port unspecified” (code 80499) as place of disembarkation, since it is unclear whether these ships headed to the Spanish mainland or one of the Spanish islands. When we are uncertain whether a ship intended to sail to the Spanish colonies at all, we impute “Americas, port unspecified” (code 80299) as intended place of disembarkation.

The Intra-American dataset rarely provides data (imputed or otherwise) on “homeport,” or where a slaving vessel continued after disembarking captives. The transatlantic database records such information because so many vessels completed clearly defined circuits (such as the “triangle trade”), with one leg of that circuit routinely carrying slaves. Given the multi-lateral commercial routes within the Americas, sometimes across imperial borders, it is difficult to identify clear starting and ending points of most intra-American slave ships’ circuits. Most commonly, the Intra-American dataset only provides data on ports of slave embarkation and disembarkation, without recording the prior and subsequent movements of vessels. (In the rare cases where we do have such data, we use the same variables as the transatlantic database to record it.)

National Carriers

Carriers of captives within a colonial empire or independent nation within the Americas were almost always from the said empire or nation, making identifying the flag of the ship unproblematic in such cases. The slave trade across colonial borders, however, raises difficulties for ascribing national identities to ships, given that smugglers often carried false papers and multiple flags.

Both the English Company of Royal Adventurers, and its successor, the Royal African Company (RAC), usually refused to deliver slaves to Spanish colonies, though they did sell them to all-comers from their factories in Jamaica and Barbados. Beginning at least as early as 1661, Spanish merchants carried off purchases from Kingston and Bridgetown in their own vessels both before and after Grillo and Lomelin, the first slave trade contractors (asientistas) of the post-1640 system, began their activities. Meanwhile, English smugglers operating outside the RAC’s monopoly also carried slaves to Spanish ports. After the mid-1690s, as English Caribbean slave entrepôts gradually became the dominant source of enslaved people for the nearby Spanish colonies, Spanish participation in the slave trade fell away. When first the Portuguese and then the French assumed the asiento between 1694 and 1713, they drew on English ports and Curaçao to acquire African men and women for delivery to Spanish America without using Spanish intermediaries. For most of the 1713 to 1739 period the South Sea Company could bring slaves into Spanish ports legally, but smugglers of various nationalities continued to operate as well. Dutch merchants dominated the slave traffic through Curaçao to Spanish colonies (though Hispanic slave traders were certainly involved) and the Portuguese played a similar role in the traffic from Brazilian ports to the Río de la Plata from 1585 through to 1777. Thereafter Spanish merchants based in the Río de la Plata came close to sharing the traffic equally with Luso-Brazilian slave traders.¹⁰

In times of war between Britain and Spain, for most of the 1790-1808 period, U.S. slave traders operated in the transatlantic traffic to Spanish colonies from Cuba to the Río de la Plata, and they participated in the intra-Caribbean traffic, as well, sometimes by using neutral flags, such as the Danish and Swedish, to avoid capture either from British or French privateers. The importance of Danish slave vessels in the internal traffic to the Spanish Caribbean from 1791 to 1810 also involved U.S. shippers using the Danish flag, for which the nationality of these vessels is somewhat disputed.

We have recorded the nationalities of vessels as recorded in official records unless we had definitive information to indicate that a vessel was flying a false flag. Given the extent of trans-imperial smuggling, researchers should approach such classifications of nationality with caution.

Tonnage

The sizes and shapes of vessels conducting intra-American slaving voyages varied greatly, even more than those plying the transatlantic slave trade, because any ship could carry one or two captives alongside merchandise between neighboring Caribbean islands or along an American coastline. Some routes could avoid open water and high seas. In transshipments within an empire, traders typically sent small numbers of captives in mixed cargoes from major ports of the slave trade to regions where slave labor was less in demand or the economy was underdeveloped. Vessels crossing imperial boundaries, by contrast, tended to be larger. They often carried nothing but captives, and sold them in prosperous ports with high demand for slaves. These receiving ports in the trans-imperial traffic typically failed to draw direct shipments of captives from Africa for geopolitical reasons.

Resistance

The intra-American dataset offers the same variables as the transatlantic database to indicate forms of resistance in slave voyages. Amistad is one of the best known slave ships in the history of the Atlantic slave trade, due to the rebellion on board and its popularization in a major motion picture and several books. This revolt actually occurred on an intra-American voyage, so the Amistad is included in this dataset (Voyage ID #112137). The two other best-known shipboard slave revolts—the Tryal (#112135; made famous by Herman Melville’s novel Benito Cereno) and the San Juan Nepomuceno (#112136)—also took place on vessels sailing between American ports rather than voyages across the Atlantic. That was also the case of a slave rebellion in the U.S. ship Creole in 1841 (#113035), as rebels diverted the ship to freedom in the Bahamas, where slavery had been outlawed.¹¹

We hope that incorporation of new data and its examination in this database will offer new light on the patterns of slave resistance in the intra-American slave trade. It is worth noting, however, that aside from these high-profile rebellions far fewer rebellions are documented for the intra-American traffic than the transatlantic trade. This may be a shortcoming of the sources available, or it may reflect some of the challenges for resistance at the intra-American stage, as captives where often physically debilitated from the brutal and disease-ridden conditions in the transatlantic crossing and any layover in an American port.¹² The frequency of rebellion in the intra-American traffic is an aspect of the slave trade that is ripe for further research.

Owners

Since the majority of voyages in the Intra-American Database were documented from port records of various kinds, which reported on the comings and goings of ships, most people or firms categorized as “owners” in the database were the owners of the ships that carried enslaved people. In many cases, such owners of ships also claimed ownership of the enslaved people on board, having speculated on purchases of enslaved people for transport on their own vessels in hopes of profitable resale in another port. In some cases, however, owners and traders of enslaved people paid freight charges to ship owners for the transport of people to an intended port of sale. Where records identify such shippers of enslaved people aboard ships owned by others, the shippers are listed on our “owner” variables as well. Often, surviving records do not make clear whether a given owner claimed possession of a ship, the enslaved people on that ship, or both, so the Intra-American Slave Trade Database does not differentiate between these types of ownership in classifying people in the “Owner” variables. However, records of the U.S. domestic slave trade often did distinguish between “owners,” “shippers,” and “consignors” for voyages that moved enslaved people. All three categories of people who profited from trafficking enslaved people are lumped together in the Intra-American Slave Trade Database’s “Owner” variables, but users can access the more refined information about such traffickers’ roles as owners, shippers, and/or consignors in the People of the Atlantic Slave Trade database.

Notes

¹ See, for example, Gregory E. O’Malley, Final Passages: The Intercolonial Slave Trade of British America, 1619-1807 (Chapel Hill: Omohundro Institute and University of North Carolina Press, 2014); Alex Borucki, David Eltis, and David Wheat, “Atlantic History and the Slave Trade to Spanish America,” American Historical Review 120, no. 2 (April 2015): 433-461.

² See, for example, Steven Deyle, Carry Me Back: The Domestic Slave Trade in American Life (New York: Oxford University Press, 2005); Walter Johnson, ed., The Chattel Principle: Internal Slave Trades in the Americas (New Haven: Yale University Press, 2004) Walter Johnson, Soul by Soul: Life Inside the Antebellum Slave Market (Cambridge, MA: Harvard University Press, 1999); Michael Tadman, Speculators and Slaves: Masters, Traders, and Slaves in the Old South (Madison: University of Wisconsin Press, 1989).

³ Michael Tadman, Speculators and Slaves: Masters, Traders, and Slaves in the Old South (Madison: University of Wisconsin Press, 1989).

⁴ Hans Rosling, Ola Rosling, and Anna Rosling Rönnlund, Factfulness: Ten Reasons We’re Wrong about the World--and Why Things Are Better Than You Think (New York: Flatiron Books, 2018), 192.

⁵ For the earliest slave trade to the Río de la Plata, see the ongoing work of Kara Schultz.

⁶ Alex Borucki, David Eltis, and David Wheat, "Atlantic History and the Slave Trade to Spanish America,” American Historical Review 120, no. 2 (April 2015): 433-461.

⁷ Kenneth J. Banks, “Official Duplicity: The Illicit Slave Trade of Martinique, 1713-1763,” in The Atlantic Economy during the Seventeenth and Eighteenth Centuries, ed. Peter A. Coclanis (Columbia: University of South Carolina Press, 2005), 229-251; Gregory E. O’Malley, Final Passages: The Intercolonial Slave Trade of British America, 1619-1807 (Chapel Hill: Omohundro Institute and University of North Carolina Press, 2014).

⁸ Gregory E. O’Malley, Final Passages: The Intercolonial Slave Trade of British America, 1619-1807 (Chapel Hill: Omohundro Institute and University of North Carolina Press, 2014), see index entries for gender and slave trade patterns, as well as for children.

⁹ Gregory E. O’Malley, Final Passages: The Intercolonial Slave Trade of British America, 1619-1807 (Chapel Hill: Omohundro Institute and Unversity of North Carolina Press, 2014), 352.

¹⁰ Alex Borucki, David Eltis, and David Wheat, "Atlantic History and the Slave Trade to Spanish America," American Historical Review 120, no. 2 (April 2015): 450-452.

¹¹ On the Amistad, see Howard Jones, Mutiny on the Amistad: The Saga of a Slave Revolt and Its Impact on American Abolition, Law, and Diplomacy (New York: Oxford University Press, 1997). On the Tryal and the San Juan, see Greg Grandin, The Empire of Necessity: Slavery, Freedom, and Deception in the New World (New York: Metropolitan Books, Henry Holt and Company, 2014). On the Creole, see Arthur T. Downey, The Creole Affair: The Slave Rebellion That Led the U.S. and Great Britain to the Brink of War (Lanham, MD: Rowman & Littlefield, 2014).

¹² On lack of evidence of rebellions in intra-American slave voyages, see Gregory E. O’Malley, Final Passages: The Intercolonial Slave Trade of British America, 1619-1807 (Chapel Hill: Omohundro Institute and University of North Carolina Press, 2014), 81-83.