David Eltis, Philip Morgan, and David Richardson

Agency and Diaspora in Atlantic History: Reassessing the African Contribution to Rice Cultivation in the Americas

Newspaper advertisement, April 26, 1760, for the sale of slaves at Ashley Ferry, near Charlestown, South Carolina. Library of Congress, Prints and Photographs Division, Washington, D.C. Reproduction number: LC-USZ6210293 (b&w film copy neg.).

Broadly speaking, two contrasting models dominate interpretations of Atlantic history. One draws on Old World influences to explain the nature of societies and cultures in the Americas, while the other assigns primacy to the New World environment. One stresses continuities, the other change. The polar extremes are persistence and transience, inheritance and experience. An emphasis on inheritance prioritizes the cultural baggage that migrants brought with them, whereas a focus on experience highlights the physical and social environments, such as climate, natural resources, and settlement processes, that they encountered. In modern parlance, one approach focuses on folkways, the other on factor endowments.

In the late nineteenth and early twentieth centuries, these two viewpoints clashed, and the debate still reverberates in modified form. An emphasis on cultural continuities was the preserve of germ theory historians, such as Herbert Baxter Adams and Edward Eggleston, who stressed what immigrants from Europe brought with them when they crossed the Atlantic. Frederick Jackson Turner most famously challenged this emphasis, arguing that an egalitarian civil society and political democracy were rooted in the expanding frontier and availability of land in temperate North America. In the course of the twentieth century, the frontier thesis gathered considerable strength. Although historians of migration no longer mention the Turner school, the new environment continues to be seen as the dominant influence, whether in terms of physical resources or the evolution of new social identities. In the Black Atlantic, the frontier thesis might seem irrelevant, but there, too, the literature on creolization, stemming most notably from the work of Sidney Mintz and Richard Price, saw the historiographical pendulum swing toward an emphasis on the discontinuity of the transatlantic experience and the critical importance of the New World environment. ¹

For students of European migration, the dichotomy reemerged when David Hackett Fischer insisted on the primacy of four British regions and their connection to four "cultural hearths" in British America, describing his argument in Albion's Seed as a "modified `germ thesis.'" His is a model of cultural diffusion. He termed it an ethnocultural exercise, involving the creation of a taxonomy of customs, ideas, and institutions—or folkways. It requires the identification of a place of origin, a place of destination, and a path between the two. It might be labeled a "transfer-conduit" or "donor-recipient" model. The principal agency of transmission, Fischer argued, was an influential immigrant elite that systematically endeavored to create a regional culture in the Americas in the image of the homeland hearth from which its members had migrated. He mentioned the "iron law of oligarchy," meaning that "small groups dominate every cultural system." ²

Fischer's diffusion model aroused much criticism, which centered most fundamentally, perhaps, on his failure to explore fully how complex forces in the New World modified the transmitted four regional cultures. Powerful elements in the settlement process—such as material conditions, interactions with other peoples, and environmental possibilities and constraints—shaped and reformulated the inherited regional cultures. Much selection and adjustment then ensued. This whole process seemed undertheorized and underrepresented in Fischer's lengthy book. It was also constantly changing, his critics alleged; by contrast, they argued, an emphasis on cultural persistence tended to create static and synchronic pictures. ³

Fischer astutely responded that these polar approaches present false choices. He noted that at one end of a broad spectrum is the stereotypical germ theory, which gives no causal importance to the environment; at the other is an exaggerated frontier thesis, which allows no role for culture. Neither is useful, he maintained. The difficulty is mediating between these two extremes. "To make a strong case for culture is not to deny or even diminish the role of environment," he said. "The great question is about the linkage of culture and environment." No simple dichotomizing method is likely to be helpful. The open environment of the New World, he emphasized, should be seen as "a fertile seedbed" for many different cultures, nourishing a dynamic pluralism more complex than any Old World society had ever been. The goal, he now maintained, was to develop a cultural model that would coexist with environmental or ecocultural history. Whether Albion's Seed quite fulfilled these prescriptions is doubtful, but Fischer proved a supple defender of a complex cultural mode of analysis. ⁴

In some ways, historians of the African American experience have been reprising the debates that raged over Fischer's book. A number of works have made claims consistent with Albion's Seed. Michael Gomez's Exchanging Our Country Marks, for example, discovers a number of "African ethnic enclaves" in North America. Thus the Chesapeake was the "preserve of the Igbo," West Central Africans were particularly numerous in the lowcountry, and "Bambaras" from Senegambia were "foundational" in Louisiana. Although Gomez argues that Africans and their descendants in North America eventually embraced a collective identity based on race rather than ethnicity—thus emphasizing the role of the New World environment—he sees the persistence of African cultures well into the early nineteenth century. Another study in a similar vein is Gwendolyn Midlo Hall's Slavery and African Ethnicities in the Americas, which also argues that distinct African regional cultures and ethnicities made "major contributions" in "particular places in the Americas." She emphasizes that most of the Africans who were brought to the New World came from only a few ethnic groups, that they clustered in various parts of the Americas, and that the earliest Africans from specific regional cultures often had "a continuing and decisive influence" on their native-born descendants and on those Africans who arrived later. ⁵

A tension regarding the question of numbers is present in many such studies. To establish the role of particular African ethnicities in the Americas, the aim is usually to point to majority influence. In late-sixteenth-century Peru, for example, an Upper Guinea influence seems likely, since more than half of all Africans who came to that region were from a twenty-thousand-square-kilometer area stretching from the Lower Casamance to the Kogon rivers. In late-seventeenth-century Brazil, a strong West Central African connection meant that Mbundu and BaKongo slaves were able to re-create specific ritual practices, such as divination ordeals, resort to spirit possession, and funeral ceremonies, that owed much to their homelands. Prominent war captives from the Akan state of Akwamu in the early 1730s later led a slave revolt in Danish St. John; their plans drew on their knowledge of Akwamu statecraft. Similarly, the rebels in the Stono Rebellion of 1739 in South Carolina were probably from Kongo; in that decade, about two-thirds of the Africans arriving in the colony came from West Central Africa, and the dancing, music making, and military strategies of the rebels can probably be attributed to native practices. Yet in other parts of the Americas, as Hall noted, the earliest migrants, who might be tiny minorities overall, could form a charter generation that had undue influence on later arrivals. She argued in an earlier work that a few thousand "Bambara" (she now prefers the term "Bamana") from Senegambia were central to understanding early Louisiana history. Ira Berlin's Many Thousands Gone valorizes the role of a small number of African creoles on whom Europeans depended to establish their position in the Atlantic. Linda Heywood and John Thornton argue that the founding generation was drawn primarily from West-Central Africa, bearers of a regional culture. In some cases, then, influence is claimed on the basis of a large majority; in others, a small crucial group, a charter generation, is said to have exercised influence disproportionate to their numbers. ⁶

Atlantic history has matured to the point where it needs to break out of the straitjacket imposed by the two models that have dominated interpretations of the historiography of the Americas. There is no need either to choose between them or to pass judgment on their appropriateness. Rather than either Old World folkways or New World environments, we need to encompass both and become much more thoroughly Atlantic. Community and cultural formation in the early Americas was a product of many forces. Rather than frame the issue as solely one of transfers and conduits, we should also think of transformations and overlapping circuits. Rather than posit that slaves and planters always acted knowingly, we should entertain the possibility that they often responded to unseen market forces. Rather than assume that migrants remained conservatively attached to traditional ways, we might also view them as experimenters and improvisers. Once such a perspective is adopted, assigning agency becomes a complex and ambiguous task. But only then are we likely to generate an accurate and plausible picture of the foundation of Atlantic societies.

Since the 1970s, what can be termed the "black rice hypothesis" has emerged in ever stronger form in successive books by Peter H. Wood, Daniel C. Littlefield, and Judith A. Carney. The major export crop of eighteenth-century South Carolina and Georgia—rice—is now seen as predominantly a creation of Africans. This African contribution to New World agriculture is epitomized by the arresting title of Carney's book: Black Rice. A direct role for Africans in American history strikes a chord at a time when the national story is becoming less parochial and is increasingly being viewed in an Atlantic or global context. Furthermore, the emphasis on African agency resonates with histories from the bottom up and with subaltern studies in general. That South Carolina's rice industry was built not just on slave labor, but also on the slaves' agricultural and technological knowledge, is an exciting and appealing revelation. In a multicultural world, it is reassuring to realize that the black contribution to American life involved more than just backbreaking muscle power. The development of American rice culture, the claim goes, marked the transatlantic migration not only of an important crop but of an "entire cultural system." It was a major African accomplishment. ⁷

The basic argument rests on three core elements. First, rice culture was indigenous to Africa and was a practice of long standing. Well before the Europeans arrived, West Africans had developed complex systems of mangrove or tidal floodplain, coastal estuarine, and upland rain-fed forms of rice cultivation. The area of greatest rice specialization centered on the Upper Guinea Coast, that part of the African littoral stretching from present-day Senegal to Liberia, but also reached into the interior, and by the seventeenth century may have extended coastwise to the western Gold Coast. ⁸ Second, in contrast to the cultivation of most plantation crops in the Americas, notably sugar and tobacco, there was never a period when free—or at least non-slave—labor could be induced to produce rice for export. The workforce engaged in cultivating rice for export was always black, although elsewhere in the world, slave labor was not the norm. Moreover, among communities of Maroon or runaway slaves, rice seems to have often become the major staple and assumed special significance. Finally, putative parallels have emerged between rice cultivation in Africa and its counterparts in the Americas. From land preparation through sowing, weeding, irrigating, threshing, milling, winnowing, and cooking, African practices seemingly left a deep imprint on New World ways of growing and processing the crop. ⁹

South Carolina (later joined by Georgia and the Cape Fear region of North Carolina) was the primary, but not the only, rice producer in the Americas. By the late eighteenth century, northeastern Brazil (the present states of Amapá, Pará, and Maranhão) had become a significant center of slave-grown rice for export. ¹⁰ There were, then, two key nodal points for commercial rice production in the eighteenth-century Americas, although one was much larger than the other. In addition, the production of rice for subsistence and as a minor plantation crop occurred in many other parts of the New World—Peru, Mexico, the Guianas, Suriname, Cayenne, El Salvador, Jamaica, and Louisiana. Surinamese Maroons grew rice as their primary food crop, and their oral traditions include stories about female ancestors who hid seed rice in their hair when moving either from Africa to Suriname or from plantation to Maroon camp; a rebellion on a Bahian sugar plantation in 1789 involved a demand by predominantly creole slaves to "be able to plant our rice wherever we wish." In short, rice became widely grown throughout the Americas, and in each case the association with black labor is evident. Rice, notes Carney, was "the signature cereal of the African diaspora." ¹¹

If the association of black labor and rice growing now seems widely accepted, the linkage between the rice-growing regions of Africa and those in the Americas is actually tenuous. Such a connection rests on some key claims, all related to flows of enslaved labor from parts of Africa to the Americas. First, and most obvious, Africans from rice-growing areas are said to have been either a significant minority or even a majority of those slaves arriving in New World regions that specialized in rice. ¹² Planter preference is largely seen to have shaped slave shipments, at least to those areas that cultivated rice commercially. Second, African agricultural expertise was highly gendered. In some places, rice was solely a women's crop; in others—usually where more elaborate systems arose for irrigating rice, requiring ditching and banking—a complex division of labor between men and women emerged; but in West Africa, women selected and sowed the seed and later processed and cooked the cereal. Given female expertise, it is claimed, slave arrivals in South Carolina included a higher percentage of women than arrivals in the Caribbean, where sugar was the predominant crop. Third, female slaves bound for American rice-growing areas allegedly commanded higher prices than in other plantation economies. In South Carolina, according to Carney, the labor of female slaves "was valued more on a par with that of male bondsmen than in the slave markets of the West Indies." ¹³

Such claims, if valid, would provide a prima facie case for the transfer of African rice-cultivating technologies to the Americas. But they have wider significance, too. With respect to gender, they represent an important caveat to the received wisdom that in the transatlantic traffic, male slaves typically outnumbered, and fetched higher prices than, females. As for commercial rice cultivation in the Americas, they seem to imply African acceptance of or even collaboration with—rather than resistance to—plantation development. At the very least, slaves are depicted as using their rice skills as a bargaining chip in active negotiation with masters. Suggestions, therefore, that rice cultivation in the Americas hinged upon the transfer of rice-growing skills and technologies through the slave trade from Africa challenge some of our most fundamental assumptions about African agency, the patterns and structures of transatlantic slavery, and working conditions on slave plantations. The degree of the African contribution to the development of American rice cultivation therefore merits close investigation.

There is no doubt that African slaves were the primary cultivators of rice and that some introduced Old World customs of sowing, threshing, and winnowing the crop into the New World. However, there is no compelling evidence that African slaves transferred whole agricultural systems to the New World; nor were they the primary players in creating and maintaining rice regimes in the Americas. Rather, a complex set of factors explains the operation of both the slave trade and the plantation system. Critical to the former were African supply conditions, Atlantic trade routes, and mercantile strategies, and to the latter, planter priorities, environmental conditions, and managerial initiatives. In neither case should the primary engine of development be reduced to planter preferences in the one and slave agency in the other. ¹⁴

A review of the African coastal provenance of slaves arriving in North America lends little support to the idea that rice planters sought slaves from the rice-growing regions of Africa. In the crucial formative period, prior to 1750, when the foundations of the lowcountry rice economy were laid and when slaves with rice-growing skills might have been expected to be in most demand, only about one-fifth of the region's Africans came from Upper Guinea. Moreover, during the same period, the tobacco-growing Chesapeake region drew on Upper Guinea to almost exactly the same degree as its lowcountry counterpart, while other North American regions received proportionately the largest share—just over half—of their slaves from Upper Guinea. In short, parts of North America other than the lowcountry have as good or an even better claim to be linked to rice-growing areas in Africa. In the first half of the eighteenth century, the relationship between the lowcountry and the rice-growing region of West Africa was weak. (See Table 1 .) ¹⁵

Table 1 Proportion of Slaves Arriving in Mainland North American Regions from Upper Guinea

	No. of Slaves from Upper Guinea	Slaves from Upper Guinea as of Arrivals from All of Africa

Before 1751
Chesapeake	9,897	22.3
South Carolina and Georgia	4,856	21.9
All other regions	4,740	52.4
1751–1775
Chesapeake	8,771	39.9
South Carolina and Georgia	35,774	58.2
All other regions	2,532	59.5
1776–1800
South Carolina and Georgia	7,158	50.5
All other U.S. regions	913	65.0
1801–1825
South Carolina and Georgia	13,886	31.6
All other U.S. regions	1,029	30.5

Source: Calculated from the Trans-Atlantic Slave Trade Database [TSTD2], http://www.slavevoyages.com.

Perhaps the origins of risiculture in South Carolina can be traced to Barbados, its cultural hearth, since so many of the early settlers and their slaves came from the island? Before 1700, however, only 6 percent of arrivals in Barbados came from Upper Guinea, and many of those were sold again, into the Spanish Americas. ¹⁶ After allowing for Upper Guinea slaves who grew millet rather than rice and for the number of children in the slave trade to Barbados, perhaps one out of a hundred slaves in this Barbadian charter generation would have known anything about rice culture. After 1700, a slave route connection between South Carolina and the eastern Caribbean continued. It involved about 10 percent of all slaves arriving in South Carolina between 1700 and 1740. Yet only 7 percent of African arrivals into the British Caribbean during this period came from Upper Guinea. ¹⁷ The possibility that lowcountry planters sought slaves with rice-growing skills via an intercolonial geographical corridor through the British Caribbean seems remote.

Indeed, at the time that South Carolina's rice economy became established, the links between the colony and Africa were tenuous in two major respects. First, before 1715 or so, most slaves arrived in the lowcountry in small shipments from the West Indies. South Carolina, as Wood famously noted, was "a colony of a colony." Not only would very few of these slaves have had much prior knowledge of rice production in Africa, but none of them would have grown the crop in Barbados. Scarce surface water and restricted provision lands made rice cultivation impossible on the island. Second, in early-eighteenth-century South Carolina, Indian slaves constituted the fastest-growing part of the population, and African slaves remained marginal at that time. ¹⁸

After 1750, the connection between South Carolina and Georgia and possible rice-growing regions in Africa became stronger. From 1750 to 1776, nearly three out of every five Africans arriving in the lowcountry came from Upper Guinea. This period also saw an indigo boom in the lowcountry, and many of these Upper Guinea slaves were undoubtedly put to growing indigo rather than rice. ¹⁹ The proportion of Upper Guinea slaves arriving in the lowcountry dropped in the last quarter of the century, but they still represented half of all arrivals. At the same time, Upper Guinea also supplied more Africans to other parts of North America than ever before. Two out of every five Africans arriving in the Chesapeake in the third quarter of the eighteenth century were from Upper Guinea; other parts of North America received as large a share of Upper Guinea slaves as did the lowcountry. Indeed, in the last quarter of the eighteenth century, Upper Guinea was once again proportionately a greater supplier to regions of the United States other than the lowcountry (although the overall numbers were small). Thus after 1750, Upper Guinea was a more significant supplier to all North American regions, not just the lowcountry, than it had ever been before.

In the early nineteenth century, when the lowcountry was the only U.S. region to resort to the African slave trade to any great degree (in one last explosion of slave trading before formal abolition), it was once again receiving only a minority of its Africans from rice-producing areas. Between 1804 and 1807, less than one in three Africans arriving in the lowcountry came from Upper Guinea. Overall, then, with the exception of the third quarter of the eighteenth century, the lowcountry did not import a majority of its slaves from rice-producing regions. Other regions in what became the United States had stronger claims to be connected to Upper Guinea than did South Carolina and Georgia.

The Caribbean, which exported no rice, also had greater connections to Upper Guinea than did the lowcountry. The islands received seven times more Africans from rice-growing areas than did South Carolina and Georgia, making the Antilles easily the most important destination of slaves from Upper Guinea after 1700. (See Table 2 .) More significant, the rising importance of Upper Guinea as a source of slaves in the third quarter of the eighteenth century is apparent for both sugar- and rice-growing regions. Between 1751 and 1775, in both the Caribbean and the lowcountry, the proportion of slaves arriving from Upper Guinea more than doubled. Still, the lowcountry and the other major North American plantation regions drew a greater share of their slaves from Upper Guinea than did any other major plantation region of the Americas. Overall, only about one in every six slaves coming into the Caribbean was from Upper Guinea. ²⁰

Table 2 Slaves from Upper Guinea Entering South Carolina, Georgia, and the Caribbean, 1676–1860

	Before 1751	1751–1775	1776–1800	1801–1860

South Carolina–Georgia	4,856	35,774	7,158	13,886
Upper Guinea captives as % of all slaves	21.9	58.2	50.5	31.4
Caribbean	83,787	192,813	122,507	67,555
Upper Guinea captives as % of all slaves	9.8	23.6	13.5	16.0

Source: Calculated from TSDT2.

South Carolina planters did not seek slaves from Upper Guinea in order to develop rice production. Once planters learned how to use land and water to grow rice—something they had achieved by the early eighteenth century, and without a critical mass of Africans to teach them—increasing numbers of rice growers were not vital to their success. The post-1750 surge of Africans who may have had some knowledge of rice cultivation can be explained without reference to planter preference and was not necessary to rice's development. Perhaps, from their homeland memories, Africans contributed the hollow logs or trunks that regulated the flow of water between rivers and fields as lowcountry planters shifted from swampland to tidal irrigation in the post-1750 period, but then again Europeans knew much about draining and embanking wetlands, and they, too, employed hollow logs. Furthermore, lowcountry planters soon adopted hanging floodgates, known only to European agriculture. The natural growth of the slave population in South Carolina after about 1750 also facilitated intergenerational transfers of rice-growing skills within the province's labor force. This development alleviated any possible dependence on African incomers just when slaves from Upper Guinea became increasingly abundant. Among the complex mix of factors that shaped the Atlantic slave trade and established the major links between Africa and the Americas, planter preferences for slaves from particular African regions played a minor role. ²¹

North American mainland markets for slaves were always peripheral to the transatlantic slave-delivery system. Between 1650 and 1780, they accounted for less than 7 percent of the slaves carried across the Atlantic, and probably no vessel in this era brought slaves to the mainland without first calling at a Caribbean island—usually Barbados or Antigua—to check out alternative marketing prospects. When slave traders planned a transatlantic expedition, they juggled a variety of factors, which included trading patterns in Africa, seasonal fluctuations in harvests in both Africa and the Americas, slave resistance patterns, international wars, and credit links between slave-trading regions on both sides of the Atlantic. ²²

Buyers of slaves in the Americas wanted a cheap supply of undifferentiated labor for field work, and transatlantic suppliers sought locations in Africa where they could obtain large numbers of slaves quickly and at reasonable cost. Buyer preferences for peoples from particular regions in Africa could be exercised only in the largest markets of the Caribbean, such as Barbados, Jamaica, and St. Domingue, where vessels from all parts of Africa arrived in large numbers. Even in these places, planter preferences were not central. Slave traders could obtain the largest volume of slaves in the shortest period of time on the Slave Coast of the Bight of Benin, the Bight of Biafra, West-Central Africa, and to a lesser extent the Gold Coast, all regions where rice was not a dominant crop. The large slave vessels from these regions first targeted major markets in the Caribbean and then moved to the fringe areas if necessary. Even then, it was well known that only the most prosperous of these secondary markets—say, Charleston after 1750—could absorb large numbers of slaves at one time. Slave trading in Upper Guinea, which lay to the north of the major slave provenance zones, required smaller vessels (and a strategy of buying small groups of slaves at scattered locations) able to supply regions in the Americas that could not easily sell the numbers per vessel carried from the major markets farther south. Before 1750, therefore, a link emerged between secondary markets in Africa and secondary markets in the Americas based on trading conditions in the Old World and the inability of North American mainland planters to compete with the sugar barons of the Caribbean. But such a link applied to both the Chesapeake and the lowcountry and had little to do with rice-growing skills. ²³

After 1750, a new, and as yet poorly understood, set of conditions reshaped the supply of slaves, the effect of which was greatly to increase the loading times of vessels trading in all the major African slave provenance zones. ²⁴ In response, more vessels began to seek slaves in Upper Guinea than had previously been the case. Minor suppliers during the era when rice cultivation was beginning, Upper Guinea locales now became more important. In fact, from the broadest perspective, Senegambia, Sierra Leone, and the Windward Coast were important in relative terms only in the third quarter of the eighteenth century, when their share of total slave departures from Africa jumped from just 9 percent before 1750 to 22 percent in 1751–1775. Thereafter, it fell back to 12 percent after 1775 and to under 10 percent in the nineteenth century. ²⁵ Bulking centers emerged on Bance (or Bunce) Island and the Iles de Los, and they began to offer loading rates that attracted larger slave vessels. Moreover, in 1759–1779, the British occupied French Senegal, thereby helping to increase the flow of slaves from Senegambia to British America in this period. The number of British vessels obtaining slaves in Upper Guinea in the third quarter of the eighteenth century was close to three times the long-run average. This transformation in trading conditions in Africa largely explains why the share of slaves from Upper Guinea arriving in South Carolina rose from just over one in five before 1751 to nearly three in five after mid-century (and to two in five in the Chesapeake). That this surge had nothing whatsoever to do with rice cultivation is suggested by the destinations of most slaves leaving Upper Guinea, which were places in the Americas not associated with commercial rice growing. Before the nineteenth century, American rice-producing regions received only one-fifth of the slaves leaving the rice-producing regions of Africa. Sugar-growing regions in the Caribbean absorbed far more slaves from Upper Guinea than North (and for that matter South) American rice-growing regions. Thus the explanation for patterns of slave provenance from Upper Guinea most probably lies in Africa rather than in the Americas.

Outside South Carolina and Georgia, Brazil was the only other part of the Americas that produced rice efficiently enough to justify shipping it to transatlantic markets—mainly Portugal. Production in Brazil centered on Maranhão and to a lesser extent Pará in the northeast. At first glance, Maranhão and Pará in Amazonia appear to offer more support for the "black rice" hypothesis than does the lowcountry. Maranhão began to export rice in the late 1760s and remained the principal source of exports in the colonial period, while from the 1760s to about 1810, almost two out of every three African slaves brought into Pará and Maranhão came from Upper Guinea. (See Table 3 .) In fact, all but a few hundred embarked at Cacheu and Bissau. Upper Guinea Africans were negligible among arrivals to all other Brazilian regions. On a cursory glance, then, a reasonable congruence between the growth of rice cultivation and a predominance of Africans from rice-producing regions emerged in northeastern Brazil.

Table 3 Proportion of Slaves Arriving in Different Regions of Brazil from Upper Guinea

	No. of Slaves from Rice-Growing Regions	Slaves from Rice-Growing Regions as % of Arrivals from All of Africa

Before 1751
Brazil north of Pernambuco	683	49.5
All other Brazilian regions	9,520	1.5
1751–1775
Brazil north of Pernambuco	16,786	76.7
All other Brazilian regions	448	0.2
1776–1800
Brazil north of Pernambuco	26,449	67.7
All other Brazilian regions	591	0.2
1801–1825
Brazil north of Pernambuco	26,202	47.9
All other Brazilian regions	2,596	0.3
1826–1850
Brazil north of Pernambuco	2,215	30.4
All other Brazilian regions	4,521	0.7

Source: Calculated from TSDT2.

Yet on closer investigation, the link between rice growing in northeastern Brazil and Upper Guinea proves flimsy. Rice played a minor role in the so-called "agricultural renaissance" of Brazil in the second half of the eighteenth century. Rice exports were small compared to those from the lowcountry. After expanding rapidly in the 1770s, exports from Amazonia leveled out at around 11 million pounds of rice per year after 1785. (See Table 4 .) By contrast, annual exports from the lowcountry were seven times this amount between 1770 and 1774, and after falling during the Revolutionary War, they quickly recovered to previous levels. ²⁶ Rice was not even the major crop of Amazonia. In Pará, cacao was far more important, and many of the slaves taken there were destined for the interior provinces of Mata Grosso and Goiás rather than the local market. Between 1760 and 1810 in Maranhão, the explosive growth of cotton rather than rice transformed a largely subsistence economy into one with a significant export sector. Cotton accounted for 80 and 75 percent, respectively, of Maranhão's exports in 1796 and 1806, and cacao for more than half of the much smaller sum of exports from Pará. Rice, unfortunately, cannot be disentangled from the category "foodstuffs" in the colonial export returns, but if it could, it would not have represented as much as 10 percent of exports from either port; it always lagged behind cotton and cacao. ²⁷ Slave arrivals in Amazonia had been only occasional before 1760. ²⁸ The region was highly unusual in the Americas for having a chartered company—the Companhia Geral do Grão Pará e Maranhão—that imported slaves after 1755 as part of an attempt to kick-start commodity exports. Slave arrivals reached almost 1,000 annually in the late 1750s, and then rose steadily to nearly 4,000 a year by the early 1800s. They went to work primarily on plantations growing cotton, not rice. ²⁹

Table 4 African Origins of Slaves Arriving in Northeast Brazil from Africa and Rice Exports from Northeast Brazil, Prior to 1810

	Upper Guinea	Gold Coast to Southeast Africa	Total Slaves of Known Origin	Column 1/Column 3 as %	Annual Rice Exports in Millions of Pounds

Pre-1756	1,975	698	2,673	73.9	0.0
1756–1760	2,008	1,742	3,750	53.5	0.0
1761–1765	3,393	3,367	6,760	50.2	0.0
1766–1770	5,280	0	5,280	100.0	0.1
1771–1775	4,812	0	4,812	100.0	1.8
1776–1780	5,168	1,938	7,106	72.7	5.8
1781–1785	4,088	881	4,969	82.3	8.2
1786–1790	6,089	3,676	9,765	62.2	11.1
1791–1795	7,415	2,568	9,983	73.7	-
1796–1800	3,889	3,582	7,471	52.1	11.2
1801–1805	6,106	9,381	15,487	39.4	8.9
1806–1810	9,989	3,630	13,619	73.3	13.5
Pre-1756–1810	60,212	31,463	91,675	65.7	-

Sources: For slaves, calculated from TSDT2; for rice, Dauril Alden, "Late Colonial Brazil, 1750–1808," in Leslie Bethell, ed., Colonial Latin America (Cambridge, 1984), 601–660, supplemented by Maranhão, cx. 45, doc. 4415; cx. 47, doc. 4595; cx. 50, doc. 4905; cx. 55, doc. 5166; cx. 61, doc. 5567; cx. 73, doc. 6288; cx. 79, doc. 6718; cx. 84, doc. 7086; cx. 86, doc. 7179; cx. 89, doc. 7404; cx. 93, doc. 7680, Arquivo Histórico Ultramarino, Lisbon.

The majority of the slaves arriving in Amazonia after 1755 came from Senegambia, where some of them would have been familiar with rice cultivation, but the winds, ocean currents, and geography of the Atlantic ensured that Amazonia would always draw on Upper Guinea whatever the nature of the crop. From the late seventeenth century, when occasional direct shipments of slaves from Africa began, down to the first export of small amounts of rice in 1767, almost half the slaves coming into Amazonia were from Upper Guinea. A variety of red rice native to the Americas was grown in Maranhão in the first half of the eighteenth century (called arroz da terra or arroz vermelho by the Portuguese), but during this period Brazil actually imported some rice from South Carolina via Lisbon. ³⁰ The trigger for an export-based risiculture in Maranhão was the Companhia Geral do Grão Pará e Maranhão's introduction of a new strain of rice from South Carolina in the late 1760s, not the arrival of slaves from Upper Guinea. ³¹ While the proportion arriving from Upper Guinea increased initially along with rice exports, this regional share also began to decline quickly as rice exports entered their most expansive phase. After 1820, only about a quarter of Amazonia's slaves came from Upper Guinea.

As in the case of South Carolina, the patterns of forced African migration into Amazonia are best explained by what was happening on the African coast. Amazonia, like South Carolina and Georgia, provided a peripheral market for slaves in the Atlantic as a whole. It faced the same tightening of slave supplies as its northern rivals in the 1760s and 1770s. Like South Carolina and Georgia, Maranhão and Pará responded to the higher prices induced by extended waits on the coast by resorting to second-tier embarkation regions. For all plantation areas in the Americas north of Pernambuco, but especially the marginal ones, regions such as Upper Guinea became more important at this time than they had been previously. The American Revolutionary War, which temporarily curtailed British slaving but barely affected Portugal, and then the massive St. Domingue slave revolt, which crippled the French trade, reduced the pressures on the African coast. This opportunity allowed marginal plantation areas, such as Amazonia, to reenter the prime slave markets in Africa, thereby easing their dependence on Upper Guinea. For Amazonia, this development accelerated in the nineteenth century as the Anglo-Portuguese treaty of 1810 and subsequent initiatives saw the British navy gradually attempt to shut down the slave trade to Brazil. These efforts initially focused on slave supply centers north of the Equator, including Upper Guinea, and while they were less than wholly successful, they helped to shift the center of gravity of slave trading southward. Thus transatlantic connections between Amazonia and Upper Guinea were forged by geography, subjected to supply conditions in Africa, and broken, or at least seriously attenuated, by British naval and diplomatic action after 1820. Planters looking for slave laborers familiar with rice-growing techniques had little influence.

For a third region of the Americas—Suriname—the focus of the "black rice" argument switches from rice as an export staple to rice as a subsistence crop. Rice thereby becomes an important part of the system that generated exports rather than the export itself. In part, this shift of emphasis is necessary because rice was never an important export from Suriname in the era of the slave trade. Rice became a major foodstuff in first English and then, after 1667, Dutch Suriname because slaves carried risiculture with them from Brazil in 1654 when their Dutch owners were expelled from Pernambuco. One-quarter of slave arrivals in early Brazil, it is argued, originated in Upper Guinea, and rice was the preferred cereal in much of the new Portuguese colony. Fresh infusions of African rice consumers and women skilled in rice production are said to have arrived in Suriname via the Dutch slave trade, not just from the Upper Guinea Coast, but from the Gold Coast, many of whose inhabitants had begun eating rice as the result of a risiculture located at Axim. ³²

There is an evidentiary base for only a small part of this narrative. There are no records of any slaves from African rice-growing regions in Brazil before the Dutch conquest. Between 1574 and 1630, the African origins are known for only sixteen slaving voyages that arrived in Pernambuco and Bahia. All sailed from Angola and São Tomé. More important, the picture of early Portuguese slave trading in Africa now emerging from the work of a new generation of scholars is at odds with the Upper Guinea focus. Before 1570, Portuguese slave traders certainly centered their slave-trading efforts on the African coast at São Tomé and the Cape Verde Islands, with the latter drawing heavily on Senegambia, but the Portuguese took the Cape Verde Island slaves to New Spain and the Caribbean, not to Brazil. The first slaving voyage direct from Africa to Brazil (probably Pernambuco) likely occurred only in 1560, and by this time the Portuguese had already begun to concentrate on Angola, albeit after sending the slaves first to São Tomé. From 1636 to 1651, the Dutch introduced a few slaves into Pernambuco from the Gold Coast, but none at all from Upper Guinea. After 1654, the old concentration on slaves originating in Africa south of the Equator was apparently reestablished. ³³

Between 1667 and 1730, Dutch Suriname drew on rice-growing regions in Africa no more than had Brazil prior to 1667. The few slaves brought over by the Dutch who originated on the Gold Coast (less than 2 percent of the Dutch slave traffic) did not come from the western parts of the region around Axim, but from farther east. Had they acquired a taste for rice distributed along the coast by Axim's rice producers? As Carney rightly notes, "Rice and millet dominated the grains that were traded," but Ray Kea's estimate of Axim's rice exports of 100 tons a year would have provided 0.5 lbs of rice per day for a mere 1,000 people. In short, rice was trivial compared to millet. There is neither hard evidence nor much probability that people with rice-growing skills arrived in Suriname on Dutch slave ships at this time. As for the provisioning of slave ships, no known slaver left the Gold Coast provisioned mainly with rice. Anomabu was the main provisioning center on the Gold Coast long before it also became the main embarkation point in the region for slaves. ³⁴ It supplied only millet (or "corn" to the English). Carney concedes that the share of slaves coming from the Gold Coast to Suriname before 1700 was small (although there were enough "to provide a critical mass for effecting the transfer of the knowledge and skills necessary to the cereal's introduction"). Leaving aside the 98 percent of slaves who must have had preferences for foods other than rice, there is no evidence that the small remainder had any "memory" of rice, while there is a great deal of circumstantial evidence that they did not. ³⁵

Of course rice came to Suriname. Carney cites two pieces of evidence to support its early introduction. In 1665, Major John Scott reported that it was among sixteen commodities grown there. In 1687, there was a small shipment of rice from Suriname to Holland. Thereafter there were no further exports until 1783. But who among Africans could have introduced rice in seventeenth-century Suriname, given the slave trade evidence? We simply do not know. What we do know is that Africans with rice-growing skills likely did not arrive in Suriname until after the removal of the Dutch West India Company's monopoly in 1730. Dutch slave traders began to draw some slaves from the Windward Coast at that point (6,077 out of 39,190 from known regions), and then more heavily after 1750, as one-third of all Suriname arrivals came from this region between 1751 and 1775. This development corresponds with the first written evidence of Maroon dependence on rice as a subsistence crop. Maroon oral histories might be used to date its introduction to 1693 (when a slave revolt occurred), but recently collected oral history seems a tenuous basis for pinpointing an event in 1693. ³⁶

The mix of peoples in the New World depended on a combination of African supply, winds and ocean currents, and strategies of competing European shippers. Preferences of the ultimate buyers and, of course, sellers of human beings inevitably influenced this as any other transaction, but such preferences have garnered too much attention from historians of the Americas. Planters in South Carolina, Georgia, Amazonia, and Suriname formed a tiny part of the overall transatlantic market for slaves, even the market for slaves from Upper Guinea. It is not even evident that their preferences actually were for slaves from Upper Guinea. In any event, they were subject to trends beyond their control—in particular, the pattern that saw the rise and then decline of Upper Guinea in the transatlantic slave trade—which affected all buyers. Poorly understood though this pattern may be, it had nothing to do with the ability of the region's slaves to grow rice.

In fact, the Upper Guinea Coast was never uniformly committed to rice production. The northernmost section, Senegambia (the area between and including the Senegal and Gambia rivers), was primarily a millet-producing region. True, Mandinka practiced a form of paddy rice agriculture along the Gambia River before the Europeans arrived, but in the Senegal River Valley, slave traders calculated the subsistence not in rice but in millet (two pounds a day). The ability of this region to export slaves was closely linked to millet production. William Littleton, who traded in the Gambia for eleven years, explained that the typical strategy of ship captains was to purchase "all the Grain we can," and he first singled out "Country Corn" or millet—and then rice. When asked whether ships could procure a "sufficient quantity of Guinea corn" for the Middle Passage, he replied, "Seldom a sufficient quantity of that alone." Similarly, another ship captain, who traded in Saint-Louis in the 1770s, observed that "Those ships going to the West Indies with Slaves were supplied by the Blacks with large quantities of corn [millet], which the Slaves preferred to any other kind of provision." In 1788, the Company of Senegal even had to ship in rice and flour from France to Saint-Louis and Gorée because of a grain shortage in the region. ³⁷

Even in the coastal Guinea-Conakry and Guinea-Bissau parts of the Upper Guinea Coast, which are usually assumed to be predominantly rice-growing areas, rice cultivation was only a part of agricultural activities—and often a small part until well into the eighteenth century. Earlier, the Balanta of Guinea-Bissau, as Walter Hawthorne demonstrates, cultivated beans, pumpkins, maize, and yams. They switched to paddy rice production for a number of reasons: the superior yield of the crop; access to the iron tools needed for ditching and embankment, which came via increased transatlantic trade; and the general rise in violence, which encouraged a movement into terrain that offered better protection. Ironically, the transatlantic slave trade facilitated the Balanta's transformation into rice farmers, but those Balanta shipped to the Americas before about the third quarter of the eighteenth century would have known little about rice cultivation. Notably, then, the nature of rice cultivation changed on both sides of the Atlantic over the course of the eighteenth century. ³⁸

Were rice farmers, especially those along the coast, particularly vulnerable to European slavers because they were sedentary and because of their proximity to European navigational routes? Many rice farmers were swept into slavery, of course, but others, such as the Diola (or Jola), managed to survive the era's violence largely because they inhabited inaccessible wetlands. As Robert Baum points out, "the Diola are considered the best wet rice cultivators in West Africa," and the evidence for their long-standing cultivation of rice in the well-watered coastal plain bisected by the Casamance River stretches over two millennia. The Diola even sold rice paddies in order to ransom relatives who had been captured by slave raiders before they were sold into slavery. The Baga-Sitem in the Rio Nunez region—a subgroup of the Baga peoples, who generally inhabited mangrove islands located between Guinea-Bissau and Iles de Los and who were said by one traveler in 1793 to be "very expert in Cultivating rice and in quite a Different manner to any of the Nations on the Windward Coast"—neither held slaves nor sold them. The Baga-Sitem and Diola were unlikely, then, to have been major sources of rice expertise in the Americas. Rather, slave raiding forced some decentralized societies to migrate to isolated swamps, where they reorganized their traditional cropping system from yams to irrigated rice. Such societies were not necessarily victims but resisted incursions of powerful state armies by moving into the riverine, marshy, tsetse fly–infested areas of the coast, where the inhospitable landscape provided something of a sanctuary. There they built defensive households and fortified villages and armed themselves. Rice cultivation did not necessarily make them "easy prey." ³⁹

Evidence of provisioning on slave ships is especially useful for yielding clues about the agricultural priorities of particular African regions. Part of the strategy for keeping valuable property alive on the transatlantic crossing was to ensure that slaves received food to which they were accustomed. Yams fed many more Africans on the Middle Passage than did rice, but even in rice-producing areas, other cereals (millet initially, and then maize in some places) were important provisions on slave ships. For twenty Royal African Company vessels that visited the Gambia and Sierra Leone between 1679 and 1688, the amount of rice and millet carried can be ascertained. Twelve of the twenty ships loaded both rice and millet, six millet alone, and only two just rice. In overall weight, millet formed 70 percent and rice 30 percent of the total provisions on these ships. All the rice on these ships was "clean" (that is, milled), as was almost all the millet; only three ships carried small quantities of millet in the husk. From this evidence, women did not mill rice on the Middle Passage, and whatever pounding by mortar and pestle occurred was more likely of millet. Indeed, the term for a domestic female slave in the Senegal River region was pileuse ("pounder of millet"). ⁴⁰

Figure 1: Baga women, one with a child on her back, transplant rice so dexterously "as to plant fifty roots singly in one minute" in "Lugars" or "flat low swamps." But, as the observer of this scene emphasized, the Baga cultivated rice differently from any other group on the Windward Coast. From the log of the slave ship Sandown, p. 55, September 24, 1793. Reprinted by permission of the National Maritime Museum, London.

For all these reasons, the number and percentage of Africans with rice-growing experience must have been far below the total number of slaves leaving Upper Guinea, as revealed in Tables 1 through 3 . Indeed, the proportion of Upper Guinea Africans with rice-growing skills was especially minimal during the years when risiculture became established in the lowcountry. "Upper Guinea" and "rice-growing regions" are far from synonymous.

Mercantile advertising of ships that transported enslaved Africans also reveals, in a somewhat refracted manner, contemporary perceptions of Africa's coastal regions. While references to the "Grain Coast" (meaning not cereals but Melegueta pepper) date from 1752, the first explicit association between an African region and the cultivation of rice came six years later. The merchant firm Austin and Laurens described the origins of the slave ship Betsey as the "Windward and Rice Coast." Thereafter, "Rice Coast" or "rice country" gradually supplanted "Grain Coast" (although the latter was still frequent) as a common designation for Upper Guinea locales. Thus in 1760, a merchant firm advertised a ship as having "200 fine Rice Coast Negroes from Sierra Leon"; the following year, two ships arrived from "Bance-Island on the Rice-Coast"; in 1769, merchants described Cape Mount on the Windward Coast as "a rice country" or as "the centre of a Rice Country"; and two years later, Gambia merited the ubiquitous term "a rice country." ⁴¹

Significantly, it was a London merchant involved in the African slave trade, although one interested in settling in the lowcountry, who made one of the earliest and most explicit references to African regional rice cultivation. In 1764, he thought that if slaves were brought to the lowcountry "from the Windward Coast where they cultivate rice they may be soon trained to plantation business." Of course, he may have meant no more than that a prior agricultural background would be useful for a prospective plantation hand. Yet, notably, he singled out African rice planting as a useful precondition, much as his Charleston counterparts were doing at roughly the same time. ⁴²

Nevertheless, the most explicit advertisements about African rice-producing origins were produced after the American Revolution and appear as enthusiastic—and quite possibly misleading—merchandising. Thus in 1784, one merchant extravagantly praised a shipment of Africans from the Gambia as "well acquainted with the cultivation of Indigo, Rice, and Tobacco," while a pair of merchants described their "Gambia Negroes" as "universally reckoned the best that can be imported, they being perfectly tractable, accustomed to labour, and well acquainted with the cultivation of rice, indigo, and tobacco." In the following year, another merchant said that Gambian slaves knew how to cultivate rice and "are naturally industrious," and yet another described two vessels from the "Windward and Gold Coast[s]" as containing Africans who "have been accustomed to the planting of rice" or "accustomed to the planting of both rice and corn." Such claims probably no more reflected the particular skills of slaves than did some of the ethnic labels that slave factors attached to them. Twenty years later, during the final flourish of slave imports into South Carolina, merchants did not mention the rice-growing skills of Africans. Having been deemed so obvious, perhaps they no longer merited attention; or possibly exuberant advertising had become counterproductive. ⁴³

The coastal region of origin of African slaves is an extremely imprecise indicator of risiculture. Slaves arriving in the Americas from Upper Guinea were more likely to be familiar with rice than those coming from other regions, but on many of the vessels from Upper Guinea, there must have been few, if any, slaves with expertise in rice cultivation, especially in the embryonic phase of the American rice industry.

Gender is an important element in the "black rice" hypothesis. Rice cultivation in Africa was typically a female activity, and the transatlantic transfer of African rice-growing technologies allegedly meant that high concentrations of females were brought in from Upper Guinea, and that a higher value was placed on female labor in American rice-growing regions than elsewhere. We can test both of these claims by placing the composition of arrivals in South Carolina and Amazonia into a wider Atlantic perspective and by comparing slave prices in South Carolina with those in other parts of British America. Once more, the "black rice" argument is found wanting.

There is no evidence for the claim that the lowcountry imported a higher than usual percentage of women. Before 1776, the proportion of males among the lowcountry's incoming Africans was 69 percent, a little higher than the proportion for the Chesapeake (68 percent) and all other North American regions (67 percent). (See Table 5 .) Furthermore, in an even broader context, the lowcountry imported fewer females than did the Caribbean, where before 1776 the male ratio was just 62 percent. Moreover, the proportion of males among Africans arriving in the lowcountry from Upper Guinea was slightly higher (71 percent) than that from all other African regions (67 percent). (See Table 6 .) The data on age are somewhat thinner than those for sex, but the breakdown for eighteen separate batches of arrivals before 1776 indicates that one-third of the females were classed as girls. In other words, only about one in every five slaves arriving in South Carolina and Charleston was a woman. Brazil and the Caribbean, by contrast, took in more women (between one in three and one in four) and fewer children. Perhaps some of the girls coming into Charleston and Savannah knew something about growing rice, but the "black rice" hypothesis suggests that women should have constituted a large proportion of slaves entering the lowcountry relative to other places. That simply was not the case.

Table 5 Male Ratios of Slaves Arriving in the North American Mainland by Broad Region of Disembarkation before 1776

	Mean	No. of vessels	Std. Deviation	No. of slaves

Chesapeake	0.683	22	0.169	3,408
South Carolina and Georgia	0.693	13	0.141	1,404
All other regions	0.674	14	0.147	2,776
Total	0.683	49	0.152	7,588

Source: Calculated from TSDT2.

Table 6 Male Ratios of Slaves Arriving in South Carolina and Georgia before 1808 by Broad African Region of Embarkation

	No. of vessels	Std. Deviation	Mean

Upper Guinea	13	0.140	0.711
All other African regions	12	0.087	0.670
All African regions	5	0.123	0.692

Source: Calculated from TSDT2.

Did lowcountry merchants seek women workers? According to Robert Pringle in 1741, "full Grown Men & Women [were] most fitt for this market," although he was prepared to concede three years later that a "parcel" of slaves from the West Indies comprising mainly "Boys & Girls of about 15 or 16 years of Age of which 2/3 Boys & 1/3 Girls" might do well. ⁴⁴ Similarly, in 1756, Henry Laurens, of the firm Austin and Laurens, noted that "tall able young People tempt many of our Folks to buy when they are in no real need of them," provided they are "very good & in full Flesh" and that the group "is composed of at least 2/3d males." Africans must be "young robust People," Laurens emphasized, and "Males sell to much more advantage than the Females." ⁴⁵ A good shipment should be at least two-thirds men between the ages of 18 and 25, he stated; the rest could be young women between the ages of 14 and 18. After reporting that most of the prime males from a shipload of slaves from Sierra Leone had sold for good prices, he added that the women had proved much harder to sell, "many of them [being] good People but nobody comes near to Ask the price." A shipment from the Gambia was "not well assorted," he declared, because it contained "more Women than Men & more Boys & Girls than are usual in Gambia Cargoes." ⁴⁶ Nothing here suggests that women were favored over men in South Carolina.

The shift to tidal rice agriculture later in the eighteenth century strengthened this pattern further. The work demands of clearing swampland, digging ditches, and building embankments were extraordinarily onerous. Ditchers were expected to dig about six hundred cubic feet a day in the eighteenth century, and in the nineteenth from five hundred to seven hundred square feet. Slaves moved at least five hundred cubic yards of river swamp for every acre of rice field in order to construct a vast series of banks, dams, canals, ditches, and drains. Planters regarded men and boys as best equipped to handle this intensive labor. They allocated spades and axes almost entirely to men. ⁴⁷

The gender balance among Africans taken to Amazonia was not much different from that in South Carolina. The proportion of females disembarked in Amazonia in the second half of the eighteenth century was above the long-term average for the transatlantic slave trade in all periods—and was greater than the share into the lowcountry—but in the years 1760 to 1810, it was no different for Amazonia than for the rest of the Americas, most of which was growing sugar at this time. (See Table 7 .) Moreover, one-third of the females were classed as girls, not women—again, a ratio little different from that in the rest of the Americas. ⁴⁸ Perhaps Amazonian planters would have preferred more females, but after 1760 they received no more women proportionately than did other American planters.

Table 7 Sex Ratios of Slaves Arriving in Northeast Brazil from Possible Rice-Growing Regions in Africa Compared to Sex Ratios of Slaves Arriving in All of the Americas from All Parts of Africa, 1761–1800

	Northeast Brazil from Upper Guinea		All of the Americas from All Parts of Africa

	Male Ratios	Sample Size (slaves)	Male Ratios	Sample Size (slaves)

1761–1765	0.688	2,225	0.622	19,237
1766–1770	0.573	2,687	0.607	90,155
1771–1775	0.600	3,172	0.610	30,484
1776–1780	0.602	2,470	0.633