The Demographic Cost of Sugar: Debates on Slave Societies and Natural Increase in the Americas

MICHAEL TADMAN

When 18 months complete their growth,
Then the tall canes' rich juices fill;
And we, to bring their liquor forth,
Convey them to the bruising-mill.
That mill, our labour, every hour,
Must with fresh loads of canes supply;
And if we faint, the cart-whip's power,
Gives force which nature's powers deny.

A. Opie, The Black Man's Lament (1826), an antislavery tract for children.

This study is concerned with explaining some remarkable population patterns and with examining the very extensive implications of these patterns. Among North American slaves, births greatly exceeded deaths, so that the slave population expanded rapidly. In sharp contrast, across the slave societies of the Caribbean and Latin America, the persistent experience was one not of natural increase but of dramatic natural decrease. Indeed, the North American pattern was probably, with a few local and sometimes short-term exceptions, unique in the history of slavery. As C. Vann Woodward wrote: "So far as history reveals, no other slave society, whether of antiquity or modern times, has so much as sustained, much less greatly multiplied, its slave population by relying on natural increase."1 Why, then, did North American slaves experience such rapid natural increase (excess of births over deaths), and why did slaves in the rest of the Americas fail to increase naturally?	1
The contrast between North America and the rest of the Americas is a fundamental one. For example, over the many years of the African slave trade, Jamaica imported some 750,000 slaves, but at the time of emancipation in 1838 its black population numbered only just over 300,000: North America, in contrast, imported only about 427,000 Africans, but at the time of emancipation in 1865 the U.S. black population had grown to more than ten times that number.2 In the antebellum period, U.S. slaves showed a natural population growth of some 25 percent per decade (and indeed, North American slaves had established a pattern of natural growth by about 1710). In sharp contrast, Caribbean and Brazilian slaves commonly suffered rates of natural decrease of 20 percent per decade.3	2
The North American slave experience is perhaps even more remarkable when compared with free white populations. From the later eighteenth century, and possibly before that even, and until the Civil War, the rate of natural growth of North American slaves was much greater than for the population of any nation in Europe, and was nearly twice as rapid as that of England. By the 1850s, the rate of natural increase was higher even than that of the white population of the United States. This was a remarkable outcome because, as Thomas Malthus suggested as early as 1798, white Americans expanded with a rapidity "probable without parallel in history."4	3
Anthropometric studies (which use the average height of a population to infer the quality and abundance of its diet) have led some researchers to underline the U.S. slave pattern even more boldly. Writers such as Robert Fogel have asserted that adult U.S. slaves were remarkably tall (and therefore well fed). Supposedly, they were taller than contemporary Africans and taller than Caribbean slaves. Not only this, anthropometricians have claimed that they were taller than European workers of the period and that they were almost as tall as white Americans, the latter apparently being the tallest population of the era. If these anthropometric findings are reliable, it might begin to seem that, in important material aspects of their lives, the experience of North American slaves was better than that of Africans, better than that of Caribbean slaves, and better even than that of the mass of free workers of Europe. It might also begin to look as though diet had a significant role in explaining the natural increase patterns of the Americas.	4
The significance of the contrasting experiences of natural increase and decrease in the Americas is far from being limited to the concerns of specialist demographers. These contrasts both grew out of special structures and attitudes in the regimes concerned and in turn set up special features in those regimes. Not surprisingly, then, these demographic patterns have been, for scholars, the starting point for fundamental assumptions about the nature of these slave societies, about relations between owners and slaves, and about relations within the slave communities. Later in this essay, much more will be said about these attitudes and experiences, but for the time being let us simply note some of the broad outlines.	5
As we shall see, evidence on positive increase and on the stature of its slaves has led some, like Fogel, to conclude that the North American slave system, though morally indefensible, was materially strong and progressive. Indeed, it seemed to Fogel to represent a pioneering model of modern efficient capitalism, where the material conditions of life were good for the free population but also surprisingly favorable for slaves, too.5 Positive natural increase was also crucial in Eugene Genovese's formulation of nineteenth-century U.S. slavery as a historically unique society, based on pre-bourgeois values and on a web of paternalistic relationships that connected slave and master. Genovese wrote: "The paternalism encouraged by the close living of masters and slaves was enormously reinforced by the closing of the African slave trade, which compelled masters to pay greater attention to the reproduction of their labor force. Of all the slave societies in the New World, that of the Old South alone maintained a slave force that reproduced itself."6	6
The literature on British Caribbean slavery has, compared with that for North America, generally placed far less emphasis on paternalism, and the supposedly greater commercial ruthlessness of West Indian owners is seen as having been closely linked to the demographics of slavery. Often, for the British Caribbean, the claim is made that owners were uninterested in natural increase, indeed, that they vigorously discouraged slave women from bearing and bringing up babies. Curiously, in the case of the literature on Latin American slavery, the significance of similar natural decrease patterns to those of the British West Indies has, until quite recent years, tended to be ignored. The traditional disinclination to link natural decrease to the broader interpretation of Latin American slavery has stemmed from a longstanding attachment to the myth that Latin America was, even under slavery, a "racial democracy," especially tolerant of color difference.7	7
The influence of the demographic experiences with which we are concerned also reached far into the lives of the slaves and into the nature of their families and communities. In order to expand its slave population, North America relied far less than the rest of the Americas on African slave importations, and this meant that the Creole (local-born) proportion of the slave population was far greater in North America than elsewhere. In 1800 and even later, Africans made up a majority of the slave populations of the British and French Caribbean and of Brazil. In North America, by contrast, there was a Creole majority by 1740, and by the last years of U.S. slavery less than 1 percent was African-born.8 This situation was of great importance for the nature of religion and slave cultures in different regimes. The demographics of slave regimes also had important implications for slave families and for the possibility of finding spouses, intensive African importation being associated with male-dominated populations and with special problems in finding marriage partners. Intensive importation of young-adult African males also had implications for white conceptions of slaves and for the nature of resistance by enslaved blacks.	8
This essay is concerned with the wider patterns and experiences associated with slave demography, but the core of the article is an attempt to establish what caused the fundamental contrast between the North American demographic experience and that of the rest of the Americas. As the next section will show, many factors influenced natural increase patterns. The central argument, however, is that plantation crop was the essential influence in determining patterns of natural increase and decrease. More specifically, I shall argue that sugar planting systematically brought together a lethal combination of factors that persistently and almost inevitably produced natural decrease among slaves. Significantly, sugar planting, while dominant in the rest of the Americas, was always of very minor significance in North America.	9
In North America, the commercial planting of sugar only began in the 1790s; it never employed more than about 6 percent slaves; and it was almost entirely confined to a group of parishes in southern Louisiana. My study will show that the demographic experience of Louisiana's sugar slaves was unique by U.S. standards. The sugar parishes stood out as a miserable island of natural decrease in the midst of the otherwise consistent pattern of natural increase that stretched across the cotton, tobacco, and rice plantations of the United States. In the Americas as a whole, however, sugar was the great plantation crop, with at least some 60 to 70 percent of all Africans who survived the Atlantic voyage to the Americas ending up on sugar plantations.9 And across the Americas, as we shall see, the demographic impact of sugar was even more severe than in the Louisiana parishes.	10
It should be noted that although this article focuses on the demographic role of sugar plantations, sugar was not the only crop associated with poor demographic results. Even so, no other crop or economic activity had the potential to create natural decrease on the scale that sugar did. After sugar, coffee was the plantation crop that seems to have produced the worst demographic rates, but its labor regime was significantly less demanding than sugar. More important still, it accounted for only a fraction of the number of slaves that sugar did. Rice, too, has been associated with poor demographic rates. Still, with rice, the pattern was one of low but positive natural increase, not natural decrease. Moreover, rice was quite a significant crop in parts of North America, and across the rest of the Americas it only accounted for a tiny percentage of slaves. Mining, mainly for gold and silver, has also been associated with poor demographic performance, indeed with natural decrease, but it was mostly a very localized activity. For instance, it was not a significant employer of black slaves in the British Caribbean, so it cannot explain natural decrease in that region.10	11
As far as the creation of natural decrease is concerned, then, this study points to the dominant role of sugar plantations and sugar planters. When turning to another part of the demographic puzzle—North America's outstandingly high rates of natural increase—this article does not support Fogel's optimistic claim that good diet was a significant factor. Instead, I shall argue that natural increase in North America was the product of circumstances equivalent in important ways to those associated with modern "Third World" populations—that is, populations where very high death rates are compensated for by even higher birth rates. North American natural increase came about despite an experience of bitter exploitation and suffering.	12
While this study acknowledges that many factors were at work in producing the demographic contrasts of the Americas, the argument is monocausal in that it maintains that only sugar is a sufficient explanation of these contrasts. The search for a sharply focused explanation seems, moreover, to be important because knowledge of the particular pattern of causation has an impact on the interpretation of many aspects of slavery—for example, on slaveowner motivation, on slave morale and culture, and on the slave family. The argument focuses on sugar, but it is not intended to amount to geographic determinism or to what might be called crop determinism. It was the setting of sugar within specific contexts that was crucial, and the particular approach taken in this article—combining quantification with a critical case study (the Louisiana sugar enclave)—seems to provide an analytical breakthrough that shows the specific nature of the sugar process.	13
What was necessary to set up the long-run pattern of natural decrease was the combination of sugar, slavery, and access to a slave trade. Slavery on its own would not have produced vast regional patterns of natural decrease, and sugar without slavery was not enough,11 nor was the combination of sugar and slavery without a slave trade.12 Although labor on sugar plantations could take a heavy toll on health and on fertility, what was lethal was the combination of sugar with the slaveowners' ability to buy slaves and to choose a male-dominated labor force, rather than being content with family labor. In other words, the demographic problem stemmed from the priorities of the sugar planter. Sugar planters, unlike the great majority of owners, calculated that they could maximize profits by continually skewing their labor force toward men, and far-reaching demographic and social consequences and costs stemmed from this. Of course, owners who farmed other crops suited their own economic interests in the way they worked and recruited slaves, but their interests were not so demographically destructive. This article, then, focuses specifically on sugar planting but is intended more broadly as an interpretation of slaveholders, economic self-interest, and the consequences of profit-based decisions for life and labor under slavery.	14
In reviewing research on the problem of slave natural increase, John J. McCusker and Russell R. Menard commented: "The task is important . . . Moreover, the explanation for this growth is intrinsically difficult, demanding the examination of numerous factors affecting natural increase . . . Data are sparse and intractable, especially for the pre-Revolutionary era. Nevertheless, some progress has been made, and given the intensity of work on the issue, more can be expected in the near future."13 C. Vann Woodward, in an earlier but still valuable review of the field, pointed to the	15
difficult . . . problems of locating and weighing the almost unlimited number of influences that played a part in determining population growth. The mention of a few such problems will serve to suggest the complexities and difficulties involved. Without regard to priority or relative importance, there come to mind the equations of man-land ratios and man-woman ratios, land fertility and human fertility, rates of mortality and natality, the incidence of disease and famine, the relative advance of medicine and hygiene, comparative health conditions in tropical and temperate zones, tribal origins and mating customs of slave populations, and religious and legal traditions of masters. Complicating all these influences would be comparative stages of economic and technical development . . . On top of these would be added the imponderables of master-slave relations, the patriarchal ethic in tradition and practice . . . Beyond these determinants lies the whole range of cultural limits to security of family [and so on].14
McCusker and Menard added a further complication to the debate: "In a sense the West Indies-mainland divergences are deceptive, however dramatic they might seem. Slave populations in all of the chief plantation areas of British America [and they could have added of the Americas more widely] exhibited a similar growth process." The process was one of long periods, sometimes centuries, of substantial slave importation and natural decrease, but of a movement into natural increase, with the transition to natural increase occurring in Maryland and Virginia by about 1710 (and probably somewhat later in South Carolina), in Barbados not until 1810 (after the ending of the slave trade), and in Jamaica only by about 1840 (after the ending of slavery itself). As McCusker and Menard suggested,	16
The distinction between the continental [North American] plantation districts and the West Indies was not a function of the ways that change occurred, which seems to have been everywhere similar, for whites as well as for blacks. Rather the differences depended on timing, on how long it took for a reproducing creole majority to emerge and to reverse the net natural decline . . . Recognition of this shared experience changes the issues to be resolved: the need is not only to account for the differences between the West Indies [together with the rest of the Americas] and the mainland but also to explain why a similar course took longer to complete in some regions than others.15
When trying to resolve the problem of slave demography in the Americas, the factors to be considered are therefore numerous, and the issues are complicated. Partly, the question is, why was the demographic experience of Caribbean and Latin American slaves so bad? Partly, the question is, why was slave natural increase in North America not just positive but spectacularly positive? But partly, too, we need to consider why, despite fundamentally different outcomes, the North American demographic pattern actually seemed, for long periods, to run in parallel with that of the rest of the Americas.	17
Although a great many factors have been cited by researchers, we can conveniently group interpretations into three basic lines of argument. First, there is the argument that natural increase was strongly influenced by factors largely extraneous to slavery, including the nature of disease environments and patterns of natural abundance. Second, there is the argument, or set of sometimes conflicting arguments, that increase patterns were determined by factors much more directly under the control of slaveowners. Here, we might include the attitudes of owners toward slave buying, toward families and children, toward welfare and profit, and owners' decisions about which crop to grow. A third type of argument, and one that has grown in influence in recent years, emphasizes the behavior of slaves, especially their attitudes toward family and toward traditions that might limit fertility. Although, in practice, historians have often combined elements from more than one of these broad classifications, the three-type division of arguments still provides a useful framework for developing the present analysis.16	18
Indispensable research by Philip D. Curtin focused significantly but not exclusively on the role of disease environments, one of the forces extraneous to slavery.17 Curtin argued that the problem with migration was not the climate that one encountered but the disease environment. He noted that an unfamiliar disease environment, in whatever climate, is potentially lethal since natural resistance will not have had the chance to build. Curtin argued that, because slaves could be bought cheaply in Africa, slaveholders in the Americas bought extensively (buying mainly males), but he added that this pattern of slave recruitment set up a train of interactions. Slaves imported into a new disease environment suffered enormously high mortality (perhaps one third dying within the "seasoning" period of the first year or so).18 Moreover, birth rates were low for those who survived, because ill health would inhibit fertility and because there were relatively few women to bear children.	19
As a result of the interconnections he outlined, Curtin argued that the "South Atlantic System" tended to develop a two-stage pattern. First, in periods of rapid economic development, there would be intensive importation of new Africans—but there would also be very marked natural decrease (because of the encounter with an unfamiliar disease environment and because of the other factors just mentioned). Theoretically, a second stage might develop, when a period of slower economic expansion brought a reduced rate of African importation and brought lower mortality, more women, a higher birth rate, and a movement toward natural increase.	20
Curtin's model is of great importance in highlighting many of the processes at work in plantation demography. We should, however, not minimize the fundamental differences between the demographic experiences of the Americas. There was in fact no inevitability about the process of moving from reliance on the slave trade. Sugar planters would always demand a male-dominated labor force and would always try to import slaves whenever a supply was open to them. In Louisiana, even when the importation was internal (from the Upper South and not from Africa), the sugar plantations still experienced persistent natural decrease.	21
Disease environment, especially following Curtin's work, forms a part of many academic explanations of slave demography, but especially in the British West Indies there has been a long tradition of emphasizing the direct and negative role of slaveholders. For example, Orlando Patterson, in his classic study of Jamaica, argued that Jamaican owners drove so hard for profit that the slave experience was close to the Hobbesian state of nature, with life being "nasty, brutish, and short." Patterson, like many other writers, argued that owners preferred to buy, not breed. Owners, supposedly, were not interested in slave families, and found child-rearing a wasteful distraction from the woman's role as worker. Consequently, Jamaican planters adopted an "anti-natalist" attitude, often punishing women who became pregnant and discouraging women from spending time rearing children. Furthermore, he argued, the rigors of the regime meant that the slaves were so demoralized that they, too, were uninterested in family and in looking after children.19	22
In the case of the United States, some historians, for example Richard Sutch, have argued that owners interfered with fertility, but in this case it has been claimed that they did so in order to promote rather than to discourage the rearing of children. Indeed, a long abolitionist tradition maintained that U.S. natural increase was so high because slaveholders, on the supposedly worn-out lands of the Upper South, made their profits by "breeding" slaves for the Lower South market.20 These and other arguments will be reviewed later in this essay.	23
In a series of publications stretching over many years, Robert Fogel and Stanley Engerman have not only rejected the idea of U.S. slaveholders as manipulative slave breeders, but they have seen owners in the United States as providing a rather benign material environment for their slaves. Indeed, the cumulative effect of Fogel and Engerman's work has been to suggest that the great demographic contrasts of the Americas (or more specifically in much of their work, the contrast between the United States and the British Caribbean) did not come about because of the negative or manipulative influence of the slaveowners of the United States and elsewhere. Instead, the contrasts are said to have been produced partly by the supposedly benign material environment that was provided in U.S. slaveholdings and partly by the influence, in the West Indies, of African cultural traditions (concerning breastfeeding), which are supposed to have greatly reduced fertility.	24
In North America, then, Fogel and Engerman argue, slaveowners encouraged families and welcomed the natural increase of their slaves.21 The diet of their adult slaves, supposedly, was abundant, and this, Fogel and Engerman argue, was reflected in the exceptional average height of U.S. slaves.22 They argue, too, that the West Indian problem was not one of exceptionally high slave mortality. Indeed, they suggest that slave mortality was only slightly higher in the Caribbean than in the United States.23 What they see as making the great difference was not mortality but fertility. West Indian slave fertility was low, and they attribute this to a great extent to the cultural traditions of imported Africans. Africans, they suggest, brought over traditions of long periods of breastfeeding (usually breastfeeding for at least two years). They suggest that this practice inhibited fertility and was to a great extent responsible for natural decrease in the West Indies.24 A later section of this essay will critically review this argument. My own interpretation differs very much from Fogel and Engerman's: it points to slaveowners, the profit motive, and to the different crop regimes of the Americas.	25
The Louisiana sugar parishes provide a very special opportunity to interpret the demographics of slavery. In slave regimes outside the United States, it is nearly always impossible to gain access to detailed and essentially reliable census data on slave populations. The Louisiana sugar parishes, however, are closely documented, decade by decade, by excellent census data. This evidence allows us to calculate crude population growth rates, ratios of children to women, and (with other evidence) numbers of slaves imported.25 From this sort of evidence, we can then establish the approximate rate of natural decrease of the sugar parishes. (For the location of the thirteen Louisiana sugar parishes that accounted for the vast majority of U.S. cane sugar production, see Figure 1.) In fact, the Louisiana case study allows us not just to establish the broad conclusion that sugar planting produced persistent natural decrease, but it also makes it possible to show how that decrease came about—with evidence on the relative contributions made by excess adult mortality, a sex imbalance, and the shortage of children.	26
FIGURE 1. The shaded area indicates the thirteen leading sugar parishes. In 1850 and 1860 respectively, these parishes produced about 80 and 70 percent of Louisiana's sugar: Ascension, Assumption, Baton Rouge West, Iberville, Lafourche, Plaquemines, St. Bernard, St. Charles, St. James, St. John Baptist, St. Mary, Terrebonne, and Jefferson.
Later sections of this essay will demonstrate in more detail the Louisiana sugar parishes' specialized and highly significant pattern of slave importation, but two essential elements should be noted at this stage. First, throughout the antebellum period, these parishes imported massively, but, unusually for the sugar regimes of the Americas, the slaves were not from Africa and the Atlantic slave trade.26 Instead, these were American slaves, bought from long-established slaveholding states such as Virginia, Maryland, and the Carolinas. The second key feature of the sugar parishes' importations is that two-thirds of the slaves were males. Such a pattern, though routine in the Atlantic slave trade, will be shown to be unique in the interregional slave trade of the United States. All other parts of the Lower South, including the neighboring cotton parishes of Louisiana, bought roughly equal numbers of males and females.	27
The sugar planters of Louisiana, unlike those who grew cotton, tobacco, or rice—the dominant U.S. staples—persistently imported far more males than females, and they did this because they wanted a male-dominated labor force. Such an unbalanced force could not be achieved naturally but only by importation. These labor requirements meant that there was a routine pattern across the Americas: everywhere, sugar planters demanded a male-dominated work force, so wherever a trade in slaves was available, they would seek to tap into that trade. The insistence on a mainly male work force came about because of the extreme physical demands that sugar planters imposed on their workers, demands that will be documented in the next section. And the combination of severe labor demands with a male-dominated population was demographically lethal. In the Louisiana sugar parishes, excessive labor meant that "net nutrition" was far from adequate.27 It also meant "excess" (above average) adult mortality, and that women were often not fit enough to produce healthy children. Of course, the regime also meant that the population had few women, so there were few potential mothers. We shall see, then, that the labor regime in the Louisiana sugar parishes meant a persistent pattern of deaths exceeding births.	28
FIGURE 2. Grinding sugar cane in a windmill. Reproduced by permission of the British Library, reference number 1786 C.9.PLV.
Local agricultural societies routinely reported intensive slave importation combined with natural decrease among slaves. Indeed, evidence later in this essay suggests that slaves on the region's plantations suffered natural decrease of about 13 percent per decade.28 We know, however, that even this experience would have been better than that of typical sugar plantations of the Caribbean and South America. In those areas, the problems that Louisiana experienced would have been greatly compounded by the fact that the imported slaves nearly always came from Africa, and so suffered the dreadfully high mortality levels of the "seasoning" period.	29
It is not new to point to the connection between sugar and natural decrease, but previous work has not been able to establish that it was essentially sugar that made the difference in the demography of slaves across the Americas. It has also not been at all clear how far the problem was sugar planting—and how far it was, for example, the African slave trade or the tropical environment in which sugar cane was usually grown. In addition, if sugar planting was to blame for natural decrease, it has not been clear why slaves in Barbados, an important sugar island, managed to achieve positive natural increase by about 1810.29 Moreover, much quite recent work, especially by Fogel and Engerman, has tended to take the focus off sugar and the severities of labor and to put it instead on African traditions and the special circumstances of the United States.	30
Barry Higman has conducted detailed and important work on slavery and demography in the British Caribbean but has pointed only very cautiously to the major role that sugar might have played in producing natural decrease. Some brief comments on his primary source base will show the problems of drawing decisive conclusions from the published primary sources of the British Caribbean. Higman's source base was the great mass of slave statistics that, in the last few years of slavery in those colonies (after the closure of the African slave trade in 1807), the British Parliament demanded. These "Registration" records, however, by covering the period after the African slave trade had been ended, represent an exceptional phase of Caribbean slavery. With the "seasoning" of new Africans no longer a factor, mortality levels would have been far lower than usual, and, with the African slave trade closed, male-female ratios would have been much more balanced (which would have tended to raise the birth rate). Higman therefore found various anomalies in the British Caribbean evidence after 1807. He observed, "Wherever slaves were not engaged in the production of sugar their chances of survival were greater." But his overall conclusion was cautious: "The complexity of the interactions involved in the strongly contrasting patterns of fertility, mortality, and natural increase found in the British Caribbean after 1807 suggests that any attempt at monocausal explanation is doomed to failure. The most that can be expected is a more precise ordering of the variables, only some of them quantifiable, and a clearer understanding of the differentials."30	31
FIGURE 3. Slaves working in a sugar-boiling house. From an eighteenth-century print.
The evidence that can be developed for the Louisiana sugar parishes has major advantages over the West Indian "Registration" evidence, and allows me to be much more certain that it was sugar that set up the great demographic contrasts between North America and the rest of the Americas. This case study allows control of the influence of factors like the African slave trade, disease environments, sex imbalance, African lactation (breastfeeding) traditions, and slaveowners' attitudes. A detailed examination of the Louisiana sugar enclave will be developed, first by examining the nature of the labor regime on sugar plantations and the overall composition of their slave populations, and then by exploring the character and scale of slave importation. By calculating the extent of its importation (and scaling down the population of the area accordingly), it will be possible to estimate the sugar parishes' rate of natural decrease. After these calculations, the final sections of this essay will return to the wider framework of the Americas, and will reflect on the overall demographic and social impact that the sugar regime had on slave societies.	32
Visitors to Louisiana often commented on the severities of the sugar plantations, and, among slaves, the cane fields were places of dread. Throughout the year, the crop involved much exhausting work in such tasks as ditching and draining, digging out old cane and making holes for new ones, spreading manure, and chopping and hauling wood for the sugar house. But the "grinding season," usually from October to January, was by far the cruelest time for sugar workers. Then, the cane cutting and hauling had to be synchronized so that grinding and boiling at the sugar mill could go around the clock. Unlike cotton, the sugar crop would not keep: it could not be stored until a neighbor might lend labor or equipment to prepare it for market. A sugar planter had to invest heavily in the machinery of the sugar mill and, to generate good profits, had to use that machinery to the optimum. In this way, the crop, when cut, would not be wasted, and the canes still standing could be milled before severe frosts destroyed them. These extreme labor demands meant that sugar planters wanted strong workers, especially adult males. As U. B. Phillips noted, "All the characteristic work in the sugar plantation called mainly for able-bodied laborers. Children were less used than in tobacco and cotton production, and the men and women, like the mules, tended to be of sturdier physique."31	33
FIGURE 4. Images of the sugar plantation from an antislavery tract. Amelia A. Opie, The Black Man's Lament (1826).
In 1833, Thomas Hamilton, a British visitor to Louisiana, recorded his impressions of the sugar plantations. He noted that the sugar cane, when ripe, was not considered safe "till it is in the mill, and the consequence is that when cutting cane begins, the slaves are taxed beyond their strength, and are goaded to labour until nature absolutely sinks under the effort." Similarly, R. W. Harris and other expert witnesses testified in evidence to the U.S. Treasury in 1846, "The cultivation of sugar requires more indefatigable labor than any other production . . . not a moment must be lost; [it] requir[es], also, about seventy days' labor, of eighteen hours each, during the boiling season." Slaves across the South had also heard of Louisiana and its sugar plantations. A slave song noted down by Edward S. Abdy, a British traveler, reflected the infamy of the sugar plantations:	34
I born in Sout' Ca'lina, Fines' country eber seen. I gwine f'om Sout' Ca'lina, I gwine to New Orleens. Ole boss he discontentum, He take de mare, Black Fanny, He buy er peddler wagon, He bound fer Lousy Anna. Chorus: Ole debble Lousy Anna, Dat scarecrow fer po' nigger, Where de sugar cane grow to pine trees, An' de pine tree turn to sugar.32
The antebellum period saw a dramatic expansion in Louisiana's sugar production and in its slave population. Table 1 plots the rise in the area's slave population from about 20,000 in 1820 to about 90,000 in 1860. The advance of the sugar kingdom was especially dramatic between 1818 and 1830, with contemporaries estimating that for much of this period sugar planters bought new slaves (from the eastern states) at the rate of 5,000 per year.33 The 1830s saw some slackening in pace, but in the 1840s there was a massive increase in the volume of the sugar crop (and a leap of 200 to 250 percent in its value). This latter growth could only have been achieved by a very high level of slave importation. The 1850s saw great annual fluctuations in sugar production, but overall there was a further doubling of output (and a 150 to 200 percent increase in value).34 Again, massive slave importation was required.	35
TABLE 1 Natural Growth Rate of the U.S. Slave Population Compared with Total Growth Rate (Including Imported Slaves) of Louisiana's 13 Leading Sugar Parishes     Source: Derived from U.S. Census.
Behind these production trends lie highly significant demographic patterns. First, in the booming 1820s and 1840s (see Table 1), the sheer number of importations into the leading sugar parishes was so great as to create crude growth rates something like twice or four times the natural increase rate of the U.S. slave population as a whole. In aggregate statistics, sugar's underlying natural decrease problems were buried by massive importation. But in the 1830s and 1850s, despite extensive slave buying, the area was not even able to achieve the growth rate expected for a population that had no importations.	36
Table 2 allows us to investigate this situation further and focuses on birth-rate problems. This table points to three key contrasts between sugar slaves and slaves in the United States as a whole. What stand out in this birth-rate evidence for the sugar parishes are a significant shortage of women, an extreme shortage of children, and very low fertility rates. All of these are intimately linked to the area's highly selective interstate importation of slaves. The shortage of women for childbearing would have restricted population growth, and the low fertility of the women who were present would have exacerbated the problem.35	37
TABLE 2 Demographic Comparisons between Slaves in Louisiana's 13 Leading Sugar Parishes and Total U.S. Slave Population     Source and notes: Derived from U.S. Census.      *for ages specified      †The 1820 census used a 0–13 years categorization, and these data were scaled to estimate the numbers of ages 0–9.
We can gain a preliminary indication of the significance of these factors by taking a snapshot of the area's population in 1860 and comparing it with the typical pattern for the South's slaves. Compared with what might have been expected in a typical southern slave population of the sugar area's size, the sugar parishes in 1860 had 13,500 fewer children. Low fertility and excess mortality among the children under ten seem to have accounted for some 60 percent of this deficiency,36 with the shortage of potential childbearing women being responsible for the remaining 40 percent.37 These figures suggest that child shortages were a major part of the sugar area's demographic problem.38 Evidence in the next section will also point to problems of excess adult mortality.	38
The special character of the sugar crop shows up very clearly in the interstate slave trade of the United States. As slave traders' records indicate, the domestic trade to all parts of the South was age selective (concentrating mostly on teenagers and on young-adult slaves).39 But a study of thousands of traders' bills of sale shows that only in the case of southern Louisiana was the domestic traffic sex selective. As I noted earlier, the trade to the cotton parishes of Louisiana, and indeed to all parts of the South except southern Louisiana, carried approximately equal numbers of males and females.40 Demographic calculations using the "survival rate" technique allow detailed sex-specific (and age-specific) estimates of interstate slave movement, and these calculations emphasize the uniqueness of domestic importations into the sugar area.41 Table 3 uses this technique for the 1850s and shows that males comprised over 67 percent of importations into the thirteen leading sugar parishes. It also shows that in the remainder of the state, because it only produced some 30 percent of the decade's sugar crop, males accounted for no more than about 54 percent of importations.	39
TABLE 3 1850s Slave Importations into Louisiana's 13 Leading Sugar Parishes Compared with Importations into the Rest of the State: Data Indicates the Sex and Age Composition but Not the Full Extent of Importations     Sources and notes: Calculations are derived from the U.S. Census, and use the "survival rate" method. This method, because of the sugar parishes' "excess mortality," will very seriously underestimate the scale of slave importation into that area (see footnote 41). It does, however, indicate the essential age and sex patterns of its importations.      *Excludes totals with minus sign.      †The minor sugar areas account for the raising of the male percentage above 50. The 13 leading sugar parishes produced about 70% of the state's sugar in 1860.
Slave trading to the sugar parishes was specialized since not only were there far more male than female slaves but the slaves had to be especially sturdy, and an exceptionally low proportion of children was carried. An important part of this traffic was a well-organized coastal trade from the Chesapeake ports to New Orleans, and the printed circulars of the several large-scale Richmond (Virginia) auctioneers regularly pointed up the need for specially selected "shipping Negroes" for the New Orleans trade.42 One such circular, in October 1850, reported that several major New Orleans traders had arrived, and "good shipping men are in demand." Similarly, a Pulliam&Davis circular of October 1854 began: "This is to inform you that negroes are selling as follows . . . No.1 young men 18–22 years mostly in demand also girls 16–20 years heavy set and very smart, suitable for shipping purposes." Thomas A. Clark, like the thousands of other traders, was well aware of the special nature of the New Orleans market, and in February 1846 wrote to a Richmond auction house: "I am sorry that I have not got any good negroes on hand that will suit the New Orleans market . . . Likely young men such as I think would suit the New Orleans market are very hard to find and also stout young women." A Betts&Gregory circular for January 1861 again reflected the trading community's awareness of the New Orleans phenomenon. It reported, "Our market continues dull except for first rate negroes. There are several persons here now making up lots for the New Orleans market and if you have any on hand now is the time to bring them in."43	40
The South Carolina trader John H. Charles also knew all about the requirements of the sugar planters, and in April 1859 wrote from New Orleans: "There is some [men] looking after [buying] negroes the last two or three days and if I had 4 or [5] good men and women [I] could have sold them at fair prices. But they are sugar planters who want them mostly and they want stout black negroes."44 The reference to "black" Negroes was also significant, planters associating the darkest skin with the toughest workers. The fact that across the South—from the Chesapeake, to the Carolinas, and to Kentucky and Tennessee—slave traders' advertisements referred to buying slaves "suited to the New Orleans market" suggests not only that this market was special but that suppliers everywhere knew what its special characteristics were.45	41
FIGURE 5. Newspaper advertisement of Kentucky slave trader buying for the New Orleans market. J. Winston Coleman, Jr., Slavery Times in Kentucky, copyright 1940 University of North Carolina Press. Used by permission of the publisher.
Appendices to this study give detailed estimates of the numbers of slaves imported into the sugar parishes in the 1840s and 1850s. These estimates take account of several supply routes. Many of the area's slaves were bought at slave pens in New Orleans—and that city was supplied not only by the coastal trade from the Chesapeake and the eastern parts of the Carolinas but also by the overland trade from the Atlantic seaboard states and by the traffic down the Mississippi Valley from Kentucky, Tennessee, and Missouri. In addition, some traders bypassed New Orleans and took their slaves directly to the sugar parishes. No single document type, taken on its own, gives anything like a full picture of the extent of the trade, but by combining various sources—including ships' manifests of the coastal trade and traders' estimates—a major part of the traffic can be reconstructed. Appendix 2 outlines the primary sources used in calculating importations.	42
A detailed calculation of the volume of importation in the 1840s is outlined in Appendix 3, and that appendix combines the totals brought by the coastal, overland, and Mississippi River routes. This calculation suggests that in the 1840s the leading sugar parishes imported a total of at least 27,000 slaves. Scaling down the area's 1850 population to allow for these 27,000 or more importations gives the level of natural decrease experienced by sugar slaves in that decade. It suggests that, instead of the typical 1840s natural increase rate of 27.8 percent, the sugar area experienced at best a 6 percent natural decrease (and in fact the situation was probably substantially worse than this).46	43
Slave importations in the 1850s must also have been high, and again in these years there must have been major natural decrease. Preliminary indications of the high levels of 1850s importations are given, as we have seen, by the fact that sugar production expanded massively in this decade (requiring much new labor),47 and heavy importation is also indicated (see Table 2) by the high proportion of males in the sugar area's working-age population of 1860. Appendix 4 gives detailed evidence on the level of importation in the 1850s and suggests that the total would have been at least 27,000. Similar calculations to those used already for the 1840s reveal that, in the 1850s, these parishes would have experienced a rate of natural decrease of over 13 percent. This contrasts dramatically with an expected positive rate of natural increase of 23.4 percent (see Table 1) for typical U.S. slaves in the 1850s.48 Evidence on importation indicates, then, that deaths greatly exceeded births among slaves in the main sugar parishes. The 1840s importation figure (of at least 27,000 slaves) probably represents a significant underestimate of the real rate of decrease because it incorporates a conservative estimate for the overland supply route. Actual rates of natural decrease in both the 1840s and 1850s, therefore, were probably at least 13 percent.	44
Having established the approximate levels of natural decrease on the Louisiana sugar plantations, we can now consider the relative importance of excess adult mortality and child shortage in bringing about this decrease.49 An extreme work load and inadequate net nutrition meant that slaves working on sugar plantations were, compared with other working-age slaves in the United States, far less able to resist the common and life-threatening diseases of dirt and poverty. An imbalance of the sexes, together with a shortage of potential mothers, would have been important in creating the sugar plantations' great shortage of children. In addition, the women who were present in the sugar area had far fewer children than those elsewhere in the United States. Indeed, they had far fewer than was the norm even for a slave-importing state, and an excessive work load would have contributed greatly to this.50 Slave women throughout the South faced the problem of insufficient relief from strenuous work during pregnancy, but on the sugar plantations the nature of work would have had a particularly damaging impact on fertility. This combination of pressures for women on sugar estates would have produced a particularly grim catalog of irregular fertility, still births, low birth weights, and poor infant survival rates.	45
Detailed calculations (see Appendix 5), taking the example of the 1850s, allow a fairly clear weighting of factors involved in producing natural decrease on Louisiana's sugar plantations. These calculations suggest that excess "adult" mortality (in this case, mortality among those who would have been age 10 years and over by 1860) accounted for some 44 percent of the gap between the sugar area's demographic performance and that of U.S. slaves generally. It follows that the shortage of children (ages 0 to 9 years by 1860) contributed the remaining 56 percent. We can also break down the child shortage into its component parts (again see Appendix 5). This disaggregation suggests that the scarcity of potential childbearing women would have been responsible for some 40 percent of the child shortage, so the combination of low fertility among the women present and high child mortality would have accounted for the remaining 60 percent. Overall, then, the child shortage was statistically more important than excess adult mortality, but both were major factors, and both were part of the extreme burdens of sugar slaves in Louisiana.	46
Moving back from the Louisiana test case to the overall debate on slave demography in the Americas, the detailed Louisiana evidence allows us to reflect on, and criticize, some of the interpretations outlined earlier in this study—the role of disease environments, anti-natalism, slave breeding, and the combination of arguments proposed by Fogel and Engerman. From the basis of the Louisiana evidence, it now seems to be possible both to reassert the case for the dominant role played by sugar planters and sugar plantations and also to resolve the puzzle concerning the timing of the transition to natural increase in different regimes.	47
First, let us consider the possibility that disease environments might explain the great demographic contrasts of the Americas. The Louisiana case study is very helpful here.51 It is true that southern Louisiana was, even for whites, not a healthy place, but it is also clear that the labor regime of the sugar crop, regardless of the local disease environment, was in itself sufficient to cause deaths to exceed births. New Orleans was known as the yellow fever capital of North America, but, even though yellow fever was dreaded in New Orleans (and in other American cities), it only rarely attacked plantation districts and accounted for only a tiny percentage of plantation deaths.52 Cholera, too, was never of major statistical significance on the sugar estates, and although it continued to affect New Orleans, it had disappeared from the sugar plantations by the early 1850s.53	48
Malaria was for Louisiana slaves a far more significant problem than either cholera or yellow fever: even so, it did not dictate the natural decrease of slaves on the sugar plantations.54 It turns out (as Figure 6 shows) that almost all of the western half of Louisiana—including a massive tract of parishes that produced cotton, not sugar—was in the same high-risk malaria belt as the sugar parishes.55 Nevertheless, the demography of the cotton parishes, in similarly malaria-prone areas to the sugar plantations, was very different. Cotton areas did not have significantly more male than female slaves. And, crucially, while the fertility rates of slave women in the sugar parishes were exceptionally low by any U.S. standards, ratios in the western Louisiana cotton parishes were very high. Despite malaria, the cotton parishes, unlike the sugar plantations, had no child shortage.56 Malaria brought many deaths and much debility, but it took the labor demands of sugar to produce natural decrease.57 The local disease environment of the Louisiana sugar parishes was not what caused the pattern of natural decrease.	49
FIGURE 6. "Malarial Disease Deaths per 1,000 Population, 1870." This map is based on the Ninth U.S. Census, and is reproduced by permission of Cambridge University Press from Frank C. Innes, "Disease Ecologies in North America," in Kenneth F. Kiple, ed., The Cambridge World History of Human Disease (Cambridge, 1993), 527.
At the start of this essay, I also noted arguments that owners directly influenced fertility, either negatively in the Caribbean by discouraging slave births or positively in the United States by adopting systematic slave breeding. There is, however, little evidence to support either of these arguments. The fact that slave infants in the British sugar plantations had a positive cash value suggests that most owners, far from being anti-natalist, would have welcomed the birth of slave children. Indeed, in the British West Indies, the price of a one-year-old slave was about 5 to 10 percent of that of an adult male, and this was very similar to the situation in the United States, as well as to that of the Americas in general.58 It is significant, too, that Patterson's references to the anti-natalism of the slaves themselves came from the most unreliable of sources—racist slaveowners who sought to shift the blame for lack of increase onto blacks. Thus owners routinely claimed that slaves were irresponsible, had no sense of family, and were not prepared to care for children.59	50
In North America, abolitionists sometimes made the claim that slavery in the Upper South depended for its economic survival on stud farms, which bred slaves for sale to the Lower South market. There is little doubt that slaveowners took a great interest in the value, and therefore number, of their "stock." The fact that I have found, in thousands of highly detailed letters by interstate slave traders, no reference at all to breeding farms, suggests, however, that the slave-selling states did not adopt such specialist child-production farms.60	51
Fogel and Engerman, sometimes in separate publications, sometimes in jointly written studies, have combined several elements in their demographic argument—an emphasis on fertility (not mortality) differences between the United States and the West Indies, a claim that the Caribbean slaves' fertility was low because of African traditions (not planters' actions), and an argument that North American slaves enjoyed a particularly good diet and that this diet contributed substantially to the health, stature, and natural increase of those slaves. Each of these important claims demands attention.	52
Fogel and Engerman's claim that mortality rates were not substantially higher in the British Caribbean than in the United States was based on an adjustment of Higman's data for Jamaica in 1817 to 1834, but this selection of evidence raises problems.61 The period chosen was one, after the ending of the slave trade in 1807, when the extreme mortality levels associated with the "seasoning" of imported slaves were no longer a factor. Fogel and Engerman's findings on mortality are also surprising since sugar was vastly more important in Jamaica than in the United States, and Fogel and Engerman themselves found that mortality was particularly high on sugar plantations. Given the importance of sugar planting in Jamaica, one would have expected to see a strong reflection of this in the U.S.-Jamaica comparison.62	53
Fogel and Engerman's second claim—that the U.S.–West Indian fertility gap was brought about by the operation of African cultural traditions in the West Indies—does not fit with the Louisiana evidence. The slaves of the Louisiana sugar plantations suffered from very low fertility, just as those in the West Indies did, but in Louisiana African lactation practices were not an issue. Louisiana had imported its slaves from the Upper South, and they would not have had significantly more contact with Africa and its traditions than would typical U.S. slaves. Fogel and Engerman based their argument on lists of children with their mothers and on the age gaps that they found between (surviving) children. The greater spacing in the Caribbean compared to the United States does not, however, seem to have been significantly influenced by special lactation practices. Instead, what would seem to account for the longer gaps between the ages of surviving West Indian children are factors that were also at work on the Louisiana sugar plantations: high child mortality, high numbers of stillborn children, poor health and net nutrition of women (and therefore a reduced ability to conceive), and interrupted fertility resulting from the death of husbands in an environment of overall high mortality.	54
Fogel and Engerman's third claim relates to the height of slaves. They have drawn on evidence from several scholars—using height samples from the coastal domestic slave trade to New Orleans, from manumission documents, and from Union Army records. These various samples have all indicated that the average height of adult male slaves in the antebellum United States was about 67 inches. These results suggest that, by any population standards of the time, U.S. slaves were tall (and therefore well fed).63 However, all three of the data sets on height seem to use atypical samples, so the anthropometric evidence does not, in practice, appear to sustain the thesis that North American slaves enjoyed outstandingly good diets.	55
For instance, the sample of the New Orleans trade was developed by Richard Steckel, using ships' manifests describing slaves brought from Virginia and the Upper South.64 There is, however, extensive evidence that slave traders generally (and even more so those supplying New Orleans) sought a carefully selected subset of the slave population. They sought "likely," "prime," strong slaves, and they screened out as difficult to sell those who were "undersized," unhealthy, or weak.65 This suggests that typical slave men in the United States were not as tall as the 67-inch average found in the ships' manifests.	56
A further height sample, using Maryland certificates of manumission, was developed by John Komlos, and these data also suggested that men averaged 67 inches.66 This sample, too, is likely to exaggerate the height of slaves, since it is probable that those who enjoyed manumission would have been among the most privileged, and therefore best fed and tallest, of U.S. slaves. Robert Margo and Richard Steckel have used a third main data set—records of the height of blacks recruited into the Union Army during the Civil War.67 Again, it seems that the sample must have had an upward bias. The army normally had a minimum height requirement, and, in its recruitment of blacks, it is likely to have discriminated against those who were significantly below average height.	57
The data sets that Fogel and Engerman based their assumptions on all, therefore, seem to carry strong upward biases. It is extremely difficult to get direct samples of the height of typical field slaves, but the process of making shirts for two 1850s slave gangs has left detailed lists of the height, age, and occupation of fifty-five slaves (all of the males age 20 years or over in these two gangs). This sample is too slender a base for confident generalizations, but still the results are interesting. The average height for these fieldhands was 64.6 inches—far shorter than that for the trade and other samples mentioned above.68 The combination of direct evidence from these lists and reservations about samples taken from the trade and elsewhere suggests that slave men would not have been as well fed as has recently been claimed by Fogel and Engerman, and suggests that they would have been significantly shorter than 67 inches on average.69 The diet of U.S. slaves does not, after all, seem to be the cause of that population's exceptional rates of natural increase.	58
We can go further than this. If we compare some key demographic indicators for slaves and free whites, the explanation for the very high natural increase of U.S. slaves seems to be far more gloomy than that suggested by Fogel and Engerman. It is true that natural increase among U.S. slaves was, by the 1850s, higher than that of U.S. whites, but the life experiences of slaves and whites contrasted profoundly, and the causes of natural increase were very different for the two populations. Slaves experienced far higher rates of infant, child, and adult mortality than did whites, and this mortality, as in "Third World" countries, was compensated for by a very high birth rate. "Third World" population patterns do not suggest that slaves were especially privileged.	59
Work by Richard Steckel has revealed the dreadful childhood experience of U.S. slaves. He found that, as a result of overworked mothers and of desperately poor net nutrition, average birth weights for U.S. slave babies were very low (typically, less than 5.5 pounds), and the rate of still births among U.S. slaves was very high (probably no less than 11 percent). He also found major contrasts between the child mortality of whites and slaves. After discounting still births, he found that 28 percent of the antebellum southern whites died before age 15, but 46 percent of slaves died before reaching that age.70	60
Moreover, it is not just in child mortality where major contrasts between U.S. whites and slaves appear. Despite far higher infant and child mortality among slaves compared with whites, the slave and free southern populations showed basically similar ratios of children to women.71 This similarly suggests that, to balance its excess child mortality, the slaves' birth rate must have been far higher than that of the whites.72 A further statistic—the fact that there were relatively few slaves age 50 or over—suggests, too, that adult mortality was significantly higher for slaves than for whites.73	61
Data on height suggest another contrast in health. Oaths sworn by former South Carolina planters in the 1860s show the height of those slaveowners to have averaged 68.6 inches. If the "shirt lists" used in this study are anything like a useful indicator of typical slave heights, they suggest that owners would, because of far better net nutrition, have towered over most of their slaves by several inches.74 U.S. slaves, then, reached similar rates of natural increase to whites not because of any special privileges but through a process of great suffering and material deprivation.	62
The evidence presented in this study does not, therefore, support arguments that focus on the disease environment, anti-natalism, African cultural traditions, or especially favorable circumstances in the United States. North American slaves were driven very hard by their owners, but the special demands imposed by the sugar planters of the Americas were even more lethal. Sugar, of course, did not account for all slaves in the rest of the Americas, and pockets of natural increase were possible. Recent work shows, for example, that in the nineteenth century the Minas Gerais region of Brazil developed a diversified economy oriented largely toward domestic markets, and that the region is remarkable for the fact that its slaves experienced natural increase.75 Crucially, Minas Gerais was not a sugar region. It seems likely that further work will show that other non-sugar regions in the Americas also experienced natural increase. Nevertheless, sugar plantations dominated the Americas, dominated its demographic patterns, and dictated too that natural decrease would dominate.	63
Sugar planters, as we have seen, combined extreme labor demands with a long-term insistence on a predominantly male work force, and the combination of burdens they imposed was lethal. The Louisiana case shows that sugar produced natural decrease even when the labor force was American-born. Where (in Brazil or Jamaica, for example) sugar planters bought their slaves from Africa, natural decrease would have been even worse than in Louisiana. Sugar and the Atlantic slave trade represented the worst of all worlds—the extreme demands of the sugar plantation, the urge to import (so as to maintain a male-dominated labor force), and the intensive sucking of Africans into an unfamiliar, and therefore hostile, disease environment.