The history and nature of the slave regimes of the New World have been examined and hotly debated in national, regional, and comparative perspectives.1 Much of the controversy has centered on the treatment of slaves in the various slave regimes and under a variety of economic conditions. This debate was greatly clarified and advanced by Eugene Genovese in his article, “The Treatment of Slaves in Different Countries,” in which he distinguished between three basic meanings of “treatment.”2 Genovese recognized that the term “treatment” has been used at different times to describe 1) the day-to-day physical conditions of the slave; 2) the existential conditions of life, the opportunities for familial, social, and religious expression; and 3) access to freedom and citizenship. While these aspects may be interrelated there is no necessary connection between the first and third. Whereas a great deal of argument rages over the comparative history of the first two of these categories, over the third there seems to be general agreement. Both those who see striking differences between Latin American and North American slavery and those who find disheartening similarity concur that access to freedom was easier in Latin America, and that the liberation of slaves was a widely practiced phenomena.3 Once this has been said, however, controversy again emerges over the motivations and functions of manumission, that is, the voluntary freeing of slaves.

It is clear that very little is known about the process or how it operated. The basic questions of who, why, how, and how many have never been studied in depth. Instead, a series of “reasonable” generalizations, based on scattered evidence or inferred from the demographic contours of the population as a whole, has characterized most statements about manumission. Perhaps it is exactly the absence of detailed studies that allowed for so much agreement. Only in the last few years have scholars begun to turn their attention to the history of freed people of color within the American slave regimes and, by extension, to the process of manumission. The few studies done thus far have begun to complicate our vision of this phenomena by demonstrating its complexity and by raising a series of still unresolved questions about the nature and functions of manumission within the slaveowning regimes.4

Historians of Brazil have so often repeated the myth that the documents on slavery were destroyed in the fervor of abolition that they have sometimes led themselves to believe it. While it is true that much documentation is now lost, there remain a wide variety of sources that simply have not been adequately exploited. Among these are the notarial registers that have survived in many places in Brazil. These volumes record much of the life of Brazil at its most basic levels. Contracts, loans, sales, and other financial arrangements were all matters of legal record, that is, they all required notarized registry. This was also true of manumission, a juridicial action in which property rights were surrendered and in which the former slave assumed a new legal personality and new legal responsibilities.

The instrument of manumission was a document usually called the carta de alforria or carta da liberdade. In this document the slaveowner identified him- or herself and then identified the slave being freed. Often this identification included the slave’s age, color, place of birth, and, on rare occasions, the slave’s occupation. It was common for the manumission charter to comment on the reasons for the award of freedom or on any subsequent limitations or conditions to be placed on that freedom. If the master received money or some other form of payment in return for the grant, it, too, would be noted. The document drawn up by the master or his legal representative was then dated, signed, and attested to by two witnesses.5Libertos (manumitted slaves) would normally keep the original letter in their possession, since the illegal enslavement of people of color was always a danger, but in order to protect themselves and to fully legalize the transition of status, the document was then taken to the nearest notary and transcribed in his register as well.6

The value of the cartas de alforria is obvious. Survival of notarial registers has assured historians an excellent opportunity to study the phenomenon of manumission as it operated in colonial and imperial Brazil. The Arquivo Público do Estado da Bahia maintains a large collection of notarial registers from the captaincy of Bahia and especially from its capital city, Salvador. This notarial collection from the city of Salvador includes 1,384 volumes covering the period from 1664 to 1915. Despite a few gaps, the series is virtually continuous from 1684 to the end of slavery in 1888.

This article is a preliminary study of the data on 1160 slaves manumitted in 1015 cartas de alforria registered in notarial offices (cartórios) of the City of Salvador during the period 1684-1745.7 It is part of a larger collaborative project that will eventually analyze the full chronological range of the Bahian letters of manumission, from colonial times to the abolition of slavery.8 The data presented here are intended as an example of the kind of materials on colonial slavery that have not yet been fully exploited.

The Social and Economic Context

Like any other aspect of a slave regime, the phenomena of manumission must be examined in relation to the prevailing sociopolitical and economic situation. The locale of this study, Salvador da Bahia, was a major Brazilian port, founded in 1549 as the colonial capital. Its nearby agricultural zone, or Recôncavo, became a major producer of sugar and tobacco, the latter in the eighteenth century. The interior of the captaincy eventually developed as a zone of livestock raising. The demands of a sugar plantation economy made Bahia one of the major importers of African slaves during the colonial period. By 1724 it was not uncommon to find that in parishes in the sugar zone, 60 to 70 percent of the population was slave. During the course of the seventeenth century the sugar economy had rapidly expanded but by the 1680s foreign competition, a series of bad harvests, a fall in the price of sugar, and certain internal contradictions in the industry created a time of crisis.9 This situation was exacerbated in the early years of the eighteenth century by the discovery of gold in Minas Gerais, which increased the demand for slaves in that region. This resulted in a steep rise of slave prices in the rest of the colony and a fierce competition for new African imports between the mining and agricultural regions. State regulation of the slave trade, a lessening demand for labor in Minas Gerais, and readjustments in the sugar zone brought a period of relative stability by the 1740s. Thus, the data on manumission presented here covers the period of transition from economic crisis to stability.

It was the plantations and their demands that made Bahia one of the major slaveholding regions in the Americas, but the interaction between the Recôncavo plantations and the city of Salvador was continual. This special relationship between the administrative-port city and the export-oriented hinterland creates certain problems for our analysis. It was not uncommon for planters to maintain residences in the city, and there was constant movement of people and goods between the urban and rural areas. Until the opening years of the eighteenth century secondary cities in the Recôncavo did not have resident notaries, so that in order to legalize contracts or transactions of various types it was necessary to do so in Salvador.10 Thus, the fact that a carta de alforria was registered in Salvador does not necessarily indicate that the master was a resident of the city, nor that the slave was engaged in “urban” occupations. I have recorded any address listed for the master, but place of residence was given in only 277 cases and, of these, 75 percent of the listed residences were located outside the city. In 883 cases no address was listed. The very high percentage of recorded residences outside of Salvador leads me to believe that when no address was recorded, as was common, the proprietor was a resident of the city and, therefore, did not need more detailed identification. If this is the case, then 17.8 percent (207/1160) of the slaves manumitted were not from Salvador, and the remaining 82.2 percent were “urban” slaves.

The data on Bahian manumission are difficult to evaluate, not only because of imprecision in such matters as the residence of the manumitted slaves but also because of deficiencies in the general demographic statistics of the Brazilian slave population. The parish counts and censes taken in the captaincy of Bahia and in the city of Salvador usually recorded sex and color distinctions among the free population but rarely ever did so for the slaves.11 In the plantation zone more care was occasionally given to the sex, color, and age of the slaves, but it would be extremely tenuous to extrapolate the statistics of urban slavery from plantation figures.

During the period under study here, Salvador had a population of approximately 30- or 40,000 inhabitants, of which perhaps half were slave. Table I presents the population of three central parishes in the city, according to an ecclesiastical account made shortly before 1724. This same account fisted a population in the Recôncavo plantation zone of 35,672, of which 22,422 or 62.8 percent were slaves.12 Unfortunately, this parish count, like most of the colonial census materials from Bahia, does not break down the slave population by color or sex.

Faced with this deficiency, a few demographic characteristics of the colonial Brazilian slave population can be set forth here as a general index of these features. Brazilian slavery depended on the constant importation of new slaves from Africa. Most authors agree that the sex ratio in the Atlantic slave trade displayed a rather constant imbalance of at least two males to every female.13 Thus, we should expect to find a disproportionate number of males in the adult slave population, since African-born slaves appear to have consistently comprised over half the total slave population. Statistics of the Portuguese slave trade also indicate that children made up only ten percent of the slave cargos.14 Therefore, slave children will be predominantly Brazilian-born. As to color, Table II presents a crude estimate for the years 1798 and 1821 which, along with census figures from Minas Gerais also for 1821, indicates that mulattoes (pardos) constituted between 10-15 percent of the slave population. Even in 1872 when the first national census was taken, mulattoes comprised only 32 percent of the slave population. Thus, during the colonial era the slave population would have been predominately black and predominantly male, but whether these general characteristics also held true in a city like Salvador remains a central and still unanswered question.15

Anyone familiar with the historiographical debate on the comparative history of slavery in the Americas is aware that an important group of historians following the path blazed by Frank Tannenbaum and Gilberto Freyre have placed great emphasis on certain cultural features of the Iberian regimes as keys to an understanding of slavery within them. Chief among these features have been the strength of a formalized Roman law tradition and the role of the Roman Catholic Church in protecting the slave.16 Unfortunately, neither of these cultural traditions offers the historian much help for an understanding of manumission, as it was commonly practiced in colonial Brazil. Despite the central position ascribed to the law as a protector of the bondsman in the Iberian regimes of slavery, Portuguese law was peculiarly silent on the nature and regulations of slavery as it operated in Brazil. The few references incorporated in the Ordenações filipinas (1603) were based on earlier codes clearly originating in a time when slavery was still associated with Moors and other non-Christians.17 These ordinances provide virtually no guide whatsoever to the nature or reality of the Brazilian slave regime. Subsequent royal legislation (leis extravagantes) and the myriad of confused and sometimes contradictory municipal ordinances draw us closer to the reality of Brazilian slavery, but these are often silent on the most commonly accepted or practiced aspects of the regime. Thus there is virtually no set of laws or ordinances that guide the historian’s approach to the question of manumission in its varied aspects, and this is especially true of the colonial era.18

The lack of a civil law slave code did not mean that the Church exercised authority in the regulation of slavery. There was, in fact, no coherent body of Church doctrine or ecclesiastical statute that regulated slavery. The synod held in Bahia (1707) codified existing practices in Brazil and set forth religious guidelines in the Constituições primeiras do arcebispado da Bahia that were quickly adopted throughout the colony. The topic of slavery was a matter of concern, and forty canons of the Constituições dealt specifically with slaves and their life. For example, slaveowners were urged to permit their bondsmen to marry and to maintain slave families intact. Other canons, however, deprived escaped slaves the right of sanctuary and explicitly stated that a slave’s acceptance of Christianity or the sacrament of marriage did not release the slave from bondage. In general, little attention was devoted to the practices of Brazilian slavery, and the question of manumission was not raised in the Constituições.19 Even when ecclesiastical authors did turn their attention to the plight of Brazilian slaves, the issue of manumission did not often concern them.20 The failure of ecclesiastical canon or civil statute to provide a guide to the manumission process forces us to turn to the cartas de alforria themselves as the most accurate introduction to the norms as well as to the process and patterns of manumission operating in colonial Brazil.

Patterns of Sex, Age, and Ethnicity

In view of the expected characteristics of the Brazilian slave population, the most striking pattern to emerge from the colonial manumission records is a constant 2:1 ratio of female to male libertos (freed slaves). This ratio is present among both the “urban” and “rural” sections of the sample, although the tendency to favor women in the manumission process is slightly elevated in the rural area. The independent studies of Mattoso of the period 1779-1850 and of Kessler for 1813-1850 demonstrate the same imbalance. Thus it appears that the 2:1 ratio of female to male libertos was a constant feature of manumission as it operated in Bahia.21 Given the apparent predominance of males in the overall slave population, we should emphasize here that females were achieving their freedom at a much greater rate than the statistical expectation, but because we do not know the ratio of females to males in the city’s slave population, the significance of this rate cannot be determined.

In Brazil slaves were distinguished by color and place of birth. Traditionally, a tripartite division separated slaves into the categories of Africans (assumed here to be black), crioulos (Brazilian-born blacks), and pardos (people of mixed racial origins). This last group included not only mulattoes but also white-Indian offspring referred to variously as mestigos, mamelucos, or caboclos. Also included here were the cabras (persons of mixed but undefinable heritage). Table IV shows the distribution of the sample of manumitted slaves among these categories. It is possible to identify the color/ethnicity of 950 libertos in the sample. Of these 54 percent were Brazilian or African-born blacks, while 43 percent were mulattoes, and 3 percent were mestiços or cabras. In other words, there was a roughly equal distribution between blacks (514, or 54%) and pardos (431, or 46%). Because the general demographic patterns of Brazilian slavery indicate that mulattoes constituted perhaps only between 10-20 percent of the slave population, it would appear that slaves of mixed origin were decidedly advantaged in the manumission process. Here it appears that the often repeated generalization about the favoritism shown to mulattoes holds true. Subsequently we shall deal with the reasons for this.

If the data is organized not by color but by place of birth, another interesting phenomenon can be seen. Brazilian-born crioulos and pardos made up 69 percent of the total freed, while Africans were only 31 percent during the period 1684-1745. This distribution appears to be an inversion of the ratio between Brazilian and foreign-born slaves in the total population.

Information concerning the specific ethnic origins of African libertos is usually so imprecise as to make impossible an analysis of the comparative ability of certain peoples to obtain ther freedom. Much of the problem stems from Portuguese confusion over African geography and ethnography. I encountered occasional references in the cartas to “negros de guiné de Angola.” Such phrases indicate a general imprecision of terminology as well as the use of the term “guiné” in a most general way. Wherever possible, the most precise definition has been used for analysis of African origins.

The African background of the libertos reflects general tendencies and patterns of the Portuguese slave trade to Brazil. In the period prior to 1725 the majority of African-born slaves who appear in the cartas are called “negros de guiné.” Despite the imprecision of this term we have followed Curtin’s definition and location as the region of Senegambia and Guinea-Bissau. Of the 105 slaves recorded as “guiné,” only nine appear in the cartas after 1725. This distribution reflects the general tendency of the slave trade to shift away from the Senegambia toward the area of the Gold Coast and the Bight of Benin. The situation of “guiné” libertos is reversed in terms of the so-called Minas. Portuguese ability to trade under license of the Dutch at El Mina and to take slaves in the ports of the Bight of Benin provided a major source of slaves in the eighteenth century. This was especially true after the establishment of a trade fort at Ouidah in 1721. The Portuguese trade from the costa da Mina became a major source of slaves in Bahia. Whereas only 30 Minas appear in the cartas before 1725, I found 82 recorded between 1726 and 1745. The other major region of slave export is the area south of the Congo River, which we have called Congo-Angola. Here, the numbers are not large, but the distribution over time is relatively even. One last feature should be noted: the sample for the period before 1745 included very few individuals (15) from the Bight of Benin area. In the later eighteenth and nineteenth centuries this region provided the large numbers of Nagos (Oyo Yoruba), Geges (Dahomeans), and Hausas who considerably altered the demography of Bahian slavery.

In evaluating the cartas de alforria no single characteristic of the libertos is more difficult to record and analyze than age. It was common in the cartas to make some statement concerning the age of the very old and the very young but the exact age of adults was not a matter of record. The problem is complicated by the custom of assigning a descriptive rather than a numerical age to the slave. Diminutives are often used to describe children or adolescents (mulatinho, crioulinha) and terms such as rapaz (young man), moga (girl), and moleque are also employed. Many of these terms then and now are by no means precise, and they present great difficulty for the creation of exact age categories. To further complicate the situation, it is not clear at what age in colonial Brazil children became functional adults. A plantation survey made at the end of the colonial era in Bahia listed no slave children above the age of nine years.22 Adulthood, or at least adult obligations, seems to have come at an early age, but whether this was as true in the city as on the plantation remains to be established. Since the majority of slaves freed were women, I have used puberty rather than any distinction based on work capacity as a division between children and adults.23 Slaves described as mulatinhos, etc. or rapaz, moleque, menino have been placed into the cohort 6-13, although it should be recognized that some may be younger than six years and a few perhaps older than thirteen. Nevertheless, when the cohorts of 0-5 and 6-13 are considered together they present a fairly accurate total of the non-adults actually manumitted.

Establishing the age of adults is even more difficult. When, however, terms such as man (homem) or woman (mulher) are used, or when it is clear that the slave is married or has children, the slave has been placed in the 14-45 age category. Here, too, the limits are arbitrary, but given the many traveler’s references to the severity of slave life and the high rate of mortality, 45 years seems to be a reasonable, if not overly generous, threshold of old age. A final complicating feature of these data is the large number of slaves for whom no age can be determined. When data on the slaves’ value was available, I have listed slaves whose age is unknown but whose value was greater than 50 milréis in the 14-45 cohort (263). Unfortunately, this still leaves 397 cases for whom no information is available. I believe that the majority of these people would fall into the 14-45 cohort and for later purposes of calculation they will be included there.24

It should be clear that the data on age are by no means exact and that at best they can provide only an indication of tendencies and trends. Table V indicates the distribution of the libertos by age. The large number of children and adolescents should be noted. Slaves younger than 14 years constituted almost 45 percent of the total for whom age could be established, and even when the 397 slaves without age designation are added in their entirety to the adult categories, children and adolescents still constitute 29.5 percent (342 cases) of the total.

Males were more likely to be freed as children than were females. Whereas only 24.9 percent of the female libertos were manumitted as children, 38.8 percent of the males freed fell into this category. If we keep in mind the overall 2:1 female to male ratio, while recognizing that males surely constituted a smaller proportion of the cohorts 0-5 and 6-13 than they did of the total slave population, the general tendency to free males as children is clear. The tendency to free male children can be explained by a combination of factors. The high rate of infant mortality probably depressed the value of slave children. This economic consideration, when added to feelings of affection toward children, probably moved slaveowners to manumit slaves at an early age. Upon reaching puberty the comparative value of young males as workers would markedly increase, and the master’s feelings of affection toward a little boy might be replaced by sexual and physical fears. This situation would help explain the continuing high rate of adult female manumission, in relation to a relative decrease in the percentage of adult male freedom.

Table VI presents the origin, age, and sex of 950 libertos for whom all these characteristics could be established. Here the general tendencies and their relationships become clear. The overall 2:1 female to male ratio can be seen in the total of all age groups combined, but when examined by age cohort and place of origin, important variations can be noted. Among the libertos of African origin, males constituted only 25 percent of the sample while among crioulos that percentage rose to 31.7 and among pardos to 38.5. In other words, while the sex ratio was 3:1 females to males among Africans, it was closer to 3:2 females to males among pardos, with crioulos in an intermediate position. In a sense, the less “African” the origin/color group of slaves, the higher the proportion of male libertos within it.

This difference can be explained in part by variations in the age structure of the three origin/color groups. I have already noted that males tended to be freed as children. As can be seen in Table VI, the sex differential among children (0-5 and 6-13) is considerably reduced so that females comprised only 56.5 percent of the libertos, although among crioulos females were still 63 percent of the two adolescent cohorts. In the adult population (14-45, 45 and over, and the unknowns) the three origin/color categories present virtually no differences in the sex ratio. Females constituted 71 percent of both the crioulo and pardo groups and only a slightly higher proportion, 74 percent, of the Africans. The very large numbers of adult females accounts in large part for the disproportionate female to male ratio among the total liberto population.

The importance of color is seen in the statistics of the pardo libertos. Among adults, pardos were 35 percent of the total, a figure perhaps double their statistical presence in the total slave population. Among children, however, pardos constituted over 78 percent of the libertos. It appears that pardos had a decided advantage in the manumission process as children and that after puberty, while they continued to gain their freedom in numbers disproportionate to their numerical importance, their comparative advantage, especially that of pardo men, sharply decreased. These observations are reinforced when we examine the ratio of children to adults among the origin/color categories. Pardo children comprised 60 percent of all pardos freed, while black children accounted for only six percent of black libertos, and with only two exceptions all of these were crioulos. This situation obviously reflected the small percentage of children carried in the Atlantic slave trade and also the importance of “feelings” of paternity and the color of the child. In comparative terms whereas virtually all Africans were manumitted as adults, 60 percent of the pardos were freed as children. Crioulos again held an intermediate position, in that 68 percent were adults and 32 percent were children. The lower percentage of crioulo children in comparison with pardo children underlines the importance of color in this age group.

A final observation about the age of libertos is in order. Traditional critics of manumission as it operated in Brazil have often argued that many of the slaves who obtained their freedom were old or infirm. Cruel masters were thus releasing their nonproductive bondsmen in order to avoid the responsibility of feeding or clothing them.25 While this did occur, it is impossible to arrive at an accurate calculation of the total number of old and infirm libertos. In only 22 instances did the cartas refer to the advanced age of the slave, and in only six was sickness or infirmity noted. Even if all 397 slaves for whom no age could be determined were added to the cohort of 46 and over, the total would still be only 36 percent of the 1160 cases. But, there is a complicating factor here. As will be shown, a large number of slaves were granted their freedom conditionally, with the obligation to remain in service until the master’s death or to continue in the employ of the master’s children. Thus, while many slaves might become legally free as children or in middle age, they might gain their freedom only in old age, and, in fact, might even die before ever having exercised their full liberty. Since it is impossible to determine when the service of libertos freed under such conditions actually terminated, we cannot positively reject the old-age thesis of manumission. Still, it is fair to point out that manumissions that provided for continued service made up only 15 percent of the total sample. In comparative terms the large number of children were probably a more important element of the manumission process in colonial Bahia than were the old and infirm. (Figure 1 summarizes the personal characteristics of the libertos.)

The Process of Manumission

The cartas de alforria cast light not only on the characteristics of the libertos but on the process of manumission and the motivations and attitudes of both masters and slaves within that process. While the complex interplay of cultural and economic considerations cannot be fully understood solely by quantitative analysis, the numerical distributions within the sample and statements made within the cartas do provide some insight into the most controversial aspects of manumission.

Why were slaves manumitted? In 47 percent of the cases direct references were made to the good or faithful service provided by the slave or the slave’s parent (usually the mother). Such words as “faithful,” “obedient,” and ‘loyal” were commonly employed in the cartas to describe the desired attributes of servitude. It is clear, however, that “good services” were not an important motive for manumission, but were instead a kind of necessary “precondition” or minimum requirement. This was especially true when the master received no payment for his release of the slave.

Despite the vaunted role of the Church in the manumission process, expressions of religious motivation, such as “for the love of God,” appeared in the cartas of only 5.7 percent (66/1160) of the slaves freed. Statements within the letters, however—and the pride with which masters granted manumission even to the old and infirm—indicate that slaveowners saw the act of manumission as a charitable gesture no matter what its conditions or terms. The absence of references to religious motivations does not mean that slaveowners derived no moral or religious gratification from their act. It does indicate, however, that other motives were more immediate in their decisions.

The bonds of affection, love, or fictive and consanguineal kinship played a vital role in the manumission process. The importance of such ties were obvious in a slaveowners freeing of his mistress (5) or his illegitimate children (18).26 Godchildren of the slaveowner or of his or her relatives were also occasionally freed (6), but there is no evidence that large numbers of children were given their liberty at the baptismal font. Slaves who performed special services, such as caring for the master during a protracted illness, were rewarded, as were slaves who had raised the master or his children (5). In numerical terms, cartas that spoke of the slave’s love for the master or the master’s love for the slave (48) were important explanations of manumission, accounting for four percent of the total sample.

As has been noted, admissions of biological paternity and the recognition of illegitimate children did play some part in the manumission process. As one man put it when in 1741 he freed his child, “he is my son not my slave.”27 At least some slaveowners recognized the contradiction that existed between those two statuses, and this statement does lay bare the feelings of paternity that did exist. On occasion slaveowners might go a step further. In 1723 Manoel da Silva Braga admitted that before marrying he had fathered two children by a slave woman and that he had baptised them and treated them as his natural offspring. Later, he and his wife not only freed the two children, but recognized them as Silva Braga’s legal heirs as well.28 While a great deal of attention has been given such sincere expressions of affection across the boundaries of color and legal status, these instances comprise only a tiny faction of the total manumissions. Slaves were recognized as heirs in four cases, and in only four others did they receive gifts along with their freedom.

By far the most common expression of affection involved what may be termed surrogate paternity or maternity. In 21 percent of the cases (246/1160) a statement such as the following was made:

I, Maria Lopes, state that among the property that I own is a little black boy named Domingos of five months of age who was born in my arms, son of my Mina slave Antonia; the said little black being the first infant born to me and because of the love I have for him I free and hold for all time as if he was born free from his mother’s womb. . . .29

The birth and upbringing of a slave child within the household, or as the documents often put it, “in my bed,” or “in my lap,” seem to be of great importance in stimulating the slaveowner’s sense of responsibility and obligation toward the slave.30

Were statements of surrogate paternity simply euphemisms used to disguise biological paternity? The evidence here is contradictory. The fathering of children by one’s slaves was not severely condemned by law or society, and there was little need to hide the fact, although a master might wish to disguise his paternity in order to avoid embarassment, or to protect his mistress from the revenge of a jealous wife. In the sample under discussion here, when the color of the mother of a liberto who had been freed with expressions of surrogate paternity could be determined, it was compared to the liberto’s own color. Of 119 pardos manumitted in these circumstances, only 27 had pardo mothers. In other words, 77 percent (92) of this group were the pardo children of black mothers. This, of course, is no sure measure of the owner’s paternity since the biological fathers may have been mulatto slaves or freedmen, or whites other than the master. This situation does indicate, however, that color was an important feature in generating the master’s feelings of paternity. Table VII demonstrates the extent to which pardos were favored by this motive of manumission.

It is only fair to point out that while 57.7 percent of these slaveowners were men and only 40.8 percent were women or married couples,31 the expressions of surrogate paternity were used almost twice as frequently by women and couples as by male slaveowners. Thus the evidence is somewhat contradictory. While miscegenation and the somatic composition of the slave surely affected the feelings of paternity expressed by owners, it is not at all clear that a large number of slaves freed with expressions of such feelings were in reality the biological children of their owners.

It has long been recognized that Brazilian slaves sometimes purchased their freedom. Some authors have seen this as proof of the openness and more humane nature of Brazilian slavery, while others have explained the phenomenon as either a response to the need for certain kinds of free laborers or as a means of social control in which slaveowners by offering the hope of eventual freedom extracted the optimum quality of labor and were eventually fully reimbursed as well.32 Between 1684 and 1745, 47.7 percent (553/1160) of the manumissions sampled were obtained by payment to the slaveowner or his legal representatives. The ratio of paid to gratis (unpaid) manumissions was the same for both men and women. In 356 of these cases no reason beyond the payment and a perfunctory reference to the slave’s “good services” was given. Thus, not only was the purchase of freedom a “common” form of manumission, but almost one out of every two libertos achieved freedom in this manner.

The ratio of paid to gratis manumissions did not remain constant during the period studied. A steady increase in the number of purchased manumissions can be noted from the 1680s to the decade of the 1720s. Then after two decades of relative stability, the percentage of purchased manumissions increased to over half of the total in the 1740s. This trend should be seen in view of the changing index of slave prices presented in Figure 3. There we note a constant rise in slave prices in the period from 1690 to the 1720s and then a leveling of the secular trend. It appears that the increasing value of slaves discussed below adversely affected the granting of gratis manumissions and that as the value of slaves rose, masters increasingly demanded payment for the granting of liberty.

Differences between the color groups in the granting of manumissions can be seen in Table VIII. Here it becomes clear that whereas pardos were only 46 percent of the libertos, they received 61 percent of the gratis manumissions. Compared by sex it can be seen that black men were the least favored group and that pardo males were comparatively advantaged. This situation is partially explained by the large number of pardo male children referred to in Table VI.

The initiatives of the slaves themselves in coping with demands of slave society is impressive. Although “kindly” masters did exist, quantitative analysis demonstrates clearly that slaves who wished to obtain their freedom by payment usually depended on themselves or on their slave relatives for the necessary sum. Godparents, charity subscriptions, the master or his relatives, and unrelated free persons together accounted for only six percent of the paid manumissions. As Table IX demonstrates, in 81 percent of the paid manumissions, the libertos themselves made the necessary payment. The sex of the liberto did not affect this situation, since men and women paid for freedom in exact proportion to their numbers.

Slaves who entered into negotiations for self-purchase often found themselves at the mercy of unscrupulous slaveowners. In disputes which resulted from these arrangements, the slave was always at a disadvantage before the law. A case from 1784 illustrates the disability of slaves in such situations. A crioula woman had obtained from her master an agreement that she would be freed upon payment of 70 milréis, but after seven years of payments the master refused to comply. He claimed that the slave woman had stolen the money from him, and even when her godmother offered to pay the sum, it was declined. Justice officials refused to intervene.33

When slaves had to depend on others to provide the money for freedom, it was usually to their parents—and especially to their mothers —to whom they turned. In 27 instances slave women were able to buy the freedom of their children, while slave fathers could do so in only two cases, as is shown in Table IX. These figures may point to the instability of the two-parent family within Brazilian slavery. Certainly, it is important to note that in 465 cartas mentioning kinship relations of the liberto, 72.9 percent (339) referred to the fact that the liberto was the child of a female slave of the same master. Thus, whatever the institutional condition of the slave family, the bonds between mothers and children were apparently strong and provided crioulo and mulatto slaves a comparative advantage in the manumission process. Of course, parental ties crossed the boundaries of slavery, and freeborn and liberto parents paid for the freedom of their children in six percent of the manumissions.

Payments made by the slave or an intercessor were usually made in common currency (dinheiro de contado), although in the eighteenth century gold dust was sometimes used as the means of exchange. In rare cases a payment in kind was made.34 Because the sums were often considerable, payment in installments was common, and cartas de alforria granted in this manner often established schedules of payment.

Two forms of payment are particularly interesting because they reveal aspects of the Brazilian slave regime that have never been discussed in any detail. Certainly, the most curious form of payment was the provision of a replacement or substitute slave. The potential liberto, or an intercessor, acquired a slave of equal value, who was then presented to the master as a substitute. Roughly, three percent (18) of the paid manumissions were made in this way. In only two cases, however, did libertos provide a replacement from the same place of origin or ethnic group. The process of substitution provokes a variety of questions. Did the law allow slaves to own slaves? There is no clear answer here in legal terms, but certainly custom recognized the practice. In a will made in 1751 the following statement appears:

I declare there is also among our property a moleque named Salvador of the gentio de guiné who is the captive of our slave Simão who owes to our estate the shipping costs and import taxes for the said moleque which I paid. . . .35

This clause indicates that the master recognized his slave’s ownership of another slave, but at the same time the need to mention this in the will may indicate the slave’s inability to legalize this ownership.36

A second type of payment, which deserves special attention, was the purchase of freedom in response to the provisions of a will. It has long been recognized that the manumission of slaves as part of a will and testament was common practice in Brazil, although no one has ever been able to show how common it actually was. Slaves freed in a master’s will were selected for special consideration. Masters did not simply release all their bondsmen but, instead, chose a few particularly “deserving” of reward. In the cartas de alforria examined in this study, 12.7 percent of the total (147) were granted as a result of clauses in wills. This proportion is probably an underestimate, however, since other cartas, while making no reference to a will, do state that the slave had been recently acquired from a deceased spouse or other relative. It should be stressed that manumission in a will did not necessarily mean manumission without payment.

In Cuba and other parts of Spanish America a series of legal mechanisms (coartación) existed, which allowed slaves to demand a just price be set on them and then to work toward their eventual selfpurchase and freedom.37 In Brazil a similar, if somewhat modified or at least less-institutionalized, form of this custom was also practiced. A number of the cartas de alforria mention slaves who are coartado or costado. Most commonly, this refers to slaves who have been designated in their master’s will in the following manner:

I declare that the black woman, Joana, my slave, and her two children, Maria and Ignacio, I leave cortadas in the amount of 100 mitréis which they will pay within five years and should these years pass and the said quantity not be paid, my executors will sell them for the best price possible.38

This arrangement placed an obligation (onus) on the slave to pay a certain amount in a defined period of time or to suffer continued enslavement. It obviously created a category of people who lived in an intermediate status between slave and free, not in the allegorical sense of the freedman but in a strict legal sense. It is interesting to note in the Villa Rica census of 1804 a number of individuals were listed as coartados, indicating a legal recognition of their intermediate status.39 Among the cartas de alforria for the period 1684-1745 there were 21 instances when slaves were noted as coartados. Invariably, these documents were granted by the executors of estates in compliance with clauses of the deceased’s will.40

From the value paid by the slave, or by a benefactor, it is possible to calculate the fluctuations in the price of manumission during the period under study. Figure 3 presents such a curve. The lack of independent price series of slave values for this period complicates an interpretation of this data, but it should be noted that the apparent fall of prices in the early 1690s coincides with the crisis of the sugar economy, while the steady rising secular trend from 1700 to 1725 reflects the price response to the increased demand for slaves in the mining zones to the south.

It is difficult to determine from this data the relationship between the price of manumission, the original purchase price, and the current market value. In a number of instances the cartas simply refer to a “just price.” In others, which involve evaluation in a will and testament, it is clear that the price set for freedom is a market price. Because almost none of the letters refer to the length of a slave’s service, it is virtually impossible to calculate the difference between the original value of the slave and the value at date of manumission.

Analysis of the phenomenon of paid manumissions provokes more questions than it resolves. The fact that almost half the manumitted slaves paid for their freedom or had someone else pay for it surely undercuts the traditional humanitarian interpretation of Brazilian manumission. But, the ability of slaves to accumulate capital raises interesting questions about the nature of the slave regime, the participation of slaves within a market economy, and the differential opportunities for freedom within the slavery. Traditional wisdom has it that manumission, and especially paid manumission, was primarily an urban phenomenon. While this is a commonsense assumption, it should be recognized that at present this is still only an hypothesis. In the cartas analyzed here, paid manumissions occurred proportionately as frequently in areas outside Salvador as within the city. Until “rural” cartórios are thoroughly examined, and until some statement can be made about the total volume of manumissions, the “urban thesis” of manumission cannot be proven.

The interpretation of paid manumission as an economic device, by which both maximum labor can be extracted from the slave and, eventually, the original or market price as well, is difficult to support in the context of the comparative history of slavery. If we accept the basic assumptions of this position, we should expect to encounter paid manumission in North America, in a regime of “unbridled capitalism.”41 Yet, despite some evidence of paid manumission in the antebellum South, the phenomena appears far more characteristic of Brazil.42 One might argue that slaveowners who felt that slaves were inherently shiftless and lazy would have little cause to believe that incentives would increase profitability. But, if that is true, then Brazilian slaveowners, moved by the desire for economic maximization, must have held different ideas about the nature of the slaves. Thus, this economic argument hinges once again on cultural differences within the slave regimes, and that, of course, is an unfinished debate.

Perhaps part of the interpretative problem lies in excessive concentration on the decisions and attitudes of the masters rather than those of the slaves. The Bahian manumissions indicate that slaves were able to maintain ties of family and kinship in the face of adversity, to enter into contractual arrangements, and to cope with the legal process. The willingness of slaves to accumulate capital for themselves and their relatives also constitutes their own comment on the nature of slavery and freedom. Whatever the disadvantages suffered by freed people of color in Brazilian society, their status was perceived by slaves as far better than the burden of slavery, and to achieve freedom men and women were willing to make great sacrifices for themselves and their loved ones.

In the context of the comparative history of slavery in the Americas, the considerable initiative of Brazilian slaves in obtaining their freedom may provide some clue to the apparently higher rate of manumission in Brazil than in the United States. The key may He not in the nature of slavery in the various regimes, but in the contours of the societies built upon slavery and the reaction of slaves to it. Given the obvious disadvantages and restrictions suffered by freedmen in North American slave society, people still in bondage may have made a shrewd calculation that the sacrifices required for the accumulation of the self-purchase price were simply not worth the end result, especially since the process rewarded the master. This is not to say that slaves in North America did not want liberty, but, rather, that they did not see purchase as a viable vehicle toward that end. Slave attitudes would undoubtedly influence slaveowner perceptions of the utility of paid manumissions, and the owners might then react by offering fewer opportunities for freedom, thus closing the circle. Brazilian slaves with a broader range of opportunities as freedmen were more willing to pursue this goal by purchase. Obviously, this argument is hypothetical, but it does at least shift the discussion of manumission away from an exclusive consideration of the cultural differences of the slaveowners to the inclusion of the perceptions and initiatives of the slaves, in response to regimes created by demographic, economic, and cultural factors.

Award of a carta de alforria, either gratis or paid, did not necessarily free the liberto from all obligations to his or her former owner. Manumissions could be granted conditionally, and throughout the period studied such grants were made. In comparative terms, however, conditional manumissions accounted for less than 20 percent of the sample. In fact, the most common phrase to appear in the cartas is “free for all time (foro para todo sempre)” or “free of all bondage (foro de toda escravidão).” This term was included in the letters of 65.9 percent (765) of the libertos, indicating the owner’s desire to release the former slave of all obligations. If we add to these cartas those others in which no statements of conditions or future commitments appear, the combined total equals 81 percent of all manumissions. Thus, while conditional manumissions were granted, they were greatly outweighed by those without limitation.

Conditional manumissions, like coartação, created a status of legal freedom but continued bondage. These conditional or limited freedoms fell into equal proportion on male and female slaves. While slaveowners displayed considerable imagination in the kind of conditions they imposed, the conditional cartas can be grouped into two categories: continued service and continued obligations. In 220 cases in which masters imposed conditions on freedom, 70 percent (154) stipulated that the slave continue to serve until the master’s death. Other variations of this pattern required the slave to serve the owner’s spouse, child/children, or relative. In ten cases a specific time limit of future service was established, while in four others service was to continue until the slave married.

Another type of conditional freedom imposed future obligations on the liberto. Slaves might be required to pay the debt of their former master or to perform certain religious obligations. In ten cases libertos had to pay for masses for their former owner’s soul. Masters sometimes showed themselves particularly reluctant to forfeit the special services or skills of their bondsmen. In 1728, a woman freed her slave, a seamstress, but required her to continue making dresses for her former mistress free of charge. In another instance, a freedwoman was required to teach her skill to a second slave before enjoying her freedom. Some slaveowners also displayed a reluctance to surrender their authority even after granting freedom. Cases exist of masters insisting to be treated as “patron” by the liberto and, in fact, in rare instances (2) cartas included a specific clause that threatened revocation of freedom in case of disobedience after freedom. Slaveholders were sometimes careful to stipulate that a liberto’s parent or child was to remain in slavery, and in one case the master insisted that any child born to his former female slave within nine months of the date of the carta was to remain his slave. Such conditions lay bare the mentality of the slaveowning class and the nature of Brazilian slavery, but they constituted a small proportion of the Bahian manumissions.

Despite the fact that the cartas de alforria commonly refer to the freedom as total, or “as if the slave was born of a free womb,” libertos apparently retained their freedom only so long as they complied with certain norms of subservience expected by the white community. In other words, all cartas were ultimately conditional in that a liberto was always subject to reenslavement. In a few cases clauses providing for revocation were included in the letter itself. In 1725, for example, a woman named Agueda de Araújo freed her mulatto slave, Marianna, for an extraordinarily high price, 400 milréis, with the condition that any future disrespect or inattention would lead to reenslavement.43 The central problem, however, was how the liberto who was not freed with such defined restrictions stood before the law. The Ordenações filipinas specifically provided that manumissions secured by fraud, or those which had deprived inheritors of their rightful property, could be revoked. More important, the law also stipulated that a liberto’s disrespect to his former master was cause for reenslavement.44 The frequency with which this law was enforced cannot be determined, but the very threat of enforcement may have been enough to produce the desired result of social control.

The disabilities of libertos and attitudes toward them are topics perhaps better suited to a discussion of Brazilian society in general, rather than an analysis of manumission, but it should be recognized that at various times attempts were made in colonial Brazil to limit manumission.45 Arguing that freeing slaves would deplete the labor force, or that freeing old people was inhumane, critics of manumission usually displayed an underlying fear that manumission would disrupt slavery and would create a volatile and hostile free colored population as well. The municipal council of Mariana in Minas Gerais argued along these lines in 1735 and explicitly claimed that paid manumissions were the result of prostitution and other crimes. The city fathers wanted paid manumissions eliminated completely.46 When, in the 1690s, the Crown had tried to intercede in the case of a badly mistreated slave woman and to order her release, Governor João de Lencastre refused to comply, claiming that undue opportunities for freedom would stimulate crime and cause general unrest.47 Such policies, however, were never general, and, for whatever reasons, manumissions continued during the long course of Brazilian slavery.


The analysis of the cartas de alforria presented here reveals the characteristics of the libertos and some of the patterns within the process of manumission as it operated in colonial Bahia. The data has enabled us to test and revise some of the traditional and at times contradictory assumptions, which for so long have dominated the discussion of Brazilian manumission. Certainly, if half the libertos obtained freedom by purchase and almost 20 percent received conditional freedom, then the purely humanitarian impulse must be seriously discounted; but similarly, if large numbers of children were manumitted, then the purely economic interpretation must also be modified. And what do we make of the ability of slaves to amass capital and to operate successfully in a market economy? In fact, the data presented here open a whole new series of questions in need of examination. What, for example, is the demographic significance for the growth of the slave population of the 2:1 female-to-male ratio of Bahian manumission? If, as we believe, the majority of females freed were of reproductive age, then this situation must have exacerbated the imbalance of the sex ratio already introduced by the slave trade. Moreover, it seems to indicate that the dependability of the slave trade and the high rate of infant mortality made a woman’s reproductive capacity of little economic value in the Bahian context. Does the steady manumission of women in childbearing age provide a partial explanation of the negative rate of increase among Brazilian slaves?48 To answer such questions we will need better statistics, not only of manumission in various regions of Brazil, but also of the Brazilian population as a whole. But better statistics will not resolve all the questions of Brazilian slavery.

It should be clear that the numbers presented in this study will not explain all of the aspects of manumission as it operated in Brazil. The motives of slaves and masters in this process cannot be fully understood by quantitative analysis alone. The cartas de alforria are filled with contradictions. The same slaveholder might treat one slave miserably and set another free. Masters moved by a pious impulse took pride in the manumission of a sick old woman, “in recognition of her long service,” and saw no inhumanity in their act. The difficulty of explaining such actions are embodied in a carta granted by Dona Branca de Meneses in 1700. In it she freed her slave “because of the good services of his mother, because he was born in the casa-grande, and I have raised him, and because of the 100 milréis he gave me to do it.”49

Traditional explanations of manumission have, to a large extent, been expressed as polar opposites of culturally determined humanitarianism or the often inhumane responses to purely economic considerations. There is no need to divide these explanations into discrete categories. The economic imperatives of slavery always operated in a cultural context. The impulse of slaveowners to maximize their profits was tempered by a variety of cultural, legal, and situational phenomena whose existence created the seeming inconsistencies of a slave regime in which profit-satisfaction, rather than profit-maximization, was the operating mode of behavior.

The contradictions and inconsistencies revealed in the cartas de alforria were not exclusively limited to the white slaveholding class. The fact that at least 2 percent of the masters who granted freedom were people of color, the evidence of slaves who owned slaves, the use of replacements as a means of buying freedom are among the most curious and disturbing aspects of Brazilian slavery. Such evidence must be weighed, of course, against the many instances of free blacks and mulattoes who purchased the freedom of others or aided runaways, or joined in slave rebellions, but we must be prepared to recognize the pervasiveness and perniciousness of the slave regime and its effects on everyone touched by it. Slavery was a system, not simply a set of economic relationships, and it is not surprising to find that it forced patterns of thought and action on slaves and former slaves, just as it did on those whose origins did not bear that stigma.


The author wishes to express his thanks to João José Reis of Bahia, who assisted in the collection of this data, and to Christine Lohr, of the University of Minnesota, who served as a Research Assistant. Thanks are also given to Professors Dauril Alden, Philip Curtin, Peter Eisenberg, Stanley Engerman, Richard Graham, Herbert Klein, Franklin Knight, John Modell, Peter Smith, and Emilia Viotti da Costa, who provided expert advice and criticism. Computer time was made available by a grant from the Graduate School of the University of Minnesota.


Eugene Genovese, “The Treatment of Slaves in Different Countries: Problems in the Application of the Comparative Method,” in Slavery in the New World, ed. Laura Foner and Eugene Genovese (Englewood Cliffs, N.J., 1969), pp. 202-210.


The status of the debate is ably summarized in Carl Degler, Neither Black Nor White. Slavery and Race Relations in Brazil and the United States (New York, 1971), pp 39-47.


See, for example, the essays in Neither Slave Nor Free, ed. David W. Cohen and Jack P. Greene (Baltimore, 1972), especially A. J. R. Russell-Wood, “Colonial Brazil,” pp. 84-134, and Frederick Bowser, “Colonial Spanish America,” pp. 19-58, which complements his treatment of manumission in The African Slave in Colonial Peru, 1524-1650 (Stanford, 1974). Bowser has also written a comparative study, “The Free Persons of Color in Lima and Mexico City: Manumission and Opportunity, 1580-1650,” in Race and Slavery in the Western Hemisphere: Quantitative Studies, ed. Stanley L. Engerman and Eugene D. Genovese (Princeton, 1974); on the French Caribbean, see Yvan Debbasch, Couleur et Liberté, 1 vol. to date (Paris, 1967). On Barbados, there is Jerome Handler, The Unappropriated People: Freedmen in the Slave Society of Barbados (Baltimore, 1973). Also important for this study is Herbert Klein, “The Colonial Freedman in Brazilian Slave Society,” Journal of Social History, 3:1 (Fall, 1969), 3-52. Richard Graham, “Brazilian Slavery Reexamined: A Review Article,” Journal of Social History, 3:4 (Summer 1g70), 431-453, points out that important recent studies of slavery in nineteenth-century Brazil, produced by the “São Paulo School” (Florestan Fernandes, Octavio Ianni, Fernando Henrique Cardoso, Emília Viotti da Costa), are weak on the problem of manumission. An interesting note is provided in Luiz Mott, “Cautelas de Alforria de duas escravas na provincia do Pará (1829-1846),” Revista de História, 47:95 (1973), 263-268.


An example of such a charter is as follows: “I, Margarida Rodrigues, a black woman, state that among the property I possess and of which I am mistress and owner with no contradiction is a black woman named Maria of the Ardahê people (gentio) and her son Estevão, a Brazilian-born boy (crioulinho) of two years of age; the two said slaves, Maria and her son Estevão, I free and consider to be free from today and for all times because of the good services I have received from the said black, Maria, as well as because of the love with which I raised the aforementioned, her son, Estevão, and because of the 100 milréis which I received from the said black woman. And thus by this letter of manumission I wish and I am pleased that the said black woman and her son enjoy and possess the said liberty because of the above-mentioned reasons and I grant this of my free will and without the pressure (constrangimento) of anyone and they can go wherever they please as freed people, free and exempt of all captivity as if they were born of a free womb. . . .” Secção Judiciária, Livro de Notas (Cidade), July 9, 1709, Arquivo Publico de Estado da Bahia, cited hereinafter as APB, 22B, fol. 133. All future citations to this collection will be to the series, “Cidade,” unless otherwise noted.


Manumitted slaves who moved to a new place of residence might re-register their letter in order to protect themselves.


These were selected from the 1759 cartas registered in 81 notarial volumes during the period 1684-1745. Because of the precarious physical condition of many of the volumes dated prior to 1720, selection in these years were based in large part on legibility. It is, therefore, not a statistically random sample, but I do not believe that the gastronomic preferences of the ants nor the acidity of the ink has significantly biased the sample. For the period 1728-1745, a systematic sample was made by recording every third letter registered. By this method 349 cartas were extracted from the 1033 registered. The most difficult problem is estimating a rate of manumission. For the period 1728-1745 an average of seventy slaves a year were registered as libertos. If, because of the gaps in the notarial series, we assume that this figure represents 50 percent of the actual manumissions, then in a slave population of 15,000 the crude rate of manumission was approximately 1 percent per annum. For the period 1819-1851, Trosko found 16,778 libertos registered in the indices of manumission. This is an average of 524 per year. In this period there were perhaps 40,000 slaves in Salvador. Thus the rate for this period is also 1 percent per annum. There is a possibility, however, that the rate was significantly higher since Trosko believes that manumissions registered in these indices do not include those recorded in the Livros de Notas. Separate registers of manumission were not kept during the period under consideration in this paper. See Barbara Trosko, “The Liberto of Bahia before Abolition,” M.A. Thesis Columbia University 1967, pp. 2-8, Appendix.


The project has been organized in collaboration with Katia M. Queirós Mattoso of the Universidade Catolica do Salvador and Arnold Kessler, University of California, Berkeley. Queirós Mattoso has published two related studies to date: “A propósito de cartas da alforria na Bahia 1779-1850,” Amis de História, 4 (1972), 23-52, and “Os escravos na Bahia no alvorecer do seculo XIX,” Revista de História (São Paulo), in press. Kessler presented his findings in “Bahian Manumission Practices in the Early Nineteenth Century,” a paper presented to the American Historical Association (San Francisco, 1973). An early study of manumission in Bahia which concentrates on the period after 1850 is Barbara Trosko, “The Liberto of Bahia.” See also Mary Karasch, “Manumission in the City of Rio de Janeiro, 1807-1831,” a paper presented to the American Historical Association (San Francisco, 1973).


I have set forth the major features of the Bahian sugar cycle in “Free Labor in a Slave Economy: The Lavradores de Cana of Colonial Bahia,” in Colonial Roots of Modern Brazil, ed. Dauril Alden (Berkeley, 1973), pp. 147-198.


The important role of notaries in Brazil has never been adequately studied. At present Dioclecio Leite de Macedo is completing a study of colonial notaries. See his compilation, Tabelliães do Rio de Janeiro (1563-1963) (Rio de Janeiro, 1965). Cf. Stuart B. Schwartz, Sovereignty and Society in Colonial Brazil (Berkeley, 1973), pp. 25, 148-149.


Thales de Azevedo, Povoamento da Cidade do Salvador, 2nd ed. (Bahia, 1969), pp. 181-201, provides the best summary of Bahian demography.


Padre Gonçalo Soares da França, “Dissertações de história eclesiastica do Brasil (1724),” Sociedade de Geografia de Lisboa MS. Reservados 1-C-147.


Philip Curtin, The Atlantic Slave Trade. A Census (Madison, 1969), pp. 41-47. See also Pierre Verger, Flux et reflux de la traite des nègres entre la golfe de Benin et Bahia de Todos os Santos du xviie au xviiie siècle (Paris, 1968).


Herbert Klein, “The Portuguese Slave Trade from Angola in the Eighteenth Century,” Journal of Economic History, 32:4 (December 1972), 894-918; “The Trade in African Slaves to Rio de Janeiro, 1795-1811: Estimates of Mortality and Patterns of Voyages,” Journal of African History, 10:4 (1969), 533-549. Klein shows that in the Rio trade children made up only 5 percent of the total imports. The Companhia Geral de Pernambuco, which imported slaves to that captaincy in the late eighteenth century, carried slave cargos in which children comprised only 1 percent of the total.


The slave population of Minas Gerais in 1821 totalled 181,892, of which 21,887 (11.9%) were mulattoes. The sex and color distribution was as follows: blacks, 104,115 (65.2%) males and 55,890 (34.8%) females; mulattoes, 12,105 (55.4%) males and 9,772 (44.6%) females. These figures are extracted from Kenneth Maxwell, Conflicts and Conspiracies: Brazil and Portugal, 1750-1808 (Cambridge, 1973), 264. My collaborator, Arnold Kessler, suggests that the percentage of females in the slave population of Salvador may have been much higher than in the general Brazilian slave population. Certainly, there is reason to believe that the demand for domestic service increased the need for female slaves in the city.


Gilberto Freyre, The Masters and the Slaves, 2nd English-language ed. (New York, 1956), originally published in Portuguese in 1933. Frank Tannenbaum, Slave and Citizen (New York, 1947). A major synthesis of their position was presented in Stanley Elkins, Slavery: A Problem in American Institutional and Intellectual Life (Chicago, 1959).


See the summary by Júlio de Freitas Brandão, “O escravo e o direito,” in Trabalho livre e trabalho escravo. Anais do VI Simpósio Nacional dos Professores Universitários de História (São Paulo, 1973), vol. I, pp. 255-284.


Agostinho Marques Perdigão Malheiro, A escravidão no Brasil, 3 vols. in 1 (Rio de Janeiro, 1866).


There is an interesting discussion of the synod by Gentil Avelino Titton, “O sínodo da Bahia (1707) e a escravatura,” in Trabalho Livre e trabalho escravo. Anais do VI Simpósio Nacional dos Professores Universitários de História (São Paulo, 1973). pp. 285-306.


For example, Jorge Benci, Economia cristã dos senhores no governo dos escravos, ed. Serafim Leite, 2nd ed. (Oporto, 1954), originally published in 1705, does not even mention the possibility of manumission.


Mattoso, “A proposito,” p. 41; Kessler, “Bahian Manumission.” It is interesting to note that Bowser’s study of Lima and Mexico City in the seventeenth century and Mary Karasch’s examination of Rio de Janeiro in the nineteenth century reveal the same sex ratio among manumitted slaves.


Census of Engenhos Guibaca, Dinheiro, and Maracangalho, APB, Cartas do Governo, 232.


Mattoso, “Os escravos” attempts to resolve the problem according to work capacity.


The average price of children in the period studied was 47.5 milréis. Because of this I have set the minimum adult price at 50 milréis.


Cf. Marvin Harris, Patterns of Race in the Americas (New York, 1964), p. 86. See also the summary in Graham, “Brazilian Slavery Re-examined,” pp. 449-450.


Numbers in parentheses indicate the number of instances encountered in the sampled letters of manumission.


APB, Livro de Notas 71, f. 178 v. (May 17, 1741).


APB, Livro de Notas 38, f. 226 (June 25, 1723).


APB, Livro de Notas 46, f. 136 (April 18, 1727). An example of a couple’s expression of these feelings reads as follows: “Say we husband and wife Domingos Gonsalves Pereira and Izabel Rodrigues that we raised a Black (crioulo) named Felipe born in our house, son of our slave named Maria, and we raised him in our bed and in our arms as our son and because of the love we have for him and the good services received of his mother it is our pleasure to free him. . . .” APB, September 4, 1701, Livro de Notas 17, f. 180.


Fifty-five libertos, or 22 percent of those freed with expressions of surrogate paternity, also paid for their freedom.


The remaining 1.2 percent were manumitted by religious institutions or parents and children. (The sum is less than 100 percent due to rounding.)


David Denslow, “Economic Considerations in the Treatment of Slaves in Brazil and Cuba,” paper read at the Mathematical Social Science Board Conference on Slavery (Rochester, N.Y., 1972); Marvin Harris, Patterns of Race in the Americas, p. 86. See also the summaries of these positions presented in Degler, Neither Black Nor White, pp. 25-47; and in David B. Davis, The Problem of Slavery in Western Culture (Ithaca, N.Y., 1966), pp. 262-273.


Juizo ordinario de Jaguaripe, 1772, APB, Cartas ao Governo 187; Traslado de auto de preto Francisco Salles, Feb. 16, 1803, APB, Cartas ao Governo 1803.


A case from 1691 is illustrative. A freeman was allowed to buy his mulatto son’s freedom for two mares and two horses. APB, Livro de Notas 7A, p. 93 (April 10, 1691).


Inventory of Eufrazia de Nascimento do Jesus, APB, Secção Judiciária, testamentos Cidade 631.


Even more enlightening about the process of acquiring slaves is testimony presented in an inquest held by the probate judge of São Francisco do Conde in 1836. A crioula slave, Lucianna Maria da Conceição, wished to purchase a slave as a dowry for her granddaughter. She gave money to a friend to be sent to Africa for this purpose, and a Nagô women named Jeronima was purchased and delivered to her. The newly acquired slave was conducted to the Engenho Cahipe where Lucianna resided but was eventually sent to the city to be “hired out” (por ao ganho). Jeronima then sent her earnings to her mistress, who continued to labor as a slave on the plantation. See the documents at Instituto Histórico e Geográfico da Bahia, pasta 28, doc. 11.


See the discussion in Herbert Klein, Slavery in the Americas. A Comparative Study of Cuba and Virginia (Chicago, 1967), pp. 196-200; cf. Franklin Knight, Slave Society in Cuba in the Nineteenth Century (Madison, 1970), pp. 93-95. The classic account is Hubert H. S. Ames, “Coartación: A Spanish Institution for the Advancement of Slaves into Freedom,” Yale Review, 17 (February, 1909), 412-431.


Testament of Jeronimo Nunes Silvany (1760), APB, Judiciária 637. For other examples, see Testament of João Alves de Azevedo (1694), Arquivo da Santa Casa da Misericórdia, Livro do Tombo 42; Processo (1818), Instituto Histórico e Geográfico da Bahia, pasta 27, doc. 9.


Herculano Gomes Mattias, ed., Um recenseamento na capitania de Minas Gerais: Villa Rica (1804) (Rio de Janeiro, 1969).


For example, “as universal executor of the deceased . . . I received from her black, Felicia, 40 milréis for which the deceased her mistress left her coartada in her testament. . . .” APB, Livro de Notas 44, fol. 226. Portuguese law stipulated that the deceased’s estate be divided into three portions, of which the testator could dispose of one as he or she saw fit, while the other two belonged to the spouse and children. Issue of a carta de alforria before actually releasing the slave was a way of insuring the slave’s freedom after the master’s death by avoiding possible involvement in a contested will. See Denslow, “Economic Considerations.”


Stanley Elkins, Slavery, argued that the lack of institutional bulwarks against the capitalist ethic of North America produced the peculiar contours of slavery in the United States. It is possible to speculate on the amount of time a slave needed to accumulate the amount needed for self-purchase. In 1751 the daily wage of a slave “ao ganho” was six vintens, or 120 réis, per day. This amounted to 3$6oo réis per year, based on a work year of 300 days. At this rate it would take a male slave over thirty-seven years to amass the 135$000 réis necessary for self-purchase. Of course, this assumes that slaves could keep all they earned, which is most likely not the case. The daily wage was estimated by the Câmara of Santo Amaro da Purificação, July 3, 1751, Arquivo da Câmara do Salvador, Cartas do Senado 28.5.


See the summary by Eugene Genovese, “The Slave States of North America,” in Neither Slave Nor Free, ed. Cohen and Greene (Baltimore, 1972), pp. 258-277.


APB, Livro de Notas 41, fol. 252. Cartas de revogação that reimpose slavery on libertos are occasionally encountered in the notarial registers.


A case from 1812 is illustrative of how the law was used as a means of social control. A liberto named Vitorino, who was living as a vagrant, had dared to insult the son of his former, and now deceased, master. The young son was not “such a child” that he could reduce the liberto to slavery as the law provides (“como hé Direito”). Vitorino finally ended up as a forced enlistment in the garrison. See Juiz de Fora de Santo Amaro, Nov. 21, 1812, APB, Cartas ao Governo 24.


See the arguments summarized by A. J. R. Russell-Wood, “Colonial Brazil,” in Neither Slave Nor Free, ed. Cohen and Greene (Baltimore, 1972), pp. 95-96, 100.


Minutes, Câmara of Mariana, May 5, 1735, APB, Ord. reg. 55, f. 99-99v.


Documentos Históricos da biblioteca Nacional do Rio de Janeiro, XXXIV (Rio de Janeiro, 1936), p. 217. For an example of the social control of libertos, see the comments of the juiz ordinario of Santo Amaro (March 26, 1809) who protected the social order by having some libertos “moderately whipped” as an example. Such techniques seem to have been most common in the period after 1780. I have tried to explain this phenomena in “Landowner Politics and the Rise of a Peasantry in Late Colonial Brazil,” a paper presented at the Johns Hopkins University Conference on Brazilian Independence (1972).


Klein suggested this in his article, “The Colored Freedmen in Brazilian Slave Society,” pp. 38-39. On the low value of reproductive capacity in the U.S. South, see Robert William Fogel and Stanley L. Engerman, Time on the Cross. The Economics of American Negro Slavery (Boston, 1974), pp. 78-86.


APB, Livro de Notas 16, f. 59 (June 25, 1700).

Author notes


The author is Associate Professor of History at the University of Minnesota.