Francisco, an African-born slave living in the city of Salvador da Bahia, Brazil, obtained his carta de alforria (letter of liberty) on May 14, 1852. His African origin or “nation” (nação) was Nagô. For his own freedom, Francisco paid the equivalent of seven hundred thousand réis (seven hundred milréis), the price his owner had paid to his former owner. But Francisco did not pay cash; rather he purchased his freedom by substituting another male slave called João whom he, Francisco, owned. João was also Nagô and subsequently took Francisco’s place working on a small boat known as an alvarenga, a lighter that ferried goods between ship and shore.1

Francisco’s letter of liberty does not tell how he accumulated enough money to purchase João, or how Francisco had come to own his own slave, albeit temporarily, presumably with the permission of his owner. All it reveals is that Francisco, employed in an urban occupation, possibly for many years, negotiated with his owner to arrive at a mutually agreeable price for João, a price equivalent to what the owner was demanding for Francisco’s freedom; and also that the owner agreed to accept João in lieu of cash payment. Equally significant is that Francisco apparently persuaded his owner to accept as a substitute a slave not only of the same gender but also the same nation.

Did Salvador’s urban setting give Francisco an advantage in obtaining the cash for self-purchase? Why did Francisco choose to “trade in” another slave in his place instead of paying cash?2 Did sharing the same nation mean belonging to the same ethnic group or sharing an ethnic identity? Did Francisco’s gender determine his way of purchasing his freedom? How common was such self-purchase of freedom, and how did such transactions change over time? Francisco’s letter of liberty poses some important questions about manumission, ethnicity, and additionally, gender in urban slavery.

A letter of liberty transferred title to property (the slave) from the owner to the slave himself. This transfer had to be formally registered and publicly notarized. Once this formality had been completed, the exslave, now called forro (freed) or liberto (liberated), was legally free. It was the ex-slave’s awesome responsibility to preserve this original letter; having the notary preserve a copy could legally protect the freed slave from re-enslavement in case the original was lost.

Letters of liberty preserved in the livros de notas in Salvador’s state archives provide useful information about the identification of each ex-slave; name; gender; birthplace or nation (for African-born) or color (for Brazilian-born); occasionally age, and more rarely occupation; and some information on the ex-owner.3 It was also common for the owner to state why and under what conditions or limitations the slave was manumitted.

This study is based on 3,516 letters of liberty registered in Salvador between 1808 and 1884. These represent the surviving letters of liberty registered in the city of Salvador for two years in each decade for the period 1810–1880.4 This information has been supplemented by 325 wills of freedpersons (157 men and 168 women) who resided in Salvador in the nineteenth century, and 2,608 legal registers of slaves purchased and sold in Sé parish of Salvador for 1838–1888.5

The aim of this essay is to discuss the manumission of slaves in a major Brazilian port city over 80 years, from the late colonial period (1808) to the abolition of slavery (1888). The study is meant to follow and supplement the studies of manumission in Bahia by Stuart B. Schwartz (for the period 1684–1745) and Kátia M. de Queirós Mattoso (for 1779–1850), and generally to call special attention to ethnicity in the New World slave community.6 First, it will examine the setting for this study: Salvador and the transatlantic slave trade, patterns of urban slaveholding, and African nations and ethnicity in the slave community. Then it will treat the practice of manumission: its patterns, its forms, and the custom of self-purchase. This section will return to the story of the Nagô slave Francisco as an example of a “trade-in,” a special form of self-purchase.

Salvador and the Transatlantic Slave Trade

The city of Salvador was founded in 1549. It was the seat of the royal captaincy of Bahia and capital of the Portuguese colony of Brazil until 1763. Sé parish, founded in 1552., was the second-oldest parish in Bahia and the most important, as the administrative and ecclesiastical center not only for the city of Salvador but also for the captaincy. Sé was also the largest parish in Salvador. In 1801, 21.8 percent of the population resided there.7

Salvador developed as the major export city for sugar and later for tobacco, with the Bahian Recôncavo as its agricultural hinterland. To meet the inexhaustible demands for labor on sugar plantations in the Recôncavo, African slaves were imported en masse to Bahia beginning in the 1570s. The major source of Bahia’s slaves in the following century was Angola, but by the end of that century, Rio de Janeiro had also emerged as a major importer of slaves from Angola for the booming mining regions of Minas Gerais. Bahia then turned to the Gold Coast. By the time of the sugar resurgence in the 1790s, the major suppliers of slaves to Bahia were the Bight of Benin and the Bight of Biafra. At the beginning of the nineteenth century, West Africans were the main component of the slave population in the city of Salvador, although the importing of limited numbers of slaves from other parts of Africa continued.8

The population of Salvador in 1807 was 51,112, of which only 28 percent was defined as white. Throughout the ensuing 80 years until the abolition of slavery (1888), some 70 percent of the population continued to be classified as nonwhite (table 1). As long as the massive influx of “human cargo” from Africa continued, the majority of Salvador’s slave population remained African-born. According to João José Reis’s estimate, 42 percent of the entire population in 1835 was enslaved and, of these, 64 percent was African-born.9 In 1872, when the first national census was taken, thewhole population of Salvador amounted to 108,138. Whereas the portion of slaves had decreased to 11.6 percent, the free population of color (whether freeborn or freed) accounted for 57.3 percent.10

The slave trade was officially banned in Brazil with the decree of November 7, 1831, but it continued illegally and on a massive scale.11 Slaves were no longer taken directly into the port of Salvador; slave ships unloaded their illegal cargoes clandestinely on the islands of the Bahia de Todos os Santos (the Bay of All Saints) or at the mouth of the Rio d’Una. The west side of the main island of Itaparica was the main depot, from which Africans were shipped further in the coastal trade or taken directly to the slave market of Salvador.12

The number of enslaved Africans imported to Bahia was 142,300 for the period 1811-1830. Over the next two decades (1831–1850), imports still totaled 98,600. Examination of the distribution of these slave imports for each year and by decades, however, demonstrates that the law of 1831 did effectively decrease the number of Africans imported to Bahia. The number of newly arrived Africans suddenly dropped from 7,000 to 1,000 in 1831 and continued lower than previously except for the years 1846–1850, when 45,000 slaves were imported.13

Since the plantations demanded a constant number of slaves for the agricultural labor force, there are reasons to believe that the city of Salvador accepted a lower percentage of new arrivals after 1831. Few documents trace the exact change in numbers, but baptismal registers of Santo Antônio parish between 1809 and 1869 show that the number of baptized African-born slaves drastically declined in the 1830s.14

Once the transatlantic slave trade ended, persistent demand brought the transfer of urban slaves, particularly men, to the rural regions, especially the booming coffee plantations of the southeast.15 This caused Salvador slave prices to rise rapidly at midcentury, though the prices of female slaves did not climb as much as those of males (figures 1 and 2).

Slaveholding Patterns in the City

In all the transatlantic slave trade to the New World, males constantly outnumbered females by a ratio of at least 2:1. This was a result of the vast demand for male laborers on the plantations as well as the prevailing local demand for slave women in sub-Saharan Africa.16 According to David Eltis’ calculation, 64.1 percent of the slaves imported to Bahia in the period 1811–1850 were male.17 Despite this gender imbalance among imported slaves, the slave population in nineteenth-century Salvador was nearly balanced by gender for both African-born and Brazilian-born, with females slightly outnumbering males. For the period 1838–1888, for slaves purchased and sold in Sé parish, male-female ratios were 48:52 for African-born and 47:53 for Brazilian-born.18 Male-female ratios in the slave population were 44:56 in the surviving regional census records of 1855 and 49:51 in the first national census, taken in 1872.19 In contrast, rural plantation areas, such as the Bahian Recôncavo, recorded a much higher proportion of male slaves as field hands.20 The gender balance in the city of Salvador is mainly attributable to the city’s high demand for female domestic servants. Among all the slaves registered in their owners’ inventories for the period 1811–1888, 19.5 percent of males and 74.1 percent of females were domestic servants.21

Slaveholding in the city was practiced on a much smaller scale than it was on the plantations, although every middle-class urban family owned at least one live-in female slave, who performed all the housekeeping. In 370 inventories of slaveowners registered in Salvador for the period 1808–1888, 86.2 percent of the owners (319 individuals) owned no more than ten slaves. When broken down further, the figures are even more revealing: 16.4 percent owned only one slave, 13.5 percent owned two slaves, and 17.3 percent owned three slaves.22 According to João Reis, who used the same source for the period 1811–1850, among 395 individuals who registered inventories, 67.1 percent (256 individuals) owned only one to ten slaves, while 13.2 percent (342 individuals) owned no slaves at all.23

Urban slavery was also characterized by the hiring-out system.24 Wageearning slaves called escravos de ganho were hired out on either a full-time or part-time basis. They were obliged to return to their owners a mutually agreed portion of their daily or weekly wages. Some even lived outside their owners’ houses. The most visible examples of escravos de ganho were peddlers of both sexes, male porters working in gangs, male artisans, and female market-stall keepers called quitandeiras.25 The city also permitted domestic slaves to hire themselves out as peddlers or prostitutes at night and on Sundays and holidays.26 This wage-earning system in urban slavery enabled the enterprising slave to accumulate money eventually to purchase freedom.27

African Nations and Ethnicity in the Urban Slave Community

Table 2 shows the distribution of slaves who were legally registered for purchase and sale in Sé parish. Among those of African birth whose place of origin was identified, the geographical distribution was 81.2 percent from West Africa, 16.8 percent from Central West Africa, and 2.0 percent from East Africa for 1838–1848; and 93.2 percent from West Africa, 5.8 percent from Central West Africa, and 1.0 percent from East Africa for 1852–1888, Therefore the majority were West Africans, among whom the nation of Nagô was the most prominent.

What was meant by an African nation? Slaves of African birth in Brazil were often identified by their “nations,” a legacy inherited from the European custom of identifying slaves in Africa by “nationalities” regardless of their specific birthplace or ethnic affiliation. “Nationalities” or “nations” were divided into two categories. One was the name of the port from which the slaves were shipped; for example, “Mina” was originally applied to a slave shipped from the Portuguese fort of El Mina on the Gold Coast (modern Ghana). This usage was extended, however, to mean anyone from West Africa in nineteenth-century Brazil. The other category was the ethnic or linguistic term linked to a much larger group. For example, “Nagô,” which had originally referred to a subgroup of the northern Yoruba, was broadly applied to all Yoruba-speaking peoples. Furthermore, some nations were loosely linked to geographical regions, such as “Congo” and “Angola.”28

The nations or nationalities were originally group labels imposed by others on enslaved Africans. Through their common Middle Passage experience, with its harsh physical conditions and high mortality, and later during their forced incorporation into New World slave systems, the African-born themselves gradually adopted these labels to construct or reconstruct self-perception, identity, and ethnic affiliation.29 In this sense, the association of those who were transported on the same ship and called one another malungos (shipmates) was a part of this process of reconstructing ethnic groups.30

On many occasions, ethnicity was collectively represented in associations of slaves largely under the labels of African nations.31 In the first three decades of the nineteenth century, Salvador experienced a series of slave uprisings, ending with the so-called Male Revolt in 1835. At first, most of the participants in these uprisings were Hausa slaves, but later Nagôs predominated. Included were small numbers of females and a very few Brazilian-born persons of color.32 In this same period, some runaway slaves from Salvador formed small-scale quilombos (maroon communities) on the outskirts of the city. The quilombo often comprised persons of the same nation.33

In their daily lives, the African-born slaves of nineteenth-century Salvador communicated with one another in their native languages.34 They also commonly referred to one another by their African names, although slaves were given Portuguese names at their Catholic baptism before or on arrival in Brazil, and were known by them to the non-African-born population.35 African names had specific meanings in their native languages, and their continued use in the slave community could have been essential for African-born persons to maintain their ethnic identity.

In the city of Salvador, each ethnic group occupied its own territory. On Sundays and holidays, the African-born population gathered at specific places for dances called batuques,36 Both male and female workers of a particular nation gathered daily at their own street corners (cantos) throughout the city.37 Furthermore, male work gangs, such as chair carriers, porters, and stevedores, were usually African-born and often composed of men of the same nation. Ethnic groups were prominent in the membership and leadership of black lay brotherhoods in the colonial period, even though such associations were usually not exclusive in terms of legal status or “color. ”38 The African-born slaves also developed economic mutual-aid associations called juntas (unions) for making loans.39

Communication within the slave community developed through these ethnic groups and gatherings. Many African-born kinspersons who had been enslaved in Africa and transported to Salvador separately managed to reunite and reconstruct their relationships. Those who had not been enslaved and shipped together, such as mothers and infants, searched for enslaved parents, children, and other kinspersons or ran into them in the streets.40 Furthermore, the choice of partners was often made among those of the same ethnic group.41

It should be emphasized that ethnic groups among slaves in the New World were not simply duplicates of the groups in their African homeland. The testimony of the slave Antônio at his trial following the Malê Revolt of 1835 exemplifies the process of reconstructing an African nation in Salvador. This Nagô defendant, a chair carrier by trade, testified, “... we all are Nagôs, but each of us has his or her own homeland [terra].”42 He made the statement in the context of his self-defense as a Nagô, the African nation with which the majority of participants in the uprising were identified.43 His remark demonstrates that in New World slavery, those who came from the same African language group but from different ethnic lineages or subgroups often associated with one another under the label imposed by the slaveowners. As a group, they sought and continually created common symbols they could share among themselves. Ethnicity was thereby symbolically recreated in the New World slave community.

How, then, was this ethnic solidarity represented in the manumission of slaves, and particularly in the process of self-purchase? The answer requires some knowledge of the general patterns of manumission in nineteenth-century Salvador.

Patterns of Manumission

Over the entire period 1808–1884, Brazilian-born slaves benefited from manumission somewhat more frequently than did their African-born counterparts. Breaking down these figures by shorter periods shows some interesting changes in the population of freedpersons (table 3).44 During the first three decades of the nineteenth century, although African-born slaves outnumbered Brazilian-born slaves by at least two to one, Brazilian-born freedpersons outnumbered their African-born counterparts. The proportion of African-born slaves in the slave population as a whole eventually declined after the slave trade was declared illegal in 1831, but African-born slaves continued to outnumber Brazilian-born slaves until midcentury.45

In the period 1831–1852, African-born freedpersons outnumbered their Brazilian-born counterparts because self-purchase by the African-born increased. For the years 1851–1884, when Brazilian-born slaves outnumbered African-born slaves by the substantial ratio of 64:36 among those purchased and sold in Sé parish, Brazilian-born freedpersons once again outnumbered African-born freedpersons, but by a much smaller ratio than that of the slave population (table 5). This suggests that members of the aging African-born enslaved population were manumitted more frequently than Brazilian-born slaves, who carried more commercial value.46

In nineteenth-century Salvador, manumission favored women only slightly over men, by a ratio of 57:43 (table 4). At the same time, the urban environment enabled a larger proportion of slave men to gain their freedom than probably could have done so on rural plantations. The male-female ratio was nearly equal for African-born freedpersons (47:53), whereas among Brazilian-born freedpersons nearly 60 percent of the adults and 70 percent of the children were females (table 4).47 Furthermore, table 4 shows some interesting changes in the distribution of freedpersons by gender. The keen shortage of male agricultural labor and the resulting higher prices of male slaves led to an abrupt decline in the percentage of male freedpersons in the period 1851–1852, but it was only temporary. For 1878–1884, African-born male freedpersons outnumbered their female counterparts by the ratio of 3:2, while among Brazilian-born freedpersons the percentage of women rose higher than ever because of the great demand for young male workers. In the last decade of slavery, owners were more willing to manumit slave men of African birth, who had worked as field hands and transporters throughout their lives and were now ill, handicapped, and aging. African-born slave women of about the same age could still work well as domestics.

Salvadoran slaveowners favored manumitting children over adult slaves. Thirty percent of Brazilian-born freedpersons for the period 1808–1884 were children.48 This seems to be a much higher percentage than that among Brazilian-born slaves; for example, the adult-child ratio was 85:15 among slaves purchased and sold in Sé parish for 1838–1888.49 A slightly higher proportion of females (33.5 percent) than males (26.0 percent) were manumitted as children.50

Among African-born freedpersons, 87.3 percent were from West Africa, 12.2 percent from Central West Africa, and 0.5 percent from East Africa for 1808–1842; and 91.2 percent from West Africa, 8.5 percent from Central West Africa, and 0.3 percent from East Africa for 1851–1884 (table 5). The Nagô nation was the most prominent (31.3 percent for 1808–1842 and 73.9 percent for 1851–1884). The geographic and ethnic distribution of African-born freedpersons corresponds to that of the African-born slaves purchased and sold in Sé parish for 1838–1848 and 1852–1888 (table 2). Thus geographic or ethnic origins apparently were not determining factors in the individual African-born slave’s attempt to gain freedom.51 For 1851–1884, less than half (48.8 percent) of the African-born freedpersons were identified by nation in their letters of liberty. The same change took place in the legal registers of slaves in Sé parish. With no more new arrivals from Africa after midcentury, specification of African origins probably no longer mattered so much, either for official use in records or for African-born persons themselves.52

Brazilian-born persons of color were classified into several categories, mainly based on the degree of skin color. Whereas prêto (black) usually meant African-born and negro (black) referred to African-born slave, at least in the early nineteenth century, the vocabulary of “color” (côr) for the Brazilian-born included crioulo (Brazilian-born black); pardo (mulatto); mulato (mulatto with pejorative connotation); cabra (mulatto), mestiço (mixed-breed in general); and caboclo (person of Indian-European ancestry). The use of these terms was often situational and poorly defined.53

The ratios of blacks to mulattos for Brazilian-born freedpersons were 64:36 for 1808–1842 and 63:37 for 1851–1884 (table 6), whereas the Brazilian-born slaves purchased and sold in Sé parish indicate higher percentages of blacks, such as 75.2 percent for 1838–1848 and 67.4 percent for 1852–1888 (table 2). Apparently, slaves of mixed origin, regardless of gender and age, attained freedom more easily than Brazilian-born blacks; but this tendency is most clearly observed in the case of female children (table 6).54

Forms of Manumission

For African-born slaves of both sexes, the most common path to freedom was through self-purchase.55 Half of them paid cash (in réis) to purchase their freedom (47.6 percent of males and 52.8 percent of females for 1808–1842; 63.3 percent of males and 51.1 percent of females for 1851–1884). By contrast, a much smaller percentage of Brazilian-born slaves (about 20 percent of both males and females for 1808–1842 and less than 30 percent for 1851–1884) were manumitted by self-purchase (tables 7 and 8). Among those of Brazilian birth who purchased freedom, only 8.1 percent of males and 7.0 percent of females were children. After midcentury, when slave prices began to rise (see figures 1 and 2), the percentage of paid manumissions in every category (male-female and African-born—Brazilian-born) increased, whereas that of unpaid manumissions decreased.56

The practice of self-purchase did not merely reflect the commercial value of an individual slave.57 African-born males were likely to be more expensive than African-born females, whereas Brazilian-born males were less expensive than Brazilian-born females until the late 1860s (figures 1 and 2). The letters of liberty show that males of African birth paid more for self-purchase than did their female counterparts, except for 1871–72 (figures 3 and 4). Although skilled Brazilian-born female slave domestics usually cost more than their African-born counterparts, many of whom had fewer occupational skills (figures 1 and 2), the latter were generally forced to pay higher prices for their freedom (figure 4). Thus the price of freedom did not necessarily correlate with the commercial value of the individual slave. The owner negotiated the price with the slave, knowing that African-born women had the capacity to pay more than their price as commodities, since many of them were successful entrepreneurs.

Unfortunately, the letters of liberty rarely mention the occupation or occupational skill of the slave to be freed. Among 3,516 manumission letters for the whole period 1808–1884, only 69 state the occupations of slaves. The letters represent 46 men and 23 women; 29 African-born, 29 Brazilian-born, and 11 whose birthplaces were not stated. The men’s occupations include 13 artisans, 9 domestics, 8 porters, 8 field hands, 4 barber-surgeons, and 4 wage earners; the women’s number 11 domestics, 10 field hands, and 2 wage earners.58

The only primary document on the daily wages of persons of African descent that could be found for this study is the table of registers of African-born freedpersons (92 men and 115 women) who lived in Santana parish in 1849. These data are too small a sample to permit valid generalizations, and they refer only to African-born freedpersons. But some trends are worth noting. At least in this sample, successful market women earned much more than skilled domestic servants. The average daily wage of 89 mercadejas (market women) was 805 réis, and that of 28 quitandeiras (female market-stall keepers) was 981 réis. Thirteen women in business (negócio) earned a higher average daily wage of 1$283 réis. By contrast, female domestic servants with special skills as laundresses (9 individuals) earned only 320 réis a day. As for men, chair carriers (25 individuals) earned the average daily wage of 794 réis, and 24 individuals whose professions were registered generally as wage earners (ganhadores) earned the daily wage of 793 réis.59

Brazilian-born slaves had kinspersons and godparents who could help them achieve freedom. For the years 1808–1842, 13 mothers, 5 fathers, 2 other kinspersons (a brother and an aunt), and 7 godparents paid cash to purchase the manumissions of Brazilian-born slaves. For the period 1851–1884, 38 mothers, 6 fathers, 2 other kinspersons (African-born grandmothers), and 9 godparents did the same.60

In some cases, those who purchased the freedom were unrelated to the slaves. For instance, two Brazilian-born sisters, Euporoma, 6 years old, and Justina, 9, daughters of their owner’s former slave, Verdina, a crioula, were liberated with 1.2 million réis donated by Manuel Joaquim Gomes Vilças because he “appreciated the services” not only of their mother but also their grandmother.61 In 1880, Guirina, a 23-year-old crioula, received her letter of liberty on payment of 800 milréis: 400 from herself and 400 from Margarida dos Sassos, who had rented Guirina for fieldwork.62 Gerge Hareis Duder paid 950 milréis for the freedom of Luís, crioulo, but on the condition that Luís continue to serve him for seven years.63

Bahian abolitionist societies, established in the 1860s, contributed money for manumissions, but only to Brazilian-born slaves and mainly to infant and child slaves. These societies not only helped pay for the slave’s self-purchase but also filed criminal lawsuits against slaveowners who mistreated their slaves. For instance, on January 3, 1887, the Sociedade Libertadora Bahiana filed legal papers on behalf of the 30-year-old, Brazilian-born slave Silvestre, resident of the engenho Agua Comprida, against his owner, Ana d’Argolo.64

In contrast to their Brazilian-born counterparts, most African-born slaves had no kinspersons or local supporters to assist them (tables 7 and 8). For the period 1808–1884, at least in the description of manumission letters, no kinspersons paid any money for the freedom of African-born slaves, and only three godfathers contributed to the manumissions of three African-born slaves, all women. This, of course, should not be taken to mean that African-born slaves lacked local assistance for self-purchase. For instance, the wills of freedpersons include six examples of slave women, all African-born, manumitted with money from their common-law husbands, who were all freedmen (five African-born and one Brazilian-born). All of the couples contracted legal matrimony once the wives obtained their letters of liberty.65

Some African-born slaves borrowed money for self-purchase from the black lay brotherhoods. Established before the turn of the nineteenth century by slaves of specific African nations, many of these groups maintained manumission pools to provide their slave members with loans. By the beginning of the nineteenth century, however, the black brotherhoods in Salvador had eliminated the provision of emancipation pools and extended their charitable activities to the urban poor in general.66 This explains why, out of all the Salvadoran manumission letters from 1808 to 1884 examined in this study, no letter mentions such manumission pools; but it does not mean that slaves could no longer rely on their brotherhoods for financial assistance in self-purchase.

Among the freedpersons who registered their wills in Salvador during this period, only one mentions that he was freed by a black brotherhood, and the purchase took place in Portugal, not Brazil. José Ignácio Joaquim, originally from Costa da Mina, accompanied his owner, Ignácio da Costa, to Lisbon on business and lived there for a time. Joaquim bought his freedom in Lisbon with funds provided by the Black Brotherhood of Our Lady of the Rosary. Later he returned to Salvador, married Ana Cruz Velozo, also from Costa da Mina, and registered his will on April 10, 1826.67

Brazilian-born slaves benefited more frequently than the African-born from unpaid manumissions, many of which were unconditionally granted for “good services,” “fidelity and obedience,” “love of God,” or “love of raising the slave” (tables 7 and 8). In most cases of conditional un paid manumissions, the owners obliged their ex-slaves to “accompany and serve” them (or sometimes their spouses, children, or siblings) until the owners died. Those manumitted in this way, now called agregados (dependents in households), continued to work for their former owners the same as they had as slaves.

Brazilian-born slaves also had the advantage of being manumitted by their owners out of feelings of paternalism and affection. Some newborn and child slaves were baptized as forros to express the owner’s gratitude to the slave mother.68 In only one of these cases did the owner become his slave’s godfather, in order to “benefit him.”69 As long as cheaper human commodity was continuously supplied from Africa, owners in Brazil usually preferred to manumit Brazilian-born slaves at birth and during childhood, unconditionally and without payment. Because of the high risk of infant and child mortality, raising those slaves was much costlier than purchasing newly arrived African-born adult slaves.

Brazilian-born child slaves were occasionally manumitted with their mothers. Among the 3,516 letters for 1808–1884, 63 describe this type of transaction.70 Two-thirds of these cases were unconditionally unpaid manumissions, because of the mother’s “good services.” One exceptional case is the unconditional liberation of a whole Brazilian-born slave family: Gormano, pardo, the father; Maria Margarida, the mother; and their children, Rodrigo, crioulo; Perigrino, Felsarda, and Maria, cabras.71 In the sample for this study, only two slave children, both mulatinhos and both the same owner’s illegitimate children by his slave Marcelina, were manumitted.72

Varieties of Self-Purchase

The mode of self-manumission usually took the form of cash (in réis).73 One exception is found among the wills of freedpersons. A Nagô woman, Gertudes Maria do Espírito Santo, paid her value to her owner, Silveiro da Silva, and his wife, Joana da Silva, with a part of her real estate, namely a house located on Rua do Gimipaneiro in Santana parish.74

One interesting form of self-purchase was substitution, as in the aforementioned case of Francisco.75 For the period 1808–1888, the sample letters of liberty record 35 cases of substitution. These include 3 cases in which the freedperson provided not only the substitute but also cash to make up the difference between the freedperson’s price and the substitute’s lower price. An example will illustrate such negotiated settlements between slaves and owners. Ventura, Mina, by occupation a bricklayer, paid three hundred milréis to his owner. Two hundred milréis was represented by the value of another slave of the same occupation named Torcano, characterized as “gentio da costa” (gentile from the coast), and one hundred milréis was paid in cash.76

These 35 cases of self-purchase through substitution constitute about 1 percent of the total number of letters of liberty, and 2.63 percent of selfpurchase (1,332), for the period 1808–1884. With one exception, in 1861, all cases occurred between 1808 and 1852. The freedpersons numbered 17 males and 18 females; 24 African-born of diverse origins (13 males and 11 females), 3 Brazilian-born (1 male and 2 females), and 8 with no indication of origin or color (3 males and 5 females).77 Substitute slaves were usually of the same gender as the freedpersons (29 out of 35 cases). One-to-one replacement was common, but in two cases, two slaves of the same gender were substituted for one freedperson.78 The common phrase used in cases of substitution was “the slave gave another slave in his [or her] place,” but specific occupations were mentioned only in four letters, including that of Francisco.79

In just two cases, the substitute had the same name as the person freed.80 Because, as noted earlier, the African-born often referred to one another by their African names, one plausible interpretation is that the freed slave registered the substitute with his or her own Catholic baptismal name to show legal ownership, but used the African name for communication.

Few letters of liberty tell how a substitute was obtained. Only two state that someone else supplied the substitute. A slave woman named Josefa received her substitute, Rita, from her niece, Marcelina Maria da Conceição. The crioula slave Bernardeira provided another female slave, Rola, Nagô, who had belonged to Bernardeira’s mother, Felicidade, slave of a different owner. In both cases, kinspersons significantly contributed to the individual manumissions.81

For Bahia, Schwartz presents some interesting testimony given before the probate judge of São Francisco do Conde in 1836. A crioula slave, Luciana Maria da Conceição, wanted to purchase a slave as dowry for her granddaughter. She sent money to a friend, who went to Africa and acquired a Nagô woman, Jeronima, who was subsequently delivered to Luciana at the engenho Cahipe. But Luciana changed her mind. Instead of contributing to her granddaughter’s dowry, she kept Jeronima as her personal slave and sent her to the city to earn wages, while Luciana herself continued to work as a slave on the plantation.82 The ending to this story is unknown. Was Luciana finally freed, possibly by using Jeronima’s accumulated earnings or by substituting Jeronima for herself? Indeed, it is not known if Luciana even made the effort to gain her own freedom. But this example does reveal that some agents, possibly African-born freedpersons, regularly traveled to Africa to trade in slaves and to make them available at reasonably inexpensive prices to buyers in Salvador.83

Substitute slaves were usually newly arrived Africans, whose prices were lower than those of skilled Brazilian-born slaves or acculturated African-born slaves.84 This explains why the practice of substitution nearly disappeared with the termination of the transatlantic slave trade. Twentyfive of the letters declare that the substitute slaves were African-born or new arrivals; the other ten provide no information about the slaves’ birthplace or nation. With the owner’s consent, a slave purchased the substitute, acculturated and trained the newcomer in special occupational skills, and finally “traded in” the substitute for the slave’s own freedom.

As the case of Francisco illustrates, some slaves purchased substitutes of the same nation. Of 19 cases in which both slave and substitute were identified, the replacements for 6 Nagô slaves were also Nagô. This could be attributed to the law of probability, because individuals defined as Nagôs constituted one-third of all African-born freedpersons. But the studies of colonial and imperial manumission by Schwartz, Mary C. Karasch, and Kathleen Joan Higgins also note the coincidence of nations between slaves and substitutes, thereby demonstrating that such coincidence was not confined to the Nagô.85

Substitution was an attractive expedient for both owner and slave. Karasch is correct in stating, “Owners willingly accepted these ‘trade-ins,’ for they did not have to acculturate them, and they solved the problem of capital depreciation. In place of an aging African woman, they might receive a teenage boy with years of service in the future. ”86 For slaves, the financial advantage of purchasing a new African-born slave much more cheaply than accumulating their own value for self-purchase was offset by the temporal consideration that the new slave needed time to acculturate before qualifying as a viable replacement. Some slaves, in their capacity as slaveowners, may have worked side by side with their own slaves, or may have sent them out into the street as escravos de ganho to earn money for several years before finally offering them as substitutes. In this context, purchasing a substitute from the same ethnic group made sense: it eliminated the problem of instructing the new arrival who did not understand the Portuguese language or Luso-Brazilian culture. Thus the common nation between freedperson and substitute was no mere coincidence, as Schwartz and Karasch imply, but rather another reflection of recreated ethnicity in the slave community of the New World, where individuals of African birth struggled daily to survive.

The slaveholding pattern of African-born freedpersons supports this revisionist interpretation.87 Many African-born freedpersons who owned both slaves of their own nation and crias (Brazilian-born slaves raised in their owners’ households) treated each group differently. The double standard is readily apparent in their wills. Whereas many Brazilian-born slaves were liberated on their owners’ death with no conditions—and even occasionally named as heirs—many African-born slaves, including those who shared their owners’ nation, were usually coartado; that is, obliged to pay a fixed price to the heirs for their freedom within a certain period of time after the owner’s death.88 This practice deviates from Karasch’s interpretation. She states: “One other factor that appears to have encouraged freedpersons to own slaves is that while the African trade continued, they had the opportunity to rescue their own people from being enslaved to Brazilians. This pattern is suggested in the negros de ganho records in which Yorubas owned other Yorubas as slaves.”89 A freed Nagô slaveholder might have taught Nagô slaves in their common language, but doing so did not necessarily mean that the owner favored them for manumission.

According to preceding manumission studies, the custom of substitution seems to have been limited to the urban environment, such as the port cities of Salvador and Rio de Janeiro and the mining town of Sabará in Minas Gerais. Substitution was not so prevalent in agricultural regions, such as Paraty (1789–1822), Paraíba (1550–1888), or even the Bahian Recôncavo of the nineteenth century.90 This comparison, however, should not be taken to imply that substitution was exclusively an urban phenomenon. At the beginning of the nineteenth century, Henry Koster observed a similar practice on a sugar plantation owned by the Benedictines in Jaguaribe, Pernambuco. More than one hundred slaves, all Brazilian-born, belonged to this plantation, whose overseer was a mulatto slave. The overseer first purchased the freedom of his mulatto wife, who had been owned by the order, and then that of his children. Subsequently, in exchange for his own freedom, he offered the friars two African-born slaves he owned. The friars rejected his proposal on the grounds that the estate could not be properly managed without him.91 Substitution rarely occurred in rural agricultural areas because of the plantation slave system and the relative lack of socioeconomic mobility of slaves who were field hands.


In nineteenth-century Salvador, the practice of slave manumission took various forms: unconditional unpaid manumission, conditional unpaid manumission, manumission paid by others, and self-purchase. Generally, Brazilian-born, female, child, and mulatto slaves found it easier to obtain manumission without cost than did African-born, male, adult, and crioulo slaves. Some slaves were freed legally with payment by their kins-persons, Active kinspersons, and others. Other slaves, particularly African-born slaves of both sexes, were often able to purchase their own freedom. Whereas ethnic-group solidarity in a slave community presumably was a factor in paid manumissions, there were also cases in which shared ethnicity in a slave society could be an ambivalent factor. One such case was substitution.

By focusing on the practice of self-purchase by slaves of African birth, this essay has explored some connections between manumission and ethnicity in New World slave communities. Individual struggles for survival in a slave society continuously created and recreated ethnic groups as well as mutual-aid associations among the slave population, but ethnic solidarity did not necessarily hamper the same individuals in pursuing their own personal gain by somehow utilizing shared ethnicity. The proportion of substitution as a form of self-purchase is certainly small, but it provides an impetus to reconsider the manumission of slaves and the nature of African slavery in the New World.

An earlier version of this essay was presented at the Sixth Annual Virginia-Carolinas Latin American Colonial History Seminar, held at the National Humanities Center in 1991. The author would like to thank Professors Richard Graham, William E. Jackson, Joseph C. Miller, A. J. R. Russell-Wood, William B. Taylor, and the anonymous HAHR reviewers for their critical readings and helpful comments at various stages of writing and revising; and the Carter G. Woodson Institute of the University of Virginia for a predoctoral fellowship (1989–1991).

Research material was drawn from the following archives: Arquivo Público do Estado da Bahia, Salvador (APB); Arquivo Municipal da Cidade do Salvador (AMCS); Arquivo da Curia Metropolitana de São Salvador da Bahia, Salvador (ACMS); Arquivo Nacional do Rio de Janeiro (ANRJ); Biblioteca Nacional do Rio de Janeiro (BNRJ).


APB, Seção Judiciária, Livro de notas da capital (hereafter cited as LNC), no. 306, May 14, 1852. The real (plural réis) was the Brazilian currency of account. The symbol $ was used to indicate thousands (milréis) and the symbol : for millions (contos).


The term trade-in is borrowed from Mary C. Karasch, Slave Life in Rio de Janeiro, 1808–1850 (Princeton: Princeton Univ. Press, 1987), 343, 358.


None of these letters states the ex-owner’s specific district or street of residence. Since notarial offices had already been established in the towns of the Recôncavo by the beginning of the eighteenth century, presumably most, if not all, of the ex-owners who registered letters of liberty at notarial offices in Salvador in the nineteenth century were residents of Salvador. See Stuart B. Schwartz, “The Manumission of Slaves in Colonial Brazil: Bahia, 1684–1745,” HAHR 54:4 (Nov. 1974), 603–35, esp. 607–8. Among the 3,156 letters of liberty examined for this study, in only 16 do ex-owners declare that they themselves were African-born freedpersons. This may reflect the limitation that a very small percentage of freedpersons could become slaveowners. For instance, 186 (89.9 percent) of 207 African-born freedpersons who resided in Conceição da Praia parish in 1846 did not own a single slave. Likewise, 263 (79.5 percent) of 331 African-born freedpersons who resided in Santana parish in 1849 were non-slaveowners. See Arollomento dos africanos libertos que rezedem nesta freguesia da Conceição da Praia, Jan. 31, 1846, APB, Seção Histórica, mago 6472; Relagáo dos africanos libertos existentes nesta freguesia com observações seguintes, Santana, Feb. 11, 1849, APB, mago 2898. On non-slaveowners in colonial and imperial Brazil, see Iraci del Nero da Costa, Arraiamiúda: um estado sobre os não-proprietários de escravos no Brasil (São Paulo: MGSP Editores, 1992).


Letters for 1878–79 and 1880–81 are in APB, Secão Histórica, maço 2880. Letters for 1881–82 could not be consulted because of a change in the way scattered registration documents were organized in the livros after 1880. The rest are in APB, LNC.


APB, Seção Judiciária, Livros de registro de testamentos da capital (hereafter cited as LRTC), 64 vols. A total of 470 wills were registered by freedpersons for the period 1790–1890. Of the testators, 112 are males and 145 females for 1790–1850; 128 males and 95 females for 1851–1890. See Maria Inês Cortes de Oliveira, O liberto: o seu mundo e os outros (São Paulo: Corrupio, 1988), 8; and AMCS, Livros de escrituras de compra e venda de escravos, freguesia da Sé (hereafter LECV), nos. 82.1–82.20. Salvador’s Municipal Archives house a great number of legal documents concerning slaves registered in the parishes of the city. Those for Sé parish have been preserved almost in their entirety.


Studies of manumission in colonial and imperial Brazil have covered regions (Bahia and Paraíba) and specific towns and cities (Rio de Janeiro, Paraty, Sabará, Campinas, and Cachoeira and São Felix of the Bahian Recôncavo). See Schwartz, “Manumission,” 603–35; Kátia M. de Queirós Mattoso, “A propósito de cartas de alforria na Bahia, 1779–1850,” Anais de História 4 (1972), 23–52; Barbara Rose Trosko, “The Liberto in Bahia Before Abolition” (M.A. thesis, Columbia Univ., 1967); Diana Soares de Galliza, O declínio da escravidão na Paraíba 1850–1888 (João Pessoa: Editora Universitaria, 1979), chap. 4; Karasch, Slave Life, chap. 11; Sidney Chalhoub, “Slaves, Freedmen, and the Politics of Freedom in Brazil: The Experience of Blacks in the City of Rio,” Slavery and Abolition 10:3 (Dec. 1989), 64–84; idem., Visões da liberdade. uma história das últimas décadas da escravidão na corte (São Paulo: Companhia das Letras, 1990); James Patrick Kiernan, “The Manumission of Slaves in Colonial Brazil: Paraty, 1789–1822” (Ph.D. diss., New York Univ., 1976); Kathleen Joan Higgins, “The Slave Society in Eighteenth-Century Sabará: A Community Study in Colonial Brazil" (Ph.D. diss., Yale Univ., 1987), chap. 4; Robert Wayne Sienes, “The Demography and Economics of Brazilian Slavery, 1850–1888” (Ph.D. diss., Stanford Univ., 1975), chap. 10; Peter L. Eisenberg, “Ficando livre: as alforrias em Campinas no século XIX,” Estados Econômicos 17:2 (May-Aug. 1987), 175–216; Fayette Darcell Wimbery, “The African Liberto and the Bahian Lower Class: Social Integration in Nineteenth-Century Bahia, Brazil, 1870–1900” (Ph.D. diss., Univ. of California, Berkeley, 1989), chap. 2. For studies of manumission in other parts of Latin America, see Frederick P. Bowser, The African Slave in Colonial Peru, 1524–1650 (Stanford: Stanford Univ. Press, 1974), chap. 10; idem., “The Free Person of Color in Mexico City and Lima: Manumission and Opportunity, 1850–1950,” in Race and Slavery in the Western Hemisphere: Quantitative Studies, ed. Stanley L. Engerman and Eugene D. Genovese (Princeton: Princeton Univ. Press, 1975); Lyman L. Johnson, “Manumission in Colonial Buenos Aires, 1776–1810,” HAHR 59:2 (May 1979), 258–79.


Luiz dos Santos Vilhena, A Bahia no século XVIII, 3 vols. (Salvador: Editora Itapuã, 1969), 2:460.


Luís Viana Filho, O negro na Bahia, 3d ed. (Rio de Janeiro: Nova Fronteira, 1988), 32; Pierre Verger, Trade Relations Between the Bight of Benin and Bahia from the Seventeenth to the Nineteenth Century, trans. Evelyn Crawford (Ibadan, Nigeria: Ibadan Univ. Press, 1976). See also Philip D. Curtin, The Atlantic Slave Trade: A Census (Madison: Univ. of Wisconsin Press, 1969); Joseph C. Miller, Way of Death: Merchant Capitalism and the Angolan Slave Trade, 1730–1830 (Madison: Univ. of Wisconsin Press, 1988).


João José Reis, Rebelião escrava no Brasil: a história do levante dos malês (1835), 2d ed. (São Paulo: Editora Brasiliense, 1987), 16.


Brasil, Diretoria Geral de Estatística, Recenseamento da população de Imperio do Brazil a que se procedeu no dia 1 de agosto de 1872 (Rio de Janeiro: Leuzinger, 1873–76), 508, 510.


David Eltis, Economic Growth and the Ending of the Transatlantic Slave Trade (New York: Oxford Univ. Press, 1987), p. 244.


Ibid., 199.


Ibid., 244.


ACMS, Livros de batizados, freguesia do Santo Antônio. The 1,235 registers consulted for this study include 211 male and 202 female African-born adult slaves: 44 males and 55 females (1809–10); 80 males and 61 females (1818–19); 80 males and 80 females (1828–29); 2 males and 2 females (1838–39); 4 females (1848–49); 5 males and 1 female (1858–59).


Thomas Merrick and Douglas H. Graham, Population and Economic Development in Brazil: 1800–Present (Baltimore: Johns Hopkins Univ. Press, 1979), 54; Robert Conrad, The Destruction of Brazilian Slavery, 1850–1888 (Berkeley: Univ. of California Press, 1972), 7–8; Herbert S. Klein, The Middle Passage: Comparative Studies in the Atlantic Slave Trade (Princeton: Princeton Univ. Press, 1978), chap. 5.


Curtin, Atlantic Slave Trade, 19, 41–47; William D. Phillips, Jr., Slavery from Roman Times to the Early Transatlantic Trade (Minneapolis: Univ. of Minnesota Press, 1985), 122–24.


Eltis, Economic Growth, 257 (table B.2).




APB, Seçâo Histórica, Série Governo, maços 1602 and 1605; Brasil, Recenseamento, 510.


For example, six Bahian engenhos (sugar plantations) in 1816, which Schwartz studied, showed 275 males for every 100 females. See Stuart B. Schwartz, Sugar Plantations in the Formation of Brazilian Society: Bahia, 1550–1835 (Cambridge: Cambridge Univ. Press, 1985), 348.


Occupations of male slaves (3,168 individuals): artisans, 738 (23.3 percent); domestics, 608 (19.2 percent); portage and transportation, 590 (18.6 percent); fieldwork, 529 (16.7 percent); wage earning, 278 (8.8 percent); maritime, 234 (7.4 percent); food processing, 52 (1.6 percent); barber-surgeons and medical service, 32 (1.0 percent); sales, 33 (1.0 percent); entertainment, 15 (0.5 percent); all services, 13 (0.4 percent); others, 46 (1.5 percent). Occupations of female slaves (2,416 individuals): domestics, 1,791 (74.1 percent); wage earning, 367 (15.2 percent); fieldwork, 193 (8.0 percent); sales, 61 (2.5 percent); all services, 2 (0.1 percent); other, 2 (0.1 percent). See Maria José de Sousa Andrade, A mão de ohra escrava em Salvador 1811–1860 (São Paulo: Corrupio, 1988), 129–30.


Only 11.4 percent (42 individuals) owned 11 to 20 slaves. See APB, Seção Judiciária, Inventários da capital (1808–1888).


Reis, Rebelião escrava, 25.


See, for instance, James Wetherell, Brazil: Stray Notes from Bahia, Being Extracts from Letters, &c., During a Residence of Fifteen Years, ed. William Hadfield (Liverpool: Webb and Hunt, i860), 16–17. On the hiring-out system of urban slavery in the U. S. South, see Richard Wade, Slavery in the Cities: The South, 1820–1860 (New York: Oxford Univ. Press, 1964), esp. 34–40, 48–49.


See, for example, Andrade, A mão de obra, 129–30.


On diverse “functions” of urban slaves in early nineteenth-century Rio de Janeiro, see Mary C. Karasch, “From Porterage to Proprietorship: African Occupations in Rio de Janeiro, 1808–1850,” in Engerman and Genovese, Race and Slavery; Karasch, Slave Life, chap. 7; Luís Carlos Soares, “Os escravos de ganho no Rio de Janeiro do século XIX,” Revista Brasileira de História 8:16 (Mar.-Aug. 1988), 107–42.


Unfortunately, this study found no statement of how many years it took for a slave to accumulate enough money to purchase freedom, either in the letters of liberty or the wills of freedpersons.


Curtin, Atlantic Slave Trade, 184–90.


Compare Karasch’s treatment of nations in early nineteenth-century Rio de Janeiro. Karasch, Slave Life, chap. 1.


One of the participants in the Malê Revolt (1835), a Nagô slave called Matheos, whose African name was Dadá, testified that two other participants, Belchior and Gaspar da Silva, were his malungos. See “Devassa do levante de escravos occorido em Salvador em 1835,” Anais do Arquivo do Estado da. Bahia (hereafter AAEB) 38 (1968), 1–142, esp. 32–38. On malungos, see Henry Koster, Travels in Brazil, 1809–1815, 2 vols. (London: Longman, Hurst, et al., 1817), 2:251; and Karasch, Slave Life, 298.


Franklin W. Knight and Margaret E. Crahan, in their discussions of ethnicity in New World slave societies, state the following: “Specific ethnic identifications did not necessarily indicate the retention of the formerly associated ethnic culture.... Representative ethnic cultures survived, but without necessarily reflecting closely the basic original mix of the African loyalty or region.” “The African Migration and the Origins of an Afro-American Society and Culture,” in Africa and the Caribbean: The Legacies of a Link, ed. Crahan and Knight (Baltimore: Johns Hopkins Univ. Press, 1979), 12.


On slave uprisings from 1807 to 1835, see R. K. Kent, “African Revolt in Bahia, 24–25 January 1835,” Journal of Social History 3:4 (Summer 1970), 334–56; Howard M. Prince, “Slave Rebellion in Bahia 1807–1835” (Ph.D. diss., Columbia Univ., 1972); Schwartz, Sugar Plantations, chap. 17; Reis, Rebelião escrava. In the Malê Revolt, the majority of the rebels were Nagôs, but unlike the preceding uprisings (1807–30), both slaves and freedpersons were involved almost equally. See Reis, Rebelião escrava, 172.


See, for example, APB, Série Ordens Régias, vol. 109, doc. 73; ANRJ, Ministro de Reino, IJJ9–318, fol. 9, and IJJ9–323, fols. 196–98; BNRJ, Seçâo de Manuscritos, II–33.22.72. Quilombo derives from the Angolan term kilombo. See Joseph C. Miller, Kings and Kinsmen: Early Mbundu States in Angola (Oxford: Clarendon Press, 1976), 162 and chap. 8.


Wetherell, Brazil: Stray Notes, 5 and 54. See also Nina Rodrigues, Os africanos no Brasil, 3d ed. (São Paulo: Companhia Editôria Nacional, 1945), chap. 6.


For example, many African-born participants in the Malê Revolt, both slave and exslave, were identified by other African-born persons only by their African names. See the trial and judicial records of the revolt, many of which were published by the State Archives of Bahia in AAEB as “Devassa do levante” (see note 30); and ‘1835 Insurreição de escravos,” AAEB 40 (1970), 9–170. See also Prince, “Slave Rebellion,” 186; and Reis, Rebelião escrava, 197 (table 7).


Rodrigues, Os africanos, 174–75. See also J. da Silva Campos, “Ligeiras notas sobre a vida intima: costumes e religião dos africanos na Bahia,” AAEB 29 (1943), 291–309.


Official correspondence from Bahia to the Portuguese court in Rio de Janeiro in 1814 observes: “In Rio de Janeiro, 10,000 or 12,000 blacks dance according to their nations but all together in the Campo de Santana, whereas here in Salvador, 100 Nagôs dance in Largo do Teatro, 50 Gêges in the Piedade, 80 Hausas in Rua de João Perreira, and thus all over the city.” Correspondencia do Présidente da Província, ANRJ, Ministro do Reino, Bahia, IJJ9–323, fol. 20. See also BNRJ, Seção de Manuscritos, II–34.6.5. (Gêge refers to the Ewe.) This correspondence provides an interesting contrast between Salvador and the other major port city, Rio de Janeiro, in which various ethnic groups shared the same public space. Salvador’s longer history of importing Africans of diverse ethnic origins into urban slavery may have established more solid ethnic patterns in the city’s slave population. On Salvador and Rio de Janeiro as the two major port cities of colonial Brazil, see A. J. R. Russell-Wood, “Ports of Colonial Brazil,” in Atlantic Port Cities: Economy, Culture, and Society in the Transatlantic World, ed. Franklin W. Knight and Peggy K. Liss (Knoxville: Univ. of Tennessee Press, 1991), 219–29.


Verger, Trade Relations, 465; Notícias da Bahia, 1850 (São Paulo: Corrupio, 1981), 64–65. See also Manoel S. Cardozo, “The Lay Brotherhoods of Colonial Bahia,” Catholic Historical Review 33:1 (Apr. 1974), 12–30; A. J. R. Russell-Wood, “Black and Mulatto Brotherhoods in Colonial Brazil: A Study in Collective Behavior,” HAHR 54:4 (Nov. 1974), 567–602; Patricia A. Mulvey, “Black Brothers and Sisters: Membership in the Black Lay Brotherhoods of Colonial Brazil,” Luso-Brazilian Review 17:2 (Winter 1980), 253–79; João José Reis, A morte é wna festa: ritos fúnebres e revolta popular no Brasil do século XIX (São Paulo: Companhia das Letras, 1991), 49–72; and Mieko Nishida, “Gender, Ethnicity, and Kinship in the Urban African Diaspora: Salvador, Brazil, 1808–1888” (Ph.D. diss., Johns Hopkins Univ., 1991), chap. 5.


The juntas are known only from the description by the contemporary mulatto historian Manuel Querino in the 1870s. No primary data on the juntas exist today, because, according to Querino, nothing about their activities was written down. See Manuel Querino, Costumes africanos no Brasil (Rio de Janeiro: Civilização Brasileira, 1938), 154–56. One of the very few references to the juntas comes in the will (1878) of an African-born freedman named Marcos Gasper, who had been treasurer of the Juntas dos Africanos. See APB, maço 3566/8, cited by Trosko, “Liberto,” 47.


See, for instance, APB, LRTC, no. 27, fols. 128–130B; no. 29, fols. 82B–85; no. 42, fols. 34–36.


For instance, 325 wills of freedpersons show 23 legally married couples in which both partners shared the same nation, while only 7 couples of African birth were of different nations. See APB, LRTC. Unfortunately, few written documents describe the choice of a partner for consensual unions, and the slave population rarely married legally. Furthermore, legal matrimony among freedpersons frequently occurred very late in life, sometimes at the deathbed of one of the spouses. See, for instance, ACMS, Livros de casamentos, freguesia de Conceição da Praia and freguesia de Penha. On family and kinship in general among people of African descent during this period, see Nishida, “Gender, Ethnicity, and Kinship,” chap. 6.


AAEB, “Devassa do levante,” 7.


Howard Prince estimates the number of total participants in the Malê Revolt at between 400 and 500. Prince, “Slave Rebellion,” 169. The police arrested a total of 326 individuals, of whom 284 were brought to trial. Of these defendants, 68.1 percent were Nagôs. See Reis, Rebelião escrava, 172.


By comparison, in Bahia as a whole for the period 1684–1745, Brazilian-born freed-persons made up 69 percent of the total freed, while African-born freedpersons constituted only 31 percent. See Schwartz, “Manumission,” 612.


For instance, 56.1 percent of the slaves in Sé parish for 1838–48 were African-born (see table 2).


On the advantage of Brazilian-born slaves in manumission, see Schwartz, “Manu-mission,” 612; Karasch, Slave Life, 352; Kiernan, “Manumíssion...Paraty,” 92; Higgins, “Slave Society,” 213; and Eisenberg, “Ficando livre,” 189–91.


On the constant 1:2 male-female ratio among freedpersons in Bahia for 1684–1745, 1779–1850, and 1813–1850, see Schwartz, “Manumission,” 611; Mattoso, “A propósito,” 41; Arnold Kessler, “Bahian Manumission Practices in the Early Nineteenth Century” (Paper presented to the American Historical Association, San Francisco, 1973), cited in Schwartz, “Manumission,” 611. The same gender ratio is found in Rio de Janeiro (1807–1831), Paraty (1789–1822), and Lima (1580–1650). See Karasch, Slave Life, 345; Kiernan, “Manumission ... Paraty,” 86; and Bowser, “Free Person,” 350. Higher proportions of women among the manumitted are also found in Sabará (1710–1809), Campinas (1799–1887), Mexico City (1580–1650), and Buenos Aires (1776–1810). See Higgins, “Slave Society,” 205–7; Eisenberg, “Ficando livre,” 184–85; Bowser, “Free Person,” 350; and Johnson, “Manumission in Colonial Buenos Aires,” 262.


APB, LNC and maço 2880. For this study, those over 15 years of age are classified as adults, on the basis of the age for work and military service for males, reproduction and domestic responsibility for females. Furthermore, those identified with diminutives such as crioulinho/crioulinha, pardinho/pardinha, and mulatinho/mulatinha, but without specified ages, are regarded as children. The letters of liberty rarely identify African-born freedpersons with their exact ages until the last two decades of the slavery regime. Likewise, many of the early nineteenth-century manumission letters for Brazilian-born persons do not state their exact ages. Considering the years African-born slaves took to gain freedom, even if they had been transported to Salvador very early in their lives, and also given the aging of the African-born population after the transatlantic slave trade terminated at midcentury, all the African-born freedpersons are considered to be adults.


AMCS, LECV. On the manumission advantage of children, see Schwartz, “Manumission,” 615–16; Kiernan, “Manumission ... Paraty,” 102; Eisenberg, “Ficandolivre,” 192; and Bowser, “Free Person,” 350–51.


APB, LNC and maço 2880.


On the nearly equal distribution of geographic and ethnic origins among manumitted African-born slaves, see Schwartz, “Manumission,” 612–14; Karasch, Slave Life, 345; and Kiernan, “Manumission ... Paraty,” 93. In the mining town of Sabará, Minas Gerais, for 1710–1809, more than twice as many Costa da Mina slaves were freed as slaves from the Congo-Angola region, even though the former outnumbered the latter only by approximately 9 percent in the inventories dating from 1725 to 1808. Kathleen Higgins suggests that the slaves from Mina had more mining experience and accumulated enough gold to buy slaves themselves. See Higgins, “Slave Society,” 216.


The same decline in references to nation after midcentury is found not only in notarial records and parish registers of baptism and marriage but also in the wills of freedpersons. See ACMS, Livras de batizados, freguesia de Santo Antonio; Livras de casamentos, freguesia de Conceição da Praia, freguesia de Penha; APB, LRTC.


On Brazilian descriptions of “color” in the colonial period, see A. J. R. Russell-Wood, “Colonial Brazil,” in Neither Slave nor Free: The Freedmen of African Descent in the Slave Societies of the New World, ed. David W. Cohen and Jack P. Greene (Baltimore: Johns Hopkins Univ. Press, 1972), 84–85. In Brazil, as in Portuguese Asia and Portuguese Africa, negro, prêto, and cafre (Kaffir) were all pejorative terms, often synonymous with escravo (slave). See Charles Ralph Boxer, Race Relations in the Portuguese Colonial Empire, 1415–1825 (Oxford: Clarendon Press, 1963), 120. In New World slave societies, as Cohen and Greene suggest, in the early phase of the emergence of free persons of color (the first several generations), the complex, multilevel color-coding system may have been operative in terms of marriage and social status. However, unlike other societies, in which these categories grew vague and confused even during the first century of slavery, in Brazil this system continued to function throughout the colonial and imperial periods, and it has survived to a considerable extent until today. See Cohen and Greene, Introduction to Neither Slave nor Free, 7.


Concerning the advantage of mulatto slaves in manumission, see Schwartz, “Manumission,” 618; Kiernan, “Manumission ... Paraty,” 92; Higgins, “Slave Society,” 215; Eisenberg, “Ficando livre,” 187; Johnson, “Manumission in Colonial Buenos Aires,” 264–65.


Before the Free Birth Law (Lei do Ventre Livre) of 1871, the prevalent practice of self-purchase had no legal grounds. Despite Henry Koster’s well-known assertion that this was a legal right of slaves, no slaveowner was bound by any legal obligation to liberate slaves who claimed their right to buy themselves out of slavery by offering a sum equivalent to their asserted value. See Koster, Travels in Brazil, 1:404; Lei N. 2040 of Sept. 28, 1871, Art. 4, Colleção das leis do Imperio do Brazil de 1871 (Rio de Janeiro: Typographia Nacional, 1871), part 1, p. 149; Conrad, Destruction of Brazilian Slavery, 191; idem., “Nineteenth-Century Brazilian Slavery,” in Slavery and Race Relations in Latin America, ed. Robert Brent Toplin (Westport: Greenwood Press, 1974), 154–55; Manuela Carneiro da Cunha, “Silences of the Law: Customary Law and Positive Law of Slaves in Nineteenth-Century Brazil,” History and Anthropology 1 (1985), 427–43.


On slave prices recorded in Salvadoran slaveowners’ inventories for 1811–88, see Andrade, A mão de obra escrava, 207–11. On manumissions in nineteenth-century Brazil, Merrick and Graham state: “Approximately one out of every two freed slaves gained their freedom through self-purchase. This proportion grew (and gratis manumissions declined) during periods of rising slave prices.” See Merrick and Graham, Population and Economic Development, 53.


On the prices of manumitted slaves in nineteenth-century Salvador, see Kátia M. de Queirós Mattoso, Herbert S. Klein, and Stanley L. Engerman, “Trends and Patterns in the Prices of Manumitted Slaves: Bahia, 1819–1888,” Slavery and Abolition 7:1 (May 1986), 59–67. On the difference between estimated value and sales price in Paraty, see Kiernan, “Manumission ... Paraty,” 137. On slave prices in colonial and imperial Brazil, see Joseph C. Miller, “Slave Prices in the Portuguese Southern Atlantic, 1600–1830,” in Africans in Bondage: Studies in Slavery and Slave Trade ed. Paul E. Lovejoy (Madison: Univ. of Wisconsin Press, 1986).


Small numbers of freedpersons mentioned their occupations in their wills. See APB, LRTC.


Relação dos africanos, Santana, Feb. 11, 1849, APB, Seção História, maço 2898.


APB, LNC and maço 2880. Examples of grandmothers: Joaquina, Nagô, purchased Anacleta, crioula, daughter of Thomara (deceased), crioula slave of the same owner, with 150 milréis. Mana, of African birth, purchased Thedolina, 11-year-old pardo. See APB, LNC, no. 297, July 1, 1851; and no. 416, Mar. 27, 1872. Godparents: only one of the godparents was identified as African-born: José de São João, Nagô, paid the price of Antônia, 17-year-old crioula (900 milréis). See APB, maço 2888, Feb. 6, 1879. On those who purchased freedom for Brazilian-born child slaves in Rio de Janeiro, see Karasch, Slave Life, 347–50.


APB, LNC, no. 416, Apr. 17, 1872.


APB, maço 2880, Feb. 16, 1880.


Ibid., Nov. 13, 1879.


APB, Seção Judiciária, Autos Crimes, 3531.8. On Bahian abolitionism and abolitionist societies, see Donald Pierson, Negroes in Brazil: A Study of Race Contact at Bahia (Chicago: Univ. of Chicago Press, 1942), 54–59. On abolitionism in Brazil, see Conrad, Destruction of Brazilian Slavery, 121–276; and Robert Brent Toplin, The Abolition of Slavery in Brazil (New York: Atheneum, 1972).


APB, LRTC, no. 10, fols. 91–96B; no. 29, fols. 103–7; no. 31, fols. 87–89B; no. 40, fols. 29B-32, no. 52, fols. 145B–149B; no. 61, fols. 168B–170.


See, for example, Compromisso da Irmandade da Nossa Senhora dos Homens Prêtos no ano de 1820, Arquivo da Igreja da Nossa Senhora dos Homens Prêtos, Salvador; Estatutos da Sociedade Protetora dos Desvalidos, titulo 1 and art. 9, published by Julio Santana Braga in his Sociedade Protetora dos Desvalidos: uma irmandade de côr (Salvador: Ianamá, 1987), 79–89.


APB, LRTC, no. 23, fols. 186–89.


Those identified as forros for 1808–1884 included 11 boys and 16 girls.


Manoel, crioulo, son of the female slave, APB, LNC, no. 164, Jan. 25, 1811.


APB, LNC and maço 2880. A Brazilian-born slave child, especially an infant or very young child, was often purchased or sold with its mother. Among 2,608 legal registers of slaves sold and purchased in Sé parish for 1838–1888, 131 (5.0 percent) could be classified as mother-and-child sales. In only four cases, all dating from the 1860s, fathers and children, all sons, were purchased and sold together. See AMCS, LECV.


APB, LNC, no. 116, Jan. 25, 1811. In Sé parish for 1838–1888, only one Brazilian-born slave family was transferred without loss or separation of any member to different ownership. AMCS, LECV, no. 82.6, Dec. 7, 1867.


APB, LNC, no. 272, Nov. 17, 1841; no. 275, Oct. 30, 1841.


Cases of self-purchase in cash include three Brazilian-born slaves freed in 1871 on the condition that they make monthly payments: Romana, 29-year-old crioula, daughter of Virginia, Angola, for six milréis a month (APB, LNC, no. 411, Aug. 8, 1871); Martinho, crioulo, for ten milréis a month (ibid., Aug. 12, 1871); and Marcelino, 26-year-old crioulo, son of the Angolan slave woman, for ten milréis a month (ibid., no. 416, Aug. 21, 1871). The manumission letters consulted for this study do not include any case in which the price for freedom was paid in silver dobra or any other type of currency.


APB, LRTC, no. 13, fols. 8B–15. In the data for Paraíba (1850–1888), some slaves purchased freedom not only with real estate but also with cattle. See Galliza, O declínio da escravidão, 150.


The practice of substitution took place in Bahia (1684–1745), the city of Rio de Janeiro (1808–1850), and the town of Sahará (1789–1822). See Schwartz, “Manumission,” 626; Karasch, Slave Life, 358; and Higgins, “Slave Society,” 247–48. Mattoso refers to substitution but does not provide any interpretation: “Slaves had to pay for their freedom in hard cash, metal coin, or paper money, either in a lump sum or by installments. Some slaves redeemed themselves by giving another slave to masters.” See Kátia M. de Queirós Mattoso, To Be a Slave in Brazil, 1550–1888, trans. Arthur Goldhammer (New Brunswick: Rutgers Univ. Press, 1986), 158. To my knowledge, substitution as a form of self-purchase in the New World was unique to Brazil. Among many studies of manumissions, the only example of substitution to be found concerned a Mandingo slave in Port-of-Spain, Trinidad, in 1830. See Carl Campbell, “John Mohammed Bath and the Free Mandingos in Trinidad: The Question of Their Repatriation to Africa, 1831-38,” Journal of African Studies 2:4 (Winter 1975), 467–95, esp. 472. The origin of substitution as self-purchase dates back to imperial Rome. A special feature of some slaves’ situation was a fund called the peculium, which the slaves themselves controlled. The slaves could use it for investments or ultimately for self-purchase; they could also use it to purchase other slaves to stand in for them (vicarius, vicaria). The peculium was usually reserved for urban, skilled slaves, not rural workers. See Phillips, Slavery from Roman Times, 28.


APB, LNC, no. 160, July 15, 1809. The same settlement: Ibid., no. 242, Apr. 10, 1832; no. 304, May 7, 1852.


In the wills of freedpersons, seven ex-slaves, all African-born, declared they had provided their owners with substitute slaves, but only one was identified in terms of nation. See APB, LRTC, no. 4, fols. 5B–9B; no. 9, fols. 101B–106; no. 16, fols. 156–61; no. 17, fols. 145B–149B; no. 28, fols. 98B–101 and 122–25B; no. 47, fols. 28–30B.


Maria and Thereza, both Nagô, in place of Maria, Nagô (APB, LNC, no. 206, May 31, 1822); two new female slaves, names and origins unspecified, in place of Anna, Nagô (no. 206, June 8, 1822).


All four letters are in APB, LNC: bricklayer, no. 160, July 15, 1809; chair carrier, no. 242, Apr. 10, 1832; fieldworker, no. 295, Feb. 3, 1851; stevedore, no. 306, May 14, 1852.


Joaquim replaced by another Joaquim, APB, LNC, no. 272, Mar. 9, 1841; Caetana, crioula, replaced by another Caetana, Nagô, in the same occupation, fieldwork, ibid., no. 295, Feb. 3, 1851.


APB, LNC, no. 293, May 22, 1821; ibid., no. 238, Feb. 16, 1832.


Instituto Geográfico e Histórico da Bahia, Salvador, pasta 28, doc. 11, cited by Schwartz, “Manumission,” 626–27, n. 36.


On African-born freedpersons in the transatlantic slave trade, see APB, LRTC, no. 19, fols. 276B–280B; no. 28, fols. 98B–101B. During his visit to Rio de Janeiro in 1829, the British traveler Robert Walsh observed: “negroes themselves who had obtained their freedom frequently sent ventures to Africa to purchase their countrymen, who were brought back to them in exchange for the beads and looking-glasses which they sent out.” Robert Walsh, Notices of Brazil in 1828 and 1829, 2 vols. (London: Fredrick Westley and A. H. Davis, 1830), 2:362.


Miller, “Slave Prices,” 45 and 47; Mattoso, Klein, and Engerman, “Trends and Patterns,” 66, n. 9.


Schwartz, “Manumission,” 626: Karasch, Slave Life, 358; Higgins, “Slave Society,” 247.


Karasch, Slave Life, 358.




On crias as heirs see, for example, APB, LRTC, no. 43, fols. 50–52. On the legal arrangement of cortação, see Schwartz, “Manumission,” 627–28. To compare the custom of coartación in Spanish America, see Hubert H. S. Aimes, “Coartación: A Spanish Institution for the Advancement of Slaves into Freedom,” Yale Review 17 (1909), 412–31; and Herbert S. Klein, Slavery in the Americas: A Comparative Study of Virginia and Cuba (Chicago: Univ. of Chicago Press, 1967), 196–200.


Karasch, Slave Life, 211.


Kiernan, “Manumission ... Paraty,” esp. 148; and Galliza, O declínio da escravidão, chap. 4. Judith Lee Allen, who conducted her archival research in Salvador and Cachoeira from 1987 to 1989, encountered only four cases of substitution in her sample of approximately 1,700 nineteenth-century manumission letters registered in the towns of the Bahian Recôncavo. Allen, personal communication.


Koster, Travels in Brazil, 2:266–67.