Small and Medium Slaveholdings in the Coffee Economy of the Vale do Paraiíba, Province of Saão Paulo

Abstract

Worldwide growth in demand for coffee during the nineteenth century led to increased production in many countries. Brazil took on a heightened role in this market, becoming a major exporter after independence in 1822. Whereas in the 1830s coffee represented 43.8 percent of Brazil’s total exports, by the 1870s it was more than half.1 During the Empire (1822–89), coffee plantations occupied large expanses of Brazil’s territory. The largest concentration of production was in the Vale do Paraíba region, located on the border between the provinces of São Paulo, Rio de Janeiro, and Minas Gerais, where soil conditions, climate, altitude, and the proximity of export ports were favorable. Classic Brazilian historians such as Gilberto Freyre, Caio Prado Júnior, and Celso Furtado felt that large slaveholding monoculture plantations, whose production was destined for the external market, determined the basis of production in both the colonial and imperial periods.2 The slave population was concentrated in the hands of a few large masters, and the Brazilian population was understood to be comprised almost exclusively of slave owners and slaves. The greatest exponents of this view asserted that the coffee economy was based on the use of large-scale slave labor in the production of goods for export. The work of Stanley Stein and Emília Viotti da Costa followed the general proposition that most production was concentrated in export-oriented latifundia with large numbers of slaves, usually around a hundred people.3A revisionist view has taken hold in recent historiography. First of all, studies of coffee production in Colombia, Costa Rica, Guatemala, El Salvador, and Puerto Rico point to labor arrangements very different from those in Brazil, such as the use of free labor.4 Steven C. Topik provides a more complex picture of coffee culture in the Americas: in many areas the latifundia was not the norm, many smallholders were able to retain or obtain land, and there were no economies of scale in the production of coffee.5In Brazil, a broad debate has taken place over the last three decades with regard to the traditional interpretation.6 New works based in local and regional manuscript sources do not entirely corroborate the general view that coffee production was largely in the hands of latifundistas. The growing incorporation of new sources allows us to qualify or even revise some of the consensual views of classic texts.7 Still, the revisitionist interpretations have not yet been consolidated, and the debate about the importance of slavery in coffee plantation agriculture continues. Nevertheless, a few pioneering studies have incorporated these documents and provide new evidence.Nominal lists (manuscript censuses) from the beginning of the nineteenth century have provided the empirical foundation for many studies of the slaveholding structure for the Vale do Paraíba in the province of São Paulo. Lucila Hermann’s 1945 thesis, the first such study to make use of these sources, pointed to greater complexity in the coffee economy, including the presence of small coffee producers, in the municipality of Guaratinguetá.8 Later, Maria Luiza Marcílio verified nonslaveholders in coffee production, and José Flávio Motta analyzed slave property in Bananal.9 Motta’s study begins in 1801, at which point non-slaveholders planted coffee. But by 1829, when coffee became well established in the region, slaveholders came to dominate cultivation. Coffee planters held an average of 13.8 slaves (compared with an average of 11.7 slaves for all slaveholders).10 In Taubaté, Armênio de Souza Rangel found an average of 7.8 slaves per coffee planter (compared with 5.1 for all slaveholders) in 1835, at a time when coffee cultivation was less extensive than in Bananal.11 Francisco Vidal Luna’s study of the municipality of Areias found an average of 12.6 slaves per slaveholder in 1836.12 My own research established an average of 6.3 slaves per slaveholder in Lorena in 1829, which at the time was less involved in coffee than Bananal and Areias.13 The average slaveholding of coffee producers in Lorena was 10.2 in 1829. A sample of 187 inventories for 1830–79 yielded an average of 22.5 slaves per coffee-producing unit. Thus, coffee planting in its early phases (up until the fourth decade of the nineteenth century) appears to have been organized in a slightly differently manner from that described by the classic historiographers: coffee growers held an average of about 10 slaves, a number similar to that of the slaveholding population as a whole. Some authors disagree with this statement. Rangel, for example, states, “[T]he municipality, in many aspects of its economic and social organization, came close to the model of the exporting plantation formulated by Caio Prado Júnior.”14When considering the Vale do Paraíba overall, we should pay attention to local variations in the economic structure. Some places show a greater proportion of large slaveholders than others: for example, Bananal and Areias in São Paulo and Vassouras in Rio de Janeiro.15 Researchers have demonstrated the importance of small and medium producers in other locations, even during the peak period of coffee cultivation. Maria Celina Whately’s study of nineteenth-century coffee production in Resende, Rio de Janeiro province, found a greater proportion of smallholders in comparison with Vassouras.16 Neither does this location seem unique. Maria Aparecida Chaves Ribeiro Papali (working with 30 inventories from São José dos Campos, São Paulo, from 1870–88) likewise observed the presence of numerous small and medium growers. Most farmers had, on average, 10–15 slave workers.17Aldeci Silva dos Santos observed, based on a sample of 553 inventories, the presence of small properties even in Vassouras: slaveholders with ten slaves or fewer comprised 33.5 percent of inventoried decedents between 1830 and 1879.18 The author went so far as to say that “the success of the coffee economy in the Paraíba Valley [of Rio de Janeiro] owes to the aid of the small property; when it could no longer reserve sufficient space in its landholdings for subsistence production, it came to depend on external supply [from smallholders] to meet its needs.”19 João Fragoso examined almost one hundred inventories from localities close to Paraíba do Sul and found an average of 18.8 slaves per owner for 1830–50 and 20.7 for 1870–75. Célia Maria Loureiro Muniz studied the agrarian property structure of the Vale do Paraíba in Rio de Janeiro during the nineteenth century and found that coffee plantations averaged between 100 and 200 alqueires (roughly 500–1,000 hectares), and many had fewer than 100 alqueires.20 Based on the forms of land appropriation, she showed the size of the properties: “Many books on the coffee economy speak of ‘large properties’ and ‘enormous latifundia.’ However, these coffee plantations were not so large; they had an average of 100 to 200 alqueires and many had fewer than 100 alqueires.”21Did the picture drawn for the valley at the beginning of the nineteenth century remain stable for the following period? The consolidation of the coffee plantation, according to Caio Prado Júnior’s model, occurred in a later period, for which there exist no manuscript censuses. Iraci del Nero da Costa seems to capture the varied reality of coffee production, stating that “small, medium, and large owners coexisted in the most varied sectors and quadrants.”22These reassessments affirm the need to incorporate the small and medium owner into our analyses of the colonial and imperial economies. In other words, we should not limit ourselves to studying the great, all-powerful slaveholders. Other recent research on the economic and demographic importance of non-slaveholders reinforces this point.23 In this regard, we would expect the slave-holding structure of the Vale do Paraíba after 1840 (the peak of coffee production in the region) to resemble the profile outlined in traditional historiography, but we should still pay attention to local variations within the valley.This article focuses on the importance of small and medium slaveholders to the coffee economy in the Vale do Paraíba. My results corroborate those found in the revisionist literature discussed above, including Herrmann, Whately, Papali, Santos, and Muniz. My data suggest the presence of a significant number of small and medium slaveholders in several different parts of the Vale do Paraíba—even during the height of the coffee boom. These owners coexisted with a reduced number of large plantations. Nevertheless, the proportion of large slaveholders varies at different locations. I will also describe the slave population of Vale do Paraíba, relating its attributes to the size of slaveholdings.In this comparative analysis, I make use of the 1872 census, the registers (matrículas) for 1871–72, and Emancipation Fund classification lists that have been unearthed for the municipalities of Bananal (1874), Lorena and Cruzeiro (1874), Paraibuna (1874), São José dos Campos (1874), São Luís do Paraitinga (1874), and Taubaté (1872).24 The Emancipation Fund dated from November 13, 1872 and was created based on the Lei do Ventre Livre (the Free Womb Law), issued September 28, 1871. The law declared children born to slave mothers to be free persons. The regulation mandated that the fund’s resources be allocated first for the freeing of slaves with certain categories of relatives, and later for others. Family preference was shown in the following order: (1) married couples owned by different owners; (2) couples who had children under the age of eight that were born free by virtue of the Free Womb Law; (3) couples who had freeborn children under the age of 21; (4) couples with minor slave children; (5) mothers with minor slave children; and (6) couples with no minor children. The remaining slaves would be shown preference in the following order: (1) mothers or fathers of free children not included in the categories above; and (2) any remaining slaves between the ages of 12 and 50, beginning with the youngest females and the oldest males (see Collecção das Leis do Império do Brasil de 1872, 1873, vol. 2, p. 1059). Not all slaves fit into these categories; for example, individuals over 50 with no relatives would not be included in the classification lists. Slaves younger than 12 or older than 50 were most likely to be omitted.I will compare demographic information on the slave population—sex, age group, and family attributes—in order to determine the representativeness of my sample. I then present a more detailed analysis of the slave cohorts named in the classification lists. Finally, I analyze the slave property in these locations and compare it with the results of other research for the beginning of the nineteenth century. There appears to be a relationship between property size and certain demographic characteristics of the slaves. Thus, we have been able to compare the reality of slavery emerging from the nominal lists of inhabitants for the first three decades to the situation in 1870 and determine whether there were changes in the pattern of slave ownership during the nineteenth century.The available evidence does not permit detailed analysis of the economic activity of all the slaveholders residing in the locations studied. However, the regional economy might be characterized as a coffee economy. That is, almost all economic activities were related, directly or indirectly, to coffee production. Many people found themselves directly linked to coffee production, while others produced goods and services for the coffee sector, even in the urban areas. As Sergio Silva notes, “The coffee economy and capital extended far beyond the plantations.”25 Several studies point to the interdependent growth of coffee production and subsistence production. Aldeci Silva dos Santos remarks, “The small proprietor is quite engaged, even if indirectly, in the coffee economy—often not as a direct producer but with the important role of supplying large producers.”26 Finally, the dividing line between coffee producers and other growers is unclear, especially for smallholders. Some authors refer to these plantations as mixed.The expansion of coffee cultivation, which reached its height in the Vale do Paraíba in the 1870s, was conditioned by and conditioned itself to demographic growth, especially of slaves. In the Vale do Paraíba, 27.9 percent of the population was enslaved. Although this percentage is slightly smaller than that found in the central region of São Paulo (36.1 percent), it dwarfed the percentage of slaves in the provincial capital, the Vale do Ribeira, and the coast.27The six locations selected for our study account for 46.0 percent of the slaves on the registers for the São Paulo portion of the Vale do Paraíba. In the 1872 census, counting both the slave and free populations, Taubaté was the most populous municipality of our six locations, with 20,000 inhabitants. The rest maintained a total population of about 15,000 citizens. Bananal’s slave population of 8,281 is the largest in the sample, while Taubaté had only half that number. According to the census, slaves were a significant percentage of the total population. The percentage of Bananal’s population that was enslaved, 53.1 percent, was higher than anywhere else in the province of São Paulo. The proportion in Paraibuna and São José dos Campos was a little over 9 percent, despite their emphasis on coffee production in the second half of the nineteenth century. Perhaps the lower percentages are due to the greater presence of free Brazilians, since there were no large numbers of foreign immigrants at the time. In Lorena and Cruzeiro, São Luís de Paraitinga, and Taubaté, slaves accounted for 14.8, 15.0, and 19.8 percent of the total population, respectively—still much lower than the percentage in Bananal.28 Thus, the relatively greater presence of slaves in the municipalities correlated with the extent of coffee production, with the exception of São Luís do Paraitinga.29Although the transatlantic slave trade ended definitively in 1850, the slave population of these municipalities grew around 88 percent between 1836 and 1872, with the highest rate of growth in Bananal (138 percent) and São José dos Campos (211 percent). Increased internal slave trade—largely originating in the Brazilian Northeast and South—and natural reproduction help to account for this increase, in the absence of trafficking from Africa.30Various reports on the slave population in the first half of the 1870s reveal differences in their totals and their composition. Table 1 presents information from the three available sources on slaves in each location. The 1872 census showed a slave population smaller than that recorded in the registers of 1871–72 for five of the six municipalities analyzed. The exception was Bananal. The biggest differences occurred in Lorena and Cruzeiro and Paraibuna.31 The classification lists I located report 19,076 slaves—87.5 percent of the slave population recorded in the registers.32Bananal, São Luís do Paraitinga, and Taubaté showed a closer match between the slave population listed in the registers and that in the classification lists. In Paraibuna, São José dos Campos, and Lorena and Cruzeiro, the classification list totals were about three-quarters of population noted in the registers. Taken together, the samples can be considered sufficiently representative of the total number of persons who inhabited the localities.We surveyed the slave classification lists for Bananal, Lorena and Cruzeiro, Paraibuna, São José dos Campos, São Luís do Paraitinga, and Taubaté. These documents contained the register number, name of the slave (from which we inferred sex), color, age, marital status, aptitude for work, profession, value, observation, and a field called “family members” that mentioned parents, children, or conjugal partners. Other information was included for some localities but not for others. For example, the prices of slaves were recorded only in Lorena.33 The field of “family members” was completed only in Lorena and São José dos Campos.34 The lists note the names of slave owners, and we were thus able to reconstruct the holdings of individual owners whose slaves were dispersed throughout the various categories.It appears that some slaves (with the same register number, owner, and similar ages) were registered in two or more documents. For example, one slave was registered on one list as the father of a freeborn child and on another list in the category of individuals between 12 and 50 years old. The rate of such duplications varied from place to place. While we found no duplicate entries in Bananal and Paraibuna and only three in Taubaté, in São José dos Campos we found (and eliminated) 181 duplicate listings, 138 in Lorena and Cruzeiro, and 50 in São Luís do Paraitinga. In some cases, we were able to establish family ties between slaves listed by means of the field “family members,” especially relationships between siblings.We were able to supplement information on some owners by consulting the Almanak da província de São Paulo para 1873.35 Cross-checking the information on owners in the Almanak with that from the classification lists allowed us to record the economic activity of the cross-listed slaveholders. However, only a small number of the slave owners named in the classification lists appeared in the Almanak, and many individuals listed in the Almanak did not appear on the classification lists—especially artists, artisans, professionals, and merchants, who usually did not own slaves.Our sample analyzes a significant proportion of all slaves present in the Vale do Paraíba at the beginning of the 1870s. However, slavery was not uniformly distributed throughout the region. The areas most involved in coffee production generally had a larger proportion of slaves in relation to the free population. The uneven distribution of the slave population seems to constitute only the most easily perceived difference between municipalities. We needed more detailed information on the slaves and their owners to carry out an in-depth analysis of the slave population; we found this information in the classification lists. The next section uses information from these lists to offer a better understanding the demographics of slavery in the region. We should remember that although the lists present a wide range of information on a large number of slaves, they do not include the entire slave population. In particular, the slaves included in the classification lists show a demographic profile somewhat different from that observed for the population recorded in other sources.As expected, the majority of slaves were males, due to the preponderance of male captives in the transatlantic traffic prior to 1850. Taubaté and Bananal recorded the greatest gender imbalance. Age range is another important factor in explaining the sex ratio among slaves. Table 2 shows a significant increase in the male-to-female ratio among older slaves for the locations in question. This result is not due to a difference in mortality rates. Instead, we should probably attribute this increase to the higher proportion of African-born slaves in the older cohorts, especially 50 or older, due to the preponderance of male slaves in the transatlantic trade.Individuals between 15 and 49 years old predominated in the sample— about two-thirds of the total slave population (see table 3).36 In São José dos Campos, the prevalence of this age range was even greater, at 73.4 percent.37 São Luís do Paraitinga exhibited a lower number of slaves 50 years old or older, just 5.8 percent of the total. On the other hand, slaves younger than 15 were a little more than a quarter of the total number of slaves in Bananal, Lorena and Cruzeiro, Paraibuna, São Luís do Paraitinga, and Taubaté.Because the Free Womb Law that gave rise to the classification lists gave preference to slaves with certain family ties and did not cover other categories of slaves (such as those over 50 without relatives), the slaves classified maintained a significant network of family ties among them (see table 4). In São Luís do Paraitinga, only 44 percent of slaves had family ties, while in Bananal it reached 54.5 percent, in Lorena and Cruzeiro 54.6 percent, and in São José dos Campos 61.5 percent.38 The largest percentage of slaves with family ties occurred in Paraibuna (63.8 percent). Perhaps this difference results from a lower degree of representativeness of classified slaves in the total on the registers of these last three localities (see table 1) and the impossibility of establishing the relationships between siblings for the first two locations. In Bananal, the proportion of men in the 15-and-older group was higher, making the constitution of families more difficult.39We should also point out that kinship ties between slaves and nonslaves were on the rise, due to changes in the law and the spread of the abolition movement. We identified 43 ingenuos (children born of slave mothers after the Free Womb Law) in Lorena and Cruzeiro, with 63 in Paraibuna and 87 in São José dos Campos.40 Although the number of ingenuos in Bananal is uncertain, our survey found 467 free persons born of slave parents in this municipality. Of these, probably 185 were ingenuos. In addition, there were 22 free couples in this city and 43 free parents.41 In São Luís do Paraitinga, we found 25 ingenuos and 5 married couples that included a free woman. Thus, we can say that a signifi-cant number of slaves maintained family relationships with free individuals.42We can also classify the slave occupations, a list that includes a rich variety of artisan activities, commercial activities, and domestic work, in addition to traditional agriculture. Nonetheless, about two-thirds of all slaves were employed in agriculture in Bananal, Lorena and Cruzeiro, Paraibuna, and São José dos Campos.43 In São Luís do Paraitinga, agriculture was the profession of a slightly smaller number (66.5 percent). This is a result of the lower numbers of women engaged in agriculture (38.2 percent of all female slaves), a smaller percentage than that of cooks (40.1 percent).In addition to the preponderance of agricultural activities, we also note an occupational specialization according to gender. Some professions were predominantly masculine—such as tailor, animal packer, basket maker, wrangler, carpenter, cart driver, delivery man, coachman, butler, foreman, blacksmith, fireworks maker, ant exterminator, vegetable producer, day laborer, miller, musician, child caretaker, farmhand, mason, shoemaker, sawyer, tile maker, and animal driver. Agriculture was also dominated by men, even though it was also listed as the occupation of a significant number of female slaves—from one-third to two-thirds of the total, depending on the location. On the other hand, certain professions were predominately female: caretaker, seamstress, cook, candy maker, starcher, spinner, laundress, chambermaid, midwife, lace maker, domestic worker, and weaver. In addition, some professions stood out as requiring greater skill or training—for example, carpenter, musician, mason, shoemaker, and animal driver.44Treating slaves as property permits an analysis of the amount and distribution of wealth in each location. The slave population can be considered (within limits) as a proxy for the fortune of individuals and even of the region, especially for the time periods considered here.45 On the other hand, uneven distribution of slaveholdings can best be understood if compared over time and space. We attempted to contrast our data with that from other studies for other locations in this period (mostly those based on similar sources) and with research on these exact municipalities at other points in time based on nominal lists of inhabitants.Table 5 shows some statistical indicators based on the classification lists. Analysis of the slaveholding structure reveals very different behavior from place to place. The average slaveholding in the 1870s was more than 15 in Bananal, while in São José dos Campos and São Luís do Paraitinga the figures were closer to 5.46 In Lorena and Cruzeiro, Paraibuna, and Taubaté, the average slaveholding was a bit higher than 6 slaves. These last three localities were thus somewhat intermediate in their average slaveholding, although they were closer to the smaller municipalities than they were to Bananal. In none of the other municipalities does the development of coffee and of the slave population seem to have come close to the size of Bananal. The average for the entire six municipalities was 8, with a Gini index of 0.689.The distribution of slaves also followed the pattern of average number of slaves, whether viewed as a coefficient of variation or through the Gini index. Using the Gini index, the most uneven distribution is found in Bananal (0.759), and the most balanced is found in São José dos Campos (0.570).47 This result corroborates existing regional differences observed at the beginning of the nineteenth century discussed in the introduction of this article; these regional differences do not change significantly in the second half of the century. The other localities have intermediate indexes: 0.611 for São Luís do Paraitinga, 0.624 for Lorena and Cruzeiro, 0.638 for Taubaté, and 0.649 for Paraibuna.We compared the average and Gini indexes we calculated with those reported by other researchers for the same localities at the start of the nineteenth century, based on the nominal lists of inhabitants of the time.48Tables 6 and 7 indicate average slaveholdings and the Gini indexes for the earlier time period. All the data we found indicate larger average slaveholdings and a higher degree of inequality in Bananal than in the other locations in the nineteenth century. Lorena and Cruzeiro and Taubaté maintained Gini indexes close to each other, except for the period 1829/30. Still, when we compare the results from 1872/74 with those from the beginning of the nineteenth century, we do note some changes. Bananal, Lorena and Cruzeiro, Paraibuna, São José dos Campos, and Taubaté show higher average slaveholdings for the period 1829/35–1872/74, mainly for the first. Finally, São Luís do Paraitinga followed a unique course of development: both average slaveholdings and inequality in distribution decreasing during this period.49If the trajectory of coffee produced identical patterns in the different locations in the valley, we would expect the results obtained for Bananal to have occurred in the other localities, since coffee was introduced across the frontier with Rio de Janeiro. In this context, the differences between the municipalities should be only a question of time. All the municipalities would have approached Bananal’s figures for average slaveholding and unequal distribution for a determined time. However, since the differences in these indicators are steady over time, we must explain the particularities of each locality.Considering the situation by grouping the size of slaveholdings in the distribution of slave property allows us a better vision of the similarities and differences among localities. The total number of slaveholders and slaves increased for all the locations between 1829/35 and 1872/74.50 Over the course of the nineteenth century, we see growing numbers of slaveholders with small slave-holdings in all municipalities, especially in Bananal (see table 8). Despite this increase, the share of the total slave population owned by smallholders decreased in all localities except Lorena and Cruzeiro. On the other hand, the holdings of large slaveholders increased, and their slaves represented a greater proportion of the entire slave population.50 This corroborates the result obtained via the Gini index.There were some differences among the localities under consideration. In 1874, 61.5 percent of slaveholders in Bananal owned 1–5 slaves, while 70.7 percent of slaveholders in Lorena owned 1–5 slaves. The slaves this group owned represented 8.1 percent and 23.1 percent of all slaves, respectively. Thus, Bananal had both the smallest proportion of smallholders and the smallest proportion of slaves owned by smallholders. These indicators were similar in Lorena and Cruzeiro, Paraibuna, São José dos Campos, São Luís do Paraitinga, and Taubaté.The professions of almost two-thirds of the slave population, as mentioned earlier, were in agriculture. By cross-referencing the slaveholders who registered their slaves on the list with those individuals reported in the Almanak da Província de São Paulo para 1873, we were able to establish the activities of some of the owners on the list.52 For the owners where we were able to obtain information on profession, agricultural activities predominated: these represent 87.6 percent of the slaveholders in Bananal, 77.6 percent in São José dos Campos, 71.7 percent in Paraibuna, 68.7 percent in Lorena and Cruzeiro, 67.7 percent in Taubaté, and 60.6 percent in São Luís do Paraitinga. According to this data, agriculture appears to have been more important in the localities of São José dos Campos and especially in Bananal. Agriculturalist owners had larger slave-holdings, on average, than did merchants, artists, and professionals (see table 9). In São Luís do Paraitinga, 67.2 percent of the slave population was owned by agriculturalists engaged in coffee production; 79.4 percent in Lorena and Cruzeiro, 78.2 percent in Bananal, 86.7 percent in Taubaté, and 96.3 percent in São José dos Campos. “Capitalists” (moneylenders) had average slaveholdings similar to those of agricultural producers and sugar millers. Finally, as expected, Bananal’s coffee growers had the largest average slaveholdings—around three times larger than the second-largest slaveholders.The distribution of slave property among coffee growers varied, since some growers owned just one slave, while other plantation owners possessed hundreds. Bananal can be contrasted with the other localities analyzed. The Alma-nak lists 40 slaveholding coffee growers in São José dos Campos; of these, 57.5 percent had fewer than 10 slaves.53 In contrast, of 57 slaveholding coffee growers in Bananal, only 9 (15.8 percent) had fewer than 10 slaves. Another data point on coffee producers in Bananal can be obtained from a 1868 list of coffee plantations, compiled when a railroad was projected from the Rio de Janeiro coast to the Paraiba Valley.54 Crossing that classification with our other data indicates 68 coffee growers who produced 310,000 arrobas of coffee and owned 4,029 slaves (53.5 percent of the total slave population). The 16 growers (23.5 percent) who owned less than 10 slaves harvested 34,500 arrobas of coffee, a little more than a tenth of the total harvest. Thus, the small slaveholders remained directly involved with coffee production at the beginning of the 1870s, ranging from around one-tenth to one-half of the coffee producing slaveholders.55Despite certain differences at the beginning of coffee development, Taubaté, Paraibuna, and Lorena converged on very similar patterns of average slave-holdings and unequal distribution by the 1870s. Our data do not indicate great differences between them, aside from the total number of slaves. On the other hand, the size of slaveholdings and degree of unequal distribution developed more slowly in São José dos Campos and remained lower than the other municipalities in the 1870s. Bananal most closely approximated the traditional model. It displays the largest contingent of slaves and the largest and most concentrated slaveholdings. Nevertheless, even Bananal witnessed a significant number of owners with fewer than five slaves—about half of all holdings. Thus, although we do not have reports for all the localities in the Vale do Paraíba, we can see that the diffusion and distribution of slave property was remarkably heterogeneous there during its golden period of coffee cultivation.In contrast to the image gleaned from classical historiography, the data I have presented clearly demonstrate the importance of small and medium slaveholders during the height of coffee production in most of Vale do Paraíba of São Paulo. In the 1870s, such small and medium slaveholders accounted for more than nine-tenths of owners and owned more than half of the slave population of the areas studied.We do see significant differences in patterns of slaveholding between the six locations studied, but these did not change over the course of the nineteenth century. Of the 18 slaveholders who owned more than one hundred slaves, 17 resided in Bananal. Bananal thus comes closest to the model proposed in the earlier literature. Patterns of coffee production and slaveholding in the Vale do Paraíba of São Paulo should, however, not be deduced from those found in Bananal, which was exceptional in our sample, but from the totality of municipalities. And even in Bananal we found a significant number of small and medium slaveholders, including many owners with only one slave. Such small and medium slaveholders were not only involved in production for the internal market but also produced coffee in significant quantities.56 Thus, the economic and demographic dynamism of the Vale do Paraíba was not restricted just to the members of the large coffee plantations.Our data indicate that slaveholdings, like other property, became more concentrated as coffee production expanded during the nineteenth century. In this process, many smallholders could have been excluded. Still, the process also allowed many people of modest resources to remain in the region and benefit from the coffee economy. Even at the point of greatest inequality in property distribution, small and medium holders still were very important to the economy. The increased concentration that accompanied the development of the coffee economy did not exclude small and medium slaveholders from production, and it is wrong to ignore them. On the contrary, the economic diversification that accompanied the expansion of coffee created opportunities for less fortunate people both in coffee cultivation and also in other occupations, agricultural or otherwise.Coffee production generated monetary income for small and medium owners through export agriculture, subsistence agriculture, or even in nonag-ricultural occupations. They exercised their purchasing power even in relation to slave labor. Our results point to the importance of the small and medium property holders in the coffee economy. Scholars have noted similar patterns in other export-producing regions of Latin America (including tobacco and cotton): even where we would expect the predominance of large plantations, researchers have found that smallholders participated in export production.57 A significant portion of Brazil’s free population owned some wealth in the form of land and slaves, which strengthened the internal Brazilian market. Analyzing the distribution of slaves in the Vale do Paraíba of São Paulo offers some insight into the overall distribution of wealth and income, a factor in fostering the development of an important internal market.