1. For a broader assessment of "legal" skin color differences also see Lanning (1944) and Moerner (1967). [BACK]

2. On the implications of these ideas in other areas see Moerner (1967) and Toplin (1981). [BACK]

3. Bowser comments on the frequency of marriage in the sixteenth and seventeenth centuries. Our sources may not be comparable, but qualitative sources at least show that marriage in the nineteenth century was an important and ever-present argument for slaves. [BACK]

4. In an analogous interpretation for the United States, Fields (1985, 29-30) writes, "Much more than simple demography worked to keep free blacks in close relationship with slaves. The vagaries of manumission did so as well. Delayed manumission was a widespread practice, embracing just over half of manumissions between 1832 and just under half thereafter. That made for awkward and anomalous family situations. It could result in a famfly composed indiscriminately of slaves for life, slaves for a term of years, and free people." [BACK]

5. For a similar assertion in the Caribbean, see Morrisey (1989, 4). [BACK]

6. By contrast, a slave in the United States who hoped to manumit a relative usually had to obtain and secure property rights in his or her own name; only a freed person could purchase slaves (Genovese 1985). This difference may well explain very different rates of self-manumission in each hemisphere. [BACK]

7. For example, Schwartz (1982:67) found for the Reconcavo in Brazil that "many of the traditional artisan skills were practiced by the slaves. Engenhos often found it more profitable to train slaves as carpenters, smiths, or coopers rather than pay for the services of free artisans"; the argument suggests an economic rationale for the presence of slave artisans as well as an aversion to manual labor among whites and mestizos . [BACK]

8. Such a paradoxical relationship also occurs in Cuba and Brazil. In Brazil, where the hiring-out system was very prevalent, Mattosso (1986, 123) states, the "possibility of shifting large numbers of slaves from one occupation to another helped to stabilize a market in which demand varied with circumstances and competition.... The system was, thus, highly flexible." In Lima owners hoping to shift workers from urban to rural markets often met resistance from their slaves. [BACK]

9. Contrast the situation for Brazil, where Mattosso states "that it would have been difficult for the slave to save much of what he earned" ( 1986, 123). [BACK]

10. In the United States, the hiring-out system was forbidden even by law; nevertheless, it was continued by custom and tradition. "In the towns and especially in the larger cities, many slaves 'hired their own time' and lived away from the masters. Although these practices were generally illegal, they were sanctioned almost everywhere, by custom and in accordance with white business interests. Consequently, free Negroes, whose freedom always was precarious, interacted every day, socially and at work, with slaves who were close to being half-free. A certain amount of intermarriage occurred, and little in the social setting generated antipathy" (Genovese 1974, 406-407). [BACK]

11. 375 ÷ 3 = 125; 125 × 8 = 2,250 and 125 × 9 = 1,125. Decrease of the slave population between 1818 and 1836 = 2,798. 2,798:2,250 = 1.24. Similarly, for the second period (1836-1845), the slave population decreased 1,291. 1,219:1125 = 1.14. The simple average of the number of slaves for the three years for which we have censuses is 7,190 between 1818 and 1836 and 5,145 between 1836 and 1845; thus: 2,250:7,190 = 31.3 percent; and 1,125:5,145 = 21.9 percent. [BACK]

12. For the same period in the United States Hart (1968 [1906], 130 ff.) calculates that in 1850 one out of 2,181 slaves became free; in 1860 it was one out of 1,309. [BACK]

13. Other slave societies did not share this long-term tendency. In the United States, in the decades between 1820 and 1840 there were higher lev-

els of manumission than between 1850 and 1860. Between 1830 and 1860, "one state after another closed off manumissions altogether and insisted on the removal of freedmen from the state. In the 1850s the position of a free Negro in New Orleans and some other cities rapidly deteriorated through keeping coffeehouses or entering special fields of employment" (Genovese 1974, 399). Until the beginning of the eighteenth century, colonial French slaveholders could do as they pleased, including granting slaves freedom. But thereafter they had to seek special state permission to liberate a slave. In 1775 manumission was taxed, and manumitting a slave woman was doubly taxed (Wirz 1984, 126). Thus, whereas in Lima the process of manumission accelerated abolition, in other areas legal and moral devices hindered abolition. [BACK]

14. Following Patterson's scheme of analysis (1982), slave formations differ by the nature of their dependence on slavery and by the degree and direction of dependence—or, more simply, by who depended on slavery and what were the conditions and consequences of the relation. My analysis of the distribution of slave ownership shows that slaves occasionally reversed the terms of dependence. [BACK]

15. Also in Brazil, "hundreds and hundreds of families have one or two slaves on whose earnings alone they live" (Schwartz 1982, 68, based on research done by Ewbank in 1856). Schwartz goes further: "[about] one-third of the households in the urban centers of São Paulo and Ouro Prêto, for example, contained at least one slave. In São Paulo the percentage decreased between 1778 and 1836, but, even at the later date, 46 percent of the free households in the town held slaves. In Ouro Prêto, capital of the old mining district of Minas Gerais, the figure was 41 percent of the households in 1804. This level of diffusion in urban areas is borne out by a published census for the parish of São Pedro in the city of Salvador in 1775, where 47 percent of the households in that central parish contained slaves, The evidence is scattered, to be sure, but it supports the impression given by foreign travelers that slavery was an ubiquitous institution in the cities and towns of Brazil" (1982, 76-77). With less precise indications, something similar seems to have been the case in Costa Rica (Olien 1980). [BACK]

16. In Rio de Janeiro, state authorities were responsible for punishing slaves. Sometimes, this charge became a pretext to use privately owned slaves for public tasks; slave owners received payment for the slaves' work; but if slaves were sent to prison, owners were asked to pay for part of their upkeep (Nogueira da Silva 1988, 150 ff.; also see Algranti 1988). [BACK]

17. In this whole setting, Rout's (1977, 93) assertion that in Hispanic America manumission was a gift and not a right, and that those slaves who enjoyed liberty were basically lucky subjects, does not hold true for Lima. There slaves weakened the system and did so not through luck but through a complex set of attitudes and options that grew from their everyday experiences. [BACK]

18. In following this question, Lombardi (1974, 168, 170) echoes Harris, Elkins, and Klein. [BACK]

19. In a recent symposium at the University of California, San Diego, Viotti da Costa, Blackburn, and Scott (all 1991) provided new insights on this theme; see also Tomich (1990) and Toplin (1981). [BACK]