This essay attempts to infer the economic performance of the free people of color of San Juan, Puerto Rico during the nineteenth century. There is no direct way to determine such performance, because of the absence of appropriate documentation. The activities of the many who served as domestic servants, for instance, were not recorded systematically. The area of the economy where free people of color functioned most effectively was crafts, but there appear to be no guild records extant, and censuses normally did not record occupation. The study emanated from a larger inquiry into the relationship between racial discrimination and the dynamics of the Spanish American market economy. Free people of color of San Juan, a large and almost completely unstudied segment of the population, comprised a legally disadvantaged caste that was required to function in an economy of commercial capitalism. The caste system may have ended in Mexico City, as Dennis Valdes argues,1 or was irrelevant in Guadalajara, Mexico, as Rodney D. Anderson has shown,2 and perhaps was on the wane elsewhere,3 but it was vital and effective in the Caribbean, and it undermined the efficacy and promise of the marketplace as arbiter of economic performance.

One pertinent measure is house ownership. The absence of such ownership does not necessarily indicate poor economic performance, since members of any group being investigated might have elected to place their capital in rural lands or in inventories, for instance, rather than in buildings. Ownership of houses, on the other hand, does suggest economic achievement. It is not easy, however, to ascertain the degree of real property ownership among free people of color in urban areas of Spanish America. Real property censuses normally did not enumerate the race of the property’s owner (and the honorific "don" in San Juan was employed so indiscriminately by census enumerators that it serves as an unreliable guide to race), and nominal censuses normally did not record property data. Hence, although there have been studies of free people of color in Spanish America, only modest attention has been paid to the matter of real property ownership.4

Several terms used in this article require elaboration. The Spanish American colonial economy, at least by the middle of the eighteenth century, was fundamentally one of preindustrial commercial capitalism. Indeed, the colonial economy and then the national economies were to varying degrees based on money, exchange, private property, and the profit motive.5 John K. Chance sees "the early development of a capitalist socioeconomic system in central Mexico."6 Thus, his study of race and class in colonial Oaxaca, Mexico "can best be understood within the context of a developing system of commercial capitalism."7 Anderson asked recently in this journal whether the economic and social stratification of Mexican society in 1821 should be seen "as … the result of economic changes accompanying the rise of early commercial capitalism,"8 and he answered in the affirmative. Both Chance and Anderson are certainly correct in their estimation of the nature of the economy.

But here our views diverge. For Chance, in colonial Antequera, now Oaxaca, "dramatic growth and the increased opportunities for trade after the Bourbon reforms rendered the sistema de castas all but obsolete as a mechanism of status definition."9 The "large miscegenated population relied primarily on its economic situation to define its place in society."10 Neither observation is true for San Juan, where, despite exceptions and differentiation within the colored caste, social status was still determined primarily by race, and the large miscegenated population enjoyed little discretion in the matter. Nor is Anderson’s view applicable to San Juan, where economic and social stratification must be understood within the context of a caste system that survived into the era of commercial capitalism. In fact, in San Juan racial disability did not permit commercial capitalism to work its sway and produce a class system. My intent is not to imply that capitalism is free of racism, but to make clear that in the Puerto Rican case racism was promulgated and perpetuated by the caste system.

One way to avoid the contentious terminology of class and caste would be to rank society according to stratification theory, which, as Magnus Mönier has observed, is "a multidimensional way of organizing society according to the variables of wealth—income, occupation, ethnicity, education, kinship, status, and power… ."11 Stratification is a useful descriptive technique to order data drawn from censuses, as long as a convincing number of variables are included.12 Unfortunately, the dearth of socioeconomic data in the San Juan censuses makes it impossible to situate the population within even such crude categories as upper, middle, and lower strata. Much less is it possible to correlate race and income or race, income, and occupation. Nor, therefore, is it possible to apply class analysis in this investigation.

Although the data are scarce, it is clear that there was a wide range of economic performance among the free people of color, to the degree that both status and interests would have divided them into several classes. Spanish racial policy, however, did not permit the dynamics of commercial capitalism to render the free people of color of San Juan into groupings of similar economic performance and social status, and, ultimately, of shared social, economic, and political values. Rather, it created and perpetuated a system in which free people of color were categorized into castes and separated legally from both whites and Indians. The three great castes—white, free colored, and Indian—were subject to differing rights and responsibilities. These castes were less rigid and less corporatist than late medieval and early modern European estates, but otherwise had much in common with them. In Puerto Rico, where there were few Indians, there were, by the end of the eighteenth century, only two significant castes—white and free colored. The caste system in Puerto Rico circumscribed the social and economic mobility of the latter, yet it did not insulate whites from downward mobility or ensure them high social status. Several scholars have noted the presence of Spaniards in low-status occupations.13 In nineteenth-century Puerto Rico there are many examples of whites occupying low-status positions in the economy.14

The legal impediments set against Puerto Rico’s free people of color, which defined their caste, were, with few exceptions, common to the rest of colonial Spanish America, and are well known to historians. Suffice it to say that ostensibly free people of color were not permitted to become physicians, lawyers, or notaries, or to enter the various bureaucracies—civil, church, and military.15 In Puerto Rico, they were permitted to serve in the military, but only in segregated militia units.16 These barriers produced functional differences between the castes; that is, certain economic positions were reserved for whites and prohibited for the colored caste. The invidious nature of such a system could come into play at any moment, as in 1815 when free people of color and whites were invited to emigrate to Puerto Rico with rewards of land, but in smaller allotments for the former than the latter.17

Our understanding of the nature of prejudice and racism in the period 1750 to 1850 is still rudimentary, but there is reason to believe that the Puerto Rican caste system, legally defined, became, in Teun A. van Dijk’s term, a sociocognitive system.18 That is, prejudice and racism in Puerto Rico had their roots both in individual cognitive processes and in "the social dimension of intergroup relations and prejudice."19 On any number of occasions, an individual in Puerto Rico might be required by the authorities to prove his or her pureness of blood, i.e., whiteness.20 Free people of color were routinely referred to as such in legal documents, and marriage to a person of color carried with it liability. In 1806, Don Manuel Hernáiz purchased at public auction the position of alderman and chief constable on the town council of San Juan. Opposition to admitting Hernáiz existed within the council on the ground that his wife was "notoriously parda," though Hernáiz was not removed from his position.21 In 1809, Don Tiburcio Durán Villafañe purchased a seat on the town council. One of the alcaldes objected to this because it was rumored publicly that Villafañe was not of completely pure blood.22 In 1810, the Audiencia ruled in favor of Villafañe, whether on the merits of the case or in support of the crown’s venal policies in the sale of bureaucratic positions. Nevertheless, the problems faced by both Villafañe and Hernáiz indicate the presence of racial values beyond legal strictures.

At this juncture, a further term must be clarified. Historically, Puerto Ricans recognized degrees of blackness, generally descending from white to pardo, to moreno, to negro. Pardos almost always were the free people of color of lightest skin color; morenos were darker skinned, and negros the darkest of all. These, nevertheless, were flexible terms, variously used, and they could be applied to both slaves and free persons. Further, the term "mulatto" was sometimes used.23 The only term of the era that fairly well defines those blacks, of whatever hue, who were legally free was gente de color.

Perhaps the extreme expression of Puerto Rican prejudice and racism in the nineteenth century was Governor Prim’s infamous decree of May 1848 against the "African race." On learning of the destructive activities of the former slaves just liberated by the French in Martinique and Guadeloupe, and considering the "ferocious stupidity of the African race," Governor Prim set out to eliminate the possibility of similar uprisings in Puerto Rico. Henceforth, all crimes committed by people of African descent, be they free or slave, would be judged by a military court. Any free colored person who used arms against a white, no matter whether justified, would have his right hand cut off. In the event that the white was wounded, the person of color would be executed by firing squad. We need not pursue this virulent decree further, except to note that it was rescinded by a new governor six months later.24

An example from Ponce illustrates the effectiveness of the caste system. In 1860, Ponce, Puerto Rico’s "second city," comprised five urban barrios, of which three that were studied for this project held 3,973 of the city’s 6,406 inhabitants (62 percent).25 Only a few free people of color functioned at the higher levels of the commercial hierarchy in the three barrios, although there was no legal restriction to prevent them from owning retail stores or the more heavily capitalized and prestigious wholesale domestic or import-export operations. Only three free people of color were counted among the merchants (comerciantes) in the three barrios, and few others could aspire to work their way up the employee ranks. Judging by the way the terms empleado and dependiente were used in other parts of Spanish America, it is likely that those listed as empleados were commercial employees of considerable importance and responsibility, more so than the dependientes, who were clerks of various descriptions, and degrees of responsibility. The chance that a free person of color might become an upper-level commercial employee and then perhaps move on to ownership is suggested by the fact that in the three Ponce barrios in 1860 none were empleados. But more indicative of the extent of commercial opportunity for the free person of color is the number of dependientes: 10 free colored males were clerks, of whom 7 could read or write. Certainly there was some opportunity. But the racial difference in opportunity is disturbingly implicit in the number of white dependientes (83). Why was there more than an 8 to 1 difference in favor of whites? Very possibly the pool of white literate youth was greater than that of free colored: that all but 2 of the 83 white dependientes could read or write reflected their opportunity for basic education. Perhaps also whites, who controlled the commercial establishment, preferred to hire whites, something that would not surprise the modern reader. There were another 2 free colored dependientes in one of the unstudied barrios, 1 of whom appears to have been a youthful female, as well as 1 colored merchant. However, there were 19 white dependientes and 29 white comerciantes.

Castes are inherently marital endogamous, and those of Puerto Rico were no exception. A useful measure of racial preference in mate selection is the endogamy ratio, which compares the observed marriages within or between the races being evaluated to what one would expect in a hypothetical situation of perfectly random selection. Where the endogamy ratio is greater than 1.0, marital parings within a race are indicated, while a ratio of less than 1.0 indicates degrees of marriage outside one’s race. Table I depicts the endogamy ratios of white and free colored married heads of households in the San Juan barrios of Santa Bárbara (in 1823), San Juan (in 1828), San Francisco (in 1833), and the first trozo (administrative unit) of the barrio of Santo Domingo (in 1846).26 Clearly, racial bias in mate selection was the rule, and this is no surprise.

What is perhaps surprising is that endogamy prevailed also within free colored subgroups. In such racially categorized subgroups as pardo and negro libre, marriage was endogamous to the degree that no endogamy ratio was less than 1.0. In fact, the lowest ratio was 2.61, indicating strong racial bias in the selection of marital partners. To be sure, these ratios are based on small numbers (small cell counts) and require substantiation from other Puerto Rican studies; they must, therefore, be considered preliminary findings. They do, however, suggest the extent to which nonwhites had absorbed the values of the caste system. In this same connection, Verena Martínez-Alier has noted several nineteenth-century Cuban instances in which families of lighter-skinned free colored children objected to their proposed marriages to partners of darker color.27 She has also adduced evidence that, when free people of color married, they "did so with their own shade… ."28 This was certainly the case in the four San Juan barrios studied.

Although the crown created the caste system, it intended that free people of color should be productive members of society and be able to earn livings and support themselves. Thus, in Puerto Rico they enjoyed considerable freedom and legal protection. They could travel freely on the island; there was no curfew; they could gather publicly in groups, and dance in the streets if they so desired. They could own stores, acquire land in whatever quantity, inherit property without restriction, and acquire an education, even if rudimentary.29 Very importantly, they were able to enter any craft, and in San Juan they entered all.

Two legal rights were especially important and were fundamental to the participation of the free colored caste in the market economy. Puerto Rico’s free people—male and female, regardless of color—had access both to the courts and to the town councils. For the great majority of the island’s population, the most important courts were the small claims courts, which were divided into two categories: juicios verbales and paz y conciliación. All civil cases involving sums of less than four hundred pesos were heard by the local alcaldes and resolved verbally. In the absence of resolution, the litigation was raised to a higher level, paz y conciliación, for adjudication. It was at the small claims level that most storekeepers, artisans, and the rest of the population would have taken their legal problems involving money.30 At the town council the free person of color could appear in person to request the purchase or rental of a piece of land owned by the town, or a group of them might appear to redress a labor grievance.31 In both instances, the courts and town councils, these were legal rights of no small importance.32

Residential Patterns

The documentation for a residential study of the free people of color of San Juan is limited. There is no single nominal census for all four of the historic barrios during the period 1750-1850. Several of the extant censuses, covering only a single barrio, are in such precarious physical condition that the next person to read them, prior to some future restoration, will unavoidably be the last to do so.33 Fortunately, however, there are three good barrio censuses that together cover a reasonably short period of time: Santa Bárbara for 1823; San Juan for 1828; and San Francisco for 1833. These censuses do not provide the coherence of a single year, but, alternatively, they permit an evaluation of a large, although not random, sample of San Juan’s population at the barrio level over the course of a decade. Furthermore, they are fairly proximate to a San Juan real property census of 1820, which provides property values and permits the addition of an important dimension to the discussion of property ownership. To enlarge the data base, I have also studied the first trozo of the fourth traditional barrio, Santo Domingo, from a census conducted in 1846.34 I would have preferred a nominal census of Santo Domingo closer to the 1823-33 period, but none exists. The first trozo was selected simply because it was in usable physical shape and was available to me during my stay in San Juan.

San Juan was a small city, both physically and demographically. The capital of the colony and seat of the royal bureaucracy supported a population of only 7,658 people in 1820.35 The reliable Pedro Tomás de Córdova, secretary to the governor, calculated the resident population in 1824 to be 8,453. For 1828, Córdova presents the figure of 9,453 residents, plus several thousand soldiers and sailors, for a total of 12,744.36 During the early nineteenth century, there were scores of provincial towns spread throughout Spanish America with populations as large or larger than San Juan’s.

Toward the end of the eighteenth century, San Juan was 682 meters long and 430 meters wide.37 Running east-west were five main through streets (calles). Running north-south were also five main through streets. De Córdova remembered that "the streets are cast in a straight line, they are of the same width, divided into … blocks of little more than one hundred yards."38 The individual streets created by the crisscrossing of the through streets were the cuadras. The blocks themselves were called manzanas. In 1820, there were 58 cuadras in San Juan.39 In that year, the capital consisted of the four barrios (often referred to as cuarteles) already mentioned. By 1859, the much less populated barrios of Ballajá, De la Marina, and Puerto de Tierra had been added.40

The four historic barrios comprised roughly the area incorporated in what today is referred to as the Old City. To the southwest was the barrio of San Juan, the administrative and commercial center of the town. Here were to be found the offices of both the royal bureaucracy and the town council as well as the town’s main plaza and market. The cathedral of San Juan was also located in this barrio of prestigious residence. As can be seen on the map, based on one drawn sometime during the period 1847-53, the barrio of San Juan contained 11 manzanas and 22 cuadras.41

At the southeast of the town and just to the east of San Juan was the barrio of San Francisco. It contained 17 manzanas, and, depending on how one interprets the map, between 27 and 32 cuadras. To the north of San Francisco was the barrio of Santa Bárbara, containing at midcentury 12 manzanas and between 24 and 27 cuadras. To the west of Santa Bárbara and north of San Juan was the barrio of Santo Domingo, which contained 12 manzanas and between 25 and 29 cuadras. In addition to the important Dominican monastery, Santo Domingo housed the military hospital.

While it is not possible to rank the four barrios in order of importance or prestige, it is possible to gain some sense of their relative status in 1820 when a real property census was compiled.42 In that year, the barrio of San Juan contained buildings and lots (solares) with a combined value of 445,500 pesos (Table II). Santo Domingo’s buildings and lots were valued at 386,025 pesos. Santa Bárbara had the least gross valuation, 290,570 pesos, and San Francisco the greatest, 774,300. But San Francisco, with the greatest number of properties, did not have the highest mean valuation. It was San Juan, just as one would expect of the administrative and commercial center of the town, which held, on average, the most valuable properties, while Santa Bárbara was the barrio where values were lowest. As can be seen from Table III, the mean valuations are much the same whether all properties or only those privately owned are considered.

The majority of the city’s houses and lots in 1820 were modestly valued, ranging between one and five thousand pesos. Nearly two hundred properties were valued at less than one thousand pesos (including the 36 vacant lots in the city), and more than four hundred were valued at less than two thousand pesos. Santa Bárbara held the greatest number of very inexpensive properties.

Free people of color resided in all of the barrios, and in substantial numbers. In 1823, approximately 59 percent of the population of the barrio of Santa Bárbara were free colored. In 1828, in the barrio of San Juan, almost certainly the town’s most prestigious barrio, approximately 38 percent of those who were free were people of color. In the barrio with the greatest number of dwellings in 1820, San Francisco, approximately 38 percent of the free population in 1833 were colored. In the first trozo of Santo Domingo it was some 43 percent in 1846.

Not only did free people of color reside in each barrio studied, they resided on every street in the barrios of Santa Bárbara and Santo Domingo (first trozo), and every block (and perhaps street) in the barrios of San Juan and San Francisco. The census of Santa Bárbara for 1823 does not group houses according to blocks or streets, so the following detailed discussion considers only San Juan, San Francisco, and Santo Domingo; but a nominal census of Santa Bárbara for 1818 makes it clear that free people of color also lived on each of that barrio’s streets.43

Although whites and free people of color were separated in many ways by law, they were clearly willing to live side by side with each other. Table IV contains the numbers of white and free colored heads of households (coresident families, including solitaries) resident on each block in the barrios of San Juan, San Francisco, and streets in the first trozo of Santo Domingo. By any useful definition, this was an unsegregated city. On several of the blocks, whites and free people of color resided in almost perfect parity. On only one, block number 6 in San Francisco in 1833, were there more free colored than white heads of households.

This apparently liberal attitude toward residential selection on the part of the whites and free people of color meant that no area of the barrios studied was overwhelmingly free colored, and thus likely to have been shunned by white storekeepers to the advantage of free colored entrepreneurs. One of the consequences of residential segregation in other parts of the Americas is that colored storekeepers and professionals sometimes have found themselves with freedom from white competition and virtually a captive market. After studying 15 cities in both the North and South of the United States during the first half of the nineteenth century, Leonard P. Curry concluded that "north of the slave states, the development of black entrepreneurial activity seems to be directly related to the degree of residential segregation."44 Boston, the most heavily segregated of the cities studied, had the second greatest percentage of blacks involved in entrepreneurial endeavors, exceeded only by New Orleans, which was also heavily segregated.45 Although a systematic study of entrepreneurial activity at the lower reaches of the economic continuum has yet to be done, my own work on the small retail grocers in San Juan and Caracas suggests at most very few free people of color among them.46 It would be ironic if a seemingly liberal attitude toward integration of the races unintentionally contributed to limited economic opportunity among San Juan’s free colored population.

Although an extreme example of an unsegregated city, ban Juan was not unique in Spanish America. Dennis Valdes found that "[r]acial and class segregation by zones or blocks … did not occur in eighteenth-century Mexico City."47 Studying the "southwest cuartel of the Spanish sector" of Mexico City in 1753, Valdes calculated that pardo (here a general term for free colored) heads of households resided on 29 of the cuartel’s 36 streets.48 Anderson has observed that in Guadalajara, Mexico in 1821 "[n]ot only did non-don Spaniards live in the same cuartel with Indians and castas, but they often lived side by side in the same block or rubbed elbows within the same household."49 (However, it should be noted that Guadalajara contained only a "small scattering" of free people of color.50) Chance has shown that in 1792 there was some clustering by peninsular Spaniards in Oaxaca, but that all other groups, including mulattos, lived throughout the city according to no pattern of segregation.51

However, this article argues that an absence of residential segregation does not mean the absence of a caste system. Chance studied segregation in Oaxaca within a hypothesis drawn "from the ecological branch of urban sociology," which, in his words, posits "a general fit between the residential distribution of race and class on the one hand, and the place of these dimensions within the overall urban social structure on the other."52 It may be that Chance is correct in his conclusion about caste and class in Oaxaca, but his sociological hypothesis is invalid in the case of San Juan.

That the San Juan areas studied were virtually unsegregated becomes more apparent when the city is compared with others outside Spanish America, which may be done using the statistic known as the segregation index. An index of dissimilarity, this is both a crude and controversial measure of segregation, but it is useful for comparative purposes if used cautiously. Simply put, the segregation index measures the percentage of whites and nonwhites who would have to be moved from one geographical unit, in this case block or street, to another in order to produce a racial proportion roughly equal to that of the city or barrio at large.53 One limitation of the segregation index is that it presumes—not wholly realistically—that factors such as income or cultural achievement, for instance, are not influential in residential selection.54 A further problem is that not all censuses measure at the same geographical level; some do not enumerate at the street level, thus limiting the usefulness of the index for comparative purposes. Yet the values for three of San Juan’s barrios are so consistent and so different from what one might have expected of a slave society that they are indeed useful.

The values expressed by the segregation index range from 0 to 100, with the lower values representing less and the higher values greater segregation. For the barrio of San Juan in 1828, the segregation index of heads of households was 15.9; for the barrio of San Francisco in 1833, it was 12.7; and for the first trozo of the barrio of Santo Domingo in 1846, it was 13.0. The segregation index at the block or street level may be seen in Table V. Clearly, the total impression is one of virtually no segregation.

Although the segregation index is useful for the purposes of comparison and generalization, like many other demographic measurements it potentially distorts historical reality because of the manner in which the heads of households are calculated. The conventional procedure is to count the first named individual in the household as its head. This method contributes to standardization of terms, and renders data amenable to comparison. However, as David Robinson notes, there is room for discretion in determining the head of household, and, in fact, censuses sometimes varied considerably in how they did this.55 Unfortunately, scholars typically do not indicate if, or how often, they have deviated from the standard of coding the first named in the household as its head, and one is left with the impression that deviation from convention is rare. In dealing with large numbers of people, this would almost certainly make no significant difference, but at the micro level distortions may result. For example, it is well known that men, often but not always bachelors, shared residences; and in San Juan there are many instances of bachelors, and sometimes married men or women without spouses present, sharing residences. In 1828, for example, in house number 86 in the barrio of San Juan two white, single musicians, one 23 and the other 22, rented an apartment (or room or rooms). Conventional coding would consider these two as one household with the 22 year old, because he is listed first, as the head. This makes no good historical sense.

Conventional coding technique overlooks many men and women, white and free colored, when such things as the segregation index and, as we shall see, household ownership are calculated. By conventional coding, one block in the barrio of San Francisco had 15 white heads of households, when in reality it probably had as many as 30. On another block, standard coding produces 103 white heads of households when it is likely that there were as many as 122. These distortions proved statistically insignificant at the barrio level when the segregation index was calculated both conventionally and according to a more historically accurate evaluation of each household. However, when the results of both coding methods are compared at the block level, it becomes clear that conventional coding sometimes masks a higher degree of segregation, just as it has the reverse effect on household ownership statistics, making it appear that a greater percentage of heads of households owned dwellings than was, I believe, actually the case. Historians concerned about the particular in their search for the general may desire to code their censuses not only in the conventional manner but also alternatively, a time-consuming but not technically difficult task.56

This caveat aside, we may use the segregation index for comparison with U. S. cities, for which extensive data are available on the spatial distribution of white and black residents. Working with individuals and wards rather than households and blocks or streets, Curry has calculated segregation indices for 15 cities in the United States between 1820 and 1850.57 In four decennial years beginning with 1820, Boston’s index ran between 46.3 and 59.2. In 1820, the index for New Orleans was 11.6, but by 1850 it had risen to 45.2. Charleston’s had increased from 10.0 in 1830 to 20.6 in 1850. Baltimore’s was relatively low, 13.1 in 1820, 11.6 in 1830, 9.9 in 1840, and 21.5 in 1850.58 Working with households and blocks, the Taeubers have computed segregation indices for 109 cities in the United States for 1940, 1950, and 1960. In almost every instance, the values were either in the 70s, 80s, or 90s. New Orleans’s, for instance, ranged between 81 and 86.3. Baltimore’s ran from 90.1 to 89.6. Among the 45 southern cities analyzed, Charleston’s were always the lowest, ranging from 60.1 to 79.5.59 The magnitude of these indices for the modern period places San Juan’s low nineteenth-century values in sharp relief.

Nowhere was the general absence of residential segregation in San Juan more apparent than at the level of the individual houses. In San Juan, whites frequently lived next door to free people of color, and, more interestingly, often resided within the same buildings, just as they did in the Mexico City cuartel of 1753 studied by Valdes.60 During the first half of the nineteenth century, there were three types of buildings in the city of San Juan. Most prevalent were casas, that is, houses. There were also a few bohíos, generally thatch-roofed shacks.61 The third type of edifice was the nonprivate building—public, church, or military—in which a few people lived. In my usage, the term "house" includes casas and bohíos.

The degree to which whites and free people of color shared houses in San Juan is impressive. There were 205 occupied houses in the barrio of Santa Bárbara in 1823, and in 78 of them (38 percent) resided both white and free colored families (including solitaries). More remarkably, perhaps, in the barrio of San Juan in 1828 there were 131 occupied houses, and the races were mixed in 83 of them (63.9 percent). Similarly, in the barrio of San Francisco in 1833 there were 275 occupied houses, and families of different races resided in 164 (61.0 percent). By a slight majority, commingling of the races also prevailed in the first trozo of the barrio of Santo Domingo in 1846, where there were 128 occupied houses, with families of different races residing in 63 of them (50.8 percent).62

The several hundred examples of houses with both white and free colored families resident are of interest for what they reveal about race relations in the slave city of San Juan. In the barrio of San Francisco in 1833, for instance, a town councilman, Don José Izquierdo, owned and lived in his own house. He rented rooms or apartments to several free colored households. In another instance, a white family was renting an apartment from a free colored owner who also resided in the building. Other such cases can easily be cited. However, to pursue this dimension of the free colored experience, I will single out the barrio of San Juan for discussion. It is precisely there, the seat of local and royal government and the center of business activity, that one might expect to find a harsh separation of the races, not the reverse.

Many houses in the city of San Juan were extremely large, some housing 30, 40, or 50 people. Often these resembled boarding houses, and perhaps some were just that, although the censuses are silent on the matter. In some instances, this "boarding house effect" may have contributed to an intermingling of the races. For instance, 37 people lived in house number 2 located on block 1 of the barrio of San Juan in 1828. The house was owned by the estate of the late Don José Ribera and in 1820 was valued at six thousand pesos. Ribera’s estate now rented or otherwise provided the house to Don Manuel Aldea, a 31-year-old white man married to Ribera’s 52-year-old widow. Aldea, or the estate, in turn rented out seven apartments. The first renting head of household was white, the second was free colored, and the third was white. The rest of the renters were all free colored. House number 5 on the same block housed 40 people. It was rented by Don Juan de Dios Cuevas, a Spanish bureaucrat. Cuevas provided apartments to 3 renters, one an 80-year-old free colored widow. House number 27 on block 3 was valued at nine thousand pesos in 1820. The owners, three siblings, lived in the house but rented it to Don Francisco Gonsález de Linares, a 50-year-old Spaniard who resided there with his family. It was he who apparently rented apartments to 7 others, among whom were 2 free people of color and their families. Together, 52 people lived in the building. House number 86 on block 6 also seems to have had the quality of a boarding house. Thirty-nine people lived in it. At its head was Doña Josefa Montenegro, the 40-year-old white widow of the dwelling’s owner. Doña Josefa rented apartments to 13 heads of separate households, 5 white and 8 free colored. Clearly, not only were some San Juan houses large and densely populated, they were highly integrated.

Other housing arrangements are especially suggestive. House number 12 on block 1 was rented by Don Mariano Taforo, a white military officer from Spain, who resided with his family. The owner of the house, also resident, was a free colored widow who lived there with her daughter and three slaves. Just a short distance away, in house 13, lived the intendant, Don Mariano Sixto from Madrid, the colony’s second ranking royal official. It appears, but is not certain, that it was Sixto who rented the house and who, in turn, rented apartments to four other people. The first three renters were white, but the fourth was free colored. If the intendant lived in a house occupied also by free people of color, it is not surprising that the practice would be commonplace. It is in the context of such manifest freedom of residential choice that one can fully appreciate the nature of house ownership among San Juan’s free people of color—both the degree to which they owned dwellings, and equally importantly, the degree to which they did not.

House Ownership

San Juan was a city of renters. Very few free people of color or whites owned houses, nor were there enough structures for many to have done so.63 However, a much higher percentage of white heads of households owned their own houses than did free people of color.64 Only in Santa Bárbara in 1823, the barrio with by far the lowest mean value of dwellings in 1820, was the percentage of ownership among heads of households fairly close between free people of color and whites (see Table VI). The barrio of Santa Bárbara clearly was a place of real property ownership opportunity for free people of color, but the data suggest a specific relationship between such opportunity and modest property values. Certainly very few free colored heads of households were capable of purchasing a dwelling in the other barrios studied.

But this is not the entire story of house ownership among free people of color, since if the data are refined to exclude those heads of households who were not also the heads of the residential units a somewhat different picture emerges. The heads of residential units were those heads of households (including solitaries) who were the owners or primary renters of the houses they lived in, and who, if theirs were multiple residences, rented one or more apartments (or rooms) to other households. In this usage, the term "residential unit" follows the Joint Oxford-Syracuse Project, and means simply a dwelling with "any number of households."65 It is further assumed that one household "serves as the main household."66 However, in the case of San Juan it is not always possible to determine which household was the main household. When several households rented within one residential unit, and the first household listed in the census manifested no decidedly superior economic status, it is unwise to hold that the first one listed was the main household. In fact, the several households might have rented independently, with no family enjoying a superior position vis-à-vis the others. If the first listed household owned the house, I have assumed that it was the main household since it rented to others; but there are also instances when the household that owned the structure or the part of the structure serving as a residence was not the first one listed in the residential unit. In those instances, I normally assumed that the first listed household was the main household under the presumption that it rented the house from the owner and, in turn, rented space to additional households (including that of the house owner, through whatever arrangement). There are, however, some instances when it is not at all clear from the census that the first listed household was in any way in a position superior to the others. In such cases, the house was considered to have had no head of the residential unit, with the result that the number of heads of residential units in Figures IIII is less than the number of houses in each barrio.67

Being the head of a residential unit frequently implied some degree of achievement and responsibility beyond what normally would be associated with simple tenantry. The head of a residential unit who rented to other households was, after all, a landlord. And when only this group is considered, the degree of house ownership among free people of color becomes both more impressive and somewhat provocative.

As one would expect, there existed a substantial degree of house ownership among free colored heads of residential units in the barrio of Santa Bárbara in 1823. In addition, approximately the same degree of house ownership prevailed among both male and female free colored heads of residential units, whereas in the case of whites there existed a greater degree of ownership among the males than white females (see Figure I). The degree of house ownership by free people of color is impressive with respect to that of the whites of both sexes, but it is no doubt significant that it was the white males who displayed the highest degree of ownership, an indication of their ability to be the basic family provider.

It is also of interest that a greater percentage of free colored female and male heads of residential units owned the houses in which they lived than did their female and male white counterparts (Figure II). This is not what one would have expected. It means that among those free colored men and women who had been successful enough in the economy and who exerted enough initiative to become or remain heads of residential units, there was an even greater degree of residential ownership and, in this regard, achievement than among whites. To be sure, this fairly widespread dwelling ownership occurred in the barrio of Santa Bárbara, where housing was the cheapest in the city in 1820 and more than likely for a long while thereafter. That more free colored female heads of residential units owned the dwellings in which they lived than did their white female counterparts may in part have been the consequence of lower marital rates among the free colored women, a reflection of severe demographic problems within the free colored community, a point which I shall discuss in a succeeding article.

Few free colored heads of residential units owned the houses in which they lived in the other barrios (Figures IIIA, IIIB, and IIIC). When compared with whites, their degree of house ownership was generally miniscule in the barrios of San Juan, San Francisco, and Santo Domingo. In each instance it was also less than occurred among free colored heads of residential units in the barrio of Santa Bárbara.68 Among whites, the numbers of heads of residential units who owned dwellings in the three barrios and (to a lesser extent) the first trozo of Santo Domingo69 are more substantial, and, in each instance, there were more white males represented than there were females. The pattern appears indicative of a healthy demographic and economic performance among whites, and, more generally, of a white dominant society.

Notwithstanding the imbalance in ownership, those free people of color who owned houses were sometimes an impressive group who belie their standard characterization. Perhaps the most famous description of housing in San Juan during the colonial period was made by Fray Íñigo Abbad y Lasierra of San Juan in 1782. Abadd recorded that the Spaniards and well-to-do residents lived in houses made of stone and covered by tiles. Some had flat brick roofs. The houses of the mulattos and people of color were made of wood. Those of the blacks and poor people were coarser and smaller, little more than straw cages, supported by canes and covered by palm leaves. These were the bohíos.70 By the 1820s, however, the picture was somewhat different. Abbad certainly did not include free people of color among the well-to-do residents, but, in the early decades of the nineteenth century, some of them owned stone dwellings, and some even owned the generally more valuable two-story buildings.

Several examples from the barrio of Santa Bárbara serve to illustrate that free people of color were not limited to the ownership of inexpensive houses. In 1820, when the mean value of a privately owned edifice in the barrio was 1,640 pesos, one 40-year-old free colored resident had inherited a house valued at 3,000 pesos.71 Bartolomé Cueto, a 40-year-old free colored, owned a house with an estimated value of 2,500 pesos. At 43, Antonio Arriago, free colored, owned three houses, apparently contiguous, with values of 1,500, 3,000, and 2,000 pesos. Don Atlántico Vargas was a 50-year-old free colored grocer. In 1820, he owned a house valued at 2,000 pesos. A man of the same name, almost certainly the same person, owned a house on the same street valued at 2,100 pesos. Among free colored women who owned property was Rita Molina, 40, whose house was valued at 1,800 pesos.

The property ownership of some free people of color suggests their considerable economic success. In 1845 Manuel Elías, a free colored silversmith, owned the three contiguous two-story houses on the street of San Francisco in the barrio of that name which he had inherited from his father, Pedro Elías.72 In 1820, when still owned by the father, the houses were valued at 3,500, 3,000, and 1,500 pesos respectively.73 In 1833, house number 154 was lived in by Manuel, along with his wife, five children, and three slaves; he rented out houses 155 and 156. In 1845, Elías not only owned the three houses on Calle San Francisco, but a fourth on Calle del Sol.74 In 1847, Antonio Cruzado inherited a house from a former slave. He quickly sold the house for 2,300 pesos.75 A month later, he also sold for 3,350 pesos a one-story stone house, which he had built at his own expense on a lot he purchased in 1841.76 In 1854, he sold another one-story stone house, this one for 2,150 pesos.77 Yet in 1854, he still owned one more stone house of the same type.78

As already mentioned, women were often house owners. A further example is Rosa Peña, free colored, who in March 1841, after requesting and receiving permission from her husband, sold a one-story stone house at the Plaza (actually Plazuela) Santiago in the barrio of San Francisco for 1,700 pesos. She had bought it in 1832 for 550 pesos.79 The following month, Peña purchased a single-story stone house on Sol for 1,600 pesos.80 In 1844, again with permission from her husband, Peña sold this new house for a handsome 3,500 pesos.81

The data presented in this essay suggest, then, that San Juan was a place of opportunity for free people of color. Both legal and social racism were moderate enough to permit them a wide choice of residential location and a broad freedom to purchase residential properties. Free people of color in San Juan lived in the same buildings as did whites, or they lived across the street or down the block from them. They purchased houses in which to live or for purposes of investment. But, although house ownership suggests considerable economic success on the part of some free people of color, opportunity in general was limited. Few whites or free people of color owned their own houses in the barrios studied, but not unexpectedly white heads of households owned proportionately more, except in the one barrio of Santa Bárbara, where both races were approximately equal in ownership. When only heads of residential units are considered, the relatively high degree of house ownership among free people of color in Santa Bárbara, the barrio of lowest dwelling values in 1820, suggests that ordinarily it was not initiative and desire to achieve that was lacking among free people of color, but rather economic means, at least in part—one can safely assume—as a consequence of centuries of caste-related restrictions. Commercial capitalism and the market economy offered a broad panoply of economic opportunity, but racism, mediated through the caste system, vitiated the promise. This becomes even clearer when one compares free colored and white marital patterns, household size, and gender and age differences, as I shall do in a future study.


Dennis N. Valdes, "The Decline of the Sociedad de Castas in Mexico City" (Ph.D. diss., University of Michigan, 1978), passim.


Rodney D. Anderson, "Race and Social Stratification: A Comparison of Working-Class Spaniards, Indians, and Castas in Guadalajara, Mexico in 1821," HAHR, 68:2 (May 1988), 209-243. Michael M. Swann has shown the weakness of caste in northern Mexico in Tierra Adentro: Settlement and Society in Colonial Durango (Boulder, 1982) and Migrants in the Mexican North: Mobility, Economy, and Society in a Colonial World (Boulder, 1989).


John K. Chance and William B. Taylor, "Estate and Class in a Colonial City: Oaxaca in 1792," Comparative Studies in Society and History, 19:4 (Oct. 1977), 454-487. See also the criticism by Robert McCaa, Stuart B. Schwartz, and Arturo Grubessich, "Race and Class in Colonial Latin America: A Critique," ibid., 21:3 (July 1979) and Chance and Taylor, "Estate and Class: A Reply," ibid., 434-442.


An exception is Swann, Tierra Adentro, 294-303. There are fragmentary data on real property ownership by free people of color in Havana, suggestive of considerable wealth. See Pedro Deschamps Chapeaux, El negro en la economía habanera del siglo XIX (Havana, 1971), 65-71, 79-80, 96, 124, 143-146, 164, 179, 183. Property ownership among the free people of color of Caracas was apparently widespread during the latter part of the colonial era. See Kathleen Waldron, "A Social History of a Primate City: The Case of Caracas, 1750-1810" (Ph.D. diss., Indiana University, 1977), 79. Patricia Seed makes an interesting observation about the relationship between "the acquisition of property" and "upward socioeconomic mobility" of mulattos in Mexico City in 1753 in "Social Dimensions of Race: Mexico City, 1753," HAHR, 62:4 (Nov. 1982), 569-606. Jerome S. Handler provides interesting data on house ownership among freed people in Barbados, although based on an admittedly very small sample (The Unappropriated People: Freedmen in the Slave Society of Barbados [Baltimore, 1974], 144-146).


For a sensible attempt to understand the term capitalism, see Fernand Braudel, Civilization and Capitalism 15th–18th Century, 3 vols., Siân Reynolds, trans. (New York, 1982-84), II, The Wheels of Commerce, 231-232.


Chance, Race and Class in Colonial Oaxaca (Stanford, 1978), 197.




Anderson, "Race and Social Stratification," 212.


Chance, Race and Class in Colonial Oaxaca, 194.


Ibid., 185.


Magnus Mörner, "Economic Factors and Stratification in Colonial Spanish America with Special Regard to Elites," HAHR, 63:2 (May 1983), 335-369.


McCaa, Schwartz, and Grubessich, "Race and Class," 421-433.


Anderson, "Race and Social Stratification," 213.


See, e.g., Archivo Municipal de Ponce (hereafter AMP), Caja 53-B, Padrón nominal … año de 1860.


For general statements on legal disabilities, see David Brion Davis, The Problem of Slavery in Western Culture (Ithaca, 1969), 281-288; Mörner, Race Mixture in the History of Latin America (Boston, 1967), 45; Herbert S. Klein, African Slavery in Latin America and the Caribbean (New York, 1986). Also of interest is Leslie B. Rout, The African Experience in Spanish America (Cambridge, 1976), 126-152. Many of the decrees are collected in Richard Konetzke, ed., Colección de documentos para la historia de la formación social de Hispanoamérica, 1493-1810, 3 vols. (Madrid, 1962). See, e.g., II, 108-109, 237, 334-335, 427-428, 543 and III, 185-187, 331-332, 553-573.

The historiography of the free people of color warrants a separate study. For the purposes of the present discussion, only a sample of what has been written will be noted. Luis M. Díaz Soler writes about Puerto Rico’s free people of color in Historia de la esclavitud negra en Puerto Rico, 3rd ed. (Río Piedras, 1970), 225-261. For Latin America, see Carlos Larrazábal Blanco, Los negros y la esclavitud en Santo Domingo (Santo Domingo, 1967); A. J. R. Russell-Wood, The Black Man in Slavery and Freedom in Colonial Brazil (New York, 1982); William F. Sharp, Slavery on the Spanish Frontier: The Colombian Chocó, 1680-1810 (Norman, 1976); Frederick P. Bowser, The African Slave in Colonial Peru, 1524-1650 (Stanford, 1974); Gonzalo Aguirre Beltrán, La población negra de México: Estudio etnohistórico, 2d ed. (Mexico City, 1972); George Reid Andrews, The Afro-Argentines of Buenos Aires, 1800-1900 (Madison, 1980); Carlos M. Rama, Los afro-uruguayos de Montevideo: El siglo ilustrado, 3rd ed. (Montevideo, 1969); and Patrick James Carroll, "Mexican Society in Transition: The Blacks in Veracruz, 1750-1830" (Ph.D. diss., University of Texas, 1975). Four important collections of essays are David W. Cohen and Jack P. Greene, eds., Neither Slave Nor Free: The Freedman of African Descent in the Slave Societies of the New World (Baltimore, 1972); Robert Brent Toplin, ed., Slavery and Race Relations in Latin America (Westport, 1976); Stanley L. Engerman and Eugene D. Genovese, eds., Race and Slavery in the Western Hemisphere: Quantitative Studies (Princeton, 1975); and Mörner, ed., Race and Class in Latin America (New York, 1971).


Noel Rivera Ayala, "Las milicias disciplinadas puertorriqueñas: Grandes períodos y el duradero valor de la institución (1765-1850)" (Master’s thesis, University of Puerto Rico, 1978), passim.


The cédula is reproduced in Boletín Histórico de Puerto Rico, I (1914), 297-307. For a useful discussion of caste and estate, see Gerhard Lenski, Power and Privilege: A Theory of Social Stratification (New York, 1966), 77-79.


Teun A. van Dijk, Prejudice in Discourse (Amsterdam/Philadelphia, 1984), 13. A standard study of prejudice is Gordon W. Allport, The Nature of Prejudice: An Analysis of Ethnic Prejudice in Cognition and Conversation (Garden City, NY, 1958).


Van Dijk, Prejudice, 16.


For instance, see Archivo General de Puerto Rico (hereafter AGPR), Ayuntamiento de San Juan, asuntos diversos, año 1818, leg. 21, "Expediente que contiene el informe sobre los antecedentes del alférez de fragata Dn. Juan Ortiz de Zárate" and AGPR, Gobernadores, Audiencia, año 1825, caja 292, "Expediente sobre el auto … declarando justa la oposición de D. Santiago Labiosa a prestar su consentimiento al matrimonio que su hijo D. José Bernardo trataba de contratar con la parda liberta María Teresa Masave."


Actas del Cabildo de San Juan de Puerto Rico (San Juan, 1945–) (hereafter ACSJ), vol. 1803-1809, pp. 222-225. See also 281-285.


Ibid., 72-76.


On racial nomenclature generally, see Mörner, Race Mixture, 56-60 and passim; Aguirre Beltrán, La población negra de Mexico, 159, 166-170, 172-173, 175; Ángel Rosenblat, La población indígena y el mestizaje en América, 2 vols. (Buenos Aires, 1954), II, 173-179; and Irene Diggs, "Color in Colonial Spanish America," The Journal of Negro History, 38:4 (Oct. 1953), 403-427.


"Bando del General Prim contra la raza africana," Cayetano Toll y Toste, Historia de la esclavitud en Puerto Rico (San Juan, 1981), 79-85, 89-91.


AMP, Caja 53-B, Padrón nominal del … año de 1860. The barrios studied were 2, 4, and 5. The other two were examined for occupational clustering, and there was none.


AGPR, Gobernadores, Municipales, San Juan, caja 563, entry 300, Padrón formado por el Comisario del Barrio del Quartel de Santa Bárbara en el presente año de 1823; Censo o padrón de los Yndividuos que habitan en el Barrio o cuartel [San Juan] en el presente año de 1828 ... ; Empadronamiento 1883-1849, Padrón del Barrio de San Francisco, año de 1833; Empadronamiento 1883-1849, Barrio de Santo Domingo. The censuses for both San Francisco in 1833 and Santo Domingo in 1846 are in one bound volume, the spine of which says Sto. Domingo. All further references to these censuses presume these citations.


Verena Martínez-Alier, Marriage, Class and Colour in Nineteenth-Century Cuba: A Study of Racial Attitudes and Sexual Values in a Slave Society (Cambridge, 1974), 93.


Ibid., 96.


To place these advantages in perspective, see the restrictions in the Código Carolino (Konetzke, ed., Colección de documentos, III, 553-573).


Unfortunately, the records for San Juan disappeared or were destroyed in a fire some years ago. Those for other parts of the island (AGPR, Tribunales) do not describe the participants by race, a healthy sign, but one which eliminates the possibility of examining these valuable records for racial patterns. I am indebted to Fr. Fernando Picó for information about the small claims courts of San Juan.


Examples of such access to a town council may be seen in ACSJ.


The civil right that might have agitated free people of color (and whites for that matter) to political activism, and then perhaps to class consciousness, suffrage was not an issue in Puerto Rico.


Notary records normally are an extremely important source of information about the lives of the general population, and, in fact, a considerable amount of valuable information was gleaned from them; however, scores of volumes in the AGPR have been taken out of circulation because of their fragile physical state.


See n. 26.


Pedro Tomás de Córdova, Memorias geográficas, históricas, económicas y estadísticas de la Isla de Puerto Rico, 6 vols., 2d ed. (San Juan, 1968), II, 307.


Córdova, Memorias geográficas, II, 13, 24-25, 397.


Adolfo de Hostos, Historia de San Juan: Ciudad murada, 1521-1898 (San Juan, 1966), 61.


Cited ibid., 79. Actually, Córdova used the terms cuadra and manzana interchangeably, observing that the streets were divided into "cuadras o manzanas …" (ibid.). See also the description by André Pierre Ledrú, Viaje a la Isla de Puerto Rico en el año 1787 (San Juan, 1971).


This is my calculation from the real property census of San Juan for 1820 (AGPR, Gobernadores, San Juan, 1816-20, caja 561). There is no single title for this census, which exists only in manuscript form. Each barrio has its own title page and title. Thus, the title for the barrio of San Juan is "Justiprecio de los Edificios existentes en el Barrio de S. Juan: Año de 1820." Santo Domingo is simply "Quartel de Santo Domingo," although a long descriptive title appears on the first page.

For examples of the use of the terminology of urban form, see David J. Robinson and Swann, "Geographical Interpretations of the Hispanic-American Colonial City: A Case Study of Caracas in the Late Eighteenth Century," in Latin America: Search For Geographical Explanations, Robert J. Tata, ed. (Chapel Hill, 1976), 1-15 and Swann, Tierra Adentro, 277.


I have based this information about the added barrios on the "Padrón o Estadística general de la riqueza urbana de la Capital—Año de 1859," AGPR, Gobernadores, leg. 123-A, pieza I. Hostos states that San Juan contained the barrios of Puerta de Tierra, La Puntilla, and San Sebastián in 1846, in addition to the historical four. By 1853, Ballajá and de la Marina were added (Hostos, Historia de San Juan, 79).


The map is found in AGPR, Archivo Fotográfico, caja/paq. 6, fol. 355. The approximate date of the map was confirmed in a personal communication to me from Aníbal Sepúlveda-Rivera. See María de Los Ángeles Castro, Arquitectura en San Juan de Puerto Rico (siglo XIX) (Río Piedras, 1980), 134.


See n. 39. The census of 1820 did not include several important public and church buildings—such as the cathedral, monasteries, and convents. It did include vacant lots—solares. It appears that all church-owned dwellings that were rented out were included. Valuations were either not given or were unknown in 14 instances. The old view that the elite resided in or around the main plaza of colonial towns, while accurate in many instances, has been proven incorrect in others (see Fred Bronner, "Urban Society in Colonial Spanish America: Research Trends," Latin American Research Review, 21:1 [1986], 7-72, 24-25 and Anderson, "Race and Social Stratification," 209-243).


AGPR, Gobernadores, Municipales, San Juan, 1816-20, caja 561, Padrón de los vecinos del Cuartel de Santa Bárbara. Año de 1818.


Leonard P. Curry, The Free Black in Urban America, 1800-1850: The Shadow of the Dream (Chicago, 1981), 24.


Ibid, and p. 56.


Jay Kinsbruner, "The Pulperos of Caracas and San Juan during the First Half of the Nineteenth Century," Latin American Research Review, 13:1 (1978), 65-85.


Valdes, "The Decline of the Sociedad de Castas," 114.


Ibid., 136-137.


Anderson, "Race and Social Stratification," 230.


Ibid., 215.


Chance, "The Ecology of Race and Class in Late Colonial Oaxaca," in Studies in Spanish American Population History, Robinson, ed. (Boulder, 1981), 93-117. Chance used a segregation index that considers several racial groups, thus his figures are not compatible with mine.


Chance, "Ecology of Race and Class," 96. For examples of the use of other variables in conjunction with place of residence, see Valdes, "The Decline of the Sociedad de Castas," 114-135 and Anderson, "Race and Social Stratification," 228.


For the segregation index I followed Karl E. and Alma F. Taeuber, Negroes in Cities: Residential Segregation and Neighborhood Change (New York, 1969), 28-35, 236-237, and passim. See also Curry, The Free Black, 54-55. The index is calculated as Segregation Index = 1/2(sumNi/N - Wi/W) where Ni = nonwhite at block or street level, N = nonwhite at barrio level; Wi = white at block or street level, W = white at barrio level


Taeuber and Taeuber, Negroes In Cities, 29.


See the comment on this coding procedure in Robinson, "The Analysis of Eighteenth-Century Spanish American Cities: Some Problems and Alternative Solutions" (discussion paper no. 4, Geography Dept., Syracuse University, 1975), 7 and 39, n. 27. See Sherburne F. Cook and Woodrow Borah, Essays in Population History, 2 vols. (Berkeley, 1971), I, 136-145, on how head of household was determined in eighteenth-century Mexico.


Both methods of coding produce unexpected results. In the alternative method, there are heads of households in their early teens and even younger. However, the use of standard coding produces similar results, for instance, a seven-year-old head of household. In fact, it may be that young people of tender age managed to live alone or were placed in their residences by a benefactor. Cook and Borah have noted that clerical enumerators in late eighteenth-century Mexico sometimes counted children above the age of 12 or 13 as heads of households (Essays in Population History, I, 138). A further refinement in coding procedures would be to code coresident siblings according to what appears to be their actual status. When the first named sibling was somewhat older than the others, or a person of some social or economic status, he or she might be coded the head of household. But when the siblings were of approximately the same age, especially when they were of the same sex, each might be coded as the head of a household, unless some mitigating information in the census suggested otherwise. Who is to say that these siblings had not simply joined together as renters of individual rooms in an apartment and maintained fairly discrete "households," while sharing some common amenities, which in any event might have been the reality among some renters of separate apartments? The Cambridge Group for the History of Population and Social Structure would have coded siblings living together without the presence of a parent or other relative from an earlier generation as a frérèche (see Peter Laslett, ed, Household and Family in Past Time: Comparative Studies in the Size and Structure of the Domestic Group Over the Last Three Centuries in England, France, Serbia, Japan, and Colonial North America [Cambridge, 1974], 30). The Cambridge Group considers "persons not evidently related" as members of a "no family" (ibid., 31).


Curry, The Free Black, 56.


Ibid. I have rounded his numbers to tenths.


Taeuber and Taeuber, Negroes in Cities, 39-41.


Valdes, "The Decline of the Sociedad de Castas," 114-135.


The barrio of Santa Bárbara contained 31 bohíos in 1823, but in that year the governor took steps to prohibit the future building of such shacks, and the town council soon ordered that all bohíos within the city’s walls be destroyed (Hostos, Historia de San Juan, 78.) These efforts seem to have been effective, since there were no bohíos in the barrio of San Juan in 1828, only one in San Francisco in 1833, and none in the first trozo of the barrio of Santo Domingo in 1846. The number of bohíos is from the census of the barrio of Santa Bárbara for 1823.


For all houses, the mean number of occupants per building was 5.7 for the barrio of Santa Bárbara; 8.9 for the barrio of San Juan; 6.9 for the barrio of San Francisco; and 6.9 for the first trozo of the barrio of Santo Domingo.


For discussions of the transition from the earlier single household to multiple household residential structures and the increase in rental housing units, see Robinson and Swann, "Geographical Interpretations," 1-15 and Swann, Tierra Adentro, 283-285.


There was also a group of people enumerated in the censuses as agregados. Scholars have characterized agregados variously as boarders, guests, or unpaid domestics. Although there sometimes were families among the agregados in the four San Juan barrios, in my definition they did not form separate households (and thus are not included in the tables and figures). In San Juan, agregados were relatives, guests, boarders, and, it seems safe to conclude, domestics, whether remunerated with wages or not. I believe we must consider some of the San Juan agregados as transients, and to a greater degree than among the general population.


Linda L. Greenow, "Spatial Dimensions of Household and Family in Eighteenth-Century Spanish America" (discussion paper no. 35, Geography Dept., Syracuse University, 1977), 10.




That it is extremely difficult to standardize terminology may be seen in Greenow’s discussion of the literature in "Family, Household and Home: A Micro-Geographic Analysis of Cartagena (New Granada) in 1777" (discussion paper no. 18, Geography Dept., Syracuse University, 1976), 17.


In Durango, New Spain, in 1778, 12.5 percent of residential units were owned by nonwhites, "and most of these [resident] household heads were mulattos" (Swann, Tierra Adentro, 301).


That so few white and free colored heads of residential units owned their dwellings in the first trozo of the barrio of Santo Domingo in 1846 is difficult to explain without knowing the value of housing at that late date. There is a real property census for the municipality of San Juan in 1859, which lists property values, but it is not divided into barrios (AGPR, Gobernadores, Municipales, Año de 1859, leg. 123-A, pieza I).


Fray Íñigo Abbad y Lasierra, Historia geográfica, civil y política de la isla de San Juan Bautista de Puerto Rico (first published 1831, 2d ed., San Juan, 1968), 94-95. Abbad’s description is repeated in Aníbal Sepúlveda-Rivera, "San Juan de Puerto Rico: Growth of a Caribbean Capital City" (Ph.D. diss., Cornell University, 1986), 174. I have both translated Abbad and used parts of Sepúlveda’s translation. See also the description of San Juan in 1797 by Ledrú, Viaje a la isla de Puerto Rico, 95 and Sepúlveda, "San Juan de Puerto Rico," 175-176.


The data in the first part of this paragraph are derived from the following documents: AGPR, Gobernadores, San Juan, Municipales, 1816-20, Real property census of 1820 and caja 561, "Padrón de … Santa Bárbara: Año 1818."


AGPR, Protocolos Notariales (hereafter PN), San Juan, caja 437, Will of Manuel Elías, Oct. 12, 1845, fols. 478-480v. See also Elías’s father’s will, caja 527, Apr. 7, 1824, fols. 279-281.


Property census of San Juan for 1820, AGPR, Gobernadores, San Juan, 1816-1820, caja 561.


AGPR, PN, caja 437, fols. 478v-480v.


Ibid., caja 439, Record of sale, Cruzado to González, June 22, 1847, fols. 422v-425v.


Ibid., caja 439, Record of sale, Cruzado to Ezguiaga, July 28, 1847, fols. 518-522. He also owned another single-storied stone house on La Luna, this one purchased in 1847 for 1,825 pesos (ibid., caja 530, Record of sale, Pozo to Cruzado, Sept. 10, 1847, fols. 200-200v and 204-205).


Ibid., caja 261, Record of sale, Cruzado to Cabrera, Feb. 4, 1854, fols. 52-54.


Ibid., caja 261, Notice of sales tax payment, Dec. 7, 1854, fols. 623-627.


Ibid., caja 528, Record of sale, Peña to Mediavilla, Mar. 4, 1841, fols. 90-93 (the second half of the volume); caja 480, Record of sale, Latorre to Peña, Mar. 23, 1832, fols. 178v-183v.


Ibid., caja 528, Record of sale, Cambián to Peña, Apr. 29, 1841, fols. 166v-169.


Ibid., caja 530 (tomo 1849), Record of sale, Peña to Ortiz de Zárate, Apr. 22, 1844, fols. 235-237.

Author notes


I wish to thank the City University of New York PSC-CUNY Research Award Program for grants which supported research in Puerto Rico. I would also like to thank the staff of the Queens College Academic Computer Center, in particular Richard White, who wrote the programs and worked closely with me in data analysis. John V. Lombardi read an early version of the article, Jay Gordon read several versions, and David J. Robinson and Michael M. Swann read the draft of the final version and, while perhaps not agreeing with all my views on coding, suggested ways to strengthen the article.