In recent years scholars have made impressive efforts to reconstruct the demographic characteristics of the population of pre- and post-Columbian Spanish America.2 But comparable progress has not been registered for Brazil where demographers, in contrast with historians, have displayed surprisingly little interest in the pre-independence centuries.3 Yet as long ago as 1938 Samuel H. Lowrie called attention to abundant census data housed in the archives of São Paulo concerning that state’s colonial and early national periods.4 But with few exceptions these records have remained unexploited, and the same is true of census reports for other parts of colonial Brazil. These records, which were apparently unknown to Professor Lowrie, are scattered in various archival and published collections. While quantitatively not as large as the demographic documentation on São Paulo, this material is sufficiently voluminous to permit a preliminary analysis of Brazil’s population in the late colonial period as to size, distribution, racial composition, and other characteristics. This paper, focused primarily on the years 1772-1782,5 is intended to illustrate the possibilities offered by this data and some of the problems that these sources pose.
Introduction
It has been suggested that the first census may have been made by a chieftain of some tribe who wanted to learn how many fighting men were at his disposal.6 One of the earliest recorded references to head counts appears in the Scriptures which tell of David’s injunction to Joab to “Go now through all the tribes of Israel, from Dan even to Beersheba, and number ye the people, that I may know the number of the people,” and of the angry deity’s displeasure when the Israelites allowed themselves to be numbered.7 Classical Athenians also recorded the number of children born each year, and upon reaching his eighteenth birthday each male citizen was registered for military service.8 Both during the Republic and the early Empire the Romans employed censuses for the apportionment of taxes and military obligations. After the Empire’s collapse the Church continued to keep partial statistics of the number of faithful in each parish. About the beginning of the fourteenth century9 secular governments in various parts of Europe began to make enumerations of their inhabitants for the purpose of levying assessments, especially poll and hearth taxes. But although the populations of cities and provinces were occasionally enumerated during succeeding centuries, no general census of large countries was successfully completed before the mideighteenth century.
The foundations of the modern census system were laid in 1749 when the government of Sweden directed church pastors to submit periodic10 reports giving births, marriages, and deaths of their parishioners according to age, sex, and occupation. Such data were consolidated and analyzed by a central Tabular Commission, and served as the basis for a periodic national census that has remained unbroken to the present. In time the Swedish example was emulated by the principal nations of Europe, including Norway and Denmark (1769), Spain (1787),11 France (1800), and England (1801).12
Population counts in colonial America long antedated the adoption of modern census procedures in Europe. As early as 1511 King Ferdinand ordered Governor Diego Columbus to report on the size of the Spanish population in the Indies, and a few years later the Council of the Indies began to call upon secular and ecclesiastical authorities in the New World to furnish numerical descriptions of the Spanish and Indian population in the new conquests. Throughout the next two and a half centuries a stream of such reports poured into the Peninsula, comprising a demographic record unmatched during this period in quantity or in abundance of detail by Portuguese Brazil13 or by any European nation with the possible exception of England. But such data always flowed across the Atlantic irregularly, and despite repeated reforms designed to improve the regularity and quality of the reportage, the Council was seldom able to obtain an accurate determination of the population of the entire Spanish empire at any one time.14 With the exception of the French province of Quebec, where periodic enumerations of a much smaller and less dispersed population began in 1665 and continued until the eve of the fall of New France,15 no successful attempts were made in the New World to collect population data on a recurring basis prior to the second half of the eighteenth century.
During the third quarter of the eighteenth century three major imperial powers in America introduced measures designed to provide the home governments with more accurate and more frequent demographic information concerning their colonies than had previously been available. The fact that each was engaged in an increasingly bitter rivalry for hegemony in the New World and required better information about the numbers of its colonists for purposes of increased taxation and military recruitment in part explains the reasons for the reforms.15a But they were also a reflection of enlightened governments’ and individuals’ preoccupation in the eighteenth century with measurements of all kinds, whether they involved the accurate determination of the longitude and latitude of key colonial cities, the maximum and minimum temperatures recorded in various parts of the empire, the elevations of mountains, or the heat of volcanos and bodies of water. It should be recalled too that during this period both Iberian powers tried (without notable success) to introduce double-entry bookkeeping systems in their colonial exchequers to improve the collection and dispersement of royal income. Both governments also established a series of so-called commercial balances in which the annual volume of trade between intra-colonial and Peninsular ports was itemized in great detail. Colonial administrators were also required to submit endless tables showing the numbers and kinds of troops garrisoned in various parts of the empire. In short, both for military and nonmilitary reasons the colonial powers were becoming increasingly statistics conscious.
The first to introduce standard census methods in its colonies was England. In 1761 the Board of Trade directed each of His Majesty’s governors in America to “from time to time give Us frequent & very full Information on the State and Condition of the Province under Your government.” The governors were to provide answers to a list of “queries” concerning the number of whites, blacks, and Indians, and the reasons for their increase or decrease during the previous biennium.16 Pursuant to this order, the governors of six of the thirteen mainland colonies submitted at least one general return between 1765 and 1776.17
In 1776 Spain and Portugal acted almost simultaneously to secure even more detailed information concerning the inhabitants in their colonies. The Spanish cedula of November, 1776, which seems to have been sent to all parts of the Indies, called upon secular authorities to submit to the Minister of the Indies annual padrones “with proper distinction of classes, [marital] status, and castes of all persons of both sexes, without omitting the infants” residing within their jurisdiction.18 While it proved beyond the capacity of the colonial bureaucracy to supply such data yearly, at least one comprehensive general census along the lines specified by the cedula was taken in New Spain, Guatemala, Nueva Granada, Peru, Chile, and in the Río de la Plata in the course of the next two decades.19
The Portuguese circular of May, 1776, extended to all parts of Brazil demographic procedures previously tried in Pará, Goiás, and São Paulo during the early ’seventies.20 They, in turn, may have been influenced by the techniques adopted for the Castilian census of 1768,21 for the Portuguese employed a similar system of age grouping. This classification, which was followed by some (though regrettably not by all) compilers of Brazilian censuses for the next quarter of a century, divided the population into the following classes:
Several features of this classification merit comment. First, unlike the coeval Spanish directive, the Portuguese circular did not require data on racial or caste status, though, as will be seen later, several governors included such information in their reports. Second, it is apparent that those who devised the classification not only recognized the fact that girls mature earlier than boys, but also (unchivalrous as it may seem!) the tendency of women to age more rapidly than men in Colonial America because of the rigors of child bearing and domestic labor.24 Third, the inclusion of Class Ix (annual births) caused confusion in the minds of some enumerators, since it duplicated data already recorded in Classes I and V (children under seven). As a result some compilers mistakenly added the figures for Class Ix to those for Classes I-VIII, and thereby inadvertently inflated their totals.25 On the other hand, at least one functionary failed to count children who had not reached their first birthday.26 Fourth, the inclusion of children under seven in the censuses required the tabulation of information not contained in the parish registers which did not list children below the confessional age.27 Presumably the purpose of Classes I and V was for projections of the colonial population in future decades, but one is at a loss to know why the king’s ministers insisted upon the names of all persons in each parish who had passed the ninetieth year, unless the Portuguese were simply following the practice of the Swedes, though both may have been inspired by the example of Roman enumerators during the late Republic who, according to Pliny, recorded the number of persons over one hundred years in parts of Italy.28
Three groups were excluded from the Portuguese censuses: regular troops (tropas pagas), ecclesiastics, and untamed Indians. The omission of the first two did not seriously affect the results, since their numbers were not appreciable.29 Though far more numerous, it was obviously impossible to enumerate unreduced Indians living in areas not effectively under Portuguese control, such as southern Espiritu Santo, western Goias, much of Mato Grosso and Rio Negro.30
Problems of Census-taking in Brazil
As a matter of fact royal officials found it difficult enough to tabulate the Christian population of Brazil. Many of their problems were the same that confronted enumerators in the Spanish and English colonies, such as the dispersion of a relatively small population over a large area, the hardships of travel particularly in remote districts, and the lack of special compensation for those charged with preparing the reports.31 If Portuguese governors did not have to cajole unfriendly colonial assemblies to co-operate in the preparation of the censuses, as did their English contemporaries,32 they did encounter passive resistance among colonials who regarded head counts as portents of future financial exactions and increased conscriptions of their sons for the unpopular militia.33 Such intransigence prompted the captain-general of Bahia de Todos os Santos to warn Salvadoreans in 1776 that their refusal to divulge the names, numbers, and ages of their children and slaves would render them liable to heavy fines and thirty days’ labor on the royal fortifications. Even so, he was obliged to post his warning a second time two years later.34
But even if Brazilians had been more co-operative than traditionally has been the case where censuses have been newly introduced, it is probable that the compilation of accurate returns would have exceeded the technical competence of the Portuguese bureaucracy, just as it did that of officials in the Spanish and English colonies and also those charged with the preparation of the early Federal censuses in the United States. In Brazil many—perhaps too many—hands were involved in the organization of the reports. As soon as the captains-general and the bishops received instructions to proceed with the stocktaking,35 they issued appropriate directions to their respective underlings. These included the governors of subordinate captaincies,36 circuit judges (ouvidores or juizes de fora) and commanders of the local militia (mestres do campo) in the secular branch, and various members of the episcopal hierarchy down to the level of the parish priest. Primary responsibility for the construction of the tables or mapas in the prescribed form37 fell upon the shoulders of the mestres do campo and the parish priests. Their data were compiled mainly from the lista de desobrigas, the register of persons who received communion at Easter. However, that list did not include children under seven, so that the local officials were obliged to determine their numbers by actual count or by estimates.
Since the secular and ecclesiastical government shared responsibility for the compilation of the census reports, the lack of symmetry between units of the two branches frequently complicated effective co-operation. In interior Minas Gerais, for example, the limits of the bishopric of Mariana were considerably smaller than those of the secular administration. As a result the parishes of Sapucai, Jacuí, and Cabo Verde pertained to the bishopric of São Paulo; those of São Romão and Paracutú were suffragan to the bishopric of Pernambuco; while the four parishes of the Minas Novas district belonged to the archbishopric of Bahia.38 Because of the involved correspondence that such arrangements necessitated, it is not surprising that the returns from such parishes often arrived late, if at all.
As the parish lists (mapas particulares) were completed, they were condensed and forwarded to the circuit judge of the appropriate comarca or county, who sent the subtotals for his district to the governor of the subcaptaincy for transmission to the captain-general. In the secretariat of the captain-general the sums for the various subdivisions of that government were entered in the appropriate columns of the master sheets which were supposedly scrutinized by the captain-general, the bishop, and other dignitaries. When the tables were found to be defective, particularly where local officials had departed from the prescribed system of classification, the mapas were sometimes returned to the parish for correction. More often, however, the master sheets were forwarded to Lisbon with an admission of their defects and assurances from the captain-general that the next reports would be free of such inadequacies.39 Sometimes a second mapa geral was forwarded the following year, but as Table I suggests, more often than not several years passed before the next census results were dispatched to the court.40
What the crown actually did with these laboriously compiled reports is unknown. It is unlikely that they were subjected to any sophisticated analysis, for Portugal, like Spain, lacked trained staffs of statisticians comparable to those in Sweden’s Tabular Commission. It seems probable that Peninsular officials were content to derive the obvious totals, and then filed the reports away to gather dust.
It must be conceded that the chaotic nature of many of the reports—their lack of uniformity, the frequency of gross errors in the recorded totals, obvious transpositions, and certain baffling lacunae— tends to discourage further examination. And yet, despite these shortcomings, a study of these materials sheds considerable light upon various features of Brazil’s population in the late colonial years. Before proceeding with some examples, it is necessary to discuss briefly the sources which serve as the basis for the tables which follow and the administrative units that the censuses represent. In the following section the sources are described according to the geographical location to which they pertain, and are identified by letter and number corresponding to the full citations provided in the appended bibliography.
The Sources
(1) Captaincy-general of Grāo Pará.
Brazil’s northernmost captaincy-general included the old captaincy of Pará and the newer captaincy of Rio Negro situated in the Upper Amazon The latter was established in 1755, but as the Sketch Map and Table II indicate, its non-Indian population was still very small two decades later. The only demographic evidence found for the two units of Grão Pará during the period covered by this study are two detailed summaries of 1772 and 1773 (A-l and A-2) sent to the Colonial Minister in 1774.40a
(2) Captaincy-general of Maranhao.
By the mid-1770’s the century-and-a-half-old State of Maranhāo had ceased to exist, its place being taken by the captaincies-general of Grão Pará and Maranhāo. However, the latter included only the senior captaincy of Maranhāo and the sparsely occupied subordinate captaincy of Piauí which had been created in 1759. The earliest demographic data found for both units (B-l) are contained in an archival copy of the captain-general’s letter of transmittal of the census of 1777, but that dispatch gives only the total number of each sex living in the two parts of the captaincy-general. A later source (B-2) provides a parish-by-parish breakdown of the population by race and civil status but curiously fails to distinguish it by sex.
(3) The northeastern captaincy-general of Pernambuco embraced the subordinate captaincies of Paraiba, Rio Grande do Norte, and Ceará, as well as the two comarcas of Pernambuco itself. The latter included settlements in Itamaracá to the North and those in Alagoas to the South. The census of 1782 (C) is in some ways the most complete demographic record available for any part of Brazil, and was included in a long economic report sent to the crown in that year. The data appears to derive mainly from the mid-1770’s, though the statistics for Ceará are clearly dated 1765. While no later censuses for the entire captaincy-general have been found, two have been discovered for Ceará (D-1 and D-2).
(4) Captaincy-general of Bahia.
For more than two centuries Bahia, or Salvador, served as the colonial capital of Brazil, but even after it had lost that honor to Rio de Janeiro, Salvador remained the administrative headquarters for the captaincy of Bahia de Todos os Santos and the subordinate captaincies of Sergipe, Ilhéus, Porto Seguro, and Espíritu Santo.41 Within the space of two decades four counts were made of the captaincy-general’s inhabitants (E-1 to E-4), each varying considerably in method of presentation and in the totals reported. The last of this series, that taken in 1780, appears to be the most complete, and therefore has been used where possible in this study.
(5) Captaincy-general of Rio de Janeiro.
The titular capital of Brazil since 1763, Rio de Janeiro served as the administrative center of the captaincy of Rio de Janeiro, the subordinate captaincies of Santa Catarina and Rio Grande de Sao Pedro, and the distant Platine outpost of Colonia do Sacramento. Presumably because of his preoccupation with the Luso-Spanish War of 1773-1777,42 the Marquis of Lavradio, Viceroy of Brazil, was unable to complete a count of the inhabitants within his captaincy-general before his retirement in 1779.43 However, his successor, Dom Luis de Vasconcelos e Sousa, included a recently completed census as part of his extremely rich statistical report on the captaincy-general in 1781 (F-1). No later censuses for eighteenth-century Rio de Janeiro have so far turned up, except for one of the capital city in 1799 (F-2).
Viceroy Vaseoncelos’ report of 1781 does not include demographic information on the southern extensions of the captaincy-general. No entirely satisfactory document on the population of Santa Catarina has come to light. The earliest known census is a rather detailed one for 1753 (G-l) which gives the total population of the island and of the adjacent mainland as 6,336 communicants, or approximately 7,180 inhabitants. In a report prepared fourteen years later (G-2) the governor of Santa Catarina stated that he had “almost 3,000 paisanos capable of bearing arms,” of whom one-third were pretos (blacks) and pardos (mulattoes). At the time that the island fell to the forces of Don Pedro de Cevallos (February 22-28, 1777), the victors reported that “according to the parish registers,” there were 6,000 inhabitants on the island, but made no estimate of those living on the mainland of Santa Catarina (G-3a). Another Spanish document of the same year (G-3b) puts the Portuguese garrison, including both regulars and militiamen, on the island at 2,962. On the basis of this and much additional information garnered from administrative correspondence of the 1770’s too extensive to cite here, I have estimated the total civilian population for insular and mainland Santa Catarina in the mid-1770’s at about 10,000, a figure that appears reasonable when compared with two later counts (G-4 and G-5) which give the sub-captaincy’s population as 20,916 in 1788 and as 23,865 in 1796.
In the late 1770’s the half-century-old Portuguese foothold in Rio Grande do Sul, or Rio Grande de São Pedro as it was called then, was limited to the periphery of the Lagoa dos Patos and the lower Jacuí-Guaíba valley as far as the fortified settlement of Rio Pardo, with a narrow extension running south to the Camaquam river. The earliest known census for the captaincy (H-1) dates from 1780 and was appended to a map of the captaincy completed just after the end of the long struggle between Portugal and Spain for mastery of the eastern part of the captaincy. Two later censuses, one for 1798 and the other for 1802 (H-2 and H-3) add much detail that is lacking in the first census.
The southernmost outpost of Brazil in the 1770’s was the famous smuggling entrepôt of Colônia do Sacramento which definitively passed into Spanish hands in June, 1777. A count of the civilian population at that time (I) revealed only 535 persons, including slaves, priests, artisans, and a surprisingly large number of women, living at the base.
(6) Captaincy-general of Sāo Paulo.
São Paulo, the mother of half a dozen captaincies, suffered repeated partitions during the course of the eighteenth century, and at one point (1748-1765) actually ceased to be an independent administrative unit. In the year of its restoration as a captaincy-general (1765), the first of two house-to-house surveys of the capital city was made and was followed by the second two years later (J-1 and J-2). Between 1772 and 1800 no fewer than fifteen censuses were taken in the captaincy-general (J-3 to J-17), far more than are known for any other part of Brazil during this time. Whether Paulista governors were especially assiduous in their reporting or whether their mapas have simply been better preserved (or calendared) than those for other parts of Brazil is difficult to say. Unfortunately the published references to most of these censuses fail to indicate adequately their contents.
(7) Captaincy-general of Minas Gerais.
The mineral-rich captaincy-general of Minas Gerais, focal point of Portuguese settlement in the Brazilian interior during the eighteenth century, is represented by three censuses taken between 1776 and 1800 (K-1 to K-3). The first appears to be the most complete, but unfortunately the contemporary copyist made several crucial transpositions which reduce the value of the source, since the editor of the published text was unable to correct the scribe’s mistakes and the original source has not been found. Neither this nor any other known enumeration provides us with a breakdown of the captaincy-general’s population by districts smaller than its four comarcas.
(8) Captaincy-general of Goiás.
Three reports are known for Goiás between 1779 and 1804 (L-1 to L-3). Of the first two we have only the governors’ letters of transmittal which merely state the total number of inhabitants in the captaincy-general. The much later census of 1804 is more helpful, but it, too, appears to be incomplete, since it reports a total population markedly lower than that given in the two earlier sources. It is, of course, possible that Goiás actually lost population during the last quarter of the eighteenth century, contrary to the general upward trend throughout Brazil, but the question remains why?
(9) Captaincy-general of Mato Grosso.
Like Goiás, Mato Grosso was a raw frontier area in the eighteenth century, and after the exhaustion of the early mining strikes, there was little to attract non-military persons to this remote, swampy land where, as in western Goiás, warlike Indians still vigorously contested Portuguese rule. Only two vilas or towns are listed in the census of 1780 (M-1), the other settlements being customs cheek points (registos), military forts, and secularized Indian communities.44 In contrast with Goiás, however, the census of 1800 (M-2) reveals that a substantial increase in Mato Grosso’s population occurred during the last twenty years of the century. Then, according to two later reports (M-3 and M-4), it leveled off for at least the next two decades.
Distribution
Table II and the Sketch Map45 show the approximate distribution of Brazil’s enumerated inhabitants during the years 1772-1782. Where the censuses list the number of persons by parishes (freguesias) and it has been possible to determine their location, such clusters have been indicated on the map by circles whose diameters vary according to the size of the cluster. Approximately half of the reports do not break down the population by localities, so that it has been impossible to show such concentrations for those areas; but where other contemporary sources furnish the names and locations of places actually occupied during this decade, they have been denoted by crosses on the Sketch Map to provide some indication of the pattern of settlement at this time.46
The methods employed in the construction of Table II require explanation. Where the totals in the sources are confirmed by the various subtotals, they have been entered as given, but where they are inconsistent with the values represented by the subtotals and the latter appear accurate, the corrected sum has been placed in parenthesis. In three eases—those of the captaincies of Rio de Janeiro, Rio Grande de São Pedro, and Mato Grosso—it is evident that the sources do not include complete enumerations of all age groups. Children under seven were omitted in the Rio Grande and Mato Grosso censuses of 1780, while the Rio de Janeiro count for the same year fails to list minors, i.e., males below the fifteenth year and females under fourteen. It is possible, however, to extrapolate the missing data by determining from other sources what percentage of the population fell within these age limits in other parts of Brazil at this time. In the seven areas for which such information is available, the censuses show that about 11.75% of the population was seven years of age or less, and that 20.7% was between zero and fourteen (females) or fifteen (males) years (see Table III). Even though these percentages appear considerably lower than might be expected for these age groups,47 they are surprisingly uniform for the seven captaincies, and I am inclined to believe that the extrapolated figures more accurately resemble the correct sums than do the original totals. Accordingly, the population of Rio Grande de Sâo Pedro and Mato Grosso have been adjusted upward by 11.75% and that of Rio de Janeiro by 20.7%. It is apparent that these adjustments do not change the ranking order of any captaincy.48
The Sketch Map and Table II suggest certain observations regarding the distribution of Brazil’s population in this period. It is apparent that most of the colony’s enumerated inhabitants continued to concentrate in the river valleys and harbors of the littoral, thus ignoring the lament of the seventeenth-century chronicler Frei Vicente do Salvador, who scorned the proclivity of the Portuguese for “living like crabs along the coast” of Brazil.49 Moreover, in spite of repeated land “rushes” during the eighteenth century, first to the mining zones of Minas Gerais, Goiás, and Mato Grosso, and later to the southern pastoral lands of Santa Catarina and Rio Grande de São Pedro, the old sugar and dyewood captaincies of the Northeast, first settled during the second quarter of the sixteenth century, continued to retain a large proportion of Brazil’s inhabitants. The censuses of the 1770’s and early ’eighties reveal that 38.8% of the population lived in the captaincies-general of Bahia and Pernambuco, compared with 20.5% residing in Minas Gerais. On the other hand, about 14% lived within the confines of the captaincy of Rio de Janeiro, and less than 16% dwelt in the captaincy-general of which it was the principal part.50 São Paulo, today Brazil’s most populous state, ranked only fifth in the 1770’s. The Sketch Map also indicates the pattern of isolated settlements and so-called “hollow frontiers” which have typified Brazil from its colonial beginnings to the present.
It is interesting to compare the calculations of Brazil’s population derived from sources used here with several contemporary assessments of the late eighteenth and early nineteenth centuries. Table IV shows three estimates for 1776: the first two are the unadjusted and the adjusted totals recorded in Table II, while the third is the calculation which the learned abbot Correa da Serra sent to Alexander von Humboldt when the latter was preparing his demographic studies of the New World.51 Writing apparently during the first decade of the nineteenth century, this distinguished Portuguese scientist52 declared that he had been permitted to examine the results of “the last state [of the population]” of Brazil sent by its bishops to the Board of Conscience and Orders in Lisbon “about the year 1776.” He added that this “state” showed a population of “somewhat more than 1,500,000 souls,” a figure that incidentally is remarkably close to the ones given here. Yet the abbot’s statement is puzzling for several reasons: first, except for the archbishopric of Salvador, there is no mention in any of the high-level administrative correspondence that I have examined of a general episcopal survey being taken in Brazil during this decade; second, the abbot seems to have been unaware of the fact that the crown was dissatisfied with the results of previous ecclesiastical censuses and therefore directed secular authorities in Brazil to take charge of future enumerations;53 and third, the mapas were forwarded not to the Board of Conscience but to the office of the Colonial Secretary.54 In any case Humboldt’s correspondent felt that the estimate derived from the alleged episcopal counts was too low since it did not include children below the confessional age or unbaptized Indians. Accordingly, he concluded “without fear of exaggeration” that the true population of Brazil in 1776 was “nearly 1,900,000 souls.”
Several estimates have been made of Brazil’s population around the turn of the century.55 Of the four indicated for 1798, the lowest is a projection of our adjusted estimate for the years 1772-1782 based upon an assumed natural increase of 1.18% per annum, a figure suggested by an analysis of vital statistics contained in the sources for the earlier years.56 However, because of their fragmentary nature and because even our adjusted estimate for the base period probably does not account for ten to twenty per cent of the population, our estimate for 1798 is very likely too low. The next highest, that of 3,250,000, is supplied by another contemporary churchman, Santa Apolónia,57 while the highest is that of the savant Corrêa da Serra. The abbot advised Humboldt that a general census of Brazil had been undertaken in 1798, and that while its results had not been officially announced, information based upon the count published in French periodicals had led him to believe that the actual population of Brazil was then about 4,000,000.58 For reasons that he did not make explicit, Humboldt scaled down the abbot’s figure to 3,800,000.59
Four millions is also the total reported by the Minister of War in 1808 upon the completion of a general census of Brazil by the newly arrived government of John VI. The details of that census have never been revealed, but one student considers the reported figure to have been a deliberate exaggeration intended to facilitate the crown’s efforts to increase the size of the colonial militia, and contrasts it with the much lower sum included in an anonymous report dated some years later which gives Brazil’s population in 1808 as only 2,473,641.60
Finally, in the last years of the colonial era two further estimates were made of the number of persons living in Brazil. The more conservative of the two is that of the French geographer Adrien Balbi, author of the well-known Essai statistique sur le royaume de Portugal et d’Algarve,61 a work undertaken for the Portuguese government which afforded Balbi access to official sources, so that he may have examined the latest available mapas from Brazil, though he does not expressly say so. One who did was Antônio Rodrigues Velloso de Oliveira who in 1819 prepared for the crown a report concerning the ecclesiastical reorganization of Brazil. In it he included statistics on the population of each captaincy based upon mapas compiled mainly between 1814 and 1817. Because of what he considered to be their notorious deficiencies, the author arbitrarily added 25% to the reported totals, and concluded that Brazil then numbered about 4,396,123 persons.62
It is evident that our estimates are considerably below those suggested by other sources. The latter are, however, undocumented, and with the exception of the report of Velloso de Oliveira, they do not particularize as to how the population represented by their assessments was distributed in Brazil. It is also apparent that there is a general correspondence in the order of magnitude between our calculations and those derived from other sources. Nevertheless, because of the probability that the enumerators of the ’seventies and early ’eighties underestimated those in the lower age groups, our estimates for c. 1776 and 1798 probably should be raised by ten to twenty per cent. This would make Brazil’s population in the ’seventies range between 1,710,720 and 1,866,240, the latter figure being quite close to the estimate of the abbot Corrêa da Serra. The revised estimates for 1798 would then be from 2,188,596 to 2,387,559. Until more evidence is available for the last years of the eighteenth century, these seem to be quite reasonable estimates of Brazil’s population during these years.63
Racial Composition
As previously noted, the royal circular of 1776 did not inquire into the caste or racial composition of Brazil’s population.64 Nevertheless, some of the governors, notably those who had reason to be concerned about the large numbers of unfree Negroes confined to their jurisdictions, directed their subordinates to include such information in their reports.65 In such cases the population was segregated into three or four primary racial strains: brancos, meaning whites or at least persons socially accepted as Caucasians ; pardos or mulattoes;66 pretos or Negroes; and (Christian) Indians. Interestingly enough, no special category was reserved for mestiços who presumably were grouped under brancos or pardos depending upon their physical appearance.67
Table V indicates the racial composition of slightly over half of Brazil’s enumerated inhabitants during the 1770’s. Two of the largest slaveholding areas, the captaincies-general of Bahia de Todos os Santos and Pernambuco, are unreported. But included are two northern captaincies-general, Pará and Maranhão, both leaders in the agricultural revival of Brazil during the second half of the eighteenth century and heavy importers of African slaves;68 three central governments, Rio de Janeiro, São Paulo, and Minas Gerais; one southern captaincy, Rio Grande de São Pedro; and the western captaincy-general of Mato Grosso. Because of transpositions in the published source, the numbers of brancos and pardos listed for Minas Gerais can be only approximated. The bracketed totals for Maranhão, São Paulo, and Mato Grosso represent extrapolations from censuses taken between 1797 and 1800. Some error is doubtless inherent in our assumption that the racial proportions in the 1770’s were the same as they were at the end of the century, but the possible distortion would appear to be minor.69
The same table shows the civil status of the inhabitants of four captaincies for which direct information on racial composition is lacking. Since Indian slavery was abolished in the 1750’s,70 it may be assumed that all persons listed as slaves were pardos or pretos. It is also likely that the majority of those counted as freemen were considered whites. Surprisingly few Indians are tabulated in the censuses, and most of those enumerated lived in the Amazon where Jesuits, Franciscans, and others had proselyted since the early seventeenth century. The high proportion of slaves to freemen in Rio de Janeiro is noticeable; undoubtedly the majority of the population of Minas Gerais, Bahia de Todos os Santos, and Pernambuco were captives at this time, though specific evidence on this point is unavailable. The presence of a higher proportion of slaves in pastoral Rio Grande de São Pedro than in agricultural Pará comes as something of a surprise, though the ratio between freemen and captivos in the southern captaincy appears to have remained about the same for at least the next four decades.71
As in Spanish America, some Negroes and mulattoes were members of the free population of eighteenth-century Brazil. Toward the end of the century one contemporary declared that out of 3,250,000 persons living in Brazil, 406,000 or 12.4% were “Negros libertos,” as compared with 1,582,000 or 49.4% of the population who were slaves.72 The censuses of 1798 and 1802 for Rio Grande de São Pedro— the only eighteenth-century reports which supply this kind of information—indicate that while over half of the pardos were freemen, only about 5% of the pretos enjoyed that status (see Table VI). On the other hand, the census of 1815 (M-3) reveals that in Mato Grosso where, in contrast with Rio Grande, the majority of the population was wholly or partly Negro, nearly four-fifths of the pardos and almost one-third of the pretos were libertos.73 Though specific evidence is lacking, it is likely that manumission was more prevalent in the great plantation captaincies-general of Pernambuco and Bahia de Todos os Santos than in Brazil’s Far South where slavery had been introduced more recently and involved far fewer numbers.74 The extent to which manumission was practiced in different parts of Brazil at this time and its effects upon the social structure of the colony deserve to be studied.
Additional Data
As Table I indicates, the censuses also furnish information on a variety of other aspects of Brazil’s population. For example, twelve of the reports tabulate the number of births and deaths recorded in the captaincy during the year covered by the census. Ten supplement the answers required by the royal circular by showing the number of hearths or family units (fogos), a vestige of earlier methods of computing population. The ratios of hearths to population for these captaincies are computed in Table VII. It is apparent that the percentages for the four subdivisions of the captaincy-general of Pernambuco are considerably below the average, but the statistics include only communicants (desobrigantes de Quaresma); consequently the omission of minor children depresses the ratios. On the other hand, a survey taken of the civilian population of Colonia do Sacramento on the eve of that bastion’s fall in 1777 yielded the following:75
This is not to suggest that Colonia’s population necessarily typified that of Brazil generally, but rather that the praça’s narrow confines permitted a more accurate enumeration than was possible in such large, sparsely occupied captaincies as Grão Pará and Mato Grosso where the size of the family units recorded seems excessively high. While it would appear that the average Brazilian family numbered close to five persons, the validity of this assertion remains to be tested by the discovery of further evidence.
As noted earlier, the principal feature of the method of enumeration introduced in Brazil in 1776 was the classification of the population by age and sex. Eight of the sources for the years 1772-1782 separate the population according to the prescribed form, but with the exception of minor children, it is impossible to compute the ratio of males to females with reasonable accuracy, since the age-bars for the two sexes are not set at comparable years.76 In later years some enumerators continued to classify their inhabitants according to the method established in 1776, while others, for seemingly arbitrary reasons, adopted different age separations.77 The early censuses are therefore not as helpful in determining the sex composition of Brazil as might be expected.
The general censuses do not provide information on a number of important demographic considerations, such as the per cent of Peninsular-, foreign-, and native-born persons living in Brazil, the ethnic origins of Indian or African elements, or the occupations of the colonists. However, two early censuses of municipal São Paulo (J-1 and J-2) do list the occupations of members of that community, and it is likely that such data can be found for other urban populations as well.78
Conclusion
A satisfactory social history of colonial Brazil cannot be written until we have a clearer understanding of the demographic characteristics of the peoples who lived there. We have too long ignored the value of the colonial censuses, particularly those of late-colonial Brazil. Despite their shortcomings, they furnish a good deal of useful information which cannot be gleaned from any other type of sources. As such distinguished historians as Francisco A. Encina and Rudolfo Barón Castro have written with respect to the coeval Spanish colonial censuses,79 they provide us with more complete and more reliable statistics on the population of late-colonial Latin America than do the writings of chroniclers, foreign travelers, and other contemporaries whose more or less informed guesses have too often commanded more respect from scholars than they deserve. And as this essay has tried to show, the censuses reveal tentative answers to such questions as the size, distribution, and racial composition of Brazil’s population during the last colonial decades.
Undoubtedly the discovery of much corroborative evidence—additional census records, parish registers, episcopal surveys, registers of arrivals of slaves and freemen in Brazil, militia and tax rolls—would make possible a more thorough study of Brazil’s population during these years than has been possible here. Such materials await the investigator in the archives of Portugal and Brazil, and if this preliminary survey stimulates their exploitation, it will have served its major purpose.
APPENDIX Bibliography of Demographic Sources
A-1. “Mappa de todos os habitantes que existem nas freguesias da capitanía do Rio Negro ao primeiro de Julho de 1772. Extracto,” Instituto Histórico e Geográfico Brasileiro, Rio de Janeiro. Arquivo Ultramarino collection of transcripts (hereafter cited IHGB/AUC) 1-1-3, fo. 335v.-336r.
A-2. “Mappa de todos os habitantes . . . da capitania do Pará ao primeiro de Julho de 1772. Extracto,” ibid., 1-1-3, fo. 333r.-335r.
B-1. Joaquim de Melo e Povoaz to Martinho de Melo e Castro, May 7, 1778, ibid., 1-1-5, fo. 81r.-v.
B-2. “Mappa dos habitantes que existem na capitanía do Maranhão no anno de 1798,” Biblioteca Nacional, Rio de Janeiro (hereafter cited BNRJ) I-7,4,30, n.1.
C. “Mappa do total de todos os habitantes eomprehendidos nas quatro capitanias . . . de Pernambuco, extrahido das relações dos parochos em o anno de 1782,” Anais da Biblioteca Nacional XL (Rio de Janeiro, 1918), 98-110 (hereafter cited ABNRJ).
D-1. Luiz Barba Alardo de Menezes, “Memoria sobre a capitania do Ceará,” Revista do Instituto Histórico e Geográfico Brasileiro (hereafter cited RIHGB), XXXIV (1871), 276, table 3.
D-2. Antônio Jozé da Silva Paulet, “Deseripção geográfica abreviada da capitanía do Ceará,” ibid., LX:1 (1897), 75-102.
E-1. José Antônio Caldas, “Noticia geral de toda esta capitania da Bahia desde o seu descobrimento até o prezente anno de 1759,” Revista do Instituto Geográfico e Histórico da Bahia, No. 57 (1931), pp. 38-41.
E-2. “População da capitania da Bahia, em janeiro de 1775,” annex to Manuel da Cunha Menezes to Melo e Castro, Mar. 3, 1775, ABNRJ, XXXII (1914), 288-289.
E-3. Estimate of the Archbishop of Salvador for his diocese, as quoted in Martinho de Melo e Castro, “Instrução para o Marquez de Valença,” Sept. 10, 1779, ibid., 437.
E-4. “Mappa da enumeração da gente e povo desta capitania da Bahia . . . ,” Dec. 5, 1780, ibid., 480.
F-1. “Mappa geral das cidades, villas e freguezias que formão o corpo interior da capitanía do Rio de Janeiro,” (1780) RIHGB, XLVII:1 (1884), 27-29.
F-2. “Resumo total da população que existia no anno de 1799, comprehendidas as quatro freguezias desta cidade do Rio de Janeiro . . .,” ibid., XXI (1858), facing p. 238.
G-1. “Mappa das freg.as q. tern a Ilha de Santa Catarina e seu continente,” Arquivo Histórico Colonial, Lisbon. Papeis avulsos. Bancroft Library Microfilm Collection (hereafter cited AHU/PA), Santa Catarina, caixa 2.
G-2. Francisco de Sousa de Menezes to Conde de Azambuja, “Informação sobre a ilha de Santa Catarina,” Dec. 8, 1767, BNRJ, 7,3,47 (orig.).
G-3(a). “Diario de las operaciones de el excercito de S.M.C. en la America meridional a las ordenes de el Thiente Guerra Dn Pedro de Ceballos” (1777), British Museum, ADD MSS 6893, fo. 57r. (Bancroft Library Microfilm Collection).
G-3(b). “Noticia de las tropas que tenia la Isla de Sta Catalina para su defensa al tiempo que vino sobre ella el Exmô S. Dn Pedro de Ceballos” (1777), British Museum, Egerton MS No. 374 (Bancroft Library Microfilm Collection).
G-4. “Ilha de S.ta Caterina, 1788-89-90 1788 [sic],” AHU/PA, Santa Catarina, caixa 2.
G-5. “Dados estatistieos sobre . . . Santa Catarina, 1797,” BNRJ, II-35,30,3, n. 2.
G-6. [Manuel] Aires de Casal, Corografia brasílica, I (Facs. ed., Rio de Janeiro, 1945), 192-193.
H-1. “Mappa das freguesias e moradores de ambos os sexos com declaração das diferentes condições e idades eom que se achão em 7 de outubro de 1780,” BNRJ, I-5,4,9, fo. 106r.; also printed in Jônatas da Costa Rego Monteiro, Dominação espanhola no Rio Grande do Sul, 1763-1777 (Rio de Janeiro, 1937), Appendix 23.
H-2. “Mappa ou numerario de todos os habitantes da cap.a do R.o Grande d. Sm. Pedro do Sul . . . no anno de 1798,” AHU/PA, Rio Grande do Sul, caixa 1.
H-3. “Mappa de todos os habitantes da capitania do Rio Grande de Sâo Pedro do Sul . . . no anno de 1802,” loc. cit.
I. “Familias que ficaram na praça da Colônia, quando esta se rendeu aos Espanhões na Guerra de 1777 . . .,” Arquivo Nacional, Rio de Janeiro, Col. 67, Liv. IV, fo. 24r.-27r.
J-1. “Listas da gente que compreende a ci[da] de de S[ão] Paulo e todo o seu termo . . .,” Documentos interessantes para a historia e costumes de São Paulo (hereafter cited DI), LXII (São Paulo, 1937), 7-256.
J-2. “Lista de todos os povos . . . do destrito desta ci[da]de de S[ão] Paulo . . .,” ibid., 257-360. For an analysis of J-1 and J-2 see Amador Florence, “Coisas de velhos censos paulistas,” IV Congreso de História Nacional, Anais, V (Rio de Janeiro, 1950), 85-245.
J-3. Cited without source by P. Geraldo José Pauwels in Revista do Arquivo Público do Rio Grande do Sul, No. 23 (Porto Alegre, 1930), p. 501 note.
J-4. “Relação de todos os habitantes da capitania de São Paulo . . .,” in Martim Lopes Lobo de Saldanha to Martinho de Melo e Castro, Mar. 2, 1776, DI, XXVIII (1898), 267-271; repr. in “Catálogo de documentos sobre a história de S. Paulo existentes no Arquivo Histórico Ultramarino de Lisboa,” RIHGB, Tomo Especial, VII (1957), 104-105. (This series hereafter cited RIHGB/TE plus volume and year).
J-5. “Relação . . . dos habitantes da capitania de São Paulo dividos nas [dez] classes . . .,” (1780), ibid., 358-359.
J-6. “Relação de todos os habitantes . . . da capitanía de São Paulo . . .,” (1781) ibid., 405.
J-7. “Relação de todos os habitantes desta capitanía [de São Paulo] no anno de 1782,” DI, XXXI (1901), 106-109.
J-8. Francisco da Cunha e Menezes to Melo e Castro, July 2, 1784, ibid., 129; repr. without details in RIHGB/TE, VIII (1958), 102.
J-9. Idem to idem, May 13, 1785, ibid., 137.
J-10. Frey Jozé Raymundo Chichorro da Gama Lobo to idem, Mar. 2, 1788, DI, XXXI, 143; repr. without details in RIHGB/TE, VIII (1958), 235.
J-11. Bernardo José [Maria] de Lorna [e Silveira] to idem, May 20 1790, RIHGB/TE, VIII, 331-332.
J-12. Idem to idem, Mar. 22, 1791, ibid., 376-377.
J-13. “Relação, feita pelo secretário do govêrno [da capitanía de Sào Paulo],” Dec. 31, 1792, ibid., IX (1958), 18.
J-14. “Relação, feita pelo secretário do govêrno [da capitania de São Paulo],” Dec. 31, 1793, ibid., IX (1958), 47.
J-15. “Relação dos habitantes da capitania de São Paulo . . .,” Dec. 31, 1794, ibid., 62.
J-16. Mappa geral dos habitantes da capitanía de S. Paulo no anno de 1797,” DI, XXXI (1901), 151-155, 157; summarized in RIHGB/TE, IX (1958), 251-252.
J-17. “Mappa dos habitantes que existem na capitania de São Paulo no anno de 1800 . . .,” n.d., ibid., X (1958), 300.
K-1. “Mappa dos habitantes actuães da capitania de Minas Gerais e dos nascidos e falecidos no anno de 1776,” Publicações do Arquivo Nacional, IX (Rio de Janeiro, 1909), 73. The subtotals of this census are also given in José João Teixeira Coelho, “Instrução para o govêrno da capitanía de Minas Gerais,” RIHGB, XV (2d ed., 1888), 260-280.
K-2. Antonio Noronha to Melo e Castro, Jan. 28, 1778, BNRJ, 2,2,24, n. 125, indicating that the mappa of 1777 was enclosed but omitting details.
K-3. Diogo Pereira Ribeiro de Vasconcellos, “Breve descripção geográphica, physica, e política da capitania de Minas Gerais,” (1806), Revista do Arquivo Publico Mineiro, VI (Belo Horizonte, 1901) 821-823.
L-1. Luiz da Cunha Menezes to Melo e Castro, July 8, 1780, IHGB/AUC, 1-2-7, fo. 245r-246r.
L-2. Idem to idem, ibid., fo. 263v-265v.
L-3. Luiz Antônio da Silva e Sousa, “Memoria sôbre o descobrimento, govêrno, população, e cousas mais notáveis da capitanía de Goiás,” (1812), RIHGB, XII (1874), 482-494.
M-1. “Relação de toda a povoação das capitanias do Matto Grosso e Cuyabá . . . [n] o anno de 1780,” IHGB/AUC, 1-2-5, fo. 78v.-80v.
M-2. Caetano Pinto de Miranda Monte Negro to Viseonde de Anadia, April 17, 1802, enclosing the mappa for 1800. RIHGB, XXVIII: 1 (1865), 125-127.
M-3. “Mappa da população da capitania de Matto-Grosso em o anno de 1815,” ibid., XX (1857), n.p. (after 292).
M-4. “Descripção estatistica da eapitania de Matto-Grosso no anno de 1817,” loc. cit.
This study is a by-product of research undertaken during the past five years concerning the administrative history of eighteenth-century Brazil, and I am grateful to the following for financial assistance at various times during these years: The Henry and Grace Doherty Charitable Foundation, Inc., New York, N. Y.; the Social Science Research Council; the University of Washington (Faculty Research Committee of the Graduate Division); and the American Philosophical Society. I also wish to thank Professors Woodrow Borah and Engel Sluiter of the University of California (Berkeley) for helpful criticisms of an earlier draft of this paper, and my wife, Alice Alden, who prepared the Sketch Map under my direction. Necessarily, I am solely responsible for the contents of this study.
I refer especially to the studies of Woodrow Borah, Shelburne F. Cook, George Kubler, José Miranda, and Lesley Byrd Simpson. For an indication of recent progress in pre- and post-Columbian demography consult the appendices of Ángel Rosenblat, La población indígena y el mestizaje en América (2d ed., Buenos Aires, 1954), I, passim.
Cf. Gigorio Mortara, “Demographic Studies in Brazil,” in The Study of Population. An Inventory and Appraisal, eds., Philip M. Hauser and Otis Dudley Duncan (Chicago, 1959), pp. 235-248, especially the bibliography on pp. 246-248. The earliest Brazilian demographic study which included estimates of Brazil’s population from c. 1770 to 1870 was by the poetaster Joaquim Norberto de Souza Silva in “Investigações sobre os recenseamentos da população geral do imperio e de cada provincia de per si tentados desde os tempos coloniães até hoje,” Relatório do Ministério dos Negocios do Império, 1870. Annex D (Rio de Janeiro, 1870). Neither Souza Silva nor Montara seems to have known of the colonial censuses described below.
“Bibliographical sources concerning population statistics in the state of São Paulo. . .,” Handbook of Latin American Studies, 1937 (Cambridge, Mass., 1938), pp. 490-501.
This period was originally selected because of my interest in determining the population of Brazil during the vice-regency of the Marquis of Lavradio (1769-1779). As will be seen from Table I below, Brazil as a whole is particularly well represented by the demographic records available for this decade.
A. M. Carr-Saunders, Population (London, 1925), p. 3.
2 Samuel, ch. 24; 1 Chronicles, ch. 2. I am indebted to my colleague, Professor Howard Kaminsky of the University of Washington, for having called my attention to these references.
Marcel R. Reinhard and André Armengaud, Histoire générale de la population mondiale (Paris, 1961), p. 31.
The Norman Doomesday Book (1086) and the first Venetian census (sometime after 1268) considerably antedate the establishment of secular enumerations in other parts of Europe.
The reports were submitted annually from 1749 to 1751, triennially from 1754 to 1772, and quinquennially from 1772 on. For a discussion of the development of census procedures in Sweden see Edvard Arosenius, “The History and Organization of Swedish Official Statistics,” in John Koren, ed., The History of Statistics . . .(New York, 1918), pp. 537-552.
A census of the kingdom of Castile was taken in 1768.
For descriptions of the evolution of the modern census see W. S. Rossitir, ed., A Century of Population Growth . . . 1790-1910. U. S. Department of Commerce. Bureau of the Census (Washington, 1909), introduction; J. A. Davies and [Walter F. Willcox], “Census,” The Encyclopedia Britanica, V (11th ed., New York 1910), 662-669; W. F. Willcox, “Census,” Encyclopaedia of the Social Sciences, III (New York, 1930), 295-300; A. M. Carr-Saunders, World Population; Past Growth and Present Trends (Oxford, 1936), ch. 1; Roger Mois, S.J. Introduction a la démographie historique des villes d’Europe du xIVe au xVIIIe siècle (3 vols., Louvain, 1954), I, 16-70; and Frank Lorimer, “The Development of Demography,” in Hauster and Duncan, pp. 124-179.
Population data on Brazil before the second half of the eighteenth century are extremely meager. For some early estimates made between 1570 and 1612 see Engel Sluiter, ed., “Report on the State of Brazil, 1612 . . .,” HAHR, xxIx (Nov., 1949), 520-521.
For a valuable introduction to census procedures, problems, and sources in colonial Spanish America to c. 1776 see Richard Konetzke, “Las fuentes para la historia demográfica de hispano-américa durante la época colonial,” Anuario de estudios americanos, V (Sevilla, 1948), 267-324, especially 289ff.
For a defense of Canada’s claim as “the first country in modern times to take a complete Census,” and for a description of census techniques in the seventeenth-century province of Quebec see A. J. Pelletir, “Canadian Censuses in the 17th Century,” in Papers and Proceedings of the Annual Meeting of the Canadian Political Science Association, II (Ottawa, 1930), 20-34.
Eris O’Brien has perceptively observed that “Depopulationism was a widespread delusion in the eighteenth century. ‘It came to be believed, with a fervency that bordered on the fanatical, that a nation’s strength depended entirely on the number of its inhabitants,’ in much the same way as the bullionists linked national welfare with gold and silver and the new industrialists thought factories and blast-furnaces the only possible gauges.” The Foundation of Australia (1786-1800). A Study in English Criminal Practice and Penal Colonization in the Eighteenth Century (2d ed., London, 1950), p. 111.
Circular of April 28, 1761 in J. H. Benton, Jr., Early Census Making in Massachusetts 1643-1765 (Boston, 1905), pp. 27-30. The oldest census in the Thirteen Colonies was that of Virginia (1624-25).
The standard, but unimaginative, compilation of census data for the English mainland colonies is Evarts B. Greene and Virginia D. Harrington, American Population before the Federal Census of 1790 (New York, 1932). For returns gathered in accordance with the circular of 1761 see Massachusetts (1765), p. 21; Connecticut (1774), p. 58; Rhode Island (1767, 1774), pp. 67-69; New Hampshire (1767, 1773, 1775), pp. 74-79; New York (1771), pp. 102-103; and New Jersey (1772), p. 112. See also pp. 4-18 for general estimates prior to 1790. According to Rossitir (p. 3) “no enumeration embracing all of the colonies was ever made, and in some of the colonies no accurate count . . . occurred during the entire colonial period.”
For the text of the cedula see José de Gálvez to Juan José de Vértiz, Nov. 10, 1776, Documentos para la historia Argentina, x, Padrones de la ciudad y campaña de Buenos Aires (1726-1810), Ricardo R. Caillet-Bois ed., with introduction by Emilio Ravignani (Buenos Aires, 1920-1955 [.sic]), 71. The order is referred to in several of the sources in note 19.
The first completed general census of New Spain was evidently that of 1793, although portions of an earlier census of 1777-1778 have survived. See Sherburne F. Cook, “The Population of Mexico in 1793,” Human Biology, xIV (Dec., 1942), 499-515; for details concerning the Guatemalan census of 1777-1778 see Rudolfo Barón Castro, La población de El Salvador (Madrid, 1942), ch. 8; there is a summary of the census of 1778 in Nueva Granada in Francisco Javier Vergara Velasco, Nueva geografía de Colombia (Bogotá, 1892), pp. DCLxII-DCLxIII; the earliest general census of Peru seems to be that of 1792 (Archivo General de las Indias, Estado, legajo 75.) I wish to thank Professor James F. King of the University of California for kindly permitting me to examine a photostatic copy of this report. The accompanying covering letter [Fray Don Francisco Gil de Toboada y Lemos] to Conde de Aranda, Nov. 5, 1792, states “hace muchos años que no se hacía el Padrón grãl, si es que alg.a ves se ha completado aquí esta obra, porque yo à lo menos no he podido hallarla.. . .” This report was evidently unknown to George Kubler when he prepared The Indian Caste of Peru, 1795-1940. A Population Study Based upon Tax Records and Census Reports. Smithsonian Institution. Institute of Social Anthropology. Pubn. 14 (Washington, D. C., 1952). For references to the first general census of Chile (1778) see Diego Barros Arana, Historia general de Chile (2d ed., Santiago, 1932), VI, 400-401, VII, 337; cf. Francisco A. Encina, Historia de Chile desde la prehistoria hasta 1891, IV (Santiago, 1945), 617. A summary of known and lost censuses of Buenos Aires is given in Nicolás Besio Moreno, Buenos Aires puerto del Río de la Plata capital de la Argentina. Estudio crítico de su población 1536-1936 (Buenos Aires, 1939), ch. 7. There appears to have been no complete census taken in the Platine viceroyalty prior to 1797 (Rosenblat, La población indígena, I, 205) but for Montevideo (1780) see Documentos para la historia Argentina, xII (Buenos Aires, 1910), 389, and for the city and campaña of Buenos Aires (1778), ibid., xI (1919), passim, xII (1919), 11-120. For reference to the “Estado grãl en extracto de la población y producciones de la Provincia de Venezuela, formado por D. Josef de Castro y Araoz . . .” (1787) see James F. King, “A Royalist View of the Colored Castes in Venezuela, 1815,” HAHR, xxxIII (Nov., 1953), 528 n. 4. Such a summary was unavailable when the French agent, François Raymond Joseph de Pons searched for demographic evidence in the archives of Caracas, and therefore he was obliged to depend upon defective parish registers for his estimate of the population of the captaincy-general. See his A Voyage to the Eastern Part of Terra Firma, or the Spanish Main . . . during the years 1801 . . . [to] 1804, tr. [Washington Irving], I (New York, 1806), 102-103, 105.
The procedures for compiling censuses in Goiás are detailed in Martinho de Melo e Castro to José de Almeida e Vasconcelos, Oct. 1, 1771, Revista do Instituto Histórico e Geográfico Brasileiro (hereafter cited RIHGB), xxVII:2 (1864), 234-259; for pre-1776 censuses of Pará and São Paulo see Table I and Appendix.
According to Barón Castro, the age grouping used in the Castilian census of 1768 was “de menos de 7 años, de 7 a 16, de 16 a 25, de 25 a 40, de 40 a 50 y de más de esta edade.” He adds, “Tiene este recuento indudable importancia para nuestro objecto, por el hace de preceder inmediatamente al ordenado para América en 1776, y es incuestionable que las experiencias adquiridas en él debieron utilizarse para la formación del primero del Nuevo Mundo.” La población de El Salvador, pp. 165-166.
The term rapariga, a dictionary synonym for mofa, is popularly reserved for prostitutes in Brazil today.
“Circular do Ministro . . . dos Negocios da Marinha, Martinho de Melo e Castro, na qual se contem diversas instruções para a organização do recenseamento da população,” May 21, 1776, Anais da Biblioteca Nacional (hereafter cited ABNRJ), xxxII (Rio de Janeiro, 1914); Melo e Castro to Manoel da Cunha e Menezes (governor of Bahia), same date, Biblioteca Nacional, Rio de Janeiro (hereafter BNRJ), 11-33,25,17 (orig.) ; idem to Marquis of Lavradio, same date, ibid. 1-2,4,7, n. 24; idem to Martim Lopes Lobo de Saldanha (governor of São Paulo), same date, Publicação official de documentos interesantes para a história e costumes de São Paulo (hereafter cited DI), xLIII (São Paulo, 1903), 65-66.
Cf. Gilberto Freyre, “Some Aspects of the Social Development of Portuguese America,” in Charles C. Grifan, ed., Concerning Latin American Culture (New York, 1940), pp. 96-97.
E.g., source J-4 cited in the Appendix.
E.g., source H-3, ibid.
João Pereira Caldas (captain-general of Pará) to Reverendo Senhor Vigario da Freguesia de—, Dec. 6, 1772, Instituto Histórico e Geográfico Brasileiro, Rio de Janeiro, Arquivo Ultramarino collection of transcripts (hereafter cited IHGB/ AUC) 1-1-3, fo. 334r.-v. This circular was sent to each vicar in Pará reminding him that while the parish register did not record children below the confessional age, such information was to be included in the census reports.
A number of the. Portuguese censuses include such information. That of Pernambuco (source C in the Appendix) even specifies both those over ninety and those past 100. For similar enumerations by the Romans see J[osiah] C[ox] Russell, Late Ancient and Medieval Population. Transactions of the American Philosophical Society, N.S., xLVIII:3 (Philadelphia, 1958), 48. Swedish practice is described by Arosenius, “History and Organization of Swedish Official Statistics,” p. 540.
Most of the regulars were stationed in the captaincy-general of Rio de Janeiro in the 1770’s. Their paper strength in 1776 was approximately 10,000, but their effective strength was considerably less. See “Quadro das forças de mar e terra existentes nas capitanias do Rio de Janeiro, Santa Catharina, Rio Grande, Minas Geraes, e na praça da Colônia, disponsiveis para a defesa da fronteira do sul em 1776,” RIHGB, xxI (1858), 181-189.
Contemporary estimates as to their numbers average about 800,000, but are mere guesswork.
Such complaints are mentioned in Joao Pereira Caldas to Melo e Castro May 15, 1773 and Feb. 14, 1774, IHGB/AUC, 1-1-3, fo. 317v.-318r„ 333r-v; Joaquim de Melo e Povoaz (captain-general of Maranhão), to idem, Sept. 27’, 1776, and May 7, 1778, ibid., 1-1-5, fo. 45v-46r., 81r-v; Luis da Cunha Menezes (captain-general of Goiás) to idem, July 8, 1780, and Aug. 9, 1781, ibid., 1-2-7, fo. 245r-246r, 263r.-v; and in Luis de Albuquerque de Melo Pereira e Caceres (captain-general of Mato Grosso) to idem, May 20, 1781, ibid., 1-2-5, fo. 78r. Cf. Rossitir, A Century of Population Growth, p. 3.
See Francis Bernard (governor of Massachusetts) to Board of Trade, Sept. 5, 1763, Benton, Early Census Making in Mass., p. 54 for an example of how the General Court frustrated the governor’s effort to answer the Board’s interrogatory.
Thus a long-time resident of Salvador, Luis dos Santos Vilhena, professor of Greek, wrote “O fazer hum mapa desta natureza neste paiz não hé tão facil como talvez se suponha, porque os paes de familias receozos de que Ihes peção tilhos para soldados não só oecultão muitos, como nem dão os nomes nos róes de confição e o mesmo praticão como os escravos, receozos de alguma capitação ou tributo, segundo o numéro de escravos que constar possuem.” Recopilação de noticias soteropolitanas e brasílicas contidas em xx cartas . . ., II (anno de 1802), (Bahia, 1921), 481. See also Marquis of Lavradio to Luis de Vasconcelos e Sousa, June’19, 1779, RIHGB, IV (2d ed., 1863), 440. For an example of the persistence of distrust of official enumerators by rural folk in modern Brazil see Marvin Harris, Town and Country in Brazil (New York, 1956), p. 207. In other parts of Latin America, too, enumerators still encounter strong local opposition to head counts. Recently in the Ecuadorian province of Tungurahua census-takers were met by rock-hurling Indians, and in the ensuing melee six persons were killed and twenty injured. The San Francisco Chronicle, Nov. 27, 1962.
“Proclamações do capitão general da Bahia Manoel da Cunha e Menezes, sobre o recenseamento da população” Bahia, Aug. 12, 1776 (orig.) and copies dated April 3, 1778, BNRJ, II, 34, 3, 3. The bando also lists and defines the ten classes for which data was being collected.
“Officio circular que Martinho de Melo e Castro dirigiu a todos os prelados de ultramar . . .,” May 21, 1776, ABNRJ, xxxII, 445; Melo e Castro to [D. José Joaquim Justiano Mascarenhas Castelo Branco (bishop of Rio de Janeiro)], same date, BNRJ, I-2,4,7, n. 25, gives the text which is lacking in the above-cited calendar.
In the past the bishops of Brazil were required to submit estimates of the population of their dioceses to the Board of Conscience and Orders in Lisbon, so it was appropriate for them to be ordered to assist in the preparation of the new enumerations. [José Francisco] Correa da Serra to Alexander von Humboldt, n.d. (c. 1804), Political Essay on the Kingdom of New Spain . . . by Alexander von Humboldt, tr., John Black, IV (London, 1814), 333; see also José Manuel Pereira da Silva, História da fundação do império brasileiro, I (Rio de Janeiro 1864), 206.
By the 1770’s Brazil contained nine captaincies-general and an equal number of subordinate captaincies (capitanias subalternas).
Copies of the circular of May 21, 1776, giving the ten groupings of the population were sent to each militia captain and parish priest.
José João Teixeira Coelho, “Instrução para o governo da capitania de Minas Gerais,” (1780), RIHGB, xV (1852), 271. Antônio de Noronha (captain-general of Minas Gerais) to Sr. Fr. Manoel, bishop of São Paulo, Dec. 31 1776 BNRJ, 2,2,24, n. 34.
There is some evidence that local officials tampered with the statistics by inflating or deflating them to protect their own jobs. See Francisco Alberto Rubim, “Breve noticia statystica da capitania do Espirito Santo . . .,” (1816), Publicações do Arquivo National, xIV (Rio de Janeiro, 1914), 110, and Antônio Rodriguez Velloso de Oliveira, “A igreja do Brasil, ou informação para servir de base a divisão dos bispados, projectada no anno de 1819 . . .,” RIHGB, xxIx:1 (1866), appendix 6. How extensive such practices were is not known.
In addition to sources already cited, the three previous paragraphs are based upon Manuel da Cunha e Menezes to Melo e Castro, Aug. 2, 1776, BNRJ, 1-4,3,16, n. 90; Antônio de Noronha to the ouvidores of Minas Gerais, Nov. 8, 1776, BNRJ, 2,2,24, n. 27; idem to Ignacio Jozé de Alvarenga, ouvidor geral of Minas Gerais [late 1777], ibid., n. 107, and materials cited in note 31.
Since writing the above I have discovered a mapa of 1775 for Rio Negro, prepared by the ouvidor e intendente geral Francisco Xavier Ribeiro de Sampaio, which gives the subcaptaincy’s population as 1,129. “Diario da viagem que em vizita, e correição das povoações da capitania de S. Josè do Rio Negro fez o . . . Francisco Xavier Ribeiro de Sampaio no anno de 1774-1775,” Anais da Biblioteca e Arquivo Publico do Pará, VI (Pará, 1907), appendix 6 (following p. 118). In addition, the “Catálogo da correspondencia dos governadores e capitães generaes do Pará com o governo da metropole,” indicates the dispatch to Lisbon of mapas for 1774, 1775, 1776, 1777, and 1778. Ibid., V (1906), 337, 338, 341, VI, 253, 256.
Espiritu Santo was the military and fiscal responsibility of the captain-general of Bahia, but under the civil and ecclesiastical jurisdiction of Rio de Janeiro.
D. Alden, “The Undeclared War of 1773-1777: Climax of Luso-Spanish Platine Rivalry,” HAHR, XLI (Feb., 1961), 55-74.
The relações parciães of the military captains of Rio de Janeiro to Viceroy Lavradio appended to his memoir to his successor are very incomplete and clearly were not compiled in accordance with the circular of 1776. They are published in RIHGB, LXXVI:1 (1913 [1915]), 289-355.
The census of 1817 [M-4] indicates that the same pattern of settlement existed at that time.
The Sketch Map has been adapted from a map entitled “Vicereinado do Brazil, 1763 . . .,” in Ministério da Agricultura, Indústria e Comércio. Directoria Geral de Estatistica. Recenseamento do Brazil realizado em 1 de Setembro de 1990, I (Rio de Janeiro, 1922), between pp. 352 and 353. We have superimposed this upon a modern map of Brazil drawn to a slightly larger scale. See “Mapa política do Brasil,” Instituto brasileiro de geografía e estatística. Atlas do Brasil (geral e regional) (Rio de Janeiro, 1959), p. 136. While the sources of the 1763 reconstruction are not indicated, much of its information can be confirmed (or corrected) from contemporary materials, and it has been used in the absence of any suitable published contemporary map for the eighteenth century. Boundaries between some of the captaincies have been intentionally left open, since it is impossible to determine them with any exactitude. Indeed, contemporary officials, like some of their modern counterparts in Brazil, spilled a great deal of ink disputing the limits of their jurisdictions.
For this purpose the following have proved helpful: Francisco Xavier Machado, “Memória relativa às capitanias do Piauhy e Maranhão” (1810), RIHGB, XVII (1854), 57-69; Antônio Jozé da Silva Paulet, “Descripçâo geográfica abreviada da capitanía do Ceará,” (c. 1816), ibid., LX:I (1897), 75-102; Marquis of Lavradio to Luis de Vasconcelos e Sousa, June 19, 1779, ibid., IV, 409-486; Sebastião Francisco Bettâmio, “Notícia particular do continente do Rio Grande do Sul” (1780), ibid., XXI (1858), 239-299; “Relação dos cargos da capitania de Minas Gerais” (1806), Revista do Arguivo Público Mineiro, VI (Belo e a praça de Santos (1783), ibid., Tomo Especial, VIII (1958), 71; Diogo Pereira Ribero de Vasconcellos, “Breve descripção geográphica, physica e política da capitanía de Minas Gerais” (1806), Revisa do Arguivo Público Mineiro, VI (Belo Horizonte, 1901), 179-794; José Joaquim da Rocha [?], “Geographia histórica . . . de Minas Gerais,” Publicações do Arguivo National, IX (1909), 81; Luis Antonio da Silva e Sousa, “Memoria sobre o descobrimento, governo, populaçâo, e cousas mais notaveis da capitania de Goiás” (1812), RIHGB, XII (2d ed., 1874), 482-494; [anon.], “Descripção geográphica da capitania de Matto-Grosso: anno de 1797,” ibid., XX (1857), 185-270. The most useful secondary account is Caio Prado Júnior, Formação do Brasil contemporáneo colonia (4th ed., São Paulo, 1953), pp. 29-64, and especially the bibliography at the end.
It is evident that the percentages of the total population represented by these age groups vary considerably among different populations. J. C. Russell finds that about half of the population of ancient and medieval Europe was under the age of twenty and one-third below age fourteen or fifteen. Ancient and Medieval Population, p. 143. On the other hand, according to Father Mois during the years 1770-1790 about 17.3% of the population of the city of Venice fell between zero and nine years, while 33.7% was between zero and nineteen. Introduction a la démographie . . . des villes d’Europe, III, 237, table 13. However, Cook found that in New Spain about 32% of the population was between zero and seven years, while about 55% was under sixteen. “The Population of Mexico in 1793 . . .,” p. 506, fig. 2.
There is also the possibility that the sources for Minas Gerais, Maranhão, Piauí, and Goiás also omitted part or all of the minors, for the covering letters of the captains-general do not make clear whether all classes of the population were enumerated.
Cf. Caio Prado’s comments concerning the disequilibrium between the settlement of the littoral and of the interior in the eighteenth century. Formaçâo, pp. 33ff.
Based upon adjusted total. Ceará not included.
Humboldt, Political Essay on New Spain, IV, 333-335.
For the abbot’s career as scientist, statesman, author, co-founder and perpetual secretary of the Academia Real das Sciencias in Lisbon and corresponding member of leading scientific societies in Europe and the United States see Innocêncio Francisco da Silva, Diccionário bibliográphico portuguez . . ., IV (Lisbon, 1860), 336-341.
For contemporary criticism of the episcopal censuses see Martinho de Melo e Castro to Marquis de Valença, Sept. 10, 1779, ABNRJ, XXXII, 437; also circular of May 21, 1776 cited in note 23.
After the arrival of the crown in Brazil the mapas were sent to the Intendencia Geral da Polícia.
I have not included here the calculation for 1780 suggested by Felix Contreiras Rodriguez, Traços de econômia social de politico do Brasil colonial (Rio de Janeiro, 1938), pp. 33-34, for it is based upon published sources such as Robert Southey’s totally unreliable guesses for the period.
I.e., the number of births in excess of deaths reported in sources A-1, A-2, B-1, C, E-4, J-4, J-6, and J-16.
As quoted without citation in Arthur Ramos, Las poblaciones del Brasil (Mexico, 1944), p. 119.
See note 51.
Political Essay, IV, 323.
Francisco José de Oliveira Vianna [?], “Resumo histórico dos inqueritos censitários, realizados no Brazil,” Recenseamento do Brazil . . . em 1920, I, 403; [anon.] “Memória estatística do império do Brasil,” (c. 1829), RIHGB, LVII:1 (1895), 93-99, which also includes figures on Brazil’s population in 1823.
(2 vols., Paris, 1822). See II, 229 and 229 n. 3.
“A igreja do Brasil,” 178-179 and appendices 1-7.
Oliveira Vianna [?], “Resumo histórico dos inqueritos censitários,” 413-414; for other estimates made between 1776 and 1872 see pp. 406-411 of the same work.
The Spanish seem to have been unique in distinguishing fractional racial permutations among their colonial inhabitants. Cf. Herbert Moller, “Sex Composition and Correlated Culture Patterns of Colonial America,” The William and Mary Quarterly, 3rd ser., II (April, 1945), 119.
E.g., Antônio de Noronha to the ouvidores and capitães mores of Minas Gerais, Nov. 8, 1776, BNRJ, 2,2,24, n. 27.
For a contemporary definition of the term pardo see Diogo Pereira Ribeiro de Vasconcellos, “Breve descripção . . . de Minas Gerais,” 777.
See note 69.
Caio Prado Júnior, Historia econômica do Brasil (5th ed., São Paulo, 1959), pp. 81-94; D. Alden, “Manuel Luís Vieira: an Entrepreneur in Rio de Janeiro during Brazil’s Agricultural Renaissance,” HAHR, XXXIX (Nov., 1959), 521-537.
In one instance where it is possible to make some sort of check, the extrapolations seem reasonably close. In a note appended to the census of 1780 the governor of Mato Grosso observed that “trez quartes partes ou mais aínda do referido numero total dos habitantes . . . são negros, mulatos, ou outros mestiços das muitas differentes especies que ha nestas paizes. . . .” The census of 1800 records 84.2% of the population of Mato Grosso as non-white. Sources M-1 and M-2.
By the laws of June 6, 1755, and May 8, 1758, which affected the States of Maranhao and Brazil respectively. Mathias C. Kiemen, “The Indian Policy of Portugal in America, with Special Reference to the Old State of Maranhao, 1500-1755,” The Americas, VI (April, 1949), 452-453.
Cf. the mapa of 1818 of Rio Grande do São Pedro in Velloso de Oliveira, “A igreja do Brasil,” appendix 2. It is interesting to note that in the neighboring captaincy of Santa Catarina there were 5,191 slaves (a percentage of 21.7) out of a population of 23,865 in 1796. Source G-5.
See note 57.
The exact figures are: brancos, 5,813 (21.3%); pardos, 9,477 (34.8%); number of free pardos, 7,908 (79.4%); pretos, 11,975 (43.9%); number of free pretos, 3,356 (28.0%).
Cf. the adjusted tables for the years 1814-1820 in Velloso de Oliveira, “A igreja do Brasil,” appendices 1, 2, and 5.
Source I. Figures do not include priests, widows, or bachelors.
Some of the returns present additional problems. Those for Pernambuco (C) define adult women (Class VII) as between 15 and 50 years, instead of 15 and 40 as the circular prescribed, while the censuses of 1798 and 1802 (H-2 and H-3) for Rio Grande de São Pedro include children between one and seven years in Classes I and V, instead of zero and seven as the circular specified.
E.g., the Maranhão census of 1800 (B-2) separates the population 0-5, 5-10 . . . 85-90, while the Mato Grosso census of 1815 (M-3) distinguishes only three age groups: 0-15, 15-45, and over 45. The Rio Grande de São Pedro mapa of 1802 (H-3) divides whites 1-7, 7-15, 15-60, and over 60, but does not break down the ages of non-whites. On the other hand, Paulista censuses until at least 1800 were organized in conformity with the royal circular of 1776.
See for e.g. the calendar notation in ABNRJ, XXXII, 290, item 8751 concerning a report on the parish of Nossa Senhora da Penha in the city of Salvador (Bahia) for 1775 which gives the modo de vida of heads of families in that freguesia. The 1799 census of Rio de Janeiro (F-2) does not report the occupations of the Cariocas, but the source in which it is published furnishes the names of merchants and the number of each type of craftmen, shopkeepers, and taverners in the city. See Antonio Duarte Nunes, “Almanac histórico da cidade de S. Sebastião do Rio de Janeiro,” (1799), RIHGB, XXI (1858), 167-171.
Encina, Historia de Chile, IV, 617; Barón Castro, La población de El Salvador, pp. 230-233.
Author notes
The author is assistant professor of history at the University of Washington.