Growth, Stagnation, and Migration: An Explorative Analysis of the Tributario Series of Anáhuac (1720-1800) Open Access

How to analyze the demographic patterns in eighteenth-century rural Central Mexico is a question that has occupied me during my research into its agrarian history. Most social and economic historians of the region would acknowledge that their work is still considerably hampered by the lack of demographic data. One almost never finds a direct answer to the questions of how many persons lived in a certain town or village, province, or district; the published parochial and census data must be considered nothing more than rough estimates. Even more challenging is the reconstruction of the precise rhythm and subregional differentiation of demographic developments during the eighteenth century. In the long run parish data may supply the missing links; in the short term the office of the tribute collectors seems to be the best place to begin building a framework for future local or parish studies. For this reason I have chosen to fill out the printed record of census data with new tributario figures.

Although the calculations based on tribute materials do not capture demographic patterns in non-Indian groups, they are in line with estimates from other demographic studies and confirm the general impression of growth, with possible extreme periodic and regional variations. Rapid growth in some central valleys and a few downslope areas seemed to coincide with stagnation around the city of Puebla. Where previous research has emphasized epidemics as the motive force behind these eighteenth-century developments, I hypothesize that migration is another factor influencing provincial and subregional fluctuations in the number of tributaries, and that urbanization and rural industrialization may have contributed to the loss of Indian tributaries in the villages. These possibilities remain hypothetical until the potential variant of differential natural increase can be removed on the basis of new parochial data.¹ The heuristic value of these hypotheses may be that, in light of regional differences, the future study of demographic material from some specific parishes will be more valid.¹

The term “Central Mexico” generally refers to a region stretching from coast to coast—from Veracruz to Acapulco and from Tampico to Mazatlán—and might in the context of this article cause misunderstanding. To avoid possible confusion I use the designation Anáhuac, as referring to a region formed by the three valleys of the central highlands, 2000-2600 meters above sea level (see Map 1). The valleys of Toluca in the west, Mexico in the center, and Puebla in the east were enclosed by rugged and impenetrable mountains, creating a degree of isolation with respect to the other regions in the viceroyalty of New Spain. The mountain area is generally referred to as faldas (slopes), and although considerably fewer people lived there than in the highlands, the faldas were economically and socially integrated in the life of the highlands. The highlands contained the two large cities, Mexico City and Puebla de los Ángeles, as well as a few smaller provincial towns like Toluca, Tepeaea, Tlaxcala, and Texeoco; the level of urbanization was remarkably high. The region consisted, according to my classification, of fifty-one provinces, most of them alcaldías mayores, most of which became subdelegaciones after the administrative reform of 1786-87. It was a densely populated area accommodating some 60 percent of the population of New Spain.³

Growth or Stagnation?

Population data in colonial Mexico are in general of three kinds: (1) census data of several parishes or provinces; (2) data based on parochial registration; and (3) tax rolls used by collectors of Indian tribute. Most of them are incomplete. There were always people who escaped registration and officials who reclassified people, who made different classifications per province, who copied previous counts or simply counted wrong. And, of course, one struggles with the thorny problem of the “coefficient”: most researchers, for example, always put the number of tributarios—tribute payers—at between four and five members per family head, without investigating the changes in family size through time. Conversion factors are variable, not constant.⁴ Although all this is bound to arouse suspicions about the overall rhythm of any curve derived from population data, as well as about the competence of cross-checking methods at the provincial level, it should not prevent us from trying. My purpose is to verify the general conviction of population growth in the region of Anáhuac, on the basis of published materials. Few data exist at present for the region in general. More is known of Oaxaca, the Mixteca Alta, Yucatan, Michoacán, or Guadalajara. Even the calculations of Sherburne Cook and Woodrow Borah cannot provide a detailed view of demographic development in the central highlands.⁵

A country’s exact population can only be determined by means of a census, but no population list will be totally accurate. (Even contemporary censuses of modern industrialized nations err. Noble David Cook reminds us of the weakness of the U.S. census of 1980, which suffers from underenumeration in major cities.) Three colonywide censuses were made in eighteenth-century New Spain: the so-called Fuenclara census of 1742-46, the Aranda census of 1776, and the Revillagigedo census of 1790-94. Besides these censuses researchers have at their disposal the very detailed Matrícula de tributarios of 1805, with date of count and a partial breakdown by age and marital status. For each area the Matrícula shows the count for indios de pueblo, indios laboríos y vagos, negros, and mulatos, but no Spaniards or other groups that did not belong to the República de Indios. According to Cook and Borah this Matrícula was the result of an improved and more centralized reporting and examination of tribute counts.⁶

The first census gives only totals of families of Indians and non-Indians in each parish. The data are grouped and published by Peter Gerhard. The reports by local administrators varied from careful to fairly cursory and show that the administrators in general only copied the latest of the regular tributary counts and “translated” the number of tributarios into the number of families, the information the viceroy had asked for.⁷ The Aranda census would have followed the 1768 census in Spain itself, but much material gathered in the Americas is now either lost or still lying undiscovered in the archives. However, some padrones for various parishes of the dioceses of Oaxaca, Puebla, and Durango, listing the population by name, have been found in Seville. The requests for information had been transmitted through the bishops and therefore to the parish priests. The data are partly published by Cook and Borah. The Revillagigedo census, which should not be confused with the contemporaneous military enumerations, was the most successful of the three censuses. By the time Revillagigedo left New Spain in July 1794, his viceregal office lacked only the returns for the intendancy of Veracruz and the jurisdiction of Coahuila. The results have been analyzed by several authors, including again Gerhard and Cook and Borah. It is clear that, although these censuses may give occasional information on the population of parishes, provinces, or regions, they can only be used to indicate a rough pattern of colonywide growth between 1742 and about 1800 or, for Oaxaca, Puebla, and Durango, between 1742 or 1776 and about 1800.⁸

Analyzing the data of these censuses allows one picture to emerge: rapid growth. For example, Gibson noted a doubling of the population in the valley of Mexico from about 120,000 Indians in 1742 to about 275,000 by 1800, an increase of roughly 3 percent per annum. Trautmann calculated a growth of almost 300 percent of the Indian population in Tlaxcala between 1742 (11,000 Indians) and 1793 (42,878 Indians), an increase of 5.7 percent per annum. For the regions in Western Mexico, including Guadalajara, parts of Michoacán, and the far North, Cook and Borah have calculated rates of population growth per annum by decade that indicate fluctuations between 2 and 3 percent per annum for each decade from 1710 to 1800. However, for the whole of colonial Mexico, Brading calculated a more moderate growth rate of 83 percent, or 1.2 percent per annum, between the 3,336,000 inhabitants in 1742 and the 6,122,000 in 1810. He attributed this growth in large measure to the resurgence of the Indian population. Slicher van Bath, who has used the data published by Gerhard, calculated a population growth rate of 0.9 percent per annum for some 129 provinces of the gobierno of New Spain from 2,094,000 inhabitants in 1742 (of whom 75 percent were registered as Indian) to 3,254,000 inhabitants in about 1800 (of whom 78 percent were registered as Indian). These figures indicate also a relative expansion of the Indian population. López Sarrelangue estimated an increase of the Indian population of about 1 percent per annum (44 percent in this period). According to Thomson the intendancy of Puebla contained 357,239 persons in 1743 (of whom 69 percent were Indian) and 506,654 persons in 1793 (of whom 74 percent were Indian); again a relative growth of the Indian population, but here by only 0.6 percent per annum. Miranda, studying the province of Ixmiquilpan in the northern faldas of Anáhuac, noted spectacular growth of the local Indian population in the eighteenth century, but his data indicate a decline of 0.1 percent per annum between 1751 and 1791, followed by a growth of 1.9 percent per annum until 1804. On the other hand, in his excellent overview of Mexican agriculture at the end of the colonial period, Van Young discusses evidence from other regions, based on data from Guadalajara, Michoacán (including the Bajío), and Oaxaca, noting a process of stagnation after about 1760. Interestingly, Cook and Borah had registered the same stagnation in late eighteenth-century Mixteca Alta.⁹

For periods in which no regular censuses were taken, data are usually to be found in tax or church registers, although in general these data were not intended to serve demographic purposes. Interesting, nevertheless, are data based on parochial registration. For the rural areas of the region of Anáhuac—Central Mexico—the four available studies all concentrated on the valley of Puebla: the villages of Zacatelco (Tlaxcala), Acatzingo (Tepeaca), and Tecali (Tecali) and the small town of Cholula. These studies stress demographic stagnation in the valley of Puebla. At first sight their figures confirm the trends in the Mixteca Alta calculated by Cook and Borah: after decades of rapid increase between 1660 and 1735, population growth was low here after the 1740s, and, in fact, during the last decade of the eighteenth century it virtually stood still. However, looking at the background of the stagnating birth rates, the authors conclude that the stagnation resulted from out-migration. Although Morin also stresses ethnic transformation, these four parishes were obviously part of an outmigration area. Research shows that in the smaller villages of the parish of Zacatelco as many as three out of every five families emigrated. It was above all the younger generation that sought its fortune elsewhere. Every generation went through an epidemic or years of dearth, especially in the late eighteenth century, and the poor in particular were affected. Malvido for Cholula, Morin for Zacatelco, and Calvo for Acatzingo conclude that the villages were losing their economic viability in the eighteenth century; particularly disastrous were the years after 1770.¹⁰

Where did these people go? The first port of call for migrants, writes Thomson, would probably have been the provincial capital of Puebla de los Ángeles. But Puebla’s population veered between stagnation and decline, while evidence provided by Thomson and others reveals the magnetic attraction of Mexico City for migrant Poblanos.¹¹ Fragmentary evidence provides additional clues as to the direction of the out-migration. The birth and death figures collected by Humboldt in the parish of Singuilucan in the province of Tulancingo between 1750 and 1801 are the mirror opposite of the figures for the valley of Puebla and are the only published parochial data in Anáhuac outside that valley I could lay hands on. While the valley of Puebla displayed a stagnating birth pattern, Singuilucan in the northeastern faldas revealed a rapid rise in the number of births toward the end of the 1780s. Calvo, Morin, and Malvido saw the stagnant birth rate as a sign of out-migration; Thomson confirmed these findings for the city of Puebla: although not necessarily so, the rapid increase in the birth rate in Singuilucan could be interpreted as a sign of in-migration. Indeed, Martin found similar signs of in-migration in the southern falda parish of Yautepec in Cuernavaca, as did Osborn in the northern falda valley of Metztitlán. However, these authors did not indicate where the migrants came from.¹²

So we are left with the contradiction between rapid growth, according to mid- and late-eighteenth-century census data, and demographic stagnation, according to some parochial information. The analysis and publication of more evidence from rural areas, especially evidence of parochial origins, would be most welcome. Tributary data, which are easier to recover and could substitute for parochial evidence, are rarely used in the overviews of regional historical developments because the registers of tribute collectors can provide only rough estimates. How do we know whether an increase reported in tribute figures results from reproductive pressures, migration, reclassification, or all in combination? Without an analysis of how the tribute data were collected, who did the enumeration in each of the villages, and what were or might have been their agendas in compiling these lists, it seems impossible to assess such evidence, much less to draw conclusions from it.

Cook and Borah state that from the 1780s to about 1810 tribute counts were made more frequently than before, usually at close to five-year intervals. I have found little trace of this specific administrative improvement, for counting in five-year intervals was attempted by the Spanish administration during the whole eighteenth century. There was, however, an effort to lower the costs of collection by using fewer collectors. Up to 1795 this reform was not executed, because of the objection of some intendants. Manuel Flon, the intendant of Puebla, for instance, calculated that it would result in a net loss of revenue. Also the projected schedule of the new collectors was not appropriate according to his calculations: the proposed registration of three hundred tributarios a day was impossible, for even an assiduous intendant, who knew his provinces with all the officials, priests, and Indian elites, could count only two hundred tributarios a day, working from sunrise until well after sunset (with a short lunch break at noon).¹³

Nevertheless, turning the pages of some legajos and expedientes in the archives, I could not help being impressed by the care tribute enumerators took to record the people in house after house, accompanied by village leaders, priests, alcaldes mayores, or corregidores. Fraud existed, since the basis of the count was fiscal, but I tend to agree with Noble David Cook that the presence of officials who had different interests at stake in the counts kept the level of fraud within limits.¹⁴ Of course, people were no less ingenious at avoiding taxes then than now, and their cleverness could probably have caused variations in tax lists that were at least partly related to the zealousness of administrators. However, paying the tribute benefited the villagers, because it gave them access to a village plot that was part of the collectively owned land, tierras de común repartimiento. Such plots were distributed for agricultural use, worked individually, and supervised, in general, by the Indian caciques in their role of village officials of the landowning communities. The rule was a simple one: no pay, no plot.¹⁵ And the ones who paid were registered.

Indeed, the tributary data show some striking similarities with the results of the above-mentioned research. In the parish of Singuilucan, for example, which formed part of the province of Tulancingo, the number of tributarios rose from 3,751 in 1730 to 5,200 in 1790, a growth of 0.6 percent per annum. In Metztitlán we find a growth of 0.7 percent per annum, from 5,079 in 1730 to 7,300 in 1790; and in the province of Cuautla Amilpas, near the parish of Yautepec, the number of tributarios grew 0.3 percent per annum, from 2,036 in 1730 to 2,439 in 1790. But in the province of Tepeaca the number fell 0.3 percent per annum, from 13,938 in 1730 to 11,081 in 1790; in Cholula it fell 0.4 percent per annum, from 6,103 in 1730 to 4,651 in 1790; and in the city of Puebla it declined from 10,047 in 1730 to 5,048.5 in 1795 (−0.8 percent per annum). In Ixmiquilpan the Indian population figures dropped by 0.3 percent per annum (4, 137.5 tributarios in 1750 to 3,670.5 in 1790), but increased in the following decade by 2.1 percent per annum (to 4,314.5 tributarios in 1800). Purely coincidence? Despite all the problems of analysis, in the absence of more accurate data the number of tributarios might be a symptom of population growth that can contribute to a more detailed picture of demographic development in Anáhuac, increase understanding of subregional differentiation, and link the known censuses with parochial data.

Tribute Counting and Provincial Comparisons

Initially the tributarios were the people belonging officially to the República de Indios. The distinction between those registered by the colonial authorities as Indians or as non-Indians stems from the administration introduced after the conquest, which divided the society of Spanish America legally into a República de Españoles and a República de Indios in order to keep the Indian and Spanish residents of the colony separate and so protect the Indians against exploitation by the conquerors. Although this strategy did not work, the viceregal state kept records of Indian taxpayers—the tributarios—to establish who was and who was not a member of the República de Indios.¹⁶ The system of classification continued almost unchanged from about the middle of the sixteenth to the middle of the eighteenth century. In principle it was an easy matter. The “Indians” were the residents of the landholding villages in the countryside (the so-called pueblos de indios) and of special districts in the towns. Every inhabitant of a pueblo had the rights and duties laid down by law, including the usufruct of village plots. In return for juridical protection and agricultural benefits, the crown collected an annual amount, the tributo. This obligation applied to the male head of the family and after 1786 also to his adult sons: each male member of the República de Indios between the ages of eighteen and fifty years, head of a household or not, was eligible to pay this tribute every year and paid it collectively with his fellow villagers. Several other changes were proposed but never fully implemented, and the tribute data I found were based mostly on the older classification.¹⁷

A small percentage of the tributarios worked as factory hands or manual laborers in the urban centers. In such cases the employer paid the tribute in advance to the treasury. The same applied to tributarios who lived on haciendas. The hacendado paid the tribute in advance, so the tributarios on the haciendas were also included in the accounts of the Spanish fiscal authorities. Throughout the eighteenth century the number of tributarios and their families amounted to roughly 80 percent of the total population. Most of these people—some 85 percent—lived in pueblos de indios. “In the Spanish conception,” Taylor writes, “Indians in the colonial system were inseparable from their pueblos.”¹⁸ This was also the Indian conception, for the link was based on the right to usufruct lands of their village. If they moved, the tributarios could be transferred from one list to another. (But I think it was more common for a tributario who settled and worked in one of the towns to be regarded there as a mestizo or even a mulatto. The original and official “ethnic” name of a resident of New Spain seems to have decreased in importance in proportion to the distance from his birth place, suggesting an open stratification.)¹⁹

Included among the tributarios were people of other ethnic origins. The archives contain many examples of mestizos or even Spaniards who were registered on the tribute lists and were therefore eventually regarded as members of the República de Indios. In the late colonial period, according to Taylor, in the region of Guadalajara “Indian” meant “village tributary.” In the sixteenth century one pueblo de indios had already been founded by black slaves, and its residents were routinely called indios in official records even though nearly all were ethnically mulattos and “afromestizos.” In order to increase the number of contributors to village fiestas and other public expenses, some pueblos allowed mestizos, mulattos, and Spaniards to register as Indian members, live within the village, and share in the lands assigned to Indian families. The situation in the region of Guadalajara as described by Taylor might be typical for other regions in New Spain as well. I have found similar examples from pueblos de indios of the central highlands of Anáhuac, where tribute-paying mestizos and Spaniards pleaded before the court that they too had rights to usufruct lands of their pueblo. Indeed, the tribute accounts listed them as Indians. Also in the Sierra Norte de Puebla in Anáhuac’s northeastern faldas, analyzed by García Martínez, groups of runaway slaves of West African descent formed pueblos de indios with the recognition of the crown. It is important to note cases like this, for, naturally, the number of tributarios would increase through such royal recognition. After 1798 the free mulattos were included among the tributarios anyway, since most of them had become residents of the pueblos de indios by then and it was virtually impossible for the government to distinguish them from other residents. This measure led to a minor increase in the number of tributarios during the following census around 1800.²⁰

The government was deeply interested in the demographic development of the República de Indios. It wanted to be able to determine the level of tribute per pueblo de indios and to have some assurance of the well-being of their inhabitants. The standard tribute procedure required a careful listing of provincial headtowns, called cabeceras, and their subordinate villages, called sujetos, and a detailed record of tribute collectors. The comisionados de matrícula, officials who assisted the alcaldes mayores or corregidores in counting the tributarios, based their counts at least partially on ecclesiastical records of baptism, marriage, and death. The priests also kept counts for other purposes, for example to keep track of those eligible for communion and of those who fulfilled their precepts annually. Tributary residents were not counted every year. As indicated above, a particular region had to be covered within a period of five years, so that five-year counts were carried out in each village.²¹ Of course, this interval was the ideal, probably implemented better in the early eighteenth century than in the seventeenth, and very well in the rest of the eighteenth century.

The tribute lists enable us to trace the changes in the numbers of tributarios in the Anáhuac region with a reasonable degree of accuracy for the eighteenth century. I was able to use half of the lists which must have been compiled in the eighteenth century: 1720, 1725, 1730, 1745, 1750, 1765, 1775, 1780, 1785, 1790, and 1800.²² In the archives the results are included for the last year of each five-year period. For instance, the statistics for the period 1773-1777 are given as the 1777 figures but do, in fact, contain mainly the counts of 1775 (see Map 2). So the statistics for 1720 include the period of 1718-22, for 1725 the period of 1723-27, for 1730 the period of 1728-32, and so on until 1800, for the period 1798-1802. The other half of the eighteenth-century series of tribute lists was missing or incomplete in the dossiers that I consulted. I have been unable to trace the lists for 1700 (1698-1702), 1705 (1703-07), and 1710 (1708-12) in the Seville or Mexico City archives. Some of the data for 1715 (1713-17) are missing or were never gathered. I suspect that the counts for 1735 (1733-37), 1740 (1738-42), and 1760 (1758-62) were only partly carried out because of major crises such as bad harvests and epidemics. There was little point in counting heads in epidemic years, since so many of the peasants were searching for means of survival or were on the road to escape the epidemic. However, this does not entirely explain the absence of the 1770 (1768-72) list, since a count was carried out in 1785-86—two years in which harvest failures and famine affected the provinces. The 1795 (1793-97) list was compiled on an irregular basis, because of the energy devoted to the Revillagigedo census of 1790-94, although this task was officially executed independently from the tribute counting.²³

The data from the counts for the region of Anáhuac are featured in the  appendix and grouped by subregion in Table 1 and Figure 1 (in terms of the geographical characteristics of Anáhuac). Because they were not included in the counts of the early decades of the eighteenth century, I have excluded all registered mulatos from the figures of the later decades. The “ethnic” mulatos cannot be separated from the tributary population, because they were not clearly identified in the tables (neither were “ethnic” mestizos or even españoles who were registered as tributaries). In this investigation 561 figures are available, of which several are copies of a preceding one. Such copying took place in 1730 in Tula and Zacatlán, in 1745 in Teotihuacán, in 1750 in Malinalco, in 1765 in Atlixco, Actopan, Zimapán, and Tlapa, in 1775 in Coyoacán, in 1780 in Cuautitlán and Metztitlán, in 1785 in Tecali, Tepeaca, San Juan de los Llanos, Tulancingo, Zacatlán, Huayacocotla, Cuautla Amilpas, Chilapa, Tlapa, Tepejí, Acatlán, Tehuacán, Orizaba, and Zimapán, in 1790 in Teotihuacán, Ecatepec, Otumba, Ixmiquilpan, Tlapa, and Córdoba, and in 1800 in Metztitlán and Tehuacán; thirty-three times, totaling almost 6 percent of the figures. Most of this copying—nineteen times—occurred during or shortly after the 1785-86 disaster; 1786 was the only famine year of the eighteenth century, which of course made the counting of the population an almost impossible undertaking.

On the graphs in Figure 1, two periods of population increase stand out; the 1720s and from the 1780s onward. Between these peaks, the number of tributarios for all provinces of Anáhuac—the alcaldías mayores and subdelegaciones—stagnated or fell. Overall totals grew by 45 percent between 1720 and 1800, for an average increase of 0.6 percent per annum. This low rate is the result of the drop between 1730 and 1765, reflecting the severe matlazahuatl epidemic of 1736-39. The great number of deaths combined with out-migration to create a general decline in the number of residents of the pueblos de indios that lasted until roughly 1765. This was a decline of about 20 percent (an average of 0.5 percent per annum). However, the falda provinces experienced slight decline. The residents of the highlands could have left for the provinces in the faldas, where it was apparently possible to accommodate the migrants. After the epidemic of the 1760s, the population in the pueblos de indios of the 51 provinces of Anáhuac began to increase again, with a growth of 52 percent (an average of 1.5 percent per annum) between 1765 and 1800, but it is not clear whether this was initiated by a wave of births or by remigration. There were incidental fluctuations, such as the notable growth in 1750 and a decline after the crisis of 1785-86. Growth between 1765 and 1785 had been at a moderate 22 percent, or about 1.1 percent per annum. After 1785 the rise in the number of tributarios continued almost uninterrupted until 1800, increasing by almost 19 percent or 1.2 percent per annum.

Although the growth of the number of tributarios coincides, for example, with the loosening of legal restrictions on non-Indians within the pueblos de indios, in light of the nature and comprehensiveness of the archival material we may conclude that few administrative changes were introduced in the latter decades of the century. The population increase in all of the 51 provinces of Anáhuac during, for example, the decade 1790-1800 was at a modest 1.4 percent per annum, the same as the increase during the decade 1765-75, only slightly higher than for the period 1745-50 (1.3 percent per annum) or the period 1780-85 (1.1 percent per annum), and somewhat lower than in the period 1725-30 (1.6 percent per annum). The odd 2.6 percent per annum for the period 1720-25 can probably be attributed to deficit counting in the early 1720s. Cook and Borah defend the reliability of the Matrícula de tributarios of 1805, and I tend to agree, for an average growth rate of 1.4 percent per annum was not unusual. This observation is underscored by a look at the annual growth rates of the individual provinces during the decade. The rates range from a low −2.3 percent per annum (the small province of Puebla) to a high 4.6 percent per annum (the provinces of Tetepango and Ecatepec). Exceptional rates of 7.4 and 8.5 percent per annum can be found for the mining provinces ofTaxco and Zimapán. Nevertheless, the moderate growth rate of the 51 provinces was influenced by the low rates of the provinces belonging to the valley of Puebla (only 0.5 percent per annum) and the provinces in the faldas, particularly the eastern (0.1 percent). The growth rates per annum for the provinces belonging to the valleys of Mexico and Toluca during this decade were impressively higher: 2.2 and 2.4 percent. These rates resemble the figures of the period of rapid growth during the 1730s, before the matlazahuatl epidemic of 1736-39. The number of residents in the pueblos de indios around 1800 was much higher than it had been in the middle of the century, thanks to a continuous increase during the latter decades of the century, an increase as substantial and widespread as that of the 1720s and 1730s.

Other figures confirm this trend. The account of 1795 (1793-97) has been preserved for 27 of these 51 provinces. In general the figures are halfway between 1790 and 1800, as could be expected. Moreover, for the valley of Mexico Gibson published figures of 1805 (1803-8), indicating still further growth.²⁴ Some examples: Xochimilco had 3,666 tributarios in 1790, 4,226 in 1795, 4,281.5 in 1800, and 4,821 in 1805, with per annum growth rates of 3.0 (1790-95), 0.3 (1795-1800), and 2.2 (1800-1805); Coatepec had 1,118.5 in 1790, 1,123 in 1795, 1,319 in 1800, and 1,596 in 1805, rates of 0.1, 0.1, 3.5, and 4.2; Coyoacán had 3,011 tributarios in 1790 and 3,272.5 in 1795, a rate of 1.7 percent, 3,722.5 in 1800, a rate of 2.7 percent, and 4,401 in 1805, a rate of 3.6 percent per annum. Most of the 1795 figures are closer to the counting of 1800 than to that of 1785. Some provinces had an even higher number of tributarios in 1795 than in 1800: Tepeaca (11,786.5 in 1795 and 11,431.5 in 1800, a rate of −0.6 percent), Zempoala (904.5 and 896.5, a rate of −0.2 percent), Cuernavaca (10,263 and 9,352.5, −1.2 percent), Córdoba (3,057 and 2,353.5, −4.6 percent), and Huejotzingo (4,200 and 4,129.5, −0.3 percent).

Despite a recognizable general pattern, notable differences exist between the subregions. The decline in the number of residents in the pueblos de indios after the epidemic of the late 1730s was particularly dramatic in the valleys of Puebla and Toluca in the highlands, with rates of −1.9 and −1.8 percent per annum, as well as in the eastern falda provinces (Orizaba, Córdoba, and Xalapa), −1.5 percent per annum (see Figure 1). The figures for the valley of Mexico do indicate a period of stagnation after the 1730s, but certainly no dramatic decline. The falda provinces to the north and south even showed a modest growth of 0.2 and 0.3 percent per annum. It is striking here that even the stagnation that followed upon the epidemic soon turned into a stronger growth in population figures. The growth of the pueblos de indios in the northern and southern faldas was caused by an increase of the inhabitants of the six silver-mining provinces that formed part of the faldas (Zimapán, Taxco, Pachuca, Tetela de Xonotla, Temascaltepec-Sultepec, and Cuautla Amilpas). These mining provinces witnessed an almost uninterrupted increase in the number of residents in the pueblos de indios, especially in the second half of the century. The interruptions in 1765 (−0.6 percent) and in 1790 (−0.7 percent) were caused by epidemics and the famine of 1786, less notable in the highland provinces. There was also considerable growth of tributaries per annum in the eastern faldas, particularly during the periods of 1745-50 (2.8 percent), 1765-75 (1.8 percent), and 1780-85 (2.8 percent). However, the high rate of 7.7 percent per annum in the valley of Toluca for the decade of 1765-75 must be looked upon with caution.

Beginning in the 1780s, the population figures were higher than the 1720 starting point and even higher than the peaks of the early 1730s, with just two exceptions: Puebla and Tlaxcala. Because of the stagnation in Puebla, it is tempting to return to the research mentioned in the previous section in which authors like Morin, Calvo, Malvido, and Thomson identified the valley of Puebla as an out-migration area. The tribute figures presented here confirm their conclusion that population growth in the valley of Puebla was low or declining between the 1730s and the 1780s and improved only moderately during the last decade of the eighteenth century. The low Indian birth rates fit a stagnating tributary population.²⁵ Also the birth and death figures collected by Humboldt in the parish of Singuilucan in the province of Tulancingo between 1750 and 1801 are confirmed as a sign of in-migration to the northeastern faldas in the late eighteenth century; it was precisely Tulancingo that emerged from my calculations as a province with an above-average population increase between 1785 and 1800. And adding another hypothesis, combining the figures presented here with the conclusions of Morin, Calvo, and Malvido: the valley of Mexico and the provinces in the northern and southern faldas could have been in-migration areas.

Particularly illustrative are the relative figures per year (see Table 1 and Figure 2). The population of the valley of Mexico represented between 16.2 and 17.7 percent of the total number of tributarios in Anáhuac. It had a regular development; no dramatic declines, no spurts. The percentage corresponding to the valley of Puebla fell from about 27.5 percent in the early eighteenth century to about 20 percent around 1800. Relatively speaking, fewer Indians were living there, especially since the 1740s and 1750s. The Indian part of the valley of Toluca had grown in 1800 to 20.2 percent of the number of tributarios in Anáhuac. Although the percentage of 1720 was somewhat lower, 18.9, by the middle of the eighteenth century it had fallen to about 16 percent, also indicating a difficult period during the 1740s and 1750s. The population again increased here during the 1770s. The “Indian” population of the faldas increased relatively more in the middle of the eighteenth century, stabilized for a while, but experienced another spurt during the 1780s. The population rates especially of the northern and southern faldas marched to a different drummer than those in the valleys.

If we look at the 51 provinces individually, we can readily identify hve patterns in the growth rates for the period between 1730 and 1800:

1. The number of tributarios increased rapidly over the entire period; this was the case for five provinces out of twelve in the valley of Mexico, one out of nine in the valley of Puebla, one out of four in the valley of Toluca, five out of twelve in the northern faldas, six out of eleven in the southern faldas, and one out of three in the eastern faldas.

2. The number of tributarios decreased after an initial period of growth and started to grow rapidly again in the latter decades of the eighteenth century; this was the case in only two provinces, Chaleo and Huayacocotla.

3. The number of tributarios stagnated or decreased during most of the eighteenth century but started to grow rapidly in the latter decades; this was the case in four provinces in the valley of Mexico, six out of nine in the valley of Puebla, two in the valley of Toluca, one in the northern faldas, and one in the southern faldas.

4. The number of tributarios increased rapidly during the eighteenth century but stagnated or decreased somewhat in the latter decades (the mirror image of variant 3); this was the case for one province in the valleys of Mexico and Puebla, five out of twelve in the northern faldas, three out of eleven in the southern faldas, and one in the eastern faldas.

5. The number of tributarios decreased during the first half of the eighteenth century, increased during most of the second half, but stagnated or decreased around the 1790s and 1800; this happened mostly in the pueblos de indios in the provinces of Xochimilco, Puebla, Cuernavaca, and Córdoba, urban areas that saw their populations leave for the nearby cities.

Map 3 shows the distribution of these five patterns by subregion. The valley of Puebla, of course, was dominated by provinces with stagnating rates during most of the eighteenth century. The northern faldas were dominated by a cluster of provinces near the valley of Mexico with strong growth, but some stagnation around the 1790s and 1800, and by a cluster of provinces farther north or northeast with continuous growth or some stagnation during an earlier period. The same pattern can be seen in the southern faldas, with the exception of Cuernavaca. In fact, Cuernavaca formed a “family” of provinces with Malinalco and Xochimilco.

To summarize the results of this part of the inquiry, there is one major provisional conclusion: the tributary figures confirm both the picture that emerged from analyses of the censuses and the results of the parochial material from the valley of Puebla. The number of tributarios in Anáhuac increased between the middle and the last decades of the eighteenth century, with an accumulative acceleration during the last decades. This growth followed a period of stagnation after the epidemics of the 1730s. The valley of Puebla never wholly recovered from the loss of the 1730s, a situation that should be attributed mainly to out-migration, probably to the faldas and some western provinces of the region.

Interpreting Provincial Shifts

To this point the ground has been fairly solid, but the footing now becomes less certain. Although the relationship is not inevitable (other possible explanations include natural increase and better counting), it is possible that a drop in one subregion was connected with a rise in another—in other words, with migration. The use of net population change between jurisdictions like the provinces of Anáhuac to indicate migration is known to be beset with problems. In fact, using tributary data, migration can never be proven. My calculations have only the hypothetical and heuristic goals indicated above, seeking to broaden the picture of Anáhuac’s demographic development, particularly with regard to subregional distinctions. Because for heuristic reasons in my hypothetical “model” migration will be an all-or-nothing proposition, the major restrictions must be noted. First, the model design is not sufficiently sophisticated to include either temporary migrants, such as individuals leaving for some years to work on haciendas (the so-called tlaquehuales) or to live in mining communities, or people who had left their pueblos but still appeared on their pueblos’ tribute rolls because they continued to pay tribute there. Second, in some provinces the number of tributarios in the pueblos de indios might have been swollen by arriving non-Indians as the crown loosened the segregation laws in the late eighteenth century. This would have been the case with “ethnic” mulattos or mestizos registered as indios. Third, some government officials might have done their counting better than others; several figures are therefore masked by a bureaucratic component. Cases of “rapid growth” (5.5, 6.5, or 8.5 percent per annum) in one period, immediately followed by rapid decline of almost the same rate per annum over the next period, might be attributed to this bureaucratic component. Last, the territorial limits of Anáhuac prevent our following the presumably great internal migration in most of the adjacent regions like the Bajío or the valley of Valladolid. In-migration provinces in the western and northwestern parts of Anáhuac in particular might have been fed both by out-migration from the central provinces of the region and by in-migration from beyond Anáhuac. Yet any interpretation of the results of province-to-province comparisons should not overlook the possibility that changes in tributary figures may be caused only partly by migration; the bureaucratic component and shifting fertility and mortality rates will have played their roles too.

The two kinds of migration I will discuss are:

1. Migration from one pueblo de indios to another, probably from the highlands to the provinces in the faldas. In principle this migration maintained a balance in the number of tributarios and may be considered primitive migration under the pressure of bad harvests and epidemics; this was a push factor.

2. Migration from a pueblo de indios to a site where the people did not belong to the República de Indios, such as neighborhoods in many provincial headtowns and cities. In this case the tributarios disappeared from the tribute lists; this might be considered a pull factor, in light of the employment opportunities provided by towns and cities.

Migration from Village to Village

In order to investigate possible migration, I refer again to the growth rates for each province (see the  appendix). However, it is not the average increase or decrease that is significant in this case, but instead any excessive deviation from that average. For that reason z scores are introduced, the standardized score showing the relative status of a score with regard to the general trend discussed above, the overall tributarios curve of Anáhuac. Each growth rate in the sample was converted into a z score, expressing the deviation from the mean in standard deviation units. The z score tells how many standard deviations x is from a mean of zero. A perhaps arbitrary but prudent margin of 20 percent (z scores beyond the +1.28 or −1.28 significance level) was adopted, and the explorative (not testing) statistical investigation was directed toward those provinces that deviated either way from the average line of development: the higher the score, the greater the deviance. In the following analysis I therefore presume that a province with a score of +1.28 or more had above-average growth, possibly due to in-migration, and that a province with a score of −1.28 or less had below-average growth, probably a fall, possibly as a result of high mortality or out-migration. These figures are underlined in the  appendix.

The pattern that results from these statistics is divided into four periods: 1720-30, 1730-65, 1765-85, and 1785-1800 (see Maps 4a-4d). Does it make sense? Two points suggest a moderately positive answer. First, clusters of provinces with “deviant” scores can be recognized. It is understandable that possible out-migration areas (with below-average performances) and possible in-migration areas (with above-average performances) were generally close to one another and that clusters of “in-migration” and “outmigration” provinces were close to one another. Only these clusters will be discussed; individual provinces with a deviant result will be ignored unless some provinces in their neighborhood had the same deviance or unless they are mining provinces, which always attracted people from all over the colony. As stated above, not included in the calculation is the in-migration from outside Anáhuac, but it would contribute to above-average performances in several provinces. Second, a superficial examination shows the effect of some well-known epidemics and agrarian crises as heavy fluctuations, caused by migration and re-migration.²⁶ The re-migration of people who left only temporarily explains why tribute populations sometimes recovered in only ten to fifteen years. Permanent in-migration would also affect tribute figures positively after a lag of about ten to fifteen years, the consequence of high birth rates. Although the counting of several of these provinces had been done in different years, the results show coherence.

The provinces with an above-average performance during the first decades of the eighteenth century were located in the northern faldas: Actopan, Ixmiquilpan, Metztitlán, Huayacocotla, and Xochicoatlán-Yahualica (see Map 4a). Also Ecatepec and Otumba had above average scores. Below average were Apan, Texcoco, Xochimilco, and Taxco. In 1729 a measles epidemic was reported in Apan and Texcoco.²⁷ The results of the counting after 1745 (1743-47) show all the traces of the ynatlazahuatl epidemic of the late 1730s (see Map 4b). Provinces with above-average growth rates can be found in the northern, southern, and eastern faldas. Provinces with below-average performance are to be found in the central highland valleys, especially in the valley of Puebla. On the other hand, the mining province of Pachuca must have had spectacular growth in the 1750s and 1760s. Some provinces evidenced considerable fluctuations. Metztitlán, for instance, had a rapid increase followed by a rapid decrease. According to the data, Tetepango saw people come, then go, and finally return again. The reverse was the case in Orizaba. Of course, it is difficult to conclude from these figures whether the “above-average growth” provinces show excessive in-migration or relatively high birth rates after the epidemics. Abrupt changes could have been caused by a complicated problem of villages that were in protest over a new tax, an abusive priest, etc.; Taylor gives evidence of villagers leaving their homes for such reasons.²⁸ But, as indicated above, fluctuations like these would mainly have been the consequence of faulty counting, harvest failures, or epidemics. Independent church counting in 110 parishes of the archbishopric of Mexico tends to confirm especially the possibility of harvest failures and epidemics.²⁹ The number of tributarios in these parishes decreased almost 34 percent, and it took them some ten years to recover. Data from this register referring to the same provinces (see Table 2) confirm the tributary accounts made by the state officials: the valley of Toluca suffered heavily in contrast to the valley of Mexico where the epidemic had not struck so badly. In sum, based on the deviant tributario figures, the epidemic resulted not only in a high mortality rate and out-migration to nearby provinces, but also in an evident migration to the faldas. Many people gradually returned home when the epidemic was over.

In the second half of the century, the spotlight again shines on the faldas more than on the central highlands. Nevertheless, more indications of both above-average growth and below-average performances occur in the valleys of Mexico and Toluca (see Maps 4c and 4d). The Indian population might have left the valley of Puebla, the northeastern faldas, and the drier provinces in the valley of Mezquital around Actopan. An agrarian crisis in 1771 would have led to such a westbound migration. Besides this, in both maps the province of Metztitlán shows signs of in-migration and prolonged demographic growth. Not far away from Metztitlán were Actopan, Tetepango, Tula, Pachuca, Xochicoatlán, Huayacocotla, and Zempoala, which had below-average growth rates during some of the years, although Tula and Tetepango recovered quickly. The people of Metztitlán might have come from these nearby provinces. During the 1770s and the early 1780s the valley of Toluca and provinces like Actopan experienced decreases in the tributario population, probably caused by epidemics.³⁰ During the early 1780s provinces directly northeast of Mexico City saw heavy demographic losses, while nearby Apan and Tacuba, like Cuernavaca somewhat later, were growing rapidly.

Most interesting should be a look at the figures for the last quarter of the eighteenth century. Harvest failures, famine, and disease combined forces to establish the late 1780s and the 1790s as one of the most disastrous periods apart from the epidemics of the 1730s.³¹ A series of poor harvests resulting from unseasonably cold weather and droughts caused massive famine among the subsistence sector of the populace, especially during 1785 and 1786, “el año de hambre.” The year of hunger led to mortality and out-migration in the provinces of the northern faldas, perhaps to migration into the valley of Mexico, for Cuautitlán, Tacuba, and Ecatepec had above-average scores in that time. The same can be said, once again, of Metztitlán and Xochicoatlán in the far north of Anáhuac, where there was still enough maize and less danger of being caught by the epidemic. The semiarid provinces to the north of the valley of Mexico all experienced harvest failures, while in the faldas people could save parts of the harvest or tried to produce a second maize crop in the winter season.³² Nevertheless, Anáhuac was much less affected by the harvest failures than were its western neighbors. It is useful to compare these patterns with a map published by McGovern-Bowen based on figures provided by Spillman, which shows that the impact of the disaster was most intensely felt in the regions of Michoacán and Guadalajara; Spillman documented a 200 to 300 percent increase in the number of burials between 1784 and 1786 in many parishes there. In general, the overall intensity of mortality during this crisis decreased with distance from the central core of Guanajuato, San Miguel de Allende, León, and Celaya. But the disaster would have caused migration to regions where the problems were still not as great, and death totals in the region of Anáhuac were much lower, leaving the central highlands relatively free from mortality above 50 percent.³³ Possible in-migration from Michoacán can be recognized in the results of the 1790 counting, with its overall increase; Metztitlán, Xochicoatlán, and Zimapán stand out.

To summarize, we may note first the confirmation of the consequences of epidemics and subsistence crises. Beyond this, the faldas have a strong profile as an overall in-migration area and the valley of Puebla as an out-migration area, especially during the middle decades of the eighteenth century. If by further research part of this pattern could definitely be established as migration, I think it will be of the first kind of migration mentioned above, from pueblo to pueblo. The people who left their pueblo de indios could be integrated into another pueblo de indios, although sometimes only temporarily. I suspect, but cannot prove at the present stage of research, that even in years of crisis the Indians showed a preference for communities of the faldas, where the indigenous languages were still spoken and popular Catholicism still flourished, i.e., where traces remained of the link with the altepeme organization—the foundation upon which the pueblos were first established.³⁴ These tributarios remained tributarios, and their migration should therefore be considered as originating from push factors: epidemics and harvest failures. The Ramo de Tributos and Ramo de Indios in the Mexican National Archive are full of reports indicating this kind of migration. The maps discussed above mirror crises in the countryside of eighteenth-century Anáhuac.

But the tributarios went to the faldas for another reason. They were probably warmly received, an issue addressed by only one study, Danièle Dehouve’s description of the separation of villages in the province of Tlapa in the southern faldas.³⁵ She traces the origins of the pueblos de indios and the division into cabeceras and sujetos, the latter administered by Indian leaders of the cabecera. During this process in the sixteenth century, especially in the thinly populated faldas, several pueblos de indios were formed by Indians of different ethnic backgrounds and united former rival villages. From the beginning of the seventeenth century the subject villages of Tlapa wanted to free themselves from domination by the cabeceras, but they could only do so if they possessed “a very decent church,” a “community house and community property,” and especially a specific minimum number of “Indian” families, i.e., tributarios. The church, community house, and community property could be arranged, but the necessity to count at least eighty Indian families according to Spanish standards remained difficult throughout the eighteenth century. Only when “Indian” migrants came to their villages and decided to stay and to contribute to the tributo could this last goal be reached. However, it did happen in the late eighteenth century. In 1767 Tlapa still had seventy sujetos, but only thirty years later the province had totally disintegrated and most of the subject villages had attained the rank of cabecera of their own pueblo de indios.

To the preliminary conclusions I can now add the following hypothesis. It can be seen that the period of bad harvests in the Anáhuac region corresponded to a period of increased migration in the countryside. Rural migration in Anáhuac has not been studied in much detail, and my provisional picture has been reconstructed with the aid of tribute lists. Despite its limitations, this picture reveals a combination of push and pull factors. It is difficult to distinguish between these factors, but it is the push factor alone—the rural disorganization resulting from the agricultural crises and relative overpopulation—that seems to fit what can be called “migration from pueblo to pueblo.” This was by and large a migration to the faldas. Many Indians did their best to continue as farmers in the larger Indian settlements in the faldas or in the highlands, even if only for the duration of the crisis.

Migration from Village to Town

The second type of migration is much more difficult to grasp and cannot be illustrated in maps. One factor that determined the migration from village to town was the hope of making a better living there. Leaving village life, of course, was a big step, not the first to come into a peasant’s mind. In New Spain it was even more difficult to cross the threshold, for it involved a change of diet (more bread, less maize), of language (more Spanish, less of one’s indigenous language), and of labor (fewer agrarian activities, more industrial). Only those peasants who abandoned the countryside for good because there was no more work for them or because they were attracted by employment possibilities in the towns were prepared to become “non-Indians” according to viceregal standards. In the course of time they became detached from the peasant setting; they became professional artisans, the majority of whom settled in a few larger villages, towns, and cities. During the last decades of the eighteenth century, many Indians, particularly young males from the peasant villages, took their chances in the cabecera or the town in the neighborhood and tried to make a living there that was no longer possible in the old villages.

How can we measure and classify the number of lower-class urban inhabitants over time? Theoretically, migration in Anáhuac must be reflected in the balance of the tributarios figures. If it had been the first kind of migration, the number of “Indian” out-migrants from the countryside would have matched the number of “Indian” in-migrants to the towns. As Cook and Borah, Seed, Valdes, and Anderson have already stated, one major flaw in this kind of research is the scarcity of urban Indian population data for the late colonial period. The military census of the 1790s excluded Indians, for they were not subject to military service. But according to Anderson and Chance we should not look for “Indians” to identify the in-migrants from the rural areas but instead look among the non-Indian groups in the towns. Chance has found traces of changing group identity among the urban Indians of Antequera de Oaxaca in southern Mexico. The curate of the village of Jalatlaco, near the city, observed in 1777 that Nahuatl was no longer spoken there and had virtually ceased to be used in three other villages. Many Indians were drawn to the city of Antequera as spinners and weavers, as masons, bakers, and artisans. Anderson convincingly argues that among those individuals labeled “Spaniards” by the census takers at that time in Guadalajara were low-status individuals virtually indistinguishable from the city’s Indian and casta residents in every significant social and economic category.³⁶ The second type of migration must thus have had an interesting effect: the number of non-Indians should have increased.

Who were the non-Indians? The social system discussed here made use of an ethnic terminology and was called the régimen de castas. The concept of casta originates in the sixteenth century as a Spanish classification of a group descended from a common ancestor or from two ancestors with different ethnic backgrounds. As such, the term can be found in contemporaneous etymological dictionaries and even in the twentieth-century Diccionario de la Real Academia. Over a period of time the term came to be used by Spanish officials to classify urbanized non-Indians who were also non-Spaniards. Most of the Spaniards possessed a Spanish birth certificate, which can be seen as a sort of “passport,” although birth certificates were rarely called for outside the elite. One could pass into the category of Spaniard if one sufficiently resembled a European. This category comprised two subgroups: the europeos, who were born in Spain— although a few came from other European countries—and the criollos, who were born in New Spain. In general in the eighteenth century, all the groups not included among the Spaniards or the Indians were referred to as castas; they possessed neither the “passport” nor the relevant somatic characteristics. They consisted of variously named subgroups: the commonest were castizos, mestizos, and mulattos, but in official documentation other names like moriscos, lobos, or negros were frequently used as well.³⁷

Legally, each main casta stratum had distinct civic and fiscal rights and obligations. For instance, the members of the mulatto category, which was quite important in western New Spain, had slave ancestry, paid tribute, and were prohibited from entering certain professions. The result was that many mulattos tried to “cross over” into another category, which in the cities would have entailed becoming a craftsman. Marriage was a suitable avenue for attaining or maintaining a higher position on the social ladder or for a step down, because where, for example, the sex ratio was unfavorable, women readily married down the ladder. There were also Spaniards who entered Indian villages, started to pay tribute, and therefore later were registered as Indians. In practice, despite legal restrictions, any combination of factors could affect one’s “ethnic” label: the color of one’s skin, the purity of one’s descent, one’s profession and standard of living, one’s personal worth, integrity, and place of birth.³⁸

An investigation of the criteria for determining calidades—“ethnic” labels—in Mexico City in the second half of the eighteenth century reveals a kind of individual assessment.³⁹ For instance, a tailor who married a mulata was regarded as a castizo by the church, while a government official classified him as a Spaniard. The priest concerned wanted to reduce the social distance between the man and woman by degrading the man, because he married beneath his calidad. Another document introduced the Spaniard Vicente Garcia, who had settled as a ranchero in the province of Cuernavaca. He was married to a mestiza and was also regarded as a mestizo by his neighbors because of his relatively low social status. The author of the document, a Spanish official, described García as “de calidad español, aunque corre por mestizo.” His profession did not yield a sufficient income for García to maintain the standard of living that matched his original calidad. He was forced to work as a muleteer in the company of two other rancheros (de calidad india), and the tribulations he met with on his travels left him looking somewhat the worse for wear, since he was referred to as a mulatto in the villages through which he passed. Humboldt states that the Mexicans were able to distinguish between muleteers from different casta origins in the darkness of the night according to their smells, because they thought that the castas of Indian and African descent had retained their peculiar bodily odors; European muleteers were called pezuñas by the people he had met, Indian muleteers poseas, and black muleteers grajos.⁴⁰ The example of Garcia shows that Humboldt’s statement was not always true.

More examples of this kind indicate that the ethnic terminology no longer reflected the original and legal distinction between different ethnic groups. Seed demonstrated that deviation in occupation from expected norms would lead to a reclassification of the calidad nearer social expectations.⁴¹ Besides this, the poorer groups in the cities and towns of Anáhuac’s central provinces did not care very much whether they were seen as mestizos, mulattos, or castizos. A census taker in Tepeaca reported that the non-Indian plebeian groups of the province left it up to whoever asked them to decide which calidad they belonged to. They were only concerned not to be classified as tributarios, for paying that tax had no advantages for them outside the pueblos de indios. In fact, the Indians in the pueblos de indios were the only ones to attach importance to their being Indian, because it entailed rights to land use (común repartimiento) in their pueblos. However, as was revealed by Aguirre Beltrán some time ago, even Indians in villages could claim non-Indian status if their economic occupation had shifted impressively. He presents an observation of a census taker writing a report of the inhabitants of Tepetlaoztoc, a pueblo de indios in the province ofTexcoco. This official had recorded the calidades of Spaniard, Indian, castizo, mestizo, and the like based on the declarations of the people themselves, and the pueblo de indios was therefore recorded as a town full of Spaniards.⁴²

The concept of calidad as discussed here could have been derived from the somatic norm image above all of the Spanish elite, which strongly influenced other social groups.⁴³ Defined as the complex of physical characteristics accepted by a group as its norm and ideal, the somatic norm image was introduced by Harry Hoetink to interpret problems of combined ethnic, sociopsychological, and economic origin in segmented societies like the Caribbean. The situation in eighteenth-century New Spain summarized in this section cannot have been much different from the one he describes. The nouveaux riches and other social climbers, but not exclusively these groups, attached importance to the “ethnic” terminology because it enabled them to outstrip their former social peers. (At the same time, and borrowing a phrase of Hoetink’s, it seems that the “white” patricians accepted “coloreds” in their midst because by their physical and cultural criteria these nouveaux élitaires were not “colored.”) The somatic norm image “belongs to the spiritual heritage of the group,” writes Hoetink, “and is comparable with such concepts as that of a norm of behavior, embodying the assumptions concerning the manner in which members of a group should behave.” As such, the calidades could have reflected an ideal of social stratification and the main indicators of social status, especially for the rich and prosperous—a way for this economic upper crust to distinguish itself.⁴⁴ The rest of society, especially the poor, the overwhelming majority of whom were not social climbers, adopted an attitude of wait and see.

What interests me in the calidades is not so much the precise way of defining stratification criteria, but the acknowledgement of the fact that the régimen de castas mirrored in one way or another the social and economic strata in New Spain. It is obvious that the terminology had become more complex by the late eighteenth century. Chance argues that this was not a consequence of the system becoming more rigid, but that “it meant only that the elite was making a belated effort to maintain its position as a dominant white minority.”⁴⁵ Such an attitude would only show up in New Spain in periods when more and more people claimed to be “non-Indians.” Indications of a growing number of castas and a further process of differentiation between the castas may be considered evidence of social mobility and an extension of the social ladder during periods of population growth. Consequently, one should note a striking increase in the number of non-Indians in the towns. This brings us to the terminology used by the census takers. A glimpse of shifts in ethnic relations in the provinces of Anáhuac can be obtained from the figures on ethnic groups in New Spain published by Slicher van Bath, discussed above. The ratio between the social/ethnic groups based on the census statistics for 1742-43 and the 1790s shows an obvious increase in the official registration of castas, indicating relative growth (see Figure 4, in which every circle reflects a figure of 100 percent). It is worth noting that the officials listed only Indians and non-Indians in 1742 while at least four ethnic groups were mentioned in the 1790s, as if the distinctions between non-Indians had indeed become a hot issue. Therefore, it could have been a sign of an extended social ladder. This is not without logic in economic history, for the well-known ecological pressure of epidemics, bad harvests, and population growth should have transformed social relations in New Spain after 1750. In fact, most authors who have done primary research on this problem indicate just such a transformation.⁴⁶

Closer examination of Slicher van Bath’s statistics (see Figure 4) reveals that the increase in the population of castas is greatest in the center of Anáhuac and diminishes if one moves toward the faldas, a trend that could be interpreted as a mirror reflection of the migration of the tributarios to the faldas mentioned earlier. It is striking that by the end of the century the valley of Puebla seems to have had a relatively large number of castas. By contrast, the residents in the provinces to the south and east of Mexico City were predominantly Indian (more than 80 percent). This characteristic is further to be found in the provinces in the faldas alone. The census data give the impression that the reduction in the Indian share in rural Anáhuac between 1742 and 1743 and the 1790s corresponded to the increase in the number of castas.⁴⁷ These castas probably lived for the most part in urban surroundings, e.g., cabeceras, small provincial towns, or the bigger cities. Mexico City, Puebla, and Pachuca were the main centers of the non-Indian population in the region, but provincial towns like Atlixco, Anan, or Teneaca could have shared the same urban character.

By 1800, the face of Anáhuac had changed. This brings me to my last hypothesis: migration within the highland itself, from pueblo to cabecera or town, had caused a growth in the urban population in which the migrant was usually no longer classified as indio but as mestizo or any other non-Indian classification. The arrival of the poor from the pueblos de indios had led to considerable disruption in the cities of Anáhuac, as the plebeians who were mainly to be found in the countryside until then now reached the suburbs of the towns. The apparent increased use of the ethnic terminology in the eighteenth century was, I think, not simply a matter of demographic rates (more people and therefore more of them of different status groups). It must have reflected the elite response, a sophistication of the ethnic terminology to design finer gradations among the calidades in the upper echelons of society and so add luster to the subtle game of social mobility.

The Main Direction of Future Research

The tributary data can be used to construct a provisional picture of population growth in rural Anáhuac, although, to paraphrase Kubler,⁴⁸ it would be futile to attempt to prove that the tributary lists were accurate, exhaustive, or unbiased. The shortcomings are illustrated by, for instance, the 50 percent decrease in the recorded number of tributarios in the provinces of Puebla-Amozoc between 1725 and 1730. Still, most of the data discussed in this essay show enough cohesion to be far from useless for our purpose, and I hope to have demonstrated that they complement existing information based on censuses or parochial archives. The data should not be evaluated separately from data reporting non-Indians or from information on the calidades. The régimen de castas, from Spaniard to Indian, influenced the number of tributarios, and a stagnation in the number of tributarios could mean excessive dying during epidemics, out-migration, or a change in calidad. Historians and anthropologists in their analyses should take into account how the social groups of New Spain used the fluid “ethnic” categories to advantage.

To construct a good picture of the migration pattern for Anáhuac, tens of parishes should be investigated; I have sought only a way to begin. Special interest must go to some falda areas, like the cotton-growing regions east and south of the central highlands, to provinces like Metztitlán and the pueblos de indios in the valley of Mezquital (Actopan, Ixmiquilpan, Tetepango), and to highland provinces like Texcoco, Teotihuacán, Metepec, and Xilotepec that repeatedly emerge as relatively fast growing areas in the records I have seen. The patterns of population shifts presented above of when, how much, where, and why “Indians” might have migrated in the central Mexican region of Anáhuac suggest a series of hypotheses that connect economic and social history.

One question remains: what factors could have stimulated overall population growth? It seems clear that the population of Anáhuac had experienced a notable increase in the second half of the eighteenth century, in both urban and rural areas. The fastest-growing regions were the faldas, and in some respects also the central provinces, while the valley of Puebla saw demographic stagnation during the middle decades as a result of outmigration. In European cases of such rural population growth, researchers almost immediately think of protoindustrialization.⁴⁹ Discussing some central European examples Fischer assumed that the specific population pattern was created by the growth of domestic industry. David Levine confirmed this finding but argued that proletarianization and landlessness did not necessarily lead to a drop in the marriage age they “only removed the disincentive to early marriage.” Only after the increase in employment offered by domestic industries did earlier and more frequent marriages occur. To be sure, Gullickson found typical protoindustrial production in the Pays de Caux of Normandy, without the occurrence of rapid population growth. Acknowledging this kind of exception, Gutmann concludes that in general the European population grew in the eighteenth century because a new social group developed that had fewer ties to the land and thus was more likely to marry even when its members could not expect to acquire a farm at all.⁵⁰ I have suggested elsewhere that an initial process of protoindustrialization swept Anáhuac in the eighteenth century.⁵¹ Such a hypothesis awaits further research, especially along the lines of work conducted by other scholars that combines analysis of parochial with economic data.

I would like to thank Pete Mason for his linguistic assistance, Catrien Bijleveld for valuable statistical assistance, Julia van der Valk for copying the tributary data of 1790 and 1795, and D. A. Brading, B. H. Slicher van Bath, and William B. Taylor for helpful criticisms on the several drafts of this study. I also pay tribute here to the collegial criticisms of three anonymous reviewers.

Tributarios in the Provinces of Anáhuac, 1720-1800