Spanish Captives in Indian Societies: Cultural Contact Along the Argentine Frontier, 1600–1835

European-indian frontiers, the intermediate zones between areas of secure European settlement and those where Amerindians maintained their autonomy, were similar in many ways in the American empires of Spain and Great Britain. In both colonial empires frontier regions were usually zones of tension and conflict, where frequent raiding sometimes gave way to open warfare. In both empires, violence between Europeans and Indians regularly led to mutual taking of captives. While literate Anglo-Americans had little interest in captured Indians and rarely bothered to record information about them, they were often interested in their fellows who had been held captive by Indians. As a result Anglo-American historical literature includes a substantial documentary record on European captives, material that provides extremely valuable, often unique, information about the societies on both sides of the frontier and their interaction.¹ Frontier relations between Europeans and Indians have been studied far less thoroughly for Spanish America, although recent work on northern Mexico and New Mexico has begun to examine this question in a Spanish American context.² This essay is intended as a contribution toward filling that gap in the historical literature for one frontier area, central and southern Argentina, using the often fragmentary but nonetheless intriguing data from the sixteenth to the early nineteenth centuries.

Patterns of Frontier Relations

The southern region of South America, the area that in the eighteenth century would become the Viceroyalty of the Río de la Plata, was, like northern Mexico, a region inhabited by several Indian societies able to fend off Spanish conquest in greater or lesser degree until the nineteenth century. The independence of these groups was in part aided by the adoption of the horse, which by the end of the sixteenth century had become an integral part of indigenous culture in southern South America. As a complement to the Indians’ warrior ethos, the animal allowed for the development of a “horse culture” and permitted Indians to imagine and sometimes to achieve military equality with the Spaniards. Indian tribes also gradually became dependent on cattle as both a source of food and an object of trade with other Indians and Spaniards alike. But it is the horse that heightened the level of conflict among Indian groups and between Indians and white society.³ In addition, these Indian groups tended to inhabit zones that were either peripheral or inaccessible to the mainstream of Spanish colonization, to live in dispersed and small communities, and to be adept at the techniques of seminomadic living and guerrilla warfare.

From the beginnings of Spanish settlement in the sixteenth century, ranches, towns, and cities were periodically threatened by Indian raids: to the north the Avá-Chiriguanos and the Calchaquíes; in the center of the region, the Chaco groups such as the Guaycurua, the Charrúa, and the Mocobí; and to the south Pampas, Pehuenche, Tehuelche, and Araucanian tribes.⁴ At times a state of endemic war existed, as Indian raids and Spanish entradas exploded along the frontier.⁵ In the middle of the eighteenth century, however, a combination of foreign and colonial policy considerations caused the Spanish crown to reexamine its defense position in Spanish America. One result was a new plan of militarization on the southern Indian frontier that combined new presidios and forts with an increasing military presence. The government established a line of military forts in the 1750s, reinforcing them during the 1770s and 1780s.⁶ The crown also encouraged the settling of civilian population close to each fort. All male inhabitants of the rural districts were required to enlist in militia units. In addition, a special military unit created to protect the frontier, the blandengues, was formed in 1751 and reorganized shortly after the founding of the viceroyalty.

The Spanish combined this line of forts and missions with a program of pacification and cooptation of hostile tribes. Indian leaders were invited to Spanish settlements, where they were entertained and presented with trinkets.⁷ Funds were also supplied by the sisa tax to pay for ransoming captives and rewarding faithful Indians. When, for example, an Indian referred to as Sinforoso and his uncle brought back a Spanish captive from the Tobas, the intendent rewarded them and their men with goods worth 104 pesos, including ponchos, hats, uniforms, a bastón, tobacco, and knives.

Sporadic incidents continued along the Indian frontier,⁸ but in general Spanish policies combined with adverse natural phenomena (such as drought), contagious disease, and widespread food shortages to weaken Indian assaults and impose an uneasy peace. Although Indian tribes in the Chaco region continued to raid each other, a buffer zone between the Spanish and hostile Indians essentially held until after independence. Moreover, the establishment of forts and coastal defensive colonies also brought the Spanish and Indians in direct contact with one another and stimulated an active and profitable, though extralegal, trade between the two groups. With their borders more or less pacified, the Spanish slowly began to open up new lands for colonization, increasing the numbers of Spaniards inhabiting rural areas and extending agriculture and stock raising. Beginning in the 1780s, the result of peace was population growth in the rural districts of Buenos Aires, Córdoba, and to a lesser degree Santa Fe, accompanied by increased production of cereals and hides.

This period of comparative peace ended in the decade following the English invasions of Buenos Aires in 1806–7. Perhaps the Indians realized that strife in the Argentine provinces now gave them the opportunity to redress the increasing encroachment on their territory. In addition, the new provincial governments, concerned first with war and peace against Spain and then with war and peace among themselves, failed to tend to the line of fortifications and neglected Indian leaders.⁹ Revolutionary forces moving north requisitioned entire companies of soldiers that had previously garrisoned the frontier. The result was widespread strife along what had been a pacified frontier. To the south, the Chilean general José Miguel Carrera joined with the Ranquel Indians and Pampas groups to raid Salto and Melincué in 1820. To the north, the Chaco Guaycuruans attacked Santa Fe and Santiago del Estero in 1821.

The newly independent Argentine provinces, dependent to a far greater degree on the export of cattle hides and other products than the viceregal colony had been, began to organize a defense. In 1819 cattle ranchers in the province of Buenos Aires created a Sociedad de Labradores y Hacendados, using their laborers as a “mobile army” to defend the more isolated ranches. The next year the provincial government of Buenos Aires, under the military leadership of the governor, Martín Rodriguez, adopted a more offensive posture with a campaign to the south modeled closely on colonial entradas. During this four-year campaign, Rodriguez and his men invaded Indian territory three times but succeeded in founding only one new fort, Fuerte Independencia (present-day Tandil).

Nine years later the new governor of Buenos Aires, General Juan Manuel de Rosas, again interested in pacifying the frontier to assure greater production of cattle products, began another “desert campaign.” Rosas was a prominent rancher and industrialist. He was also a consummate, ruthless politician committed to extending the grazing lands of the province of Buenos Aires and willing to ally himself to so-called “friendly” Indians to achieve his ends. His goal was to free those lands between the Salado River to the north and the Colorado and Negro rivers and the cordillera to the south from hostile Indian encroachment.¹⁰

Rosas was successful. He added to Buenos Aires province an area extending two hundred leagues west to the Andes and south beyond the Río Negro.¹¹ Moreover, as leader of one of three divisions involved in the 1833-34 campaign, he was able to return with a large number of people previously captured by the Pampas Indians. Rosas’ division, responsible for the left flank of the invasion, advanced from his ranch at Los Cerrillos to the island of Choele-Choel on the Río Negro and from there to the mouth of the Río Colorado. Here the troops split, with one group continuing south along the coast to the Río Negro and then up this river to the confluence of the Limay and the Neuquén. Other troops marched inland, following the Colorado into areas “never before seen by the Christians.” The captives that Rosas brought back to “civilization” were all encountered in this large region.

The Taking of Captives

The fact that Rosas encountered Spanish captives in his “desert campaign” is hardly surprising, for one of the constants of Spanish-Indian warfare in the Río de la Plata throughout four centuries had been the taking of captives by both parties. Whether it was the Avá-Chiriguanos to the north, the tribes of the Chaco that assaulted Spanish settlements, or the Indian groups to the south, all seemed to be interested in two types of booty—livestock and human captives.¹² As early as 1577, the Avá-Chiriguanos attacked the newly founded Spanish settlement of Tarija and carried off some 40 people.

By capturing Spaniards and mestizos the Avá-Chiriguanos were in fact continuing their traditional cultural patterns, for they had always taken captives from other Indian tribes.¹³ Although officially forbidden by Spanish law, these Indian prisoners of war were sold as slaves to Spanish and mestizo landowners. Spanish and mestizo captives, such as two very young girls and a young mestizo freed in 1590, were either ransomed or recaptured by the Spanish and returned to Spanish society. More than two centuries later, the Avá-Chiriguanos were still resisting Spanish encroachment and still taking captives. In 1809, the Avá-Chiriguano chief Cumbay, angered that five captives had been turned over to the comandante of Santa Cruz, mentioned that “since olden times, it has been the custom to ransom [captives] for one silver peso apiece.” Although in the peace treaty signed with the Spaniards the same year Cumbay promised to turn over all Christian captives within two years, by 1811 this clause had not yet been honored.

In the south, in the pampas region, probably the first reference to taking captives concerned a Spanish sailor captured by the Tehuelche Indians in the early 1600s.¹⁴ By the beginning of the eighteenth century, if not earlier, most Indian groups included some captives among their population. Reports of myriad incursions by hostile Indians normally included mention of the taking of captives, while Spanish entradas often freed at least one or two. It was from the ranks of ex-captives who had learned the languages and customs of their captors that the Spanish often recruited interpreters and scouts.

Sporadic hostage taking continued along with sporadic raiding. Some of these hostages were incorporated into Indian society, but others escaped, and still others were ransomed back to Spanish society. Governors and viceroys were often called upon to contribute to funds for the ransom of captives. In 1788, for example, Viceroy Loreto donated a total of 663 pesos 3½ reales to ransom from Indian captivity Spaniards who had probably been captured in the 1786 San Luis raids. From the entries in his account books, the price for rescuing a captive seems to have ranged from 50 pesos 4½ reales paid for a woman in April 1788 to 100 pesos paid for a man two months earlier.¹⁵ The viceroy also paid 512 pesos 7 reales for eight captives freed in Salinas in December 1788, an average of 64 pesos per individual.

Such relatively large private gifts were not the only source of money for freeing Spaniards. After receiving government permission, individual citizens also ransomed members of their families who had been taken captive. In addition, all people drawing up wills in colonial Río de la Plata donated at least two reales to the Fund for the Redemption of Captives, one of the mandas forzasas originally envisioned to aid in freeing captives in the Holy Land. In the Río de la Plata this money was used to ransom local people.

Indians were not the only ones to take captives. Spanish officials were not averse to holding Indians as hostages in an attempt to coerce local tribes. After learning of an Indian raid in 1582, for example, Pedro de Segura, corregidor of Tomina, held hostage a group of Avá-Chiriguano chiefs who had come to visit.¹⁶ It is also clear that throughout the Río de la Plata many Indians whom the Spanish managed to capture in battle were enslaved. In 1750, the Jesuit Andreu mentioned a government official in Salta who had captured some Indian children and was selling them for 100 pesos apiece. In addition, Indians held in prison by Spanish authorities or working in religious institutions or private homes in Spanish towns and cities commonly believed themselves to be captives of the Spaniards. The fine distinctions between captivity and imprisonment or between captivity and Christianization escaped many Indians.¹⁷ From time to time, captives were exchanged. In 1783 Pedro Pablo Maldonado was sent to the Luján fort by his Indian captors to deliver a message offering to exchange two Christians for two specific Indians. If the exchange were effected, the Indians would consider it a sign of peace, but if the Spaniards failed to release these two captives, the Spaniards would be attacked when they made their yearly visit to the Salinas salt flats.

Spanish society probably took captives as booty, for profit, and to teach a lesson to those whom they considered “heathen savages.”¹⁸ Although the Spanish colony often suffered from a scarcity of labor, Indian slaves were too few in number and too intractable to offer a viable solution to Spanish society. Indian societies probably took captives for profit, to gain a medium of exchange with other Indian groups and the Spanish, and to increase their labor force. Adult male captives were often enslaved, or at least thought of themselves as being in some type of serfdom. In the surviving captivity declarations, both men and women refer to their “amo,” their Indian master.

Surviving documents further attest to the use of the captives, especially children, as a medium of exchange or as goods to be bought, sold, or bartered.¹⁹ In 1790, for example, the Auca Indians approached the small Spanish garrison in Patagonia offering to trade “a girl aged 4 or 5, daughter of Christians,” for “aguardiente, flannel baize, yerba, yellow tin foil, shaving bowls, and other goods.” The government responded by supplying blue glass beads, baize, little mirrors, small bells, ribbons of various colors, and blue woolen stuff from the royal warehouse, and purchasing aguardiente, hats, spurs, bridles, small tin basins, thimbles, large rings for reins, tobacco, yerba, and dried figs for the Indians. The entire shipment, worth 295 pesos 6 reales, was dispatched south, while funds to cover this expense were transferred from the Fund for the Redemption of Captives to the War Department. Preparations were also made to receive the child in the Buenos Aires orphanage while waiting to see if her parents or any relative claimed her. Although the government warned against “having captivity become a branch of commerce,” that, in part, is what it had always been. The very fact that the Royal Warehouse stocked such items as glass beads and small bells attests to an ongoing trade fueled by the ransoming of Spanish captives. In addition, captives were occasionally used in intertribal trade; Pampas Indians, who did not themselves practice formal bondage, provided the Araucanians with slaves.²⁰ Captives, ransomed back to the Spaniards or exchanged between aboriginal groups, provided a medium of exchange for Indian commerce.

The periodic return of captives to Spanish society could sometimes serve either as a ritual demonstration that an Indian group was willing to enter into peace negotiations with the local authorities or as an affirmation of that peace.²¹ Captives were also used by Indian groups as a vehicle for signaling their willingness to come into missions. In the mid-eighteenth century, for example, as soon as a provisional peace was signed between the Spaniards and an Indian group in which the Indians requested that a missionary be sent to them, they voluntarily released any captives living among them. As Governor José de Andonaegui reported to the Spanish government, when the Indians requested a reducción, “they bring with them, at the same time, a large number of Christian captives.”²² Indeed the Spaniards lost no time in interpreting this conduct as an indication of the Indians’ willingness to leave all “savage” customs behind.

The more than six hundred captives freed during Rosas’ lengthy campaign were taken from Pampas, Tehuelche, and Araucanian groups.²³ Military officials interrogated the captives one by one, asking a fixed set of questions to elicit basic information about them and their experience. Upon his return to Buenos Aires, the governor had a list of those freed printed and widely distributed in the hope of helping these men, women, and children to find their kinfolk. The published list is an excellent source of information on the demography of captivity. Each captive is described by name, sex, age, years in captivity, and ability to speak Spanish. Several entries are enhanced by more detailed physical descriptions. Because of the uniformity of the questions asked, the list of freed captives provides comparable and quantifiable data on the entire group. There is every reason to believe that those freed by Rosas were representative of a typical group of captives.

Demographic Analysis of Rosas’ List of Freed Captives

The most striking characteristic of those ex-captives who had not been born in captivity was that women outnumbered men by almost two to one. Of the total of 634 such individuals, 389 (61 percent) were female; 245 (39 percent) were male. Indeed the 1833 group of captives probably had a larger proportion of men than most captive groups, suggesting the possible presence of Spanish renegades among those Rosas classified as captives. In 1764, for example, the outgoing governor of Tucumán referred to 33 Spanish raids into the Chaco that had freed “20 male Christian captives . . . [and] . . . 240 women and young children.”²⁴ Another 73 unspecified “children born in captivity” (presumably mestizos) were freed in the 1833 campaign, bringing the total number returned to creole society to 707.

The overwhelming predominance of women in the captive group can in part be attributed to the Indians’ systematically taking women and children while killing men.²⁵ In the words of a mid-eighteenth-century Scottish traveler to the Río de la Plata, “the death of the men is certain if, by some misfortune, they are captured by the savages, for they sacrifice all the Spanish men except the children.” The same treatment of Spanish male captives was still in effect in 1803; according to a freed male captive, the Indians raided “to rob from the haciendas and to take the boys they encounter captive, bringing them up according to their customs, and killing the adults.” Twenty-five years earlier the viceroy of the Río de la Plata reported that “the Indians are so inhumane that they delight in killing, making no exceptions because of age or sex, and only sometimes reserving the life of the women, whom they take with them in order to indulge their abominable vices.”

Specific or approximate ages are given for 97.8 percent of the captives. The mean age for women is 21.3 years, while that for men is only 13.1. The same type of age discrepancy can be seen in the median age, 19 for women and 13 for men. When men and women are divided into ten-year age groups, below the age of 10 there were more males than females (see Table 1). The largest group of captives falls into the 10–19 age group; this group is also the modal group for both female and male captives. Above the age of 19 the male and female profiles differ greatly. For example, between the ages of 30–39 and 40–49 there are sizable groups of female captives but virtually no males.

Regrouping the data into two sets—0 to 14 (childhood) and 15+ (adulthood)—we can again see that while there were slightly more male children than female among the captive group, in the adult population women predominated (Table 2).

Only 35 percent of the female captives were children. The rest, indeed the largest group in captivity, were adult women. The next major group were white males below the age of 15. Among males only 38 percent were adults. While there was a slightly larger number of males among the total under-15 age group, the over-15 group was dominated by females. Thus Indian captives consisted of women of all ages and young boys. Even among the over-15 male captives, only four were above the age of 25. Interestingly, these four “older” men were somewhat atypical: two were Paraguayans and two, Chileans.

The Indian preference for female captives was probably based on a combination of sexual, strategic, and economic reasons. Possibly, women could help the Indian tribes replenish their population. Spanish women, like their Indian counterparts, were economically productive members of native society. They were more docile and physically easier to manage. Once captured by Indians they showed little tendency to escape back to Spanish society with reports of Indian military preparations, as did Spanish men. Of course, those Spanish women who had borne children while in captivity would have been even less willing to escape, as that would have obliged them to leave their children behind.

The data on age at the time of capture are far scantier, in part a result of the long years of captivity that dimmed the memory of those taken captive young (see Table 3). The average female was 16.2 years old at the time of her capture, while the average male was only 7.6 years old. It is interesting to note that female respondents had a much higher rate of recall, in part a function of their usually being older than males when taken captive. While 62 percent of the females questioned could give the approximate length of time of their captivity, only 37.3 percent of the males could supply the same information. Nevertheless, the data indicate that males were overwhelmingly boys below the age of 10 at the time of capture. Young children, both male and female, were attractive to the Indians because they could be more fully acculturated into Indian society; yet the data show a relative preference for capturing male rather than female children.²⁶ In other words, females were at risk to be taken captive at any age, while the older a male was, the more probable it was that he would be killed rather than captured.

Because of the relatively greater age at time ot capture among the female population, it is not surprising that a sizable number of women captured at age 15 or above were already married (21 percent or 52/ 248) or widowed (another 11 percent or 28/248) at the time they were taken. Indian raiders displayed no cultural bias against taking women who had been previously married or women with children. Indeed, women of proven fecundity might have been more attractive as prospective sexual partners.

Did female gender help assure better treatment once captured? At least one source suggests that neither native nor captive women were well treated, both being flogged “in a most barbarous manner” if they lost any of the animals under their care.²⁷ On the other hand, Spanish captive women were often taken as wives or concubines by a cacique or warrior among both the Chaco and Pampas tribes, although among certain groups, such as the Chaco Guaycuruans, captives had such a low status that only men who could do no better took them as mates. The practice of polygamy in Chaco and Pampas Indian society as well as among the Patagonian-based Tehuelches, especially among the caciques, made it easier to absorb women into the native family structure. Indeed among certain Pampas groups, having “many wives, many head of cattle and much silver” were all signs of power and wealth, and therefore of social standing. In some tribes the availability of Spanish women as mates allowed men to avoid the payment of the “bride price” that they would have had to pay for an Indian woman. Seen in this light, captive women represented an attractive alternative for Indian men of marriageable age.

The practice of Indian men taking Spanish wives was beginning to change somewhat in the early nineteenth century as larger groups of Araucanians from Chile came to dominate the pampas, restructuring the indigenous pampa tribes in the process of “Araucanizing the Desert.”²⁸ The Araucanians displayed a widespread cultural proclivity for creating male-centered myths about the sexual skills of females of another culture. They prized Spanish women for their special erotic talents and as a result tended to incorporate female Spanish captives into their society as slave-concubines, rather than as wives. Nevertheless, women held by Araucanized tribes as consorts or slaves also provided power, wealth, and status to their captors. The net result was that Spanish women, through one form of sexual liaison or another, formed bonds with their Indian captors that were usually not created between Spanish men and Indian women. Captured Spanish adult men were rarely allowed to take Indian wives, but rather forced to endure involuntary celibacy.

Even though captured in harrowing raids, many Spanish women came to identify with their captors, preferring to live among the Indians rather than return to “civilization.” This was especially true of women captured as young girls. The aforementioned anonymous Scotsman alluded to the case of two girls who were captured as young children and subsequently ransomed, but who soon afterward escaped from Spanish society to rejoin the Indians.²⁹ As early as the end of the sixteenth century Spanish soldiers came across captive Spanish women who had been completely acculturated into Indian society and who, when given the chance, preferred to remain with their so-called captors.

An instructive episode is provided by Luis de la Cruz, a Spanish military officer sent to survey a trans-Andean route between southern Chile and Buenos Aires in 1806.³⁰ Twenty days after leaving Santiago, between Guacaque and Puelee, a woman whom he first believed to be an Indian was brought to de la Cruz. Upon looking more closely the officer realized that she had Spanish features, and he proceeded to question her. Her name was Petronila Pérez, and she was a native of Pergamino, one of the forts along the Buenos Aires frontier. She was a captive of the Pehuelches and the wife of the Indian Mariñan, having been previously married to Carrilon, brother of the cacique, who had since died. Petronila recounted how she had been taken captive as a young child along with a sister and two stepbrothers in a raid along the Buenos Aires post road, in which her mother and stepfather had been killed by the Indians. De la Cruz, amazed at her ability to speak Spanish, asked her how she had come to learn it. “I’ve had dealings with other women captives who taught me how to speak as they did,” Petronila responded, testifying not only to the existence of a group of Spanish women captives within Indian society but also to their awareness of being linguistically and culturally different from their captors.

While the first part of de la Cruz’s interview with Petronila suggests a self-conscious attempt by Spanish women captives to preserve and transmit their culture, their subsequent conversation reveals other levels of complexity. It is interesting to note that de la Cruz himself could not decide whether to treat Petronila as a Spaniard or an Indian. He enticed her to return for further questioning by offering her “many gifts,” the traditional Spanish approach to influencing Indians. Petronila in captivity had lived in the Salinas area, a region traversed by annual Spanish expeditions to the salt marshes and a zone of increasing Spanish encroachment. She admitted that over the years she had seen several Spaniards, and that in fact every year her two brothers, who had subsequently been freed, came to visit her at her home. Clearly the frontier was a permeable zone with Indians visiting Spanish settlements and Spaniards visiting Indian ones. At this point de la Cruz could no longer contain his amazement. “Why didn’t you join them and return to the Christians?” “I didn’t want to leave because I love my children,” was her most human answer.

We do not know how Indian social mores dictated that a man treat his wife, or whether women, either Indian or Spanish, had any say in choosing their marriage partners.³¹ If captured while still young, as the above examples demonstrate, Spanish women could be integrated into Indian society well enough that they preferred it to the “Christian” world. This preference probably resulted from their loyalty to their Indian husbands and children, and from fear of returning to a Spanish world that might brand them as social outcasts.

Regardless of their motivation, their behavior was inexplicable to European men, who could only interpret it as a sign of feminine sexual passion and weakness. “They prefer to live like slaves and satisfy their passions, than reside among those of their race (so corrupt is human nature).”³² While women who preferred Indian life were licentious and corrupt, men who chose “captivity” over “freedom” were seen as outlaws or traitors. To the Spaniards, captivity was furthermore a punishment ordained by God; one female captive reported that her daughter had spent the last years as a beata in the House of Religious Retreat in Buenos Aires beseeching God that her mother be freed, and doing penance.

Both men and women captured by the Indians were expected to participate in the Indian economy. Among the Guaycuruans to the north, Indian women and captives of both sexes participated in spinning, weaving, preparing wild honey and carob beans for fermentation into intoxicants, and other domestic chores.³³ To the south female captives worked along with Indian women at herding livestock, mounted on horseback to tend the cattle and sheep day and night. Among those tribes that practiced agriculture, Spanish women were involved in cultivating wheat, barley, and beans. They probably also joined in the preparation of raw hides, wool, skins, tallow, grease, and ostrich feathers for trade to Spanish markets, as well as in artesanal production of woven fabrics, leather goods, and silver objects. Native women and captives were also responsible for all housekeeping chores, including cooking food, saddling horses, and setting up the tents (toldos) that served as native housing.

The Indians certainly chose their captives with a view toward who could best serve their needs when acculturated into their society. A modicum of physical preference may also have been at work in determining who would be captured or at least who would survive. Rosas’ list provides physical descriptions for 34 percent of the women (133/389) and 41 percent of the men (99/245) to aid in identification (See Table 4). An analysis demonstrates a strong preference for people described by the soldiers freeing them as fair-skinned and/or blond (rubio). Blue eyes (ojos azules) were also a popular feature. This description of the captive population is rather startling given the overwhelming predominance of dark-skinned (trigueño or moreno), dark-eyed settlers along the frontier. Analyzing the physical attributes by the sex of the captives, there is a suggestion that fair complexion, probably equated with exotic physical beauty, was even more prized in the choice of female than male captives.

In addition to those women described as “dark” were two slave women (one negra and the other morena), a morena ex-slave, and a parda. Among the men, one was classified as a mulatillo and another as a black. The captive group also included three hispanized male Indians and a woman described as having been born in the Abipon Reducción. At most this group of non-españoles numbered ten. The vast majority (98.5 percent) of the captives perceived themselves as racially Spanish.

An analysis of geographical zones supplying captives shows that the largest group of captives were people born in the province of Buenos Aires (Table 5). Providing half as many captives was San Luis province to the west of Buenos Aires. The next-largest number of captives were born in Chile and Córdoba. The small numbers of paraguayos, tucumanos, and san juaninos freed in the Bosas campaign is not surprising given that the captives found were all in an area to the south of Buenos Aires province, and thus relatively far from the northern Chaco areas. But the small number of mendozinos is surprising, especially in contrast to the relatively large number of captives born in neighboring Chile.

The vast majority of the captives were country people, inhabitants of the agricultural and stock-raising zones opening along the frontier. Only sixteen individuals (nine women and seven men) had been born in a city; all the others listed rural towns, estancias, and chacras as their places of birth. Their modest origins show in that only eight of them refer to their father by the title “Don,” a universal sign of respect, social standing, and at least a modicum of wealth in the society. Only one captive made any reference to owing property herself, and another identified her husband as a wagon driver and owner.”³⁴ Three city-born women, two of whom were related to arrieros, were taken while traveling from one city to another. On the whole, the captives were typical representatives of the rural population of the Spanish pampas, people of modest means who tended cattle or raised crops for an absentee landowner or perhaps themselves held small parcels of land. They differed from the rural population at large only in the overrepresentation of women in their midst.³⁵

Comparing information on place of birth and place of capture offers some insights into the rural population of the pampa (Table 6). Just as most of the captives had been born in Buenos Aires or San Luis, most were taken captive there. Those few listed as city dwellers were captured in the campo. The great majority of captives were country people taken captive in the very zone or region where they had been born, a reflection of low geographical mobility for the population at large. Seventy-one percent of the women for whom information is complete were captured in the place of their birth (127/180); for men the number was 64 percent (44/69). A group of male and female rural migrants from Santiago del Estero, Mendoza, and Paraguay had moved to the Buenos Aires-Córdoba-San Luis frontier in the hope of finding better economic conditions. In spite of the presence of female migrants, captures of women tended to occur in the region of their birth, suggesting less geographical mobility for the female rural population.

One hundred ninety-five respondents supplied even more specific information on where they had been captured. Overwhelmingly they had been taken while on an estancia or chacra (156 individuals), in a rural chapel (6), or along a road (8), that is, in the countryside. Another group had been captured in or near a fort (5) or in post houses (postas) (11). Only 7 captives described the place where they were seized as “in town,” while another 2 were found hiding in a coal shed. Those taken captive were overwhelmingly rural people, performing rural tasks. Their capture had probably taken place in much the same way that Andrés had been taken in 1803.

While it is difficult to determine what psychological processes the captives underwent during their capture and early captivity, the list of freed captives and other evidence provides some interesting suggestions as to the ability of the captives to survive as culturally Spanish.

One important indicator of Spanish cultural persistence was the retention of spoken Spanish. Although less than a perfect indication of culture, it is a surrogate variable. Among those freed in the Rosas expedition 106 people (or 16.7 percent of the group) could not speak one word of Spanish (Table 7). Another 77 were limited to at most a few Spanish words.

More striking is the difference in language retention between male and female captives. While at least 28 percent of the male captives (69/ 245) had suffered total language deprivation, the comparable percentage for females was only 11.6 percent (45/389). Females, who represented 61.5 percent of the entire group, were only 38 percent of those who had suffered language deprivation. Here, three factors seem to have been of capital importance: age at time of captivity, exposure to a sizable group of captives within Indian society, and the captor society’s attitude toward the group. Those captured young quickly forgot not only their native language but even the names of their mother and father. Conversely, those held with other captives were able to maintain their language in spite of youth and long years among the Indians.³⁷ Finally, Indian societies deemed women’s language to be different from, if not inferior to, that of men and seem to have tolerated Spanish women continuing to speak a different tongue.

There is much indirect evidence that some groups of Spanish women who spent much of their adult lives in captivity never lost their consciousness of being Spanish and their use of the Spanish language. The above-mentioned testimony of the captured Petronila Pérez, the woman who could speak Spanish because “other women captives” taught it to her, is evidence of the existence of groups of captives aware of their linguistic heritage and working to preserve it among other Spaniards. In the 1833 group, at least eight women testified that they knew their names, the names of their parents, or details of their capture, as well as their native language, because of information given to them by their “compañeras.” In some areas Spanish women captives seem to have been so numerous they almost formed their own subsociety, but apparently the same cultural or information network never functioned among male captives.

The Spanish language was also maintained by captives kept with other members of their families. Although this situation was rare, at least 85 captives were taken with at least one other family member. The largest family group freed was that of Doña Felipa Ortiz, a native of Antuco, Chile, and the wife of Don Pablo Castro. She was freed along with their four daughters and two sons, ranging in age from 22 to 6.³⁸ More frequent were the cases of mothers taken captive with one or two small children.

Given the predominance of women among the captives, it is not surprising that a group of children was born in captivity to Spanish mothers and Indian fathers. In addition to the 634 men, women, and children listed in the inventory, another 73 young children “who are at the side of their respective mothers” were also freed, and at least 2 more were left behind with the Indians.³⁹ Unlike those detailed in the published list, these children had been born in captivity.

Was one function of captives to help Indian tribes recover from their demographic losses? The data supplied by the 1833 list, while too incomplete to allow for sophisticated demographic calculations, provide some possible answers. The above-mentioned 75 children probably represent most of the surviving offspring of the female captives, as there is little reason to believe that Rosas was willing to leave more than a handful of these children with the Indians. Calculating the ratio of these surviving children to the number of women (210) between the ages of 15 and 39—the childbearing years—yields a rough estimate of .36 children born and surviving for each woman.

Information on age at captivity and age at return provides a rough idea of the number of woman-years spent between the ages of 15 and 39 among the Indians; that is, the number of years the fecund captive women were “at risk” to be impregnated by Indian males. The number of surviving children, 75, when divided by the total number of woman-years, 1,148, gives a fertility ratio of women to surviving offspring of .065. In other words, in any one year a captive woman had an almost 7 percent chance of bearing a child and having that child survive. While admittedly a rough calculation, this fertility ratio and the above-mentioned child-woman ratio suggest that captured Spanish women did not significantly alter the demography of indigenous society because of either low fertility or high infant mortality. The data also suggest two other possibilities. Perhaps captured women, although culturally assimilated through marriage into Indian society, were not particularly attractive as sexual partners for Indian men. It is also possible that Indian society was more concerned with loss of resources than with loss of population. If this were true, the Indians might have practiced some sort of birth control, perhaps infanticide or abortion, to prevent a surplus population from straining reduced resources or limiting the physical mobility of a nonsedentary tribe. Unfortunately we do not have enough ethnographic information to test these hypotheses.

While our data do not provide direct information as to whether, once captured, women had a better chance of surviving because of favored treatment, information supplied by the captives does allow us to calculate the average length of time spent in captivity (see Table 8). The average term of captivity for the entire group was 8.8 years. If we analyze time in captivity by sex we find little difference between the two groups, with women averaging 8.9 years and men 8.6. This suggests that once admitted into native society men and women experienced similar survival rates, perhaps the result of similar treatment.

Information on length of their captivity also allows us to trace an approximate chronology of Indian raids in the pampas. If Indian raids had been constant through the years, each succeeding year should show slightly fewer captives, due to the effects of mortality, especially among the older female population. But as Table 9 shows, the largest group of captives was those held for 10 to 15 years and taken during the tumultuous early years of the 1820s. Indeed, 54 captives (15.1 percent of the group) had been in captivity for 14 years (see Table 8). This group represents the survivors of those men and women taken during the Carrera-Ranqueles invasion of 1820, perhaps the most dramatic Indian attack on white settlements. The number of captives who had been among the Indians for 5 to 9 years was markedly smaller than the 10-14 or the 0-4 cohort, an indication that the raids had tapered off during the middle of the 1820s. The large number of captives taken after 1828 reflects the growing number of Indian attacks occasioned in part by a major drought that severely affected both the Indian and the Spanish economies of the pampas. These attacks prompted Rosas to undertake the 1833 campaign. Ironically, some of those taken captive in 1829 were victims of the Pampas Indian raids, which were probably carried out with the tacit support of the Rosas government.⁴⁰

Analyzing the captives by sex and length of captivity, it appears that Indian preferences for male or female captives changed over time. The sex ratio is approximately .3 males to every female for those held from 5 to 9 and 15 to 19 years. A very different pattern is found among those in captivity from o to 4 and from 10 to 14 years; that is, those captured between 1820 and 1824 or 1830 and 1834, years of intense combat along the frontier. During this period male captives are found in greater numbers, with more than .4 males for each female. Indeed during the 1820-24 period the number of males rises to almost .5 males for each female. These differing ratios suggest that during relatively peaceable periods Indians were mainly interested in taking female captives, while during periods of war, they took more male captives. These were, of course, young males who, the Indians perhaps hoped, could be assimilated and trained as warriors in a relatively short time. Taking adult male captives in time of war was never in the Indians’ best interest because of problems of physical control.

Conclusions

In the main, captives were people of rural origin seized in or near the place of their birth. The group was predominantly female, but male and female captives displayed markedly different age patterns. Males were usually captured young, while women of all ages seemed desirable to Indian captors. As a result, and perhaps also as a result of different patterns of socialization once captured, females seemed to retain the Spanish language and culture better than males. Paradoxically, women were probably better accepted by Indian societies, marrying native men and bearing their children.

The relative length of captivity experienced by all members of this group raises questions about their ability to readapt to Spanish society. Once freed, could these ex-captives reincorporate themselves into the world they had come from? This is a complex issue, dependent on the reactions of both the ex-captives and Spanish society. Throughout the period under study, male ex-captives seemed to experience little difficulty in reentering white society. Many of them were able to take advantage of skills learned during their years of captivity; they settled near the frontier, where they served as interpreters and guides. Their experience among the Indians equipped them to perform a vital service to the Spanish frontier communities.

In general, male captives seemed more eager to return to Spanish society than their female counterparts. Although far fewer men than women were held captive, among the group of Spaniards who over the years managed to escape from the Indians of their own volition, the vast majority were men.⁴¹ While this imbalance perhaps reflects a greater daring on the part of males, it also suggests that Spanish women were less unhappy in their condition of “cautivas” than their male counterparts. Women, in general, seemed less anxious to return to Spanish society, perhaps because this transfer—moving from a position as the wife of an Indian chief to that of a simple peasant—meant a loss in status. It is also doubtful whether those women, victims of “Indian captivity and sensuality,” received a warm welcome when they returned to Spanish society, with or without their halfbreed children. At least a handful of women always attempted to go back to the Indians after their so-called rescue. Ironically, the captive women apparently spoke more Spanish and probably remembered Spanish society better than did the men. Women had stronger ties to both sides of the frontier and must have faced a far more difficult choice when offered their freedom.

To what degree did Spanish families actively attempt to ransom their children from captivity? We have little direct evidence from either earlier captives or the Rosas group, and what we have is often contradictory. Some parents actively sought the release of their children from the beginning, welcomed the return of these children from captivity, and probably helped them to readapt to the Spanish world.⁴² But many of the female captives freed by Rosas were unable to reestablish links with their families and were placed under the charge of the Sociedad de Beneficencia in Buenos Aires.⁴³ The sex of the individual, the age of capture, the years spent among the Indians, the bearing of children fathered by Indian men, and the degree to which an individual had been integrated into native society all influenced the eagerness with which he or she sought to reidentify with Spanish society and the ease with which this society accepted the returnee.

The economic effects of captivity on either the Indian or the Spanish societies are hard to ascertain. At least one scholar has suggested that the labor of captive women helped Indian society to overcome a labor short-age, but little numerical data on either the numbers of captives or the size of Indian societies during the seventeenth and eighteenth centuries make this suggestion difficult to explore. The same author also points out that the 1833-34 ransoming of 707 captives from Araucanian tribes, which themselves totaled about 8,000, appears to have seriously crippled the native economy.⁴⁴ It should be remembered that only 340 of those captives were above the age of 14. Nevertheless, the loss of this number of productive adults may well have crippled native groups who were at the edge of subsistence most of the time.

How the loss of these people affected the economy of their home regions is even more difficult to ascertain, in part because of the lack of viable data on the population. The analysis presented in this essay shows that the rural population lost more individuals to captivity than did cities. The analysis further suggests that the people taken captive tended to be peones and small landowners, individuals who made an important contribution to the local labor force but were not necessarily perceived as essential. Furthermore, the scattered nature of the raids tended to diminish their economic impact. The province of Buenos Aires, for example, with a total frontier population of 9,239 in 1836, provided only 134 captives, less than 1.5 percent of its population, to the 1833-34 group.⁴⁵

Perhaps most perplexing is the relative lack of dramatic reaction to the continuous loss of settlers to captivity during the entire period under consideration. This silence prevailed perhaps because those most at risk to be taken captive were rural people, illiterate folk with little or no political power. Furthermore, because so many of the captives were women, their loss did not represent a dramatically visible reduction of the rural work force. Nevertheless, the fear of Indian raids, with their resultant death, destruction, and captive taking, served to dissuade frontier settlement. Although captives are infrequently mentioned in Spanish sources, their loss had a powerful psychological effect on frontier society. Regardless of the actual risk, Spanish settlers felt weak and vulnerable to Indian attack, and to the appalling alternatives of death or captivity, until the final decades of the nineteenth century.

The author would like to thank James Saeger, Kristine Jones, Juan Carlos Garavaglia, and John Juricek for their helpful comments on earlier versions of this paper.