INTRODUCTION
Despite the sophistication of sector-based theories of change, diversification of rural household incomes among various economic activities appears to be the norm throughout the developing world (deJanvry 1981, Connell et al. 1976,). Increasingly this diversification involves a spatial component in the sense that one or more of the household members migrates. Nowhere is this pattern of rural household change more apparent than in the Ecuadorian Andes. The land reforms of the 1960s and early 1970s that increased the number of smallholder agricultural households (minifundistas) not only led to a changing rural landscape but also encouraged the diversification of farm labor, initially focusing on short-term labor migration (circulation) to generate income to augment that of smallholder households. Starting in the 1970s and accelerating through the 1980s and 1990s, however, the nature of migration in south-central Ecuador changed dramatically as males began to emigrate illegally to the United States. Currently tens of thousands of people (mostly men) from the provinces of Callar and Azuay alone are working and living in the New York metropolitan area.

Here, the process, patterns, and scale of the international emigration phenomenon in south-central Ecuador are briefly detailed and interpreted.¹ This phenomenon is placed in the historical context of agrarian change and patterns of migration and circulation² in the sierra, drawing on field work from two subregions in Cafiar Province and linking it to the Ecuadorian census data to grasp the regional implications. United States Immigration and Naturalization Service (INS) data demonstrate the prevalence of Ecuadorian immigrants in the U.S. The character of the phenomenon is changing through time, spreading spatially and changing demographically as more indigenous highland households send their youth to the United States (New York) and more women emigrate from the area to join their male relatives.

SOUTH-CENTRAL ECUADOR
Ecuador is divided traditionally into three regions: the humid coastal lowlands (costa), the [end p. 63] sierra, and the Amazonian lowlands (oriente) (Figure 1). Guayaquil is the largest city as well as the principal economic and industrial center of the costa.

The region is dominated by export and plantation agriculture, especially bananas, coffee, and cacao. More recently, shrimp farming has increased among coastal mangrove swamps. The oriente, which is mostly lowland tropical rainforest, is sparsely populated, although the government has moved thousands of people into the region through [end p. 64] various colonization programs. The northern oriente is economically important because a large proportion of Ecuador's petroleum reserves are located there. Separating the two lowland regions is the sierra which consists of two rugged cordilleras, reaching over 6,000 m at the highest peaks. Quito and Cuenca, the second and third largest Ecuadorian cities, as well as most of the sierra population, are located in intermontaine basins. The costa and sierra share equally some 96 percent of Ecuador's 11.5 million population (1995).

The provinces of Cañar and Azuay, located in south-central Ecuador (Figure 1), have a predominantly rural population, despite the presence of Cuenca (250,000, 1990). Their combined 1990 population totaled 695,437 (Azuay 506,090, Cañar 189,347). Non-farm economic activity is concentrated in Cuenca, located at 2,500 m in the Tomebamba River valley, and attracts tens of thousands of permanent migrants and circulators from its rural hinterland.

The rural landscape of the two provinces is heterogeneous. Medium to large sized dairy operations occupy river valleys south of Cuenca. Outside the valleys, smallholder, semi-subsistence agriculture (minifundio) dominates, with landholdings averaging between one and four hectares (Chiriboga 1988).³ Most hillsides below 3,000 m are planted annually in intercropped maize and legumes, and in lower elevations (2,300-2,500 m), planted beneath fruit trees. Minifundio maize production is particularly prevalent north of Cuenca, in mestizo communities surrounding Azogues. Additionally, this minifundio maize subregion is also a zone of commercial handicraft production. For more than a century the inhabitants, especially women, have woven sombreros de paja toquilla, known colloquially as Panama hats. Though the practice has fallen off considerably in recent decades, some women still spend much of their day weaving Panama hats and other commercial handicraft items. The area of Panama hat weaving and minifundio maize production lying between approximately 2,300 m and 3,000 m is here referred to as Lower Cañar (Figure 2), although the population (predominantly mestizo) is more integrated into, and socially identified with, Azuay Province (Cuenca) than Cañar. North of Azogues lies a sparsely populated area of medium to large dairy operations that separates the minifundio maize zone from a distinctive area of highland, minifundio agriculture, referred to here as Upper Cañar. The majority of the minifundio population in Upper Cañar, who plant rotations of potatoes, grains (barley and maize), legumes (fava and kidney beans), and improved pasture, is indigenous.⁴ Agriculture is extended up to and sometimes into the páramo (moist, cold, ecozone commonly 3,600 m and above). Commercial handicraft production in Upper Cañar is minimal. The distinction between Lower and Upper Cañar is useful because each subregion has a distinct land reform history, ethnic identity, agricultural production, pattern of circulation and, as will be seen, pattern of international migration.

LAND REFORM: FROM SEMI-FEUDALISM TO MIGRATION AND WAGES
Sierra landholding is dominated by latifundio-minifundio, a highly inegalitarian remnant of semi-feudalism. For centuries, land was held by relatively few people in haciendas,⁵ while indigenous families (huasipungueros) worked four or five days on the hacendado's land in exchange for a small, often steep, unproductive piece of land called the huasipungo. In addition to the huasipungo, which typically was passed down through inheritance, huasipungueros also gained access to other resources such as firewood, pasture, and water. Many sierra households were not obligated by the hacienda/huasipungo system directly, but participated in sharecropping arrangements with the hacendado. This semi-feudal arrangement effectively controlled most land and labor in the countryside, and created a "captive population" (Carrasco 1990) that was spatially tied to the countryside. Only a relatively small population was not bound to the hacienda and was free to circulate to off-farm employment. And, where the hacienda/huasipungo system dissolved early (pre-1950), as in Lower Cañar, a higher percentage of the male population began to circulate, typically to coastal plantation jobs. Throughout the sierra, however, pre-1960s labor circulation was the exception.

The semi-feudalistic system began to unravel in the 1950s as indigenous protests increased and Ecuador's domestic market expanded. A semi-feudal labor force was deemed incapable of responding to food demands for an expanding urban population (Commander and Peek 1986). Land reform [end p. 65] legislation in 1964 outlawed the huasipungo system (although it persisted on many government-owned estates), and turned over huasipungo plots to the families who had farmed them (Waters 1995).

Many hacendados, anticipating a land reform, subdivided their estates into medium-sized estates (20-99 ha). By ending huasipungo labor relations and converting their estates into "productive" capital-intensive dairy farms (especially near Quito) which could be managed with minimal wage labor, hacendados kept the best lands and were exempted from land reform expropriation (Zamosc 1994). Land reform continued with a 1973 law and other land transfers during the 1970s, but only a small percentage of the rural population benefited from land reform (Waters 1995).⁶ The extreme concentration of landholdings [end p. 66] (latifundio-minifundio) in the sierra remained relatively unchanged. In 1974, for example, 78 percent of the rural land parcels were five hectares or less in size, but these holdings accounted for only 13 percent ofthe land (Waters and ButteI 1987). In 1984 approximately one third of all sierra agricultural lands were in holdings of 100 ha or greater and the average plot size was approximately 3 hectares (Chiriboga 1988).

Thus, while the 1964 land reform effectively terminated semi-feudal labor relations it maintained an inegalitarian land tenure structure, and created a large minifundio population. Because most minifundios did not provide full employment (Commander and Peek 1986) minifundistas were forced to find off-farm employment. These circumstances, combined with the draw of urban jobs, encouraged thousands of rural residents to migrate permanently to urban areas.

In Ecuador rural to urban migration increased considerably since 1950. Pichincha Province (Quito) in the sierra and Guayas Province (Guayaquil) on the coast received the most immigration. Pichincha is the only sierra Province that has experienced a positive net immigration since 1960; all other provinces have suffered a net emigration (Delaunay 1987, INEC 1991, 1985). In 1990, just over one fifth of the population (2.5 million) of Guayas Province and one third population (1.7 million) of Pichincha Province had been born elsewhere in Ecuador. More than 31,000 of Azuay and Cañar's population live in Pichincha, and nearly 52,000 live in Guayas. This permanent migration (which has become both rural to urban and urban to urban) continues. Between 1985 and 1990 more than 5,000 people from Azuay and Cañar moved to Pichincha, and more than 3,600 moved to Guayas (INEC 1991h, 1991p).

Notwithstanding this urbanward migration, most minifundistas remained in the countryside, worked small plots of land, and earned an increasing percentage of their income from off-farm seasonal or temporary migration (Commander and Peek 1986). Ecuador's oil-led economic growth of the 1970s created a demand for seasonal, unskilled labor in construction and service sectors in urban centers (mostly Guayaquil and Quito). Previously, men not bound to rural haciendas circulated to coastal locations to work at plantation jobs. With the disintegration of the hacienda/huasipungo system, circulation became an increasingly important component of reproducing the household, canceling debt, and maintaining agricultural production. Some communities reported that more than 70 percent of households had at least one person circulating, although the occurrence varied by community (Waters and Buttel 1987). Circulation permitted some households to purchase agricultural lands (Carrasco 1990, Jokisch 1993), but typically, migrants did not circulate to increase agricultural production through land purchase or major improvements. Rather, minifundistas have and continue to circulate during low labor demands in the countryside (or high off-farm demand) to augment income and to maintain agricultural production (Pachano 1988).

Circular migration is blamed for numerous rural pathologies resulting, most commonly, from household labor scarcity (Collins 1988, Painter 1986). Labor shortages are said to encourage a neglect of agriculture (Bebbington 1993) or a poverty of agricultural innovation, environmental degradation, damage to social organization that sustains agriculture, and over-burdening those who remain in the countryside (most commonly women) with labor that prevents them from performing all other necessary tasks.

The process of releasing men and women from the confines imposed by the hacienda/huasipungo system did not occur evenly across space, nor affect everyone uniformly. Brown and colleagues, examining the proposition that Ecuadorian land reform created "spatially differentiated effects on enterprises and individuals" (1988, 148), combined place variables (agrarian structure) with individual characteristics to determine how agrarian structure influenced decisions to migrate (permanent) or circulate (temporary). They found that circular migration was more common from areas with semi-subsistence agriculture, large farms, and "traditional" settings (rural, poorly educated, and highly agricultural), whereas migration was associated more with market agriculture, medium and small farms, and "modern" socioeconomic settings (better educated, larger service economy).

Circular migration may be associated with particular personal attributes and agrarian structural characteristics, but all types of migration are fluid [end p. 67] processes. As demand for temporary, unskilled labor moves from one economic sector to another, minifundistas develop a variety of off-farm employment strategies.⁷ The destination and the duration of circulation changes with the ebbs and flows of the Ecuadorian economy as well as by kinship, community networks, and gender roles. For example, male minifundistas from Azuay and Lower Cañar provinces (Figure 2) developed a tradition of circulating to coastal Ecuador to work on sugar (zafreros) or banana plantations dating to the 1950s (and earlier in some places) because Ecuador was in the midst of an agro-export boom, temporary labor was in demand, and the ties of hacienda/huasipungo had been dissolved. During the 1970s, export agriculture suffered and many jobs were replaced by mechanization. As plantation jobs declined, Ecuador urbanized rapidly with a demand for temporary, unskilled labor in cities (Commander and Peek 1986). Men from Cafiar and Azuay adjusted their circulation accordingly to work as albaniles (construction laborers) in Cuenca and Guayaquil.

Elsewhere in Latin America sex ratios (number of men to women) indicate that more women than men migrate to, or circulate through, urban centers for employment (Chant and Radcliffe 1992). This pattern is not as evident in Ecuador. Guayaquil has a nearly one to one ratio, and Quito has only a slightly larger percentage of women than men (INEC 1991 h, 1991p). Research shows that men constitute the majority of those circulating (Brown et al., 1988, 158) and migrating internationally (Borrero 1992, 80, Carpio 1992, 90), creating a considerable sex imbalance in Away and Canar Provinces. As will be demonstrated below, this imbalance is due more to male international emigration and less to female rural-to-urban migration or circulation.

Women do work off the farm, however. The recent rise in non-traditional export agriculture exports (NTAE) in Ecuador is an example of women's increasing participation in the rural work force (Thrupp 1994). Two NTAE's are shrimp and flowers, particularly roses. Men from Upper Cañar and other sierran locales circulate to work on shrimp farms (Jokisch 1993), but thousands of women (and men) commute (returning home in the evening) to "flower" plantations (Thrupp 1994, Waters and Salamea 1994).⁸ Interestingly, when men circulate or migrate, women are left to devise household and agricultural strategies, yet when men remain on the farm, female household members are always present.

Labor circulation strategies have been significant to the migration phenomenon in the sierra, and thousands of households remain engaged in it. Over the past 20 years, however, this strategy has declined in south-central Ecuador (Carpio 1992), even in the more remote regions of Cañar. In its stead, households turn to long-term, undocumented (illegal) international migration to the United States.

INTERNATIONAL EMIGRATION FROM SOUTH-CENTRAL ECUADOR
Starting in the 1970s, and increasing through the 1980s and 1990s, thousands of households from south-central Ecuador have sent a member of the family to the United States, almost exclusively to metropolitan New York. A 1990 survey reported that 33.8 percent of all households in the region had at least one family member in the United States (Instituto de Investigaciones Sociales [IDIS], 1990 survey in Borrero 1992).⁹ This family member was, in most cases, relatively young and male (Borrero 1992, Carpio 1992). The process of traveling undocumented to the United States has always been dangerous, financially risky, and expensive, but these elements have worsened in the past five years. Relatively early emigrants, those who left during the 1970s and early 1980s, needed approximately $1,200 (US 1980) to pay a pasador (synonymous with coyote, a smuggler of people) to get them across the Mexican-U.S. border and on to New York. The trip usually entailed a flight to Mexico, for which the emigrant had a visa, and then a trip across the border in California, followed by a flight to La Guardia airport in Queens, New York. ¹⁰

Conditions have changed of late. In 1995, would-be emigrants had to pay a pasador a minimum of $6,000 (US) for a more difficult trip. Most rural (or urban) Ecuadorian households cannot procure $6,000 without taking a loan, which means incurring monthly interest rates of 4 percent to 8 percent! Loans are acquired a number of ways. Most commonly, money is borrowed from chulqueros (informal economy money-lenders), who are relatively wealthy people or return-migrants, often those who earlier migrated the U.S., and may have legal residency there. Alternatively, would-be migrants borrow from family and/or friends living [end p. 68] in the United States. Flying legally to Mexico before crossing the U.S. border is no longer an option because migrants cannot obtain Mexican visas. Most pasadores now take migrants through a number of Central American countries, where they border-hop, passing from one pasador to another. Once at the Mexico-U.S. border, emigrants become part of the well-known border "dashers," hoping not to get caught by Immigration and Naturalization Service (INS) border patrols. Once across, emigrants are increasingly driven across the country to New York because airports are watched by INS officials. Successful migrants arrive in New York where they use kinship and community networks to find employment, which ranges from menial, service oriented tasks (washing dishes) to sweatshops. The immigrant can then send money to his wife, or mother if unmarried, to pay the loan acquired for travel, build a house, and perhaps purchase land. ¹¹ If the migrant is caught along the way, he may be sent back. Lacking the means to pay-off the $6,000 debt, another loan and attempt to emigrate to the United States is often undertaken. Should the migrant fail to get employment in New York, his family's debt burden in Ecuador increases rapidly, putting his family at risk of losing whatever they used as collateral; if the migrant is from rural Ecuador, family landholdings invariably constitute the collateral.

EVIDENCE OF INTERNATIONAL MIGRATION
Ecuadorian emigration can be detected both in the source region and in the destination, New York. Kyle and Sawyer (1992) estimated that in the early 1990s between 200,000 and 350,000 Ecuadorians lived in metropolitan New York City, which would make it the third largest urban concentration of Ecuadorians, after Guayaquil and Quito. The United States granted residency to 180,252 Ecuadorians between 1961 and 1995. Of that, 16,292 were granted amnesty under the Immigration, Reform, and Control Act (IRCA) of 1986 (INS 1992, 72) (Table 1). IRCA granted residency to two groups of people: 1) persons who could prove that they entered the United States prior to 1982 and had lived continuously for two years in the United States from that date, and 2) seasonal agricultural workers (SAW) who were employed in seasonal agricultural labor for a minimum of 90 days in the previous year.¹²

The INS estimates that 55,000 Ecuadorians were living illegally in the United States in 1996 (INS 1997, 183), which would bring the total number of first generation Ecuadorian immigrants (1961-1996) to 235,252. Including those granted residency under IRCA, these tIumbers indicate that only 71,292 Ecuadorians had immigrated illegally since the 1970s. This estimate is considerably lower than the 1990 University of Cuenca (IDIS) survey findings, which concluded that 80,000-100,000 Ecuadorians from Cañar and Azuay provinces alone had emigrated, mostly undocumented, to the United States. Part of this discrepancy may be explained by the INS underestimating the number of Ecuadorians who have entered the United State without visas. Of the 55,000 Ecuadorians living illegally in the United States, the INS estimates 50,000 resulted from visa overstays and only 5,000 crossed the border illegally (Warren 1997). Yet, in 1995 alone, 1,029 Ecuadorians were detained at the border or otherwise reported to have entered the country without [end p. 69] inspection (INS 1997, 165). This figure suggests that thousands of Ecuadorians are entering the United States illegally each year and that the number who have entered the states by crossing the border illegally is considerably higher than 5,000. Therefore, I suspect that the INS estimate of 55,000 is low, and the total number of Ecuadorians living in the United States without documentation must be closer to 80,000, boosting the total number of Ecuadorians in the United States to a minimum of 260,000, consistent with Kyle and Sawyer's estimate.

The majority of Ecuadorian emigrants originate from Azuay and Cañar provinces. This phenomenon is well-known in the area, and research (Borrero 1992, Carpio 1992, Kyle and Sawyer 1992) confirms that this area is anomalous for the high number of people who have left. These studies also show that the demographics of international migrants resembles that of circulators, relatively young and male, indicating that the cohort who circulated in the past has instead emigrated to the United States (Borrero 1992, Carpio 1992). In a 1988 study, Borrero found that 78 percent of international emigrants were male and most between the ages of 18 and 40. Carpio (1992) found that among the return migrants (from the U.S.) he interviewed, 88.9 percent were male and mostly between the ages of 17 and 35.

DETECTING INTERNATIONAL MIGRATION AT THE ORIGIN
Because international emigration has been predominantly male, sex ratio imbalances have developed in origin communities. This change is revealed in census data. When sex ratios, the number of men per 100 women, are mapped by province, and canton (county), a "crude" geography of migrant origins is revealed (See Figures 3-5).¹³ As shown below, Cañar and Azuay counties have sex ratios well below the sierra (or national) county average; in 1990, Cañar and Azuay reported an average county sex ratio of 89.9, while the sierra county average was 96.2 (lNEC 1991).¹⁴

In 1982, five counties in Azuay reported a sex ratio of 92 or below, one of which was below 80. In Cañar, two of four counties reported sex ratios below 92 (Figure 3).¹⁵ This pattern, or geography of migrant origins, indicates what anecdotal information and empirical studies support: emigration prior to 1982 was concentrated in Lower Cañar and Azuay Province, from predominantly mestizo populations (both rural and urban).¹⁶

In all, the 1990 census indicated more than 40,000 more women than men lived in Azuay and Cañar. Significantly, the census reported only 8,307 Cañar and Azuay residents residing temporarily in other provinces (INEC 1991). Thus, if we assume that all of these individuals were men (unlikely) and the census data are relatively accurate (likely), then the sex imbalance is adjusted to approximately 32,000. For the population, such an imbalance is anomalous to the rest of the country, providing indirect support for the estimates of international migration.

In 1990, falling sex ratios indicate that male emigration intensified in counties and parishes already demonstrating high emigration (Azuay Province), and this emigration spread to neighboring counties, most notably to Cañar Province (Figure 4). The Azuay and Cañar county average fell from 92.7 in 1982 to 89.9 in 1990. Five of thirteen counties reported between 84 and 91.9 men per 100 women, and another four counties reported below 84 men. To put this in perspective, only one county from elsewhere in the Ecuadorian sierra (54 counties) reported a sex ratio below 90. ¹⁷[end p. 70]

A more startling sex imbalance appears when the age group 20-39 is mapped (Figure 5). Five counties reported sex ratios below 70, meaning that in many communities less than 40 percent of the population between the ages 20-39 were men. The extreme was Biblian county, in Lower Caiiar, which reported nearly twice as many women as men (sex ratio = 52.3).

This evidence strongly suggests that international emigration increased considerably between 1982 and 1990, both in the "core" areas and in newer communities. Older emigrant communities continue to witness increasing male emigration, to the extent that few men in their 20s are left, and more communities are experiencing male emigration. As the sex imbalance grows in Upper Cañar, the census indicates what field work confirms; more indigenous men (and women) are emigrating from Cafiari rural communities.

In summary, many communities in Lower Cañar and Azuay province have experienced medium to long-term and extensive male emigration. By 1990, this pattern of male emigration intensified in Lower Cañar and can be detected increasingly in Upper Cañar. In addition to the spatial extension of this pattern, more women from Azuay and Cañar, but predominantly Lower Cañar, are joining this international emigration stream.

SUBSEQUENT EMIGRATION
The most recent evidence for regional international emigration data is provided by the 1990 census. International migration has continued, if not accelerated, since then. Although recent region-wide data are lacking, there is additional evidence to support the contention that more rural, indigenous men from Upper Cañar are emigrating to the United States. During field work in 1994-95, I witnessed nascent international migration networks develop in Upper Caiiar. Compared to Lower Cañar and Azuay the number of international migrants from Upper Cañar is low, but increasing. In one research commmunity in Upper Cañar (Chuguin), the number of male emigrants more than doubled during 1995. Just as telling is the recent opening of travel agencies/courier services in the town of Cañar.¹⁸ From November, 1994 to December, 1995 the first three agencies opened, although branch offices have operrated in Lower Cañar (Azogues) and Azuay for years.

A second, subsequent development, is the increasing number of women who are now emigrating to join male relatives in metropolitan New York. They have more commonly than not, migrated legally with their spouse (or father) sponsoring them sponsoring them from New York. In 1995, of the [end p. 71] 6,397 Ecuadorians granted temporary residency, 3,124 (48.8%) were sponsored by family members living in the United States, and 1,427 of these were the resident's spouse (INS 1997,47).¹⁹ Also, more women are emigrating illegally from Cañar and Azuay. Women face compelling reasons to emigrate, even though they are usually caring for children, managing a household, and if rural, continuing agricultural production. First, many men do not return from New York as the family expects or desires. People report bitterly that men abandon their spouses and families once they arrive in the United States, sometimes before even their debt is canceled. Because of the fear this generates, women take loans or solicit money from their spouses in New York to pay for illegal transportation. In addition to hoping to prevent abandonment in New York, female emigration is also an important household economic strategy designed to improve their economic standing in Ecuador, should they decide to return.

CONCLUSION
Ecuadorian land reform ended servile labor relations in the sierra, but did little to alter extreme land concentration or promote job creation in the countryside. Most rural households devised circular labor strategies to augment income and maintain agricultural production. This study demonstrates, however, that in south-central Ecuador, this pattern has been replaced by international emigration (mostly undocumented) to the United States. International emigration has snowballed in the past two decades and enveloped the entire region, leaving considerable sex imbalances in Azuay and Cañar counties.

Although international emigration replaced circular migration for many households, it constitutes a marked departure from circulation in several important ways, and should not be considered merely an extension, or natural progression, of circulation. First, circulation is commonly taken, in the absence of local economic alternatives, to reproduce the household and maintain agricultural production. International migration, however, offers the possibility of radically improving the economic position and cultural status of migrant households, something circulation has never accomplished. While some international emigrant households have sunk deeper into poverty from failed migrant trips or as a result of abandonment, most have benefitted economically, some substantially.

Second, the economic and cultural consequences of international emigration are numerous and much greater than circular migration. Household and community disruption caused by international emigration is much greater and longer-lasting than that caused by circular migration. Families are forced to adjust to the migrant's extended absence, and until the debt is canceled, to serious financial pressures. Yet, a migrant economy has developed, complete with chulqueros, pasadores, and travel agencies/courier services. Everyday, more people make plans to emigrate, motivated by the consequences of an migrant economy, and enticed by the migrant symbols of success, including automobiles, and in particular, modern, brick, or cement-block houses. These symbols demonstrate an improved economic standing, and prompt people, particularly the relatively young, to look to emigration for the future. This study has shown the development and growth of this emigration pattern, and because of the cultural and economic momentum for international emigration, thousands more Ecuadorians will seek their future in the United States. Although the future of emigration is uncertain, due in no small part to the relative success of immigrants in New York, the pattern will likely encompass more communities and include more indigenous people and women.

ACKNOWLEDGMENTS
I am grateful to numerous people in Ecuador who helped me during field work, particularly Enrique Duy and Byron Jaramillo of the Fundación de Desarrollo Agropecuaria (Fundagro), James Keese, the Hector Bravo family, and to the residents of Chuguin, Luis Cordero, and Cojitambo, Cañar. I am also grateful to Billie Lee Turner II, Sharon Moran, and David J. Robinson for their comments on earlier drafts of this article. I also wish to thank Nick Haan and Anne Gibson for their work on the graphics and Paul Robbins for critical comments. This material is based on a larger project supported by the National Science Foundation under grant number SBR-942304. Any mistakes are mine alone. [end p. 72]

NOTES
1. The causes of such massive population movements are the subject of considerable debate, and has been explained from nearly every social theoretical standpoint. This paper does not attempt to explain these causes or locate them within an explanatory perspective, something that is considered as part of a larger work not reported here.

2. Migration is a broad term used to describe numerous patterns of human mobility. Circulation is short-term (2-6 months), often seasonal, labor migration that does not constitute a permanent change in residence. This is distinct from permanent migration, which entails a permanent change of residence, and international emigration, which in this case, is long-term, although not necessarily permanent.

3. This 1984 average is calculated from plots 20 ha and smaller (Chiriboga 1988). Subsequent discussion of minifundio landholdings is also calculated this way. The average size varies considerably through the sierra.

4. The population in Lower Cañar is principally mestizo while Upper Cañar is predominantly indigenous (Cañari). The term, indigenous, is often a floating category (Knapp 1996), meaning that there are no clear and acceptable criteria and indigenous identity is often relative to who is making the description. Knapp classifies those areas where at least 33 percent of the population speaks Quichua as indigenous (Knapp 1991). This criterion categorizes rural Upper Cañar as indigenous, but not Lower Cañar, where Quichua is seldom used, even by elder residents. By self-description, outside identification, and extensive use of Quichua, the rural UpperCafiar population is heavily indigenous. The population of the small city of Cañar, located in Upper Cañar, considers itself and is considered mestizo.

5. Various diocese of the Catholic church also owned large agricultural estates that were nonetheless run in a similar fashion. In 1895, the "liberal revolution" of Eloy Alfaro expropriated many church lands in an effort to reduce the church's power, although the church continued to own haciendas throughout the sierra. The transfer of land from church to state did not end servile labor arrangements or concentrated land tenure. The government rented haciendas to arrendatarios (renters, in this case local elites), who maintained the privileges of hacendados, well into the 1960s.

6. Less than 15 percent of the rural population benefited from agrarian reform, which includes colonization of the oriente and sierra land reform (Waters 1995,5). The percentage of the rural population who benefited directly from sierra land reform is therefore less than 15 percent.

7. Circulation should not be seen only as a set of responses (adaptations) to a changing economy however. The focus of this paper is on migratory strategies taken by minifundistas and not on the structure of the Ecuadorian economy (nor international division of labor) which creates the demand for temporary, low-paying jobs. Nevertheless it is important to note that having a large minifundio population, which needs off-farm employment to reproduce the household, benefits employers. Temporary wages at construction sites and export agriculture are commonly below minimum wage and what would be demanded from a permanent labor force (Pachano 1988). This lower pay is possible because minifundista households rely on their own small plot of land to defray much of the cost of food (de Janvry 1981, Meillassoux 1981).

8. See Thrupp 1994 and Waters & Salamea 1994 for elaboration on Ecuador as an example of the "feminization of the rural wage force." Some women in Lower Cañar are working in low paying non-traditional agricultural export jobs, but this economic trend is much more common near Quito. Women commonly have a short-term contract, if at all (Waters and Salamea 1994).

9. This survey, conducted by researchers at IDIS (Universidad de Cuenca), interviewed 789 households in 1990. As discussed, the authors calculated that between 80,000 and 100,000 people from Azuay and Cañar had emigrated to the United States (Borrero 1992).

10. Some emigrants paid for forged documents, often without success.

11. Migrants usually send remittances to Ecuador by purchasing a cashiers check or money order and then sending it through a travel agency in New York that has a corresponding office in south-central Ecuador. In addition to international mail/courier service, these agencies serve other functions for migrant families including international phone service and electronic money transfers. Wiring money electronically to Ecuador, however, is prohibitively expensive for most migrants. After the debt acquired for travel has been canceled, migrants sometimes send objects of value, such as a VCR or winter clothes, instead of money, through these agencies. The number of such agencies has proliferated in Azuay and Cañar in recent years. Unfortunately, they will not divulge the magnitude (dollar figure) of the remittances they receive in New York and wire to Ecuador.

12. The number of IRCA and SAW residency recipients is accurate as of Feb. 13, 1992. Due to duplication of applications, the final number of recipients may be slightly lower than reported here (INS 1992, 72).

13. This technique may not be as illustrative in the next census because through time, more women are migrating, both undocumented, and documented, as they are sponsored by male relatives from the United States. The number of migrants will increase, but the sex ratio may decrease slightly.

14. Most oriente and costa provinces and counties report a greater percentage of men than women.

15. In Cafiar, two counties reported slightly more men than women, one of which is La Troncal. This county, while still pertaining to Cañar, is part of the coastal lowlands. It attracts circulators to plantation agriculture, thereby skewing the average sex ratio. In Azuay Province, one county reports more men than women and data were missing for 2 counties.

16. When examined at the parroquia (parish) level, the geography of migrant origins is even more noticeable. Eleven parishes (out of 20) reported a sex ratio below 92, nine of which are in Azuay Province.

17. Only three counties registered between 89 and 92 men per 100 women. One of these counties is Otavalo, Imbabura. Many men of the Otavalo indigenous group travel internationally selling sweaters and other handicraft items. This supports the argument that a significantly higher percentage of women than men is indicative of male international migration.

18. As mentioned, these agencies are the conduit for remittances and gifts from New York to Cañar and Azuay (as well as dozens of other countries for other immigrants). Numerous such agencies exist in metropolitan New York. [end p. 73]

19. As with other Hispanic/Latino groups, until recently, relatively few permanent residents have applied for U.S. citizen status. In 1977 4,063 Ecuadorians were admitted, but by 1993 only 1319 (32.5 percent) had become citizens (INS 1997). Approximately one quarter of all Ecuadorians residing legally in the U.S. are citizens of the U.S. (Carpio 1992,33).

REFERENCES
Bebbington, Anthony 1993. "Sustainable Livelihood Development in the Andes: Local Institutions and Regional Resource Use in Ecuador," Development Policy Review 11: 5-30.

Borrero, Ana Luz Vega 1992. "La Migración y la Movilidad en la Provincia del Azuay," Revista de Investigaciones no. 5, (Universidad del Azuay: Cuenca, Ecuador) 75-159.

Brown, Lawrence, Brea, Jorge A. and Andrew R. Goetz 1988. "Policy aspects of development and individual mobility: Migration and Circulation from Ecuador's Rural Sierra," Economic Geography 64: 147-170.

Carpio, Patricio Benalcazar 1992. Entre Pueblos y Metropolis; La Migración Internacional en Comunidades Austroandinas en el Ecuador, (Instituto Latinoamerica de Investigaciones Sociales (ILDIS Quito, Ecuador).

Carrasco, Hernán 1990. "Migración Temporal en La Sierra: una Estrategia De Recampesinizinación" In El Campesino Contemporáneo: Cambios recientes y los paises Andinos, ed. Fernando Bernal (Tercer Mundo Editores: Bogota, Colombia): 151-184.

Chant, Sylvia, Radcliffe, Sarah A. (eds.) 1992. Gender and Migration in Developing Countries, (Belhaven Press: London, New York).

Chiriboga v., Manuel 1988. "La Reforma Agraria Ecuatoriana y los Cambios en la Distribución de la Propiedad Rural Agrícola, 1974-1984," in Transformaciones Agrarias en el Ecuador, Gondard, Pierre et al. eds., (Centro Ecuatoriano de Investigación Geográfica: Quito, Ecuador): 39-57.

Collins, Jane L. 1988. Unseasonal Migrations; The Effects of Rural Labor Scarcity in Peru, (Pinceton, Princeton University Press).

Commander, Simon, Peek, Peter 1986. "Oil Exports, Agrarian Change and the Rural Labor Process: The Ecuadorian Sierra in the 1970s," World Development 14 (1): 79-96.

Connell, John et aI., 1976. Migration from Rural Areas: the evidence from Village Studies, (Oxford University Press).

deJanvry, Alain 1981. The Agrarian Question and Reformism in Latin America, (The Johns Hopkins University Press: Baltimore and London).

Delaunay, Daniel 1987. Migraciones Internas en el Ecuador 1950-1982, Documento de Investigación serie Demografía y Geografía de la Población no. 4. (Centro Ecuatoriano de Investigación Geográfica (CEDIG: Quito, Ecuador).

INS (Immigration and Naturalization Service), Statistical Yearbook of the INS 1990-1992, 1993, 1994. (U.S. Department of Justice: Washington D.C.) .

INEC (Instituto Nacional de Estadística y Censos) 1983. IV Censo Nacional de Población y III de Vivienda 1982: Resultados Definitivos (Quito, Ecuador).

____ . 1991a. V Censo de Población y IV de Vivienda 1990, Resultados Definitivos: Provincia del Azuay, July (Quito, Ecuador).

____ . 1991b. V Censo de Población y IV de Vivienda 1990, Resultados Definitivos: Provincia del Bolivar, November (Quito, Ecuador).

____ . 1991c. V Censo de Población y IV de Vivienda 1990, Resultados Definitivos: Provincia del Cañar, November (Quito, Ecuador).

____ . 1991d. V Censo de Población y IV de Vivienda 1990, Resultados Definitivos: Provincia del Carchi, November (Quito, Ecuador).

____ . 1991e. V Censo de Población y IV de Vivienda 1990, Resultados Definitivos: Provincia del Chimborazo, October (Quito, Ecuador).

____ . 1991f. V Censo de Población y IV de Vivienda 1990, Resultados Definitivos: Provincia del Cotopaxi, November (Quito, Ecuador).

____ . 1991g. V Censo de Población y IV de Vivienda 1990, Resultados Definitivos: Provincia del Esmeraldas, October (Quito, Ecuador).

____ . 1991h. V Censo de Población y IV de Vivienda 1990, Resultados Definitivos: Provincia del Guayas, August (Quito, Ecuador).

____ . 1991i. V Censo de Población y IV de Vivienda 1990, Resultados Definitivos: Provincia del Imbabura, November Quito, Ecuador).

____ . 1991j. V Censo de Población y IV de Vivienda 1990, Resultados Definitivos: Provincia del Loja, October (Quito, Ecuador).

____ . 1991k. V Censo de Población y IV de Vivienda 1990, Resultados Definitivos: Provincia del Los Rios, October (Quito, Ecuador).

____ . 1991l. V Censo de Población y IV de Vivienda 1990, Resultados Definitivos: Provincia del Morona Santiago, October (Quito, Ecuador).

____ . 1991m. V Censo de Población y IV de Vivienda 1990, Resultados Definitivos: Provincia del Manabi, September (Quito, Ecuador).

____ . 1991n. V Censo de Población y IV de Vivienda 1990, Resultados Definitivos: Provincia de El Oro, September (Quito, Ecuador).

____ . 1991o. V Censo de Población y IV de Vivienda 1990, Resultados Definitivos: Provincia del Pastaza, November (Quito, Ecuador).

____ . 1991p. V Censo de Población y IV de Vivienda 1990, Resultados Definitivos: Provincia del Pichincha, September (Quito, Ecuador).

____ . 1991q. V Censo de Población y IV de Vivienda 1990, Resultados Definitivos: Provincia del Sucumbios, November (Quito, Ecuador). [end p. 74]

____ . 1991r. V Censo de Población y IV de Vivienda 1990, Resultados Definitivos: Provincia del Tungurahua,September, Quito Ecuador)

Jokisch, Brad D. 1993. "Migrations and Women's Participation in Ecuadorian Agriculture: a Community Study," Mimeograph prepared for FUNDAGRO (Fundación para el Desarollo Agropecuaria).

Knapp, Gregory 1991. Geografía Quichua de La Sierra del Ecuador, 3rd edition, (Ediciones Abya- Yala: Quito, Ecuador).

Knapp, Greg 1996. Personal communication.

Kyle, David A., Sawyer, Ann, 1992. "The Health Impact of Return Migration from New York City to Ecuador," Migration World, vol. 21 (2/3) 39-42.

Meillassoux, Claude 1981.Maidens, Meal and Money: Capitalism and the Domestic Community, (Cambridge University Press: Cambridge).

Pachano, Sim6n 1988. "Campesinado y Migración: Algunas Notas Sobre El Caso Ecuatoriano," in Población, Migración y Empleo en el Ecuador (lnstituto Latinamericano de Investigaciipnes Sociales (ILOIS: Quito, Ecuador): 127-224.

Painter, Michael 1986. "The Value of Peasant Labour Power in a Prolonged Transition to Capitalism," The Journal of Peasant Studies, 13 (4) 221-239.

Thrupp, Lori Ann 1994. "New Harvests, Old Problems: Feeding the Global Supermarket", North American Congress on Latin America, vol. 28 (3) Nov/Dec. 22-27.

Warren, Robert 1997. INS, personal communication.

Waters, William F. and Lucia Salamea 1994. "Gender Issues in the Restructuring of Ecuadorian Agriculture," Paper prepared for presentation at the XVIII International Congress of Latin American Studies, (March 10--12: Atlanta, Ga.).

Waters, William F. Buttel, Fredrick H. 1987 "Diferenciación sin Descampesinización: Acceso a la Tierra y Persistencia del Campesinado Andino Ecuatoriano," Estudios Rurales Latinoamericanos, vol. 10 (3), 356-381.

Waters, William F. 1995. "The Agrarian Reform Debate and Indigenous Organization in Ecuador," Paper prepared for presentation at the XVIII International Congress of Latin America Studies Association, (September 28-30: Washington D.C.).

Zamosc, León 1994. "Agrarian Protest and the Indian Movement in the Ecuadorian Highlands," Latin American Research Review vol. 3, (3) 37-68.

RESUMEN
La estructura agraria del Ecuador ha cambiado profundamente en las últimas tres décadas. La terminación del huasipungo (sistema feudal) y substitución con relaciones capitalistas ha obligado minifundistas buscar trabajo transitorio afuera del campo, en las plantaciones de la costa y las ciudades serranas. Comenzando al final de los 1970, y aumentando durante los 1980 y 1990, la carácter de las migraciones cambió cuando hombres de las provincias Azuay y Cañar emigraron (la mayoría sin documentos) con destino en los Estados Unidos (Nueva York). Actualmente, miles de hombres (y un menor número de mujeres) de centro-sur del Ecuador viven y trabajan en la ciudad de Nueva York.
Este artéculo describe el proceso, el patron, y la escala de la migración internacional. Este fenómeno está presentado en el contexto de transformaciones agrarias y estrategias de migración (permanente y transitorio) en la sierra del Ecuador. [end p. 75]