ABSTRACT
The Brazilian Amazon is marked by intra-regional hypermobility. Two major population trends arise: the occupation of remote areas and a rush to the cities. Roraima state is testimony to both trends. Despite its remote location, Roraima was the fastest growing Brazilian State during the 1980s. Most of its growth, however, took place in urban areas. This paper casts light on the occupation and urbanization of this remote area. A survey conducted in the capital city, Boa Vista, provides information on migration histories and migrants' characteristics. Five major migrant types are delineated: peasants, garimpeiros, first time movers, repeat urban movers, and professionals. Current views of urbanization of the Amazon are focused on the displacement factor that development and state geopolitics have upon rural settlements and neglect the presence of some of these migrant types.

INTRODUCTION
The settlement history of Brazil engendered an uneven population distribution and a heavy concentration along the coast. Attempts were made by various administrations to occupy the empty interior for geopolitical reasons and to relieve the population pressure in the poverty-stricken Northeast region. The construction of Brasilia epitomizes these efforts prior to the 1964 military coup. The military governments of the late 1960s and 1970s took a more aggressive role in relocating the population by developing a daring plan for the occupation of vacant lands of the Amazon. The new strategy was centered on massive road building, free distribution of lots of land adjacent to the new roads, and financial incentive packages for those willing to invest in the region (Dawsey, 1992; Foresta, 1992; Fearnside, 1989; Sternberg, 1987; Sena Filho, 1983).

At a first stage, small-scale landless farmers benefited. By 1974, however, policies changed dramatically favoring large scale speculative investors from southern Brazil. As little was done to ameliorate the land tenure system and the overall infrastructure of the Amazonian migrant source regions, landless peasants continued to arrive in the region only to discover that land was no longer widely available. These land-hungry peasants settled unoccupied tracts of land, regardless of their ownership status, generating several conflicts (Sena Filho, 1983; Wood and Schmink, 1983; Becker, 1990).

The battle over the occupation of the Brazilian Amazon is being fought from two major stands: peasant front, composed of landless farmers; and capitalistic front, epitomized by national and multinational corporations as well as large scale farmers and ranchers (Martins, 1980). These two fronts permeate the far corners of the Amazon, advancing at different rates in different settings. When the two fronts clash, those fighting on the less powerful peasant side inevitably are displaced. These recurrent clashes engendered widespread human movement in the region culminating in two major population trends: the advancement of the peasant front to the most isolated sections of the Amazon (away from the tentacles of the capitalist front) and a rush towards the burgeoning urban centers.

Roraima (Figure 1), the northernmost Brazilian state, is testimony to such trends. The state is the most remote area of the country due to its distance from the great Brazilian centers and traditional [end p. 51]

migrant source regions (the Northeast and the South), and its poor transportation network. This faraway land experienced a tremendous population growth during the 1980s, primarily via migration, surpassing that of any other Brazilian state. This growth, however, has a paradoxical character: despite the rural nature of the pull factors (distribution of free lots of agricultural land, and remote gold and diamond alluvial mines), the bulk of the population growth took place in urban settings, primarily in the capital city of Boa Vista.

The urbanization of Roraima is not an isolated event, but part of a massive move towards Amazonian cities documented in a rapidly increasing literature (Aragon, 1980 and 1983; Martine, 1981 and 1984; Mougeot and Aragon, 1983; Mougeot, 1983 and 1986; Silva, 1986; Ferreira, 1987; Lavinas, 1987; Sawyer, 1987; Becker, 1990; Browder and Godfrey, 1990; Sawyer et al., 1990; Godfrey, 1990 and 1992). Notwithstanding impressive population growth, little is known of migration to Roraima; most studies dealing with the urbanization of the Brazilian Amazon are focused on Amazonas, Para and Rondonia states, and scarcely contain any reference to Roraima.

This paper illuminates migration toward and within Roraima during the 1980s and early 1990s. The myriad of studies on the Brazilian Amazon display a bias for environmentally related investigations. At this juncture we know more about the ecological impact of settlement than we know about the agents responsible for environmental change. Coming to grips with migrants and migration in the Amazon is of paramount importance if sound conservation plans and stable forms of settlement are to be developed.

RORAIMA
Roraima state encompasses 230,103 square kilometers, about 2.71 percent of Brazil. It is located in the northern part of the country, containing a portion of the Amazon rainforest, and sharing borders with Venezuela and Guyana. This forgotten frontier of Northern Brazil (Riviere, 1972) long characterized by sparse settlements and cattle ranching activities has been overwhelmed by a recent demographic surge. Roraima was the fastest growing state in Brazil during the 1980s, experiencing an average annual growth rate of 9.55 percent, compared with a national growth rate of 1.89 percent. The population almost tripled from 79,848 in 1980to 214,790 in 1991. The bulk of the population is concentrated in the capital city, Boa Vista, which accounted for 66.2 percent of Roraima's population in 1991 (lEGE, 1981 and 1992).

Primary economic activities are the basis of the economy in Roraima, employing the largest number of people. Agricultural production is based on subsistence agriculture. Properties are small ranging from 60 to 100 acres and for the most part have poor soils. Farmers are mostly migrants with low levels of educational attainment, small amounts of capital and no background or education on high yield techniques (Silveira and Gatti, 1988). Roraima State is endowed with various minerals, with current extraction restricted to gold and diamonds. These minerals, usually located in alluvial deposits with low exploitation costs, have been extracted by individuals rather than by licensed firms (Governo de Roraima, 1989).

Until the mid 1980s most population increase was planned via federal or state funded colonization projects, where migrants had access to free lots of land. It is estimated that an average of 2,400 migrants arrived per annum in the state during the 1980s attracted by the colonization programs (Silveira and Gatti, 1988). With the discovery of gold and diamond mines in the state in 1987 migration flows have been largely spontaneous, leading to uncontrolled forms of settlement (Governo of Roraima, 1989; Godfrey, 1992; Furley and Mougeot, 1994). An estimated 40,000 migrants were involved in the 1987-1990 gold rush, and many more worked in related activities (MacMillan, 1995).

Roraima mirrors many features and problems of other Amazonian areas: it has experienced fast development, massive road building, colonization programs, competition for land, destruction of natural vegetation, conflicts between indigenous groups and settlers, and between mining companies and garimpeiros (miners) (Furley and Mougeot, 1994). Roraima's growth reproduced at a smaller scale Rondonia state's boom during the 1970s. Roraima's boom is impressive given its isolation from major migrant-supplying regions and agriculture-consumer markets, as well as the less advertised nature of its colonization schemes (lEGE, 1992; MacMillan, 1995). [end p. 53]

MIGRATION CONTEXT
A groups of scholars advocate that development is the major force behind urbanization in the Amazon (Martins, 1975; Foweraker, 1981; Godfrey, 1990; Bowder and Godfrey, 1990. These scholars understand the development of frontier areas in the Brazilian Amazon as a multi-staged phenomenon. In the process local backward economies are gradually absorbed by the national one, and pre-capitalist societies are transformed into capitalist ones. A common assumption is that the settlement of remote areas is initially fostered by the exploitation of natural resources, which draws numerous migrants. With improved transportation networks and information flows, landless farmers looking for homesteads migrate into the region, pulled primarily by the abundant land. The exhaustion of natural resources displaces earlier settlers and shifts the core of the economy to agriculture, which becomes increasingly capitalized, and land ownership progressively concentrated. Later with the arrival of large-scale agricultural and ranching projects, former small land holders are displaced and forced to migrate into other remote areas of the region or into nearby towns.

Among the development factors leading to the urbanization of the Amazon are the mechanization of the local agriculture, the seasonal labor demands associated with the introduction of cash crops and ranching activities, the concentration of land tenure and widespread conflicts over land. According to this view, development leads to the rapid closing of the agrarian frontier vis-a-vis the consolidation of large land holdings, leaving settlers little choice but to subsist precariously in the towns. Migration becomes a survival strategy (Aragon, 1983; Martine, 1984; Mougeot, 1986; Silva, 1986; Ferreira, 1987; Lavinas, 1987; Sawyer, 1987 and 1989; Godfrey, 1990; Sawyer et al. 1990).

The government is undoubtedly the primary catalyst of development in the region responsible for road building, colonization programs, and large scale mineral and energy projects. But besides fomenting development, Becker (1985 and 1990) sees the government as the primary responsible for the urban character of the Amazon frontier. She postulates that the urbanization of the Amazon is an outcome of the government's strategy of fast occupation and development of the region. According to her, the government induced migration into the Amazon, offering job opportunities in large scale mineral projects, and highway and dam construction. Once such projects were accomplished, a great mass of jobless workers sought employment in nearby cities. The colonization programs developed by the government also attracted various migrants, and although access to land in these programs was relatively easy, most never gained access to land titles. This situation hampered peasants' access to credit, leading to the loss of their land and subsequent migration into urban areas. Moreover, suitable infrastructure was never fully developed by the government in most colonization areas, leading to massive departure to cities.

With this strategy, the government sought the development of congregation centers for displaced migrants in urban areas of the Amazon (Becker, 1985 and 1990). Such urban places are conceptualized as centers of concentration and redistribution of Amazonian displaced migrants, a work-force characterized by its mobility and flexibility. This flexible work force greatly facilitates the occupation and development of the region, as they can easily circulate among various governmental projects and private enterprises, performing different activities.

These views about the urbanization of the Amazon connote the presence of widespread movement in the region. Underlying this massive movement are processes of chain, step and repeat migration. The notion of chain migration implies that there are active (primary) and passive (secondary) migrants, leaders and followers (White and Woods, 1980). The primary group is comprised of innovators who make the first moves from a given origin. These innovators after achieving "stable" conditions at destination are followed by members of the secondary group and "migration builds gradually to a crescendo" (Stearnman, 1985). Lee (1966) argues that migration tends to take place within well-defined streams. Overcoming a set of intervening obstacles by early migrants lessens the difficulty of passage for later migrants.

Ravenstein (1885 and 1889) maintained that migration takes place in steps. In step migration, the migrant often seeks intermediate "steps" along the rural-urban continuum which afford a gradual adaptation to the urban environment. Accordingly [end p. 54] the migration system is comprised of a series of moves, which may be rural-rural, rural to small town, small town to larger city or large city to metropolis (Stearman, 1985). As migrants move from placeplace though the urban hierarchy, they become exposed to information about opportunities occurring at the next level in the hierarchy. Step migration has been described by many researchers in different regions (Ghersi and Dobyns, 1963; Pool, 1968; Whiteford, 1972; Thomas, 1972; Orellana, 1973; Stearnman, 1985; Sewastynowicz, 1986).

Although involving multiple moves, repeat migration does not necessarily entail movement up the urban hierarchy. Repeat migration involves lateral moves and moves from the urban to the rural end of the continuum. The latter moves are epitomized by the occupation of frontier lands and colonization programs, where individuals move from a rural place to an even more remote location. Settlement of certain frontier areas in Bolivia (Henkel, 1971), Venezuela and Colombia (Crist and Nissly, 1973), and Brazil (Margolis, 1973; Muller, 1980; Mougeot and Aragon, 1983; Moran, 1984; Wood and Wilson, 1984; Henriques, 1984,1985 and 1986;) took place in this fashion.

DATA AND METHODS
Literature is scarce on Roraima migration, and mobility data gathered during the 1991 demographic census are yet to be disclosed. In order to examine migration to and within Roraima, I personally interviewed 146 migrants in Boa Vista during the summer of 1993. Based on a stratified random sample, interviews were conducted in 27 subareas of the city in order to provide information across social classes and to minimize sampling bias. Clusters were associated with each of the 27 suburbs of the city and subsamples were drawn proportionately to their population. Once the number of interviews to be conducted in each suburb was established, buildings were randomly selected from each suburb with the help of a grid pattern and detailed city maps provided by SUCAM, the government agency responsible for malaria control. Buildings found empty during the survey were replaced by other drawn buildings. The survey was restricted to household heads, found in the selected buildings.

RESULTS
Echoing the migration experiences of other Amazonian states, the bulk of Roraima migrants (61.6%) was born in the Northeast. The Amazon was also an important source region (Table 1). Two prominent migration flows originating in the Northeast overwhelm the system: one from the rural areas of Maranhao state (Figure 1), where over one third of migrants were born; and another from Ceara state, where 17 percent of migrants were brought up. These streams developed during the 1960s and 1970s with the advertisement of colonization programs, have continued during the 1980s (Silveira and Gatti, 1988). Although the majority of Roraima migrants were born in the Northeast, the bulk of them moved into Roraima from other states in the Amazon region, suggesting a great deal of repeat migration (Table 1). In fact, most migrants made at least one move before arriving in Boa Vista; the average number of moves among migrants was 2.76.

These findings contrast with Bentes' (1986) study of migration to Manáus, Amazonas state. The major migrant source regions for Manáus were cities in the Amazonas state (56.7%), followed by Pará (11.5%) and Ceará (9.9%) states. Also Sawyer and Carvalho's (1986) study of the demographic surge in Rio Branco, Acre, point to a significant intra-state migration. Fifty-one percent of Rio Branco migrants originated within the state of Acre; whereas only 20.6 percent of the migrants came from the Northeast. These discrepancies testify to the profound changes taking place in the Amazon. Massive migration into Amazonas and Acre dates back to the rubber cycle of the late nineteenth and early twentieth centuries (Reis, 1966), in contrast to the more recent rush to [end p. 55] Roraima state. Amazonas and Acre experienced the actual occupation of the peasant and rubber tapping fronts for decades before their primate cities began receiving extensive number of migrants from rural areas. Boa Vista, on the other hand, attracted its migrants from both rural and urban areas of Roraima, but primarily from other states. The fast pace of development undergone by Boa Vista and Roraima in recent years is underlined by connections with a diverse set of places.

Migrants to Boa Vista are typically in their late twenties (average age 29.25). Most have little or no schooling with an average of 4.08 school years, and most come from a rural background. These findings echo the characteristics of migrants to other Amazonian states. Rondônia's migrants also had very low educational levels (average of 4.7 years), came from a rural background, and had an average age of 30.4 upon arrival in the state (Henriques, 1984, 1985, 1986). Such overall low human capital levels, make individuals more susceptible to the political and economic inconstancies of the frontier, leading to a significant number of moves.

Sixty percent of migrants had acquaintances in Boa Vista prior to their moves, and over half lived with acquaintances prior to, or at the time of the survey. Moreover, when inquired about the reasons for relocating in Boa Vista, 34.5 percent of migrants revealed that the presence of friends and family members in the city was the primary determinant of the move. Thus, chain migration is an important phenomenon embedded in the urbanization of Roraima, as it determines the destination of a sizable number of individuals.

Migrants' dependence upon acquaintances diminishes with length of residence in Boa Vista (Table 3). There is a clear length of residence threshold after which dependence on acquaintances diminishes considerably. After two years in Boa Vista migrants become better acquainted with local economic opportunities and better paying jobs, and are more likely to live independently from acquaintances. [end p. 56]

Boa Vista migrants are far from being a homogenous group and vary widely in terms of personal characteristics, migration histories, and rationale for moving into the city. After a careful examination of migration histories, five major migrant groups were derived: first movers, professionals, repeat urban, peasants, garimpeiros. These groups encompass about 88 percent of Boa Vista migrants. The remaining 12 percent fall in a residual category, from which no clear migration pattern, nor migrant characteristics can be delineated.

First Time Movers

Unlike their more peripatetic counterparts, first time movers are experiencing their first migratory event. First time movers are essentially followers, tracking the steps of their acquaintances and represent about one-fourth of Boa Vista migrants (Table 4). The vast majority of first time movers admitted having friends or family members in Boa Vista prior to the move and actually dwelled with them for different periods of time. First time movers are the youngest of all Boa Vista migrants with an average age of22 years upon arrival in the city. They are the second best educated group with 5.22 years of schooling and migrated into Boa Vista individually. The majority of first time movers arrive in Boa Vista from major urban centers, mostly state capitals of the Northeast and Amazon regions. A sizable minority, however, migrate into the city from rural areas of Maranhao state.

There is a bifurcation in this group. Those coming from a rural background tend to belong to families of displaced peasant, who after settling rural areas of the Amazon relocated in Boa Vista. Rural first time movers perform primarily low pay manual jobs in the city like construction work, gardening; while some even venture in gold mining. Their low educational backgrounds (3.3 years of schooling) hamper their access to high paying jobs in Boa Vista and relocation may follow as a survival strategy. Thus, moving into Boa Vista may represent only the first of a series of potential moves. First time movers originating in urban areas, on the other hand, are much better educated (6.4 years of schooling), and for the most part do not engage in any form of paying job in the city. Instead they move into the city to join family members and to study.

The currents of thought focused on the structural factors leading to the urbanization of the Amazon fail to account for the presence of first time movers in the region. Their moves into Boa Vista are determined primarily by the presence of acquaintances in the city. Nevertheless, geopolitical maneuvers and the advance of capitalism may have determined the move of first time movers' acquaintances into Boa Vista, influencing their arrival indirectly.

Professionals
The presence of professionals in Boa Vista challenges the stereotype of Amazonian cities as being exclusively a haven for the negatively selected and the displaced. Professionals encompass around 10 percent of Boa Vista migrants, and their educational attainment surpasses that of any other group (Table 4). They are likely to have a bachelor's degree or at least some level of technical training. They move irrespective of the presence of acquaintances at destination, and are drawn to Boa Vista to supply the demand generated by the growth of the city. They originate from a much broader set of places than other migrant types, including major cities of the South, Southeast, Amazon, and Northeast [end p. 57] regions. For the most part they migrate into the city accompanied by their siblings, and work as computer technicians, dentists, physicians and professors.

Prior to arriving in Boa Vista, most professionals made at least one other move, generally motivated by educational purposes such as the pursuit of a degree or a change in school. Given their young age (average of 29.57 years) and the fact that most of them have graduated recently, their migration into Boa Vista is essentially economically driven, being determined by the less competitive nature ofthe local markets and vast demand for their services. Like first time movers, the presence of professionals has also been overlooked by the theories and models of the urbanization of the Amazon. Nevertheless, the development of Boa Vista's labor market undoubtedly is the chief responsible for the arrival of such individuals.

Repeat urban
Repeat urban migrants arrive in Boa Vista in their late twenties (Table 4) and have moderate educational attainment levels, generally around 6 years of schooling. They originate in urban places of the Amazon and Northeastern Brazil and before arriving in Boa Vista make on average two other moves. Destinations are primarily cities of the Amazon and the Northeast regions. These individuals move irrespective of acquaintances and the primary force driving their moves is the quest for economic gains. They tend to be small scale businessmen or salespersons, who relocated in Boa Vista to take advantage of the wealth generated by gold mining.

One striking factor about their migration histories is the fact that they move into Boa Vista primarily from other major cities in the Amazon region, notably from Manaus and Belem. This repeat migration pattern among cities of the Amazon region substantiates current views of the determinants of urban Amazonia. The incipient development of the region, fostered primarily by the government has created uneven economic opportunities among Amazonian cities, and repeat urban movers move to maximize their gains.

Peasants
One in five Boa Vista migrants moved into the city from Roraima's colonization programs. Peasants arrived in the city with an average of 32.90 years of age, constituting the oldest migrant group (Table 4).

These individuals also display extremely low educational attainment, with an average of 2.93 schooling years. They originate from the rural areas of Maranhao and Ceara states and made on average two moves prior to arriving in the Boa Vista.

Facing severe economic hardship and having access to suitable arable land denied by the highly concentrated land tenure system of the Northeast, peasants leave for the Amazon. Their moves are driven by the hope of gaining access to land, generated by the accounts of acquaintances and advertisement campaigns for colonization programs. Migration chains are predominant among peasants, as the presence of acquaintances at potential destinations in the Amazon influences the destination choice. Distance is not a major barrier to the move, as they are willing to go as far away as necessary to get access to land. After a few days on the road, using a mix of transportation systems, these individuals along with their families arrive at one of the Amazonian peasant fronts.

Peasants remain at these locations until they are forced out by the arrival of the capitalist front, by environmentally related factors such as poor soils and malaria, or by the overall lack of infrastructure. As a survival strategy, peasants abandon their homesteads and move into even more remote areas. Peasants are likely to undergo repeated displacements (two among Boa Vista migrants) in search of land. After experiencing the volatility of the agricultural frontier, these individuals try their luck at the regional urban centers, where access to low paying urban jobs in the formal and informal sectors is relatively easy. Despite the urban character of Boa Vista, displaced peasants are not completely dissociated from rural activities, as on a seasonal basis they fill in the demand for labor in the adjacent cash crop farms or ranches.

When queried about the reasons for leaving rural areas of Roraima and other Amazonian states, peasants maintained that lack of land ownership and market for their products, low infrastructure levels, and the harsh environment were some of the major displacement factors. These claims corroborate with the displacement power of development and government neglect in providing suitable infrastructure in the region, embedded in the current views of urbanization of the Amazon. [end p. 58]

Garimpeiros
Garimpeiros (miners), display different migration patterns. They tend to originate in both rural and urban areas of the Northeast and come from different occupational backgrounds. They are the second oldest migrant group, and display meager educational attainment (Table 4). After hearing stories about auspicious miners, these industrious individuals migrate into the mining camps of the Amazon region guided by the hope of hitting the jack pot. Migration chains predominate during the first moves when incipient miners are learning their trade, but during latter moves garimpeiros relocate irrespective of the presence of family members and acquaintances.

Their lives are defined by the quest for gold, and they are peripatetic in nature, always on the move from mining camp to mining camp. Along their string of moves (3.9 on average), they display a certain degree of occupational mobility, as some may engage in activities other than mining. Most have contracted malaria multiple times, owing to poor sanitary conditions at the mining camps. Besides the Anopheles mosquitoes, garimpeiros have other bloodsuckers to deal with. Given the remote location of the mining camps, garimpeiros are heavily dependent upon the donos de garimpo (garimpo owners), who make fortunes providing them with overpriced goods and services such as air transportation, prostitutes, and malaria medication.

With the expiration of the precious metals and stones and the mechanization of mines, garimpeiros move on. Mining in the Amazon for the most part is undertaken illegally in Indian reservations, national parks, and private properties. Conflicts with local Indians, government agents, large scale farmers and ranchers have also the potential to displace garimpeiros. Mining in Roraima epitomizes the illegal character of the activity, given the heavy presence of garimpeiros in the Yanomami reservation during the late 1980s and early 1990s.

Rather than relocating their households in the surrounding mining areas, and eventually moving into Boa Vista, Roraima garimpeiros establish residence directly in the capital city. Boa Vista is a reference point for them, and the place where their families are settled and supplies are acquired. The city also works as the primary market for gold and diamond. Garimpeiros display a complex circulation pattern constantly moving between mines and Boa Vista (Mac Millan, 1995). Their migratory behavior cannot be explained directly by state geopolitics and development of the region. Instead, garimpeiros' mobility seems to be more influenced by the market for precious metals, and by the availability and accessibility of natural resources. They display the most complex migration histories. Migration chains are important during early moves, but diminish in importance in latter ones. Although step may occur among them, repeat migration and circulation are constants, as miners are continually on the move.

CONCLUSION
This study uncovered relevant aspects of migration in the latest booming area of the Brazilian Amazon: Roraima state. A survey conducted in Boa Vista revealed that migrants moved into the state primarily seeking free land, gold and diamond mining, but also to take advantage of the wealth generated by the mining activities in the city. Three-fourths of the migrants to Boa Vista are repeat movers. For the most part these individuals have low levels of education, and are in their late twenties. Roraima migrants are primarily born in rural areas of the Northeast, but migrate into the state essentially from within the Amazon region itself. Nevertheless, much variation exists among Boa Vista migrants in terms of socio-economic characteristics, migration histories and rationale for moving into the city. Five major migrant groups are delineated: first time movers, professionals, repeat urban, peasants and ganmpeiros.

Migration chains are important mechanisms in the occupation of Roraima, creating well defined migration streams between specific origins in rural Maranhao and Ceara states and Boa Vista. This process is more common among certain groups, notably peasants and first time movers, and virtually absent among professionals and repeat urban migrants. Identifying major migrant source regions can greatly assist government and private enterprises boost or curtail migration into a given area.

This paper illuminates the volatility of settlement in rural Amazonia, epitomized by peasants' migration histories. The insensible position adopted by the government to the recurring clashes between peasant and capitalist fronts in the Amazon, coupled with its negligence in providing suitable infrastructure in [end p. 59] colonization programs have fostered hypermobility, as individuals are constantly being displaced. Unless Brazilian authorities take a more active role in changing the current land tenure system in rural Brazil, granting peasants and small scale farmers suitable infrastructure and training, peasants are bound to have long sad migration histories.

The advancement of the capitalist front in Amazonia, besides displacing peasants to the cities, attracts a variety of migrant types to urban Amazonia. The stereotype of urban Amazon as a haven for negatively selected individuals just does not fit the reality. The current views of the urbanization of the Amazon are focused primarily on the movement of those engaged in resource extraction and subsistence agriculture, overlooking the presence and the determinants of the in migration of other groups into Amazon cities. Besides displaced peasants and garimpeiros, Boa Vista also attracts professionals, first time and repeat urban movers, from a vast array of places. Scholars studying this phenomenon should incorporate in their models and theories, the presence of this diverse pool of migrants rushing to the local cities and the determinants of their movement. Future studies should also explore the circulation at urban centers, b.elieved to be expressive among displaced peasants and garimpeiros.

REFERENCES
Aragón, L. 1980. Mobilidade dos Migrantes no Norte de Goiás. Seminário promovido pelo Setor de Pesquisa (SEPEQ) do NAEA (Nucleo de Altos Estudos Amazônicos) em 31 de agosto de 1978. Série Seminários e Debates, #4. Belém: NAEA.

____ . 1983. "Mobilidade Geográfica e Ocupacional no Norte de Goiás: Urn Exemplo de Migração por Sobrevivência," In 0 Despovoamento do Território Amazônico: Contribuçiões Para Sua lnterpretação. Mougeot, L. and Aragon, L. (eds.) (Belem: UFPAlNAEA) 91-122.

Becker, B. 1985. "Fronteira e Urbanizção Repensadas," Revista Brasileira de Geografia. (Rio de Janeiro) 47 (3/4):357-371.

__________ . 1990. Amazônia. (Sao Paulo: Atica).

Bentes, R. 1986. "Zona Franca, Desenvolvimento Regional e o Processo Migratório Para Manaus," in Migrações lnternas na Amazonia: Contribuições Teóricas e Metodológicas. Aragon, L and Mougeot, L. (eds.) (Belem:UFPA, NAEA, CNPq.) 220-254.

Bowder, J. and Godfrey, B. 1990. "Frontier Urbanization in the Brazilian Amazon: A Theoretical Framework for Urban Transition," Conference of Latin American Geographers, Vol. 16: 56-66.

Crist, R. and Niss1y, C. 1973. East from the Andes. University of Florida Social Sciences Monograph no. 50. (Gainesville: University Presses of Florida).

Davanzo, 1. and Morrison, P. 1986. "Return and Other Sequences of Migration in the United States," Demography 18:85-101.

Dawsey, C. 1992. "Migration in Brazil: Research During the 1980s," Yearbook, Conference of Latin Americanist Geographers, vol. 17/ 18: 109-116.

Fearnside, P. 1989. "Projetos de Colonização na Amazônia Brasileira: Objetivos Conflitantes e Capacidade de Suporte Humano," Cadernos de Geociências, vol. 2: 7-25.

Ferreira, I. 1987. "Expansão da Fronteira Agrícola e Urbanização," In A Urbanização da Fronteira. Lavinas, Lena (Editor). Series Monográficas #5, vol. 2. (Rio de Janeiro:lPPUR/ UFRJ):1-26.

Foresta, R. 1992. "Amazonia and the Politics of Geopolitics," The Geographical Review. vol. 82 (2): 128-142.

Foweraker, P. 1981. The Struggle for Land. (Cambridge: Cambridge University Press).

Furley, P. and Mougeot, L. 1994. Perspectives in the Forest Frontier, Settlement and Change in Brazilian Roraima. Peter Furley (ed.) (New York: Routledge).

Ghersi, H. and Dobyns, H. 1963. "Migración por Etapas: EI Caso del Valle del Virú," Migración e lntegración en el Perú, Editorial Estudios Andinos. Henry F. Dobyns and Mario C. Vásquez (eds.) (Lima: Monografía Andina no. 2): 152-59

Godfrey, B. 1990. "Boom Towns of the Amazon". The Geographical Review. vol. 80 (2): 103-117.

__________ . 1992. Migration to the Gold-Mining Frontier in Brazilian Amazonia". The Geographical Review, vol. 82 (4): 458-469.

Governo of Roraima, 1989. Perfil sócio-econômico do Estado de Roraima. (Boa Vista).

Henkel, R. 1971. The Chapare of Boliva : A Study of Tropical Agriculture in Transition. Unpublished PhD. dissertation. (University of Wisconsin).

Henriques, M. 1984. A Política de Colonização Dirigida no Brasil: Um Estudo de Caso, Rondonia. Revista Brasileira de Geografia (Rio de Janeiro), 46 (3/4): 393-423. [end p. 60]

__________ . 1985. "A Dinámica Demográfica de Uma Area de Fronteira: Rondônia," Revista Brasileira de Geografia (Rio de Janeiro) 47 (3/4): 317-356.

__________ . 1986. "Os Colonos de Rondônia: Conquistas e Frustrarações," Revista Brasileira de Geografia. (Rio de Janeiro) 48 (I): 3-42.

IBGE, 1981. Censo Demográfico de 1980. (Rio de Janeiro:Fundar;ao Instituto Brasileiro de Geografia e Estatfstica).

__________ . 1992. Resultados Preliminares do Censo Demográfico de 1991. (Rio de Janeiro: Fundação Instituto Brasileiro de Geografia e Estatística).

Lavinas, L. 1987. "A Agro-Urbanização da Fronteira," In A Urbanização da Fronteira. Lavinas, Lena (ed.). Series Monográficas #5, vol. 1. (Rio de Janeiro: IPPUR/UFRJ): 91-108.

Lee, E. 1966. "A Theory of Migration," Demography 3:47-57.

MacMillan, G. 1995. At the End of the Rainbow? Gold, Land and People in the Brazilian Amazon. (London: Earthscan Publications LTD).

Margolis, M. 1973. The Moving Frontier. University of Florida Latin America Monograph Series, No. 11. Gainesville: University Presses of Florida.

Martine, G. 1981. A Migração Repetida e a Busca de de Sobrevivencia: Alguns Paddies Brasileiros. Projeto de Planejamento de Recursos Humanos. Relatório Técnico #52. (Brasília).

__________ . 1984. Frontier Expansion, Agricultural Modernization and Population Trends in Brazil (Brasilia: IPEN IPLANI CNRH). Martins, J. 1975. Frente Pioneira: Contribuição para uma Caracterização Sociológica, in Capitalismo e Tradicionalismo, ed. Otavio Velho (Sao Paulo: Editora Pioneira): 43-5.

__________ . 1980. Expropriariio & Violencia; A Questiio Politica no Campo. (São Paulo, Hucitec).

Moran, E. 1984. "Colonization in the Transamazon and Rondonia," In Frontier Expansion in Amazonia, Marianne Schmink and Charles Wood (eds.) (Gainsville: University Press of Florida, University of Florida). 285-306.

Orellana, C. 1973. "Mixtec Migrants in Mexico City: A Case Study of Urbanization," Human Organization 32 (Fall): 273-282.

Pereira, A. 1990. Garimpo e Fronteira Amazônica: As Transformações dos Anos 1980, Belo Horizonte, CEDEPLAR, Master's Thesis.

Pool, D. 1968. The Number and Type of Conjugal Unions as Correlates of Level of Fertility and Attitudes to Family Limitations in Ghana. Paper presented to the annual meeting of the Population Association of America, Boston.

Ravenstein, E. 1885. "The Laws of Migration," Journal of the Royal Statistical Society, 48: 167-277.

__________ . 1889. "The Laws of Migration," Journal of the Royal Statistical Society, 52: 241-302.

Reis, A. 1966. A Amazonia e a lntegridade do Brasil. (Manaus: Edições Governo do Estado do Amazonas).

Riviere, P. 1972. Forgotten Frontier: Ranchers of Northern Brazil. New York: Holt, Rinehart and Winston.

Sawyer, D. 1987. "Urbanização da Fronteira Agrícola no Brasil," In A Urbanização da Fronteira. Lavinas, Lena, (ed.) Series Monográficas #5, vol. 1 (Rio de Janeiro: IPPURlUFRJ): 43-60.

__________. 1989. Urbanization of the Brazilian Frontier. Papaer presented in a Seminar on Urbanization in Large Developing Countries, IUSSP and Gujarat Institute of Area Planning, Ahmedabad, India, 28th of September to October 1st.

Sawyer, D. and Carvalho, J. 1986. "Os Migrantes em Rio Branco, Acre: Uma Análise a Partir de Dados Primarios," In Migrações lnternas na Amazonia: Contribuições Teóricas e Metodológicas, Aragon, Land Mougeot, L. (eds.) (Belem: UFPA, NAEA, CNPq): 112-147.

Sawyer, D. Torres, H. Pereira, A. and R. Abers. 1990. Fronteiras na Amazonia: Significado e Perspectivas. Belo Horizonte, CEDEPLAR.

__________.1987. "Amazon: A Backdrop to Contemporary Security and Development Issues,"Yearbook, Conference of Latin Americanist Geographers. vol. 13: 25-35.

Sena Filho, M. 1983. "Contribuções Para Uma Teoria de Migração Para a Amazonia. Notas de Discussão", In O Despovoamento do Território Amazonico: Contribuções Para Sua Interpretação. Mougeot, L. and L. Aragón (eds.) (Belem: UFPA/NAEA): 27-53.

Sewastynowicz, 1986. "Two-step migration and Upward Mobility on the Frontier? The safety valve Effect in Pejibaye, Costa Rica", Economic Development and Cultural Change, 34 (4): 731-753.

Silva, R. 1986. "Retenção e Seleção Migratorias em Rio Branco, Acre". In Migrações Internas na Amazonia. L. Aragon and L. Mopugeot (eds.) (Belém: UFPA, NAEA, CNPq): 182-219.

Silveira, I., and M. Gatti. 1988. "Notas Sobre a Ocupação de Roraima, Migração e Colonização," Boletim do Museu Paraense Emilio Goeldi, Antropologia, 4 (1): 43-64.

Stearman, A. 1985. Camba and Kolla, Migration and Development in Santa Cruz, Bolivia (Orlando: University of Central Florida Press).[end p. 61]

Sternberg, H.1987. "Manifest Destiny" and the Brazilian Amazon: A Backdrop to Contemporary Security and Development Issues," Confernece of Latin Americanist Geographers, vol. 13: 25-35.

Thomas, R. 1972. "The Migration System of Guatemala City: Spatial Inputs," The Professional Geographer 26 (2): 105-112.

White, P. and Woods, R. 1980. The Geographic Impact of Migration. New York: Longman Inc.

Whiteford, S. 1972. Bolivian Migrant Labor in Argentina: A Second Cybernetics Case. Paper presented at the 71st annual meeting of the American Anthropological Association, Toronto.

Wood, C. and Schmink, M. 1983. "Culpando a Vitima: Pequena Produção Agricola em um Projeto de Colonização na Amazonia. In O Despovoamento do Território Amazonico: Contribuções Para Sua Interpretação. Mougeot, L. and L. Aragón (eds.) (Belem: UFPA/NAEA): 70-90.

Wood, C. and Wilson, J. 1984. "The Magnitude of Migration to the Brazilian Frontier," In Frontier Expansion in Amazonia, Marianne Schmink and Charles Wood (eds.) (Gainsville: University Press of Florida, University of Florida): 143-152.

RESUMO
A Amazônia brasileira é caracterizada por urn alto nével de migrações intra-regionais. Duas tendências populacionais emergem: a ocupação de remotas áreas da região e movimento massivo para as cidades locais. Roraima é testimonha destas duas tendências. Apesar de ser uma area remota, Roraima foi o estado brasileiro de mais rápido crescimento populacional durante a década de 1980. A maior parte deste crescimento ocorreu em áreas urbanas. Este trabalho enfoca a ocupação e urbanização de Roraima. Uma amostragem junto aos imigrantes da capital, Boa Vista, proporciona informações sobre histórias migratórias e caracteristicas dos imigrantes. Cinco tipos de imigrantes são delineados. As visões sobre a urbanização da Amazonia, baseadas no poder desalojador do desenvolvimento e da geopolítica da região negligenciam a pesença de alguns tipos de migrantes identificados neste trabalho.[end p. 62]