Chapter I: Introduction
The Amazon Basin of South America is the watershed of the world's greatest river. It is roughly conterminous with the world's largest tropical rain forest, also known as the Amazon. About half of the rain forest is located in Brazil, the rest being shared by seven other countries. Because the largest part of the basin is in Brazil, the problems of the Amazon region are usually discussed in connection with this country, rather than Peru, Bolivia or one of the other Amazon countries. Since it was first encountered by Europeans in 1500, the Amazon has attracted explorers from throughout the world, has aroused scientific interest in many quarters, and has spawned fant astic stories of headhunters and cannibals. The Amazon has long been viewed both as a wild and mysterious place and as a store of vast riches that only awaited harvesting. During the closing quarter of the twentieth century, its ecological importance has been recognized, and it is seen by some as the last battle ground in a fight to save the planet.

This dissertation presents research into deforestation and the growth of urban places in Rondônia, a state in the western portion of the Brazilian Amazon that is widely know for its burning forests, cattle ranches, and gold rushes, but whose urban places are little-known, even within Brazil. This introductory chapter begins with a description of Rondônia’s physical and human geography, followed by a general discussion of tropical deforestation, including an overview of its principle causes in the Amazon region. The introductory chapter concludes with a discussion of the theoretical perspectives that helped to guide the research project and a pr esentation of the research question. Chapter II, "Setting Context," presents a review of issues relevant to the emergence and expansion of urban places in Rondônia, setting the stage for discussion of the impacts of that urban growth. Chapter III describes the methodologies used in the study. Findings are presented in Chapter IV. Chapter V presents conclusions and a discussion of where future research could be directed.

1. Rondônia, Brazil

 

Rondônia comprises 238,512 square kilometers, about the size of the U.S. state of Arizona and approximately five percent of the land surface of the Amazon region (Figure 1). It is bordered to the north by the state of Amazonas, to the east by the sta te of Mato Grosso, and to the south and west by the country of Bolivia. The lowland terrain of the valleys of the Rio Madeira in the north and the Rio Mamoré in the west give way to rolling, granitic hills in the center of the state. The eastern and southern portions of the state are dominated by river valleys ranging from 200 to 500 meters elevation. Soils include latossolo (oxisols) covering 45 percent of the state and podzólico (ultisols) covering 31 percent (EMATER-RO 1996, 1 9). The climate is tropical-wet, with summer-maximum precipitation. The dominant vegetation in Rondônia is upland wet forest, although varzea (flood plain) forest and cerrado (savanna) occupy nine percent and seven percent of the state, respectively (Martine 1990).

Note: Cities shown were município seats as of 1983, except for Rolim de Moura, which is shown here because it is discussed in some detail in the dissertation. No further municípios were added until 1990.

Sources: Geomapas, "Mato Grosso e Rondônia," 1:1,365,000, (São Paulo: Geomapas Produções Cartográficas Ltda., 1986) and Instituto Brasileiro de Geografia e Estatística (IBGE), " Estado do Rondônia," 1:1,000,000, (Rio de Janeiro: Centro do Serviços Gráficos do IBGE, 1985).

The capital of Rondônia is Porto Velho, situated on the Rio Madeira one thousand kilometers from its confluence with the Amazon and 2,200 kilometers upstream from the Atlantic Ocean. Even this far inland, the Rio Madeira is a formidable river (Photograph 1), a tributary of which, the Rio Mamoré, is navigable several hundred kilometers further upstream, where it forms the bo rder with Bolivia. Porto Velho is effectively the head of navigation, however, because of a series of twenty low but treacherous waterfalls, which begin with Cachoeira Santo Antonio, just upstream of the city. The "Old Port" from which Porto Velho gets its name was located at Santo Antonio, where the Portuguese had abandoned a fort built during its war with Paraguay in the 1860s. The city itself dates only from 1907, when the U.S. firm of May, Jekyll & Randolph began construction of the legend ary and infamous Madeira-Mamoré Railway (Silva Filho 1995). A telegraph link created in the same year established Porto Velho as an unusual node: linked by water to the Atlantic, by rail to Bolivia, and by wire to Rio de Janeiro.

Photograph 1: Rio Madeira at Porto Velho

Between 1960 and 1991, the population of Rondôni a increased almost twenty-fold, from 70,000 to 1.3 million (Table 1). The majority of the population is now urban, and the urban population is no longer limited -- as it once was -- to Porto Velho. Brazil’s military government had decided in the 1960s to promote rural development in Rondônia by paving the tenuous road that linked the old telegraph stations. When Rondônia achieved statehood in 1977, six of its seven município (county) seats were located on the new highway, the famous BR-36 4. As of 1996, the majority of Rondônia’s 52 município seats were found on or close to the same road.

Table 1: Rural and Urban Populations of Rondônia, Amazonia, and Brazil

 

	Population (1,000s)				Change
Region and Zone	1960	1980	1991	1960 to 1991
Brazil	70,070	119,002	146,917	+ 110%
Rural	38,767	38,566	36,042	- 7%
Urban	31,303	80,436	110,876	+ 250%
Percent urban	45	< P ALIGN="RIGHT">68	75
Amazonia	2,562	5,880	10,257	+ 300%
Rural	1,604	2,843	4,326	+ 170%
Urban	958	3,037	5,932	+ 520%
Percent urban	37	52	58
Rondônia	70	491	1,131	+1500%
Rural	39	263	473	+1100%
Urban	30	229	658	+2100%
Percent urban	38	47	58
Sources: Based on Godfrey (1990, 107) and IBGE (1992, 207).

1. Tropical Deforestation
1. Tropical Rain Forests

 

Tropical rain forests are the most productive of the terrestrial biomes, but they are typically found on oxisols, which are particularly nutrient-poor soils. These ancient soils are formed in warm, wet envir onments that receive up to several meters of precipitation each year. Accelerated leaches results, removing virtually all organic matter and mineral nutrients. Oxisols are characterized by high aluminum and iron oxide content, poor soil structure, and low soil fertility. A principal adaptation of rain forests to the poor soil environment is that most biological activity takes place above ground, so that nutrient cycling does not depend on the soil. The soil acts primarily as a mechanical substrate for the f orest, which is typically very complex and diverse, allowing nutrient exchanges without benefit of nutrient sinks in the soil. As a result, recovery or succession following a disturbance may be concsiderably slower than in a temperate forests, if indeed it is possible at all.

1. Causes of Deforestation in the Amazon

 

 
 
 
 
 

The Amazon rain forest has historically supported human populations only at low densities. Shifting cultivation has been the dominant adaptation to the ecological conditions of the rain forest described above. Also known as slash-and-bur n or milpa agriculture, it involves clearing a small part of the forest, releasing nutrients to the soil by burning the downed trees, and growing crops in the clearing for a year or two. The soil is typically depleted of nutrients within two to three years, and the process is repeated elsewhere. Where population densities are low, this approach works well, because the small, abandoned clearings can be re-colonized by the forest before they are needed again for cultivation.

Throughout the Amaz on, deforestation has resulted from the influx of large human populations attempting to engage in agriculture. Although small holders settling in the region have engaged in practices that resemble the slash-and-burn techniques, the greatly increased population density resulted in clearing that was not only quantitatively but also qualitatively different from the clearing achieved by indigenous populations because large patches are not as readily recolonized by wind-blown seeds as are small patches.

Settlement by growing numbers of small holders from other parts of Brazil, however, is certainly not the only cause of deforestation in the Amazon. Several other kinds of activities are responsible for significant losses of forest cover. These include commercial grazing, mining, hydroelectric power generation, logging, and even the production of charcoal as a fuel for heavy industry. The resources of the Amazon have been viewed as the means of solving problems in overpopulated parts of the country. Not only can excess population be shifted to the Amazon, but the powerful rivers and abundant minerals can also provide both energy and much-needed foreign exchange (Christian Science Monitor, September 22, 1988).

Cattle ranching has been a common cause of deforestation in the Amazon. The soils are not suitable for sustained grazing, but low land costs have enabled ranchers to buy land, clear it, use it for pasture for a few years, and then abandon it. Some of the inexpensive beef available in the fast-food and processed-food markets of the United States has come from land cleared of rain-forest cover, but selling beef is sometimes only a second or third motive for ranching in the Amazon region. The money to be made from land speculation or government subsidies has sometimes been more than the ranchers would make from selling the beef they produce (Hecht 1985; Hecht and Cockburn 1989). During early efforts to settle the region, for example, the Amazonian Development Superintendency (SUDAM) increased the advan tage of large operations by offering the equivalent of US $10 billion for projects of large agro-businesses in the Amazon (Mendes, 1985).

Mining and related activities have been another cause of deforestation in the region. Examples include the Grande Carajás iron mine in the eastern portion of the Amazon Basin (discussed further in Chapter II) and the many small gold mines operating in Rondônia, particularly during the 1980s. Even where the mining activity itself is not extensive, defo restation can result from the attendant short-term population growth, the opening of transportation corridors, and other ancillary activities.

Hydroelectric dams have been built that inundate vast areas of forest, displacing both wildlife and indigenous people. Although the discharge of the dammed rivers is great, the reservoirs are constructed in relatively flat areas, where vast areas must be inundated to achieve economically viable levels of hydraulic head. The environmental damage from such proj ects has tended to outweigh the economic benefits (Mendes 1985; Chernela 1988). Under international pressure, the World Bank curtailed funding for environmentally-damaging projects of this kind in the mid-1980s (Goodland 1985; Skillings 1985; Barbosa 1993).

1. The Importance of Tropical Deforestation
1. Reasons for International Concern

 

Attention to deforestation also stems from humanitarian concerns. The waste of primary productivity inherent in deforestation threatens the long-term economic development and food security of the poor (Colchester and Lohmann 1992). Penetration of rain forest by outside cultures has led to the demise of indigenous cultures (Perdigão and Bassegio 1992). Conflict over scarce resour ces in areas experiencing deforestation is often violent (Branford and Glock 1985; Mendes 1990). Finally, because a significant proportion of pharmaceuticals already are based on tropical plants, deforestation is threatening a vast pool of genetic information that may be useful agains such diseases as AIDS and cancer (Prance, 1985; Linden 1989).

The prospect of increased exposure to tropical diseases is a further reason for concern about tropical deforestation. In the 1990s, public attention was dra wn to tropical diseases -- such as Ebola -- that have surfaced within affluent countries of the temperate latitudes, including the United States and Germany (Garrett 1994; Preston 1994). Outbreaks of such diseases are likely to become more common, as the tropical rain forest regions in which many of them originate become further integrated into the world space economy. Rain forest regions throughout the tropics are connected to the rest of the world, both by the penetration of roads and by increasingly freq uent air travel. The increased vulnerability of wealthy populations in the temperate zones to the diseases of the impoverished tropics has helped to increase concern in the North about tropical deforestation.

1. Tropical Deforestation in a Global Context

1. Conceptual Framework

 

 
 
 
 
 

1. Realism
1. Important Attributes of the Realist Framework

 

Realism is a philosophy of science that takes as given the existence of social structures that constrain or enable individual actions. A realist analy sis seeks to explain the properties of these objects that "enable them to produce or suffer particular kinds of change" (Lawson and Staeheli 1990, 13). These structures are taken to have attributes that must be understood in order for the researcher to discover "what produces change, what makes things happen, [and] what allows or forces change" (Lawson and Staeheli 1990, 13). Even where these structures exist at a large spatial scale, their influence may be exhibited at the local level. Through qualitative research, local outcomes are understood in terms of the mechanisms that link them to the given structures (Lawson and Staeheli 1990, 14). The theories that best explain the large-scale structures may not be suitable for explaining the local outcomes. In order to explain processes that transcend spatial scales, therefore, a realist approach may require the incorporation of more than one theoretical perspective.

Realism presumes that social systems are open. Each social object stan ds in one of two kinds of relations to other social objects: necessary or contingent. A relation is necessary if an object cannot exist without it, as in the relation between landlords and tenants, neither of which can exist outside of the relation between them. Relations or conditions that are not inherent to the identity of a social object are contingent, and can yield variable results from a given social object exhibiting a single set of causal powers. In the landlord-tenant relation, for example, contin gent circumstances might include the race or sexual orientation of a tenant, which is contingent to the landlord-tenant relation, but from which may arise a variety of individual outcomes having nothing to do with the inherent properties of the landlord-tenant relationship. That is, whatever power relations are inherent between landlord and tenant, that relation will have implications that vary according to contingent circumstances. It is precisely the existence of these contingent relations, which are not predictable from the inherent nature of objects, that makes social systems open. Within a realist framework, therefore, social systems are not seen as subject to the experimental controls that are needed for prediction.

A degree of uncertainty is accepted as inevitable when attempting to predict the changes that will occur in one element of a social system as a result of change in another. Rather than attempting to make predictions about the future, therefore, realist research focuses on "unde rstanding how elements of the system are causally related" (Lawson and Staeheli 1990, 14). The openness of social systems requires that causal relations not be seen as laws within a realist framework, but rather as tendencies. Because a given social object will not always exist under the same set of influences and conditions, a variety of outcomes may arise from a single mechanism under different conditions. For this reason, realist research is not limited to formal patterns, but emphasizes processes a s well.

Some realists (e.g., Sayer 1984, cited in Lawson and Staeheli 1990) extend the distinction between global and local scales to imply that the global is abstract and the local concrete. Lawson and Staeheli reject this distinction, insisting that "abstractions are as important for analyzing local phenomena as for global" (1990, 16).

1. Methodologies in Realist Research

 

 
 
 
 
 

According to Sayer (1992, 1), meth odologies that rely on "hypothesis formation and testing, the search for generalizations and so on" persist, even though they may be incompatible with developments in theories of knowledge, which point to a completely different view of causality. A second problem for Sayer is that "so much depends in social research on the initial definition of our field of study" (1992, 2). The problem, he argues, is that causation is assumed to be "a matter of regularities between events," an d that any failure to model such regularities is seen as "allegedly inferior, `ad hoc’ narratives." This exposes an

erroneous view of causation. Realism replaces the regularity model with one in which objects and social relations have causal powers which may or may not produce regularities, and which can be explained independently of them. In view of this less weight is put on quantitative methods for discovering and assessing regularities and more on methods of establishing the qua litative nature of social objects and relations on which causal mechanisms depend (Sayer 1992, 2-3).

Realism provides an alternative to traditional research methods, in which what is concluded about a phenomenonmay be constrained by the framing of the research question and the delineation of a particular methodology. For Sayer, a given macro-scale socio-economic and political process will manifest itself in different ways under different conditions in the real world. The properties of s uch a process are necessary (inherent), but the conditions under which the process is articulated in the existential world are variable (contingent). Because social systems are open, a given process operating in two different places will not necessarily generate two similar outcomes.

In the present research project, it is important to extend Sayer’s analytical social-science framework to include what is commonly known as the "natural environment." Because environmental systems are in most instances closely linked with open social systems, contingent (that is to say, unpredictable) outcomes in environmental systems may arise from structures within social systems. In this context, generalization about such questions as the causes of deforestation may be impossible, so that research should focus on a qualitative explanation of the phenomenon in the place under study, even if the explanation will not directly explain similar outcomes in other places.

1. The Practice of Realist Research

 
At the methodological level, Sayer is careful to describe the differences between the extensive techniques required for generalization and the intensive methods associated with concrete research (Sayer 1992, 241-251). Extensive research seeks to identify patterns of common forms within a population of individual cases, whereas intensive (or concrete) research is concerned with "how some causal process works out in a particular case or a limited number of cases" (Sayer 1992, 242). Further differences between intensive and extensive research are presented in Table 2.

Table 2: Comparison of Intensive and Extensive Research
 

	INTENSIVE	EXTENSIVE
Research questions	How does a process work in a particular case? What produces a certain change? What did the agents actually do?	What are the common patterns of a population? How widely are certain characteristics distributed?
Relations	Substantial relations of connection	Formal relations of similarity
Types of groups studied	Causal groups	Taxonomic groups
Type of account produced< /TD>	Explanation of the production of certain events, though not necessarily representative ones	Descriptive, representative generalizations
Typical methods	Study of individual agents in their causal contexts, interactive interviews, ethnography; qualitative analysis	Large-scale survey of population or re presentative sample, formal questionnaires, standardized interviews; statistical analysis
Limitations	Unlikely to be generalizable	Limited explanatory power
Appropriate tests	Corroboration	Replication

The adoption of a realist perspective, then, may have important implications for methodology, particularly if an intensive, concrete project is undertaken. The differences begin with the research question, which will be focused on the explanation of a single outcome, rather than a comparison among numerous outcomes. Objects or groups will be identified on the basis of their connections to each other, rather than on similarities they may share. For this reason, where extensive research may rely on standardized interviews among representatives of a class of subjects, intensive research is not concerned with the representativeness of research subjects. Techniques such as rolling interviews, in which each interview subject might lead the researcher to the next subject, may be preferred. The result is an explanation of events that may not be generalizable to other cases, but which provides an explanation of the causes of the case in question.

Although Sayer acknowledges a role for ex tensive research, he advocates intensive methods of research as more likely to overcome the problems in social-science research that led him to the realist perspective. The qualitative methods of intensive research are not limited to ethnography. They include "structural and causal analysis, participant observation and/or informal and interactive interviews" (Sayer 1992, 244). "Instead of relying upon the ambiguous evidence of aggregate formal relations among taxonomic classes," as in ex tensive research, these methods allow the researcher to analyze causality "by examining actual connections" (Sayer 1992, 244). The advantage of a less standardized, interactive interview is that

the researcher has a much better chance of learning from the respondents what the different significances of circumstances are for them. The respondents are not forced into an artificial, one-way mode of communication in which they can only answer in terms of the conceptual grid given them b y the researcher. This also enables the researcher to refer to and build upon knowledge gained beforehand about the specific characteristics of the respondent, instead of having to affect ignorance (tabula rasa) in order to ensure uniformity or `controlled conditions’ and avoid what might be taken as `observer-induced bias.’ (Sayer 1992, 245)

Another key difference is the way in which individual cases are chosen for study or interview. In extensive research, samples must be selected in advance and according to a rigid set of criteria. Intensive research allows the study of individuals who may not even be representative of a larger population. Moreover, "they may be selected one by one as an understanding of the membership of a causal group is built up" by the researcher in a process of exploration, in which the research design is established, to a certain extent, as the research progresses (Sayer 1992, 244). Sayer, warns however:

This is not intended as a j ustification for empty-headed `fishing expeditions.’ It is just a counter to the rather peculiar idea that researchers should specify what they are going to find out about before they begin and an acknowledgment of the need to develop research procedures which do not inhibit learning-by-doing. (Sayer 1992, 244-245)

Elsewhere, Sayer acknowledges that "methodology needs to be critical and not merely descriptive," but he argues that often there are "no superior alternatives& quot; to explorative research (Sayer 1992, 4).

The adoption of a realist framework, particularly the intensive research methods described above, can inform the substantive practice of geographic research. A realist framework is appropriate for geographic research, because it links different spatial scales to "explain the social world in place at particular points in time" (Lawson and Staeheli 1990, 18). The adoption of the realist framework does impose some constraints on the researcher, t hough. Because "realists recognize a one-to-many correspondence between cause and effect, ... realists are forced to focus on processes rather than on patterns in their research" (Lawson and Staeheli 1990, 18). This precludes reliance on such extensive techniques as the analysis of spatial covariation or the use of formal surveys as central aspects of the research. This does not mean that quantitative methods must never be used, as "iteration demands that realists combine methodologies as dif ferent rounds raise different questions, some of which require qualitative and some of which require quantitative techniques" (Lawson and Staeheli 1990, 18). The ways in which these constraints have informed the present research are discussed as part of the presentation of the research question at the end of this chapter and again as part of the methodological discussion in Chapter III.

1. Theoretical Perspectives U sed in This Study

 

This dissertation describes a research project that integrates the qualitative methods of intensive research with the quantitative methods of generalization, both within the realist framework described above (Figure 2). As such, the research begins with qualitative research into the causal relations between supra-regional structures and processes operating at local levels and uses the findings of this analysis to create a series of hypotheses that are then used to test generalizations at a regional level that arise from the qualitative research. Lawson and Staeheli (1990, 15) have recognized the advantage of realism’s flexibility, which allows for the incorporation of aspects of several theoretical perspectives in a given research undertaking, as long as the theories used are compatible at some basic level. This section describes political economy and political ecology perspectives, which are used in this study, and carrying capacity, which is not. It al so describes substantive research on urbanization and resource use that guided the selection of variables for use in this study.

Urbanization is considered here as a process that is situated within a particular structure (the capitalist space-economy) and that may contribute to landscape change in Rondônia, Brazil. In order to describe that contribution, the project identifies specific mechanisms by which urbanization and landscape change are connected. This approach is similar to that taken i n much of the literature on Amazonia, particularly where political ecology is employed. The existing literature on change in the Amazon region, particularly in Rondônia, is therefore used to situate the process of urbanization within a broader framework. The empirical project of examining the mechanisms by which urbanization has led to change in the local landscape complements existing studies (described in Chapter II) that describe the causes of urbanization in the region.

1. Political Economy

 

Wallerstein’s world systems perspective, especially as described by Taylor (1989), is useful here, especially in three areas: the role of the State as part of an inter-state system, Brazil's status as a semi-peripheral state and the historical nature of geopolitics. Taylor describes the modern nation-state as part of a system that has evolved to promote the interests of industrial capitalism. In contrast to the instrumentalist view that each state serves the interests of groups within its territory, Taylor views each state as but one part of an interstate system, upon which the operation of the world economy collectively depends. In contrast to previous periods in world history, the modern (post-1500 or so) period has been characterized by the constant expansion of a single world economy until it has included nearly the entire world. The existence of sepa rate states, which are both variable in their factor endowments and compatible in their basic economic relations, facilitates both the flexibility and the mobility of international capital.

In the world-systems view, two major kinds processes in the development of the world economy have been core processes, by which capital accumulates at the expense of labor, and peripheral processes, by which labor creates surplus to be accumulated elsewhere by capital. It is these processes that have resulted in the rapid expansion of economies at the core of the world-system at the same time as peripheral economies were underdeveloped. Peripheral processes began with colonialism, but have persisted in other forms of economic domination, such as the accrual of interest on international loans (Barbosa 1993, 111-112). There are no semi-peripheral processes, but there are places -- and Brazil is one of them -- which appear to experience aspects of both kinds of processes. Semi-peripheral regions generate surplus for t he core but at the same time accumulate surplus extracted from the periphery.

Shannon’s (1989) review of world-systems theory includes some critiques that are instructive. It has been argued, for example, that world-systems theory suffers from a lack of historical accuracy. It has not been demonstrated that the world economy has operated to the real benefit of the center nor to the real impoverishment of the periphery; in fact, states often act in ways that are detrimental to capitalists. Moreover, world-systems theory implies a teleology or even a conspiracy that is difficult to accept. That is, if both the world economy and the world political system operate so consistently to benefit one small group, must there not be some coordinated effort? Is it possible that capitalists, each of whom is primarily concerned to maximize individual profits, can be operating with such a unity of purpose? An additional shortcoming of world-systems analysis is that it gives inadequate attention to resistance. The pro gress of the system and the furtherance of the goals of actors in the core are seen as inexorable and unamenable to change by historic processes. Shannon’s critique is well-taken, although it does not diminish the utility of world systems as an explanatory tool.

The realist approach taken in this dissertation allows world systems to be employed as a mode of explanation, without accepting the tendency toward teleology to which Shannon refers. The question of whether processes of exploitation actually benefit the core economies is not relevant to the question of whether these processes actually exist, and the existence of processes of exploitation does not eliminate the possibility of resistance. In fact, resistance is discussed in Chapter V of this dissertation in a way that is informed by the world-systems perspective. Brazil's position in the world economy is important to this research because it helps to explain economic and political decisions that have in turn resulted in important changes in the landscape of the Amazon region in general, and of Rondônia in particular. These decisions and the resultant changes are described in Chapter II.

1. Political Ecology

 

Deforestation in the Amazon region may represent an extreme dislocation in the relationship between humans and the environment, but it is not useful to cleave the reality of the place into the commonly received categories of "human" and "environmen tal." Rather, the reality of change in the Amazon region can best be understood as a singular manifestation of the many ways in which the development of the world political economy interacts with environmental systems at a local scale (Hewitt 1983; Millikan 1992). A political ecology perspective extends the analysis of political economy to encompass relationships with the non-human environment.

Bryant’s (1992) early review of the literature in political ecology is an effort to identify the majo r theoretical issues in this inter-disciplinary area. Political ecology is variously defined as the political economy of the environment or the international politics of the environment. In some sense, it is the logical extension of ecology to supra-natural systems. It is concerned, as is this dissertation, with interactions among three major spheres: the economy, politics, and the environment. Because the specific theoretical content of political ecology continues to defy definition, the term is employed i n the current work with respect to the kinds of interactions that take place among these three spheres.

Bryant sees three major kinds of questions emerging in the political ecology literature: the political context of environmental change, the struggle for resources, and the political consequences of environmental change. Each of these issues is very much current in the Amazon. For example, Brazil's internal politics have affected the environment in the Amazon both directly through the establishment of incentives to clear land and indirectly through the promotion of economic development in the region. Brazil's position as a debtor nation has been relevant to deforestation because the need to generate foreign exchange has been one of the motivations for these policies. The struggle for access to resources has included peasants or indigenous people who fight to preserve environmental resources under threat from miners, ranchers or others. The struggle has also involved peasants and indigenous people who want to continue to use their resources while outside environmentalists strive to protect it. The third area, political consequences of environmental change, is exemplified by the mobilization of rubber tappers in the Amazon and by the increased importance of the environmental debate in Brazil as a whole.

In this dissertation, the political-economic context of environmental change is emphasized. The link between the political-economic context on the one hand and environmental change on the other i s described in two segments. The first connects the political-economic context to the phenomenon of urbanization, and the second connects urbanization to environmental change. The first segment, which is well established in existing literature, is the focus of Chapter II. The second segment is the focus of the orginal research in this dissertation, and is the focus of the subsequent chapters.

1. Carrying Capacity

Carrying capacity is a precursor to the notion of sustainable development, and has at its core some of the same shortcomings (Worster 1995). It can be defined in a variety of ways, but most simply put, carrying capacity is the size of population that can be sustained on a given area of land without diminishing the capacity of that land to sustain the population. In range management terms, it is the number of livestock that can be sustained on a given parcel of rangeland without consuming more biomass than is produced on the land over a specific time interval.

The concept provides a simple metric by which to assess the impact of human settlement on the natural resources of a region. Even in the case of rural settlement, however, the concept poses a variety of difficulties, beyond the obvious ones of having to measure biomass production on the one hand and consumption by humans on the other. By focusing on quantities of resources, potentia l qualitative changes in the environment are ignored. This is arguably of less importance in the case of rangeland than in the case of rain forest, given the difference in biological diversity between the two. More importantly, the concept of carrying capacity implies a rather limiting dualism, in which nature and humans are seen as distinct categories. This dualism makes the concept of carrying capacity incompatible with the political ecology paradigm. As mentioned above, a realist perspective often requir es the use of several theories, but incompatible theories cannot be forced together ad hoc.

Finally, the paradigm of carrying capacity is not usable in the current research project because urban populations by their very nature draw on such a geographically dispersed set of resources that it would be impossible to make any direct connections between the productivity of local environmental systems and the sustainable size of human populations, even if all of the philosophical ramifications of such an exercise could be ignored.

1. Empirical Connections - Micro Level

The WRI report in cludes a section devoted to the impacts of urban growth on land conversion, including deforestation. It documents fuel consumption as the cause of increased deforestation in areas surrounding Dar es Salaam and cities in India and Burkina Faso. On the positive side, urban places worldwide occupy just one percent of the earth’s total land surface, and are therefore responsible for less direct land conversion than are agriculture, forestry, and grazing. For this reason, it is possible that urbanization may act ually help to ameliorate the process of land conversion in some situations. In the Republic of Korea, for example, rural households occupy six times more land per capita than do urban households. The pace of urban growth and the previous use of the land into which a city grows determine the impact of urbanization on land conversion adjacent to the city.

1. The Research Question

Until very recently, the small literature on urbanization in the Amazon has been focused on the causes of urbanization and social and economic conditions in the new - or newly expanded - urban places. This literature suggests the importance of "inter-urban and rural-urban linkages of the Amazon settlement system" and "the nature and the direction of landscape change" (Browder and Godfrey 1990, 63), but -- as with the literature on other frontier regions in developing countries -- it has not yet adequ ately addressed these issues (Brown et al 1994). That is to say, although rapid urbanization has been documented, neither its structure nor its impact on deforestation had not yet been addressed directly at the time the research reported in this dissertation was undertaken. Since the fieldwork for this dissertation was completed, however, Godfrey and Browder (1996, and Browder and Godfrey 1997) have addressed the questions raised in their previous work cited above. Their findings are described in the literature review in Chapter II, and are compared to the present work in Chapter V.

The focus of this research is the relationship between urbanization and deforestation in Rondônia. The principal research question is whether urbanization in Rondônia has ameliorated or exacerbated deforestation in Rondônia. This question arises from my reading of the literature on deforestation in Rondônia, in which migration of settlers to the cities of Rondônia has, in effect, remove d them from the scene as potential agents of deforestation. My reading has been that although they had quit the rural activities areas in which deforestation occurs, they had not quit the region entirely, and that their potential impact as urbanites in the region should be addressed. In other words, should the migration from rural to urban places be seen as the end of the deforestation story, or simply another chapter?

The question as it arises from this reading of the literature is not easy to answ er, even in principle. As the question is asked, it invites the response, "Compared to what?" If the comparison is to be between the impacts of urban growth and the impacts that would exist with no growth, then the answer would be clear without further inquiry: urbanization must exacerbate deforestation. If, however, the comparison is made between the urbanization and a comparable level of rural growth, then the question is at least answerable in principle, even though it is impossible to measure, since only one of the outcomes has actually occurred, and hence no concrete basis exists for comparison.

The realist framework described above suggests a third approach, focused on the linkages between the macro-scale processes and structures that led to urbanization and local-scale processes of landscape change. The selection of potential linkages suitable for study is informed by the World Resources Institute study mentioned above. Rather than compare the actual impacts of urbanization with theor etical alternatives, an effort is made here to assess the impact of urbanization on deforestation in Rondônia relative to the total amount of deforestation. I examine four mechanisms that link urban population growth to deforestation:

• land conversion for urban use
• food consumption in urban areas, leading to the clearing of additional land for agriculture
• wood consumption for housing in urban areas
• power consumption in urban areas, leading to the i nundation of land for hydroelectric power generation