<%BANNER%>

Farming and Patterns of Agrobiodiversity on the Amazon Floodplain in the Vicinity of Mamiraua, Amazonas, Brazil


PAGE 1

FARMING AND PATTERNS OF AG ROBIODIVERSITY ON THE AMAZON FLOODPLAIN IN THE VICINITY OF MAMIRAU, AMAZONAS, BRAZIL By MORGAN J. SCHMIDT A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2003

PAGE 2

Copyright 2003 by Morgan J. Schmidt

PAGE 3

ACKNOWLEDGMENTS This research was supported by a Charles Wagley Research Fellowship through the Center for Latin American Studies and Tropical Conservation and Development Program at the University of Florida. Graduate studies were supported by a Graduate Alumni Fellowship from the University of Florida through the Department of Geography. The research was further supported by the Mamirau Sustainable Development Institute. I would especially like to thank Jomber Chota Inuma who introduced me to the Mamirau reserve and contributed much encouragement, advice, and support during the research. I am grateful for the hospitality of Jombers family, Dina and Ortila, who gave me a home in Tef. I would also like to thank Niele Peralta Bezerra, Jos Mrcio Ayres, Edila Moura, and Ana Claudese Nascimento for their support of the research; and Jos Elizon Rocha da Silva, for his tireless efforts in the field. I am grateful to all of the other assistants, informants, farmers, and families (too numerous to name) for their help, cooperation, and wonderful hospitality. Finally, I would like to thank my family, whose support made my graduate career possible; and my committee members (Dr. Nigel Smith, Dr. Hugh Popenoe, and Dr. Michael Heckenberger) for their guidance, advice, and support. iii

PAGE 4

TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iii LIST OF TABLES.............................................................................................................vi LIST OF FIGURES...........................................................................................................ix ABSTRACT.......................................................................................................................xi CHAPTER 1 AGRICULTURAL BIODIVERSITY AND SHIFTING CULTIVATION: THE RESEARCH IN CONTEXT.........................................................................................1 Introduction...................................................................................................................1 Our Stake in Agricultural Biodiversity.........................................................................3 Call for Conservation....................................................................................................6 Swidden Model.............................................................................................................9 Why Study Swidden?.................................................................................................12 Swidden Agriculture in Time and Space....................................................................14 2 FOCUSING ON THE MIDDLE SOLIMES REGION...........................................18 Prehistoric Agriculture and Settlement on the Amazon River...................................18 Practical Aspects of Swidden Research in Amazonia................................................20 Amazon Caboclo........................................................................................................27 Middle Solimes Region............................................................................................28 Connservation of the Vrzea Ecosystem: The Mamirau and Aman Sustainable Development Reserves...........................................................................................30 Data and Methods.......................................................................................................31 Research Considerations of the Flood Season............................................................36 Life on the Vrzea......................................................................................................38 3 STRATEGIES FOR FARMING ON THE FLOODPLAIN: THE ROA...............43 Characteristics of the Roa.........................................................................................45 Farming Praias and Lamas..........................................................................................49 Diversity and Management of Restinga Baixa Fields................................................51 Restinga Alta: The Preferred Habitat for Planting.....................................................55 Crop Losses from Flooding, Pests, and Diseases.......................................................62 iv

PAGE 5

Storage and Loss of Planting Material........................................................................67 4 CAPOEIRAS: ENRICHED FORESTS AND MANAGED FALLOWS..................71 Restinga Baixa............................................................................................................73 Restinga alta................................................................................................................76 Bananal.......................................................................................................................78 Capoeira......................................................................................................................83 Capoeira Bananal........................................................................................................87 Capoeira Cacaual........................................................................................................90 Capoeira Fruteira........................................................................................................90 Frutal...........................................................................................................................93 Sitio.............................................................................................................................94 5 DIVERSITY IN HOME GARDENS.......................................................................100 Diversity in Middle Solimes Vrzea Home Gardens.............................................104 Losses from Flooding...............................................................................................114 Size of Home Gardens..............................................................................................115 Livestock in Home Gardens.....................................................................................117 Terra Firme Home Gardens......................................................................................118 6 A MORE STABLE ENVIRONMENT: TERRA FIRME........................................126 Field and Fallow Management on Terra Firme........................................................130 Agrobiodiversity in Terra Firme Roas....................................................................138 Agrobiodiversity in Terra Firme Capoeiras..............................................................143 Other Non-Roas......................................................................................................148 7 LESSONS FOR CONSERVATION OF AGROBIODIVERSITY IN THE MIDDLE SOLIMES..............................................................................................150 Distribution of Agrobiodiversity..............................................................................151 Theoretical Implications...........................................................................................155 Directions for further research..................................................................................158 APPENDIX A USEFUL SPECIES OCCURRING IN FIELDS AND FALLOWS.........................166 B USEFUL SPECIES OCCURRING IN HOME GARDENS....................................174 LIST OF REFERENCES.................................................................................................186 BIOGRAPHICAL SKETCH...........................................................................................199 v

PAGE 6

LIST OF TABLES Table page 2-1 Number of home gardens, fields, and fallows visited in 19 communities on uplands and floodplains in the middle Solimes region........................................32 3-1 Summary of statistics for low levee fields.............................................................54 3-2 The eight most common species remaining in sixteen harvested fields on low levees......................................................................................................................55 3-3 Summary of statistics for high levee fields............................................................56 3-4 The 19 most common useful species found in planted fields on high levees with corresponding number of harvested fields.....................................................59 3-5 Varieties of bitter manioc (mandioca) growing in 70 fields on high levees..........61 3-6 Varieties of sweet manioc (macaxeira) growing in 70 fields on high levees........61 3-7 Varieties of banana growing in 70 fields on high levees.......................................62 3-8 Perennials that were reported lost in fields and fallows........................................66 4-1 Summary of statistics for fallows on low levees...................................................74 4-2 The 25 useful species occurring in 17 capoeiras on low levees............................75 4-3 Local designations of work areas, number of areas visited, and their description..............................................................................................................77 4-4 Average size by type of capoeira...........................................................................77 4-5 Average number of useful species by type of capoeira.........................................78 4-6 The 22 useful species occurring in 17 banana groves (bananais)..........................82 4-7 Summary of statistics for banana groves (bananais)..............................................82 4-8 Banana varieties occurring in 10 banana groves (bananais)..................................82 4-9 Summary of statistics for fallows (capoeiras) on high levees................................84 vi

PAGE 7

4-10 Banana varieties occurring in fallows (capoeiras) on high levees.........................84 4-11 Twenty-five most frequently encountered species and their use in capoeiras on high levees with a comparison of former roas and former bananais...............85 4-12 Summary of statistics for banana fallows (capoeira bananal)...............................88 4-13 The 30 useful species found in 21 banana fallows (capoeira bananal)..................88 4-14 Banana varieties in banana fallows (capoeira bananal).........................................89 4-15 Summary of statistics for fruit fallows (capoeira fruteiras)...................................91 4-16 Banana varieties in six fruit fallows (capoeira fruteiras).......................................91 4-17 The 29 useful species most frequently encountered in 12 fruit fallows................92 4-18 The 25 useful species found in two fruit orchards (frutal).....................................94 4-19 The twenty-four most frequently encountered species and their use in five sitios.......................................................................................................................99 5-1 Average number of useful species in 163 vrzea home gardens by community...........................................................................................................108 5-2 Fruit, nut, and beverage species in vrzea home gardens....................................109 5-3 Number of banana varieties in vrzea home gardens..........................................113 5-4 Banana varieties in vrzea home gardens............................................................113 5-5 Home garden plants that were killed during the 1999 high flood on the vrzea...................................................................................................................114 5-6 Average size of 155 varze home gardens in 14 communities............................116 5-7 Average size of 30 terra firme home gardens in 3 communities.........................116 5-8 Average number of species in 30 terra firme home gardens...............................119 5-9 Number of banana varieties in terra firme home gardens....................................120 5-10 Banana varieties in terra firme home gardens......................................................120 6-1 Summary of local land-use designations by terra firme farmers.........................130 6-2 Summary of statistics for roas on terra firme.....................................................134 6-3 Summary of statistics for capoeiras on terra firme..............................................135 vii

PAGE 8

6-4 The 51 useful species encountered in 87 terra firme roas..................................140 6-5 The 31 varieties of bitter manioc (mandioca) found in 84 terra firme roas.......142 6-6 The 5 varieties of sweet manioc (macaxeira) found in 84 terra firme roas........142 6-7 The 11 banana varieties found in 46 terra firme roas.........................................143 6-8 The 16 banana varieties documented in non-roa areas on terra firme and the frequency in each land-use type...........................................................................145 6-9 The 68 useful species encountered in fallows on terra firme and the frequency they were recorded in each land-use type...........................................146 7-1 Summary of species diversity by land use type on the vrzea and terra firme of the Middle Solimes Region...........................................................................153 A-1 Useful species occurring in fields and fallows....................................................167 B-1 Useful species occurring in home gardens...........................................................175 viii

PAGE 9

LIST OF FIGURES Figure page 1-1 Map of the Brazilian Amazon centered on the middle Solimes region...................3 2-1 Landsat TM satellite image (bands 543) of the study region from 1999 with the locations of 19 communities included in the survey................................................33 2-2 Seasonal fluctuation in water level on the middle Solimes, 1992-2000................39 2-3 Floodplain community during low water.................................................................42 3-1 The bank of the Japur River during low water.......................................................44 3-2 Number of species in 82 planted fields (roas) on high levees................................58 3-3 Number of species in 31 harvested fields (roas) on high levees............................58 3-4 Number of manioc varieties in fields on high levees...............................................60 3-5 Number of banana varieties in fields on high levees...............................................60 3-6 Bundles of manioc stems placed upright in restinga alta field.................................70 3-7 Storage of manioc stems on a raised platform in a high levee field.........................70 4-1 Banana fallow on a high levee on the Solimes River.............................................89 5-1 Yard in the vrzea located on a paran connecting the Japur River to the Solimes.................................................................................................................103 5-2 The number of species in 163 home gardens on the vrzea...................................106 5-3 The number of species in 30 home gardens on terra firme....................................106 5-4 Scatter diagram showing number of species in 163 vrzea home gardens in relation to size in square meters .............................................................................116 5-5 Backyard with animals in a Solimes River community.......................................121 5-6 Maromba constructed to provide shelter for animals during the flood..................122 ix

PAGE 10

5-7 Cattle being fed the floating canarana grass during the flood season....................123 5-8 Terra firme home garden........................................................................................124 5-9 Well developed home gardens grade into diverse fruit orchards...........................125 6-1 Length of the previous fallow for 39 fields (roas) on terra firme.........................135 6-2 Age of 63 fallows (capoeiras)................................................................................135 6-3 Number of consecutive crops planted in 51 fields (roas) on terra firme..............135 6-4 Number of consecutive crops previously planted in 25 fallows (capoeiras) on terra firme...............................................................................................................136 6-5 Number of previous fallows for 66 fields (roas)..................................................136 6-6 Number of cropping cycles for 35 fallows (capoeiras) on terra firme...................136 6-7 Eight-month-old terra firme field (roa) cleared from old-growth forest..............137 6-8 Roa developing into fruit fallow (capoeira fruteira).............................................137 6-9 Number of useful species in 87 fields (roas) on terra firme.................................141 6-10 Number of manioc varieties in 84 fields (roas) on terra firme.............................141 6-11 Number of banana varieties in 46 fields (roas) on terra firme..............................141 6-12 Number of useful species in 43 fallows (capoeiras) on terra firme........................144 6-13 Number of banana varieties in 17 fallows (capoeiras) on terra firme....................144 7-1 Map of fields and home gardens of Betania...........................................................162 7-2 Preliminary classification of satellite image from 7-1...........................................163 7-3 Map of fields and home gardens of So Paulo do Corai......................................164 7-4 Participatory mapping of levees with fields and fallows in So Paulo do Corai................................................................................................................165 x

PAGE 11

Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science FARMING AND PATTERNS OF AGROBIODIVERSITY ON THE AMAZON FLOODPLAIN IN THE VICINITY OF MAMIRAU, AMAZONAS, BRAZIL By Morgan J. Schmidt December 2003 Chair: Nigel J. H. Smith Major Department: Geography An examination of agrobiodiversity on the flood plains and adjacent uplands along a stretch of the Amazon river (Middle Solimes) illustrates the enormous variability between different land use strategies. Farmers in the region use five principal habitat types for agriculture; beaches, mudflats, low levees, high levees, and uplands. High levees and uplands contain the greatest diversity in both land use types and useful plants. The research revealed 24 designations of land use by local residents when divided along lines of habitat, current status (field or fallow), and management. Home gardens are an important repository of agrobiodiversity because they contain almost the entire spectrum of useful species including many not found in other land use types and almost every home has one. xi

PAGE 12

CHAPTER 1 AGRICULTURAL BIODIVERSITY AND SHIFTING CULTIVATION: THE RESEARCH IN CONTEXT Introduction Agricultural biodiversity is paramount in meeting the needs of Earths growing human population. Wild relatives and traditional landraces of todays major crops hold the key to developing new, more robust, and better-adapted varieties that allow greater productivity in selected environments and better resistance to pests and diseases (Harlan 1975a, 1976; Vavilov 1987). Little-known crops are a major resource that can contribute new foods and other products with higher productive potential, greater efficiency, and increased nutritional value; or simply add welcome variation to our diet. A newly developed crop can strengthen local economies and create major industries. Maintaining diversity means keeping our options open to develop new crops and varieties or improve existing ones. This will help us meet the challenge of adapting agriculture to changing environments; and developing efficient, sustainable agricultural ecosystems. Hundreds of fruits, nuts, tubers, seeds, medicinals, ornamentals, timber, and other plants are virtually unknown outside of the regions where they are grown. These plants hold great untapped potential; and many are in danger of being lost to us forever with increased alteration of both natural environments and traditional agricultural ecosystems (National Research Council 1993). In recent years, researchers and policy makers have increasingly called attention to the importance of studying the agricultural systems of traditional farmers (Posey 1981, 1

PAGE 13

2 1985; Smith 1996a; Srivastava et al. 1999). It is thought that development plans can benefit from traditional farmers intimate knowledge of the environment and time-tested farming techniques. Traditional systems of shifting cultivation in the tropics almost invariably contain high agricultural biodiversity; and are well-adapted to local environments. The farming methods of the caboclo population (mixed descendants of indigenous peoples and immigrants in the Amazon region) in the middle Solimes region of Amazonia are no exception. The Solimes river is the section of the Amazon river in Brazil stretching from the Peruvian border to the confluence of the Negro river. The middle Solimes region is approximately centered on the confluence with the Japur river, a northern tributary coming from Colombia (Figure 1-1). Caboclos utilize a number of land-use systems and manage a myriad of useful plants in the diverse landscape along a stretch of the greatest river in the world. Our goal was to document the distribution and variability of agricultural biodiversity in the region as it relates to land management and habitat. Chapter 1 gives a background on agricultural biodiversity and shifting cultivation and their significance. Chapter 2 describes practical aspects of studying traditional agriculture in the Amazon region; the research area (including its people, environment, and conservation reserves); data and methods; and life on the vrzea (Amazon floodplain). Chapters 3 through 6 describe the results of the project. The chapters are organized by land-use and habitat. Chapter 3 describes the roa or manioc field and chapter 4, the capoeira or fallow, both on the vrzea. Chapter 5 focuses on the diversity of home gardens in both the vrzea and terra firme (uplands). Chapter 6 summarizes

PAGE 14

3 findings from the survey on terra firme agriculture. Chapter 7 summarizes what was learned and gives directions for future research. Figure 1-1. Map of the Brazilian Amazon centered on the middle Solimes region. The box delimits the study area. Our Stake in Agricultural Biodiversity Development of agriculture is arguably the most significant occurrence in human history. This major factor has allowed our population to rise at current rates and we now must increase food production to keep pace with population growth. Agricultural biodiversity (agrobiodiversity) is the biological resource that directly and indirectly contributes to crop and livestock production. It has been and will continue to be fundamental in our efforts to intensify agriculture (Srivastava et al. 1999). For thousands of years, humans have been domesticating new crops and expanding the range of existing ones by borrowing from neighboring groups or taking them along

PAGE 15

4 when they migrate to new regions (Harlan 1975b). Repeatedly, crops that began in localized areas have proliferated around the globe to become staple foods or the basis of industry. Hundreds of little-known crops and varieties exist that could someday be grown on a much larger scale. Tropical agricultural ecosystems (agroecosystems) are especially diverse in both food and nonfood species; with current and potential uses including medicines, construction materials, industrial materials, and ornamentals (Alcorn 1984; Denevan and Padoch 1987; Smith et al. 1995). Agroecosystems often contain plants in the process of domestication (proto-domesticates) from which exciting new crops may be recruited (Smith 1996b). Material for crop improvement is obtained from diverse landraces, wild crop relatives, and even unrelated species (Pimentel et al. 1997; Smith 1999; Smith et al. 1992). Plant breeders use these genetic resources to develop new varieties that are of greater quality, higher yielding, resistant to insects and disease, tolerable of stress, amenable to mechanical harvesting, and adapted to different environments (Harlan 1976; Plucknett et al. 1983). Traditional crops and varieties, proto-domesticates, and wild crop relatives, are a rapidly shrinking resource as modern agriculture advances, more farmers enter the market economy, cultural ecological knowledge is lost, and natural habitats are transformed (Altieri et al. 1987; Harlan 1975a; Nabhan 1985). The expansion of green revolution technology can have a negative impact on genetic diversity. In the push for greater productivity, a handful of crops have gained importance and come to dominate world agricultural production, often resulting in the loss of traditional crops and varieties (Altieri et al. 1987; Harlan 1975a). For industrialized farms to be viable, farmers seek the highest yielding varieties available.

PAGE 16

5 Commercial seed, developed to be compatible with agrochemicals, gives farmers limited choices as to which varieties will be planted (Clawson 1985). Mechanization requires that crops be of uniform size and shape; mechanization also makes intercropping of multiple species and varieties impractical (Chang 1977). The need for expensive technology and specialized knowledge has caused modern farms to focus on usually only one or two crops in order to compete successfully. Market forces dictate which crops will be grown as farmers cannot afford to spend time and energy on those that will not fetch the highest prices. The conversion of small, more diverse farms to large, mono-cropped farms and cattle ranches can eliminate traditional crops, varieties, and proto-domesticates (Smith 1999). In addition, the expansion of modern agriculture onto new lands puts populations of many wild crop relatives at risk of extinction, posing a threat to the crops that we rely on most (Damania 1994; Harlan 1976). All of these factors together have the effect of homogenizing agriculture around the world, by squeezing out crops and varieties that are less commercially important. These factors are critical, however, to the future development of agriculture and to human welfare. Globalization brings rapid change to many cultures, often resulting in the disappearance of cultural knowledge of ecosystems, plant uses, and traditional agriculture (Altieri et al. 1987; Nabhan 1985; Steinberg 1998). This knowledge, acquired through many generations of experience and experimentation, is as important as genetic diversity for the development of sustainable agroecosystems (Altieri 1995; Gliessman et al. 1981). The passing away of traditional knowledge and management practices often goes hand in hand with vanishing crop diversity.

PAGE 17

6 The loss of agrobiodiversity and traditional knowledge has grave implications for the safe and equitable future of the growing human population. As global food consumption increases, productive land is being lost by urbanization and degradation caused by unsustainable management. This necessitates the development of higher yielding varieties and more productive and sustainable agricultural systems. We must constantly contend with new pest and disease outbreaks that can devastate entire crops (Adams et al. 1971). Awareness is also increasing of the threat to agriculture posed by changing climates that may cause temperature and rainfall regimes to shift (Zandstra 1993). Our ability to expand food production and adjust to ecological and market changes stems largely upon the available genetic diversity. Call for Conservation During the 1960s, concern for the vulnerability of our major crops led to increasing efforts at collecting and preserving crop germplasm (Harlan 1975a; National Academy of Sciences 1972; Plucknett et al. 1983). A global network of gene banks was constructed with facilities for short, medium, and long term storage of seeds and limited maintenance of vegetatively propagated crops. The importance of preserving genetic diversity was emphasized by Harlan (1976), In the future, the need for genetic variability and sources of resistance shall drive us to a much fuller exploitation of all the genetic resources we can assemble. Although the practice of storing germplasm in ex situ gene banks is crucial for the future of agriculture, there are several recognized shortcomings in fully relying upon this method alone to conserve valuable genetic resources. First, there is insufficient representation of the full range of genetic diversity for many crops and a large number of less important crops are not represented at all (Harlan 1975a, 1976; Plucknett et al 1983).

PAGE 18

7 Second, storage facilities may be vulnerable to natural disasters and political change. Third, the plants are taken out of the ecosystem where they have continuously co-evolved with animals, other plants, and environmental stress including rapidly evolving pests and diseases. For these reasons a number of researchers have argued for the importance of conserving genetic resources in situ (Altieri and Merrick 1987; Brush 1989). This can be accomplished if a way is found to preserve the agroecosystems in which they are found (Altieri 1989). Tropical agroecosystems often contain a multitude of crops, varieties, proto-domesticates, and wild crop relatives including plants with a number of non-food uses. To conserve this resource we must understand where high genetic diversity is maintained and how management affects the abundance of crops and varieties (Alcorn 1984; Altieri and Merrick 1987; Smith 1996a). In tropical shifting cultivation (swidden) systems, it is pertinent to identify which processes of cultivation result in a secondary forest enriched with useful species, both planted and volunteer. Some useful plants may become more numerous when they are protected by the farmer during clearing, burning, and weeding or when they are favored by repeated cycles of cultivation (Bale 1994). Enriched fallows and mixed orchards may result from the practice of swidden-fallow agroforestry where fields are not only left fallow but are planted to a diverse array of useful species (Colfer 1997; Coomes and Burt 1997; Denevan and Padoch 1987). For traditional farmers, a diverse crop and variety repertoire has advantages in the field, on the farm, and in the region by providing security and greater efficiency (Clawson 1985). Maintaining a heterogeneous agricultural landscape of fields, fallows, natural areas, and extractive areas benefits families by providing game, fish, fruit,

PAGE 19

8 construction materials, medicines, and other products that contribute to subsistence and income (Altieri et al. 1987). When several crops and varieties are planted in a field or diverse landscape, they provide natural barriers that can act to slow the spread of pests and diseases (Adams et al. 1971) and can result in increased biological insect pest control (Altieri et al. 1987). However, one drawback is that the forest may act as a source of disease or pests for some native crops. Farmers are better able to cope with losses or plunging market prices if they are not dependent on a single crop. They can benefit by harvesting several crops that mature at different times, spreading their workload and income more effectively throughout the year and utilizing land more efficiently. A selection of crops and varieties also allows them to exploit small-scale environmental variation. A good example of this is the use of different crops by floodplain farmers according to micro-relief and soil texture (Denevan 1984). Studying the ways that traditional farmers interact with and manage their environments will enable us to merge the benefits of traditional agriculture with modern green revolution technology, leading us down a path toward a sustainable agriculture. The plant diversity of many traditional agroecosystems remains largely unexplored. There is a danger of losing the valuable genetic resources and sustainable management practices evolved over many generations as these agricultural landscapes are transformed. By documenting this diversity, plans may be devised to conserve it. It is certain that the more diversity we are able to maintain, the better equipped we will be to confront the challenges that await us. This research attempts to document the diversity of plants considered useful by the local population of the middle Solimes region. The focus is on the distribution of useful plants within the agricultural ecosystem of the Amazon vrzea

PAGE 20

9 (floodplain) and adjacent terra firme (uplands). By understanding how agricultural practices influence the diversity of useful plants in the landscape, resource managers and local farmers can design management plans that more effectively conserve genetic resources. Swidden Model For many, the terms slash and burn, swidden, or shifting cultivation carry negative connotations and conjure images of poverty stricken peasants laying waste to large tracts of virgin rainforest. This type of farming has often been blamed for the current crisis of deforestation around the globe, particularly in the species diverse rainforest areas. It has been described as primitive, haphazard, simple, wasteful, and inefficient by scholars (Conklin 1954, 1961; Gregor 1977; World Bank 1992) and even outlawed by some governments (Finley and Churchill 1913; UNESCO 1983). Finley exemplifies the dominant view of swidden agriculture during the first half of the twentieth century in his 1913 ethnographical and geographical sketch of the Subanu, an upland tribe on Mindanao in the Philippines, who practice this type of cultivation known locally as kaingin. He explains that agricultural development is seriously retarded by want of proper methods and the lack of efficient labor (Finley and Churchill 1913:15) Finley goes on to say, The kaingin method of farming involves a great waste of labor and materials and must be eventually interdicted by appropriate laws, rigidly enforced. Echoes of this pessimistic view of shifting cultivation can still be heard (Lavelle 1987). Today, most swidden scholars agree that the system is not inherently unsustainable (Kleinman et al. 1995). In fact, it is a very effective adaptation to the environment, especially where soils are infertile. Swidden is a relatively efficient form of agriculture

PAGE 21

10 in terms of food production per unit of labor (Beckerman 1987; Carneiro 1968). The problems begin when human populations, market influences, or circumscription of territory exert too much pressure on the available land, causing farmers to shorten fallow periods (Boserup 1965; Carneiro 1970; Food and Agriculture Organization 1994; Pratap 2000; Steinberg 1998; Vasey 1979). Shorter fallows may result in lower yields, and soil degradation. Although interest has grown enormously since Harold Conklin (1954, 1961) brought these issues to the forefront in the middle of the twentieth century, researchers are only beginning to understand the intricacies of slash and burn. This ancient form of agriculture is known by many names. In English, most commonly, the terms shifting, slash and burn, or swidden are combined with cultivation, agriculture, or horticulture to refer to the general model. In this thesis, I will use these names interchangeably. A widely accepted definition has been given by Harold Conklin (1961) as any continuing agricultural system in which impermanent clearings are cropped for shorter periods in years than they are fallowed. The process is fundamentally the same worldwide independent of culture, climate, or ecosystem. A patch of forest is cut, allowed to dry, and then burned. The elements that were locked in the living vegetation are then released as ash providing a pulse of nutrients that are easily utilized by crops. The area is planted for one or more seasons becoming a productive field. During the cropping period nutrients are depleted from the system, weeds invade, crop yields and efficiency drop, and the field begins a metamorphosis back to a state of higher biomass. This transformation may be left to the hand of nature by simply abandoning the field and allowing the forest to regenerate itself, or the fallow may be carefully managed by farmers, effectively increasing the

PAGE 22

11 productivity and efficiency of the system (Denevan and Padoch 1987; Eden and Andrade 1987; Hammond et al. 1995). Although swidden systems are usually associated with the burning of cleared vegetation, there are exceptions in areas of high rainfall and certain floodplain habitats where fire is not always used (Beckerman 1987; Orejuela 1992; Pinedo-Vazquez et al. 1996). The standard model of swidden farming hides infinite variation (Brookfield 2001; Denevan 1971). Differences include crop configuration, number of species and varieties, number of consecutive years planted, number of years left in fallow, whether or not it is burned, whether or not fallows are managed, amount of labor investment, and season of planting and harvest (Conklin 1954). The reasons behind this variation are also diverse. They include environmental, economic, demographic, and cultural factors that influence responses by farmers (Conklin 1954, 1961). Environmental conditions such as climate, vegetation, soils, altitude, pests and diseases, and flooding regimes shape the management systems at local and regional scales. Tropical ecosystems, in particular, can exhibit significant small-scale heterogeneity that requires numerous adaptations for successful cropping. Economic considerations including access to land, labor, and capital, the degree of local and regional market integration, and available transport are extremely relevant to the decisions a farmer makes such as the amount of land cleared and planted, choice of crops, time in fallow, and degree of fallow management. Cultural factors control, to a great extent, the methods used in this agricultural system. Though farmers are often innovative, they tend to farm the way they are taught by their relatives, neighbors, or extension workers. Farming techniques may be lost or learned as cultures come into

PAGE 23

12 contact with one another, market opportunities arise, and new technologies and crops are introduced. Finally, individual preferences play a part in all of these variables. The knowledge and practices of individuals may vary significantly within a single community. The enormous diversity encountered in swidden systems around the world means that generalizations made in the past are becoming obsolete. Why Study Swidden? Research on swidden agriculture is valuable from both a theoretical and practical standpoint. From the theoretical angle, its study holds important clues for our understanding of agricultures distant origins and development (Sauer 1952). Although the traditional forms of shifting cultivation we encounter today may not be a window into the past, examination of its myriad forms and techniques can provide insights and help form hypotheses to be tested with the analysis of empirical data. As human populations and market integration increase, swidden systems are transformed in numerous ways (Behrens et al. 1994; Henrich 1997; Serro and Homma 1993). Information on these changes is valuable in providing us with a better understanding of the mechanisms and impacts of agricultural intensification (Boserup 1965; Turner et al. 1977; Vasey 1979). From a practical standpoint, research on the sustainability of swidden systems is essential in our efforts at rainforest conservation in the twenty-first century (Serro 1993). Destruction of tropical rainforests around the world is one of the most pressing environmental issues that we face today (Whitmore 1990). Impact on biodiversity and the global climate are two effects of deforestation that have relevance for everyone, regardless of where we live (National Research Council 1993). There are also various local and regional problems that may result from deforestation including, soil erosion and

PAGE 24

13 degradation, loss of habitat for fish and wildlife (Goulding 1999), rainfall reduction, and increased sedimentation of rivers and other water bodies. Shifting cultivation has been cited as one of the main causes of tropical deforestation in many regions (Food and Agriculture Organization 1994). In a 1982 FAO study, this form of agriculture was identified as the most important reason for the loss of tropical forests (Lanley 1982). It was estimated to be responsible for 35 percent of deforestation in the American tropics, 49 percent in Asia, and 70 percent in Africa. Because of the species diversity of the rainforest, its threatened situation, and the fact that it is the environment where shifting cultivation dominates today, the question persists whether this form of agriculture is compatible with conservation. Can it be practiced in such a way as to benefit the natural environment by preserving biodiversity or even increasing it at the landscape scale? There is some indication that this might actually be the case. William Bale (1989, 1994), in his work among the Kaapor tribe that practice swidden gardening in the Eastern Amazon, found that old fallows contain significantly different species than old-growth forest suggesting an enhancement of biodiversity at regional scales. How can a system like the Kaapors be adapted to areas with higher population density and greater pressure on the land? Only a better understanding of contemporary swidden systems can begin to shed light on this question. Slash and burn is a highly adaptive and sustainable agricultural system when cropping intensity is not so great as to preclude adequate fallow periods (Kleinman et al. 1995). However, when population increase or territorial circumscription necessitates raising the cropping intensity and shortening the fallow period, the ecologically sound

PAGE 25

14 system can break down, becoming unsustainable (World Bank 1992). By evaluating the many forms of shifting cultivation in different environments, we hope to find a way to use our rainforests rationally and intelligently, thus preserving them for future generations. Swidden Agriculture in Time and Space We do not know when the practice of swidden cultivation first came into use. We can be fairly confident, however, to suggest that it is among the oldest forms of agriculture in existence and it is, in that sense, a truly primitive form. Though conventional wisdom holds that agriculture began in the Fertile Crescent about 10,000 years ago, it is quite likely that it developed much earlier at various times and places rather than in one single origin event (Harlan 1975; Vavilov 1987). We must not rule out the tropical forest or tropical floodplain as possible settings for an early beginning of agriculture (Sauer 1952; Smith 1999). Several conditions combine to make the tropical moist environment conducive to the development of plant domestication. The first is the climate, being very amenable to plant growth with abundant sunlight and rainfall. Second, there is a rich abundance of species with a large number of potential uses. A third characteristic that makes planting easier are the soft, friable soils of the forest floor and the annually renewed, nutrient rich floodplain soils. No matter what environment it was first initiated in, humans eventually learned that they could kill large trees by girdling to allow sunlight to reach the forest floor and burn vegetation to provide nutrients needed for plant growth. Besides being one of the oldest forms of agriculture, swidden is one of the most widespread forms even today. It has been used on every continent except Antarctica and thousands of islands around the world. Although it is primitive in the spirit of its great

PAGE 26

15 age and continuity, swidden agriculture has evolved to a high level of diversity and sophistication in the multitude of cultures and environments where it is utilized. Innovative farmers from around the world have contributed to the evolutionary development of swidden giving it an abundance of forms. It is not a static and rigidly defined set of methods but rather, variations are determined by individual preferences and cultural history in addition to environmental factors. Today we think of swidden as primarily a tropical agricultural system. In the past, however, it was practiced throughout the temperate zone as far north as present day Canada and Scandinavia (Clark 1952). During the Neolithic, shifting cultivation was used extensively in Europe and apparently resulted in immense ecological change by deforesting large areas. Comparing the average fire interval over time, Lehtonen and Huttunen (1997) show that the slash-and-burn technique was commonly used for cereal production for about 2000 years in southern Finland, being the most important cultivation technique in the eastern part of the country until the 1900s. They point out that fires have a great influence on forest structure and swidden cultivation has greatly influenced the history of forest fires. In another study in neighboring southern Sweden, Lageras (1996) examines the pollen record over the past 7000 years to document vegetation and land-use history. He determines that the first indication of human impact by forest clearance occurs around 6000 years ago. Pasture has been the principal agricultural land-use in the area beginning about 5000 years ago and slash-and-burn cultivation was probably introduced to the area around 1200 AD. In the last century the land has once again become covered by forest

PAGE 27

16 making it easy for us to forget that the landscape was once dominated by human activities, namely forest clearance and burning. This supports the idea that many areas, including parts of Amazonia, often considered pristine wilderness are, in fact, the product of human manipulation (Denevan 1992; Erickson 2000, 2001; Heckenberger 1996, 1998; Heckenberger et al. 1999, 2003; Raffles 1999; Smith 2002). Swidden farming was a common practice by prehistoric Amerindians in many places throughout North America (Lopinot and Woods 1993; Woods 1987). Woods (1987) examined the ethnographic record to characterize prehistoric settlement patterns and concluded that many were determined largely by the reliance on maize agriculture. For native North Americans from the east coast to the Mississippi valley, corn was one of the most important crops and the ash from burned vegetation was an essential addition of nutrients to the soil. Because of corns high demand on nutrients, all but alluvial soils, replenished annually by floods, had to eventually be put into fallow. Many settlements were located on high ground near rivers where fields could be planted on the floodplain as well as uplands. Many fallow fields were present and an infield-outfield system was often practiced with smaller gardens near homes and larger fields further away. In many cases, villages would relocate periodically at 10 to 30 year intervals due to exhaustion of agricultural land and lack of fuelwood. European colonists also used shifting cultivation up through the 1700s in the United States (Matlock 1997) and Southern Ontario (Clark 1952) resulting in the clearance of much of the eastern forests of North America. The tropical moist forest is not an environment conducive to the preservation of ancient crop remains or human artifacts that were often made from wood and other perishable material (Harris 1972). Because of this fact, along with the paucity of

PAGE 28

17 archaeological research, relatively little is known about the origins of agriculture in the tropics as compared with Southwest Asia. However, evidence is mounting for the antiquity of agriculture in the tropics. Recent research is firmly establishing Mesoamerica as one of the cultural regions that was an early center of crop domestication. In the lowlands of coastal Mexico, Pope et al. (2001) found evidence for extensive forest clearing and maize cultivation by 6,800 years ago. It is believed that the Maya civilization was based on some form of shifting cultivation that allowed a relatively dense population in a humid tropical region (Harris 1972). Remains of domesticated manioc have been found in archaeological sites on the coast of Peru dating from 2,800 years ago (Ugent et al. 1986). This, together with recent phylogeography research that places the origin of cultivated manioc in the southern border region of the Amazon Basin, suggests that the crop was grown throughout the American tropics by 3000 years ago (Olsen and Schaal 1999). Swidden agriculture clearly has an immense history in tropical forest regions around the world.

PAGE 29

CHAPTER 2 FOCUSING ON THE MIDDLE SOLIMES REGION Prehistoric Agriculture and Settlement on the Amazon River We know that human populations have been settled along the Amazon for many thousands of years. Paleoindian remains were found by Anna Roosevelt and colleagues at a site called Caverna de Pedra Pintada near Monte Alegre on the left bank of the Amazon downstream from Santarem (Roosevelt et al. 1996). It is a well dated cave site with cultural material from about 8,500 to 11,200 years ago and rock art thought to be from the same period because of paint specs found in the stratigraphy. The site is especially significant because of food remains of tropical forest species that were found including many edible tree fruits and a variety of fauna showing that the people subsisted on foods from the rainforest and river and not on big game like other paleoindian groups. Roosevelt suggests that the many species adapted to disturbance indicates there may have been some forest clearance by these early inhabitants, however, no remains of crop species have yet been found. The river bluffs adjacent to the Amazon floodplain have yielded archaeological evidence for the earliest pottery yet known in the Americas and confirms specialized exploitation of river fauna by these early populations (Roosevelt et al. 1991, 1996). Caverna de Pedra Pintada and an Archaic shell-midden called Taperinha on the opposite side of the river from Monte Alegre excavated by Anna Roosevelt in 1987, both contain ceramics with dates over 7000 years ago. 18

PAGE 30

19 Evidence for manioc agriculture is present in the archaeological record at Caverna de Pedra Pintada in the form of bowls and thick griddles like those used today for cooking manioc. The pottery relates to the lowland South American Formative period from about 4000 to 2000 years ago, thought to represent the diffusion of ceramics and root horticulture (Roosevelt 1980; Roosevelt et al. 1996). Evidence from various locations in northern South America indicates that manioc was a staple food for a number of groups by 3000 years ago (Lathrap 1970). Ethnohistorical documents describe large populations densely settled along the main branch of the Amazon River when Europeans first arrived (Carvajal 1934; Porro 1994). Preliminary archaeological evidence lends credence to these reports (Heckenberger et al. 1999; Smith 1980). Most of the floodplain cultures were completely destroyed in the first 150 years of European contact. It is likely that these populations subsisted on a combination of aquatic resources and agriculture, both on the vrzea and terra firme (Roosevelt 1980; Smith 1999). There is much uncertainty and debate as to the relative proportions of these resources in the subsistence economy of prehistoric riverine cultures. Roosevelt (1980) theorized that the introduction of maize in the Orinoco and Amazon basins acted as a springboard for the development of dense populations. She reasons that manioc is ill-suited to the floodplain and makes inefficient use of the relatively fertile alluvial soils. Roosevelt claims that short maturing maize is a more appropriate crop for the unpredictable bottomlands and the high-protein grain makes optimal use of the nutrient-rich soils. The examination of contemporary agricultural systems, resource use, and ecological perceptions on the floodplains and adjacent uplands

PAGE 31

20 can help us to evaluate existing theories and provide insights into the subsistence economies of prehistoric riverine inhabitants. Amerindian populations today are a vestige of what they once were (Denevan 1976a). In Amazonia, like elsewhere in the western hemisphere, the population has fallen drastically since first contact with Europeans in the sixteenth century due to disease, warfare, and slavery (Denevan 1976b; Heckenberger 1998; Meggers 1992). In pre-contact times, intensive agriculture, with significant transformation of the landscape, was well underway in several regions of the Amazon (Denevan 1970; Ericson 2000, 2001; Heckenberger 1996, 1998, 1999, Heckenberger et al. 2003). It is clear that Amerindian subsistence economies, including swidden cultivation, described in the ethnographic record are a remnant of their former range of variation (Beckerman 1987). Research on the swidden systems in use along the Amazon river today can add to our overall understanding of this subsistence activity in the region. Practical Aspects of Swidden Research in Amazonia National governments view the development of their Amazonian territories as imperative to the growth of their economies and modernization of their countries. Conservationists hope they will seek sustainable ways to develop natural resources. Conventional agricultural methods including mechanical clearing and tillage of the soil, large plantations, and mono-crops have so far not proven successful in Amazon rainforest areas (Fearnside 1987, 1988). These methods often cause soil erosion, degrade ecosystems, and deplete biodiversity (Gascon and Lovejoy 1998; Milliken 1992; Serro et al. 1993). The search continues for forms of development that maintain the integrity of ecosystems and, at the same time, contribute to markets and raise the standard of living for local people.

PAGE 32

21 Many researchers have emphasized that indigenous peoples have an important contribution to make to the question of how to develop the Amazon (Anderson and Posey 1989; Posey 1981,1985,1993; Schmink et al. 1992). Smith (1999) argues that native knowledge and modern science should be combined when trying to make rational use of the Amazons resources. The National Research Council has urged that information on traditional agricultural systems of the humid tropics be researched and evaluated before this valuable knowledge is lost (National Research Council 1982). Since the 1960s, the Brazilian government has implemented an aggressive development strategy for its Amazon territory. A perceived threat from other nations who covet the Amazon could be ameliorated by a clear Brazilian presence. Furthermore, the migration of large numbers of people to the vast, unoccupied frontier region was seen as a safety valve for social problems in the rest of the country. The opening up of Rondonia to colonist settlement brought thousands of landless people from Southern Brazil. In 1970, the Transamazon highway colonization scheme was begun with hopes that large numbers of migrants would come from the drought stricken northeast to settle in the Amazon and build a strong economy based upon agriculture, mining, and timber extraction (Moran 1984; Smith 1981). Rather than settling along the main rivers, the colonists inhabited upland areas along the new highways. They brought their notion of agriculture from the sub-tropical and temperate south and the dry northeast to a foreign environment. The colonization project made little effort at ecological zoning according to soil type or other indicators of fertility. The colonists were handed a piece of land that often lay on extremely poor soils. Many of the colonists failed and returned home while others found a livelihood in the growing urban centers. Some learned how to

PAGE 33

22 adapt their agricultural methods to the rainforest environment and were able to make a living. The development of diverse agroforestry systems is seen by many as a viable option for Amazonian farmers (Smith 1996b). The maintenance of biodiversity and better preservation of ecosystem services are two potential advantages of agroforestry and some argue that the cutting of old growth forests would be slowed by having a more permanent system in place (Denevan 1980; Smith et al. 1995). Swidden agriculture frequently includes agroforestry as an integral part of forest management (Alcorn 1990; Coomes and Burt 1997; de Jong 1996; Denevan and Padoch 1987; Eden and Andrade 1987; Hammond et al. 1995; Unruh 1990). Indigenous peoples throughout the tropics often practice some form of fallow management that increases the productivity of a given field. In indigenous swiddens, cultivation is rarely stopped abruptly but instead, useful plants are harvested for many years after the field is abandoned. Denevan and Padoch (1987) and their colleagues studied what they aptly termed swidden-fallow agroforestry in a native Bora community of the Peruvian Amazon. In this particular case and many others, swidden fallow management is based on successive planting and harvesting of different crops. For example, a staple crop such as manioc may be interplanted with longer maturing crops like bananas and some fruit trees. As the crops are harvested, others may be planted in their place. Eventually, the field can become a secondary forest enriched with useful species or a fruit orchard. The advantage is that a farmer can return to his fallows to continue harvesting fruits, timber, medicines, and other products for years to come. Another benefit of having a diversity of crops growing in managed fallows is that farmers can spread production throughout the year.

PAGE 34

23 This research will show that traditional caboclo farmers of the middle Solimes also practice various degrees of swidden-fallow agroforestry and forest enrichment. Traditional systems could be adapted by colonists to develop a more viable agricultural system. Indeed, innovative colonists are already experimenting with various consortiums of crops on their own (Smith et al. 1995). The land use systems of traditional inhabitants can certainly provide some useful lessons for colonists to improve the way they utilize and conserve resources in Amazonia. This research examines the extent that traditional farmers on the middle Solimes utilize fallow management by documenting the distribution of useful plants in the agricultural landscape and the practices that result in that distribution. In Brazil, while most of the nations attention was focused on the terra firme as the main area of development in Amazonia, the vrzea with its relatively small extent of about 2% of the basin, was largely ignored. The vrzea is the floodplain of white water rivers that carry fertile sediment from the Andes Mountains. The realization that terra firme soils are generally poor in nutrients and do not support intensive cultivation of cereal crops without substantial external inputs has sparked a renewed interest in the vrzea as a promising area for modern agricultural development (Junk 1982; Smith 1981). Because of the influx of nutrients brought by the flood, good prospects for irrigation, and ease of transport, some have advocated the development of arable farming or irrigated rice on the vrzea. However, developers and policy makers generally do not know how to implement large-scale mechanized agriculture on land where massive, unpredictable floods are the norm.

PAGE 35

24 Several attempts have been made to grow irrigated rice on the vrzea with varying levels of success. A project at the mouth of the Jari River attempted to grow irrigated rice on a large scale in the 1970s. Their efforts were hampered by problems with soil fertility, pests, fungal diseases, and weeds (Fearnside and Rankin 1985). Junk et al. (2000) discourage large-scale arable farming on the vrzea for environmental, economic, and social reasons. Instead, they support the integral management of aquatic and terrestrial resources through the zoning of economic activities and decentralized and participatory management by the local population. There are four major economic activities on the vrzea that vie for control over limited resources; fishing, forestry, agriculture, and ranching (Junk et al. 2000; Smith 1999). The conflicts of interest are exemplified by the possible adverse relation between fishing and the other three land uses that result in deforestation of the floodplain. Forest clearance for timber, agriculture, and pasture poses a threat to the productivity of aquatic resources by reducing their habitat and food sources (Fearnside 1990, 1995; Goulding et al. 1996; Goulding 1999). Deforestation is also a threat to the many endemic plants and animals of the vrzea that are so far not adequately protected. In addition, the growth of ranches and large-scale agriculture can result in the loss of crops, varieties, proto-domesticates, and crop relatives that are vital for the future development of the vrzea (Smith 1999). With the increasing interest in intensifying agriculture on the vrzea, it is pertinent to examine the resource use and perceptions of the local inhabitants. There are very few Amerindian groups that utilize vrzea resources today. Having endured the first and longest sustained contact with their European conquerors, the Amerindians who lived along the Amazon River were virtually wiped out by disease,

PAGE 36

25 warfare, and slavery. Three Amerindian groups that currently practice agriculture on the floodplain are the Shipibo and Cocama who live in the Peruvian Amazon and the Tikuna (they have moved to the riverside from the uplands since the time of European contact) who live in the frontier region of Colombia and Brazil (Bergman 1980; Shorr 1999). However, they are relatively acculturated groups and are a small minority along the Amazon River. The caboclos of Brazil or ribereos of Peru comprise most of the inhabitants along the major rivers today. They are people of mixed Amerindian, European, and African ancestry (mestizo) who rely largely on subsistence techniques similar to those of indigenous Amazonians (Moran 1993; Sutlive et al. 1985; Wagley 1953). Indeed, there is a continuity of Amerindian culture and resource management in the caboclo way of life. Relatively little attention has been paid to the living strategies of the mestizo riverine population (Padoch and de Jong 1991; Parker 1989). Hiraoka (1992) provides a brief review of the research on caboclo and ribereo resource management. Here, I will highlight some of the literature pertaining to caboclo/ribereo agriculture that is particularly relevant to this thesis. Denevan (1984) elaborated the concept of horizontal ecological zonation of agriculture on the floodplain, likening it to altitudinal zonation on mountain slopes. His article was inspired by Bergmans (1980) study of the Shipibo in Peru and work done by Judith Gunn who studied agricultural scheduling on the Ilha dos Purus in the Amazon above Manaus. Hiraoka (1985a, 1985b) described the ecological zonation of floodplain agriculture of a ribereo community near Iquitos in the Peruvian Amazon. Several studies have focused mainly on mestizo terra firme agriculture along the river near Iquitos (Chibnik 1994; de Jong 1996; Hiraoka 1986;

PAGE 37

26 Padoch and de Jong 1992; Padoch et al. 1985). Frechione and colleagues (1989) examined caboclos perceptions of biotopes and ecological zonation at Lake Coari on the Solimes. Anderson et al. (1995) studied resource management of caboclos in the Amazon estuary near the mouth of the Tocantins River. Swales (1999) compared land use dynamics and agriculture in the uplands and floodplain in the lower Amazon. Preliminary results of research on floodplain farming in the Brazilian state of Amap are given by Padoch and Pinedo-Vasquez (1999). Some ongoing work is being done on agricultural systems and change on the vrzea in the Mamirau Sustainable Development Reserve where the data for this thesis was collected (Padoch et al. 1996; Pinedo-Vasquez et al. 1996, 1999). Finally, a comprehensive work edited by Junk et al. on actual use and options for sustainable management of the vrzea includes information on current management practices by caboclos (Junk et al. 2000). One way that caboclo agriculture can contribute to questions of vrzea development are the many crops and varieties that they plant. Generations of farmers have selected varieties and individuals that are tolerant to flooding and resistant to pests and diseases. Development plans would be prudent to incorporate crops that are adapted to conditions on the vrzea. Caboclo agricultural practices can also inform vrzea development planners. We know that Amazon floodplain farmers utilize a number of habitats. What management strategies do they use to cope with the varying environmental conditions of those habitats? What crops do they plant in which habitats? The imperative of conserving vrzea biodiversity and natural resources compels us to analyze the agroecosystems of caboclos in detail to determine which aspects of their management can contribute to the sustainable development of the Amazon floodplain.

PAGE 38

27 Amazon Caboclo The people who make their living along the middle Amazon today are known as caboclos, a designation whose significance has evolved and been modified during the history of colonial occupation (Lima-Ayres 1992). The term caboclo conveys different meanings to different people. An important distinction is made between the colloquial use of the word and its application in the academic literature. In the colloquial use, caboclo can be a racial category, economic class, or rural resident. In elementary school, Brazilian children learn that caboclo means a mixture of Amerindian and European ancestry. Caboclo also indicates a low social class, the rural or urban poor, and illiteracy. The term is generally derogatory and few people will refer to themselves as caboclo. It would be used by the urban upper class to refer to the urban poor, and by the urban poor for people from the interior or rural areas. Rural inhabitants, in turn, would not call themselves caboclos but use it to refer to Amerindians. In other words, it is always the other and rarely oneself (Lima-Ayres 1992). In the academic literature, caboclo is not a racial category but rather, a social category, the indigenous Amazonian peasantry (Lima-Ayres 1992; Parker 1985; Ross 1978). A distinction is made between the caboclo and the other two major groups in the Amazon, Amerindians and recent migrants or colonists. Caboclos are made up of a mixture of people from three main regions; Amazonia, Europe, and Africa. Many caboclos are descendents of the indigenous inhabitants that lived along the Amazon when Francisco de Orellana first sailed down the river from Ecuador to the Atlantic in 1542. Amerindians were captured as slaves or enticed into missions and incorporated into the European colonial system. As time went on, Amerindians continued to migrate from tributary and headwater regions to fill the demographic void created by the depopulation

PAGE 39

28 of the vrzea region. Miscegenation resulted in a population dominated by people of European (mainly Portuguese) and native Amazonian descent. Included in the mix are Africans who were originally brought to Brazil as slaves to work on plantations and also migrants who came to the Amazon from Northeast Brazil and other regions during the rubber boom from 1850 to 1920. Wagley (1953) and Galvo (1952) made the first anthropological studies of the caboclo and characterized caboclo culture as a combination of Amerindian and Iberian cultural traits. Many aboriginal economic strategies have persisted in the caboclo population long after the tribal peoples were gone. Amerindian methods of agriculture, hunting, and fishing, as well as customs and religious beliefs are still practiced in order to make a living in the volatile, often unpredictable environment of the middle Amazon (Moran 1974; Parker 1985). Middle Solimes Region The Solimes River is the section of the Amazon River in Brazil stretching from the Peruvian border to the confluence of the Negro River near the city of Manaus. The middle Solimes region is approximately centered where the Japur River joins the Solimes (Figure 1-1) (Lima-Ayres 1992). Tef is the largest town and the commercial center of the region. According to the 1996 census, the municipality of Tef had a population of 62,000 with 76% living in urban areas (IBGE 1998). The surrounding municipalities have a much lower urban population. In the nearby municipality of Uarini, with a 1996 population of 10,500, 20% are urban. The regional population is growing rapidly with a high rate of natural increase and rural to urban migration. Two degrees south of the equator, the climate is tropical with little year long variation in average temperatures. Average monthly temperatures range from a daily

PAGE 40

29 minimum of between 21 and 23 degrees Celsius to a daily maximum of between 30 and 33 degrees Celsius (Ayres 1995). There is considerable variation in precipitation throughout the year and from one year to the next. December through March is the rainy season with about three times the amount of precipitation as during the dry season from July to October (Ayres 1995). Average yearly rainfall from 1977 to 1981 for the city of Tef was 2,373 millimeters with the lowest and highest rainfall for those years being 2,190 and 2,632 millimeters. Fishing, agriculture, and timber extraction are the predominant economic activities in the region. According to the Brazilian Institute of Geography and Statistics, manioc is the most important crop (IBGE 1998). In the municipality of Tef (23,800 km 2 ), 2,238 out of 2,255 (99%) farms surveyed produce some manioc. Approximately 64,000 tons were harvested on 6000 hectares. However, only 33 tons were reportedly sold indicating that most manioc is grown for subsistence only. Bananas are the second most important crop in the region and, unlike manioc, most are sold on the market. Other crops that were included in the survey that are produced to a much lesser extent include corn, beans, sugarcane, oranges, tomatoes, and rice. Cattle ranching has not yet taken off in the region. A total of 5000 head of cattle were reported in the four municipalities of Tef, Alvares, Mara, and Uarin with the majority in Alvares and Tef near the larger urban centers. Some hogs and poultry are also produced. Many rural families keep a few ducks or chickens around the home. Only a few of the households surveyed for this thesis owned cattle or hogs. Virtually all settlements in the middle Solimes region are located near a river. Most rural inhabitants live in scattered settlements of fifteen houses on average and a few

PAGE 41

30 live in isolated homes. Many of the small settlements in the middle Solimes region began with the decline of the rubber era when rubber tappers moved in from more isolated extraction areas (Sociedade Civil Mamirau 1996). The communities in the region are located either on the river bluffs surrounding the vrzea, on the vrzea itself, or on large lakes formed at the mouths of tributaries of the Solimes such as Lake Tef, Lake Uarini, and Lake Aman. Communities differ in their monetary wealth, degree of isolation, market connections, government support, and available habitats for agriculture and resource extraction. On the terra firme, the agricultural cycle follows the change in seasons from wet to dry. The cycle of activities on the vrzea follow the pulse of the flood. Connservation of the Vrzea Ecosystem: The Mamirau and Aman Sustainable Development Reserves The enormous productive potential and inumerable endemic species of the vrzea have spurred researchers to underscore the importance of rationally managing floodplain resources (Goulding et al. 1996; Junk et al. 2000; Padoch et al. 1999; Smith 1999). The newly created Mamirau Sustainable Development Reserve (MSDR) is the only conservation area specifically dedicated to the preservation of biodiversity and management of resources on the vrzea in Brazil (Sociedade Civil Mamirau 1996). Covering over one million hectares above the confluence of the Solimes and Japur Rivers, the area was first established as an Ecological Station in 1990. The recognition that people live in and utilize the resources of Mamirau prompted a new category of conservation unit to be created and the area was reclassified as a sustainable development reserve in 1996. In 1999, the Aman SDR was created adjacent to Mamirau to link up with Ja National Park creating a central Amazon conservation corridor.

PAGE 42

31 The goal of the Mamirau project is to integrate conservation of biodiversity with the social development of local communities (Lima 1999; Howard et al. 1995; Sociedade Civil Mamirau 1996). Scientists work with residents of the reserve to research, monitor, and protect natural resources and to develop a management plan that encompasses both conservation and development objectives. They must determine which activities threaten natural resources and seek new activities that are compatible with conservation goals. Our study builds on others from the Mamirau Reserve (Chota 1999; Padoch et al. 1996; Pinedo-Vasquez et al. 1996; Pinedo-Vasquez et al. 1999) to show the enormous resource of agrobiodiversity and management diversity used by small farmers within the conservation reserves. It is hoped that it will be valuable as a jumping off point for further research on agrobiodiversity in and outside the reserves and useful when considering new policies or projects that may affect agrobiodiversity in the region. Data and Methods To understand the spatial distribution of agrobiodiversity across different habitat types and management systems, the home gardens, fields, and fallows were visited in nineteen communities in the middle Solimes region (Table 2-1 and Figure 2-1). An inventory was conducted in each home garden that included all food, timber, and medicinal plants. The number of individuals of each species was recorded, the size of the home garden was estimated, and a position was taken with a GPS receiver. Fields and fallows were visited and the following information was gathered. GPS position Type of work area (field, fallow, orchard, etc.) Estimated size Habitat (terra firme, high levee, low levee, mudflat, or beach) Useful species (including food, timber, medicinal, and other) Varieties of manioc and banana

PAGE 43

32 Name of owner Other information gathered for some fields and fallows including but not limited to: crop age, field age, number of consecutive seasons planted, number of previous fallows, age of last fallow, and status or use of the area before (fallow, old growth forest, etc.) An attempt was made to visit all fields and fallows in each community. However, this was not always possible because of the great number of fallows in some communities, the logistics of reaching some of them, and the question of recruiting informants who were willing to take time out of their busy schedules to help us. The objective of the study was to visit a large number of fields in communities located throughout the region, therefore, quick estimates of size were made rather than taking the time to carefully measure each field. Size was estimated by site and Table 2-1. Number of home gardens, fields, and fallows visited in 19 communities on uplands and floodplains in the middle Solimes region Community Map id Vrzea/ terra firme Num. home gardens visited Num. fields visited Num. fallows visited Acari 9 vrzea 0 1 0 Barroso 8 vrzea 11 16 9 Bate Papo 10 vrzea 0 2 3 Bela Vista do Manguary 11 vrzea 2 5 9 Betania 15 vrzea 16 35 16 Betel 16 vrzea 10 26 13 Boa Esperana 19 terra firme 17 55 45 Jarau 18 vrzea 11 31 12 Jubar 13 terra firme 5 15 8 Manacabi 17 vrzea 7 14 12 Marirana 7 vrzea 10 13 5 Nossa Senhora da Ftima 2 terra firme 8 16 22 Pentecostal 12 vrzea 9 8 6 Porto Praia 4 vrzea 15 7 12 So Francisco do Aiuc 5 vrzea 15 48 25 So Joo 6 vrzea 16 10 6 So Paulo do Corai 14 vrzea 12 22 48 Sitio Fortaleza 3 vrzea 11 15 34 Vila Alencar 1 vrzea 19 19 36 In these two communities, fields were visited on both vrzea and terra firme

PAGE 44

33 Figure 2-1. Landsat TM satellite image (bands 543) of the study region from 1999 with the locations of 19 communities included in the survey

PAGE 45

34 occasionally checked by pacing. Estimates of the size of smaller fields should be accurate to within a couple of meters while the error may increase proportionately as the size of the field increases. In cases where it was impractical to measure the size by pacing and it could not be estimated by site because of thick canopy, the owner or informant gave his estimate of the size, leaving room for error. Therefore the size, especially for the larger fallows, should be considered rough estimates. The distinction between habitats of terra firme and vrzea is more clear than that between high and low levees and low levees and mudflats. It would have been possible to draw an arbitrary line between the habitat types based on the height of the land in relation to the high water level of the previous years flood but this would be possible only where there are trees with visible high water marks. In the interest of time this was not done. Instead, the decision was made to use the informants classification of the habitat. Local farmers classify habitats based on their observations of the depth of flooding and length of time underwater each season as well as the vegetation and soil types. They rely on their own experience from their particular local environment and so these classifications can be somewhat subjective. Farmers classify four major floodplain habitats used for agriculture; restinga alta (high levee), restinga baixa (low levee), lama (mudflat or silt bar), and praia (beach or sand bar). At the time of fieldwork, high levees were usually above water but some were flooded to a depth of usually not more than forty centimeters. The vegetation is well developed forest. Areas that farmers classified as low levees were flooded from one to several meters and vegetation is usually dominated by the imbaba tree (Cecropia spp.). Mudflats are typically dominated by canarana grass (Echinochloa spp.) and beaches are devoid of vegetation.

PAGE 46

35 The inventory of the fields and fallows recorded all species that were considered useful by the inhabitants. These included tubers, fruits, vegetables, nuts, timber, and medicinal plants. A few others with uses such as fishbait, latex, and thatch were also included. The inventories were made by walking around in the field or passing slowly along the riverbank in a boat. The informants were relied upon to name the useful species and much effort was made on our part to insist that all of them be recorded. The error here is most likely to consist of some species not being recorded for some fallows, especially the very large ones. Varieties of banana and manioc were recorded in many cases but not all. Sometimes the variety was unknown to the informant. Many of the fields that were visited had already been harvested and were often flooded. In this case the crops that had been harvested were reported. The information on management was gathered, when possible, in informal interviews with the informant while visiting the field. Occasionally, details on yield, weeding, markets, or management were learned in this way. Finally, an informal interview was conducted in the home after taking the home garden inventory. The residents were asked about the number and size of their fields and fallows, the crops they had harvested, and the produce they had sold to market. They were also asked about losses of crops and home garden plants due to the flood or pests and diseases. The information gathered in this way is supplementary only. The interviews were not structured and the same information was not given by every interviewee. The data was put into a database and incorporated into a GIS using the geographical positions recorded in the field. To do this, a Landsat Thematic Mapper

PAGE 47

36 image was geo-referenced and the points were plotted. This has been given to the Mamiraua Sustainable Development Institute to be incorporated into their larger GIS database. Areas that were designated as high levee or restinga alta were flooded to a maximum depth of one meter in the year 2000. Most of them were flooded less than 40 centimeters and the majority still had dry land, with many elevated up to 50 centimeters above the high water level for the year. In this survey the classification of habitat type was based on the informants designation of the area. Therefore, it is assumed that some overlap exists between habitat type designations. In other words, areas that were called restinga alta in one community with one meter of water may be termed restinga baixa in another community. For example, in So Paulo do Corai located on a side channel (paran) of the Japur, some fields that had not flooded and were up to 50 cm. above the flood in 2000 were termed restinga baixa while in Betel, a community on the Japur, restinga baixa fields were submerged in one meter of water. In the interest of time, the decision was made to rely on local designations of habitat types rather than attempting to measure the height of the land. Due to the extremely dynamic nature of the vrzea landscape, the height of the land can change from year to year due to erosion and deposition of sediment. There is also the issue of differential flooding. Areas near the main channel of the Solimes probably receive higher floodwaters than areas near the Japur or side channels. Research Considerations of the Flood Season Field research was conducted in the summer of 2000 (May to August) during the peak of the flood. Ideally, to study agricultural biodiversity on the vrzea, it would be best to observe the agricultural fields through an entire cycle or, better yet, several cycles.

PAGE 48

37 There would be no single part of the year on the floodplain in which all diversity could be captured. The agricultural calendar is complicated and dynamic with crops being planted and harvested during several seasons throughout the year. At first consideration, it would seem that the flood season is the worst possible time to study agricultural diversity since the majority of the fields are harvested or even under water. But it is at precisely this season that we can get an idea of which species and varieties are retained in situ even during the flood. It is also significant that this research followed three consecutive years of high floods and the previous year, 1999, experienced an exceptionally high flood (Figure 2-2). The flood of 1999 was one of the five highest in the last century and caused substantial losses of seasonal crops, planting material, and perennials for residents of the Mamirau and Aman Reserves. This research is especially pertinent because it can indicate the level of agrobiodiversity that remains in the agricultural system after a period of several high floods. In addition, although the specific objective of this research was not to gather data on crop losses resulting from the high flood, a significant amount of information was collected during the inventories and interviews. It is enough to give a good, if not comprehensive, indication of which species are more susceptible to drowning and to what extent crops are damaged by high floods. Crops including corn, beans, squash and others are probably somewhat underrepresented by this study since most of the areas where they are grown, mudflats (lamas) and beaches (praias), were under several meters of water at the time of the survey. Although the attempt was made to visit all of the areas where crops were planted, including those underwater, it is likely that some areas were missed in the survey due to the informants misunderstanding, forgetting, or downplaying the importance of these

PAGE 49

38 areas. Additionally, some of the non-harvested manioc fields that were visited may have contained additional crops that were harvested before the survey and went unreported by the informant. These factors combined generate substantial room for error. It is important to emphasize that the present study is simply a snapshot in time and a much more comprehensive, long-term study would be required to capture the full range of agrobiodiversity and its dynamic nature and resilience on the vrzea. Life on the Vrzea The stretch of floodplain in the middle Solimes region is among the widest along the entire length of the Amazon river. Inhabitants of the vrzea must often travel many kilometers to reach the terra firme. The difference in level from low to high water is the highest in the Brazilian Amazon, exceeding fourteen meters in some years. Even the highest ground, the floodplain levees (restingas), may be inundated up to several months each year imposing special living conditions on residents who must cope with water covering their yards and fields for weeks on end. The vrzea landscape is one of the most dynamic on earth. The shifting river channels continually form and erode levees. High levees that are used for agriculture and settlements collapse into the river while elsewhere sandbars are colonized by pioneer vegetation and new islands and levees are formed. The time scales at which these morphological changes take place are extremely fast by geological standards with levees and islands potentially being molded or erased in less than a human lifetime. Agricultural fields concentrated in one area may shift to another area at intervals of several years in order to adjust to the rapid morphological changes in the landscape such as the closing off of channels as they become choked with vegetation and sediment and the building and destruction of levees.

PAGE 50

39 Figure 2-2. Seasonal fluctuation in water level on the middle Solimes, 1992-2000. Rise and fall (in meters) of the river near the town of Tef. In 1999, there was approximately 14 meters difference between low and high water. Source: SDR Mamirau 2000 Vrzea dwellers have adapted their lifestyles to contend with the inevitable yearly flooding. Settlements on the floodplain are located on the high levees to avoid prolonged periods of inundation (Goulding et al. 1996). Most are along major river channels for access to river traders and collective transportation. Families on the floodplain usually build their homes on stilts to avoid being inundated by the rising waters. During the weeks or months with water under their homes people rely on their canoes for even a quick trip next door. Often the water is only several inches to knee deep and trudging through mud is commonplace. Children are frequently seen playing and splashing in the river right outside their front door. Animals must be kept in the house or on floating rafts if no dry ground is available.

PAGE 51

40 The enchente or high water period is unanimously considered to be a time of hardship. Crops may not be grown on flooded land and fishing is less productive since fish are dispersed throughout the igap or flooded forest. The greatest difficulty comes when they get a high flood like the flood of 1999 (Figure 2-2). They are accustomed to this ordeal and cope with it in a variety of ways. Many leave their homes and temporarily stay with relatives in towns on the river bluffs. Others stick it out and simply build a new floor above the water level, continuing to raise it when necessary to keep pace with the rising deluge. Residents along the main channels complained of huge wakes from passing boats. They would leave their doors open in order to diffuse the force of large waves breaking against the front of their homes. Instead, the powerful waves would sweep through their living room and crash into the back wall of the house. Some avoid the invading waters by living in a floating house with the added benefit of being more mobile, even able to tow their residence to a new location. People are very resilient and bounce back quickly from such events. They rely on friends, neighbors, and kinship networks to get them through these tough times. The exceptional floods like the flood of 1953, the highest of the twentieth century, are remembered long afterwards. High floods can be quite a setback for residents. It is after one of these disruptions to their lives that families sometimes decide to make a move to terra firme. The people are highly mobile and the turnover rate for residents of the vrzea is high (Sociedade Civil Mamirau 1996). Fishing and agriculture are the predominant economic activities for residents in this part of the Amazon floodplain. To earn an income, some devote all of their energy to fishing and others focus on agriculture but the majority depend on both activities for their

PAGE 52

41 livelihood. Other activities such as the extraction of timber, firewood, and fruit also serve to supplement their income but are generally only secondary in importance. The aai (Euterpe spp.) fruit, used to make a popular drink, is commonly collected and sold. Other fruits such as yellow mombim (Spondia lutea) and genipap (Genipa americana) are common on floodplain levees and are sometimes brought to market or sold to passing river traders. Honey from native stingless bees is occasionally harvested from the forest. Another possibility to earn extra money is hunting caiman and selling the meat dried or salted. Hunting of other animals may also be an important subsistence activity especially for the communities on or near terra firme. However, selling wild game products, including caiman, is illegal in Brazil so there is not a market for game meat as in some other regions of the Amazon, such as Peru. Turtles, turtle eggs, and manatee meat fetch a high price but those who try to sell them risk getting caught and fined by IBAMA, the federal environmental protection agency. Working within or outside the community for a wage is an option for some individuals. Working as a day laborer for a neighbor is more common in terra firme communities where production of farinha or manioc flour is practiced on a relatively larger scale. Here, farmers sometimes hire day labor to weed their fields or they may contract someone with a chainsaw to clear some land. They typically pay a few dollars per day for this type of work. Some people also find temporary work in nearby towns.

PAGE 53

42 Figure 2-3. Floodplain community during low water. The building in the foreground is a floating house and the high water mark from the previous flood is half way up the walls of the homes on the bank. Boca do Mamirau, November 1999.

PAGE 54

CHAPTER 3 STRATEGIES FOR FARMING ON THE FLOODPLAIN: THE ROA Vrzea farmers contend with unpredictable floods, relentless pests and diseases, and an extremely heterogeneous environment that is constantly changing. Crop losses are frequent and can cause significant setbacks. As a means of risk management, they employ numerous strategies and take advantage of the varied terrain by utilizing several different habitats for planting crops. Reliance on rapidly maturing varieties as well as crops that are able to withstand weeks or months in standing water help farmers cope with the inevitable yearly flood. They use their knowledge of the environment to select appropriate areas for planting crops. There is an element of risk as farmers attempt to predict the magnitude and timing of the coming deluge. Residents, therefore, rely on one another for assistance when losses occur. The floodplain habitats used for agriculture are divided by farmers into four main types: beaches (praias), mudflats (lamas), low levees (restingas baixas), and high levees (restingas altas). The main factor that they use to differentiate them from one another is their altitude and, therefore, the degree to which they are subject to flooding. The distinction that is made between these habitat types can be rather arbitrary. Actually, a continuum exists between the habitat types and management depends on other factors as well such as predominant vegetation, sediment deposition, and individual preference. Farmers who have access to land on terra firme often plant there as well. In the vrzea community of Sitio Fortaleza, for example, several families have been planting on terra firme that is some distance away (one hour by motor boat or three hours rowing). 43

PAGE 55

44 Figure 3-1. The bank of the Japur River during low water. Shows a typical gradation from sandy beach to grassy mudflat to forested low levee (Nov. 1999). Several members of the community expressed their opinion about the hardships of life on the floodplain and the relative security of farming on terra firme. They have decided to move their families to high ground and others in the community may soon follow. Jubar, a terra firme community on the bank of the Japur river, also utilizes both vrzea and terra firme for agriculture. Fields, fallows, and home gardens are the three major categories of land use that farmers manage. All places where crops are planted or were planted in the past can be referred to as work areas (areas de trabalho). Annuals such as corn and beans or the semi-perennial crop manioc are planted in fields with the frequent inclusion of perennial fruit trees. Fallows range from a simple abandoned field with no useful species through a mono-cultural banana grove on up to a mixed fruit orchard. They harbor a large selection

PAGE 56

45 of useful plants and are a significant resource for subsistence and the market and often especially rich in construction materials. Home gardens contain a diverse range of fruit, vegetable, medicinal, and ornamental plants. They provide a range of foods and other useful products to the household. When an area is selected for cultivation, farmers take advantage of pre-existing valuable plants by sparing them during field clearance. They also practice selective weeding to protect useful volunteer plants, some of which are favored by forest disturbance. These farming methods alter the mix of species in regenerating forests to create a heterogeneous forest environment on the floodplain, effectively increasing the occurrence of plants that are useful to humans and, in some cases, benefiting wildlife (Pinedo-Vasquez et al. 1999, Padoch et al. 1996). It is possible that overall diversity in the region is greater as a result of the variety of management strategies employed by vrzea farmers (Bale 1994). Characteristics of the Roa In the vernacular language of the region the term roa refers specifically to a manioc field. Plots dominated by a crop other than manioc are generally referred to by another name. For example, a field dominated by corn, watermelon, or banana would be called a milheral, melancial, or bananal, respectively. The manioc tuber is planted in habitats ranging from mudflats to high levees but never on low-lying sandy beaches that are the first to flood. A limited repertoire of crops is planted on beaches; mainly beans and watermelon. In this thesis, roas are designated as work areas used to plant any annual crop as well as the semi-perennial crop manioc. Roas may be monocultures with one or several varieties of a crop but are frequently polycultures containing a number of species and varieties. Perennial fruit trees (usually commercially valuable) are

PAGE 57

46 sometimes planted with the manioc crop to be left growing when cultivation is discontinued. In a study of agricultural management systems in three communities in the Mamirau Reserve, Pinedo-Vasquez et al. (1996) described two types of planting configurations that are primarily used. The first, called randomly planted (plantio mixturado), is a system under which two or more crops are interplanted. There is often one main crop and one or more secondary crops. In the second system, stratified planting (plantio dividido), roas are divided into sections where different crops are grown based on soils and topography. According to their research, the frequency with which each system is used is dependent upon the intensity of the flood in the previous year. Farmers tend to employ the stratified planting system more often after high floods as a means of risk management. They expect high floods two years in a row so they carefully choose the highest areas of the field to plant manioc and bananas and tend to increase production of fast growing crops like squash. Roas may be cleared from forest (mata, sometimes referred to as mata bruta or mata virgem indicating its status as old-growth, mature, or virgin forest) or from any work area type included in the fallow category. Levees (restingas) are most often covered by closed canopy forest that necessitates burning in the first season. Once fields are established, burning is often unnecessary in the following seasons even when clearing young fallows. In some cases, what is called forest (mata) may actually be old secondary regrowth. Residents usually do not have knowledge of a particular forest beyond about twenty years as many of the communities on the vrzea of Mamirau are less than twenty years old. It

PAGE 58

47 is quite possible that much of the forested levees have been cultivated at some time in the past (in historic or prehistoric times), effectively altering the natural mix of species in floodplain forests. Valuable trees are sometimes, but not always, spared during forest clearance. Large timber or fruit trees can sometimes be found standing in the middle of a roa but this is fairly rare on the vrzea. It is a more usual sight in terra firme roas that are typically much larger. Very large trees are sometimes left standing because of the danger or time it takes to cut them down. They often die when surrounding forest is removed but wood can still be exploited years afterward. A common way to preserve valuable trees in both ecosystems is to clear the forest just up to the useful tree and stop, leaving it just on the border of the field locally called the aero. Farmers reported weeding levee fields up to four times during an eight-month growing season. This is done by crouching or sitting on the ground among the manioc stems and removing unwanted plants by hand or with a machete. Plants perceived of as valuable are often protected during this process. The three techniques; sparing trees during forest clearance, selective weeding, and the planting of perennials all contribute to the increased value of swidden fallows for the subsistence and income of farming families in the region. Communities differ greatly in the distance to and amount of each habitat type available to them. In some communities forest on high levees is in short supply. Residents from the community of Manacabi reported that almost all of the high levees around the community has been cultivated and there are only a few small pieces of old

PAGE 59

48 forest left. Farmers in some communities travel great distances from their homes to clear fields from old forest on high levees. Manioc is the principal staple crop of the region and the main commercial crop for many farmers who sell their surplus of processed manioc flour (farinha de mandioca). The crop is typically grown during a six to eight month season on the vrzea. According to reports by farmers, some varieties of manioc that take one year to mature on terra firme take just six months in the fertile floodplain soils. The decision is usually made to harvest when the rising waters threaten to drown the crop. Farmers harvest the lowest lying roas first in order to keep pace with the rising flood. They also begin with the most flood prone areas within a roa. They hurry to pull the tubers from the ground before they are inundated and it is not uncommon for them to fall behind and harvest a crop that is already under a few inches of water. If flooded, a crop must be rescued quickly before it rots. This usually happens within about two days. If it is apparent that the river has stopped going up, the decision is often made to leave the crop in the ground for up to twelve months to thicken (engrossar). If farmers decide to continue planting a roa for another season, they frequently begin planting immediately after or during harvest as long as they believe that the river has stopped rising. After manioc, banana is the most important crop for the majority of vrzea farmers. Bananas are sold to market by many families and are a frequent contribution to the diet of local residents. A common strategy on the floodplain is to interplant a roa with manioc and banana for one or two years and then discontinue manioc production, leaving the field in some type of banana fallow. The density of this banana plantation can vary greatly. It might be only a few plants growing in a regenerating forest, simply termed

PAGE 60

49 capoeira or fallow by farmers. With a slightly greater density of planting and intensity of management including periodic weeding, it is called a capoeira bananal or banana fallow by the owner. The farmer refers to a dense, carefully weeded banana grove as a bananal. This example, which will be explored further in the next chapter, shows that there is no clear distinction between management types or the designations that farmers give them. Because of the diversity of management practices used in banana production, the bananal was not included in the roa category but instead was placed in the next major category, the capoeira. Farming Praias and Lamas The distinction between praias (beaches or sandbars) and lamas (mudflats or silt bars) is not always clear. Beaches and mudflats are often found together with the sandy beach grading upwards to the mudflat. The ground is uneven in most floodplain fields with high and low spots consisting of varying grades of sediment that can change on a yearly basis. Crops are segregated according to topography and soil texture. The sandy beach, nearly devoid of vegetation at low water, is typically utilized to plant beans and watermelon. No preparation of the land is necessary. In one instance the informant took us to an immersed sandbar in the Solimes River. Though the land surface was several meters underwater in the middle of the huge, swiftly flowing river, he was able to show us the boundaries of the field where three families work together to cultivate watermelon and two varieties of beans, Manteiga and Comum (Preto). The size of the field is approximately 180 meters long by 40 meters wide divided into sections for each family. When the crop is nearing maturity, the families camp out on the beach to protect the crop from theft by passing boats. They take advantage of this time to collect turtles, turtle eggs, and to fish for catfish that are then salted and later sold. Farmers most

PAGE 61

50 often exploit the beaches and mudflats immediately in front of their village or on the opposite side of the river for convenience and protection from theft. Where the sandy beach grades into silt or mud, farmers grow a variety of crops including beans, corn, bitter and sweet manioc, gherkin, and squash. Most areas that were classified as mudflats by residents were covered with canarana grass (Echinochloa spp.) and a few had low Cecropia trees. In these areas a thick layer of mud is deposited when the flood subsides. Management of mudflats varies between communities and even among farmers of the same community. Of the nine fields where information on management was collected, only two were burned. For an additional two fields, it was reported that the farmers had wanted to burn but were unable to because of the weather. Of the two fields that were burned, one was burned before the flood leaving the area clean when the waters receded and the other was burned after the flood. In the second instance the canarana grass was burned both before and after slashing. In another field the grass was cut before the flood and the current carried the slash away leaving the area clean. The remaining fields were cleared using the roa-tomba-junta method (Pinedo-Vasquez et al. 1999). This means that the field is not burned. At least 2 informants reported that the vegetation is cut and thrown into piles and then carried to the edge of the field and dumped. None reported that the slashed material was used as mulch (cobertura morta) used to control weed invasion as Pinedo-Vasquez and colleagues found but the question was not consistently asked in this survey. Their research also found that mulched fields should be left standing to dry out for 25 to 30 days before planting to avoid problems of insect attack and that farmers wait to plant their fields until they see

PAGE 62

51 that certain insects have emerged as adults. This fascinating aspect of floodplain agriculture warrants further investigation. Fields on beaches and mudflats are often planted year after year without the need for fallow. For 5 fields on beaches, the average number of consecutive years planted was three with a minimum of two and a maximum of four. An additional field with beach and mudflat had been planted every year for 25 years according to residents. The average number of consecutive years planted for 15 fields on mudflats was five with a minimum of one and a maximum of twelve. The species harvested were recorded for a total of 23 fields made on beaches and mudflats. The mean number of species harvested per field was two with a maximum of five. The most commonly planted crops in sand were beans and watermelon. On mudflats, the most common is manioc. Only two of the mudflat fields contained useful species after harvest; each had one samaumeira (Ceiba pentandra), a valuable and greatly exploited timber species in Amazonia. Diversity and Management of Restinga Baixa Fields The flood brings a thick layer of rich sediment that is deposited on restinga baixa or low levees each year. This allows farmers, in many cases, to plant their fields for a number of seasons with no reported decline in productivity due to soil fertility. More than twice as many fields were visited as fallows on low levees (106 fields and 50 fallows). This is the opposite of what was observed for high levees where there were many more fallows than fields (about 113 fields and 176 fallows). This could be a reflection of the fact that farmers do not need to clear new fields as often on low levees. Actual results from fallows for the number of years previously planted was about 3.75 years for low levees and 3.5 years for high levees. An alternative hypothesis for the

PAGE 63

52 greater number of fields on low levees is that, because fallows on low levees hold relatively few useful species, the location of these areas is quickly forgotten and they are, therefore, underrepresented in the survey. Out of ninety fields where the information was collected, the average number of consecutive years planted was 3.1 with a maximum of fifteen. Management varies on low levee fields depending on location. Describing the specific factors that determine how the fields are managed is outside the scope of this research. However, it is likely to depend on influences such as depth of flood, sediment load and deposition, speed of current, vegetation, and weather. Other factors that are not determined by the environment may include available labor, available time, knowledge, custom, and personal preference. The number of seasons that farmers cultivate low levee fields is apparently at least partially dependent on the maintenance of soil fertility through the annual deposition of sediment. In the community of Jubar on the Japur River they are able to plant just two seasons before production begins to fall and the disease tuber rot (podre de batata) affects the manioc crop. According to one informant, they typically plant two years and leave the field fallow for seven years. This appears to corroborate what one farmer said from the village of Betania, also on the Japur River. According to him, on the banks of the Japur where the water carries a lower sediment load they typically plant three to five years and then leave the field fallow. However, on the side channel (parana) where the water carries more sediment, they can plant every year and the tuber does not rot. The method of field preparation on low levees varies. Fire is used in some cases but usually only when opening a field for the first time. Jubar farmers reported that

PAGE 64

53 nobody burns low levee fields; they are only slashed and cleaned. In Betania and Marirana at least some low levee fields are burned after clearing only in the first season and it is not necessary the following seasons. In the village of So Paulo, it was observed that some farmers slash a field on low levees before the flood. Pinedo-Vasquez et al. (1996) found that a small proportion of fields were burned in 1993-94 in the three communities of Vila Alencar, Jaraua, and Barroso and those that were burned had been cleared too late in the dry season. Most fields on low levees were made on land that had been fallow. Out of 82 fields, 55 (67%) were reportedly made on land that had previously been fallow and 27 (33%) were reportedly cut from virgin forest (mata) (Table 3-1). At least 11 of the fields cut from virgin forest had been imbaubais, stands consisting almost exclusively of Cecropia trees and canarana grass. For 17 of the fields that had previously been in fallow, the average length of time in fallow was 6.3 years with a minimum of two years and a maximum of twenty. Crops harvested from low levee fields in 2000 include beans, corn, gherkin, bitter and sweet manioc, okra, squash, tobacco, and watermelon. The average number of species that were harvested from 58 fields was two with a minimum of one and a maximum of seven. It is not uncommon that farmers alternate crops, planting one crop one year and another crop the next. There were 13 fields that had not yet been harvested. At least two of these were being harvested at the time of fieldwork. Although these fields were classified as low levees by informants, they fall on the high end of the spectrum since most low levee fields were flooded to a depth greater than one meter. All of them were planted with

PAGE 65

54 manioc and only four contained other species. These include two fields with mulateiro (Calycophyllum spruceanum), and one each with banana (Musa sp.), sugar cane (Saccharum sp.), and ing (Inga sp.). Eight fields held from one to five manioc varieties with an average of 2.6. Ninety-three of the fields visited on low levees had been harvested and only 16 of those (17%) had useful species remaining. Of these sixteen fields, there was an average of 2.4 species in each. The eight species that were found in at least two fields are shown in table 3-2. All of these are relatively flood tolerant except papaya (Carica papaya), a weedy species easily spread by birds. Only banana, guava (Psidium guajava), and cocoa (Theobroma cacao) were likely planted here. Aai preto (Euterpe precatoria), genipap (Genipa Americana), mulateiro, and the kapok tree (Ceiba pentandra) are volunteer species frequently encountered in fields and fallows. Other species that were found in only one field each include: cashew (Anacardium occidentale), sugarcane, cubiu (Solanum sessiliflorum), cuia (Crescentia cujete), feijo de praia (Vigna unguiculata), jauari (Astrocaryum jauari), jit (Guarea sp.), corn (Zea mays), capsicum pepper (Capsicum sp.), pitomba (Talisia esculenta), and taxi (unknown). Three of these are volunteers; jit and taxi are timber trees and jauari is a common vrzea palm used for fish bait. One field also contained a small patch of manioc that the farmer had failed to harvest and one had already been replanted with manioc. Table 3-1. Summary of statistics for low levee fields N Mean Min. Max. S Consecutive Years Planted 90 3.1 1 15 2.6 Size (m 2 ) 1 105 1643 100 6000 1247 Last Fallow Period (years) 17 6.3 2 20 4.6 Species Harvested 58 2.4 1 7 1 Manioc Varieties 24 2 1 5 2 1 the largest field (16,000 m 2 with four owners) was dropped

PAGE 66

55 Table 3-2. The eight most common species remaining in sixteen harvested fields on low levees Rank Local name English name No. fields (16) % 1 Banana Banana 5 31 2 Goiaba Guava 4 25 3 Aai Aai 3 19 3 Cacau Cocoa 3 19 3 Jenipapo Genipap 3 19 3 Mulateiro Mulato wood 3 19 3 Samauma Kapok tree 3 19 4 Mamo Papaya 2 13 Restinga Alta: The Preferred Habitat for Planting Diversity of crops and management systems is greatest in the restinga alta or high levee. Several years can pass without flooding making them the most favorable environment for planting bananas and other perennials. Informants repeatedly voiced their opinion that high levees are the preferred areas for clearing fields. Local farmers designated eleven types of work areas on high levees (not including home gardens) with a range of crops and management strategies. There are two main types of fields on high levees, the roa and the corn field (milheral). These fields are predominately planted with manioc and/or corn but are often mixed with banana and other annuals and perennials. The other nine types of work areas will be discussed in chapter four on fallows. The generic designation of fallow is only for convenience as many are not simple abandoned fields. They range from having no management and no useful species to having intensive management and dozens of useful species. Some high levee fields are planted year after year while others are planted just two or three seasons before being put in fallow. The number of consecutive years planted averaged 2.8 for 80 fields (Table 3-3). For comparison, some fields owned by residents of Betania had been planted up to 15 years in a row while in another area residents of

PAGE 67

56 Jarau reported that fields could only be planted two seasons before fallow was necessary. Both areas are high levees but by examining the location one difference is clear; the Betania fields are along the banks of the Japur where a much greater amount of sediment is deposited while the Jarau fields are on a paran or side channel that carries little sediment. Table 3-3. Summary of statistics for high levee fields N Mean Min. Max. S Consecutive years planted 80 2.8 1 15 2.4 Size (m 2 ) 132 2573 150 20000 2490 Last fallow period (years) 18 6.9 4 20 3.9 Number of species (planted fields) 82 4.1 1 14 3.3 Number of species (harvested fields) 53 2.8 1 8 2.1 Number of manioc varieties 70 2.3 1 6 1.2 Number of banana varieties 72 1.4 1 4 0.7 According to the survey, high levee fields are cleared from forest (mata) about half the time. Out of 91 roas, 40 had been cleared from forest, 11 from banana groves (bananal), and 40 from fallows (capoeira). Available data for the age of 18 capoeiras that had been cleared for roas yielded an average of 6.9 years. The ages ranged from 4 to 20 years. Reports from informants indicate that, in general, roas are only burned in the first season after they are cleared from forest or old secondary forest (capoeira velha). In subsequent years, farmers generally use the roa-tomba-junta method for clearing the fields. The average size of 132 roas was 2,573 square meters. They ranged from 150 square meters to two hectares in size. The 135 fields visited on high levees are divided into two groups for comparison; planted and harvested. The number of useful species in 82 planted fields ranged from 1 to 14 with an average of 4.1. Figures 3-2 and 3-3 show graphically the number of useful species encountered in planted and harvested roas. Planted fields with a single useful

PAGE 68

57 species are relatively infrequent, amounting to just 15 (18%). It is much more common for fields to be interplanted with two or more crops. A little more than half of the roas on high levees were found to still contain useful species after harvest. Out of 53 harvested roas, 31 (58%) of them held an average of 2.8 useful plants with a maximum of eight. Manioc is the principle crop planted in high levee fields. The tuber was found growing in all but one of the planted fields. The crop is usually grown in association with banana. Fifty-six of the planted roas (68%) contained these two crops. Figure 3-3 shows that the majority of harvested fields are left with a single crop. This reflects the common practice of collecting the manioc, leaving only banana. Twenty out of the 31 harvested fields with useful plants contained banana, by far the most frequently encountered useful species left in fields after the manioc is collected. Table 3-4 lists the 19 most common species found in planted fields on high levees with the corresponding number of harvested fields. It shows that the number of planted and harvested fields with bananas is relatively consistent at 70% and 65% respectively. The slightly lower number of harvested fields with bananas is probably a reflection of the fact that they are more likely to be on lower lying areas than planted fields. A greater proportion of planted perennials and non flood tolerant species such as aai, cocoa, lime (Citrus aurantifolia), and capsicum pepper are observed in planted fields, lending support to this idea. Corn was often left in harvested fields because of ill-developed ears or attacks by pests. Occasionally, a small patch of manioc remained in the roa that had either not been harvested in time or was yet to be collected. A number of other useful species that had

PAGE 69

58 either been planted, protected, or had sprouted spontaneously remained in fields. Lime, sugarcane, cocoa, and possibly aai were planted. 0510152025301234567891011121314Number of SpeciesNumber of Fields Figure 3-2. Number of species in 82 planted fields (roas) on high levees 0246810121412345678Number of SpeciesNumber of fields Figure 3-3. Number of species in 31 harvested fields (roas) on high levees The 31 manioc varieties recorded growing in high levee roas are shown in table 3-5 and 3-6. Nine of these are sweet manioc (macaxeira) while the rest are bitter manioc (mandioca). Sweet manioc is typically grown in small patches within roas of bitter manioc or sometimes in home gardens. Out of 70 roas, 25 (36%) had only one variety of manioc while the majority had two or more varieties (Figure 3-4). The number of varieties ranged from one to six with an average of 2.3. By far, the two most common were the bitter varieties Pacu and Valdevina. At least one of these two were found in over 75% of all roas. Qualidades, as locals refer to varieties, may be called different names by different communities or individuals. In some cases more than one name may

PAGE 70

59 be given for the same variety. Further study is called for on the diversity of manioc varieties used by varze farmers. Table 3-4. The 19 most common useful species found in planted fields on high levees with corresponding number of harvested fields Rank Local name English name Planted fields n=82 % Harvested fields n=31 % 1 Mandioca Manioc 81 99 6 19 2 Banana Banana 57 70 20 65 3 Aai Aai 14 17 3 10 4 Cacau Cocoa 13 16 2 6 5 Mulateiro Mulato wood 12 15 5 16 6 Samauma Kapok tree 11 13 3 10 7 Mamo Papaya 10 12 2 6 8 Cana Sugarcane 9 11 1 3 9 Jenipapo Genipap 6 7 3 10 9 Limo comum Lime 6 7 1 3 9 Tapereb Yellow mombim 6 7 5 16 10 Cedro Cedar 5 6 2 6 10 Jit Jit 5 6 1 3 10 Ucuba Virola 5 6 1 3 11 Car New world yam 4 5 0 0 11 Jerimum Squash 4 5 0 0 11 Milho Corn 4 5 5 16 11 Muiratinga Muiratinga 4 5 2 6 11 Pimenta Capsicum pepper 4 5 0 0 There are two species of aai but aai preto is most common on the vrzea. Nine varieties of banana were recorded in 74 fields on high levees (Table 3-7). In this case, just over half (56%) of the fields held only one variety of banana (Figure 3-5). The greatest banana diversity found in a single field was four types with an average of 1.5. The most common variety, banana prata, was encountered in 84% of all roas. The second most abundant type was banana comprida, a variety of plantain or cooking banana, found in 44% of roas. Certain varieties are more resistant to flooding, pests, and diseases.

PAGE 71

60 051015202530123456Number of Manioc VarietiesNumber of Fields Figure 3-4. Number of manioc varieties in fields on high levees 010203040501234Number of Banana VarietiesNumber of Fields Figure 3-5. Number of banana varieties in fields on high levees

PAGE 72

61 Table 3-5. Varieties of bitter manioc (mandioca) growing in 70 fields on high levees Rank Variety Num. Fields % 1 Pacu 28 40 2 Valdevina 25 36 3 Olho roxo 8 11 4 Calai 7 10 5 Casca grossa 6 9 5 Ouro 6 9 6 Azulona 5 7 6 Pretona 5 7 6 Sisa 5 7 7 Antinha 4 6 7 Samauma 4 6 7 Tartarugo 4 6 7 Traira 4 6 8 Canjiru 3 4 9 Ona 2 3 10 Amendoem 1 1 10 Auraninha 1 1 10 Azulo 1 1 10 Colaia 1 1 10 Ourinho 1 1 10 Pelonia 1 1 Rank is combined for varieties bitter and sweet manioc. Table 3-6. Varieties of sweet manioc (macaxeira) growing in 70 fields on high levees Rank Variety Num. fields % 4 Po 7 10 5 Macaxeiro 6 9 7 Amarela 4 6 7 Branca 4 6 7 Pretinha 4 6 9 Preta 2 3 10 Cabral 1 1 10 Colombiana 1 1 10 Pagoa 1 1 Rank is combined for varieties of bitter and sweet manioc.

PAGE 73

62 Table 3-7. Varieties of banana growing in 70 fields on high levees Rank Variety No. fields (N=70) % 1 Prata 59 84 2 Comprida 31 44 3 Ma 8 11 4 Nadja 4 6 4 Pacovo 4 6 4 Pacovinha 4 6 5 Guariba 2 3 6 Baiazinha 1 1 6 Urucuri 1 1 Crop Losses from Flooding, Pests, and Diseases Locals view the yearly flood, especially the occasional high flood, as a major difficulty of surviving on the floodplain. Even in normal years, losses can occur when the erosive power of the current causes levees to collapse and fields are partially or entirely swept away. Farmers usually attempt to harvest their crops just ahead of the rising waters. The flood can be unpredictable, coming too fast or at the wrong time, leaving farmers with insufficient means to harvest their crops. They usually cultivate several fields in different locations and are sometimes faced with the decision of which fields will be sacrificed to the river and which ones will be saved. When manioc fields are drowned, the tubers quickly rot and farmers sustain additional losses when their planting stock for the following season is destroyed. Families must depend on relatives, friends, neighbors, and even neighboring villages to recover from these losses. A number of families indicated that they had lost partial or entire crops during the high flood of 1999. Families were not comprehensively interviewed concerning the loss of crops but at least ten farmers reported that they had lost all or nearly all of their manioc roas to the flood the previous year. The example of one farmer in Vila Alencar may clarify the magnitude of the losses that can accrue in a high flood event. Anselmo lost a

PAGE 74

63 roa with the equivalent of about 200 kilograms of flour (farinha) because he didnt have time to collect it. He also lost 500 banana plants and a good many fruit trees that he had planted. In addition, at least 25 individuals of nine species were drowned in his home garden. His troubles carried over into the subsequent year because he lost most of his planting material with the loss of his crops and was only able to plant two small fields. Although crop losses are potentially much greater during a high flood, farmers may still lose parts of their fields during normal years when they are unable to harvest them in time. At least two fields that were visited had small patches of un-harvested manioc standing in water. One farmer stated that he had lost his crop of beans when the water took it away. At least two farmers from two different communities complained that corn they had planted failed to sprout. In the community of Manacabi, one informant reported that his family did not plant corn last season. He had planted some seed the year before that the Mamirau project had given them and it had not sprouted. He also confirmed that corn could not be planted on the left bank of the channel because brown capuchin monkeys (macaco prego, Cebus apella) destroy it. They only plant it on the right bank (actually an island) where there are no monkeys. Another farmer explained that he normally sells a lot of squash but he could not plant it this year because of the large amount of sand deposited by the previous flood. When a high flood occurs, farmers can suffer profound losses of perennial crops. Bananas, an important source of income for many families in the region, can be wiped out as many floodplain dwellers experienced in 1999. Bananas and plantains are a convenient crop for floodplain farmers because they can harvest them year after year.

PAGE 75

64 When they are drowned, however, residents must replant large areas and it may take several seasons to build up their banana holdings once again. Young banana plants or suckers are sometimes destroyed along with the mature plants leaving planting material in short supply. Banana losses depend on the length of time an area remains inundated and certain varieties are more resistant to flooding than others. According to one informant the Prata variety can resist up to three months, Ma 2 months, and Comprida only 1 month. Another resident reported that the variety Peruana does not die in the flood. Suckers, locally called filhos (offspring), may stay under water more than three months and still escape the flood but this also probably depends on the variety. Several farmers reported that all of the banana suckers had been lost in the flood along with the crop. During the interviews, at least 20 farmers reported that they had lost their banana plantations in the high flood the previous year. Antonio of Vila Alencar reported that out of 1000 banana plants, he was left with only 90 banana prata and 150 banana comprida. Francisco of Jarau lost his banana crop and no suckers survived for replanting. In Sitio Forteleza, Antonio lost 110 of his 122 banana plants in a high levee plantation. He is one of several farmers in the community who has decided to give up agriculture on the vrzea and move to terra firme. Francisco of Vila Alencar lost his entire crop of banana that he had planted on a low levee this season without harvesting any. He plans to replant the area again this year. Perennial fruit trees that are not adapted to flooding are frequently killed in fields, fallows and home gardens. Recently planted seedlings are especially susceptible to drowning. Several villagers cited this reason for not planting fruit trees. They risk

PAGE 76

65 having all of their work wasted with the next year's flood. Table 3-8 lists the perennials that were drowned the previous season and the number of times they were reported. It does not reflect the number of individuals that were killed. Avocado (Persea Americana), lime (Citrus aurantifolia), and orange (Citrus sinensis) were the three most commonly cited fruit trees that had drowned in fields and fallows during the 1999 flood season. An interesting comment was made by several informants concerning the survivability of fruit trees in flooded conditions. These residents have observed that many fruit trees can survive when flooded as long as they are not bumped by a passing canoe or floating log. Apparently, the shock caused by the hit is enough to kill a tree that would otherwise have survived a couple feet of water over its roots. Similarly, fruit trees in calm water are much more likely to survive than those exposed to a strong current especially since the current can carry floating objects that may bump the tree. The economies of many communities have suffered from attacks on their crops by pests and diseases. According to local informants, communities along the lower Japur River relied heavily on bananas and made a good income until about five years ago when mal (bad) struck. Mal is the generic term given to unknown pests or diseases that strike their crops. Reports indicated that many families had formerly derived the major part of their income from bananas before they were hit by disease in 1994. It is common for a familys main source of income to shift depending on the season and also year by year. As Swales (1999) observed for the lower Amazon, there are no aid programs for farmers when crop losses occur and no extension workers to identify the diseases and teach how best to combat them. It is possible that the banana crop is suffering from black

PAGE 77

66 Table 3-8. Perennials that were reported lost in fields and fallows Local name English name Num. reported losses Local name English name Num. reported losses Abacate Avacado 12 Ing Ing 2 Limo Lime 11 Manga Mango 2 Laranja Orange 10 Puxuri Puxuri 2 Cupuau Cupuau 7 Abacaxi Pineapple 1 Abiu Abiu 5 Aai Aai 1 Caj Cashew 3 Caco Cocoa 1 Jambo Malay apple 3 Camu camu Camu camu 1 Lima Lima 3 Carambola Carambola 1 Pupunha Peach palm 3 Seringa Rubber 1 Andiroba Andiroba 2 sigatoka, a disease that had spread south from Central America to Colombia by 1981 (Smith et al. 1992) and may have continued spreading down the Japur River. Another disease that affects banana in the region is Moko disease (febre da banana) (Pinedo-Vasquez et al. 1996). Infected planting material can be brought into vrzea plantations when it is borrowed from the terra firme after a high flood leaves farmers in short supply. Farmers have discovered that certain varieties of banana have greater resistance to disease. Whether the bananas are grown in sun or shade also seems to make a difference. Several farmers reported that bananas grown in shade were less liable to contract a disease. Lemon is another important crop that has suffered from disease in the last several years. The history of one roa on a high levee in the community of Manacabi can illustrate the impact of both flooding and disease particularly on banana production. The farmer cleared the roa from 6-year-old secondary regrowth, burned, and planted manioc and banana. He used the common strategy in which the field is left in banana after the manioc is harvested. The high flood of 1994 drowned his banana crop. The next year he planted it again with manioc and the Ma variety of banana. Mal arrived and again his

PAGE 78

67 banana crop was killed. According to him, the disease dries the leaves, the bananas do not develop, and after about 3 months it dies. He replanted with a new variety, Prata, with a greater resistance to the disease. Last years flood destroyed that crop. Now he says that the disease is beginning to affect the Prata variety. Another example of the devastating effects of crop losses comes from the low-lying terra firme community of Boa Esperana located on the shore of Lake Aman. Some residents reported earning a relatively prosperous living in previous years by selling banana, lemon, and avocado. They had money in the bank and the harvest and sale of fruit was comparatively easier work than growing manioc and processing manioc flour. Then, in the mid-nineties, a new disease struck the bananas and greatly curbed production. Shortly afterward, lemon trees were hit with a disease outbreak and most of the trees were lost. They were left with their avocado trees that still brought in a good income. A further blow was struck in 1999 when the community lost hundreds of trees from its fields and orchards along with a substantial part of the manioc crop during the unusually high flood. The high water killed most of the avocado trees and the community had no choice but to intensify manioc production in the year 2000, reportedly initiating an increase in forest clearing. Storage and Loss of Planting Material The semi-perennial crop manioc is reproduced vegetatively. On the vrzea of the middle Solimes manioc is planted by cutting the stems into approximately 30 cm sections, making a shallow hole, and horizontally placing the sections in the ground. At harvest, leaves are stripped from some of the stems which are then saved for the following season. Depending on their method of storage, a high flood can cause crop loss not only for that year but also the following year. Planting material for the next

PAGE 79

68 season is typically stored in the field by setting bundles of stems upright on the ground (Figure 3-6). Any part that stays above water will be viable planting material but if submerged it is ruined. A safer strategy is to place the bundles of stems on a raised platform (Figure 3-7). In most years this method is successful, but the occasional high flood can submerge the platform and destroy the planting stock. The surest way to eliminate these losses is to store manioc stems on a floating raft but few people take the trouble to do this as it would entail either building a raft near the field or transporting the bulky bundles of stems to the village. Material for building the rafts, the lightweight assacu (Hura crepitans) trunks used for floats, are becoming scarce as timber extraction continues in the region (Albernaz and Ayres 1999). When planting material is lost, people must turn to their relatives, friends and neighbors for help. Fortunately, most residents are willing to share and divide up their stock among those in need. Francisca explained that during the previous season there was a large deficit of planting stock in the community of Vila Alencar. Only a little of hers was lost and she was able to donate manioc stems to eight or nine families including some in a nearby community. It was still not enough for everyone and another farmer from the same village claimed that he had planted only part of one field and left another large field fallow because he lacked manioc and banana stock. A resident of Manacabi reported that he had planted manioc in less than half of his fields because he divided his planting material with others who had none. If farmers have connections with villages on the terra firme, they may take a trip there to borrow some stock. One farmer from Barroso acquired stems from his church brothers (irmos da igreja) on the upland banks of the Japur River. He said that he was

PAGE 80

69 lucky not to have a problem with tuber rot (podre de batata) as often happens when terra firme varieties are planted on the floodplain. A farmer from Vila Alencar with a shortage of planting material said he could have obtained some in Alvares, a nearby terra firme town, but he did not have available transport. A problem that was repeatedly voiced by vrzea residents was the lack of adequate transport for situations like this and for getting their produce to market. In the year 2000, just after the high flood of 1999, many farmers claimed that they were unable to plant all of the land that they would have normally cultivated because not enough planting stock was available. Often fields were planted to only a third or a half of their size. The survey documented 16 roas that had not been planted completely because the farmer lacked planting material. The average area not planted was approximately 1500 square meters per field. Seeds of annual crops are dried and stored in bottles or cans for the following planting season. Corn is hung out to dry in the sun and may be stored hanging under the eves of the house where it can last for a year. Bananas are planted from the suckers or young shoots that sprout next to mature plants. The suckers must be taken from an existing field and brought to a new area for planting. Once the suckers are taken out of the ground they must be kept shaded and moist until planting. As was mentioned in the previous section, the suckers may also be killed when a field is inundated for an extended period. They are usually able to survive for periods of up to three months while immersed in floodwater.

PAGE 81

70 Figure 3-6. Bundles of manioc stems placed upright in restinga alta field. The water is approximately 0.5 meters deep. Community of Jarau. Figure 3-7. Storage of manioc stems on a raised platform in a high levee field. Bundles of planting material are sometimes seen stored on the side of a floating house or on a nearby raft. Community of Jarau.

PAGE 82

CHAPTER 4 CAPOEIRAS: ENRICHED FORESTS AND MANAGED FALLOWS The popular perception that swidden agriculture entails two or three years of cultivation followed by an abrupt transition to a long fallow period is not supported by actual research. In many traditional swidden systems there is a gradual transition from field to managed fallow or orchard in which production continues for years beyond abandonment (Denevan and Padoch 1987; Eden and Andrade 1987; Hammond et al. 1995). There is often a steady progression of crops that in some ways mirror the natural regeneration of the forest through the development of vegetation strata. The rural population of the middle Solimes region manage the development of fallow fields to a great extent. Farmers also manage forests to enrich them with useful species including fruits and construction materials. This chapter documents the rich agrobiodiversity across a range of habitats and managed fallows in the middle Solimes vrzea. In the Brazilian Amazon, the term capoeira means fallow or secondary forest. Farmers in the middle Solimes refer to some areas as capoeiras even when they contain useful species and receive some form of management. When the concentration of useful plants and/or degree of management is greater, they use other descriptive terms such as bananal (banana grove) or frutal (fruit orchard) to refer to these work areas located principally on high levees where flooding is occasional and less intense. Documenting the size and management history of fallows is often more difficult than gathering the same information for roas. The planted area of multiple ex-fields is sometimes lumped together by owners and informants as one capoeira that can have parts 71

PAGE 83

72 of different ages. In some cases a capoeira can be best thought of as a work area where different activities are taking place through time. A piece of the fallow may be occasionally used for a roa or bananal, or managed for other products. Small manioc patches on the order of 10x10 meters or less within a fallow are also sometimes planted. Partially as a result of this lumping together, the average size of capoeiras in table 4-4 is almost 6000 square meters compared to roas with an average of only 2600 square meters. Management practices of forest enrichment with bananas and fruits also use more extensive areas than a typical roa (some of these areas were not previously manioc fields according to the knowledge of the informant). Information on specific field divisions and management is forgotten over time and fallows can be left when residents move away or be created by farmers from other communities. With strong kinship ties in neighboring villages, farmers sometimes plant their fields in the territory of nearby communities (the territories of communities are generally not rigidly divided). Close relatives often plant their roas side by side to take advantage of the collective labor in clearing the forest and planting. In the three Mamirau communities studied by Inuma, Padoch, and Pinedo-Vasquez, fallow management or agroforestry is a specialized activity used only by a few households (Padoch et al. 1996, Pinedo-Vasquez et al. 1996). They reported that the number of managed fallows was about 40% of the total. Managed fallows were divided into two categories: 75% used the criao-da-mata (forest creation) agroforestry system which are developed from planted and protected volunteer species in a cleared field and 25% used the enriquecimento-da-mata (forest enrichment) system in which trees of useful species were protected during field clearance. The researchers stress the point that

PAGE 84

73 there is a tremendous variation in fallow management among individual farmers. They detail 12 techniques that are used in managing the stand, species, and the individual plant. These include thinning, girdling, selective cutting, removal of vines, leaving forest cover, removing forest, planting, seedling selection, pruning, weeding, insect nest removal, and bleeding. Inuma, Padoch, and Pinedo-Vasquez (1996) also describe three classes of techniques for managing forest stands with a high density of a specific valuable species. Mata demarcada (demarcated forest) involves keeping a buffer of low shrubby vegetation around the managed forest, mudana de mata (changing of the forest) entails the creation of forest gaps to protect and encourage the growth of valuable trees and shrubs, and mata mantida (maintained forest) refers to the maintenance of dense stands of fruits such as bacur-par (Rheedia sp.) or camu-camu (Myrciaria dubia). The managed forest stands are named according to the dominant economic species and include bacabais (Oenocarpus bacaba), bacuriais (Rheedia sp.), camu-camuais (Myrciaria dubia), pau mulatais (Calycophyllum spruceanum), and urucuriais (Attalea minor). The three communities tend to specialize in one or two types of forest stands that they have within their territory. Restinga Baixa Capoeiras on low levees (restinga baixa) are generally not managed and contain few useful species because they are flooded frequently and for a relatively long duration. None of the more intensely managed land use classes were encountered on low levees. Table 4-1 summarizes the data for size of fallows, number of useful species, and the number of consecutive years that the area was cultivated before the current fallow period. With a mean average of about 2900 square meters, fallows on low levees tend to be

PAGE 85

74 Table 4-1. Summary of statistics for fallows on low levees N Mean Min. Max. S Median Size 49 2894 200 17,500 3009 2000 Num. of useful species 17 2.8 1 7 1.7 3 Age of fallow (years) 50 2.8 1 6 1.47 3 Num. of consecutive years planted previously 12 3.75 1 8 2.2 3 smaller than those on high levees. This is probably because the more land extensive fallow management techniques are little used on low levees. The size reflects more accurately the smaller size of roas on low levees. Compared to the 2600 square meter average high levee roas, a low levee roa is about 1600 square meters or just over half the average capoeira size. This discrepancy between the size of roa and capoeira is most likely a result of sometimes lumping individual roa parts into one capoeira. The average age of fallows ranged from one to six with an average of just under three. The number of consecutive crops planted previously varied from one to eight years with an average of almost four. Only 34% of the fallows contained useful species. Seventeen out of 50 low levee capoeiras held from one to seven species with an average of three. Table 4-2 shows the 25 useful species with nine uses encountered in low levee fallows. There is a large spread of useful species with low frequency counts. By making more observations the list of useful species would certainly continue to grow. The two most common, aa (Euterpe oleracea), used for making a beverage, and the Kapok tree (Ceiba pentandra), a valuable timber, were each encountered in five out of 17 fallows. It is interesting to note that these two highly esteemed trees, occur in low levee fallows in only two communities each and tend to be found near one another. The different environments where each community plant their fields partially explains why some species tend to be found in higher concentrations in some communities and not others.

PAGE 86

75 Table 4-2. The 25 useful species occurring in 17 capoeiras on low levees Local name English name Use Num. Capoeiras (17) % Aa Aa 2 2 5 29 Samama Kapok tree 3 5 29 Tapereb Yellow mombim 1 4 24 Jenipapo Genipap 1 3 18 Lauro inamu Lauro inamu 3 3 18 Mulateiro (Pau mulato) Mulato wood 3 3 18 Cacau Cocoa 1,2 2 12 Castanha de macaco Castanha de macaco 7 2 12 Inga Ing 1 2 12 Muiratinga Muiratinga 3 2 12 Puru Puru 1 2 12 Taxi Taxi 3 2 12 Andiroba Andiroba 3 1 6 Bacaba Bacaba 2 1 6 Bacuri coroa Bacuri coroa 1 1 6 Bacuri liso Bacuri liso 1 1 6 Banana Banana 1 1 6 Cedro Cedar 3 1 6 Goiaba Guava 1 1 6 Jacareuba Jacareuba 3 1 6 Jauari Jauari 5 1 6 Limo comum Lime 1 1 6 Macacaricuia Macacaricuia 3 1 6 Paricarana Paricarana 3 1 6 Ucuba Virola 3 1 6 Use categories: 1 Fruit, 2 Beverage, 3 Timber, 5 Fish bait, 7 Nut, 9 Medicinal, 10 Latex, 11 Vegetable, 12 Livestock feed There are two possible species of aa; aa do Par (E. oleracea), and aa preto (E. precatoria). Numbers three and four on the list, yellow mombim (Spondias mombim) and genipap (Genipa Americana), are two large fruit trees that seem to thrive in swidden fallows on the floodplain. They are rarely planted but often protected. In some cases however, they will be sacrificed in the interest of clearing a manioc roa. The next two on the list, lauro inamui (Ocotea cymbarum) and mulateiro (Calycophyllum spruceanum), found in three fallows each, are both valuable timber trees that sprout easily in manioc roas and are often protected and managed.

PAGE 87

76 Restinga Alta In contrast to low levee fallows that are flooded yearly for several months, fallows on high levees (restinga alta) are often managed and contain a much greater diversity of beneficial plants. Middle Solimes farmers use a variety of terms to designate work areas on high levees. The terms range from the fallow with no useful plants to the sitio (site), an area that is comparable to a home garden in the great diversity of useful species. A list of the work area designations and their descriptions is shown in table 4-3. The number of work areas visited gives an idea of the relative proportions in which they occur. The majority (70%) are simple fallows that have a lower average number of useful species than most of the other more intensively managed fallow types that made up 60 out of 202 (30%) non-roa work areas on high levees. Table 4-4 compares the size of non-roa work areas. The complexity of management practices and diverse land use histories and the ambiguity of designations of land use types by different farmers makes it difficult to precisely divide the work areas into discrete categories, resulting in some overlap. Of all the areas not designated as roas, 75% contain useful species. The intensity of management and number of species vary widely across these work areas (Table 4-5). Some fallows contain only useful plants that were either spared when the roa was cleared or sprouted spontaneously while others are planted with a few dominant crops such as banana and aa. Still others contain a diverse array of carefully tended fruits, vegetables, timber, or medicinal plants. Management techniques such as weeding and pruning vary in intensity from frequent to none. The non-roa work areas can be split into two general groups; those that are considered capoeiras and those that are not. A capoeira may be specified further by local

PAGE 88

77 Table 4-3. Local designations of work areas, number of areas visited, and their description Local Term English translation Num. of areas visited Description Bananal Banana Grove 17 Banana plantation or forest enriched with banana Cacaual Cocoa Grove 0 Area planted with cocoa Capoeira Fallow 142 Fallow field Capoeira bananal Banana Fallow 22 Fallow enriched with banana or former bananal Capoeira cacaual Cocoa Fallow 2 Former cacaual Capoeira fruteira Fruit Fallow 12 Fallow enriched with fruit trees Frutal Fruit Orchard 2 Area enriched with a diversity of fruit trees Sitio Site 5 Area of intensive, long-term management and/or a high diversity of useful plants No cacaual areas were encountered during the survey but a few capoeiras were said to be former cocoa groves. Table 4-4. Average size by type of capoeira Type N Mean size (m 2 ) Min. (m 2 ) Max. (m 2 ) S (m 2 ) All Areas 1 188 5978 600 50000 7899 Bananal 16 4222 400 10000 3420 Capoeira 2 133 5901 500 50000 7924 Capoeira bananal 21 5362 400 40000 8362 Capoeira cacaual 2 2525 1050 4000 Capoeira fruteira 3 9 10856 1400 48000 16630 Frutal 2 36000 12000 60000 Sitio 5 4760 2000 8000 2385 1 The five largest and smallest were dropped in calculating the average. 2 The two largest and smallest were dropped in calculating the average. 3 The largest and smallest were dropped in calculating the average.

PAGE 89

78 residents as a capoeira bananal (banana fallow), capoeira cacaual (cocoa fallow), or capoeira fruteiras (fruit fallow) based on the relative amount of work that has gone into enriching them with valuable plants and the density of planted species. Three types of land use, bananal (banana grove), frutal (fruit orchard) and sitio (site) are not considered to be capoeiras by farmers. The owner has generally invested much more time and energy into these areas and the density of planted species is relatively higher. Table 4-5. Average number of useful species by type of capoeira Type N Mean num. species Min. Max. S Bananal 17 2.4 1 7 2.2 Capoeira 88 3.5 1 18 3.1 Capoeira bananal 21 3.6 1 12 2.9 Capoeira cacaual 2 1.0 1 1 Capoeira fruteira 12 10.2 5 19 4.3 Frutal 2 16.0 13 19 Sitio 5 20.8 11 34 Only capoeiras that hold useful species are included. Bananal Banana is found growing in all types of work areas on high levees including home gardens. It is also planted on low levees where it is harvested after six months just before the flood (or just after) and completely replanted the next season. Although banana may still come in after manioc in importance as a regional staple, it was overall the most frequently observed species in the survey with 325 total observations compared to manioc which came in second with 210 observations. Its ubiquitousness across many types of habitats and management areas underscores its versatility and adaptability to the floodplain environment and its status as an important commercial and subsistence crop. Bananas are produced in a monocultural grove or intercropped with manioc. They are also found growing scattered here and there among roas, capoeiras, or home gardens.

PAGE 90

79 The bananal is formed using one of three techniques. A common strategy is to interplant banana with manioc in the roa on high levees. When manioc cultivation is discontinued, the area is left as a bananal often for several years until a high flood destroys the bananas at which time it may be cleared to plant a roa once again or left as a capoeira. The second technique is to clear an area specifically for planting bananas. The area may be cleared from either forest (mata) or secondary forest of any age. It may or may not be burned. The third way is to clear only the undergrowth and select trees in a mature forest or secondary forest leaving large trees standing to provide shade for the crop. The area is often cleared strategically with undesirable trees removed and valuable ones left in place. Several informants told us that, by growing bananas in the shade, some disease can be avoided. In all three of these systems, various fruit trees may also be added creating a bananal with several economically valuable species in the mix. Out of 10 banana groves (bananais plural) where the information was collected, half had been planted to roa at least one year at the beginning of the cropping sequence. In the other five, only banana was planted. At least three of these five areas were not completely cleared of trees. Instead, the underbrush was cut and the bananas were planted under shade. A particularly interesting bananal was visited on a very wide high levee near the community of Vila Alencar. This area was referred to as mata bruta (mature forest or regrowth) and was apparently very old secondary forest. Only a few trees were cut and some trees were ringed but there was still sufficient canopy cover to shade out grasses. Weeds of only a few families (Musaceae, Piperaceae, and Solonaceae), soft-stemmed and easily weeded, are scattered throughout the field. According to the informant, after the

PAGE 91

80 bananas are grown and weeded, the weeds do not come back and some disease is avoided. In another field it was explained that roa was planted only the first year. After harvest, the secondary forest was allowed to grow back and provide shade for bananas and other fruit trees planted under the canopy. Besides banana, two other species (aa and cocoa (Theobroma cacau) were planted and several timber species had been spared including muiratinga (Maquira spruciana), virola (Virola surinamensis), and the kapok tree. The area is occasionally weeded and other seedlings may be added. When bananas are killed by drowning like in the 1999 flood season, surviving suckers are gathered, the undergrowth is slashed, and the bananal is refurbished. We visited 17 areas that were designated as bananal by informants. The average number of species and average size was slightly lower than the other capoeiras but the sample size is insufficient to conclude that there is a significant difference between the two types of land uses (Tables 4-4 and 4-5). The average age of 11 bananais was 3.4 years with a maximum of 7 years (Table 4-7). High floods necessitate the periodic refurbishment of a bananal which may occur multiple times during the period of cultivation. Farmers sometimes change the location of the bananal when disease arrives in the hopes of achieving a disease free crop for a couple of seasons. Leaving a bananal fallow for a time can also help in controlling disease. Bananais contained the lowest average number of species of any capoeira type except the capoeira cacaual. The planted species include aa, banana, cocoa, sugarcane, cupuau (Theobroma grandiflorum), breadfruit (Artocarpus altilis), and manioc (Table 4-6). Other species were either protected during forest clearance or sprouted (or

PAGE 92

81 resprouted) spontaneously. One farmer informed us that he spared castanha de macaco (Couroupita guianensis), cedar (Cedrella fissilis), and muiratinga (Maquira spruciana) in his bananal at the time of clearing. A number of naturally occurring trees are economically valuable timber species that are harvested in the region. These include cedar, muiratinga, mulateiro, kapok, taxi, and virola (Albernaz and Ayres 1999). Ing and yellow mombim (Spondius mombim) are two fruit trees that also occur spontaneously on the floodplain. Itu (Gnetum nodiflorum) is a vine which bears fruit that can be roasted and eaten and is also utilized as fish bait. Varietal diversity was not found to be very great in bananais (Table 4-8). The average number of banana varieties in 10 fields was 1.3 with seven containing only a single variety. Only three varieties of banana were registered in these areas. The Prata variety has replaced Ma as the most popular variety in the region due to its elevated resistance to disease. It was found in eight out of ten bananais. The Comprida variety was second, occurring in four bananais while Ma was found in only one. Bananais were greatly affected by the previous flood season possibly lowering the average number of varieties found in the fields but this may not be the primary reason why there exists a relative dominance of a few select varieties. In his study of agricultural intensification by a Tikuna Amerindian community in the Upper Solimes, Shorr (1999) revealed the increasing dominance of a few varieties of manioc and banana in recent decades. According to his research, it was increasing market participation that caused the shift in their crop repertoire. The Comprida variety was more frequently planted due to its durability in transport, greater marketability, and higher price. It is

PAGE 93

82 likely that a similar process has been occurring in the Middle Solimes as farmers have become increasingly integrated into local markets. Table 4-6. The 22 useful species occurring in 17 banana groves (bananais) Local name English name Use Num. Bananais % Banana Banana 1 17 100 Aa Aa 2 2 3 18 Cacau Cocoa 1, 2 2 12 Cedro Cedar 3 2 12 Muiratinga Muiratinga 3 2 12 Cana Sugarcane 2 1 6 Carapanaba Carapanaba 3, 9 1 6 Castanha de macaco Castanha de macaco 7 1 6 Cupuau Cupuau 1 1 6 Fruta pu Breadfruit 1 1 6 Ing Inga 1 1 6 Itu Itu 1, 5 1 6 Mandioca Manioc 8 1 6 Mulateiro Mulato wood 3 1 6 Murumuru Murumuru palm 5 1 6 Namuirana Namuirana 3 1 6 Pau ferro Pau ferro 3 1 6 Samama Kapok tree 3 1 6 Tanibuca Tanibuca 3 1 6 Tapereb Yellow mombim 1 1 6 Taxi Taxi 3 1 6 Ucuba Virola 3 1 6 Use categories: 1 Fruit, 2 Beverage, 3 Timber, 4 Bowl, 5 Fish bait, 6 Thatch, 7 Nut, 8 Tuber, 9 Medicine There are two possible species of aa; aa do Par (E. oleracea), and aa preto (E. precatoria). Table 4-7. Summary of statistics for banana groves (bananais) N Mean Min. Max. S Median Age of bananal 11 3.4 1 7 2.0 3 Num. of species 17 2.41 1 7 2.24 1 Num. of banana varieties 10 1.3 1 2 0.48 1 Table 4-8. Banana varieties occurring in 10 banana groves (bananais) Variety No. Bananais % Prata 8 80 Comprida 4 40 Ma 1 10

PAGE 94

83 Capoeira Farmers distinguish between two classes of secondary regrowth, old secondary regrowth (capoeira velha) and young secondary regrowth (capoeira nova) but age stipulations can vary by community or individual. For example, one informant may designate a 1 to 3-year-old fallow as young and anything older than 3 old while an informant from another community may consider fallows up to 5 or even 8 years old. There is never a sharp distinction between the two designations. In the areas designated as capoeira, 88 of 142 (62%) were found to contain an average of 3.5 useful species. Maintenance of swidden fallows consists of occasional weeding around bananas and fruit trees by slashing encroaching vegetation with a machete. The capoeiras can be divided into two distinct types; those that were formerly roas and those that were formerly bananais. Information on previous land use was collected for 111 fallows. Most of the areas designated as former bananais were not planted with manioc at the start, agreeing with the information on management history of bananais presented above. A few of the areas were planted with roa the first year and then planted to banana (and sometimes other fruits) in the following years. It is useful to compare these two types of fallows to see which form of management may lead to a greater diversity of useful plants and the nature of the beneficial species that are retained. Table 4-9 shows data on the number of useful species found, age, and number of consecutive years formerly under cultivation for the two types of capoeiras. Fifty-two of 79 (82%) former roas contained an average of 3.5 species. Less of the former bananais were found to contain useful species, only 20 out of 32 (60%). The average number of species in the former bananais were slightly higher with 4.4 species.

PAGE 95

84 Table 4-9. Summary of statistics for fallows (capoeiras) on high levees N Mean Min. Max. S Median Useful Species 52 3.52 1 18 3.33 3 Banana Varieties 20 1.3 1 3 0.57 1 Age of Fallow (years) 77 3.4 1 13 2.5 3 Years Planted Before 53 3.4 1 15 3.2 2 Former roa Size 1 (m 2 ) 72 5196 800 18000 4770 4000 Useful Species 20 4.35 1 13 3.25 4 Banana Varieties 9 1.1 1 2 0.33 1 Age of Fallow (years) 31 3.4 1 15 3.3 2 Years Planted Before 22 3.6 1 12 2.7 2 Former bananal Size 2 (m 2 ) 30 6891 1000 28000 7615 3950 1 Two largest and smallest values were dropped in calculating the mean. 2 Largest and smallest values were dropped in calculating the mean. Table 4-10. Banana varieties occurring in fallows (capoeiras) on high levees Variety Num. capoeiras (41) % Num. former roas (23) % Num. former bananais (13) % Prata 31 76 17 74 10 77% Ma 8 20 4 17 4 31% Comprida 5 12 3 13 2 15% Nadja 3 7 1 4 2 15% Pacovinha 1 2 1 4 0 0%

PAGE 96

Table 4-11. Twenty-five most frequently encountered species and their use in capoeiras on high levees with a comparison of former roas and former bananais Local name English name Use 1 Capoeiras (88) % Former roas (52) % Former bananal (20) % Banana Banana 1 41 47 23 44 13 65 Aa Aa 2 2 26 30 15 29 9 45 Cacau Cocoa 1 24 27 13 25 6 30 Mulateiro Mulato wood 3 14 16 10 19 3 15 Tapereb Yellow mombim 1 14 16 7 13 5 25 Lauro inamui Lauro inamui 3 12 14 5 10 6 30 Ucuba Virola 3 11 13 7 13 4 20 Tacaca Tacaca 3 10 11 5 10 4 20 Assacu Assacu 3 8 9 6 12 1 5 Bacaba Bacaba 2 8 9 4 8 4 20 Limo comum Lime 1 8 9 6 12 1 5 Manga Mango 1 8 9 4 8 3 15 Muiratinga Muiratinga 3 7 8 4 8 1 5 Cedro Cedar 3 6 7 4 8 2 10 Cuia Calabash goard 4 6 7 3 6 0 0 Cupuau Cupuau 1 5 6 3 6 2 10 Jit Jit 3 5 6 3 6 2 10 Cana de aucar Sugarcane 2 4 5 1 2 0 0 Fruta po Breadfruit 1 4 5 3 6 0 0 Ing Inga 1 4 5 3 6 1 5 Jambo Malay apple 1 4 5 3 6 1 5 Jauari Jauari 5 4 5 3 6 1 5 Jenipapo Genipap 1 4 5 2 4 0 0 Mamo Papaya 1 4 5 2 4 0 0 Urucuri Urucuri palm 6 4 5 1 2 2 10 85 Use categories: 1 Fruit, 2 Beverage, 3 Timber, 4 Bowl, 5 Fishbait, 6 Thatch There are two possible species of aa; aa do Par (E. oleracea), and aa preto (E. precatoria).

PAGE 97

86 The results for both the average age of the fallow and the average number of years formerly under cultivation were three and a half years; virtually the same for both land use types. A total of 68 useful species were recorded in 88 capoeiras on high levees. Table 4-11 shows the 25 most frequently encountered and gives an indication of their relative dominance in former roas and former bananais. Although the sample of former bananais is small, it can give us an indication of possible differences between the two land use types. Banana is the most common valuable plant growing in fallow areas. The percentage of former bananais containing banana was higher with 65% compared to former roas with 44%. Aa also occurred 16% more frequently in former bananais. The fact that yellow mombim occurs 12% more frequently in former bananais would makes sense since it is a tree that would more likely be sacrificed when clearing land for a roa because of the need for full sunlight in manioc production. It is more convenient to spare fruit trees like yellow mombim and genipap or valuable timber in a bananal where large trees are often preserved to provide some shade and the banana plants are often scattered over a wide area. Of the 25 most frequently encountered useful plants in capoeiras, the majority (11) are fruit trees, eight are timber trees, three are used for beverages, and one each are used for fish bait, thatch, or bowls. Five varieties of banana were found growing in capoeiras on high levees (Table 4-10). In 29 capoeiras, the average number of varieties was 1.2. Only six capoeiras (21%) had more than one variety; five had two varieties, and one had three. Five of these are former roas and one is a former bananal. Again, prata is the variety that dominates in these areas. This variety was found in 76% of the 41 capoeiras where information on

PAGE 98

87 variety was collected. It occurred in almost the same percentage of former roas and former bananais. Ma is the second most common variety, found in 20% of 41 capoeiras. Comprida is third with 12 percent. Two relatively uncommon varieties, Nadja and Pacovinha, were also found. Capoeira Bananal Local residents sometimes distinguish the capoeira bananal (banana fallow) from a regular capoeira. It is relatively enriched with bananas and often with other fruits. It can result from several possible origins. After a high flood kills much of a crop in a more densely planted bananal the status of the work area is downgraded to capoeira bananal. The area is often replanted to renew its status as a bananal. The capoeira bananal can also result from a harvested roa that was interplanted with bananas. It may be only partially planted with bananas. In that case it is more like a small bananal within a larger capoeira. It can also be capoeira of any age that is enriched with scattered banana plants. Out of 13 capoeira bananais (plural) surveyed, six were formerly roas, four were mata (forest), two were bananais, and one was capoeira. The ages of capoeira bananais range from 2 to 18 years with an average of almost four (Table 4-12). The number of years capoeira bananais was previously planted with crops, reported in only three cases, was two. The capoeira bananal is a range along a continuum of plant diversity in the agricultural ecosystem of the middle Solimes vrzea. They contained from 1 to 12 useful species with an average of over three. The most frequently encountered species are those that commonly sprout spontaneously in restinga alta fields and fallows (Table 4-13). These include aa, genipap, papaya, mulatto wood, and yellow mombim.

PAGE 99

88 Table 4-12. Summary of statistics for banana fallows (capoeira bananal) N Mean Min. Max. S Median Age of capoeira bananal 16 3.88 2 18 4.00 3 Num. of species 21 3.57 1 12 2.87 2.5 Num. of banana varieties 17 2 1 4 1.06 2 Table 4-13. The 30 useful species found in 21 banana fallows (capoeira bananal) Num. capoeira bananal (21) Local name English name Use % Banana Banana 1 20 91 Mulateiro Mulato wood 3 6 27 Aa Aa 2 2 5 23 Mamo Papaya 1 5 23 Tapereb Yellow mombim 1 5 23 Cuia Calabash gourd 4 3 14 Jenipapo Genipap 1 3 14 Goiaba Guava 1 2 9 Manga Mango 1 2 9 Muiratinga Muiratinga 3 2 9 Samama Kapok tree 3 2 9 Alfavaca Alfavaca 9 1 5 Assac Assac 3 1 5 Bacaba Bacaba 2 1 5 Bacuri coroa Bacuri coroa 1 1 5 Cacau Cocoa 1 1 5 Caj Cashew 1 1 5 Capim santo Lemon grass 9 1 5 Cubiu Cubiu 1 1 5 Goiaba araa Goiaba araa 1 1 5 Graviola Soursop 1 1 5 Ing Inga 1 1 5 Jacareba Jacareba 3 1 5 Jambo Malay apple 3 1 5 Jauari Jauari 5 1 5 Jit Jit 3 1 5 Lauro abacate Lauro abacate 3 1 5 Limo comum Lime 1 1 5 Majirio Majirio 9 1 5 Mandioca Manioc 8 1 5 Use categories: 1 Fruit, 2 Beverage, 3 Timber, 4 Bowl, 5 Fishbait, 6 Thatch, 7 Nut, 8 Tuber, 9 Medicine There are two possible species of aa; aa do Par (E. oleracea), and aa preto (E. precatoria).

PAGE 100

89 Table 4-14. Banana varieties in banana fallows (capoeira bananal) Variety Num. banana fallows (17) % Prata 16 94 Comprida 9 53 Ma 5 29 Guariba 2 12 Peruana 2 12 Figure 4-1. Banana fallow on a high levee on the Solimes River. Banana, aai and several other fruits are growing in partial shade in an area about ten meters wide extending 200 meters. Cuia, a planted and carefully tended shrub whose giant gourd fruits are used as bowls, cups, and ladles was found in three of 21 capoeira bananais. Of course, a capoeira bananal should have bananas. It is the presence and density of bananas that determine the areas classification. They occur with up to four varieties in all but one of the 21 surveyed with an average of two. Of five varieties, banana prata is the most prolific, occurring in 94% of the fallows surveyed, Comprida (plantain) was

PAGE 101

90 found in half, and Ma, Guariba, and Peruana were found in a few (Table 4-14). The single capoeira bananal with no bananas had been planted with bananas until the previous years flood killed them all. Capoeira Cacaual The designation of cocoa fallow (capoeira cacaual) is relatively rare among land use types on the vrzea. Only two areas of this type were encountered. One of the sites was apparently an old cocoa plantation that had been planted about 20 years ago. It was reported to be a very large area, about 2000 meters long and 500 meters wide, and the only planted species it contains is cocoa. No information on additional species, management, or number of individuals was collected but there is almost certainly at least some other useful species in such a large area. The other site is an eight year old cocoa fallow that was cleared from mature forest 10 years ago and planted to manioc and banana for two consecutive seasons before being planted to cocoa and left fallow. The area is about 100 by 40 meters and cocoa is the only useful species. Capoeira Fruteira A fruit fallow (capoeira fruteira) is created when a farmer decides to invest a significant amount of energy in forming a swidden fallow replete with fruits and other useful species. These areas are typically located conveniently close to the village to facilitate care and protection of crops and easy transport of bulky fruits back to the home. The development of the fruit fallow begins in the roa phase. Fruit trees are planted among the manioc crop offering excellent protection for the developing seedlings by providing a moist, shady environment. Seedlings may be planted during the first or second crop of manioc or at any time throughout the life of the capoeira. During the fallow phase, optimal conditions are maintained for the production of fruits by managing

PAGE 102

91 secondary forest trees to provide the right amount of sun or shade for each crop. By manipulating the species composition of the regenerating forest in favor of valuable subsistence and market crops, a farmer creates a fallow rich in fruits and other useful plants and increases land-use efficiency by significantly boosting production over the fallow cycle. The agroforestry system thus installed acts the same as the naturally regenerating forest in functioning to restore the fertility of the soil for a future cropping sequence. Table 4-15. Summary of statistics for fruit fallows (capoeira fruteiras) N Mean Min. Max. S Median Age of capoeira fruteira 9 7.78 1 30 8.66 5 Num. of species 12 10.17 5 19 4.28 9.5 Num. of banana varieties 6 2.17 1 3 0.98 2.5 Table 4-16. Banana varieties in six fruit fallows (capoeira fruteiras) Variety Num. capoeiras fruteiras (6) % Prata 6 100 Comprida 2 33 Ma 2 33 Peruana 2 33 Guariba 1 17 If the density of valuable fruits reaches a high level, the area may stay in fallow indefininately. A small sample revealed a mean age of eight years and a median of five years with one fallow 30 years old (Table 4-15). The farmer may continue to enrich the fruit fallow over time with the addition of seedlings, vegetables, and medicinal plants and trotection of naturally occuring plants. Reaching a certain density of valuable plants, a famer may begin to refer to the fruit fallow as a sitio. When trees are old, a farmer can decide to clear the land (or certain portions of it) again for manioc production. Also, a high flood may kill many of the trees making roa production advantageous once again.

PAGE 103

92 Table 4-17. The 29 useful species most frequently encountered in 12 fruit fallows Local name English name Use Num. capoeira fruteiras (12) % Banana Banana 1 8 67 Aa Aa 2 2 6 50 Fruta po Breadfruit 1 6 50 Tapereb Yellow mombim 1 6 50 Cacau Cocoa 1 5 42 Jenipapo Genipap 1 5 42 Cuia Calabash gourd 4 4 33 Cupuau Cupuau 1 4 33 Goiaba Guava 1 4 33 Inga Ing 1 4 33 Limo comum Lime 1 4 33 Mamo Papaya 1 4 33 Manga Mango 1 4 33 Araa boi Araa boi 1 3 25 Buriti Buriti palm 2 3 25 Caj Cashew 1 3 25 Laranja Orange 1 3 25 Muiratinga Muiratinga 3 3 25 Mulateiro Mulato wood 3 3 25 Pupunha Peach palm 1 3 25 Seringa Rubber 10 3 25 Abiu Abiu 1 2 17 Bacuri liso Bacuri liso 1 2 17 Cacau jacar Cacau jacar 1 2 17 Jambo Malay apple 1 2 17 Lima Lima 1 2 17 Limo caiana Limo caiana 1 2 17 Ucuba Virola 3 2 17 Urucu Annatto 11 2 17 Use categories: 1 Fruit, 2 Beverage, 3 Timber, 4 Bowl, 5 Fishbait, 6 Thatch, 7 Nut, 8 Tuber, 9 Medicine, 10 Latex, 11 Colorant There are two possible species of aa; aa do Par (E. oleracea), and aa preto (E. precatoria). Twelve fruit fallows were encountered in 6 communities. Five of these belong to the residents of Vila Alencar. The encouragement of researchers from the Mamiraua project, who contributed technical support and seedlings of fruit trees, resulted in several of these areas especially rich in fruits (Table 4-17). Banana was most frequent with up to three varieties in eight of twelve (67%) capoeira fruteiras and a total of five varieties

PAGE 104

93 present (Table 4-16). Approximately twenty out of the 29 most frequently observed species in the fruit fallows were planted. Others such as papaya and ing may be planted or protected. Frutal A fruit orchard (frutal) can be thought of as a well developed fruit fallow with a relatively dense planting of fruit trees and a high diversity of species. Two areas designated as fruit orchards were visited on two long high levees behind the community of Sitio Forteleza. Both are owned by the same family. One is approximately 300 meters long and the other is 2000 meters long. The width varies from 20 to 40 meters. The first is three years old. It was previously forest (mata) with many valuable species. Thirteen useful species were inventoried in this fruit orchard (Table 4-18). They spared all trees with a use. These include aa preto used for making a beverage, the bacuri coroa (Rheedia sp.) fruit, and the timber trees envira vassourinha (Xylopia calophyllum), jacareba (Calophyllum brasiliense), manixi (Brosimopsis oblongifolia), marirana (Coepia sp.), virola (Virola surinamensis) (also a good fish bait), and others. The method used here is to clean underneath the trees and plant fruits. Some of the fruits that were planted include ara boi (Eugenia stipitata), bacuri liso (Rheedia macrophylla), breadfruit, cocoa, cupuau, ing, Malay apple (Eugenia malaccensis), and mango (Mangifera indica). At the time of my visit they were clearing undergrowth to plant bananas and had already planted some. They will not plant roa here. The second, much larger fruit orchard begins directly behind the community at the back of the home gardens and stretches along a high levee beside the natural canal (paran) for approximately 2000 meters. It is several years old and contains 19 useful species (age was not obtained). The villages two small herds of cattle and water buffalo

PAGE 105

94 have free reign in this area. The owner, Francisco, claims that roa has never been planted here. A number of banana and orange trees were lost in last years flood. Table 4-18. The 25 useful species found in two fruit orchards (frutal) Local name English name Use Num. frutals (2) Aa* Aa 2 2 2 Ara boi Ara boi 1 2 Bacuri coroa Bacuri coroa 1 2 Cacau Cocoa 1 2 Limo comum Lime 1 2 Samama Kapok tree 3 2 Ucuba Virola 3 2 Andiroba Andiroba 3, 9 1 Bacuri liso Bacuri liso 1 1 Cedro Cedar 3 1 Cupuau Cupuau 1 1 Envira vassourinha Envira vassourinha 3 1 Fruta po Breadfruit 1 1 Ing Inga 1 1 Jacareba Jacareba 3 1 Jambo Malay apple 1 1 Jit Jit 3 1 Manixi Manixi 1 Manga Mango 1 1 Marirana Marirana 3 1 Muiratinga Muiratinga 3 1 Mulateiro Mulato wood 3 1 Puru Puru 1 1 Seringa Rubber 10 1 Tapereb Yellow mombim 1 1 Use categories: 1 Fruit, 2 Beverage, 3 Timber, 4 Bowl, 5 Fishbait, 6 Thatch, 7 Nut, 8 Tuber, 9 Medicine, 10 Latex, 11 Colorant There are two possible species of aa; aa do Par (E. oleracea), and aa preto (E. precatoria). Sitio When a farmer proudly proclaims a place my sitio it carries with it an intimate meaning of a site that receives special care and attention. It may even be a secluded retreat where the family can go to take a break from the tediousness of village life. A diverse orchard of mature fruit trees is the core vegetative component of the sitio. This

PAGE 106

95 cultural forest has been carved out of the surrounding mata bruta (old forest) through years of patient investment in time and energy. In addition to the fruit orchard, a small roa or maize patch, or a garden of vegetables and medicinal herbs is sometimes planted in an area with adequate sunlight. Other useful timber, medicinal, or fish bait trees may be protected or planted in the sitio. The designation of sitio implies something of permanence. As much as any location on the ever changing vrzea can be considered permanent, the sitio is quite possibly the longest lasting cultural feature of the floodplain forest. It can stay in the family, passed on from father to son. It is also one of the few areas with enough invested to give the land monetary value. The sitio may be bought and sold. The sitio can be formed in a number of ways. It sometimes begins as a roa that is continually enriched with valuable plants until it eventually becomes a mature, productive orchard. At some point there is a conscious decision by the farmer that this is where he will concentrate a portion of his energy; this will be his sitio. The energy invested in management can be reduced when the orchard is mature and forms a thick canopy that blocks out the sun. A sitio may also start out as a vacant residents home garden. When people change residences on the vrzea they often dismantle their homes to take advantage of the lumber and reassemble them in a new location. The former home garden may then be kept as a sitio, passed to a family member, or sold to a neighbor. The sitios visited on the vrzea had several different origins and management regimes. A brief description of four of them can provide a picture of the diversity of land use on the vrzea even in this single type. The first sitio, located in the community of

PAGE 107

96 Betel on a very high levee along the bank of the Japur River, started out as someones home garden. The current owner purchased it from the previous owner who took his house with him when he left. This impressive orchard had 17 species of fruit packed into an area about 10 meters wide and 300 meters long. There was a large number of cocoa, cupuau, citrus, and guava (Psidium guajava) trees, thriving among many other species. The owner earns some income from the sitio by selling cupuau and cocoa. I could hardly take notes as my guide repeatedly broke open cocoa fruits and shoved them into my hands, a welcome refreshment in the hot equatorial sun. There were bundles of manioc stored here that had started to take root and hundreds of small fruit tree seedlings scattered around that had sprouted on their own. This is an important source area of seedlings for enriching fields and fallows. Many of the seedlings drowned in the 1999 flood including avacado (Persea Americana), abiu (Pouteria caimito), banana, cupuau, lima (Citrus spp.), and peach palm (Bactris gasipaes). Some seedlings of cupuau and pupunha survived. These are two terra firme species that normally do not resist flooding very well. By continuously selecting more resistant seedlings for replanting, varieties might evolve over time that are better adapted to conditions on the vrzea. The second sitio is located along a side channel of the Japur. The owner also purchased this area from another owner. He paid $RS400 or approximately US$235 for this 120 by 40 meter site. It now has 11 species after three were lost during last years flood. All of the abiu, avacado, and banana were lost along with some orange trees. A third sitio in the community of So Paulo has a manioc processing house (casa de farinha) on the site. It sits on a high levee that did not flood in the year 2000. They have

PAGE 108

97 started some watermelon seedlings to plant on the mudflat in front as soon as the flood recedes. Besides the twenty odd species of fruits, vegetables, and medicinal plants they have planted, several valuable trees have been spared during forest clearance including castanha de macaco (Couroupita guianensis), macacarecuia (Couroupita sp.), muiratinga, and the kapok tree. Interestingly, they also planted cator (Crataeva benthami), a favorite fruit of the prized tambaqui fish (Colossoma macropomum)(Araujo-Lima and Goulding 1997). There is a harvested manioc field at the site located conveniently close to the manioc processing house. At a forth site, named Sitio Isaias near the village of Sitio Forteleza, the family constructed a small house, not a manioc processing house, but an elevated shelter from the midday sun or a weekend retreat. The family told me they like to come here occasionally and stay for a few days just to get away from the village. It is a place of rest and relaxation where they enjoy roasting fish or harvesting and eating any fruit that is in season. They can keep busy working in the nearby fields or in one of two small vegetable and herb gardens at the site. When I visited they were making the delicious beverage from the aa palm. Often mixed with manioc flour or tapioca, it is a hearty snack that is greatly revered by people of the region and provides a welcome break in the days activities when friends and neighbors are invited to share the refreshing treat. This sitio is about 20 years old and was developed from a former cocoa grove (cacaual). There is a banana fallow on one side that extends for approximately 100 meters. The large size (8000 m2) and heterogeneity of the site with the addition of two small vegetable and herb gardens gives this sitio the greatest number of useful plants of any work area that was visited with 34 species.

PAGE 109

98 The five sitios that were encountered in the survey contained an average of 20.8 species. This is a significant jump in useful plant diversity over the other types of land use, more comparable with home gardens than fields or fallows. Actually, the diversity is higher than the average vrzea home garden probably due to the greater size of the sitio and because some homeowners do not take an interest in developing home gardens. The average sitio is more than four times larger than the average home garden. The sitio with the lowest number of species held 11 while that with the highest held 34. A total of 61 species were logged in the five sitios. Table 4-19 shows all species that were found in at least two sitios. The most common were cocoa and Malay apple, occurring in all five sitios, and aa occurring in four. Fruit trees account for most of the diversity in sitios. Out of the 24 most common species, 13 are fruits, three are edible palms used for making beverages (aa, bacaba (Oenocarpus bacaba), and buriti (Mauritia flexuosa)), three have medicinal uses (andiroba (Carapa guianensis), boldo (Pneumus boldus), guava), two are timber trees (andiroba, kapok), one is a latex (rubber (Hevea brasiliense)), one is a vegetable (tomato (Lycopersicum esculentum)), one is a tuber (manioc), and one is used as a bowl (calabash gourd (Crescentia cujete)). Other species that were found in only one of the five sitios also include fruit, nuts, timber, condiments, and medicinal plants.

PAGE 110

99 Table 4-19. The twenty-four most frequently encountered species and their use in five sitios Local name English name Use Num. sitios (5) % Cacau Cocoa 1 5 100 Jambo Malay apple 1 5 100 Aa Aa 2 2 4 80 Bacaba Bacaba 2 3 60 Banana Banana 1 3 60 Buriti Buriti palm 1, 2 3 60 Cuia Calabash gourd 4 3 60 Cupuau Cupuau 1 3 60 Fruta po Breadfruit 1 3 60 Laranja Orange 1 3 60 Lima Lima 1 3 60 Limo comum Lime 1 3 60 Manga Mango 1 3 60 Puru Puru 1 3 60 Andiroba Andiroba 3, 9 2 40 Ara goiaba Ara goiaba 1 2 40 Boldo Boldo 9 2 40 Goiaba Guava 1 2 40 Mamo Papaya 1 2 40 Mandioca Manioc 8 2 40 Pupunha Peach palm 1 2 40 Samama Kapok tree 3 2 40 Seringa Rubber 10 2 40 Tomate Tomato 11 2 40 Use categories: 1 Fruit, 2 Beverage, 3 Timber, 4 Bowl, 5 Fish bait, 6 Thatch, 7 Nut, 8 Tuber, 9 Medicine, 10 latex, 11 vegetable There are two possible species of aa; aa do Par (E. oleracea), and aa preto (E. precatoria).

PAGE 111

CHAPTER 5 DIVERSITY IN HOME GARDENS An examination of agrobiodiversity in agricultural ecosystems of Amazonia would be sorely deficient if it did not include a serious look at home gardens. These resource islands packed with useful plants are found almost ubiquitously around homes in the region. Padoch and Jong (1991) emphasized that most studies of home gardens had focused on world regions with high population density and Amazonia, with its extensive available land, did not seem a likely area to find this complex and intensively managed system. Their work in the Peruvian community of Santa Rosa, however, illustrated that home gardens of the Amazon basin contain a high level of diversity even among the non-tribal riverine population whom, it was sometimes felt, had lost the intimate plant knowledge of tribal Amerindians. Smith (1996b) points out that even recently arrived colonists are quick to develop diverse backyard collections of plants. Recent interest in tropical home gardens has been driven by the enormous diversity of plants they contain, their importance for family subsistence and household income, and their promise for agroforestry. Smith and colleagues stressed the value of home gardens as a largely untapped resource for agroforestry development in the Amazon and their role in crop domestication (Smith 1996b, 1999; Smith et al. 1995, 1996). They note that several crops such as Barbados cherry (Malpighia glabra) and cupuau, previously grown only in the backyard for subsistence, are now successful commercial crops. Chota (1999) describes the rise of the native backyard tree, umar (Poraqueiba cericeae) to principal crop in the village of Tamshiyacu, 30 kilometers from the city of Iquitos, Peru. 100

PAGE 112

101 Varieties of the tree were carefully selected from home garden stock to plant on a commercial scale. Today there are approximately 4,000 hectares of umar growing in Tamshiyacu. Padoch et al. (1985) found that more than 60% of the $5,000 average annual family income comes from the sale of umar. There is clearly an enormous potential for developing new crops from the immense diversity of plants grown in home gardens. Home gardens are an important venue for plant domestication. Plants are sometimes recruited from the forest and carefully tended for their medicinal, ornamental, or food producing qualities. Pre-existing and spontaneous plants are commonly protected. Many exotic species from the African and Asian tropics are also introduced. Once established in the home garden, the process of farmer selection and propagation can begin. Seedlings were observed growing in home gardens on many occasions in the study area and elsewhere in the Amazon. Seeds from especially good quality fruits are saved and planted. They may be found growing in planters or seedbeds with vegetables and medicinal plants. Spontaneous volunteers are found at the base of mature home garden trees where they have sprouted from fallen seeds. There is often an abundance of young plants that provide farmers with a large selection for replanting in other areas. Farmers select especially vigorous specimens for replanting. A natural selection process also takes place in home gardens as more resistant seedlings survive flooding while others perish. Seedlings are stored in the home garden where they can be closely watched until they are transferred to a roa or capoeira, usually at the beginning of the rainy season.

PAGE 113

102 Tropical home gardens are noted to have an intimate relation with the family and residents affirm that the home is not complete without one (Jose and Shanmugaratnam 1993). They are often among the most well tended components of the agricultural system. Plants are kept nearby where they can be easily looked after and protected. Many receive special attention that includes watering, soil amendments, protection from animals, and the use of planters or pots. Charcoal or other organic materials are commonly mixed with soil to enhance fertility. Especially rich soil is selected for use in planters and pots. A section of the garden used for growing vegetables and medicinal plants is often protected from chickens and other animals with wire or netting. Home gardens furnish shade and protection from the wind for people, domestic livestock, and pets. Activities such as the processing of manioc and other foods and the production of charcoal are often performed in the comfort of the home garden. Children especially benefit by the diverse foods grown in the home garden and by the knowledge developed from close proximity and observation of the many species and varieties of plants. Whenever there is something to be had, kids climb the trees or reach up with long poles to dislodge fruits, ripe or unripe. Home gardens often make a large contribution to family subsistence by providing many foods in season at various times throughout the year or at critical times when other sources of food are scarce, notably during the flood season. Home gardens are important areas for leisure as well as meeting places for neighbors (Lima 1994). The shaded area provides welcome relief from the hot afternoon sun and wooden benches are sometimes placed to enjoy the view of the river and passing boats. A sizeable area around the house is often kept weeded and swept clean. The usual

PAGE 114

103 method of dealing with grass or weeds is to remove them with a hoe or machete, scraping the ground smooth and clean. Frequent sweeping with a broom made of palm fronds keeps the ground free of forest litter which is then swept into piles and burned. In a study on Careiro Island near Manaus, Lima (1994) found that the fires can be intentionally used to disperse pests from trees. Other areas are kept in grass and leaves may be left in place to provide mulch, conserving moisture and adding organic material to the soil. Many yards on the vrzea are kept in grass and several communities in the study area had cattle, buffalo, or goats that kept it nicely trimmed. Figure 5-1. Yard in the vrzea located on a paran connecting the Japur River to the Solimes. On the floodplain, home garden plants are often kept principally in the area behind the house while the front of the home is kept in grass. Cattle affect the way that plants can be grown in the village. They often destroy young trees and other plants that are left unprotected. The levee here was about one meter above the high water mark for 2000. Community of So Paulo do Corai.

PAGE 115

104 Diversity in Middle Solimes Vrzea Home Gardens Contrary to a popular perception that the floodplain is not amenable to growing perennial crops, recent research has shown that home gardens on the Amazon vrzea contain levels of diversity comparable to those on terra firme (uplands) (Lima 1994; Lima and Saragoussi 2000; Chota 1999; Smith 1996b). Smith directly compared home gardens on the uplands with those on the middle Amazon vrzea and found a greater number of perennial species in a smaller sample of floodplain gardens. In the middle Solimes vrzea, an area with the highest range in river level from low to high water for the entire length of the Brazilian Amazon (sometimes more than 16 meters), it appears that in terms of average number of species, floodplain home gardens are somewhat less diverse than upland gardens. A total of 176 useful species (excluding ornamentals) were recorded in 163 home gardens on the vrzea with an average of 12 species per household. This compares to 135 useful species found in 30 terra firme home gardens in the same region with an average of 21 plant species per garden. There were 99 species common to both the vrzea and terra firme out of 212 total home garden species. Average diversity is lower in home gardens on the floodplain because of environmental variation within communities and the difficulties of planting perennials in areas subject to annual flooding. Differences in socio-economic situation (more fishermen with less interest in planting?) or higher mobility of residents could also be factors that lead to less diverse home gardens on the floodplain. Although vrzea home gardens are slightly less diverse than their counterparts on terra firme, a large number of useful species on the vrzea are found exclusively in home gardens illustrating their value as a repository of agrobiodiversity. Out of 204 total species recorded on the vrzea, 87 were found exclusively in home gardens. By

PAGE 116

105 comparison, there were just 29 species not found in home gardens. This indicates that almost the complete repertoire of useful plants in the agroecosystem of the middle Solimes is represented to some extent in home gardens, an important fact to remember when considering the conservation of agrobiodiversity in the region. The wide range of uses for home garden plants underscores its importance for family subsistence and well-being. Approximately 20 different uses were reported for plants in the middle Solimes. The use was recorded only when reported by residents or in the scientific literature. Home gardens were inventoried in 14 vrzea communities. The average number of homes in each village is 12. Of the 171 homes visited, only 8 did not keep any useful plants. The most common use is fruit (including nut) or beverage with 62 species (44%). All except 16 vrzea gardens had at least one fruit tree growing. Home gardens contained an average of 4.8 species and 14.6 individual trees in the fruit and beverage category (Table 5-1). Medicinal is the second largest use category with 39 species (28%) and timber is third with 19 species (14%). At least 19 species have more than one use. Other uses are vegetable (11 species), fishbait (6 species), condiment (4 species), calking (3 species), latex (2 species), livestock feed (2 species), snack (2 species, includes chocolate and sugarcane), tuber (2 species), and 1 species each used for a bowl, dye, fiber, food colorant, grain, and mulch. Firewood is also gathered from the home garden. The number of useful species in floodplain home gardens varies widely. Figure 5-2, shows the frequency that a given number of species occurred. The number of useful species ranges from one to 29 and the average vrzea home garden contains 12 species.

PAGE 117

106 0246810121357911131517192123252729Number of SpeciesNumber Home Gardens Figure 5-2. The number of species in 163 home gardens on the vrzea 0123414710131619222528313437404346Number of SpeciesNumber Home Gardens Figure 5-3. The number of species in 30 home gardens on terra firme There are several possible reasons for homes that do not have gardens or have only a few species. Like fields, many villages lie on uneven ground. Some residents have low-lying yards and strong currents can kill all except the most adapted of floodplain plants. Home gardens within communities usually lie adjacent to one another but it is not unusual for groups of homes to be separated by larger distances when divided by low lying areas. A home garden is very personalized and each member of the family can contribute to its development. Exceptionally diverse gardens are usually managed by avid plant collectors. Padoch and Jong (1991) studied upland home gardens in the village of Santa Rosa in the Peruvian Amazon and found that plant assemblages varied according to the special needs and interests of families. Some individuals do not have much interest in

PAGE 118

107 raising home gardens. This is also true in the middle Solimes region. Several homeowners claimed they were primarily fisherman and not interested in agriculture. In the Mamirau reserve, a small proportion of families in some villages live on floating houses. They have no yard (quintal) but frequently care for several types of medicinal plants or vegetables in planter boxes on the side of the house or on separate floating rafts. New homes had few plants and young couples or single men can often rely on mature home gardens of their family, friends and neighbors. In a large family there are more individuals to contribute to the maintenance of the garden and long-term residents have had more time to acquire diverse collections of plants. However, there is no proven correlation between homeowner age and species richness in home gardens (Padoch and Jong 1991). People with undeveloped gardens usually expressed their desire of planting more trees but many also explained their frustration with raising perennials on the vrzea. Some considered it a waste of time. Seedlings can be less tolerant to flooding than mature trees and are often killed during the first few years after planting. Some of the gardens visited had fruit seedlings underwater and it was unknown which ones would come out of the flood alive. A number of residents reported that older trees, ordinarily able to survive a few feet of water, can be killed from the impact of a floating log or canoe. This effect was also reported by floodplain farmers on Careiro Island near Manaus (Lima 1994). A process of natural selection is likely operating here if, frequently, only the most tolerant individuals survive. The assumption follows that vrzea home gardens would be promising places to search for crop plants that are more tolerant to flooding for the expansion of agroforestry on the floodplain.

PAGE 119

108 A large variation in home garden diversity was found among communities, ranging from 7.4 to 17.3 average species. Table 5-1 shows the average number of species broken down by community. Minimum numbers of species by community range from 1 to 12 and maximums range from 17 to 29. The variation may be due in a large part to differences between communities in terms of impact from flooding (depth and speed of current), socio-economic situation, age of households, and personal preference. The two communities with the lowest average number of species are Barroso, a low-lying community on the Solimes, and Jarau, another village with a relatively greater impact from flooding where residents have developed a fishing cooperative. The four communities with the highest average diversity, Betel, Manacabi, Marirana, and Porto Praia are all on relatively higher levees and were completely dry at the time of the survey. More research is needed to learn how the location, environment, economy, and social factors influence diversity in home gardens. Table 5-1. Average number of useful species in 163 vrzea home gardens by community Community N* Mean Min. Max. S Median All communities 163 12.0 1 29 6.3 11 Betel 11 17.3 12 27 5.1 17 Manacabi 8 16.9 12 26 4.2 17 Marirana 6 16.5 8 24 5.7 17 Porto Praia 16 14.9 3 24 6.7 13 So Francisco do Aiuc 17 13.0 4 27 6.1 12 Betania 15 12.3 6 24 5.5 12 Bela Vista do Manguary 2 12.0 5 19 So Paulo do Corai 12 11.5 3 29 7.0 10.5 So Joo 11 11.3 3 22 6.1 9 Sitio Fortaleza 11 10.3 2 17 4.4 11 Vila Alencar 21 10.1 3 23 5.6 9 Pentecostal 10 9.4 6 17 3.4 9 Barroso 11 8.4 2 19 5.7 9 Jarau 12 7.4 1 21 6.4 6.5 *Homes without a home garden were omitted (8 homes surveyed had no useful species).

PAGE 120

Table 5-2. Fruit, nut, and beverage species in vrzea home gardens Vrzea home gardens Terra firme home gardens Local name Num. of indiv. Num. of home gardens (n=163) Ave. indiv. per garden observed Num. of indiv. Num. of home gardens (n=30) Ave. indiv. per garden observed Goiaba Guava 566 107 5.3 91 21 4.3 Cacau Cocoa 261 52 5.0 32 17 1.9 Banana Banana 257 56 4.6 73 21 3.5 Aa 1 Aa 1 186 46 4.0 120 20 6.0 Cana de aucar Sugarcane 162 29 5.6 5 4 1.3 Ing au Ing au 118 63 1.9 15 6 2.5 Graviola Soursop 80 46 1.7 11 9 1.2 Camu camu Camu camu 71 28 2.5 4 1 4.0 Bacaba Bacaba 69 24 2.9 38 17 2.2 Cubiu Cubiu 68 29 2.3 13 5 2.6 Mamo Papaya 62 17 3.6 60 18 3.3 Ing cip 2 Ing cip 2 60 29 2.1 31 11 2.8 Limo comum Lime 55 28 2.0 19 8 2.4 Cco Coconut 45 33 1.4 18 12 1.5 Caj Cashew 45 23 2.0 28 13 2.2 Manga Mango 41 23 1.8 56 16 3.5 Tapereb Yellow mombim 40 27 1.5 4 3 1.3 Jambo Malay apple 40 23 1.7 26 17 1.5 Ara goiaba Ara goiaba 31 20 1.6 4 3 1.3 Buriti Buriti 30 19 1.6 17 11 1.5 Castanha sapucaia Castanha sapucaia 30 11 2.7 0 0 0 Cupuau Cupuau 28 11 2.5 54 16 3.4 Bacuri liso Bacuri liso 25 14 1.8 0 0 0 Jenipapo Genipap 22 16 1.4 4 3 1.3 109

PAGE 121

Table 5-2. Continued Vrzea home gardens Terra firme home gardens Local name Num. of indiv. Num. of home gardens (n=163) Ave. indiv. per garden observed Num. of indiv. Num. of home gardens (n=30) Ave. indiv. per garden observed Puru Puru 21 16 1.3 5 3 1.7 Maracuj Passionfruit 20 17 1.2 8 6 1.3 Lima Lima 20 13 1.5 8 5 1.6 Ing Inga 20 12 1.7 5 1 5.0 Azeitona Olive 17 11 1.5 2 2 1.0 Marirana Marirana 17 10 1.7 9 6 1.5 Pupunha Peachpalm 15 10 1.5 41 15 2.7 Laranja Orange 14 10 1.4 54 8 6.8 Limo tangerina 3 Limo tangerina 3 13 10 1.3 12 6 2.0 Frutapo Breadfruit 13 7 1.9 0 0 0 Bacuriliso Bacuriliso 12 9 1.3 0 0 0 Mutamba Mutamba 12 8 1.5 0 0 0 Goiaba ara Goiaba ara 11 8 1.4 0 0 0 Abiu Abiu 10 6 1.7 35 9 3.9 Castanha de macaco Castanha de macaco 9 6 1.5 0 0 0 Castanholeira Castanholeira 8 8 1.0 0 0 0 Bacuri coroa Bacuricoroa 7 5 1.4 2 2 1.0 Abacaxi Pineapple 7 2 3.5 31 6 5.2 Araboi Araboi 5 4 1.3 16 4 4.0 Cacau jacar Cacau jacar 5 4 1.3 3 3 1.0 Cacaurana Cacaurana 4 3 1.3 1 1 1.0 Pitomba Pitomba 3 2 1.5 2 1 2.0 Ara do igapo Ara do igapo 2 2 1.0 0 0 0 110

PAGE 122

Table 5-2. Continued Vrzea home gardens Terra firme home gardens Local name Num. of indiv. Num. of home gardens (n=163) Ave. indiv. per garden observed Num. of indiv. Num. of home gardens (n=30) Ave. indiv. per garden observed Carambola Carambola 2 2 1.0 1 1 1.0 Ingigap Ingigap 2 2 1.0 0 0 0 Ingsapo Ingsapo 2 2 1.0 0 0 0 Ingsinho Ingsinho 2 2 1.0 0 0 0 Marimari Marimari 2 2 1.0 7 3 2.3 Abacate Avocado 1 1 1.0 31 9 3.4 Aerola Barbados cherry 1 1 1.0 0 0 0 Apu Apu 1 1 1.0 0 0 0 Bacabinha Bacabinha 1 1 1.0 3 2 1.5 Birib Birib 1 1 1.0 8 5 1.6 Cajurana Cajurana 1 1 1.0 0 0 0 Ingcurta Ingcurta 1 1 1.0 0 0 0 Limocida Limocida 1 1 1.0 0 0 0 Tangerina Tangerina 1 1 1.0 6 4 1.5 111 Note : The frequency (number of home gardens) and abundance (number of individuals) of 61 fruit, nut, or beverage species in 171 home gardens on the vrzea and their corresponding number in terra firme home gardens. (See appendix 2 for the complete list of home garden species with their scientific name, frequency, and use) 1 There are two possible species of aa; aa do Par (E. oleracea), and aa preto (E. precatoria). 2 also called ing comprida 3 also called limo de suco

PAGE 123

112 Counting the number of individuals of each species gives an accurate picture of crop abundance (Table 5-2). There appears to be some positive correlation (r=0.7) between the frequency a species is encountered in home gardens and its abundance (the number of individuals). The 5 most frequently encountered perennials in the fruit, nut or beverage category (guava, cocoa, banana, aai, and sugarcane) were also the most abundant with 4 to 5 individuals per garden. The list of fruit, nut, or beverage species in table 5-2 is a mix of native and exotic species. The top ten most abundantly occurring species include two exotics from the Asian tropics, banana and sugarcane. Of the 61 species listed, 48 are native to tropical America and 13 come from tropical Asia. The occurrence of species in home gardens is clearly influenced by economic and environmental reasons as well as personal preferences. Guava, occurring in 66% of home gardens, is tolerant to flooding and is a favorite, vitamin rich, snack fruit and also medicinal plant. Although guava is a relatively unimportant commercial crop in the area, it is both more than twice as frequent and abundant as cocoa, the second most abundant crop. Cocoa is commercially valuable, flood tolerant, and, besides being made into chocolate, is a delicious fruit or juice. Bananas, the third most abundant fruit, are found in 34% of vrzea home gardens. Table 5-3 shows that there is significant variation among communities in the extent that bananas are grown in home gardens. Bananas are a refreshing treat, a frequent food at mealtime, and are also readily sold. In addition, bananas are somewhat tolerant to flooding. Many banana plants can survive one to three months in saturated ground. They usually produce before being drowned by the flood and it is not uncommon for them to

PAGE 124

113 be harvested by canoe in several feet of water. Suckers can usually survive longer when immersed and will continue growing the following season. Table 5-3. Number of banana varieties in vrzea home gardens Community N Gardens with banana present % gardens with banana Average num. of varieties* Average num. of individuals** All Communities 163 43 26 1.2 6.0 Betel 11 5 45 1.2 4.2 Manacabi 8 4 50 1.5 8.3 Marirana 6 6 100 1.0 10.2 Porto Praia 16 3 19 1.0 2.0 So Francisco do Aiuc 17 2 12 2.0 3.0 Betania 15 4 27 1.0 3.0 Bela Vista do Manguary 2 0 0 So Paulo 12 0 0 So Joo 11 3 27 1.0 6.3 Sitio Fortaleza 11 0 0 Vila Alencar 22 3 14 1.0 3.0 Pentecostal 10 7 70 1.9 7.6 Barroso 11 6 55 1.2 6.2 Jarau 12 0 0 *Average calculated only for home gardens with banana present. **Average number of torceiras calculated only for home gardens with banana present. Table 5-4. Banana varieties in vrzea home gardens Variety Num. individuals Num. home gardens Avg. num. of individuals per home garden observed Comprida 107 23 4.7 Prata 107 19 5.6 Urucuri (sapo) 14 5 2.8 Peruana 9 1 9.0 Paco vo 7 2 3.5 Japur 3 1 3.0 Bai 2 1 2.0 Ma 2 2 1.0 Inaj 2 1 2.0 The nine varieties of banana reported in vrzea home gardens are given in table 5-4. Two of these dominated home gardens on the floodplain. The Comprida variety, a plantain, and Prata, a common commercial variety, were both found in about 20 gardens

PAGE 125

114 with a total of 107 individuals each. Urucuri, also known as sapo or frog, came in third with 14 individuals in 5 home gardens. The other six varieties were found in only one or two home gardens each with very few individuals. Home gardens on the vrzea should also be a promising place to search for new crop germplasm to be used in the development of varieties with greater resistance to flooding. Losses from Flooding Losses in home gardens resulting from high floods primarily consist of species that are intolerant to flooding and seedlings (Table 5-5). Species that are adapted to terra firme such as cashew, cupuau, and peach palm are high on the list of drownings. Table 5-5. Home garden plants that were killed during the 1999 high flood on the vrzea Local name English name Num. gardens Local name English name Num. gardens Caj Caj 10 Pimento Green pepper 2 Limo comum Lime 9 Urucu Annatto 2 Abiu Abiu 6 Alfavaca Alfavaca 1 Cupuau Cupuau 6 Ara boi Ara boi 1 Jambo Malay apple 6 Araticum Araticum 1 Laranja Orange 6 Azeitona Olive 1 Pimenta Capsicum pepper 6 Bacaba Bacaba 1 Aai Aai 5 Bau bau Bau bau 1 Ing Inga 4 Cacau Cocoa 1 Manga Mango 4 Cubiu Cubiu 1 Pupunha Peach palm 4 Goiaba Guava 1 Abacate Avocado 3 Limo caiana Limo caiana 1 Graviola Soursop 3 Limo de suco Limo de suco 1 Maracuj Passion fruit 3 Mastruez Mastruez 1 Tomate Tomato 3 Namui Namui 1 Banana Banana 2 Pimenta doce Pimenta doce 1 Carambola Carambola 2 Seringa Rubber 1 Note: not all families were interviewed about losses in home gardens

PAGE 126

115 Fruits from the Asian tropics including citrus, Malay apple, and mango are also frequently lost when high floods occur. Species that are usually flood tolerant such as aai and ing are easily drowned when young seedlings are completely immersed in water for extended periods during high floods. Size of Home Gardens The size of home gardens on the floodplain range from a mere 50 square meters to greater than half a hectare (Table 5-6). The average size is about 900 square meters. When broken down by community, the average size varies widely. The village Pentecostal had the smallest average size. Its gardens are consistently small with an average size for the community of only 261 square meters. On the other end of the spectrum lies Marirana with gardens ranging in size from 800 to 6000 square meters and an average of over 2800. There is slight positive correlation between size and number of species in vrzea home gardens (r=0.31). The scatter diagram in figure 5-4 shows number of species in 163 vrzea home gardens plotted against the size given in square meters. It is obvious that a larger home garden does not necessarily mean more species. In Inumas (1999) study of home gardens in the three villages of Vila Alencar, Jarau, and Barroso, it was found that the size, as well as the sale of products, is inversely proportional to the distance from markets. It was also observed that larger home gardens held greater numbers of livestock. The average size of terra firme home gardens in table 5-7 is only slightly greater than those on the vrzea at 971 square meters. Nossa Senhora da Fatima, a community established on the high bluff of the Solimoes river, has the largest gardens and the greatest average species diversity. The average size of home gardens in the relatively

PAGE 127

116 Table 5-6. Average size of 155 varze home gardens in 14 communities Community N Mean size (m 2 ) Min. Max. S Median All Communities 155 898 50 6400 962 600 Marirana 10 2833 800 6000 1586 2450 So Francisco do Aiuc 15 2047 600 6400 1530 1625 Bela Vista do Manguary 2 1500 1000 2000 Porto Praia 15 1242 225 3200 684 1250 So Joo 16 910 200 2400 794 500 Vila Alencar 19 815 250 2800 712 600 Sitio Fortaleza 11 759 200 1500 433 600 Jarau 11 572 50 1750 481 600 Barroso 11 550 100 1250 337 600 Betel 10 505 225 1000 249 425 Manacabi 7 489 225 750 220 500 Betania 16 470 105 2400 557 300 So Paulo do Corai 12 348 150 500 116 338 Pentecostal 9 261 100 450 125 225 Table 5-7. Average size of 30 terra firme home gardens in 3 communities Community N Mean size (m 2 ) Min. Max. S Median All Communities 30 971 25 5400 1367 300 Nossa Senhora da Ftima 8 2316 300 5400 1926 1800 Jubar 5 944 240 1600 580 1000 Boa Esperana 17 333 25 2500 584 150 0510152025303501000200030004000500060007000SizeNumber of Species Figure 5-4. Scatter diagram showing number of species in 163 vrzea home gardens in relation to size in square meters. There is a slight positive correlation between the size of home gardens and number of useful species (r=0.31).

PAGE 128

117 large upland community of Boa Esperanca is much lower. However, back yards in this village commonly grade into diverse fruit orchards or capoeira fruteiras. There is sometimes a fuzzy division between the end of the backyard and the start of a fruit fallow or orchard. In cases where the border of the home garden was not obvious, the division was drawn according to the opinion of the homeowner. Livestock in Home Gardens There is often an animal component to home gardens on the floodplain. Many families raise livestock and pets (Figure 5-5). According to Inuma (1999), the principal animal component in the three communities studied is poultry. Chickens (Gallus sp.) make up the largest part. They are also the home garden product that is most frequently sold. Ducks are less common because they are frequent victims of attacks by predators including piranhas which bite their feet while they swim. According to one family in Marirana that raises a higher than average number of ducks and chickens in a fenced in area of the yard, they lose a few animals per year to anacondas and caiman, the main predators. The family plants corn to feed the livestock. Chickens and other animals also forage on kitchen scraps, fallen fruits, insects, weeds and often roost in low home garden trees and bushes. Inuma found that in Vila Alencar, in 1999, three families were raising sheep and four were raising cattle. Floating platforms like the one pictured in figure 5-6 called marombas are constructed to shelter livestock during the flood and they are fed the floating canarana grass pictured in figure 5-7. They are sometimes transported to the nearby town of Alvares on the river bluffs to wait for the water to recede. Five out of the 14 villages surveyed had households that raised cattle. Some of the communities voted that cattle would not be allowed because of the disruption caused to crops and

PAGE 129

118 home gardens. Cultivation was discontinued in some areas because of the destruction caused by roaming cattle and water buffalo. Planters were forced to move their fields to the opposite side of the river. Water buffalo are especially adept at swimming across streams or around fences. Terra Firme Home Gardens Thirty-four home gardens were inventoried in three widely separated terra firme communities set in different environments. Nossa Senhora da Ftima is a river bluff community located on the right bank of the Solimes, a sediment laden white water river, upstream from Tef. Jubar sits low on the left bank of the Japur, a brown water river with a lower sediment load. Boa Esperana is a relatively isolated low lying village on the shore of Lake Aman, one of the large lakes formed near the mouths of tributaries when sea level was lower. The latter two communities are not immune from the effects of an abnormally high flood. Substantial areas of the village and agricultural fields were engulfed during the previous season causing large crop losses in both Jubar and Boa Esperana. It is not known what losses may have accrued in the home gardens of these two communities as residents were not interviewed. This leads to the realization that these communities, lying on the edge of the vrzea, are set in a transitional zone where they must still contend with the effects of the flood. The majority of terra firme home gardens are consistently diverse with a mean average of 21.4 species. All of the 34 home gardens surveyed contained at least ten useful species with five exceptions. One yard in Boa Esperana owned by a young man held just three individuals of three species. Four homes in Jubar did not have gardens. These houses are built on stilts in low areas that are seasonally flooded. When

PAGE 130

119 households do not have a home garden they typically share produce with a nearby family member or neighbor. It makes sense that Nossa Senhora da Ftima would have the greatest number of species since it lies on the high bluff and is unaffected by flooding. On the other hand, a number of home gardens in the other two communities are occasionally impacted by exceptionally high floods. In the survey, a total of 135 species were found in terra firme home gardens. Ninety-nine (73%) of these were also encountered in vrzea gardens leaving 36 species that were exclusive to the terra firme. The ten most abundant fruit and beverage species were aai (includes two species E. oleracea and E. precatoria), guava, banana, papaya, mango, cupuau, orange, peach palm, bacaba, and abiu. From table 5-2, we see that species with a low tolerance to flooding were much more common in terra firme home gardens such as avocado, pineapple (Ananas comosus), abiu, cupuau, orange, and peach palm. Banana was a frequent member of terra firme gardens as the third most abundant fruit and beverage species. Overall, half of the gardens contained banana with an average of 1.5 varieties and 5.1 individuals (Table 5-9). The most abundant types are two varieties of plantains (Table 5-10). Out of nine varieties recorded, five are plantains. Table 5-8. Average number of species in 30 terra firme home gardens Community N* Mean Min. Max. S Median All communities 30 21.4 3 47 10.1 20 Nossa Senhora da Fatima 8 28.3 16 42 8.8 30.5 Boa Esperana 17 18.9 3 47 10.8 15 Jubar 5 18.6 12 22 4.0 20 Homes without a home garden were omitted. (4 homes surveyed did not have useful species).

PAGE 131

120 Table 5-9. Number of banana varieties in terra firme home gardens Community Num. of samples Gardens with banana present % gardens with banana Avg. num. varieties* Avg. num. individuals** All communities 30 15 50 1.5 5.1 Nossa Senhora da Ftima 8 3 38 1.7 5.3 Boa Esperana 17 11 65 1.5 5.3 Jubar 5 1 20 1.0 2.0 Average calculated only for home gardens with banana present. ** Average number of torceiras calculated only for home gardens with banana present. Table 5-10. Banana varieties in terra firme home gardens Variety Num. individuals Num. home gardens Avg. num. individuals per home garden observed Paco vo* 33 5 6.6 Comprida* 17 4 4.3 Prata 15 4 3.8 Comprida 3 palmas* 4 2 2.0 Bai 1 1 1.0 Comprida po* 1 1 1.0 Comprida rabo de bandeira* 1 1 1.0 Guariba 1 1 1.0 Ma 1 1 1.0 Variety of plantain or cooking banana.

PAGE 132

121 Figure 5-5. Backyard with animals in a Solimes River community. The village was inundated with up to one meter of water and divided by a swift flowing furo or canal. This quintal was the only patch of dry ground left in the village for animals to seek shelter. Note that there is an obvious lack of plant diversity in this home garden with many animals. Several families also maintain a small herd of cattle and water buffalo. Community of Sitio Fortaleza.

PAGE 133

122 Figure 5-6. Maromba constructed to provide shelter for animals during the flood. Bela Vista do Manguari; two isolated homes on the Japur River.

PAGE 134

123 Figure 5-7. Cattle being fed the floating canarana grass during the flood season. Bela Vista do Manguari.

PAGE 135

124 Figure 5-8. Terra firme home garden. Raised planters are used to grow vegetables and herbs and netting is used to protect sensitive plants from animals. Numerous species and varieties of useful plants are densely planted in a small area around the house. Boa Esperana, Lake Aman.

PAGE 136

125 Figure 5-9. Well developed home gardens grade into diverse fruit orchards. Boa Esperana, Lake Aman.

PAGE 137

CHAPTER 6 A MORE STABLE ENVIRONMENT: TERRA FIRME Data was gathered on land-use management and agrobiodiversity in four communities on the uplands or terra firme. This is useful as a comparison in order to help resolve academic issues concerning the nature of agricultural systems on the vrzea as well as practical problems that can contribute to the development of agroforestry in the region. It also shows that we must consider not only the floodplain but also the surrounding terra firme if our goal is to protect and conserve agrobiodiversity in the region. The terra firme or uplands can be can be defined as all land that is not subject to flooding. This is contrasted with the vrzea or floodplain that is subject to the powerful forces of Amazonian rivers that persistently act to reshape the landscape. There are advantages for agriculture on both landforms. The principal benefit for agriculture on the floodplain is the seasonal or occasional replenishment of soil fertility in the form of fresh sediment deposited by the Solimes and Japur Rivers. This may allow for repeated cropping on the same location many times in succession. The terra firme field, on the other hand, often has low, quickly declining soil fertility and rapid weed invasion. As a consequence, it is usually only planted with one or two consecutive crops before production drops and a fallow period is initiated. The great advantage of the terra firme, as was often voiced by local residents, is its stability. Although more land may be required for a sustainable swidden rotation, 126 farmers need not contend with the unpredictability of the flood and the disruption to their

PAGE 138

127 lives it can cause. They may still lose production due to weed invasion, pests and diseases, and declining soil fertility but one of the major variables determining a good harvest on the vrzea, the flood, is taken out of the equation for terra firme farmers. Upland farmers have the option of building more permanent orchards or agroforestry systems that include crops intolerant to flooding. These are often ones that have a ready market and a high value as fruit or in the pulp industry. They include such crops as avocado, Barbados cherry, cupuau, lime, orange, papaya, pineapple, and soursop (Annona muricata). Local farmers often aspire to raise cattle and the uplands is perceived as an environment much more amenable to this purpose than the floodplain where there is the ever present problem of what to do with the herd during the flood. Cattle farmers on the river bluffs have the advantage of utilizing both upland and floodplain habitats for grazing. Unlike some relatively ephemeral floodplain communities, sometimes obliged to shift location due to the changing landscape of the vrzea, terra firme villages stay in a fixed location. With the relative permanence of the uplands comes stability for the family. There is often easier access to better schools with more regular schedules and higher quality health care. The data on fields, fallows, and home gardens on uplands was collected in four communities at dispersed locations throughout the study area. These communities differ dramatically with respect to the environment and access to markets, the habitats used for agriculture, and the relative emphasis on fruit or manioc production.

PAGE 139

128 Nossa Senhora da Ftima is located on a high bluff on the right bank of the Solimes upstream from the municipal center of Tef and slightly downstream from the town of Uarini. It is at the mouth of a small stream (igarap) that greatly facilitates access to fields during the high water season. Residents of Nossa Senhora da Fatima do not sustain damage from the flood but it greatly controls the ease of access to their fields. They typically travel by boat during high water to reach their port where they walk the remaining distance to the field. It is not uncommon to travel 30 minutes by boat and up to one hour on foot to reach the roa. Families in this area focus heavily on the production of manioc flour (farinha de mandioca) for the market. Many farmers cultivate several large fields totaling several hectares annually. They often rely on credit to purchase equipment like motors and chainsaws and hire labor for weeding and forest clearance. Boa Esperana is a relatively isolated community on the shores of Lake Aman. Although this village is at a much greater distance from markets than Nossa Senhora da Fatima, their access to the market is similar. There seemed to be a difference however, in the relative importance of manioc and fruit between the two communities. Residents of Boa Esperana had a greater reliance on fruits (avocado, banana, and lemon) until disease and flooding drastically cut the production of these crops. It is noteworthy that this community, although considered terra firme, suffered substantial losses due to flooding during the 2000 season. The lakeshore does not have high bluffs and low-lying areas are subject to flooding as the lake is connected to the Japur River by parans or natural canals.

PAGE 140

129 Jubar is on the left bank of the Japur River. There are no high bluffs in this area and residents reported that a number of upland fields were flooded with some loss of crops in 1999. Several homes in low-lying areas are built on stilts to avoid flooding and have no home garden plants. Farmers can cultivate both the vrzea and terra firme giving them greater options for choosing field locations. Finally, the community of Sitio Fortaleza lies on the floodplain a good distance from the river bluffs (one hour by motorboat or three hours rowing). Several households from this village regularly make fields on the bluffs in the territory of another community where they have relatives. Some of them are planning to permanently move to the terra firme in order to avoid the disruption to their lives caused by the flood. Like the farmers of the vrzea, cultivators of terra firme lands use a variety of descriptive terms to refer to land-use zones in different stages of the cropping cycle and with different crop compositions. The land-use designations used by farmers of the terra firme are summarized in table 6-1. Farmers conception of land-use on terra firme is essentially the same as on the high levees of the vrzea. Roa is a synonym for manioc field and it usually contains several additional minor crops. If a roa is interplanted with a significant amount of banana is may be referred to as manioc with banana (roa com banana), otherwise, there are no special designations for different types of roas. Non-roa work areas or fallows are divided into several types depending on the relative amount of investment in the site or the dominant planted species. A capoeira bananal or capoeira fruteira is a fallow that has been enriched with banana or other fruit trees to a greater density than a simple capoeira. Going one degree further, the designation of abacatal, aazal, bananal, or frutal imply an even denser planting of

PAGE 141

130 Table 6-1. Summary of local land-use designations by terra firme farmers Land-use designation English translation Num. visited Description Abacatal Avocado Grove 3 Area enriched primarily with avocado Aazal Aa Grove 1 Area enriched primarily with aa Bananal Banana Grove 3 Area enriched primarily with banana Capoeira Fallow 70 Fallow area Capoeira Bananal Banana Fallow 1 Fallow enriched primarily with banana Capoeira Fruteira Fruit Fallow 11 Fallow enriched with a variety of fruit trees Frutal Fruit Orchard 1 Area enriched with a variety of fruit trees Roa Manioc Field 87 Field planted to manioc and other crops. Sitio Site 4 Area of intensive, long-term management and/or a high diversity of useful plants avocado, aa, banana, or fruit trees, respectively. The sitio on terra firme is usually a mature fruit orchard surrounding the casa de farinha or manioc processing house when it is located away from the home site. Field and Fallow Management on Terra Firme Some of the most obvious differences one notices between roas on the vrzea and terra firme are their size, shape, and configuration. A typical roa on the vrzea is ten meters wide and 50 to 100 meters long while a typical roa on terra firme is much larger, measuring 100 by 100 meters or one hectare. The average size of terra firme roas is 6738 square meters (0.67 ha.) compared to only 2573 square meters (0.26 ha.) for roas on high levees. Whereas vrzea roas are long, narrow, and sometimes irregularly shaped following the contour of the levees, terra firme roas are typically rectangular with straight edges and distances measured in arm spans (braos).

PAGE 142

131 The average size of capoeiras is more than double that of roas because of the lumping together of fields into a single fallow, discussed in chapter four. The mean size of fallows is 1.4 hectares while the median is one hectare. There are some very large fallows of seven or eight hectares that presumably contain parts of different ages. The most that any one family would have in production at a time would be about four hectares. Most vrzea roas are confined to levees that are accessible by fluvial transport, usually along the major rivers or well-traveled side channels. Terra Firme roas may also be located on lakeshores or streams where they are accessible by boat. However, they are more commonly reached on foot along an array of trails carved from the forest. Trails are formed in roas while they are in production and are extended outward as new fields are cleared further from the village. Like the vrzea, most of the terra firme roas surveyed had been cleared from secondary forest as opposed to so-called mata virgem or virgin rainforest. Out of 68 roas, 44 (65%) were cleared from fallow and 24 (35%) were forged from old-growth forest. The age of fallows cleared for roas, reported in 39 cases, ranged from one to 15 years (Figure 6-1). Fields were left in fallow a mean of 5.1 years with a median of four. The length of the fallow period from the previous cropping cycle, also recorded for eight fallows, is consistent with this. The average was 5.8 years with a range from four to ten years. We can compare this to available data for high levees on the floodplain. The ages of 15 high levee fallows at the time of clearance ranged from four to 20 years with a mean of 7.9. Although it is a small sample, this indicates a possibility that fields are left fallow for longer periods on the vrzea, a question that warrants further research.

PAGE 143

132 The actual age was recorded for 63 fallows. They ranged from one to 20 years old with a mean average of almost four years and a median of three. From the histogram in figure 6-2, we see that the majority of fallows surveyed were one or two years old. There are two possible explanations for the propensity of young fallows on the uplands. One is that older fallows are larger, consisting of smaller fallows lumped together. There is a slight positive correlation (r=0.27) between the size of fallows and their age. The other possibility is that manioc agriculture is expanding in the communities surveyed. This second hypothesis is supported by preliminary satellite image processing work on manioc agriculture in the area of Nossa Senhora da Fatima and by interviews with farmers. It is likely a combination of both. Once forest is cleared, farmers usually plant one or two crops of manioc before the next fallow period. The average number of consecutive crops planted in roas is 1.6 (Figure 6-3). However, these roas are still in production. To determine the actual average number of consecutive crops planted we must look at the data for capoeiras and the number of previous crops planted (Figure 6-4). This data was reported for 25 fields and yielded a mean of 2.24 with a range of one to six crops and a median of two. This is less than the results for vrzea high levees. The number of previous crops planted on high levees ranged from one to 15 with a mean of 3.4 and median of two. Vrzea farmers are able to plant more crops, on average, than terra firme farmers. This is presumably due to differences in soil fertility and/or weed invasion. Farmers were asked about the number of cropping cycles on a particular piece of land. This refers to the number of times that the land has gone through a complete cycle from field to fallow and back to field again. The information was gathered for 33 terra

PAGE 144

133 firme fields (Table 6-2). The maximum number of cropping cycles recorded in the survey was five and the average was about two. Figure 6-5 shows the number of previous fallows for 66 terra firme roas. Adding one to the number of previous fallows would indicate the current cropping cycle. Figure 6-6 shows the number of cropping cycles for 35 capoeiras. Most of the current roas are in their second cropping cycle while most capoeiras are in their first cropping cycle. A common scenario is the clearance of old-growth forest for one or two consecutive manioc crops followed by a short fallow period of one to five years and another crop of manioc. Farmers decide if another crop will be planted based on growth and yield of the previous crop and the degree of weed invasion as well as field history, soil quality, and the presence of pests or disease. The field would then be allowed to rest in fallow for a longer period, on the order of eight to ten years or more. Farmers usually judge when a fallow is ready to enter a new cropping cycle by the height and thickness of the regenerating forest. Secondary forests accumulate living biomass at different rates depending on the soil quality. The number of consecutive crops and length of the fallow period also depends on available labor. In one instance, when asked why a particular field had been planted five times in a row, the woman replied that she was a widow and did not have the means to clear a new field from forest. The yield was greatly reduced because the land was not given time to recuperate. Crops typically take longer to grow on terra firme because of generally lower soil fertility. Bananas often take 12 months to produce as opposed to just eight months on the floodplain. An advantage of farming manioc on the terra firme is that it can be left in the

PAGE 145

134 ground until the farmer is ready to harvest rather than being forced to harvest before the flood. The longer it is left in the ground, the larger the tubers will be. Crops are normally left in the ground for one to two years. Fields can be partially harvested according to need and availability of time for processing. The average age of the upland manioc crop was almost eight months. The oldest crop that was recorded had been growing for 22 months. Losses of both manioc and other perennials in terra firme fields and fallows were sometimes reported. Farmers in both Boa Esperana and Jubar lost crops in low-lying areas in 1999. Some were forced to harvest their crop early and at least one field was totally lost. For example, a 0.24-hectare roa with 5,800 stems planted (covas) harvested early at seven months yielded seven sacks of tubers. Had the crop been left to grow for 12 months, an estimated 40 sacks would have been harvested. Several farmers in Nossa Senhora da Fatima reported significantly lower yields due to tuber rot (podre de batata). Crops also suffer occasional damage from peccaries and other wild animals such as deer and agouti. Table 6-2. Summary of statistics for roas on terra firme N Mean Min. Max. S Median Size (square meters) 79 6738 300 60000 9753 3500 Num. of species 87 4.17 1 15 2.85 4 Manioc varieties 84 2.76 1 12 1.86 2 Banana varieties 46 1.59 1 5 0.93 1 Crop age (months) 65 7.57 1 22 3.43 8 Consecutive crops 51 1.61 1 5 0.78 1 Num. cropping cycles 33 2.21 1 5 0.96 2 Last fallow (years) 39 5.12 1 15 3.05 4

PAGE 146

135 Table 6-3. Summary of statistics for capoeiras on terra firme N Mean Min. Max. S Median Size (square meters) 56 14,105 300 80,000 16,908 10,000 Num. of species* 43 3.93 1 10 2.35 3 Num. of banana varieties 17 1.76 1 5 1.25 1 Age of fallow (years) 63 3.8 1 20 3.3 3 Consecutive crops planted previously 25 2.24 1 6 1.23 2 Num. cropping cycles 35 1.77 1 5 1.14 1 Last fallow (years) 8 5.8 4 10 2.14 5 *Only capoeiras with useful species are included. 0246810123456789101112131415Length of Last FallowNumber of Fields Figure 6-1. Length of the previous fallow for 39 fields (roas) on terra firme 05101520135791113151719Age of Fallow in YearsNumber of Fallows Figure 6-2. Age of 63 fallows (capoeiras) 05101520253012345Consecutive CropsNumber of Fields Figure 6-3. Number of consecutive crops planted in 51 fields (roas) on terra firme

PAGE 147

136 05101520123456Consecutive CropsNumber of Fallows Figure 6-4. Number of consecutive crops previously planted in 25 fallows (capoeiras) on terra firme 051015202530354001234Previous FallowsNumber of Fields Figure 6-5. Number of previous fallows for 66 fields (roas). Adding one to the number of previous fallows indicates the number of planting cycles. 051015202512345Planting CyclesNumber of Fallows Figure 6-6. Number of cropping cycles for 35 fallows (capoeiras) on terra firme

PAGE 148

137 Figure 6-7. Eight-month-old terra firme field (roa) cleared from old-growth forest. Manioc is planted among fallen tree trunks. Community of Boa Esperana, Lake Aman. Figure 6-8. Roa developing into fruit fallow (capoeira fruteira). Boa Esperana, Lake Aman.

PAGE 149

138 Agrobiodiversity in Terra Firme Roas The dominant crops on the terra firme differ from those found on the vrzea but the pattern of diversity across land-use types is similar. Roas in both environments hold an average of four species. The histogram in figure 6-9 showing the number of species occurring in terra firme roas is comparable to the one for planted roas on vrzea high levees (Figure 3-2). The great majority of fields have two or more species with two being the most common. Just 14 out of 87 roas (16%) contained only a single species (manioc). The number of species in each upland field ranged up to 15 compared with 14 in restinga alta fields. Terra firme fields include many species common to the vrzea and several more that are relatively rare on the flood affected lowlands. A total of 51 useful species were documented in terra firme roas (Table 6-4). Crops that are grown principally in the uplands include pineapple, new world yam (Dioscoria trifida), Brazil nut, cupuau, and peach palm. Banana and the new world yam (car, a tuber similar to a potato) are the most frequently occurring secondary crops inter-planted in the manioc field. These two crops are encountered in half of all roas. They are followed by aa and pineapple, found in 24% and 21% of roas, respectively. Two valuable crops, Brazil nut and cupuau are both found in 15% of roas. Two palms, peach palm and tucum (Astrocaryum vulgare) are represented in 10% of roas. Peach palm is usually planted while tucum is often dispersed by animals. Various other fruits and, to a lesser extent, timber trees are also planted or protected in fields. Nineteen of the 50 useful species recorded in table 6-4 were encountered only once. A total of 31 varieties of bitter manioc (mandioca) and 5 varieties of sweet manioc (macaxeira) were found in 84 terra firme roas. Table 6-5 and 6-6 list the name, rank,

PAGE 150

139 and the number of fields where each variety occurred. The number of varieties in each field, shown graphically in the histogram in figure 6-10, ranged from one to 12. Most fields were planted with three, two, or one variety, respectively. The average was nearly three. The most common variety of manioc was Leonsa, found in 52% of roas, and second was Sete anos, found in 33% of roas. The repertoire includes five varieties of sweet manioc. The sweet varieties have a high risk of being destroyed by peccaries and agoutis so farmers avoid planting a large amount. It must be noted that the two most common varieties were found only in Boa Esperana, the largest community visited (63% of the terra firme roas surveyed were in this community). They were absent from the other three communities. It must be remembered that the same variety could be known by a different name in another community. It is imperative to initiate comprehensive research to document the diversity of manioc in the region in order to protect this valuable resource. Eleven banana varieties were recorded in 46 terra firme roas. Table 6-7 lists the varieties and their frequency of occurrence. Like fields and home gardens on the vrzea, the crop is represented by a few dominant varieties with others occurring at a low proportion. As with terra firme home gardens, plantains are the most common. Two varieties of plantain, Comprida and Pacovo, are the most frequently planted. They are found in 52% and 37% of fields, respectively. The number of varieties in each field ranged from one to five with an average of 1.6 (Figure 6-11).

PAGE 151

140 Table 6-4. The 51 useful species encountered in 87 terra firme roas Local name English name Num. roas % Local name English name Num. roas % Mandioca Manioc 87 100 Cupu Cupu 2 2 Banana Banana 48 55 Gangelim Gangelim 2 2 Car New world yam 44 51 Mamo Papaya 2 2 Aa (2 species) Aa 21 24 Manga Mango 2 2 Abacaxi Pineapple 18 21 Muiratinga Muiratinga 2 2 Abacate Avacado 14 16 Sova Sova 2 2 Castanha do Par Brazil nut 13 15 Alfavaca Alfavaca 1 1 Cupuau Cupuau 13 15 Andiroba Andiroba 1 1 Pupunha Peach palm 9 10 Angelim Angelim 1 1 Tucum Tucum 9 10 Araruta Araruta 1 1 Caj Cashew 8 9 Bacabo Bacabo 1 1 Bacaba Bacaba 6 7 Capitiu Capitiu 1 1 Goiaba Guava 5 6 Carapanaba Carapanaba 1 1 Jambo Malay apple 4 5 Envira Envira 1 1 Limo comum Lime 4 5 Gingibre Ginger 1 1 Piqui Piqui 4 5 Man Man 1 1 Cana de aucar Sugarcane 3 3 Marimari Marimari 1 1 Cedro Cedar 3 3 Maba Maba 1 1 Ing Inga 3 3 Pimenta Capsicum pepper 1 1 Jerimum Squash 3 3 Pimento Green pepper 1 1 Laranja Orange 3 3 Seringa Rubber 1 1 Milho Corn 3 3 Sucupeira Sucupeira 1 1 Pun Pun 3 3 Toari Toari 1 1 Tanimbuco Tanimbuco 3 3 Tucupi de arara Tucupi de arara 1 1 Cacau Cocoa 2 2 Uxi liso Uxi liso 1 1

PAGE 152

141 0 2 4 6 8 10 12 14 16 18 123456789101112131415 Number of SpeciesNumber of Fields Figure 6-9. Number of useful species in 87 fields (roas) on terra firme 051015202530123456789101112Number of Manioc VarietiesNumber of Fields Figure 6-10. Number of manioc varieties in 84 fields (roas) on terra firme 0510152025303512345Number of Banana VarietiesNumber of Fields Figure 6-11. Number of banana varieties in 46 fields (roas) on terra firme

PAGE 153

142 Table 6-5. The 31 varieties of bitter manioc (mandioca) found in 84 terra firme roas Rank Variety Num. roas % 1 Leonsa 44 52 2 Sete anos 28 33 3 Baiana 23 27 3 Tartaruga 23 27 4 Joo Gonzalo 14 17 5 Ar 12 14 6 Azulo 9 11 7 Parau 8 10 8 Antinha 7 8 8 Lira 7 8 9 Amarelinha 5 6 9 vila nova 5 6 10 Pe de lopes 4 5 11 Pretinha 3 4 12 Bod 2 2 12 Castanha 2 2 12 Farau 2 2 12 Pelonia 2 2 12 Seis meses amarela 2 2 12 Seis meses branca 2 2 12 Tucum 2 2 13 Aauria 1 1 13 Ajur 1 1 13 Amarela 1 1 13 Castinha 1 1 13 Corai 1 1 13 Geralda 1 1 13 Gonzalo 1 1 13 Maguari 1 1 13 Moura 1 1 13 Saca 1 1 Table 6-6. The 5 varieties of sweet manioc (macaxeira) found in 84 terra firme roas Rank* Variety Num. roas % 7 Caboquinha 8 10 11 Preta 4 3 12 Amarela 2 2 13 Africana 1 1 13 Roxa 1 1 rank is based on all varieties of bitter and sweet manioc

PAGE 154

143 Table 6-7. The 11 banana varieties found in 46 terra firme roas Rank Variety Num. roas % 1 Comprida* 24 52 2 Pacovo* 17 37 3 Prata 8 17 4 Tres palmas* 7 15 5 Ma 6 13 6 Baia 2 4 6 Guariba 2 4 6 Guariba branca 2 4 6 Nadj 2 4 6 So Tom 2 4 7 Guariba vermelha 1 2 variety of plantain Agrobiodiversity in Terra Firme Capoeiras Ninety-six non-roa areas were surveyed on terra firme. Out of these, 70 were called simple capoeiras by their owners. The remaining areas were referred to by other descriptive terms depending on crop diversity and management. Refer back to table 6-1 for a summary of the land-use types and their explanations. This section will cover the results of the survey for simple capoeiras. The next section will describe the other land-use types referred to by terra firme farmers. The histogram in figure 6-12 shows that, out of 70 capoeiras, 43 (61%) held useful species. Most of them contained between one and six. If we look at the 21 most commonly occurring species in table 6-9, those found in five or more fallows, we see that 18 are in the fruit, beverage, and nut categories. The three most common fruits are aa, avocado, and banana. There is a strong market for aa in urban centers. Although two species are grown on terra firme, aa preto and aa do Par, the latter is preferred for planting and marketing. Avocado is easily collected and transported and is sold at a high price in towns.

PAGE 155

144 051015202530012345678910Number of SpeciesNumber of Fallows Figure 6-12. Number of useful species in 43 fallows (capoeiras) on terra firme 02468101212345Number of Banana VarietiesNumber of Fallows Figure 6-13. Number of banana varieties in 17 fallows (capoeiras) on terra firme Banana was found in nearly half of the capoeiras. Most contain only one variety (Figure 6-13). Fourteen varieties of banana were documented (Table 6-8). Several of the fallows are managed for Brazil nut (castanha do Par), a valuable nut that fetches a good price at market. One farmer reportedly sells about 60 boxes of nuts per year for $RS 12 per box. Many of the other fruits on the list are also marketable and often contribute to a familys income. Other useful species common in terra firme fallows include such products as New World yam (car), a tuber commonly planted in roas that looks like a purple potato. It continues to grow in fallows long after the roa phase has ended. According to one informant, Car never dies. It can even be found growing in a 15-year-old fallow. Manioc also spouts again after harvest and is commonly found in terra firme fallows for some time after cultivation is discontinued. Unlike the vrzea, there is no flood to kill the

PAGE 156

145 plant and farmers sometimes take advantage of spontaneously sprouting manioc. Cacau jacar (Theobroma mariae) (caiman cocoa), a species of wild cacau, ing, a sweet leguminous fruit, and tucum, an oily palm fruit, are common volunteer trees in capoeiras. Seringa, the rubber tree, can be used to extract latex and is also a good fish bait. Like the vrzea, some terra firme fallows are reportedly managed for timber. Although we must go to number 25 on the list before we reach the first timber species (samauma), there are at least 20 species that are exploited for construction materials. The last, ubim (Geonoma sp.), is a palm whose fronds are utilized for roofing material. There is a wide variety of timber species with a low frequency of occurrence. Table 6-8. The 16 banana varieties documented in non-roa areas on terra firme and the frequency in each land-use type Local name All non-roa n=67 Capoeira n=43 Capoeira fruteira n=11 Bananal n=3 Frutal n=1 Sitio n=4 All varieties 30 18 3 3 1 3 Prata 12 5 2 3 1 1 Ma 10 5 2 2 Pacovo* 9 4 3 2 Comprida* 8 5 1 Tres palmas* 6 1 2 2 So Tom 4 3 1 Baia 3 1 1 1 Costela de vaca* 2 1 Guariba 2 2 Rabo de bandeira 2 2 1 Baia grande 1 1 Duas palmas 1 1 Guariba vermelha 1 1 Nadja 1 1 variety of plantain

PAGE 157

146 Table 6-9. The 68 useful species encountered in fallows on terra firme and the frequency they were recorded in each land-use type All non-Capoeira Local name English name roa n=67 Capoeira n=43 fruteira n=11 Bananal n=3 Frutal n=1 Sitio n=4 Aa (2 species) Aa 34 19 5 3 1 4 Abacate Avocado 30 13 9 3 1 1 Banana Banana 30 18 3 3 1 3 Pupunha Peach palm 25 9 8 2 1 4 Castanha do Par Brazil nut 23 13 3 2 1 3 Cupuau Cupuau 21 5 9 1 4 Limo comum Lime 18 8 5 1 3 Ing Inga 14 6 3 1 3 Abacaxi Pineapple 13 8 3 1 Car New world yam 13 8 3 1 1 Bacaba Bacaba 12 4 4 1 3 Tucum Tucum 12 9 1 1 Manga Mango 11 1 4 1 1 3 Cacau Cocoa 10 3 2 1 4 Caj Cashew 9 5 1 3 Goiaba Guava 8 2 1 2 1 1 Seringa Rubber 8 2 3 1 2 Mandioca Manioc 7 3 1 1 1 Cacau jacar Cacau jacar 5 2 1 2 Jambo Malay apple 5 3 2 Lima Lima 5 1 1 2 Cana de aucar Sugarcane 4 2 1 Laranja Orange 4 1 1 2 Piqui Piqui 4 2 1 1 Samama Kapok tree 4 1 1 2 Abiu Abiu 3 1 2 Berib Berib 3 1 1 1 Cupui Cupui 3 1 2 Graviola Soursop 3 1 1 1 Marimari Marimari 3 2 1 Puru Puru 3 1 2 Azeitona Olive 2 1 1 Caj a Caj a 2 1 1 Cedro Cedar 2 1 1 Mamo Papaya 2 1 1 Murucututu Murucututu 2 2

PAGE 158

147 Table 6-9. Continued All non-roa CapoeiraCapoeira fruteira Bananal FrutalSitio Local name English name n=67 n=43 n=11 n=3 n=1 n=4 Par par Par par 2 2 Pimenta Capsicum pepper 2 1 1 Sucuba Sucuba 2 2 Tangerina Tangerina 2 1 1 Tapereb Yellow mombim 2 2 Angelim Angelim 1 1 Anoir Anoir 1 1 Apui Apui 1 1 Ara Ara 1 1 Ara goiaba Ara goiaba 1 1 Arac pel Arac pel 1 1 Assac Assac 1 1 Bacabo Bacabo 1 1 Cap Cap 1 1 Carambola Carambola 1 1 Cedrorana Cedrorana 1 1 Lacre Lacre 1 1 Lauro abacate Lauro abacate 1 1 Lauro amarelo Lauro amarelo 1 1 Leitaba Leitaba 1 1 Limo cida Limo cida 1 1 Manga ma Manga ma 1 1 Maracuj Passion fruit 1 1 Maxixi Maxixi 1 1 Muirataua Muirataua 1 1 Pachiuba Pachiuba 1 1 Perereca Perereca 1 1 Tacana Tacana 1 1 Tanimbuco Tanimbuco 1 1 Taxi Taxi 1 1 Tinta arana Tinta arena 1 1 Ubim Ubim 1 1

PAGE 159

148 Other Non-Roas Farmers on the terra firme, like those on those on the vrzea, use descriptive terms to distinguish special work areas from simple capoeiras. These are areas where they have invested a relatively greater amount of time and energy in planting fruits and other crops (Table 6-8) and usually receive some degree of continuing management such as weeding and additional enrichment planting. They make up one-quarter of all non-roa work areas. Because of the considerable investment necessary for the production of tree crops and the high concentration of useful plants, they are often taken out of the roa cycle for an indefinite period. Old trees may be cut down and replaced but the area is rarely destroyed completely in order to plant a roa. Farmers sometimes plant orchards consisting of one principal crop. When a single species dominates, the land-use area is named for that crop. For example, an area especially rich in avocado, acai, or banana is referred to as an abacatal, acaizal, or bananal. Three avocado groves and one aa grove were documented in the community of Boa Esperana. The avocado groves are not densely planted orchards but avocado trees interspersed in a capoeira matrix. They are small areas containing up to four other perennial fruit crops. The aa grove is an area 50 x 30 meters recently planted (2 years) with 1000 aa, 50 cupuau seedlings, and several plantains. The bananal is rare on terra firme. Farmers tend to combine cultivation of banana with the roa rather than clearing an area solely for that purpose, a common practice on the vrzea. Only three were encountered in the terra firme survey, all in Boa Esperana. Unlike banana groves on the vrzea that are primarily monocultures, these were interplanted with a number of other crops. Two of them held more than ten species each while the third contained four. The more diverse banana groves were formed during an

PAGE 160

149 earlier roa phase. One of these was planted with banana and lemon. The lemon was wiped out by disease. The third was created in five-year-old capoeira using an interesting system. The undergrowth is cleared first and suckers are planted. After planting, the larger trees are chopped down and the area is later burned. It is much more common for farmers to plant a diverse mix of crops than a single species. Besides the roa and capoeira, the most frequent land-use designation used by farmers on the terra firme is fruit fallow (capoeira fruteiras). Eleven were visited; ten in Boa Esperana and one in Nossa Senhora da Fatima. The number of species ranged from three to 17 with an average of eight. Most of the capoeira fruteiras were developed from roas, either interplanted with manioc or enriched with fruits after a short fallow period. A number of lemon, avocado, and banana plants were lost in these areas from flooding and disease. One particularly diverse area, designated as a fruit orchard (frutal), contained 22 species including, according to the owner, 800 aa seedlings. It was cleared from forest eight years previously and, during that time, partially planted to manioc twice. There were four areas visited on terra firme that the owner was proud to call, my sitio. Three of them are in Boa Esperana and one is in Nossa Senhora da Fatima. They are located near the home and, in the case of Nossa Senhora, at the boat landing on the way to the fields. The latter consists of a diverse array of fruit trees growing around the manioc processing house. The ages range from six to 14 years and all of them were planted with one or two crops of manioc at the start. These mature fruit orchards are consistently diverse, with the number of species ranging from 13 to 21. A few valuable timber species are also included in the mix.

PAGE 161

CHAPTER 7 LESSONS FOR CONSERVATION OF AGROBIODIVERSITY IN THE MIDDLE SOLIMES In the coming decades, humans will continue to develop and intensify agriculture in order to feed our rapidly growing population. Farmers and scientists will strive to improve agricultural systems, making them more sustainable and better adapted to shifting markets and environments. Agrobiodiversity is crucial in this endeavor. Researchers and policy makers have singled out the Amazon vrzea as an area of great potential for increased arable farming and the development of agroforestry systems. To do this, we must take stock of the agrobiodiversity in existing farming systems and how it relates to management and habitat. Ultimately, we will want to protect this diversity. One way to do this is to initiate policies that encourage the conservation of biological diversity in the agricultural landscape and avoid policies that affect the loss of this valuable resource. If we wish to protect and conserve agrobiodiversity in the middle Solimes region, we must first understand its distribution across the agroecosystem. The objective of this research is to answer the question, Which systems of management and which environments are repositories of agrobiodiversity? This knowledge will help those concerned develop policies that lead to conservation of the diverse useful plant repertoire and the continuation of management practices that maintain high levels of plant diversity. 150

PAGE 162

151 Distribution of Agrobiodiversity What have we learned about the distribution of agrobiodiversity in the Middle Solimes region? It comes as no surprise that there is a great diversity of useful plants in the tropical agroecosystem of the Middle Solimes. Studies in other regions of the tropics, including Amazonia, have yielded similar results. This survey documented a total of 263 species used by local residents excluding ornamentals. The wide range of habitats that are utilized by farmers in the Middle Solimes is impressive. Farmers conceive of five principal habitats used for agriculture; beaches (praias), mudflats (lamas), low levees (restingas baixas), high levees (restingas altas), and uplands (terra firme). It must be emphasized that these categories are not rigidly divided, but instead form a continuum. Furthermore, there exist widely varying environmental characteristics between areas in the same category in terms of soils, vegetation, sediment deposition, impact from flooding, pests and diseases, and accessibility. The habitat determines the diversity of useful species to a great extent. This can be couched in terms of impact from flooding. The highest areas on the floodplain, the high levees, are flooded for a relatively short duration and many do not flood at all in normal years. This allows farmers to build up holdings of perennials, even those that are exotic to the vrzea and intolerant to flooding. For this reason work areas on floodplain high levees contain levels of diversity comparable to those on uplands. There is a myriad of management strategies within habitat types. The research revealed 24 designations of land use by local residents when divided along lines of habitat, current status (field or fallow), and management. The most diverse floodplain habitat in terms of land use types is the high levee with nine types recorded. Several additional classifications of managed capoeiras were described by Pinedo-Vasquez et al.

PAGE 163

152 (1996) that were not dealt with in this survey. The most species diverse land use types are also on high levees resulting in the bulk of agrobiodiversity on the varzea being found here. Ten land use categories were distinguished on terra firme. Again it must be stressed that, like habitat types, land use categories are anything but rigid and unchanging. Roas and capoeiries are extremely varied in management and dynamic in form. Table 7-1 helps to conceptualize the agroecosystem as a whole in terms of agrobiodiversity as it relates to habitat and land use. The number of work areas visited gives us a good idea of the proportions that each land use type is found. Combining this with the total number and average number of species gives a useful quick measure of the value of each land use type as a storehouse for agrobiodiversity. The number and percent with useful species indicate the proportion of each land use type that houses useful plants. By definition most of the land use types must contain useful species but this statistic is particularly useful for examining the number of fields and simple capoeiras that house economically valuable plants during the fallow period. Home gardens are the heavyweight champions of agrobiodiversity on both the vrzea and terra firme. Almost the complete repertoire of useful plants in the agroecosystem of the Middle Solimes is represented to some extent in home gardens, an important fact to remember when considering the conservation of agrobiodiversity in the region. Out of 204 total species with 19 possible uses documented on the vrzea, 176 (86%) of them were observed growing in home gardens. Of these, 87 species were found exclusively in home gardens. Likewise on terra firme, 135 out of 176 total species (77%) were present in home gardens with 92 species found only there. It is abundantly clear

PAGE 164

153 Table 7-1. Summary of species diversity by land use type on the vrzea and terra firme of the Middle Solimes Region Total Avg.* Num. with % with Land use type Num. visited num. species num. species useful species useful species Roa low levee harvested 93 23 2.4 16 17 Roa high levee planted 82 64 4.1 82 100 Roa high levee harvested 53 37 2.8 31 58 Roado high levee** 7 22 6.6 5 71 Capoeira low levee 50 25 2.8 17 34 Bananal 17 22 2.4 17 100 Capoeira high levee 141 68 3.5 88 62 Banana fallow 21 30 3.6 21 100 Cocoa fallow 1 1 1.0 1 100 Fruit fallow 12 41 10.2 12 100 Fruit orchard 2 25 16.0 2 100 Sitio 5 61 20.8 5 100 Home garden 171 176 12.0 163 95 Vrzea All vrzea 655 204 460 70 Home garden 34 135 21.4 30 88 Roa 87 51 4.2 87 100 Fallow 70 47 3.9 43 61 Avocado grove 3 10 4.0 3 100 Aai grove 1 3 3.0 1 100 Bananal 3 16 8.7 3 100 Banana fallow 1 5 5.0 1 100 Fruit fallow 11 29 7.9 11 100 Fruit orchard 1 18 18.0 1 100 Sitio 4 34 17.5 4 100 Terra Firme All terra firme 215 176 184 86 All work areas 872 263 644 74 Averages are calculated only for work areas containing useful species. ** Roado is a cleared field before planting. The useful species in a roado consist of those that are spared at the time of forest clearance. Unfortunately the sample size is small because the season of clearing fields was just beginning at the time of the research. Note: The habitats praia (beach) and lama (mudflat) are not included in the table since they rarely contain useful species after harvest.

PAGE 165

154 that home gardens represent a critical repository of agrobiodiversity in the Middle Solimes Region and a crucial resource for in situ conservation. Several other land use types boast levels of diversity comparable to home gardens. These areas include the fruit fallow (capoeira fruteiras), fruit orchard (frutal), and sitio. They are rarely developed by farmers. Only a few with sufficient knowledge and experience are willing to invest the time and energy necessary to create a diverse fruit orchard or sitio. Unlike home gardens they are located off the home site and often cover extensive areas. Fields contain multiple useful species more frequently than a single species. It was found that 82% of planted roas on high levees hold two or more species with an average of 4.1. They consist of a mix of planted and protected species. Farmers tend to plant polycultures even in low-lying beaches and mudflats where the land is flooded for many months each year, although these areas rarely contain useful species after harvest. On high levees, however, it is a common practice to interplant the manioc roa with banana and other perennial fruit trees that are left after manioc cultivation is discontinued. Even after harvest, high levee roas contained an average of 2.8 useful species. Fallows are also sometimes further enriched with fruits. In this way secondary forests are itinerantly enriched from cyclical planting and fallow periods when economically valuable species are spared, protected and planted in fields and fallows. This process could have the net effect of increasing overall biodiversity across the landscape. Results indicated a rich selection of manioc and banana varieties in the Middle Solimes. Approximately 65 varieties of manioc and 20 varieties of banana were reported. However, a few varieties of manioc and banana typically dominate while

PAGE 166

155 others are represented in only a small proportion of fields and gardens. The average number of varieties planted across all work areas is approximately 2.3 for manioc and 1.5 for banana. With the exception of the former cocoa grove (cacaual), the bananal is the land use type with the lowest average number of useful species. It is also not particularly diverse in varieties. Are we seeing, in these two land use types, the effects of streamlining production for the market? The bananal is typically a mono-cultural, mono-varietal planting that produces a product with a high demand in the marketplace. However, the management technique of selectively clearing forest (primarily undergrowth) and planting across an extensive area in partial shade could hold promise as a low impact form of agriculture on the floodplain. For effective conservation of agrobiodiversity in the Middle Solimes Region it is necessary to consider not only the vrzea but also the surrounding terra firme. Out of the 263 total species recorded in the survey 86 are exclusive to the vrzea and 59 are exclusive to the terra firme. Some plants that were given the same common name in the survey could also actually be separate species, adding to the list. It is common in the tropics for the same name to be given to a number of closely related species. Theoretical Implications The results of this research raise questions concerning some current theories about the nature of prehistoric agriculture and settlement on the Amazon river. According to Roosevelt (1980), fast growing annuals such as corn and beans were the principal crops grown on the vrzea in prehistoric times. She theorized that maize cultivation on fertile vrzea soil was the mechanism that propelled the development of dense populations and complex chiefdoms. Manioc is seldom mentioned as having been an important crop on

PAGE 167

156 the vrzea. In fact, Roosevelt claims it is ill-suited for floodplain agriculture and makes inefficient use of the fertile soils. Several lines of indirect evidence call into question Roosevelts ideas on the importance of manioc for prehistoric vrzea farmers. Through interviews with informants we learned of vrzea varieties that are clearly distinct from those of the terra firme. They are faster growing varieties adapted to the shorter growing season on the floodplain. The great number of varieties that are grown lends support to the idea of a deep history of manioc agriculture on the vrzea. Furthermore, when terra firme varieties are brought to they vrzea they frequently develop diseases, namely podre de batata or tuber rot. Research is needed to understand the diversity of manioc on the vrzea. A second point for consideration is that current practices seem to refute claims of manioc as an inappropriate crop for the vrzea. Local farmers have shown that manioc agriculture is viable on the vrzea by growing the crop in a range of habitats including mudflats, low levees, and high levees. When one sees the prevalence of manioc production on the middle Solimes vrzea today and the relatively unimportant role of maize, one wonders at the possibility of maize having once been the dominant crop. Further arguments against manioc as an important crop on the vrzea can also be refuted. The idea that manioc is a poor source of protein and corn, with its greater nutrient content, uses floodplain soils more efficiently and is more appropriate for sustaining dense populations is negated by the super abundance of animal protein on the vrzea. Some have also argued that manioc is not an appropriate crop for the production of surplus which is seen as a prerequisite for the development of complex societies. However, ethnohistorical accounts report that manioc tubers were buried in the ground on

PAGE 168

157 the floodplain and preserved for extended periods during the flood (Acua 1641). Clear ethnographic examples demonstrate that processed manioc can be stored for months. In the upper Xingu region of southern Amazonia, Amerindians store dried, processed manioc in large silos inside the home (Carneiro 1983). Manioc is also an ideal crop to store in the field since it does not need to be harvested until the farmer is ready. When flood waters stop rising, manioc fields on high levees are often left for several extra months to increase production by harvesting larger tubers. This research has shown the wide range of fruits that can be grown on the vrzea. Some of them may have been important crops for Amerindian populations that provided variation to the diet, food at different seasons, and rations for animals such as turtles. Most fruits on the floodplain ripen during the high water period to be dispersed by fish; at precisely the time when manioc production stops and fishing is least productive because fish are dispersed throughout a wide area. Fruits may have been an important food source during this time of scarcity. It is true that the dynamic vrzea is not a promising place to find evidence for the economies of prehistoric societies but, aside from Lathrap (1968) in the upper Amazon of Peru, there has been virtually no archaeological work done on the floodplain above Maraj Island. Intensive management of natural resources such as is evidenced in other regions of Amazonia (Ericson 2000, 2001; Heckenberger 1996; Heckenberger et al. 2003), and in the Solimes region itself from ethnohistorical documents (Carvajal 1934; Porro 1994), could have supported a much larger population on the vrzea than there is today. Further research on the living strategies of the local population along with

PAGE 169

158 archaeological and botanical studies must be undertaken in order to learn how prehistoric Amerindians exploited the vrzea environment. Directions for further research The conservation of agricultural biodiversity should be a priority in the Amazon Region and particularly on the vrzea. Local caboclo farming systems should be recognized as a research and conservation priority within the Mamirau project. Policies should be devised that strive to conserve agrobiodiversity in situ and care should be taken not to implement projects that result in its loss on a large-scale. While the principal concern of the Mamirau and Aman Sustainable Development Reserves is to conserve plant and animal diversity and not necessarily crop diversity, they are in an excellent position to consider conservation of agrobiodiversity. An infrastructure is in place for research, education, and extension work. Communities are sufficiently organized to discuss possibilities for resolving the problems of preservation and there is a growing sensitivity to issues concerning the conservation of natural resources especially if it concerns their livelihood. The Mamirau Institute has encouraged the cultivation of annual crops on less biodiverse beaches and non-forested levees as a way of taking pressure of the species-rich floodplain forests on high levees. They have also initiated programs to promote the planting of commercially valuable fruits and timber. Further research on agrobiodiversity could help in this effort. Questions about agricultural sustainability and compatibility with goals of biodiversity conservation should also be asked. Information on the abundance of useful species in fields and fallows is of critical importance. Except in the case of home gardens, this research has simply recorded the presence or absence of useful species in fields and fallows. More information is needed

PAGE 170

159 on the density of these species to more precisely estimate their actual abundance or scarcity in the agroecosystem. More work needs to be done concerning the origin of useful species in fields and fallows. Which species are planted and which are protected? Preliminary work has been done by Inuma (1999) in home gardens but much more research needs to be done concerning the management of useful species. Preliminary work in three communities by researchers at the Mamirau Institute has shown that fallow management and diverse agroforestry systems are only developed by a select few farmers with specialized knowledge and experience (Padoch et al. 1996; Pinedo-Vasquez et al. 1996). The research should be expanded to include a larger sample of communities. How many families actually practice agroforestry and fallow management? What reasons are given for not doing it? Floodplain farmers could greatly benefit by the development of agroforestry systems using flood tolerant crops or those that are native to the vrzea. Several questions need to be asked: Which crops with an existing market are most suitable for growing on the vrzea? Which crops have potential for developing and expanding markets? What is the potential for communities to create a value added product from their production? Possibilities for the manufacture of frozen fruit pulp, jams, and other products are already being discussed by residents of Mamirau. Critical research needs to be done on crop varieties. How many varieties are out there and what is their distribution? This research has barely touched on this problem but results have indicated that production consists of a few dominant varieties and other

PAGE 171

160 varieties are much less common. Are these heirloom varieties being lost in the interest of streamlining production? What can be done to conserve them in the field? If we want to further understand the spatial distribution of agrobiodiversity and the dynamics and sustainability of the agricultural system it is necessary to examine individual communities more closely. A combination of field data, interviews and satellite data can be used to answer some important questions: How do communities vary in their distance to, availability, and use of different habitat types? What is the distribution of agrobiodiversity among communities? How do families differ in the amount of land they use for agriculture and the amount of forest cleared annually? How do communities differ? What is the potential for expansion of annual crop production on low-lying areas? Are there fallows that were missed in the survey because they are old and forgotten or in areas difficult to reach? Is agriculture expanding on the floodplain? How much old growth forest is left? Fields are usually cleared close to major rivers and channels. Are there species endemic to these areas that are at risk of extinction? How do agriculture and forest dynamics differ on terra firme? An example of further work that can be done is shown here for the community of Betania. Figure 7-1 shows the resulting map of the territory of Betania with GPS positions of fields and home gardens. Thirty-four roas were visited in Betania for a total area of 80,500 m2 (8.05 ha.) among 17 owners. The average size of the roas is 2368 m2. The smallest is just 600m2 while the largest is one hectare (10,000m2). A total of 45 capoeiras were visited with a total area of 293,570 m2 (29.36 ha.) among 19 owners. The average size of the capoeiras is 6827 m2. The smallest is 480 m2 and the largest is 50,000

PAGE 172

161 m2. No useful species were encountered in 10 capoeiras. The other 35 have a mean of 4.4 useful species with a maximum of 16. The average age is 3.8 years with a maximum age of 13 years. We also know which habitats these fields and fallows are found. By classifying satellite images we can estimate the amount of land available for agriculture in each habitat type (Figure 7-2). Putting these together, we have an idea of the proportion of land used for agriculture and natural areas. Further progress could be made by doing more detailed participatory mapping of agriculture and resource use in individual communities (Figures 7-3 and 7-4). Improved GPS technology and the removal of selective availability allow accurate maps of agricultural fields, fallows, and resource extraction areas.

PAGE 173

162 Figure 7-1. Map of fields and home gardens of Betania. GPS positions are overlain on Landsat TM image, bands 5,4,3, August 1999. Color code for points as follows: green = home garden, red = restinga alta roa, black = restinga baixa roa, gold = lama, burgundy = praia.

PAGE 174

163 Beta n i a W h i t e w a t e r B l a c k w a t e r B e a c h L o w V e g e t a t i on M e a d o w M e d i u m V e g etation H i g h l e v e e U n c l a s s i f i e d 5 0 510Kilom e t e r s N Figure 7-2. Preliminary classification of satellite image in the vicinity of Betania from figure 7-1. High levees can be seen somewhat clearly but more work needs to be done to differentiate habitat types and land use using satellite image processing. New sensors with higher resolution will help in this effort.

PAGE 175

164 Figure 7-3. Map of fields and home gardens of So Paulo do Corai. Color code for GPS points is the same as figure 7-1.

PAGE 176

165 Figure 7-4. Participatory mapping of levees with fields and fallows in So Paulo do Corai. The community is in the center. Roas are in red and capoeiras are in green. Compare to satellite and GPS map in figure 7-3.

PAGE 177

APPENDIX A USEFUL SPECIES OCCURRING IN FIELDS AND FALLOWS

PAGE 178

Bacabo Oenocarpus sp. Arecaceae Beverage Bacabinha Oenocarpus mapora, O. minor Arecaceae Beverage 167Table A-1. Useful species occurring in fields and fallows Local name Scientific name Family Use Abacate Persea americana Lauraceae Fruit Abacaxi Ananas comosus Bromeliaeae Fruit Abiu Pouteria caimito Sapotaceae Fruit Acasou ? ? Aa do Par Euterpe oleracea Arecaceae Beverage Aa preto Euterpe precatoria Arecaceae Beverage Alfavaca Ocimum micrathum Lamiaceae Medicine Algodo Gossypium sp. Malvaceae Caulking Amacaca ? ? Amor crescido Portulaca sp. Portulacaceae Medicine Andiroba Carapa guianensis Meliaceae Medicine, timber Angelim Hymenolobium sp. Leguminosae Medicine, timber Anoir ? ? Timber Apu Clusia sp. ? Fruit Apurui ? ? Ara Eugenia sp. Myrtaceae Fruit, fish bait Ara boi Eugenia stipitata Myrtaceae Fruit, fish bait Ara do igapo Eugenia sp. Myrtaceae Fruit, fish bait Ara goiaba Eugenia sp. Psidium sp. Myrtaceae Fruit, fish bait Ara pel Eugenia sp. Psidium sp. Myrtaceae Fruit, fish bait Araruta Marantha arundinacea Marantacea Tuber, medicinal Assac Hura crepitans Euphorbiaceae Timber, rafts Azeitona Eugenia cuminii Myrtaceae Fruit Bacaba Oenocarpus distichus, O. bacaba, O. multicaulis Arecaceae Beverage

PAGE 179

Castanha do Par Bertholletia excelsa Lecythidaceae Nut Castanha sapucaia Lecythis pisonis Lecythidaceae Nut 168Table A-1. Continued Local name Scientific name Family Use Bacuri Rheedia sp., Platonia insignis, Moronobea canoidaGutiferacea Fruit Bacuri coroa Rheedia sp. Gutiferacea Fruit Bacuri liso Rheedia macrophylla, R. brasiliensis Gutiferacea Fruit Banana Musa sp. Musaceae Fruit Birib Rollinia deliciosa Annonaceae Fruit Boldo Pneumus boldus Monimiacea Medicinal Bolaxeira Apeiba sp. ? Buriti Mauritia flexuosa Arecaceae Fruit, beverage Cacau Theobroma cacao Sterculiaceae Fruit, chocolate Cacau jacar Theobroma mariae Sterculiaceae Fruit Cacaurana Theobroma speciosum, T. mariae Sterculiaceae Caj Anacardium occidentale Anacardiaceae Fruit Caj a Anacardium giganteum Anacardiaceae Fruit Cajurana Pouteria aff. elegans Sapotaceae Camapu Physalis angulata Solanaceae Vegetable Camu camu Myrciaria dubia Myrtaceae Fruit Cana de aucar Saccharum sp. Graminea Snack, beverage, medicinalCapim santo Cymbopogon citratus Poaceae Beverage, medicinal Capitiu, ca-piti, (caapitiu?) Siparuna guianensis, (Renealmia occidentalis?) Siparunacea, (Zingerberacea) Medicinal Cap, (capuchina?) (Tropaeslum majus?) (Trepaeolacea?) Timber Car Dioscorea trifida Dioscoriaceae Tuber Carambola Aberrhoa caranbola Oxalidaceae Fruit Carapanaba Aspidosperma nitidum Apocynaceae Timber Castanha de macaco Couroupita guianensis (Lecythis usitata?) Lecythidaceae Nut

PAGE 180

Goiaba ara Psidium acutangulum Mytaceae Fruit Graviola Annona muricata Annonaceae Fruit 169Table A-1. Continued Local name Scientific name Family Use Castanha de tracaj ? ? Nut Castanholeira Terminalia catappa Combretaceae Nut Cator Crataeva benthami Capparidaceae fFsh bait Cebolinha Allium fistulosum Liliaceae Condiment Cedro Cedrella fissilis Meliaceae Timber Cedrorana Cedrelinga catenaeformis Leguminosae: Mimosoideae Cco Cocos nucifera Arecaceae Nut Coentro Coriandrum sativum Umbelliferae Condiment Copaba Copaifera sp. Leguminosae Medicinal Couve Brassica oleraceae Brassicaceae Vegetable Cravo Dianthus sp. Cariofilaceae Cubiu Solanum sessiliflorum Solanaceae Fruit Cuia Crescentia cujete Bignonaceae Bowl Cupuau Theobroma grandiflorum Sterculiaceae Fruit Cupu Theobroma subincanum Sterculiaceae Fruit Envira Rollinia sp. Annonaceae Timber Envira cacau Rollinia sp. Annonaceae Timber Envira vassourinha Xylopia cf. calophyllum Annonaceae Timber, broom Envira torrada ? Annonaceae Timber Feijo de praia (manteiginho) Vigna unguiculata Fabaceae Vegetable Fruta po Artocarpus altilis Moraceae Fruit Gergelim Sesamum indicum Pedaliacea Condiment Gingibre Gingiber officinale Zingiberaceae Condiment Goiaba Psidium guajava Mirtaceae Fruit

PAGE 181

Limo tangerina, Limo de suco Citrus sp. Rutaceae Fruit Macacaricuia Couroupita sp., Eschweilera sp. ? 170Table A-1. Continued Local name Scientific name Family Use Hortel Mentha sp. Lamiaceae, Labiatae Medicinal Hortelzinho Mentha piperita Lamiaceae, Labiatae Medicinal Hortel roxa Mentha sp. Lamiaceae, Labiatae Medicinal Ing Inga sp. Mimosaceae Fruit Itu Gnetum nodiflorum Gnetaceae Medicinal, food Jacareba Calophyllum brasiliense, C. angulare Clusiaceae, Gutiferacea Timber Jambo Eugenia malaccensis, E. jambos Myrtaceae Fruit Jauari Astrocaryum jauari Arecaceae Fish bait Jenipapo Genipa americana Rubiaceae Fruit Jerimum Cucurbita sp. Cucurbitaceae Vegetable Jit Guarea sp. ? Timber Lacre Vismia sp. Guttiferae Laranja Citrus sinensis Rutaceae Fruit Lauro abacate Aniba sp. ? Timber Lauro amarelo Nectandra sp. ? Timber Lauro chumbo ? ? Timber Lauro jacar ? Lauraceae Timber Louro inamu Ocotea cymbarum Lauraceae Timber, medicinal Leitaba ? ? Lima Citrus spp. Rutaceae Fruit Limo comum Citrus aurantifolia Rutaceae Fruit Limo de Caiena Averrhoa bilimbi Oxalidaceae Fruit Limo caipirinha ? Rutaceae Fruit Limo cida (cidra?) (Citrus medica?) Rutaceae Fruit

PAGE 182

Murumuru Astrocaryum murumuru Arecaceae Murupi ? ? 171Table A-1. Continued Local name Scientific name Family Use Macacaba Platymiscium ulei, P. duckei Leguminosae: Papilionoideae Timber Majirio Ocimum sp. Lamiaceae Medicinal Mamo Carica papaya Caricaceae Fruit Man Fraxinus ornus Oleaceae Medicinal Mandioca Manihot esculenta Euphorbiaceae Tuber Manixi Brosimopsis oblongifolia Moraceae Manga Mangifera indica Anacardiaceae Fruit Manga maa Mangifera sp. Anacardiaceae Fruit Mapati Pourouma cecropitolia Moraceae Medicinal Maracuj Passiflora edulis Passifloraceae Fruit Maracuj do igap ? ? Fruit Marimari Cassia leiandra, C. spruceana, C. grandis Leguminosae: Caesalpinioideae Fruit, medicinal Marirana Coepia sp. Chrysobalanaceae Mastruz Chenopodium ambrosioides Chenopodiaceae Medicinal Maba Clinostemon mahuba (Licaria mahuba) Lauraceae Timber Maxixe Cucumis anguaria Cucurbitaceae Vegetable Melancia Citrullus vulgaris Cucurbitaceae Fruit Milho Zea mays Poaceae Food Mucura ca Petiveria alliacea Phytolacaceae Medicinal Mucuba ? ? Muirataua, Muirajuba Apuleia molaris Leguminosae: Caesalpinioideae Timber Muiratinga Maquira spruciana (M. coriaceae?) Moraceae Timber Mulateiro Calycophyllum spruceanum Rubiaceae Timber Murucututu ? ?

PAGE 183

Tapereb Spondias mombim Anacardiaceae Fruit Taxi ? ? Timber 172Table A-1. Continued Local name Scientific name Family Use Namuirana ? ? Pachiuba ? ? Timber Par par Jacaranda copaia, Cordia umbraculifera Bignonaceae Medicinal, timber Paricarana Pithecellobium corymbosum ? Timber Pau ferro ? ? Timber Perereca ? ? Timber Pimenta (5 varieties) Capsicum sp. Solanaceae Condiment Pimento Capsicum annum Solanaceae Condiment Piqui Caryocar villosum Caryocaraceae Fruit, timber Pitomba Talisia esculenta Sapindaceae Fruit Pun ? ? Timber Pupunha Bactris gasipaes Arecaceae Fruit Puru Boroja sorbilis Rubiaceae Fruit Samauma Ceiba pentandra Bombacaceae Timber Seringa Hevea brasiliense Euphorbiaceae Latex Seringa barriguda Hevea spruceana Euphorbiaceae Latex Seringa murupita ? ? Latex Sorva Couma utilis Apocynaceae Latex Sucupira Andira micrantha Leguminosae: Papilionoideae Timber Sucuba Himatanthus tarapotensis, H. sucuuba Apocynaceae Timber Tacaca Sterculia sp. Sterculiaceae Timber Tacana ? ? Tangerina Citrus nobilis var. deliciosa Rutaceae Fruit Tanibuca Terminalia tanibouca Combretaceae Timber

PAGE 184

173Table A-1. Continued Local name Scientific name Family Use Tinta arana ? ? Toari ? ? Timber Tomate Lycopersicum esculentum Solanaceae Tucum Astrocaryum vulgare, A. tucuma, A. aculeatum Palmacea, Arecaceae Fruit Tucupi de arara ? ? Ubim Geonoma sp. Arecaceae Ucuba Virola surinamensis Myristicaceae Timber, fishbait Uru Cordia nodosa Boraginaceae Fish bait, livestock feed Urucu Bixa orellana Bixaceae Food colorant Urucuri Attalea minor Arecaceae Roofing material Uxi liso Endopleura uchi Humiriaceae Fruit Sources: Araujo-Lima and Goulding 1997; Brucher 1989; Cavalcante 1974; Cid 1978; Cointe 1947; Fernandes et al. 1952; Gentry 1993; Inuma 1999; Lima 1994; Popenoe 1920; Purseglove 1968a, 1968b, 1972a, 1972b; Ribeiro 1999; Smith 1999; Smith et al. 1992, 1995; Stasi 1989. Note: Scientific names and families are a best guess as no samples were collected in the field. The objective of this research was to document the spatial distribution of agrobiodiversity in the middle Solimes region rather than to identify and name each plant. This list includes all plants found in vrzea and terra firme home gardens considered useful by residents except ornamentals.

PAGE 185

APPENDIX B USEFUL SPECIES OCCURRING IN HOME GARDENS

PAGE 186

Araticum Annona sp. Annonaceae medicinal 0027Arruda Ruta graveolens Rutaceae medicinal 3227Assac Hura crepitans Euphorbiaceae medicinal, timber 5300 175Table B-1. Useful species occurring in home gardens Local Name Scientific Name Family Use Num. varzea gardens (163) % Num. terra firme gardens (30) % Abacate Persea americana Lauraceae fruit 11930 Abacaxi Ananas comosus Bromeliaeae fruit 21620Abiu Pouteria caimito Sapotaceae fruit 64930Aa do Par, Aa preto (2 species) Euterpe oleracea, E. p recatoria Arecaceae beverage 46282067Acerola Malpighia glabra Malpighiaceae fruit 1100Alface Lactuca sativa Asteraceae vegetable 3200Alfavaca Ocimum micrathum Lamiaceae medicinal 34211033Algodo Gossypium sp. Malvaceae caulking 5313Algodo branco Gossypium barbadenseMalvaceae caulking 2100Algodo roxo Gossypium sp. Malvaceae caulking 161013Alho Allium sativum Liliaceae condiment, medicinal 0013Amor crescido Portulaca sp. Portulacaceae medicinal 6427Anador planta ? ? medicinal 4200Andiroba Carapa guianensis Meliaceae medicinal, timber 21310Apu Clusia sp. Moraceae fruit 1100Ara boi Eugenia stipitata Myrtaceae fruit, fishbait 42413Ara do igap Eugenia sp. Myrtaceae fruit, fishbait 2100Ara goiaba Eugenia sp. Psidium sp. Myrtaceae fruit, fishbait 2012310Ara pel Eugenia sp. Psidium sp. Myrtaceae fruit, fishbait 00413Ara pera Eugenia estipitata Myrtaceae fruit, fishbait 0013Arati ? ? 3200

PAGE 187

Cacaui Theobroma speciosum, T. sylvestre Sterculiaceae fruit 0013Cacau jacar Theobroma mariae Sterculiaceae fruit 42310 176Table B-1. Continued Local Name Scientific Name Family Use Num. varzea gardens (163) % Num. terra firme gardens (30) % Assacui Euphorbia cotinoides Euphorbiaceae timber 0013 Azeitona, Ameixa Szyzgium cumini Myrtaceae fruit 11727Bacaba Oenocarpus distichus, O. bacaba Arecaceae beverage 24151757Bacabinha Oenocarpus mapora, O. minor Arecaceae beverage 1127Bacuri Rheedia sp., Platonia insignis, Moronobea canoida Gutiferacea fruit 9600Bacuri coroa Rheedia sp. Gutiferacea fruit 5327Bacuri liso R eedia macrophylla, R. brasiliensis Gutiferacea fruit 14900Banana (13 varieties) Musa sp. Musaceae fruit 56342170Batata doce (3 varieties) Ipomea batatas Convulvulaceae tuber 11517Bau Bau ? ? medicinal 1100Birib Rollinia deliciosa Annonaceae fruit 11517Boldo Pneumus boldus Monimiacea medicinal 106310Boldo folha grauda ? ? medicinal 1100Boieira Apeiba sp. ? 5300Buriti Mauritia flexuosa Arecaceae fruit, beverage 19121137Cacamo ? ? 1100Cacau Theobroma cacao Sterculiaceae fruit, chocolate, medicinal 52321757

PAGE 188

Castanha sapucaia Lecythis pisonis Lecythidaceae nut 11700Castanholeira Terminalia catappa Combretaceae nut 8500Cator Crataeva benthami Capparidaceae fish bait 2100 177Table B-1. Continued Local Name Scientific Name Family Use Num. varzea gardens (163) % Num. terra firme gardens (30) % Cacaurana Theobroma speciosum, T. sylvestre Sterculiaceae fruit 3213 Caf Coffea arabica Rubiaceae beverage 00310Caj Anacardium occidentale Anacardiaceae fruit 23141343Caj a Anacardium giganteumAnacardiaceae fruit 0013Cajurana Pouteria aff. elegans Sapotaceae fruit 1100Calipi ? ? 0013Camu camu Myrciaria dubia Myrtaceae fruit 281713Cana de aucar Saccharum sp. Graminea snack, beverage 2918413Capeba Pothomorphe sidaefolia Piperacea 0013Capim santo Cymbopogon citratus Poaceae beverage, medicinal 4528620Car Dioscorea trifida Dioscoreaceae tuber 0027Carajuru Arrabidea chica Bignonaceae medicinal 00413Carambola Averrhoa caranbola Oxalidaceae fruit 2113Carapanaba Aspidosperma nitidum Apocynaceae timber 1100Caraucu ? ? 1100Cariru, Caruru Amaranthus flavus Amarantaceae vegetable 138620Carrapateira ? ? medicinal 0013Castanha eletaca ? ? nut 0013Castanha de macaco Couroupita guianensis, Lecythis usitata Lecythidaceae nut 6400Castanha do Par Bertholletia excelsa Lecythidaceae nut 00310

PAGE 189

Cuiamanza Polyscias sp. Araliaceae 3227Cuminho (Cuminho bravo) (Pectis elongata) (Composta) condiment 3213 178Table B-1. Continued Local Name Scientific Name Family Use Num. varzea gardens (163) % Num. terra firme gardens (30) % Catinga de mulata Leucas martinicensis Labiatae medicinal 64517 Cauacu ? ? 0013Caximguba Ficus sp. Moraceae mulch, fishbait 12713Cebolinha Allium fistulosum Liliaceae condiment 90552067Cedro Cedrella fissilis Meliaceae timber 10600Cheiro verde ? ? condiment 2127Chicria Eryngium ekmanii, Cichorium intybus Umbelliferae, Composta condiment 159827Chifre de corno ? ? medicinal 0027Cibalena ? Zingiberaceae medicinal 1100Cidreira Melissa officinalis Lamiaceae medicinal 15927Cidrela Lippa alba Verbenaceae 0027Cip-alho Mansoa alliacea Bignoniaceae 21310Coco Cocos nucifera Arecaceae beverage, nut 33201240Coentro Coriandrum sativum Umbelliferae 8527Corama B ryophyllum pinnatum, B. calycimum Crassulaceae medicinal 0013Corisa ? ? 2100Couve Brassica oleraceae Brassicaceae vegetable 1610930Cravo Dianthus sp. Cariofilaceae 74413Cravo-de-defunto Tagetes erecta, T. p atula Asteraceae 4200Cubiu Solanum sessiliflorum Solanaceae fruit 2918517Cuia Crescentia cujete Bignonaceae bowl 93571240

PAGE 190

Ing curta Inga sp. Mimosaceae fruit 1100Ing do igap Inga sp. Mimosaceae fruit 2100Ing sapo Inga sp. Mimosaceae fruit 2100 179Table B-1. Continued Local Name Scientific Name Family Use Num. varzea gardens (163) % Num. terra firme gardens (30) % Cupuau Theobroma grandiflorum Sterculiaceae fruit 1171653 Envira Rollinia sp. Annonaceae 1100Envira cacau Rollinia sp. Annonaceae 2100Envira-vassorinha Xylopia cf. calophyllum Annonaceae 2100Envirinha ? Annonaceae 1100Erva doce Pimpinella anisum ? medicinal 00413Eucalypto (arbusto) ? ? medicinal 1100Feijo de corda Vigna sinensis Fabaceae vegetable 1100Fruta-po Artocarpus altilis Moraceae fruit 7400Gergelim Sesamum indicum Pedaliacea medicinal 3200Goiaba Psidium guajava Mirtaceae fruit 107662170Goiaba ara Psidium acutangulum Mytaceae fruit 8500Graviola Annona muricata Annonaceae fruit 4628930Hortel Mentha sp. Lamiaceae, Labiatae medicinal 85620Hortel folha grada Plectramthus amboicus, Mentha viridis Lamiaceae, Labiatae medicinal 149310Hortel folha mida Mentha sp., Mentha villosa Lamiaceae, Labiatae medicinal 4200Hortelzinho Mentha piperita Lamiaceae, Labiatae medicinal 64517Ing au Inga cinnamoea Mimosaceae fruit 6339620Ing cip / comprida Inga edulis Mimosaceae fruit 29181137

PAGE 191

Limo de suco 106620Macucu Aldina heterophylla, Ilex macoucoua Leguminosae: Papilionoideaa medicinal, dye 1100 180Table B-1. Continued Local Name Scientific Name Family Use Num. varzea gardens (163) % Num. terra firme gardens (30) % Ing sp. Inga sp. Mimosaceae fruit 12713 Ingsinho Inga sp. Mimosaceae fruit 2100Jacareba Calophyllum brasiliense, C. angulare Clusiaceae, Gutiferaceatimber 2100Jaismia de gato (Jasmim?) ? ? 1100Jambo Eugenia malaccensis, E. jambos Myrtaceae fruit 23141757Jambu Spilanthes acmella Compositae medicinal 12713Japana Eupatorium ayapana Compositae medicinal 1113Jauari Astrocaryum jauari Palmae fish bait 2113Jenipapo Genipa americana Rubiaceae fruit 1610310Jerimum Cucurbita sp. Cucurbitaceae vegetable 4200Jit Guarea sp. ? timber 1100Juc Caesalpinia ferrea Luguminosae medicinal 0013Lacre Vismia sp. Guttiferae 1100Laranja Citrus sinensis Rutaceae fruit 106827Louro Varium Lauraceae medicinal, timber 5300Louro-inamu Ocotea cymbarum Lauraceae medicinal, timber 4200Lima Citrus spp. Rutaceae fruit 138517Limo comum Citrus aurantifolia Rutaceae fruit, condiment 2817827Limo cida (cidra?) (Citrus medica?) Rutaceae fruit 1100Limo tangerina, Citrus sp. Rutaceae fruit

PAGE 192

spruceanum 10600Munguba Pseudobombax munguba Bombacaceae mulch 11700 181Table B-1. Continued Local Name Scientific Name Family Use Num. varzea gardens (163) % Num. terra firme gardens (30) % Majirio Ocimum sp. Lamiaceae medicinal 271713 Malvasco, Ca peba Piper peltatum Malvacea medicinal 4213Mamo Carica papaya Caricaceae fruit 17101860Mandioca, Macaxeira Manihot esculenta Euphorbiaceae tuber 32723Manga Mangifera indica Anacardiaceae fruit 23141653Mangarataia Zingiber officinalis Zingiberaceae condiment, medicinal 106413Manixi Brosimopsis oblongifolia Moraceae 1100Maracuj Passiflora edulis Passifloraceae fruit 1710620Marcela Pluchea quitoc Compositae medicinal 0013Mari-mari Cassia leiandra, C. spruceana, C. grandis Leguminosae: Caesalpinioideae fruit, medicinal 21310Marirana Coepia sp. Chrysobalanaceae fruit 106620Mastruz Chenopodium ambrosioides Chenopodiaceae medicinal 1912620Maxixe Cucumis anguaria Cucurbitaceae vegetable 11700Melancia Citrullus vulgaris Cucurbitaceae fruit 0013Melhoral Piper cavalcantei Piperacea medicinal 0013Menta ? ? 1100Milho Zea mays Poaceae food, livestock feed 1100Mucura ca Petiveria alliacea Phytolacaceae medicinal 85723Muiratinga M aquira spruciana, M. coriaceae Moraceae timber 221300Mulateiro Calycophyllum Rubiaceae timber

PAGE 193

varieties) 89552480Pimenta doce (2 varieties) Pimenta docia Myrtaceae condiment 53827 182Table B-1. Continued Local Name Scientific Name Family Use Num. varzea gardens (163) % Num. terra firme gardens (30) % Mungubarana Pachira sp., Bombax p araense Bombacaceae 0013 Murucuri ? ? 1100Murupita Sapium hippomane, Sapium sp. Euphorbiaceae 3200Mutamba Guazuma unifolia Esterculiacea medicinal, fruit, fiber, livestock feed 8500Mututi Paramachaerium ormosioides Leguminosae 1100Orelho de porco ? ? 3200Palheteira Clitoria racemosa Leguminosae timber 8527Palmaceiteira ? ? 0013Panquil ? ? medicinal 1100Paric Virola theiodora, Peptadenia peregrina Myristicaceae, Numosacea 3200Parigorico ? ? 1113Patchouli Pogostemon patchoulyLabiatae medicinal 0013Pau ferro Acacia catehu ? timber 1100Pau de ralo ? ? 1100Peo branco Jatropha curcas Euphorbiaceae medicinal 3723620Peo roxo Jatropha gossyfiifolia Euphorbiaceae medicinal 39241447Pepino Cucumis sativus Cucurbitaceae vegetable 2100Pico Bidens bipinnatus Compositae 0013Pimenta ardosa (6 Capsicum frutescens Solanaceae condiment

PAGE 194

Taioba Xanthosoma sp. Aracea 1100Tamanqueira Zanthoxylum sp. Rutaceae timber 1100Tmara ? (palm) 1100 183Table B-1. Continued Local Name Scientific Name Family Use Num. varzea gardens (163) % Num. terra firme gardens (30) % Pimento Capsicum annum Solanaceae vegetable 1710517 Piqui Caryocar villosum Caryocaraceae fruit 00310Pirarucu ca Kalanchoe pinnata Crassulaceae medicinal 138310Pitomba Talisia esculenta Sapindaceae fruit 2113Pupunha Bactris gasipaes Arecaceae fruit 1061550Puru Boroja sorbilis Rubiaceae fruit 1610310Purum Pourouma guianensis Cecropiaceae 1127Quebra-pedra Phyllanthus corcovadensis Euphorbiaceae medicinal 0013Quiabo Abelmoschus sp. Malvacea vegetable 8513Repolho Brassica oleraceae var. capitata Apiaceae vegetable 3200Sapota (Sapot?) (Achras sapota?) (Sapotaceae?) 0027Se de sangria / bacuralzinho ? ? medicinal 0013Senna Senna sp. Leguminosae: Caesalpinioideae medicinal 6427Seringa Hevea brasiliensis Euphorbiaceae latex 7413Seringa barriguda Hevea spruceana Euphorbiaceae latex 1100Sucuba Himatanthus tarapotensis, H. sucuuba Apocynaceae timber 1127Samama Ceiba pentandra Bombacaceae timber 6400Samaumarana ? ? 1100

PAGE 195

Vick Mentha spicata Lamiaceae medicinal 0013 184Table B-1. Continued Local Name Scientific Name Family Use Num. varzea gardens (163) % Num. terra firme gardens (30) % Tangerina Citrus nobilis var. deliciosa Rutaceae fruit 11413 Tanibuca Terminalia tanibouca Combretaceae timber 5300Tapereb Spondias mombin Anacardiaceae fruit, medicinal 2717310Taxi Triplares schomburgkiana, Sclerolobium goeldianum Polygonaceae, Leguminosae: Casalpiniacea timber 1912310Taxi branco Pterocarpus sp. Polygonaceae, Luguminosae: Casalpiniacea timber 1100Taxi vermelho Triplares surinamensisPolygonaceae timber 0013Tecura ? ? medicinal 0013Tomate Lycopersicum esculentum Solanaceae vegetable 3622310Trevo roxo Hyptis atrorubens Lamiaceae medicinal, ornamental 2113Tucum Astrocaryum vulgare, A. tucuma, A. aculeatum Palmacea, Arecaceae fruit 00413Turima ? ? 3200Ucuba Virola surinamensis Myristicaceae timber 5300Urucu Bixa orellana Bixaceae food colorant 1271550Urucurana Sloanea sp. Elaeocarpaceae 2100Urucuri Attalea phalerata Arecaceae 11713Vence-tudo ? ? medicinal 1127

PAGE 196

185Sources: Araujo-Lima and Goulding 1997; Brucher 1989; Cavalcante 1974; Cid 1978; Cointe 1947; Fernandes et al. 1952; Gentry 1993; Inuma 1999; Lima 1994; Popenoe 1920; Purseglove 1968a, 1968b, 1972a, 1972b; Ribeiro 1999; Smith 1999; Smith et al. 1992, 1995; Stasi 1989. Note: Homes without a home garden were omitted. (8 home gardens surveyed on the varzea and 4 on terra firme did not have useful species). Scientific names and families are a best guess as no samples were collected in the field. The objective of this research was to document the spatial distribution of agrobiodiversity in the middle Solimes region rather than to identify and name each plant. This list includes all plants found in vrzea and terra firme home gardens considered useful by residents except ornamentals.

PAGE 197

LIST OF REFERENCES Acua, C. (1891). Nuevo Descubrimento del Gran Rio de las Amazonas en El Ao 1639. Garca, Madrid. Adams, M. W., A. H. Ellingboe, and E. C. Rossman (1971). "Biological Uniformity and Disease Epidemics." BioScience 21(21): 1067-1070. Albernaz, A. L. K. M., and Ayres, J. M. (1999). Selective Logging along the Middle Solimoes River. Vrzea: diversity, development, and conservation of Amazonia's white water floodplains. C. Padoch, J. M. Ayres, M. Pinedo-Vasquez, and A. Henderson editors. New York, The New York Botanical Garden Press. 13: 135-152. Alcorn, J. B. (1984). "Development Policy, Forests, and Peasant Farms: Reflections on Huastec-managed Forests' Contributions to Commercial Production and Resource Conservation." Economic Botany 38(4): 389-406. Alcorn, J. B. (1990). Indigenous Agroforestry Strategies Meeting Farmers' Needs. Alternatives to Deforestation: Steps Toward Sustainable Use of the Amazon Rain Forest. A. B. Anderson editor. New York, Colombia University Press: 141-151. Altieri, M. A. (1989). "Rethinking Crop Genetic Resource Conservation: A View From the South." Conservation Biology 3(1): 77-79. Altieri, M. A. (1995). Agroecology: The Science of Sustainable Agriculture. Boulder, Westview Press. Altieri, M. A., M. K. Anderson, and L. C. Merrick (1987). "Peasant Agriculture and the Conservation of Crop and Wild Plant Resources." Conservation Biology 1(1): 49-58. Altieri, M. A., and L. C. Merrick (1987). "In Situ Conservation of Crop Genetic Resources through Maintenance of Traditional Farming Systems." Economic Botany 4(1): 86-96. Anderson, A. B., P. Magee, A. Gely, and M. A. G. Jardim (1995). "Forest Management Patterns in the Floodplain of the Amazon Estuary." Conservation Biology 9(1): 47-61. 186

PAGE 198

187 Anderson, A. B., and D. A. Posey (1989). Management of a Tropical Scrub Savanna by the Gorotire Kayap of Brasil. Resource Management in Amazonia: Indigenous and Folk Strategies. D. A. Posey and W. Bale editors. New York, The New York Botanical Garden. 7: 159-173. Araujo-Lima, C. and M. Goulding (1997). So Fruitful a Fish: Ecology, Conservation, and Aquaculture of the Amazons Tambaqui. New York, Columbia University Press Ayres, J. M. (1995). As Matas de Vrzea do Mamirau, Sociedade Civil Mamirau. Bale, W. (1989). The Culture of Amazonian Forests. Resource Management in Amazonia: Indigenous and Folk Strategies. D. A. Posey and W. Bale editors. New York, The New York Botanical Garden. 7. Bale, W. (1994). Footprints of the Forest: Ka'apor Ethnobotony the Historical Ecology of Plant Utilization by an Amazonian People. New York, Columbia University Press. Beckerman, S. (1987). Swidden in Amazonia and the Amazon Rim. Comparative Farming Systems. S. Brush editor. Guilford, CT, Guilford Press. Behrens, C. A., M. G. Baksh, and M. Mothes (1994). "A Regional Analysis of Bari Land Use Intensification and Its Impact on Landscape Heterogeneity." Human Ecology 22(3): 279-316. Bergman, R. W. (1980). Amazon Economics: The Simplicity of Shipibo Indian Wealth, Department of Geography, Syracuse University. Boserup, E. (1965). The Conditions of Agricultural Growth; the economics of agrarian change under population pressure. London, G. Allen & Unwin. Brookfield, H. (2001). Exploring Agrodiversity. New York, Columbia University Press. Brucher, H. (1989). Useful Plants of Neotropical Origin. Berlin, Springer-Verlag. Brush, S. B. (1989). "Rethinking Crop Genetic Resource Conservation." Conservation Biology 3(1): 19-29. Carneiro, R. L. (1956). Slash-and-burn Agriculture: A Closer Look at its Implications for Settlement Patterns. Men and Cultures. A. F. C. Wallace editor. Philadelphia, University of Pennsylvania Press. Carneiro, R. L. (1968). Slash-and-burn Cultivation among the Kuikuro and its implications for cultural development in the Amazon Basin. Man in Adaptation: The Cultural Present. Y. A. Cohen editor. Chicago, Aldine Publishing Company: 131-145. Carneiro, R. L. (1970). "A Theory of the Origin of the State." Science 169: 733-738.

PAGE 199

188 Carneiro, R. L. (1983). The Cultivation of Manioc Among the Kuikuru of the Upper Xingu. Adaptive Responses of Native Amazonians. R. B. Hames and W. T. Vickers editors. New York, Academic Press. Carvajal, de (1934). The Discovery of the Amazon. New York, American Geographical Society. Cavalcante, P. B. (1974). Frutas Comestveis da Amazonia II. Belem, Museu Goeldi. Chang, J. (1977). "Tropical Agriculture: Crop Diversity and Crop Yields." Economic Geography 53(3): 241-254. Chibnik, M. (1994). Risky Rivers: The Economics and Politics of Floodplain Farming in Amazonia. Tucson, University of Arizona Press. Cid, P. (1978). Plantas Medicinais e Ervas Feiticeiras da Amazonia. Sao Paulo, Atlantis. Clark, J. G. D., Ed. (1952). Prehistoric Europe: The Economic Basis. London, Methuen & Co. Ltd. Clawson, D. L. (1985). "Harvest Security and Intraspecific Diversity in Traditional Tropical Agriculture." Economic Botany 39(1): 56-67. Colfer, C. J. P. (1997). Beyond Slash and Burn: Building On Indigenous Management of Borneo's Tropical Rain Forests. New York, The New York Botanical Garden. Conklin, H. C. (1954). "An Ethnoecological Approach to Shifting Cultivation." Transactions of the New York Academy of Sciences, Series II 17(2): 133-142. Conklin, H. C. (1961). "The Study of Shifting Cultivation." Current Anthropology 2(1): 27-61. Coomes, O. T., and G. J. Burt (1997). "Indigenous Market Oriented Agroforestry: Dissecting Local Diversity in Western Amazonia." Agroforestry Systems 37: 27-44. Dagon, R. R. (1967). Current Agricultural Practices Among the Waiwai. Montreal, Department of Geography, McGill University. Damania, A. B. (1994). "In Situ Conservation of Biodiversity of Wild Progenitors of Cereal Crops in the Near East." Biodiversity Letters 2: 56-60. Denevan, W. M. (1970). "Aboriginal Drained-Field Cultivation in the Americas." Science 169: 647-654. Denevan, W. M. (1971). "Campa Subsistence in the Gran Pajonal, Eastern Peru." The Geographical Review 61(4): 496-518.

PAGE 200

189 Denevan, W. M., Ed. (1976a). The Native Population of the Americas in 1492. Madison, University of Wisconsin Press. Denevan, W. M. (1976b). The Aboriginal Population of Amazonia. The Native Population of the Americsa in 1492. W. M. Denevan editor. Madison, University of Wisconsin Press. Denevan, W. M. (1980). Swiddens and Cattle Versus Forest: The Imminent Demise of the Amazon Rain Forest Reexamined. Where Have All the Flowers Gone? Deforestation in the Third World. Williamsburg, Department of Anthropology, College of William and Mary. Denevan, W. M. (1984). Ecological Heterogeneity and Horizontal Zonation of Agriculture in the Amazon Floodplain. Frontier Expansion in Amazonia. M. Schmink and C. H. Woods editors. Gainesville, University of Florida Press: 311-336. Denevan, W. M. (1992). "The Pristine Myth: The Landscape of the Americas in 1492." Annals of the Association of American Geographers 82(3): 369-385. Denevan, W. M. (1996). "A Bluff Model of Riverine Settlement in Prehistoric Amazonia." Annals of the Association of American Geographers 86(4): 654-681. Denevan, W. M., and C. Padoch, Ed. (1987). Swidden-Fallow Agroforestry in the Peruvian Amazon. Advances in Economic Botany. New York, The New York Botanical Garden. Eden, M. J., and A. Andrade (1987). "Ecological aspects of swidden cultivation among the Andoke and Witoto Indians of the Columbian Amazon." Human Ecology 23(3): 335-356. Ericson, C. (2000). An Artificial Landscape-scale Fishery in the Bolivian Amazon. Nature:408: 190-193. Ericson, C. (2001). Pre-Columbian Roads of the Amazon. Expedition 43(2): 21-30. Food and Agriculture Organization (1994). Land degradation in south Asia: Its severity, causes and effects upon the people. Food and Agriculture Organization of the United Nations. Rome. Fearnside, P. M. (1987). "Rethinking Continuous Cultivation in Amazonia." BioScience 37(3): 209-214. Fearnside, P. M. (1988). "Jri at Age 19: Lessons for Brazil's Silvicultural Plans at Carajas." Interciencia 13(1): 12-24.

PAGE 201

190 Fearnside, P. M. (1990). Predominant Land Uses in Brazilian Amazonia. Alternatives to Deforestation: Steps toward the Sustainable Use of the Amazon Rain Forest. New York, Columbia University Press Fearnside, P. M. (1995). "A Conservation Gap Analysis of Brazil's Amazonian Vegetation." Conservation Biology 9(5): 1134-1147. Fearnside, P. M., and J. M. Rankin (1985). "Jri Revisited: Changes and the Outlook for Sustainability in Amazonia's Largest Silvicultural Estate." Interciencia 10(3): 121-128. Fernandes, F., C. P. Luft, F. M. Guimaraes (1952). Dicionario Brasileiro Globo. Sao Paulo, Globo. Finley, J. P. and W. Churchill (1913). The Subanu: Studies of a Sub-Visayan Mountain Folk of Mindanao. Washington D. C., Carnegie Institution. Frechione, J., D. A. Posey and L. F. da Silva (1989). The Perception of Ecological Zones and Natural Resources in the Brazilian Amazon: An Ethnoecology of Lake Coari. Resource Management in Amazonia: Indigenous and Folk Strategies. D. A. Posey and W. Balee, editors. New York, New York Botanical Gardens. 7: 260-282. Gajaseni, J. and N. Gajaseni (1999). "Ecological rationalities of the traditional homegarden system in the Chao Phraya Basin, Thailand." Agroforestry Systems 46: 3-23. Galvo, E. (1952). The Religion of an Amazon Community: a study in culture change, Columbia University. Gascon, C. and T. E. Lovejoy (1998). "Ecological impacts of forest fragmentation in central Amazonia." Zoology 101: 273-280. Gentry, A. H. (1993). A Field Guide to the Woody Plants of Northwest South America (Colombia, Ecuador, Peru). Chicago, The University of Chicago Press. Gliessman, S. R., R. Garcia E., and M. Amador A. (1981). "The Ecological Basis for the Application of Traditional Agricultural Technology in the Management of Tropical Agro-Ecosystems." Agro-Ecosystems 7: 173-185. Goulding, M. (1999). Introduction to the Section 1: Fish and Fisheries. Vrzea: diversity, development, and conservation of Amazonia's whitewater floodplains. J. M. Ayres, C. Padoch, M. Pinedo-Vasquez, and A. Henderson editors. New York, The New York Botanical Garden Press: 3-6. Goulding, M., N. J. H. Smith and D. J. Mahar (1996). Floods of Fortune: Ecology and Economy Along the Amazon. New York, Columbia University Press.

PAGE 202

191 Gregor, T. (1977). Mehinaku: The Drama of Daily Life in a Brazilian Indian Village. Chicago, University of Chicago Press. Hammond, D. S., P.M. Dolman, and A.R. Watkinson (1995). "Modern Ticuna SwiddenFallow Management in the Colombian Amazon: Ecologically Integrating Market Strategies and Subsistance-Driven Economies?" Human Ecology 23(3): 335-356. Harlan, J. R. (1975a). "Our Vanishing Genetic Resources." Science 188: 618-621. Harlan, J. R. (1975b). Crops and Man. Madison, American Society of Agronomy and Crop Science Society of America. Harlan, J. R. (1976). "Genetic Resources in Wild Relatives of Crops." Science 16: 329-333. Heckenberger, M. J. (1996). War and Peace in the Shadow of Empire: Sociopolitical Change in the Upper Xingu of Southeastern Amazonia, A.D. 1400-2000. Doctoral dissertation. Department of Anthropology. University of Pittsburgh. Heckenberger, M. J. (1998). "Manioc Agriculture and Sedentism in Amazonia: the Upper Xingu Example." Antiquity 72: 633-48. Heckenberger, M. J., A. Kuikuro, U. T. Kuikuro, J. C. Russell, M. Schmidt, C. Fausto, and B. Franchetto (2003). "Amazonia 1492: Pristine Forest or Cultural Parkland?" Science 301: 1710-1714. Heckenberger, M. J., J. B. Petersen, and E. G. Neves (1999). "Village Size and Permanence in Amazonia: Two Archaeological Examples From Brazil." Latin American Antiquity 10(4): 353-376. Henrich, J. (1997). "Market Incorporation, Agricultural Change, and Sustainability Among the Machiguenga Indians of the Peruvian Amazon." Human Ecology 25(2): 319-351. Hiraoka, M. (1985a). "Floodplain Farming in the Peruvian Amazon." Geographical Review of Japan 58 (Ser. B)(1): 1-23. Hiraoka, M. (1985b). "Mestizo Subsistence in Riparian Amazonia." National Geographic Research Spring: 236-246. Hiraoka, M. (1986). "Zonation of Mestizo Riverine Farming Systems in Northeast Peru." National Geographic Research 2(3): 354-371. Hiraoka, M. (1992). Caboclo and Ribereno Resource Management in Amazonia: A Review. Conservation of Neotropical Rainforests: Working from Traditional Resource Use. K. H. Redford and C. Padoch editors. New York, Columbia University Press.

PAGE 203

192 Howard, W. J., J. M. Ayres, D. Lima-Ayres, and G. Armstrong (1995). "Mamiraua: a case study of biodiversity conservation involving local people." Commonwealth Forestry Review 74(1): 76-79. The Holocene 7: 223-228. IBGE (1998). Base de Informaes Municipais, Instituto Brasileiro de Geographica e Statstica, So Paulo. Inuma, J. C. (1999). Caraterizaco das Praticas de Manejo dos Sistemas Agroflorestais em Quintais pelos productores na Reserva de Desenvolvimento Sustentvel de Mamirau, Tef (AM). Unpublished thesis. Manaus, INPA/UA. Johnson, A. (1983). Machiguenga Gardens. Adaptive Responses of Native Amazonians. R. B. H. a. W. T. Vickers. New York, Academic Press: 29-63. Jong, W. de (1996). "Swidden-Fallow Agroforestry in Amazonia: Diversity at a Close Distance." Agroforestry Systems 34: 277-290. Junk, W. J. (1982). "Amazonian Floodplains: Their Ecology, Present and Potential Use." Rev. Hydrobiol. Trop. 15(4): 285-301. Junk, W. J., J. J. Ohly, M. T. F. Piedade and M. G. M. Soares, Ed. (2000). The Central Amazon Floodplain: Actual Use and Options for a Sustainable Management. Leiden, Backhuys Publishers. Kleinman, P. J. A., D. Pimentel, R. B. Bryant (1995). "The Ecological Sustainability of Slash-and-Burn Agriculture." Agriculture, Ecosystems and Environment 52: 235-249. Lageras, P. (1996). "Farming and Forest Dynamics in an Agriculturally Marginal Area of Southern Sweden, 5000 BC to Present: A Palynological Study of Lake Avegol in the Smaland Uplands." The Holocene 6: 301-314. Lanley, J. P. (1982). "Tropical Forest Resources." Food and Agriculture Organization of the United Nations, Rome, Forestry Paper 30. Lathrap, D. (1968). Aboriginal Occupation and Changes in River Channel on the Central Ucayali, Peru. American Antiquity 33(1):62-79. Lathrap, D. (1970). The Upper Amazon. New York, Praeger. Lavelle, P. (1987). "Letters." BioScience 37(9): 639. Le Cointe, P. (1947). Amazonia Brasileira III: Arvores e Plantas Uteis. Sao Paulo, Companhia Editora Nacional. Lehtonen, H. and P. Huttunen (1997). "History of Forest Fires in Eastern Finland from the Fifteenth Century AD: The Possible Effects of Slash-and-burn Cultivation."

PAGE 204

193 Lima-Ayres, D. M. (1992). The Social Category Caboclo: History, Social Organization, Identity and Outsider's Social Classification or the Rural Population of an Amazonian Region (The Middle Solimoes). Unpublished dissertation. Social Anthropology, University of Cambridge. Lima, D. M. (1999). Equity, Sustainable Development, and Biodiversity Preservation: Some Questions about Ecological Partnerships in the Brazilian Amazon. Varzea: Diversity, Development, and Conservation of Amazonia's Whitewater Floodplains. C. Padoch, J. M. Ayres, M. Pinedo-Vasquez, A. Henderson, editors. New York, The New York Botanical Garden Press. 13: 247-263. Lima, R. M. B. (1994). Descrico, Composico e Manejo dos Cultivos Mixtos de Quintal na Vrzea da "Costa do Caldero", Iranduba, AM. Manaus, INPA/UA. Lopinot, N. H. and W. I. Woods (1993). Wood Overexploitation and the Collapse of Cahokia. Foraging and Farming in the Eastern Woodlands. C. M. Scarry. Gainesville, University of Florida Press: 206-231. Matlock (1997). "Four Centuries of Forest Clearance and Regeneration in the Hinterland of a Large City." Journal of Biogeography 24: 281-295. Meggars, B. J. (1992). Prehistoric Population Density in the Amazon Basin. Disease and Demography in the Americas. J. W. V. a. D. H. Ubelaker. Washington D. C., Smithsonian Institution Press. Milliken, B. (1992). "Tropical deforestation, land degradation and society." Latin American Perspectives 72(19): 45-72. Moran, E. F. (1974). The Adaptive System of the Amazonian Caboclo. Man in the Amazon. C. Wagley. Gainesville, The University Presses of Florida: 136-159. Moran, E. F. (1984). Colonization in the Transamazon and Rondonia. Frontier Expansion in Amazonia. M. S. a. C. H. Wood. Gainesville, University of Florida Press: 285-303. Moran, E. F. (1993). Through Amazonian Eyes: The Human Ecology of Amazonian Populations. Iowa City, University of Iowa Press. Nabhan, G. P. (1985). "Native American Crop Diversity, Genetic Resource Conservation, and the Policy of Neglect." Agriculture and Human Values 2: 14-17. National Academy of Sciences (1972). Genetic Vulnerability of Major Crops. Washington D. C., National Academy of Sciences. National Research Council (1982). Ecological Aspects of Development in the Humid Tropics. National Research Council, Washington D. C., National Academy Press.

PAGE 205

194 National Research Council (1993). Sustainable Agriculture and the Environment in the Humid Tropics. National Research Council, Washington D. C., National Academy Press. Resource Management in Amazonia: Indigenous and Folk Strategies. D. A. Posey and W. Balee. New York, The New York Botanical Garden. 7: 249-259. Oliveira, L. A., F. W. Moreira, N. P. Falcao, and V. S. G. Pinto (2000). Floodplain Soils of Central Amazonia: Chemical and Physical Characteristics and Agricultural Sustainability. The Central Amazon Floodplain: Actual Use and Options for a Sustainable Management. W. J. Junk, J. J. Ohly, M. T. F. Piedade and M. G. M. Soares, editors. Leiden, Backhuys Publishers. Olsen, K. M., and B. A. Schaal (1999). "Evidence on the Origin of Cassava: Phylogeography of Maniot Esculenta." Proceedings of the National Academy of Sciences 96: 5586-5591. Orejuela, J. E. (1992). Traditional Productive Systems of the Awa (Cuaiquer) Indians of Southwestern Colombia and Neighboring Ecuador. Conservation of Neotropical Forests: Working from Traditional Resource Use. K. H. Redford and C. Padoch. New York, Columbia University Press: 58-81. Padoch, C. and W. Jong (1992). Diversity, Variation, and Change in Ribereno Agriculture. Conservation of Neotropical Forests. K. Redford and C. Padoch, editors. New York, Columbia University Press. Padoch, C. and W. Jong (1991). "The house gardens of Santa Rosa: diversity and variability in an Amazonian agricultural system." Economic Botany 45(2): 166-175. Padoch, C., J. C. Inuma, and W. Jong, and J. Unruh (1985). "Amazonian Agroforestry: a market-oriented system in Peru." Agroforestry Systems 3: 47-58. Padoch, C., M. Pinedo-Vasquez, and J. C. Inuma (1996). Atividades Produtivas e de Manejo, para a Conservaco de Recursos e Incremento da Renda Familiar, na Reserva de Mamirau. Tef, Instituto de Desenvolvimento Sustentvel Mamirau. Padoch, C. and M. Pinedo-Vasquez. (1999). Farming Above the Flood in the Varzea of Amap: Some Preliminary Results of the Projeto Vrzea. Vrzea: Diversity, Development, and Conservation of Amazonia's Whitewater Floodplains. J. M. Ayres, C. Padoch, M. Pinedo-Vasquez and A. Henderson, editors. New York, New York Botanical Garden Press. 13. Parker, E. (1985). The Amazon Caboclo: An Introduction and Overview. The Amazon Caboclo: Historical and Contemporary Perspectives. N. A. V. H. Sutlive, M. D. Zamora, V. Kerns. Williamsburg VA, Dept. of Anthropology, College of William and Mary. Parker, E. P. (1989). A Neglected Human Resource in Amazonia: The Amazon Caboclo.

PAGE 206

195 Pimentel, D., C. Wilson, C. McCullum, R. Huang, P. Dwen, J. Flack, Q. Tran, T. Saltman, and B. Cliff (1997). "Economic and Environmental Benefits of Biodiversity." BioScience 47(11): 747-757. Sons, Inc. Pinedo-Vasquez, M., C. Padoch, and J. C. Inuma (1996). Identifying and Understanding Agricultural, Agroforestry and Forest Management Systems and Techniques Practiced in Mamirau. Tef, Sociedade Civil Mamirau. Pinedo-Vasquez, M., C. Padoch, and J. C. Inuma (1999). Change in the Agricultural Patterns of Communities and Households of the Mamirau Sustainable Development Reserve. Tef, Sociedade Civil Mamirau. Plucknett, D. L., N. J. H. Smith, J. T. Williams, N. Murthi Anishetty (1983). "Crop Germplasm Conservation and Developing Countries." Science 220: 163-169. Pope, K. O., M. E. D. Pohl, J. G. Jones, D. L. Lentz, C. von Nagy, F. J. Vega, I. R. Quitmyer (2001). "Origin and Environmental Setting of Ancient Agriculture in the Lowlands of Mesoamerica." Science 292: 1370-1373. Popenoe, W. (1920). Manual of Tropical and Subtropical Fruits. New York, Macmillan. Porro (1994). Social Organization and Political Power in the Amazon Floodplain. Amazonian Indians from Prehistory to the Present. A. Roosevelt. Tucson, University of Arizona Press: 79-94. Posey, D. A. (1981). Indigenous ecological knowledge and development of the Amazon. The Dilemma of Amazonian Development. E. F. Moran, editor. Boulder, Westview: 225-250. Posey, D. A. (1985). Native and indigenous guidelines for new Amazonian development strategies: Understanding biodiversity through ethnoecology. Change in the Amazon. J. Hemming editor. Manchester, Manchester University Press: 156-181. Posey, D. A. (1993). Indigenous knowledge in the conservation and use of world forests. World Forests for the Future: Their Use and Conservation. K. R. a. G. M. Woodwell. Binghamton, Yale University Press: 156. Pratap, A. (2000). The Hoe and the Axe: An Ethnohistory of Shifting Cultivation in Eastern India. New Delhi, Oxford University Press. Purseglove, J. W. (1968a). Tropical Crops: Dicotyledons 1. New York, John Wiley & Sons, Inc. Purseglove, J. W. (1968b). Tropical Crops: Dicotyledons 2. New York, John Wiley & Sons, Inc. Purseglove, J. W. (1972a). Tropical Crops: Monocotyledons 1. New York, John Wiley &

PAGE 207

196 Purseglove, J. W. (1972b). Tropical Crops: Monocotyledons 2. New York, John Wiley & Sons, Inc. Raffles, H. (1999). Exploring the Anthropogenic Amazon: Esturine Landscape Floodplain of the Upper Brazilian Amazon. Unpublished dissertation. Anthropology, Indiana University. Transformations in Amapa, Brazil. Varzea: Diversity, Development, and Conservation of Amazonia's Whitewater Floodplains. Padoch, C., J. M. Ayres, M. Pinedo-Vasquez, A. Henderson editors. New York. The New York Botanical Garden Press. Pp. 355-370. Ribeiro, J. E. L. de S., M. J. G. Hopkins, A. Vicentini, C. A. Sothers, M. A. da S. Costa, J. M. de Brito, M. A. D. de Souza, L. H. P. Martins, L. G. Lohmann, P. A. C. L. Assunco, E. da C. Pereira, C. F. da Silva, M. R. Mesquita, L. C. Procopio (1999). Flora da Reserva Ducke: Guia de Identificacao das Plantas Vasculares de Uma Floresta de Terra-firme na Amazonia Central. Manaus, INPA-DFID. Roosevelt, A. C. (1980). Parmana: Prehistoric Maize and Manioc Subsistence along the Amazon and Orinoco. New York, Academic Press. Roosevelt, A. C., M. L. Costa, C. L. Machado, M. Michab, N. Mercier, H. Valladas, J. Feathers, W. Barnett, M. I. da Silveira, A. Henderson, J. Silva, B. Chernoff, D. S. Reese, J. A. Holman, N. Toth, and K. Schick (1996). "Paleoindian Cave Dwellers in the Amazon: The Peopling of the Americas." Science 272(5260): 373-384. Roosevelt, A. C., R. A. Housley, M. Imazio da Silveira, S. Maranca, and R. Johnson (1991). "Eighth Millennium Pottery from a Prehistoric Shell Midden in the Brazilian Amazon." Science 254(1621-1624). Ross, E. B. (1978). "The Evolution of the Amazon Peasantry." Journal of Latin American Studies 10(2): 193-218. Sauer, C. O. (1952). Agricultural Origins and Dispersals: The Domestication of Animals and Foodstuffs. Cambridge, MIT Press. Schmink, M., K. H. Redford, and C. Padoch (1992). Traditional Peoples and the Biosphere: Framing the Issues and Defining the Terms. Conservation of Neotropical Forests: Working form Traditional Resource Use. K. H. R. a. C. Padoch. New York, Columbia University Press: 3-13. Sociedade Civil Mamirau (1996). Mamirau Management Plan. Tef, Sociedade Civil Mamirau. Serro, E. A. S. and A. K. O. Homma (1993). Brazil. Sustainable Agriculture and the Environment in the Humid Tropics. National Research Council. Washington D.C., National Academy Press: 265-351. Shorr, N. (1999). Agricultural Intensification in a Large Tikuna Community on the

PAGE 208

197 Smith, N. J. H. (1980). "Anthrosols and Human Carrying Capacity in Amazonia." Annals of the Association of American Geographers 70(4): 553-566. Smith, N. J. H. (1981). "Colonization Lessons from a Tropical Forest." Science 214: 755Department of Geography. University of Florida. 761. Smith, N. J. H. (1996a). Effects of Land-Use Systems on the Use and Conservation of Biodiversity. Biodiversity and Agricultural Intensification: Partners for Development and Conservation. Washington D. C., The World Bank. Smith, N. J. H. (1996b). "Home Gardens as a Springboard for Agroforestry Development in Amazonia." International Tree Crops Journal 9: 11-30. Smith, N. J. H. (1999). The Amazon River Forest: A Natural History of Plants, Animals, and People. New York, Oxford University Press. Smith, N. J. H. (2002). Amazon Sweet Sea: Land, Life, and Water at the Rivers Mouth. Austin. University of Texas Press. Smith, N. J. H., I. C. Falesi, P. de T. Alvim, and E. A. S. Serro (1996). "Agroforestry trajectories among smallholders in the Brazilian Amazon: Innovation and Resiliency in Pioneer and Older Settled Areas." Ecological Economics 18: 15-27. Smith, N. J. H., T. J. Fik, P. de T. Alvim, I. C. Falesi and E. A. S. Serro (1995). "Agroforestry Developments and Potential in the Brazilian Amazon." Land Degradation & Rehabilitation 6: 251-263. Smith, N. J. H., J. T. Williams, D. L. Plucknett, and J. P. Talbot (1992). Tropical Forests And Their Crops. London, Cornell University Press. Srivastava, J. P., N. J. H. Smith and D. A. Forno (1999). Integrating Biodiversity in Agricultural Intensification: Towards Sound Practices. Washington D. C., The World Bank. Stasi, L. C. di, E. M. G. Santos, C. M. dos Santos, C. A. Hiruma (1989). Plantas Medicinais na Amazonia. Sao Paulo, UNESP. Steinberg, M. K. (1998). "Mopan Maya Forest Resources in Southern Belize." The Geographical Review 88(1): 131-137. Sutlive, V. H., N. Altshuler, M. D. Zamora, and V. Kerns, Ed. (1985). The Amazon Caboclo: Historical and Contemporary Perspectives. Studies in Third World Societies. Williamsburg, Department of Anthropology, College of William and Mary. Swales, S. (1999). Dynamics of Land Use and Agricultural Practices on the Uplands and Adjacent Flood Plain in the Lower Amazon. Doctoral dissertation. Gainesville.

PAGE 209

198 Turner, B. L., R. Q. Hanham, and A. V. Portararo (1977). "Population Pressure and Agricultural Intensity." Annals of the Association of American Geographers 67(3): 384-396. Ugent (1986). "Archaeological Manioc (Maniot) from Coastal Peru." Economic Botany 40(1): 78-102. UNESCO (1983). Swidden Cultivation in Asia. Bangkok, UNESCO. Unruh, J. D. (1990). "Iterative Increase of Economic Tree Species in Managed Swidden-Fallows of the Amazon." Agroforestry Systems 11: 175-197. Vasey, D. E. (1979). "Population and Agricultural Intensity in the Humid Tropics." Human Ecology 7(3): 269-283. Vavilov, N. I. (1987). Origin and Geography of Cultivated Plants. Cambridge, Cambridge University Press. Wagley, C. (1953). Amazon Town: A study of Man in the Tropics. Oxford, Oxford University Press. Whitmore, T. C. (1990). An Introduction to Tropical Rainforests. Oxford, Clarendon Press. Woods, W. I. (1987). Maize Agriculture and the Late Prehistoric: A Characterization of Settlement Location Strategies. Emergent Horticultural Economies of the Eastern Woodlands. W. F. Keegan, Center for Archaeological Investigations, Southern Illinois University. Occasional Paper No. 7: 275-293. World Bank (1992). World Development Report: Development and the Environment. Oxford, World Bank. Zandstra, H. G. (1993). Preserving the Options: International Research for Sustainable Agriculture. 13th Agricultural Sector Symposium, Washington D. C., The World Bank.

PAGE 210

BIOGRAPHICAL SKETCH Morgan Schmidt graduated from the University of North Dakota with a B.S. in geography (and a minor in space studies) in 1993. He pursued a wide range of interests as an undergraduate, including physical geography, environmental science, space science, geology, remote sensing, GIS, anthropology, and Chinese language. After graduation, he decided to postpone graduate school in order to travel. His travels eventually took him to the Brazilian Amazon where he decided to pursue a career in research and conservation. In Manaus, Brazil, Morgan worked on several internships and taught English. The last job was at the Biological Dynamics of Forest Fragments Project jointly sponsored by the Smithsonian Institute and the National Institute for Amazon Research. He worked as part of a NASA/LBA team to set up a GIS laboratory and begin to incorporate data into a GIS from the previous 20 years of the project. He left this job to enroll in a Ph.D. program in geography at the University of Florida. After finishing masters research, Morgan became involved in an ethno-archaeology project in the Upper Xingu of southeastern Amazonia, Mato Grosso, Brazil. His dissertation research examines the formation and distribution of anthrosols and how they can contribute to an understanding of prehistoric complex society and land use in the region. 199


Permanent Link: http://ufdc.ufl.edu/UFE0002726/00001

Material Information

Title: Farming and Patterns of Agrobiodiversity on the Amazon Floodplain in the Vicinity of Mamiraua, Amazonas, Brazil
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0002726:00001

Permanent Link: http://ufdc.ufl.edu/UFE0002726/00001

Material Information

Title: Farming and Patterns of Agrobiodiversity on the Amazon Floodplain in the Vicinity of Mamiraua, Amazonas, Brazil
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0002726:00001


This item has the following downloads:


Full Text












FARMING AND PATTERNS OF AGROBIODIVERSITY ON THE AMAZON
FLOODPLAIN IN THE VICINITY OF MAMIRAUA, AMAZONAS, BRAZIL



















By

MORGAN J. SCHMIDT


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA


2003


































Copyright 2003

by

Morgan J. Schmidt
















ACKNOWLEDGMENTS

This research was supported by a Charles Wagley Research Fellowship through the

Center for Latin American Studies and Tropical Conservation and Development Program

at the University of Florida. Graduate studies were supported by a Graduate Alumni

Fellowship from the University of Florida through the Department of Geography. The

research was further supported by the Mamiraua Sustainable Development Institute.

I would especially like to thank Jomber Chota Inuma who introduced me to the

Mamiraua reserve and contributed much encouragement, advice, and support during the

research. I am grateful for the hospitality of Jomber's family, Dina and Ortila, who gave

me a home in Tef6. I would also like to thank Niele Peralta Bezerra, Jose Marcio Ayres,

Edila Moura, and Ana Claudeise Nascimento for their support of the research; and Jose

Elizon Rocha da Silva, for his tireless efforts in the field. I am grateful to all of the other

assistants, informants, farmers, and families (too numerous to name) for their help,

cooperation, and wonderful hospitality.

Finally, I would like to thank my family, whose support made my graduate career

possible; and my committee members (Dr. Nigel Smith, Dr. Hugh Popenoe, and Dr.

Michael Heckenberger) for their guidance, advice, and support.















TABLE OF CONTENTS

page

A C K N O W LE D G M EN T S ................................................................... ......................... iii

LIST O F TA B LES .................. .................. ............................ .................... vi

LIST OF FIGURES .............................................................................. ix

ABSTRACT..... ............................. .... ...................... xi

CHAPTER

1 AGRICULTURAL BIODIVERSITY AND SHIFTING CULTIVATION: THE
RESEARCH IN CONTEXT................................... ............................1

Intro du action ............................... ......................... 1
Our Stake in Agricultural Biodiversity.................................................3
C all for C conservation ...................................................................... .................... 6
Sw hidden M odel .................................................................. ......... .. ......... ......9
W hy Study Sw hidden? ...................................... ............ ........ .................... 12
Swidden Agriculture in Time and Space............................................................14

2 FOCUSING ON THE MIDDLE SOLIMOES REGION....................... ...............18

Prehistoric Agriculture and Settlement on the Amazon River ...................................18
Practical Aspects of Swidden Research in Amazonia..............................................20
A m azon C ab oclo ............... .................................................................................... 27
M iddle Solim 6es R egion .................... ......... ..... ..... ..........................28
Conservation of the Varzea Ecosystem: The Mamiraua and Amana Sustainable
Developm ent Reserves................................. ..............................30
D ata and M methods ....................... ... ..... .................. .. ... .............. ...... 31
Research Considerations of the Flood Season.............................. ....................36
Life on the V rzea ........................ ............... .......................... .................... 38

3 STRATEGIES FOR FARMING ON THE FLOODPLAIN: THE RO9A ...............43

C characteristics of the R o a ................................................................ .................... 45
Farming Praias and Lamas....................... ..... ..........................49
Diversity and Management of Restinga Baixa Fields .............................................51
Restinga Alta: The Preferred Habitat for Planting ....................... ....................55
Crop Losses from Flooding, Pests, and Diseases .................... ......................62


iv










Storage and Loss of Planting Material................................................................67

4 CAPOEIRAS: ENRICHED FORESTS AND MANAGED FALLOWS..................71

R estinga B aixa ................... ... ..................................... ........... .................. 73
R estinga alta .................... ... ........................................ ............... .................. 76
B ananal ............................................................ 78
C apoeira ........................................................ 83
Capoeira Bananal.................................................... ......... ...... ......... 87
Capoeira Cacaual .................... ............ .................. ..... .................... 90
C apoeira Fruteira ............................................................ ............ ........90
F ru tal...... . ............................................................... ........................................9 3
S itio ............................................................9 4

5 DIVERSITY IN HOME GARDENS ................................. .... ..................100

Diversity in Middle Solim6es Varzea Home Gardens ...........................................104
L losses from Flooding ........... ...................................................... .................... 114
Size of H om e G gardens ............ .................................................... .................... 115
Livestock in H om e G ardens ................................. ................... ... ... .............. 117
Terra Firm e Hom e Gardens ............................ ........................ ....................118

6 A MORE STABLE ENVIRONMENT: TERRA FIRME.....................................126

Field and Fallow Management on Terra Firme ....................................................... 130
Agrobiodiversity in Terra Firme Ro9as.............................................................138
Agrobiodiversity in Terra Firm e Capoeiras ............................................................143
O th er N on -R o as ............... ................................................................................. 148

7 LESSONS FOR CONSERVATION OF AGROBIODIVERSITY IN THE
M ID D LE SO LIM O ES ..................... ......... ............................. .................... 150

D distribution of A grobiodiversity ................................... ......................................... 151
Theoretical Im plications ................................................................. .................... 155
D directions for further research.................... .................................................... 158

APPENDIX

A USEFUL SPECIES OCCURRING IN FIELDS AND FALLOWS.....................166

B USEFUL SPECIES OCCURRING IN HOME GARDENS .................................174

LIST O F REFEREN CES ......... ................. ......... ........................................ 186

BIO G RA PH ICA L SK ETCH ......... ...... ........... ............................. ........................ 199
















LIST OF TABLES

Table Page

2-1 Number of home gardens, fields, and fallows visited in 19 communities on
uplands and floodplains in the middle Solim6es region......................................32

3-1 Summ ary of statistics for low levee fields..........................................................54

3-2 The eight most common species remaining in sixteen harvested fields on low
lev e es.................................. ...................................................... .................... 5 5

3-3 Summary of statistics for high levee fields.....................................................56

3-4 The 19 most common useful species found in planted fields on high levees
with corresponding number of harvested fields...................................................59

3-5 Varieties of bitter manioc mandiocaa) growing in 70 fields on high levees..........61

3-6 Varieties of sweet manioc (macaxeira) growing in 70 fields on high levees. .......61

3-7 Varieties of banana growing in 70 fields on high levees ..................................62

3-8 Perennials that were reported lost in fields and fallows. .....................................66

4-1 Summary of statistics for fallows on low levees. ................................................74

4-2 The 25 useful species occurring in 17 capoeiras on low levees ..........................75

4-3 Local designations of work areas, number of areas visited, and their
d description ........................................................................................ 7 7

4-4 Average size by type of capoeira................................ ........................................ 77

4-5 Average number of useful species by type of capoeira. .....................................78

4-6 The 22 useful species occurring in 17 banana groves (bananais)..........................82

4-7 Summary of statistics for banana groves (bananais)............................................82

4-8 Banana varieties occurring in 10 banana groves (bananais)................................82

4-9 Summary of statistics for fallows (capoeiras) on high levees..............................84










4-10 Banana varieties occurring in fallows (capoeiras) on high levees.......................84

4-11 Twenty-five most frequently encountered species and their use in capoeiras
on high levees with a comparison of former rogas and former bananais...............85

4-12 Summary of statistics for banana fallows (capoeira bananal). ............................88

4-13 The 30 useful species found in 21 banana fallows (capoeira bananal) ................88

4-14 Banana varieties in banana fallows (capoeira bananal)...................................... 89

4-15 Summary of statistics for fruit fallows (capoeira fruteiras).................................91

4-16 Banana varieties in six fruit fallows (capoeira fruteiras).....................................91

4-17 The 29 useful species most frequently encountered in 12 fruit fallows. ...............92

4-18 The 25 useful species found in two fruit orchards (frutal)...................................94

4-19 The twenty-four most frequently encountered species and their use in five
sitios. ........................................................ ...................... 99

5-1 Average number of useful species in 163 virzea home gardens by
com m u n ity .............. ............................................................. ............ .. ..... 10 8

5-2 Fruit, nut, and beverage species in virzea home gardens..................................09

5-3 Number of banana varieties in virzea home gardens. .......................................1.13

5-4 Banana varieties in virzea home gardens. ...................................................113

5-5 Home garden plants that were killed during the 1999 high flood on the
varzea .................... ........................ ........................... 114

5-6 Average size of 155 varzea home gardens in 14 communities............................116

5-7 Average size of 30 terra firme home gardens in 3 communities .........................116

5-8 Average number of species in 30 terra firme home gardens. ..............................119

5-9 Number of banana varieties in terra firme home gardens..................................20

5-10 Banana varieties in terra firme hom e gardens......................................................120

6-1 Summary of local land-use designations by terra firme farmers.......................130

6-2 Summary of statistics for rogas on terra firme ..................................................134

6-3 Summary of statistics for capoeiras on terra firme...........................................135









6-4 The 51 useful species encountered in 87 terra firme rocas..................................140

6-5 The 31 varieties of bitter manioc mandiocaa) found in 84 terra firme rocas.......142

6-6 The 5 varieties of sweet manioc (macaxeira) found in 84 terra firme rocas........142

6-7 The 11 banana varieties found in 46 terra firme rocas....................................... 143

6-8 The 16 banana varieties documented in non-roya areas on terra firme and the
frequency in each land-use type........................................................................ 145

6-9 The 68 useful species encountered in fallows on terra firme and the
frequency they were recorded in each land-use type ........................................146

7-1 Summary of species diversity by land use type on the varzea and terra firme
of the M iddle Solim 6es Region ........................................................................ 153

A-1 Useful species occurring in fields and fallows. .................................................167

B-1 Useful species occurring in hom e gardens.........................................................175
















LIST OF FIGURES


Figur a

1-1 Map of the Brazilian Amazon centered on the middle Solim6es region .................3

2-1 Landsat TM satellite image (bands 543) of the study region from 1999 with the
locations of 19 communities included in the survey..................... ....................33

2-2 Seasonal fluctuation in water level on the middle Solim6es, 1992-2000. ...............39

2-3 Floodplain community during low water. .....................................................42

3-1 The bank of the Japura River during low water.....................................................44

3-2 Number of species in 82 planted fields (rogas) on high levees................................58

3-3 Number of species in 31 harvested fields (rogas) on high levees. ...........................58

3-4 Number of manioc varieties in fields on high levees..............................................60

3-5 Number of banana varieties in fields on high levees. ...........................................60

3-6 Bundles of manioc stems placed upright in resting alta field.................................70

3-7 Storage of manioc stems on a raised platform in a high levee field.........................70

4-1 Banana fallow on a high levee on the Solim6es River ..........................................89

5-1 Yard in the virzea located on a paranA connecting the Japura River to the
S olim 6es. ..................................................... ...................... 103

5-2 The number of species in 163 home gardens on the virzea................................106

5-3 The number of species in 30 home gardens on terra firme. ..................................106

5-4 Scatter diagram showing number of species in 163 virzea home gardens in
relation to size in square meters ....... ...........................................116

5-5 Backyard with animals in a Solim6es River community.......................................121

5-6 Maromba constructed to provide shelter for animals during the flood................122










5-7 Cattle being fed the floating canarana grass during the flood season ..................123

5-8 Terra firm e hom e garden ............................................................ ... ............... 124

5-9 Well developed home gardens grade into diverse fruit orchards.........................125

6-1 Length of the previous fallow for 39 fields (ro9as) on terra firme.........................135

6-2 Age of 63 fallows (capoeiras). ...........................................................135

6-3 Number of consecutive crops planted in 51 fields (ro9as) on terra firme ............135

6-4 Number of consecutive crops previously planted in 25 fallows (capoeiras) on
terra firm e. .......................................................... ........... ... ... ..... ....... 136

6-5 Number of previous fallows for 66 fields (ro9as). ...............................................136

6-6 Number of cropping cycles for 35 fallows (capoeiras) on terra firme.................136

6-7 Eight-month-old terra firme field (roca) cleared from old-growth forest ............137

6-8 Ro9a developing into fruit fallow (capoeira fruteira) ..........................................137

6-9 Number of useful species in 87 fields (ro9as) on terra firme.................................141

6-10 Number of manioc varieties in 84 fields (ro9as) on terra firme.............................141

6-11 Number of banana varieties in 46 fields (ro9as) on terra firme..............................141

6-12 Number of useful species in 43 fallows (capoeiras) on terra firme........................144

6-13 Number of banana varieties in 17 fallows (capoeiras) on terra firme..................144

7-1 M ap of fields and home gardens of Betania..........................................................162

7-2 Preliminary classification of satellite image from 7-1 ........................................163

7-3 Map of fields and home gardens of Sao Paulo do Corai...................................... 164

7-4 Participatory mapping of levees with fields and fallows in Sao Paulo
d o C ora9 i. .................................................... ...................... 16 5
















Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

FARMING AND PATTERNS OF AGROBIODIVERSITY ON THE AMAZON
FLOODPLAIN IN THE VICINITY OF MAMIRAUA, AMAZONAS, BRAZIL

By

Morgan J. Schmidt

December 2003

Chair: Nigel J. H. Smith
Major Department: Geography

An examination of agrobiodiversity on the flood plains and adjacent uplands along

a stretch of the Amazon river (Middle Solim6es) illustrates the enormous variability

between different land use strategies. Farmers in the region use five principal habitat

types for agriculture; beaches, mudflats, low levees, high levees, and uplands. High

levees and uplands contain the greatest diversity in both land use types and useful plants.

The research revealed 24 designations of land use by local residents when divided along

lines of habitat, current status (field or fallow), and management. Home gardens are an

important repository of agrobiodiversity because they contain almost the entire spectrum

of useful species including many not found in other land use types and almost every

home has one.















CHAPTER 1
AGRICULTURAL BIODIVERSITY AND SHIFTING CULTIVATION: THE
RESEARCH IN CONTEXT

Introduction

Agricultural biodiversity is paramount in meeting the needs of Earth's growing

human population. Wild relatives and traditional landraces of today's major crops hold

the key to developing new, more robust, and better-adapted varieties that allow greater

productivity in selected environments and better resistance to pests and diseases (Harlan

1975a, 1976; Vavilov 1987). Little-known crops are a major resource that can contribute

new foods and other products with higher productive potential, greater efficiency, and

increased nutritional value; or simply add welcome variation to our diet. A newly

developed crop can strengthen local economies and create major industries. Maintaining

diversity means keeping our options open to develop new crops and varieties or improve

existing ones. This will help us meet the challenge of adapting agriculture to changing

environments; and developing efficient, sustainable agricultural ecosystems. Hundreds

of fruits, nuts, tubers, seeds, medicinals, ornamentals, timber, and other plants are

virtually unknown outside of the regions where they are grown. These plants hold great

untapped potential; and many are in danger of being lost to us forever with increased

alteration of both natural environments and traditional agricultural ecosystems (National

Research Council 1993).

In recent years, researchers and policy makers have increasingly called attention to

the importance of studying the agricultural systems of traditional farmers (Posey 1981,









1985; Smith 1996a; Srivastava et al. 1999). It is thought that development plans can

benefit from traditional farmers' intimate knowledge of the environment and time-tested

farming techniques. Traditional systems of shifting cultivation in the tropics almost

invariably contain high agricultural biodiversity; and are well-adapted to local

environments. The farming methods of the caboclo population (mixed descendants of

indigenous peoples and immigrants in the Amazon region) in the middle Solim6es region

of Amazonia are no exception. The Solim6es river is the section of the Amazon river in

Brazil stretching from the Peruvian border to the confluence of the Negro river. The

middle Solim6es region is approximately centered on the confluence with the JapurA

river, a northern tributary coming from Colombia (Figure 1-1). Caboclos utilize a

number of land-use systems and manage a myriad of useful plants in the diverse

landscape along a stretch of the greatest river in the world. Our goal was to document the

distribution and variability of agricultural biodiversity in the region as it relates to land

management and habitat.

Chapter 1 gives a background on agricultural biodiversity and shifting cultivation

and their significance. Chapter 2 describes practical aspects of studying traditional

agriculture in the Amazon region; the research area (including its people, environment,

and conservation reserves); data and methods; and life on the virzea (Amazon

floodplain). Chapters 3 through 6 describe the results of the project. The chapters are

organized by land-use and habitat. Chapter 3 describes the roya or manioc field and

chapter 4, the capoeira or fallow, both on the virzea. Chapter 5 focuses on the diversity

of home gardens in both the virzea and terra firme (uplands). Chapter 6 summarizes









findings from the survey on terra firme agriculture.

learned and gives directions for future research.



Equator


Chapter 7 summarizes what was


Figure 1-1. Map of the Brazilian Amazon centered on the middle Solim6es region. The
box delimits the study area.

Our Stake in Agricultural Biodiversity

Development of agriculture is arguably the most significant occurrence in human

history. This major factor has allowed our population to rise at current rates and we now

must increase food production to keep pace with population growth. Agricultural

biodiversity (agrobiodiversity) is the biological resource that directly and indirectly

contributes to crop and livestock production. It has been and will continue to be

fundamental in our efforts to intensify agriculture (Srivastava et al. 1999).

For thousands of years, humans have been domesticating new crops and expanding

the range of existing ones by borrowing from neighboring groups or taking them along









when they migrate to new regions (Harlan 1975b). Repeatedly, crops that began in

localized areas have proliferated around the globe to become staple foods or the basis of

industry. Hundreds of little-known crops and varieties exist that could someday be

grown on a much larger scale. Tropical agricultural ecosystems (agroecosystems) are

especially diverse in both food and nonfood species; with current and potential uses

including medicines, construction materials, industrial materials, and ornamentals

(Alcom 1984; Denevan and Padoch 1987; Smith et al. 1995). Agroecosystems often

contain plants in the process of domestication (proto-domesticates) from which exciting

new crops may be recruited (Smith 1996b).

Material for crop improvement is obtained from diverse landraces, wild crop

relatives, and even unrelated species (Pimentel et al. 1997; Smith 1999; Smith et al.

1992). Plant breeders use these genetic resources to develop new varieties that are of

greater quality, higher yielding, resistant to insects and disease, tolerable of stress,

amenable to mechanical harvesting, and adapted to different environments (Harlan 1976;

Plucknett et al. 1983). Traditional crops and varieties, proto-domesticates, and wild crop

relatives, are a rapidly shrinking resource as modem agriculture advances, more farmers

enter the market economy, cultural ecological knowledge is lost, and natural habitats are

transformed (Altieri et al. 1987; Harlan 1975a; Nabhan 1985).

The expansion of green revolution technology can have a negative impact on

genetic diversity. In the push for greater productivity, a handful of crops have gained

importance and come to dominate world agricultural production, often resulting in the

loss of traditional crops and varieties (Altieri et al. 1987; Harlan 1975a). For

industrialized farms to be viable, farmers seek the highest yielding varieties available.






5


Commercial seed, developed to be compatible with agrochemicals, gives farmers limited

choices as to which varieties will be planted (Clawson 1985). Mechanization requires

that crops be of uniform size and shape; mechanization also makes intercropping of

multiple species and varieties impractical (Chang 1977). The need for expensive

technology and specialized knowledge has caused modem farms to focus on usually only

one or two crops in order to compete successfully. Market forces dictate which crops

will be grown as farmers cannot afford to spend time and energy on those that will not

fetch the highest prices. The conversion of small, more diverse farms to large, mono-

cropped farms and cattle ranches can eliminate traditional crops, varieties, and proto-

domesticates (Smith 1999). In addition, the expansion of modem agriculture onto new

lands puts populations of many wild crop relatives at risk of extinction, posing a threat to

the crops that we rely on most (Damania 1994; Harlan 1976). All of these factors

together have the effect of homogenizing agriculture around the world, by squeezing out

crops and varieties that are less commercially important. These factors are critical,

however, to the future development of agriculture and to human welfare.

Globalization brings rapid change to many cultures, often resulting in the

disappearance of cultural knowledge of ecosystems, plant uses, and traditional agriculture

(Altieri et al. 1987; Nabhan 1985; Steinberg 1998). This knowledge, acquired through

many generations of experience and experimentation, is as important as genetic diversity

for the development of sustainable agroecosystems (Altieri 1995; Gliessman et al. 1981).

The passing away of traditional knowledge and management practices often goes hand in

hand with vanishing crop diversity.









The loss of agrobiodiversity and traditional knowledge has grave implications for

the safe and equitable future of the growing human population. As global food

consumption increases, productive land is being lost by urbanization and degradation

caused by unsustainable management. This necessitates the development of higher

yielding varieties and more productive and sustainable agricultural systems. We must

constantly contend with new pest and disease outbreaks that can devastate entire crops

(Adams et al. 1971). Awareness is also increasing of the threat to agriculture posed by

changing climates that may cause temperature and rainfall regimes to shift (Zandstra

1993). Our ability to expand food production and adjust to ecological and market

changes stems largely upon the available genetic diversity.

Call for Conservation

During the 1960s, concern for the vulnerability of our maj or crops led to increasing

efforts at collecting and preserving crop germplasm (Harlan 1975a; National Academy of

Sciences 1972; Plucknett et al. 1983). A global network of gene banks was constructed

with facilities for short, medium, and long term storage of seeds and limited maintenance

of vegetatively propagated crops. The importance of preserving genetic diversity was

emphasized by Harlan (1976), "In the future, the need for genetic variability and sources

of resistance shall drive us to a much fuller exploitation of all the genetic resources we

can assemble."

Although the practice of storing germplasm in ex situ gene banks is crucial for the

future of agriculture, there are several recognized shortcomings in fully relying upon this

method alone to conserve valuable genetic resources. First, there is insufficient

representation of the full range of genetic diversity for many crops and a large number of

less important crops are not represented at all (Harlan 1975a, 1976; Plucknett et al 1983).









Second, storage facilities may be vulnerable to natural disasters and political change.

Third, the plants are taken out of the ecosystem where they have continuously co-evolved

with animals, other plants, and environmental stress including rapidly evolving pests and

diseases. For these reasons a number of researchers have argued for the importance of

conserving genetic resources in situ (Altieri and Merrick 1987; Brush 1989). This can be

accomplished if a way is found to preserve the agroecosystems in which they are found

(Altieri 1989).

Tropical agroecosystems often contain a multitude of crops, varieties, proto-

domesticates, and wild crop relatives including plants with a number of non-food uses.

To conserve this resource we must understand where high genetic diversity is maintained

and how management affects the abundance of crops and varieties (Alcorn 1984; Altieri

and Merrick 1987; Smith 1996a). In tropical shifting cultivation (swidden) systems, it is

pertinent to identify which processes of cultivation result in a secondary forest enriched

with useful species, both planted and volunteer. Some useful plants may become more

numerous when they are protected by the farmer during clearing, burning, and weeding

or when they are favored by repeated cycles of cultivation (Bale6 1994). Enriched

fallows and mixed orchards may result from the practice of swidden-fallow agroforestry

where fields are not only left fallow but are planted to a diverse array of useful species

(Colfer 1997; Coomes and Burt 1997; Denevan and Padoch 1987).

For traditional farmers, a diverse crop and variety repertoire has advantages in the

field, on the farm, and in the region by providing security and greater efficiency

(Clawson 1985). Maintaining a heterogeneous agricultural landscape of fields, fallows,

natural areas, and extractive areas benefits families by providing game, fish, fruit,









construction materials, medicines, and other products that contribute to subsistence and

income (Altieri et al. 1987). When several crops and varieties are planted in a field or

diverse landscape, they provide natural barriers that can act to slow the spread of pests

and diseases (Adams et al. 1971) and can result in increased biological insect pest control

(Altieri et al. 1987). However, one drawback is that the forest may act as a source of

disease or pests for some native crops. Farmers are better able to cope with losses or

plunging market prices if they are not dependent on a single crop. They can benefit by

harvesting several crops that mature at different times, spreading their workload and

income more effectively throughout the year and utilizing land more efficiently. A

selection of crops and varieties also allows them to exploit small-scale environmental

variation. A good example of this is the use of different crops by floodplain farmers

according to micro-relief and soil texture (Denevan 1984). Studying the ways that

traditional farmers interact with and manage their environments will enable us to merge

the benefits of traditional agriculture with modem 'green revolution' technology, leading

us down a path toward a sustainable agriculture.

The plant diversity of many traditional agroecosystems remains largely unexplored.

There is a danger of losing the valuable genetic resources and sustainable management

practices evolved over many generations as these agricultural landscapes are transformed.

By documenting this diversity, plans may be devised to conserve it. It is certain that the

more diversity we are able to maintain, the better equipped we will be to confront the

challenges that await us. This research attempts to document the diversity of plants

considered useful by the local population of the middle Solim6es region. The focus is on

the distribution of useful plants within the agricultural ecosystem of the Amazon varzea









(floodplain) and adjacent terra firme (uplands). By understanding how agricultural

practices influence the diversity of useful plants in the landscape, resource managers and

local farmers can design management plans that more effectively conserve genetic

resources.

Swidden Model

For many, the terms slash and burn, swidden, or shifting cultivation carry negative

connotations and conjure images of poverty stricken peasants laying waste to large tracts

of virgin rainforest. This type of farming has often been blamed for the current crisis of

deforestation around the globe, particularly in the species diverse rainforest areas. It has

been described as primitive, haphazard, simple, wasteful, and inefficient by scholars

(Conklin 1954, 1961; Gregor 1977; World Bank 1992) and even outlawed by some

governments (Finley and Churchill 1913; UNESCO 1983).

Finley exemplifies the dominant view of swidden agriculture during the first half of

the twentieth century in his 1913 ethnographical and geographical sketch of the Subanu,

an upland tribe on Mindanao in the Philippines, who practice this type of cultivation

known locally as kaingin. He explains that "agricultural development is seriously

retarded by want of proper methods" and "the lack of efficient labor" (Finley and

Churchill 1913:15) Finley goes on to say, "The kaingin method of farming involves a

great waste of labor and materials and must be eventually interdicted by appropriate laws,

rigidly enforced." Echoes of this pessimistic view of shifting cultivation can still be

heard (Lavelle 1987).

Today, most swidden scholars agree that the system is not inherently unsustainable

(Kleinman et al. 1995). In fact, it is a very effective adaptation to the environment,

especially where soils are infertile. Swidden is a relatively efficient form of agriculture









in terms of food production per unit of labor (Beckerman 1987; Careiro 1968). The

problems begin when human populations, market influences, or circumscription of

territory exert too much pressure on the available land, causing farmers to shorten fallow

periods (Boserup 1965; Careiro 1970; Food and Agriculture Organization 1994; Pratap

2000; Steinberg 1998; Vasey 1979). Shorter fallows may result in lower yields, and soil

degradation. Although interest has grown enormously since Harold Conklin (1954,

1961) brought these issues to the forefront in the middle of the twentieth century,

researchers are only beginning to understand the intricacies of slash and burn.

This ancient form of agriculture is known by many names. In English, most

commonly, the terms shifting, slash and bur, or swidden are combined with cultivation,

agriculture, or horticulture to refer to the general model. In this thesis, I will use these

names interchangeably. A widely accepted definition has been given by Harold Conklin

(1961) as "any continuing agricultural system in which impermanent clearings are

cropped for shorter periods in years than they are fallowed."

The process is fundamentally the same worldwide independent of culture, climate,

or ecosystem. A patch of forest is cut, allowed to dry, and then burned. The elements

that were locked in the living vegetation are then released as ash providing a pulse of

nutrients that are easily utilized by crops. The area is planted for one or more seasons

becoming a productive field. During the cropping period nutrients are depleted from the

system, weeds invade, crop yields and efficiency drop, and the field begins a

metamorphosis back to a state of higher biomass. This transformation may be left to the

hand of nature by simply abandoning the field and allowing the forest to regenerate itself,

or the fallow may be carefully managed by farmers, effectively increasing the









productivity and efficiency of the system (Denevan and Padoch 1987; Eden and Andrade

1987; Hammond et al. 1995). Although swidden systems are usually associated with the

burning of cleared vegetation, there are exceptions in areas of high rainfall and certain

floodplain habitats where fire is not always used (Beckerman 1987; Orejuela 1992;

Pinedo-Vazquez et al. 1996).

The standard model of swidden farming hides infinite variation (Brookfield 2001;

Denevan 1971). Differences include crop configuration, number of species and varieties,

number of consecutive years planted, number of years left in fallow, whether or not it is

burned, whether or not fallows are managed, amount of labor investment, and season of

planting and harvest (Conklin 1954). The reasons behind this variation are also diverse.

They include environmental, economic, demographic, and cultural factors that influence

responses by farmers (Conklin 1954, 1961).

Environmental conditions such as climate, vegetation, soils, altitude, pests and

diseases, and flooding regimes shape the management systems at local and regional

scales. Tropical ecosystems, in particular, can exhibit significant small-scale

heterogeneity that requires numerous adaptations for successful cropping. Economic

considerations including access to land, labor, and capital, the degree of local and

regional market integration, and available transport are extremely relevant to the

decisions a farmer makes such as the amount of land cleared and planted, choice of crops,

time in fallow, and degree of fallow management. Cultural factors control, to a great

extent, the methods used in this agricultural system. Though farmers are often

innovative, they tend to farm the way they are taught by their relatives, neighbors, or

extension workers. Farming techniques may be lost or learned as cultures come into









contact with one another, market opportunities arise, and new technologies and crops are

introduced. Finally, individual preferences play a part in all of these variables. The

knowledge and practices of individuals may vary significantly within a single

community. The enormous diversity encountered in swidden systems around the world

means that generalizations made in the past are becoming obsolete.

Why Study Swidden?

Research on swidden agriculture is valuable from both a theoretical and practical

standpoint. From the theoretical angle, its study holds important clues for our

understanding of agriculture's distant origins and development (Sauer 1952). Although

the traditional forms of shifting cultivation we encounter today may not be a window into

the past, examination of its myriad forms and techniques can provide insights and help

form hypotheses to be tested with the analysis of empirical data. As human populations

and market integration increase, swidden systems are transformed in numerous ways

(Behrens et al. 1994; Henrich 1997; Serrao and Homma 1993). Information on these

changes is valuable in providing us with a better understanding of the mechanisms and

impacts of agricultural intensification (Boserup 1965; Turner et al. 1977; Vasey 1979).

From a practical standpoint, research on the sustainability of swidden systems is

essential in our efforts at rainforest conservation in the twenty-first century (Serrao

1993). Destruction of tropical rainforests around the world is one of the most pressing

environmental issues that we face today (Whitmore 1990). Impact on biodiversity and

the global climate are two effects of deforestation that have relevance for everyone,

regardless of where we live (National Research Council 1993). There are also various

local and regional problems that may result from deforestation including, soil erosion and









degradation, loss of habitat for fish and wildlife (Goulding 1999), rainfall reduction, and

increased sedimentation of rivers and other water bodies.

Shifting cultivation has been cited as one of the main causes of tropical

deforestation in many regions (Food and Agriculture Organization 1994). In a 1982 FAO

study, this form of agriculture was identified as the most important reason for the loss of

tropical forests (Lanley 1982). It was estimated to be responsible for 35 percent of

deforestation in the American tropics, 49 percent in Asia, and 70 percent in Africa.

Because of the species diversity of the rainforest, its threatened situation, and the fact that

it is the environment where shifting cultivation dominates today, the question persists

whether this form of agriculture is compatible with conservation. Can it be practiced in

such a way as to benefit the natural environment by preserving biodiversity or even

increasing it at the landscape scale? There is some indication that this might actually be

the case.

William Bal6e (1989, 1994), in his work among the Ka'apor tribe that practice

swidden gardening in the Eastern Amazon, found that old fallows contain significantly

different species than old-growth forest suggesting an enhancement of biodiversity at

regional scales. How can a system like the Ka'apor's be adapted to areas with higher

population density and greater pressure on the land? Only a better understanding of

contemporary swidden systems can begin to shed light on this question.

Slash and burn is a highly adaptive and sustainable agricultural system when

cropping intensity is not so great as to preclude adequate fallow periods (Kleinman et al.

1995). However, when population increase or territorial circumscription necessitates

raising the cropping intensity and shortening the fallow period, the ecologically sound









system can break down, becoming unsustainable (World Bank 1992). By evaluating the

many forms of shifting cultivation in different environments, we hope to find a way to

use our rainforests rationally and intelligently, thus preserving them for future

generations.

Swidden Agriculture in Time and Space

We do not know when the practice of swidden cultivation first came into use. We

can be fairly confident, however, to suggest that it is among the oldest forms of

agriculture in existence and it is, in that sense, a truly primitive form. Though

conventional wisdom holds that agriculture began in the Fertile Crescent about 10,000

years ago, it is quite likely that it developed much earlier at various times and places

rather than in one single origin event (Harlan 1975; Vavilov 1987). We must not rule out

the tropical forest or tropical floodplain as possible settings for an early beginning of

agriculture (Sauer 1952; Smith 1999).

Several conditions combine to make the tropical moist environment conducive to

the development of plant domestication. The first is the climate, being very amenable to

plant growth with abundant sunlight and rainfall. Second, there is a rich abundance of

species with a large number of potential uses. A third characteristic that makes planting

easier are the soft, friable soils of the forest floor and the annually renewed, nutrient rich

floodplain soils. No matter what environment it was first initiated in, humans eventually

learned that they could kill large trees by girdling to allow sunlight to reach the forest

floor and burn vegetation to provide nutrients needed for plant growth.

Besides being one of the oldest forms of agriculture, swidden is one of the most

widespread forms even today. It has been used on every continent except Antarctica and

thousands of islands around the world. Although it is primitive in the spirit of its great









age and continuity, swidden agriculture has evolved to a high level of diversity and

sophistication in the multitude of cultures and environments where it is utilized.

Innovative farmers from around the world have contributed to the evolutionary

development of swidden giving it an abundance of forms. It is not a static and rigidly

defined set of methods but rather, variations are determined by individual preferences and

cultural history in addition to environmental factors.

Today we think of swidden as primarily a tropical agricultural system. In the past,

however, it was practiced throughout the temperate zone as far north as present day

Canada and Scandinavia (Clark 1952). During the Neolithic, shifting cultivation was

used extensively in Europe and apparently resulted in immense ecological change by

deforesting large areas.

Comparing the average fire interval over time, Lehtonen and Huttunen (1997) show

that the slash-and-bum technique was commonly used for cereal production for about

2000 years in southern Finland, being the most important cultivation technique in the

eastern part of the country until the 1900s. They point out that fires have a great

influence on forest structure and swidden cultivation has greatly influenced the history of

forest fires.

In another study in neighboring southern Sweden, Lageras (1996) examines the

pollen record over the past 7000 years to document vegetation and land-use history. He

determines that the first indication of human impact by forest clearance occurs around

6000 years ago. Pasture has been the principal agricultural land-use in the area beginning

about 5000 years ago and slash-and-bum cultivation was probably introduced to the area

around 1200 AD. In the last century the land has once again become covered by forest









making it easy for us to forget that the landscape was once dominated by human

activities, namely forest clearance and burning. This supports the idea that many areas,

including parts of Amazonia, often considered pristine wilderness are, in fact, the product

of human manipulation (Denevan 1992; Erickson 2000, 2001; Heckenberger 1996, 1998;

Heckenberger et al. 1999, 2003; Raffles 1999; Smith 2002).

Swidden farming was a common practice by prehistoric Amerindians in many

places throughout North America (Lopinot and Woods 1993; Woods 1987). Woods

(1987) examined the ethnographic record to characterize prehistoric settlement patterns

and concluded that many were determined largely by the reliance on maize agriculture.

For native North Americans from the east coast to the Mississippi valley, corn was one of

the most important crops and the ash from burned vegetation was an essential addition of

nutrients to the soil. Because of corn's high demand on nutrients, all but alluvial soils,

replenished annually by floods, had to eventually be put into fallow. Many settlements

were located on high ground near rivers where fields could be planted on the floodplain

as well as uplands. Many fallow fields were present and an infield-outfield system was

often practiced with smaller gardens near homes and larger fields further away. In many

cases, villages would relocate periodically at 10 to 30 year intervals due to exhaustion of

agricultural land and lack of fuelwood. European colonists also used shifting cultivation

up through the 1700s in the United States (Matlock 1997) and Southern Ontario (Clark

1952) resulting in the clearance of much of the eastern forests of North America.

The tropical moist forest is not an environment conducive to the preservation of

ancient crop remains or human artifacts that were often made from wood and other

perishable material (Harris 1972). Because of this fact, along with the paucity of









archaeological research, relatively little is known about the origins of agriculture in the

tropics as compared with Southwest Asia. However, evidence is mounting for the

antiquity of agriculture in the tropics.

Recent research is firmly establishing Mesoamerica as one of the cultural regions

that was an early center of crop domestication. In the lowlands of coastal Mexico, Pope

et al. (2001) found evidence for extensive forest clearing and maize cultivation by 6,800

years ago. It is believed that the Maya civilization was based on some form of shifting

cultivation that allowed a relatively dense population in a humid tropical region (Harris

1972).

Remains of domesticated manioc have been found in archaeological sites on the

coast of Peru dating from 2,800 years ago (Ugent et al. 1986). This, together with recent

phylogeography research that places the origin of cultivated manioc in the southern

border region of the Amazon Basin, suggests that the crop was grown throughout the

American tropics by 3000 years ago (Olsen and Schaal 1999). Swidden agriculture

clearly has an immense history in tropical forest regions around the world.















CHAPTER 2
FOCUSING ON THE MIDDLE SOLIMOES REGION

Prehistoric Agriculture and Settlement on the Amazon River

We know that human populations have been settled along the Amazon for many

thousands of years. Paleoindian remains were found by Anna Roosevelt and colleagues

at a site called Caverna de Pedra Pintada near Monte Alegre on the left bank of the

Amazon downstream from Santarem (Roosevelt et al. 1996). It is a well dated cave site

with cultural material from about 8,500 to 11,200 years ago and rock art thought to be

from the same period because of paint specs found in the stratigraphy. The site is

especially significant because of food remains of tropical forest species that were found

including many edible tree fruits and a variety of fauna showing that the people subsisted

on foods from the rainforest and river and not on big game like other paleoindian groups.

Roosevelt suggests that the many species adapted to disturbance indicates there may have

been some forest clearance by these early inhabitants, however, no remains of crop

species have yet been found.

The river bluffs adjacent to the Amazon floodplain have yielded archaeological

evidence for the earliest pottery yet known in the Americas and confirms specialized

exploitation of river fauna by these early populations (Roosevelt et al. 1991, 1996).

Caverna de Pedra Pintada and an Archaic shell-midden called Taperinha on the opposite

side of the river from Monte Alegre excavated by Anna Roosevelt in 1987, both contain

ceramics with dates over 7000 years ago.









Evidence for manioc agriculture is present in the archaeological record at Caverna

de Pedra Pintada in the form of bowls and thick griddles like those used today for

cooking manioc. The pottery relates to the lowland South American Formative period

from about 4000 to 2000 years ago, thought to represent the diffusion of ceramics and

root horticulture (Roosevelt 1980; Roosevelt et al. 1996). Evidence from various

locations in northern South America indicates that manioc was a staple food for a number

of groups by 3000 years ago (Lathrap 1970).

Ethnohistorical documents describe large populations densely settled along the

main branch of the Amazon River when Europeans first arrived (Carvajal 1934; Porro

1994). Preliminary archaeological evidence lends credence to these reports

(Heckenberger et al. 1999; Smith 1980). Most of the floodplain cultures were completely

destroyed in the first 150 years of European contact. It is likely that these populations

subsisted on a combination of aquatic resources and agriculture, both on the varzea and

terra firme (Roosevelt 1980; Smith 1999). There is much uncertainty and debate as to the

relative proportions of these resources in the subsistence economy of prehistoric riverine

cultures.

Roosevelt (1980) theorized that the introduction of maize in the Orinoco and

Amazon basins acted as a springboard for the development of dense populations. She

reasons that manioc is ill-suited to the floodplain and makes inefficient use of the

relatively fertile alluvial soils. Roosevelt claims that short maturing maize is a more

appropriate crop for the unpredictable bottomlands and the high-protein grain makes

optimal use of the nutrient-rich soils. The examination of contemporary agricultural

systems, resource use, and ecological perceptions on the floodplains and adjacent uplands









can help us to evaluate existing theories and provide insights into the subsistence

economies of prehistoric riverine inhabitants.

Amerindian populations today are a vestige of what they once were (Denevan

1976a). In Amazonia, like elsewhere in the western hemisphere, the population has

fallen drastically since first contact with Europeans in the sixteenth century due to

disease, warfare, and slavery (Denevan 1976b; Heckenberger 1998; Meggers 1992). In

pre-contact times, intensive agriculture, with significant transformation of the landscape,

was well underway in several regions of the Amazon (Denevan 1970; Ericson 2000,

2001; Heckenberger 1996, 1998, 1999, Heckenberger et al. 2003). It is clear that

Amerindian subsistence economies, including swidden cultivation, described in the

ethnographic record are a remnant of their former range of variation (Beckerman 1987).

Research on the swidden systems in use along the Amazon river today can add to our

overall understanding of this subsistence activity in the region.

Practical Aspects of Swidden Research in Amazonia

National governments view the development of their Amazonian territories as

imperative to the growth of their economies and modernization of their countries.

Conservationists hope they will seek sustainable ways to develop natural resources.

Conventional agricultural methods including mechanical clearing and tillage of the soil,

large plantations, and mono-crops have so far not proven successful in Amazon rainforest

areas (Fearside 1987, 1988). These methods often cause soil erosion, degrade

ecosystems, and deplete biodiversity (Gascon and Lovejoy 1998; Milliken 1992; Serrao

et al. 1993). The search continues for forms of development that maintain the integrity of

ecosystems and, at the same time, contribute to markets and raise the standard of living

for local people.









Many researchers have emphasized that indigenous peoples have an important

contribution to make to the question of how to develop the Amazon (Anderson and Posey

1989; Posey 1981,1985,1993; Schmink et al. 1992). Smith (1999) argues that native

knowledge and modem science should be combined when trying to make rational use of

the Amazon's resources. The National Research Council has urged that information on

traditional agricultural systems of the humid tropics be researched and evaluated before

this valuable knowledge is lost (National Research Council 1982).

Since the 1960s, the Brazilian government has implemented an aggressive

development strategy for its Amazon territory. A perceived threat from other nations

who "covet" the Amazon could be ameliorated by a clear Brazilian presence.

Furthermore, the migration of large numbers of people to the vast, "unoccupied" frontier

region was seen as a safety valve for social problems in the rest of the country. The

opening up of Rondonia to colonist settlement brought thousands of landless people from

Southern Brazil. In 1970, the Transamazon highway colonization scheme was begun

with hopes that large numbers of migrants would come from the drought stricken

northeast to settle in the Amazon and build a strong economy based upon agriculture,

mining, and timber extraction (Moran 1984; Smith 1981). Rather than settling along the

main rivers, the colonists inhabited upland areas along the new highways. They brought

their notion of agriculture from the sub-tropical and temperate south and the dry northeast

to a foreign environment. The colonization project made little effort at ecological zoning

according to soil type or other indicators of fertility. The colonists were handed a piece

of land that often lay on extremely poor soils. Many of the colonists failed and returned

home while others found a livelihood in the growing urban centers. Some learned how to









adapt their agricultural methods to the rainforest environment and were able to make a

living.

The development of diverse agroforestry systems is seen by many as a viable

option for Amazonian farmers (Smith 1996b). The maintenance of biodiversity and

better preservation of ecosystem services are two potential advantages of agroforestry

and some argue that the cutting of old growth forests would be slowed by having a more

permanent system in place (Denevan 1980; Smith et al. 1995). Swidden agriculture

frequently includes agroforestry as an integral part of forest management (Alcorn 1990;

Coomes and Burt 1997; de Jong 1996; Denevan and Padoch 1987; Eden and Andrade

1987; Hammond et al. 1995; Unruh 1990). Indigenous peoples throughout the tropics

often practice some form of fallow management that increases the productivity of a given

field. In indigenous swiddens, cultivation is rarely stopped abruptly but instead, useful

plants are harvested for many years after the field is "abandoned."

Denevan and Padoch (1987) and their colleagues studied what they aptly termed

"swidden-fallow agroforestry" in a native Bora community of the Peruvian Amazon. In

this particular case and many others, swidden fallow management is based on successive

planting and harvesting of different crops. For example, a staple crop such as manioc

may be interplanted with longer maturing crops like bananas and some fruit trees. As the

crops are harvested, others may be planted in their place. Eventually, the field can

become a secondary forest enriched with useful species or a fruit orchard. The advantage

is that a farmer can return to his fallows to continue harvesting fruits, timber, medicines,

and other products for years to come. Another benefit of having a diversity of crops

growing in managed fallows is that farmers can spread production throughout the year.









This research will show that traditional caboclo farmers of the middle Solim6es also

practice various degrees of swidden-fallow agroforestry and forest enrichment.

Traditional systems could be adapted by colonists to develop a more viable

agricultural system. Indeed, innovative colonists are already experimenting with various

consortium of crops on their own (Smith et al. 1995). The land use systems of

traditional inhabitants can certainly provide some useful lessons for colonists to improve

the way they utilize and conserve resources in Amazonia. This research examines the

extent that traditional farmers on the middle Solim6es utilize fallow management by

documenting the distribution of useful plants in the agricultural landscape and the

practices that result in that distribution.

In Brazil, while most of the nation's attention was focused on the terra firme as the

main area of development in Amazonia, the virzea with its relatively small extent of

about 2% of the basin, was largely ignored. The virzea is the floodplain of "white water"

rivers that carry fertile sediment from the Andes Mountains. The realization that terra

firme soils are generally poor in nutrients and do not support intensive cultivation of

cereal crops without substantial external inputs has sparked a renewed interest in the

virzea as a promising area for modern agricultural development (Junk 1982; Smith

1981). Because of the influx of nutrients brought by the flood, good prospects for

irrigation, and ease of transport, some have advocated the development of arable farming

or irrigated rice on the virzea. However, developers and policy makers generally do not

know how to implement large-scale mechanized agriculture on land where massive,

unpredictable floods are the norm.









Several attempts have been made to grow irrigated rice on the varzea with varying

levels of success. A project at the mouth of the Jari River attempted to grow irrigated

rice on a large scale in the 1970's. Their efforts were hampered by problems with soil

fertility, pests, fungal diseases, and weeds (Fearnside and Rankin 1985). Junk et al.

(2000) discourage large-scale arable farming on the virzea for environmental, economic,

and social reasons. Instead, they support the integral management of aquatic and

terrestrial resources through the zoning of economic activities and decentralized and

participatory management by the local population.

There are four major economic activities on the virzea that vie for control over

limited resources; fishing, forestry, agriculture, and ranching (Junk et al. 2000; Smith

1999). The conflicts of interest are exemplified by the possible adverse relation between

fishing and the other three land uses that result in deforestation of the floodplain. Forest

clearance for timber, agriculture, and pasture poses a threat to the productivity of aquatic

resources by reducing their habitat and food sources (Fearnside 1990, 1995; Goulding et

al. 1996; Goulding 1999). Deforestation is also a threat to the many endemic plants and

animals of the virzea that are so far not adequately protected. In addition, the growth of

ranches and large-scale agriculture can result in the loss of crops, varieties, proto-

domesticates, and crop relatives that are vital for the future development of the virzea

(Smith 1999). With the increasing interest in intensifying agriculture on the virzea, it is

pertinent to examine the resource use and perceptions of the local inhabitants.

There are very few Amerindian groups that utilize virzea resources today. Having

endured the first and longest sustained contact with their European conquerors, the

Amerindians who lived along the Amazon River were virtually wiped out by disease,









warfare, and slavery. Three Amerindian groups that currently practice agriculture on the

floodplain are the Shipibo and Cocama who live in the Peruvian Amazon and the Tikuna

(they have moved to the riverside from the uplands since the time of European contact)

who live in the frontier region of Colombia and Brazil (Bergman 1980; Shorr 1999).

However, they are relatively acculturated groups and are a small minority along the

Amazon River. The caboclos of Brazil or ribereios of Peru comprise most of the

inhabitants along the major rivers today. They are people of mixed Amerindian,

European, and African ancestry (mestizo) who rely largely on subsistence techniques

similar to those of indigenous Amazonians (Moran 1993; Sutlive et al. 1985; Wagley

1953). Indeed, there is a continuity of Amerindian culture and resource management in

the caboclo way of life.

Relatively little attention has been paid to the living strategies of the mestizo

riverine population (Padoch and de Jong 1991; Parker 1989). Hiraoka (1992) provides a

brief review of the research on caboclo and riberefo resource management. Here, I will

highlight some of the literature pertaining to caboclo/riberefo agriculture that is

particularly relevant to this thesis. Denevan (1984) elaborated the concept of

"horizontal" ecological zonation of agriculture on the floodplain, likening it to altitudinal

zonation on mountain slopes. His article was inspired by Bergman's (1980) study of the

Shipibo in Peru and work done by Judith Gunn who studied agricultural scheduling on

the Ilha dos Purus in the Amazon above Manaus. Hiraoka (1985a, 1985b) described the

ecological zonation of floodplain agriculture of a riberefo community near Iquitos in the

Peruvian Amazon. Several studies have focused mainly on mestizo terra firme

agriculture along the river near Iquitos (Chibnik 1994; de Jong 1996; Hiraoka 1986;









Padoch and de Jong 1992; Padoch et al. 1985). Frechione and colleagues (1989)

examined caboclo's perceptions of biotopes and ecological zonation at Lake Coari on the

Solim6es. Anderson et al. (1995) studied resource management of caboclos in the

Amazon estuary near the mouth of the Tocantins River. Swales (1999) compared land

use dynamics and agriculture in the uplands and floodplain in the lower Amazon.

Preliminary results of research on floodplain farming in the Brazilian state of Amapa are

given by Padoch and Pinedo-Vasquez (1999). Some ongoing work is being done on

agricultural systems and change on the virzea in the MamirauA Sustainable Development

Reserve where the data for this thesis was collected (Padoch et al. 1996; Pinedo-Vasquez

et al. 1996, 1999). Finally, a comprehensive work edited by Junk et al. on actual use and

options for sustainable management of the virzea includes information on current

management practices by caboclos (Junk et al. 2000).

One way that caboclo agriculture can contribute to questions of virzea

development are the many crops and varieties that they plant. Generations of farmers

have selected varieties and individuals that are tolerant to flooding and resistant to pests

and diseases. Development plans would be prudent to incorporate crops that are adapted

to conditions on the virzea. Caboclo agricultural practices can also inform virzea

development planners. We know that Amazon floodplain farmers utilize a number of

habitats. What management strategies do they use to cope with the varying

environmental conditions of those habitats? What crops do they plant in which habitats?

The imperative of conserving virzea biodiversity and natural resources compels us to

analyze the agroecosystems of caboclos in detail to determine which aspects of their

management can contribute to the sustainable development of the Amazon floodplain.









Amazon Caboclo

The people who make their living along the middle Amazon today are known as

caboclos, a designation whose significance has evolved and been modified during the

history of colonial occupation (Lima-Ayres 1992). The term caboclo conveys different

meanings to different people. An important distinction is made between the colloquial

use of the word and its application in the academic literature. In the colloquial use,

caboclo can be a racial category, economic class, or rural resident. In elementary school,

Brazilian children learn that caboclo means a mixture of Amerindian and European

ancestry. Caboclo also indicates a low social class, the rural or urban poor, and illiteracy.

The term is generally derogatory and few people will refer to themselves as caboclo. It

would be used by the urban upper class to refer to the urban poor, and by the urban poor

for people from the "interior" or rural areas. Rural inhabitants, in turn, would not call

themselves caboclos but use it to refer to Amerindians. In other words, it is always "the

other" and rarely oneself (Lima-Ayres 1992).

In the academic literature, caboclo is not a racial category but rather, a social

category, the indigenous Amazonian peasantry (Lima-Ayres 1992; Parker 1985; Ross

1978). A distinction is made between the caboclo and the other two major groups in the

Amazon, Amerindians and recent migrants or colonists. Caboclos are made up of a

mixture of people from three main regions; Amazonia, Europe, and Africa. Many

caboclos are descendents of the indigenous inhabitants that lived along the Amazon when

Francisco de Orellana first sailed down the river from Ecuador to the Atlantic in 1542.

Amerindians were captured as slaves or enticed into missions and incorporated into the

European colonial system. As time went on, Amerindians continued to migrate from

tributary and headwater regions to fill the demographic void created by the depopulation









of the varzea region. Miscegenation resulted in a population dominated by people of

European (mainly Portuguese) and native Amazonian descent. Included in the mix are

Africans who were originally brought to Brazil as slaves to work on plantations and also

migrants who came to the Amazon from Northeast Brazil and other regions during the

rubber boom from 1850 to 1920.

Wagley (1953) and Galvao (1952) made the first anthropological studies of the

caboclo and characterized "caboclo culture" as a combination of Amerindian and Iberian

cultural traits. Many aboriginal economic strategies have persisted in the caboclo

population long after the tribal peoples were gone. Amerindian methods of agriculture,

hunting, and fishing, as well as customs and religious beliefs are still practiced in order to

make a living in the volatile, often unpredictable environment of the middle Amazon

(Moran 1974; Parker 1985).

Middle Solim6es Region

The Solim6es River is the section of the Amazon River in Brazil stretching from

the Peruvian border to the confluence of the Negro River near the city of Manaus. The

middle Solim6es region is approximately centered where the Japura River joins the

Solim6es (Figure 1-1) (Lima-Ayres 1992). Tef6 is the largest town and the commercial

center of the region. According to the 1996 census, the municipality of Tef6 had a

population of 62,000 with 76% living in urban areas (IBGE 1998). The surrounding

municipalities have a much lower urban population. In the nearby municipality of

Uarini, with a 1996 population of 10,500, 20% are urban. The regional population is

growing rapidly with a high rate of natural increase and rural to urban migration.

Two degrees south of the equator, the climate is tropical with little year long

variation in average temperatures. Average monthly temperatures range from a daily









minimum of between 21 and 23 degrees Celsius to a daily maximum of between 30 and

33 degrees Celsius (Ayres 1995). There is considerable variation in precipitation

throughout the year and from one year to the next. December through March is the rainy

season with about three times the amount of precipitation as during the dry season from

July to October (Ayres 1995). Average yearly rainfall from 1977 to 1981 for the city of

Tef6 was 2,373 millimeters with the lowest and highest rainfall for those years being

2,190 and 2,632 millimeters.

Fishing, agriculture, and timber extraction are the predominant economic activities

in the region. According to the Brazilian Institute of Geography and Statistics, manioc is

the most important crop (IBGE 1998). In the municipality of Tef6 (23,800 km2), 2,238

out of 2,255 (99%) farms surveyed produce some manioc. Approximately 64,000 tons

were harvested on 6000 hectares. However, only 33 tons were reportedly sold indicating

that most manioc is grown for subsistence only. Bananas are the second most important

crop in the region and, unlike manioc, most are sold on the market. Other crops that were

included in the survey that are produced to a much lesser extent include corn, beans,

sugarcane, oranges, tomatoes, and rice. Cattle ranching has not yet taken off in the

region. A total of 5000 head of cattle were reported in the four municipalities of Tef6,

Alvaraes, MaraA, and Uarini with the majority in Alvaraes and Tef6 near the larger urban

centers. Some hogs and poultry are also produced. Many rural families keep a few ducks

or chickens around the home. Only a few of the households surveyed for this thesis

owned cattle or hogs.

Virtually all settlements in the middle Solim6es region are located near a river.

Most rural inhabitants live in scattered settlements of fifteen houses on average and a few









live in isolated homes. Many of the small settlements in the middle Solim6es region

began with the decline of the rubber era when rubber tappers moved in from more

isolated extraction areas (Sociedade Civil Mamiraua 1996). The communities in the

region are located either on the river bluffs surrounding the virzea, on the virzea itself, or

on large lakes formed at the mouths of tributaries of the Solim6es such as Lake Tef6,

Lake Uarini, and Lake Amana. Communities differ in their monetary wealth, degree of

isolation, market connections, government support, and available habitats for agriculture

and resource extraction. On the terra firme, the agricultural cycle follows the change in

seasons from wet to dry. The cycle of activities on the virzea follow the pulse of the

flood.

Conservation of the Varzea Ecosystem: The Mamiraua and Amani Sustainable
Development Reserves

The enormous productive potential and inumerable endemic species of the virzea

have spurred researchers to underscore the importance of rationally managing floodplain

resources (Goulding et al. 1996; Junk et al. 2000; Padoch et al. 1999; Smith 1999). The

newly created Mamiraua Sustainable Development Reserve (MSDR) is the only

conservation area specifically dedicated to the preservation of biodiversity and

management of resources on the virzea in Brazil (Sociedade Civil Mamiraua 1996).

Covering over one million hectares above the confluence of the Solim6es and JapurA

Rivers, the area was first established as an Ecological Station in 1990. The recognition

that people live in and utilize the resources of Mamiraua prompted a new category of

conservation unit to be created and the area was reclassified as a sustainable development

reserve in 1996. In 1999, the Amana SDR was created adjacent to Mamiraua to link up

with Jai National Park creating a central Amazon conservation corridor.









The goal of the Mamiraua project is to integrate conservation of biodiversity with

the social development of local communities (Lima 1999; Howard et al. 1995; Sociedade

Civil Mamiraua 1996). Scientists work with residents of the reserve to research, monitor,

and protect natural resources and to develop a management plan that encompasses both

conservation and development objectives. They must determine which activities threaten

natural resources and seek new activities that are compatible with conservation goals.

Our study builds on others from the Mamiraua Reserve (Chota 1999; Padoch et al.

1996; Pinedo-Vasquez et al. 1996; Pinedo-Vasquez et al. 1999) to show the enormous

resource of agrobiodiversity and management diversity used by small farmers within the

conservation reserves. It is hoped that it will be valuable as a jumping off point for

further research on agrobiodiversity in and outside the reserves and useful when

considering new policies or projects that may affect agrobiodiversity in the region.

Data and Methods

To understand the spatial distribution of agrobiodiversity across different habitat

types and management systems, the home gardens, fields, and fallows were visited in

nineteen communities in the middle Solim6es region (Table 2-1 and Figure 2-1). An

inventory was conducted in each home garden that included all food, timber, and

medicinal plants. The number of individuals of each species was recorded, the size of the

home garden was estimated, and a position was taken with a GPS receiver.

Fields and fallows were visited and the following information was gathered.

* GPS position
* Type of work area (field, fallow, orchard, etc.)
* Estimated size
* Habitat (terra firme, high levee, low levee, mudflat, or beach)
* Useful species (including food, timber, medicinal, and other)
* Varieties of manioc and banana










* Name of owner
* Other information gathered for some fields and fallows including but not limited to:
crop age, field age, number of consecutive seasons planted, number of previous
fallows, age of last fallow, and status or use of the area before (fallow, old growth
forest, etc.)

An attempt was made to visit all fields and fallows in each community. However,

this was not always possible because of the great number of fallows in some

communities, the logistics of reaching some of them, and the question of recruiting

informants who were willing to take time out of their busy schedules to help us.

The objective of the study was to visit a large number of fields in communities

located throughout the region, therefore, quick estimates of size were made rather than

taking the time to carefully measure each field. Size was estimated by site and

Table 2-1. Number of home gardens, fields, and fallows visited in 19 communities on
uplands and floodplains in the middle Solim6es region
Num.
Varzea/ home Num. Num.
terra gardens fields fallows
Community Map id firme visited visited visited
Acari 9 virzea 0 1 0
Barroso 8 virzea 11 16 9
Bate Papo 10 virzea 0 2 3
Bela Vista do Manguary 11 virzea 2 5 9
Betania 15 virzea 16 35 16
Betel 16 virzea 10 26 13
Boa Esperanga 19 terra firme 17 55 45
JarauA 18 virzea 11 31 12
JubarA 13 terra firme 5 15 8
Manacabi 17 virzea 7 14 12
Marirana 7 virzea 10 13 5
Nossa Senhora da Fatima 2 terra firme 8 16 22
Pentecostal 12 virzea 9 8 6
Porto Praia 4 virzea 15 7 12
Sao Francisco do AiucA 5 virzea 15 48 25
Sao Joao 6 virzea 16 10 6
Sao Paulo do Coragi 14 virzea 12 22 48
Sitio Fortaleza 3 virzea 11 15 34
Vila Alencar 1 virzea 19 19 36
* In these two communities, fields were visited on both virzea and terra firme



















































0 20 40 Kilometers


Figure 2-1 Landsat TM satellite image (bands 543) of the study region from 1999 with
the locations of 19 communities included in the survey









occasionally checked by pacing. Estimates of the size of smaller fields should be

accurate to within a couple of meters while the error may increase proportionately as the

size of the field increases. In cases where it was impractical to measure the size by

pacing and it could not be estimated by site because of thick canopy, the owner or

informant gave his estimate of the size, leaving room for error. Therefore the size,

especially for the larger fallows, should be considered rough estimates.

The distinction between habitats of terra firme and varzea is more clear than that

between high and low levees and low levees and mudflats. It would have been possible

to draw an arbitrary line between the habitat types based on the height of the land in

relation to the high water level of the previous years flood but this would be possible only

where there are trees with visible high water marks. In the interest of time this was not

done. Instead, the decision was made to use the informant's classification of the habitat.

Local farmers classify habitats based on their observations of the depth of flooding

and length of time underwater each season as well as the vegetation and soil types. They

rely on their own experience from their particular local environment and so these

classifications can be somewhat subjective. Farmers classify four major floodplain

habitats used for agriculture; resting alta (high levee), resting baixa (low levee), lama

(mudflat or silt bar), and praia (beach or sand bar). At the time of fieldwork, high levees

were usually above water but some were flooded to a depth of usually not more than forty

centimeters. The vegetation is well developed forest. Areas that farmers classified as

low levees were flooded from one to several meters and vegetation is usually dominated

by the imbatiba tree (Cecropia spp.). Mudflats are typically dominated by canarana grass

(Echznochloa spp.) and beaches are devoid of vegetation.









The inventory of the fields and fallows recorded all species that were considered

useful by the inhabitants. These included tubers, fruits, vegetables, nuts, timber, and

medicinal plants. A few others with uses such as fishbait, latex, and thatch were also

included. The inventories were made by walking around in the field or passing slowly

along the riverbank in a boat. The informants were relied upon to name the useful

species and much effort was made on our part to insist that all of them be recorded. The

error here is most likely to consist of some species not being recorded for some fallows,

especially the very large ones. Varieties of banana and manioc were recorded in many

cases but not all. Sometimes the variety was unknown to the informant. Many of the

fields that were visited had already been harvested and were often flooded. In this case

the crops that had been harvested were reported.

The information on management was gathered, when possible, in informal

interviews with the informant while visiting the field. Occasionally, details on yield,

weeding, markets, or management were learned in this way.

Finally, an informal interview was conducted in the home after taking the home

garden inventory. The residents were asked about the number and size of their fields and

fallows, the crops they had harvested, and the produce they had sold to market. They

were also asked about losses of crops and home garden plants due to the flood or pests

and diseases. The information gathered in this way is supplementary only. The

interviews were not structured and the same information was not given by every

interviewee.

The data was put into a database and incorporated into a GIS using the

geographical positions recorded in the field. To do this, a Landsat Thematic Mapper









image was geo-referenced and the points were plotted. This has been given to the

Mamiraua Sustainable Development Institute to be incorporated into their larger GIS

database.

Areas that were designated as high levee or resting alta were flooded to a

maximum depth of one meter in the year 2000. Most of them were flooded less than 40

centimeters and the majority still had dry land, with many elevated up to 50 centimeters

above the high water level for the year. In this survey the classification of habitat type

was based on the informant's designation of the area. Therefore, it is assumed that some

overlap exists between habitat type designations. In other words, areas that were called

resting alta in one community with one meter of water may be termed resting baixa in

another community. For example, in Sao Paulo do Corayi located on a side channel

(paranA) of the Japuri, some fields that had not flooded and were up to 50 cm. above the

flood in 2000 were termed resting baixa while in Betel, a community on the Japura,

resting baixa fields were submerged in one meter of water. In the interest of time, the

decision was made to rely on local designations of habitat types rather than attempting to

measure the height of the land. Due to the extremely dynamic nature of the virzea

landscape, the height of the land can change from year to year due to erosion and

deposition of sediment. There is also the issue of differential flooding. Areas near the

main channel of the Solim6es probably receive higher floodwaters than areas near the

Japura or side channels.

Research Considerations of the Flood Season

Field research was conducted in the summer of 2000 (May to August) during the

peak of the flood. Ideally, to study agricultural biodiversity on the virzea, it would be

best to observe the agricultural fields through an entire cycle or, better yet, several cycles.









There would be no single part of the year on the floodplain in which all diversity could be

captured. The agricultural calendar is complicated and dynamic with crops being planted

and harvested during several seasons throughout the year.

At first consideration, it would seem that the flood season is the worst possible time

to study agricultural diversity since the majority of the fields are harvested or even under

water. But it is at precisely this season that we can get an idea of which species and

varieties are retained in situ even during the flood. It is also significant that this research

followed three consecutive years of high floods and the previous year, 1999, experienced

an exceptionally high flood (Figure 2-2). The flood of 1999 was one of the five highest

in the last century and caused substantial losses of seasonal crops, planting material, and

perennials for residents of the Mamiraua and Amana Reserves. This research is

especially pertinent because it can indicate the level of agrobiodiversity that remains in

the agricultural system after a period of several high floods. In addition, although the

specific objective of this research was not to gather data on crop losses resulting from the

high flood, a significant amount of information was collected during the inventories and

interviews. It is enough to give a good, if not comprehensive, indication of which species

are more susceptible to drowning and to what extent crops are damaged by high floods.

Crops including corn, beans, squash and others are probably somewhat

underrepresented by this study since most of the areas where they are grown, mudflats

(lamas) and beaches (praias), were under several meters of water at the time of the

survey. Although the attempt was made to visit all of the areas where crops were planted,

including those underwater, it is likely that some areas were missed in the survey due to

the informants misunderstanding, forgetting, or downplaying the importance of these









areas. Additionally, some of the non-harvested manioc fields that were visited may have

contained additional crops that were harvested before the survey and went unreported by

the informant. These factors combined generate substantial room for error. It is

important to emphasize that the present study is simply a snapshot in time and a much

more comprehensive, long-term study would be required to capture the full range of

agrobiodiversity and its dynamic nature and resilience on the virzea.

Life on the Varzea

The stretch of floodplain in the middle Solim6es region is among the widest along

the entire length of the Amazon river. Inhabitants of the virzea must often travel many

kilometers to reach the terra firme. The difference in level from low to high water is the

highest in the Brazilian Amazon, exceeding fourteen meters in some years. Even the

highest ground, the floodplain levees (restingas), may be inundated up to several months

each year imposing special living conditions on residents who must cope with water

covering their yards and fields for weeks on end.

The virzea landscape is one of the most dynamic on earth. The shifting river

channels continually form and erode levees. High levees that are used for agriculture and

settlements collapse into the river while elsewhere sandbars are colonized by pioneer

vegetation and new islands and levees are formed. The time scales at which these

morphological changes take place are extremely fast by geological standards with levees

and islands potentially being molded or erased in less than a human lifetime.

Agricultural fields concentrated in one area may shift to another area at intervals of

several years in order to adjust to the rapid morphological changes in the landscape such

as the closing off of channels as they become choked with vegetation and sediment and

the building and destruction of levees.











1800

1600

1400

1200

1000

00



400

200

000



Figure 2-2 Seasonal fluctuation in water level on the middle Solimes, 19922000 Rise
and fall (n meters) of the river nearthe town ofTefe In 1999, there was
approximately 14 mteers difference between low andhighwater Source
SDR Mamraua 2000

Varzea dwellers have adapted their lifestyles to contend wit the inevitable yearly

flooding Settlements on the floodplain are located on the high levees to avoid prolonged

periods of inundation (Goulding et al 1996) Most arealong major ver channels for

access to nver traders and collective transportation Families on the floodplain usually

build their homes on stilts to avoid bng inundated by the isng waters During the

weeks or months with water under their homes people rely on their canoes for even a

quick tnp next door Often the water is only several inches to knee deep and trudging

through mud is commonplace Children are frequently seen playing and splashing in the

nver rght outside their front door Animals must be kept in the house or on floating rafts

if no dry ground is available









The enchente or high water period is unanimously considered to be a time of

hardship. Crops may not be grown on flooded land and fishing is less productive since

fish are dispersed throughout the igap6 or flooded forest. The greatest difficulty comes

when they get a high flood like the flood of 1999 (Figure 2-2). They are accustomed to

this ordeal and cope with it in a variety of ways. Many leave their homes and temporarily

stay with relatives in towns on the river bluffs. Others stick it out and simply build a new

floor above the water level, continuing to raise it when necessary to keep pace with the

rising deluge. Residents along the main channels complained of huge wakes from

passing boats. They would leave their doors open in order to diffuse the force of large

waves breaking against the front of their homes. Instead, the powerful waves would

sweep through their living room and crash into the back wall of the house. Some avoid

the invading waters by living in a floating house with the added benefit of being more

mobile, even able to tow their residence to a new location. People are very resilient and

bounce back quickly from such events. They rely on friends, neighbors, and kinship

networks to get them through these tough times.

The exceptional floods like the flood of 1953, the highest of the twentieth century,

are remembered long afterwards. High floods can be quite a setback for residents. It is

after one of these disruptions to their lives that families sometimes decide to make a

move to terra firme. The people are highly mobile and the turnover rate for residents of

the varzea is high (Sociedade Civil Mamiraua 1996).

Fishing and agriculture are the predominant economic activities for residents in this

part of the Amazon floodplain. To earn an income, some devote all of their energy to

fishing and others focus on agriculture but the majority depend on both activities for their









livelihood. Other activities such as the extraction of timber, firewood, and fruit also serve

to supplement their income but are generally only secondary in importance. The a9ai

(Euterpe spp.) fruit, used to make a popular drink, is commonly collected and sold.

Other fruits such as yellow mombim (Spondia lutea) and genipap (Genlpa americana)

are common on floodplain levees and are sometimes brought to market or sold to passing

river traders. Honey from native stingless bees is occasionally harvested from the forest.

Another possibility to earn extra money is hunting caiman and selling the meat dried or

salted. Hunting of other animals may also be an important subsistence activity especially

for the communities on or near terra firme. However, selling wild game products,

including caiman, is illegal in Brazil so there is not a market for game meat as in some

other regions of the Amazon, such as Peru. Turtles, turtle eggs, and manatee meat fetch a

high price but those who try to sell them risk getting caught and fined by IBAMA, the

federal environmental protection agency.

Working within or outside the community for a wage is an option for some

individuals. Working as a day laborer for a neighbor is more common in terra firme

communities where production of farinha or manioc flour is practiced on a relatively

larger scale. Here, farmers sometimes hire day labor to weed their fields or they may

contract someone with a chainsaw to clear some land. They typically pay a few dollars

per day for this type of work. Some people also find temporary work in nearby towns.


































Figure 2-3 Floodplain community during low water The building in the foreground is a
floating house and the high water mark from the previous flood is halfway
up the walls of the homes on the bank Boca do Mamiraua, November 1999















CHAPTER 3
STRATEGIES FOR FARMING ON THE FLOODPLAIN: THE ROCA

Varzea farmers contend with unpredictable floods, relentless pests and diseases,

and an extremely heterogeneous environment that is constantly changing. Crop losses

are frequent and can cause significant setbacks. As a means of risk management, they

employ numerous strategies and take advantage of the varied terrain by utilizing several

different habitats for planting crops. Reliance on rapidly maturing varieties as well as

crops that are able to withstand weeks or months in standing water help farmers cope

with the inevitable yearly flood. They use their knowledge of the environment to select

appropriate areas for planting crops. There is an element of risk as farmers attempt to

predict the magnitude and timing of the coming deluge. Residents, therefore, rely on one

another for assistance when losses occur.

The floodplain habitats used for agriculture are divided by farmers into four main

types: beaches (praias), mudflats (lamas), low levees (restingas baixas), and high levees

(restingas altas). The main factor that they use to differentiate them from one another is

their altitude and, therefore, the degree to which they are subject to flooding. The

distinction that is made between these habitat types can be rather arbitrary. Actually, a

continuum exists between the habitat types and management depends on other factors as

well such as predominant vegetation, sediment deposition, and individual preference.

Farmers who have access to land on terra firme often plant there as well. In the

varzea community of Sitio Fortaleza, for example, several families have been planting on

terra firme that is some distance away (one hour by motor boat or three hours rowing).






44

























Figure 3-1 The bank of the Japura River during low water Shows a typical gradation
from sandy beach to grassy mudflat to forested low levee (Nov 1999)

Several members of the community expressed their opinion about the hardships of life on

the floodplain and the relative secunty of farming on terra firme They have decided to

move their families to high ground and others in the community may soon follow

Jubara, a terra firme community on the bank of the Japura nver, also utilizes both varzea

and terra firme for agnculture

Fields, fallows, and home gardens are the three major categories of land use that

farmers manage All places where crops are planted or were planted in the past can be

referred to as work areas (areas de trabalho) Annuals such as corn and beans or the

semi-perennial crop mamoc are planted in fields with the frequent inclusion of perennial

fiuit trees Fallows range fiom a simple abandoned field with no useful species through a

mono-cultural banana grove on up to a mixed fruit orchard They harbor a large selection









of useful plants and are a significant resource for subsistence and the market and often

especially rich in construction materials. Home gardens contain a diverse range of fruit,

vegetable, medicinal, and ornamental plants. They provide a range of foods and other

useful products to the household.

When an area is selected for cultivation, farmers take advantage of pre-existing

valuable plants by sparing them during field clearance. They also practice selective

weeding to protect useful volunteer plants, some of which are favored by forest

disturbance. These farming methods alter the mix of species in regenerating forests to

create a heterogeneous forest environment on the floodplain, effectively increasing the

occurrence of plants that are useful to humans and, in some cases, benefiting wildlife

(Pinedo-Vasquez et al. 1999, Padoch et al. 1996). It is possible that overall diversity in

the region is greater as a result of the variety of management strategies employed by

varzea farmers (Bal6e 1994).

Characteristics of the Rosa

In the vernacular language of the region the term roya refers specifically to a

manioc field. Plots dominated by a crop other than manioc are generally referred to by

another name. For example, a field dominated by corn, watermelon, or banana would be

called a milheral, melancial, or bananal, respectively. The manioc tuber is planted in

habitats ranging from mudflats to high levees but never on low-lying sandy beaches that

are the first to flood. A limited repertoire of crops is planted on beaches; mainly beans

and watermelon. In this thesis, rocas are designated as work areas used to plant any

annual crop as well as the semi-perennial crop manioc. Ro9as may be monocultures with

one or several varieties of a crop but are frequently polycultures containing a number of

species and varieties. Perennial fruit trees (usually commercially valuable) are









sometimes planted with the manioc crop to be left growing when cultivation is

discontinued.

In a study of agricultural management systems in three communities in the

Mamiraua Reserve, Pinedo-Vasquez et al. (1996) described two types of planting

configurations that are primarily used. The first, called randomly planted (plantio

mixturado), is a system under which two or more crops are interplanted. There is often

one main crop and one or more secondary crops. In the second system, stratified planting

(plantio dividido), rogas are divided into sections where different crops are grown based

on soils and topography. According to their research, the frequency with which each

system is used is dependent upon the intensity of the flood in the previous year. Farmers

tend to employ the stratified planting system more often after high floods as a means of

risk management. They expect high floods two years in a row so they carefully choose

the highest areas of the field to plant manioc and bananas and tend to increase production

of fast growing crops like squash.

Ro9as may be cleared from forest (mata, sometimes referred to as mata bruta or

mata virgem indicating its status as old-growth, mature, or virgin forest) or from any

work area type included in the fallow category. Levees (restingas) are most often

covered by closed canopy forest that necessitates burning in the first season. Once fields

are established, burning is often unnecessary in the following seasons even when clearing

young fallows.

In some cases, what is called forest (mata) may actually be old secondary regrowth.

Residents usually do not have knowledge of a particular forest beyond about twenty years

as many of the communities on the varzea of Mamiraua are less than twenty years old. It









is quite possible that much of the forested levees have been cultivated at some time in the

past (in historic or prehistoric times), effectively altering the natural mix of species in

floodplain forests.

Valuable trees are sometimes, but not always, spared during forest clearance.

Large timber or fruit trees can sometimes be found standing in the middle of a roya but

this is fairly rare on the varzea. It is a more usual sight in terra firme royas that are

typically much larger. Very large trees are sometimes left standing because of the danger

or time it takes to cut them down. They often die when surrounding forest is removed but

wood can still be exploited years afterward. A common way to preserve valuable trees in

both ecosystems is to clear the forest just up to the useful tree and stop, leaving it just on

the border of the field locally called the a9ero.

Farmers reported weeding levee fields up to four times during an eight-month

growing season. This is done by crouching or sitting on the ground among the manioc

stems and removing unwanted plants by hand or with a machete. Plants perceived of as

valuable are often protected during this process. The three techniques; sparing trees

during forest clearance, selective weeding, and the planting of perennials all contribute to

the increased value of swidden fallows for the subsistence and income of farming

families in the region.

Communities differ greatly in the distance to and amount of each habitat type

available to them. In some communities forest on high levees is in short supply.

Residents from the community of Manacabi reported that almost all of the high levees

around the community has been cultivated and there are only a few small pieces of old









forest left. Farmers in some communities travel great distances from their homes to clear

fields from old forest on high levees.

Manioc is the principal staple crop of the region and the main commercial crop for

many farmers who sell their surplus of processed manioc flour (farinha de mandioca).

The crop is typically grown during a six to eight month season on the varzea. According

to reports by farmers, some varieties of manioc that take one year to mature on terra

firme take just six months in the fertile floodplain soils. The decision is usually made to

harvest when the rising waters threaten to drown the crop. Farmers harvest the lowest

lying rogas first in order to keep pace with the rising flood. They also begin with the

most flood prone areas within a roga. They hurry to pull the tubers from the ground

before they are inundated and it is not uncommon for them to fall behind and harvest a

crop that is already under a few inches of water. If flooded, a crop must be rescued

quickly before it rots. This usually happens within about two days. If it is apparent that

the river has stopped going up, the decision is often made to leave the crop in the ground

for up to twelve months to thicken (engrossar). If farmers decide to continue planting a

roca for another season, they frequently begin planting immediately after or during

harvest as long as they believe that the river has stopped rising.

After manioc, banana is the most important crop for the majority of varzea farmers.

Bananas are sold to market by many families and are a frequent contribution to the diet of

local residents. A common strategy on the floodplain is to interplant a roca with manioc

and banana for one or two years and then discontinue manioc production, leaving the

field in some type of banana "fallow." The density of this banana plantation can vary

greatly. It might be only a few plants growing in a regenerating forest, simply termed









"capoeira" or fallow by farmers. With a slightly greater density of planting and intensity

of management including periodic weeding, it is called a "capoeira bananal" or banana

fallow by the owner. The farmer refers to a dense, carefully weeded banana grove as a

"bananal." This example, which will be explored further in the next chapter, shows that

there is no clear distinction between management types or the designations that farmers

give them. Because of the diversity of management practices used in banana production,

the bananal was not included in the roya category but instead was placed in the next

major category, the capoeira.

Farming Praias and Lamas

The distinction between praias (beaches or sandbars) and lamas mudflatss or silt

bars) is not always clear. Beaches and mudflats are often found together with the sandy

beach grading upwards to the mudflat. The ground is uneven in most floodplain fields

with high and low spots consisting of varying grades of sediment that can change on a

yearly basis. Crops are segregated according to topography and soil texture.

The sandy beach, nearly devoid of vegetation at low water, is typically utilized to

plant beans and watermelon. No preparation of the land is necessary. In one instance the

informant took us to an immersed sandbar in the Solim6es River. Though the land

surface was several meters underwater in the middle of the huge, swiftly flowing river, he

was able to show us the boundaries of the field where three families work together to

cultivate watermelon and two varieties of beans, Manteiga and Comum (Preto). The size

of the field is approximately 180 meters long by 40 meters wide divided into sections for

each family. When the crop is nearing maturity, the families camp out on the beach to

protect the crop from theft by passing boats. They take advantage of this time to collect

turtles, turtle eggs, and to fish for catfish that are then salted and later sold. Farmers most









often exploit the beaches and mudflats immediately in front of their village or on the

opposite side of the river for convenience and protection from theft.

Where the sandy beach grades into silt or mud, farmers grow a variety of crops

including beans, corn, bitter and sweet manioc, gherkin, and squash. Most areas that

were classified as mudflats by residents were covered with canarana grass (Echznochloa

spp.) and a few had low Cecropia trees. In these areas a thick layer of mud is deposited

when the flood subsides.

Management of mudflats varies between communities and even among farmers of

the same community. Of the nine fields where information on management was

collected, only two were burned. For an additional two fields, it was reported that the

farmers had wanted to bur but were unable to because of the weather. Of the two fields

that were burned, one was burned before the flood leaving the area clean when the waters

receded and the other was burned after the flood. In the second instance the canarana

grass was burned both before and after slashing. In another field the grass was cut before

the flood and the current carried the slash away leaving the area clean.

The remaining fields were cleared using the ro9a-tomba-junta method (Pinedo-

Vasquez et al. 1999). This means that the field is not burned. At least 2 informants

reported that the vegetation is cut and thrown into piles and then carried to the edge of the

field and dumped. None reported that the slashed material was used as mulch (cobertura

morta) used to control weed invasion as Pinedo-Vasquez and colleagues found but the

question was not consistently asked in this survey. Their research also found that

mulched fields should be left standing to dry out for 25 to 30 days before planting to

avoid problems of insect attack and that farmers wait to plant their fields until they see









that certain insects have emerged as adults. This fascinating aspect of floodplain

agriculture warrants further investigation.

Fields on beaches and mudflats are often planted year after year without the need

for fallow. For 5 fields on beaches, the average number of consecutive years planted was

three with a minimum of two and a maximum of four. An additional field with beach and

mudflat had been planted every year for 25 years according to residents. The average

number of consecutive years planted for 15 fields on mudflats was five with a minimum

of one and a maximum of twelve.

The species harvested were recorded for a total of 23 fields made on beaches and

mudflats. The mean number of species harvested per field was two with a maximum of

five. The most commonly planted crops in sand were beans and watermelon. On

mudflats, the most common is manioc. Only two of the mudflat fields contained useful

species after harvest; each had one samaumeira (Ceiba pentandra), a valuable and greatly

exploited timber species in Amazonia.

Diversity and Management of Restinga Baixa Fields

The flood brings a thick layer of rich sediment that is deposited on resting baixa or

low levees each year. This allows farmers, in many cases, to plant their fields for a

number of seasons with no reported decline in productivity due to soil fertility. More

than twice as many fields were visited as fallows on low levees (106 fields and 50

fallows). This is the opposite of what was observed for high levees where there were

many more fallows than fields (about 113 fields and 176 fallows). This could be a

reflection of the fact that farmers do not need to clear new fields as often on low levees.

Actual results from fallows for the number of years previously planted was about 3.75

years for low levees and 3.5 years for high levees. An alternative hypothesis for the









greater number of fields on low levees is that, because fallows on low levees hold

relatively few useful species, the location of these areas is quickly forgotten and they are,

therefore, underrepresented in the survey. Out of ninety fields where the information was

collected, the average number of consecutive years planted was 3.1 with a maximum of

fifteen.

Management varies on low levee fields depending on location. Describing the

specific factors that determine how the fields are managed is outside the scope of this

research. However, it is likely to depend on influences such as depth of flood, sediment

load and deposition, speed of current, vegetation, and weather. Other factors that are not

determined by the environment may include available labor, available time, knowledge,

custom, and personal preference.

The number of seasons that farmers cultivate low levee fields is apparently at least

partially dependent on the maintenance of soil fertility through the annual deposition of

sediment. In the community of Jubara on the Japura River they are able to plant just two

seasons before production begins to fall and the disease tuber rot (podre de batata) affects

the manioc crop. According to one informant, they typically plant two years and leave

the field fallow for seven years. This appears to corroborate what one farmer said from

the village of Betania, also on the Japura River. According to him, on the banks of the

Japura where the water carries a lower sediment load they typically plant three to five

years and then leave the field fallow. However, on the side channel (parana) where the

water carries more sediment, they can plant every year and the tuber does not rot.

The method of field preparation on low levees varies. Fire is used in some cases

but usually only when opening a field for the first time. Jubara farmers reported that









nobody bums low levee fields; they are only slashed and cleaned. In Betania and

Marirana at least some low levee fields are burned after clearing only in the first season

and it is not necessary the following seasons. In the village of Sao Paulo, it was observed

that some farmers slash a field on low levees before the flood. Pinedo-Vasquez et al.

(1996) found that a small proportion of fields were burned in 1993-94 in the three

communities of Vila Alencar, Jaraua, and Barroso and those that were burned had been

cleared too late in the dry season.

Most fields on low levees were made on land that had been fallow. Out of 82

fields, 55 (67%) were reportedly made on land that had previously been fallow and 27

(33%) were reportedly cut from virgin forest (mata) (Table 3-1). At least 11 of the fields

cut from virgin forest had been imbaubais, stands consisting almost exclusively of

Cecropia trees and canarana grass. For 17 of the fields that had previously been in

fallow, the average length of time in fallow was 6.3 years with a minimum of two years

and a maximum of twenty.

Crops harvested from low levee fields in 2000 include beans, corn, gherkin, bitter

and sweet manioc, okra, squash, tobacco, and watermelon. The average number of

species that were harvested from 58 fields was two with a minimum of one and a

maximum of seven. It is not uncommon that farmers alternate crops, planting one crop

one year and another crop the next.

There were 13 fields that had not yet been harvested. At least two of these were

being harvested at the time of fieldwork. Although these fields were classified as low

levees by informants, they fall on the high end of the spectrum since most low levee

fields were flooded to a depth greater than one meter. All of them were planted with









manioc and only four contained other species. These include two fields with mulateiro

(Calycophyllum spruceanum), and one each with banana (Musa sp.), sugar cane

(Saccharum sp.), and inga (Inga sp.). Eight fields held from one to five manioc varieties

with an average of 2.6.

Ninety-three of the fields visited on low levees had been harvested and only 16 of

those (17%) had useful species remaining. Of these sixteen fields, there was an average

of 2.4 species in each. The eight species that were found in at least two fields are shown

in table 3-2. All of these are relatively flood tolerant except papaya (Carica papaya), a

weedy species easily spread by birds. Only banana, guava (Psidium guajava), and cocoa

(Theobroma cacao) were likely planted here. Ayai preto (Euterpe precatoria), genipap

(Genipa Americana), mulateiro, and the kapok tree (Ceiba pentandra) are volunteer

species frequently encountered in fields and fallows. Other species that were found in

only one field each include: cashew (Anacardium occidentale), sugarcane, cubiu

(Solanum sesslhflorum), cuia (Crescentia cujete), feijao de praia (Vigna unguiculata),

jauari (Astrocaryumjauarz), jit6 (Guarea sp.), corn (Zea mays), capsicum pepper

(Capsicum sp.), pitomba (Talisia esculenta), and taxi (unknown). Three of these are

volunteers; jit6 and taxi are timber trees and jauari is a common varzea palm used for fish

bait. One field also contained a small patch of manioc that the farmer had failed to

harvest and one had already been replanted with manioc.

Table 3-1. Summary of statistics for low levee fields
N Mean Min. Max. S
Consecutive Years Planted 90 3.1 1 15 2.6
Size (m2)1 105 1643 100 6000 1247
Last Fallow Period (years) 17 6.3 2 20 4.6
Species Harvested 58 2.4 1 7 1
Manioc Varieties 24 2 1 5 2
1 the largest field (16,000 m' with four owners) was dropped









Table 3-2. The eight most common species remaining in sixteen harvested fields on low
levees
Rank Local name English name No. fields (16) %
1 Banana Banana 5 31
2 Goiaba Guava 4 25
3 Agai Ayai 3 19
3 Cacau Cocoa 3 19
3 Jenipapo Genipap 3 19
3 Mulateiro Mulato wood 3 19
3 Samauma Kapok tree 3 19
4 Mamao Papaya 2 13

Restinga Alta: The Preferred Habitat for Planting

Diversity of crops and management systems is greatest in the resting alta or high

levee. Several years can pass without flooding making them the most favorable

environment for planting bananas and other perennials. Informants repeatedly voiced

their opinion that high levees are the preferred areas for clearing fields. Local farmers

designated eleven types of work areas on high levees (not including home gardens) with a

range of crops and management strategies. There are two main types of fields on high

levees, the roya and the corn field (milheral). These fields are predominately planted

with manioc and/or corn but are often mixed with banana and other annuals and

perennials. The other nine types of work areas will be discussed in chapter four on

fallows. The generic designation of fallow is only for convenience as many are not

simple abandoned fields. They range from having no management and no useful species

to having intensive management and dozens of useful species.

Some high levee fields are planted year after year while others are planted just two

or three seasons before being put in fallow. The number of consecutive years planted

averaged 2.8 for 80 fields (Table 3-3). For comparison, some fields owned by residents

of Betania had been planted up to 15 years in a row while in another area residents of









Jaraua reported that fields could only be planted two seasons before fallow was

necessary. Both areas are high levees but by examining the location one difference is

clear; the Betania fields are along the banks of the Japuri where a much greater amount

of sediment is deposited while the Jaraua fields are on a parana or side channel that

carries little sediment.

Table 3-3. Summary of statistics for high levee fields
N Mean Min. Max. S
Consecutive years planted 80 2.8 1 15 2.4
Size (m2) 132 2573 150 20000 2490
Last fallow period (years) 18 6.9 4 20 3.9
Number of species (planted fields) 82 4.1 1 14 3.3
Number of species (harvested fields) 53 2.8 1 8 2.1
Number of manioc varieties 70 2.3 1 6 1.2
Number of banana varieties 72 1.4 1 4 0.7

According to the survey, high levee fields are cleared from forest (mata) about half

the time. Out of 91 rogas, 40 had been cleared from forest, 11 from banana groves

(bananal), and 40 from fallows (capoeira). Available data for the age of 18 capoeiras that

had been cleared for rogas yielded an average of 6.9 years. The ages ranged from 4 to 20

years. Reports from informants indicate that, in general, rocas are only burned in the first

season after they are cleared from forest or old secondary forest (capoeira velha). In

subsequent years, farmers generally use the ro9a-tomba-junta method for clearing the

fields. The average size of 132 rocas was 2,573 square meters. They ranged from 150

square meters to two hectares in size.

The 135 fields visited on high levees are divided into two groups for comparison;

planted and harvested. The number of useful species in 82 planted fields ranged from 1

to 14 with an average of 4.1. Figures 3-2 and 3-3 show graphically the number of useful

species encountered in planted and harvested ro9as. Planted fields with a single useful









species are relatively infrequent, amounting to just 15 (18%). It is much more common

for fields to be interplanted with two or more crops. A little more than half of the rogas

on high levees were found to still contain useful species after harvest. Out of 53

harvested rogas, 31 (58%) of them held an average of 2.8 useful plants with a maximum

of eight.

Manioc is the principle crop planted in high levee fields. The tuber was found

growing in all but one of the planted fields. The crop is usually grown in association with

banana. Fifty-six of the planted rocas (68%) contained these two crops. Figure 3-3

shows that the majority of harvested fields are left with a single crop. This reflects the

common practice of collecting the manioc, leaving only banana. Twenty out of the 31

harvested fields with useful plants contained banana, by far the most frequently

encountered useful species left in fields after the manioc is collected.

Table 3-4 lists the 19 most common species found in planted fields on high levees

with the corresponding number of harvested fields. It shows that the number of planted

and harvested fields with bananas is relatively consistent at 70% and 65% respectively.

The slightly lower number of harvested fields with bananas is probably a reflection of the

fact that they are more likely to be on lower lying areas than planted fields. A greater

proportion of planted perennials and non flood tolerant species such as a9ai, cocoa, lime

(Citrus aurantifoha), and capsicum pepper are observed in planted fields, lending support

to this idea.

Corn was often left in harvested fields because of ill-developed ears or attacks by pests.

Occasionally, a small patch of manioc remained in the roya that had either not been

harvested in time or was yet to be collected. A number of other useful species that had











either been planted, protected, or had sprouted spontaneously remained in fields. Lime,


sugarcane, cocoa, and possibly a9ai were planted.


30
25
20
LI-
Z15

E 10



1 2 3 4 5 6 7 8 9 1011 1213 14
Number of Species

Figure 3-2. Number of species in 82 planted fields (rocas) on high levees


14
12
10

6
4
z
2

1 2 3 4 5 6 7 8
Number of Species

Figure 3-3. Number of species in 31 harvested fields (rocas) on high levees

The 31 manioc varieties recorded growing in high levee rocas are shown in table


3-5 and 3-6. Nine of these are sweet manioc (macaxeira) while the rest are bitter manioc


mandiocaa). Sweet manioc is typically grown in small patches within rocas of bitter


manioc or sometimes in home gardens. Out of 70 ro9as, 25 (36%) had only one variety


of manioc while the majority had two or more varieties (Figure 3-4). The number of


varieties ranged from one to six with an average of 2.3. By far, the two most common


were the bitter varieties Pacu and Valdevina. At least one of these two were found in


over 75% of all ro9as. Qualidades, as locals refer to varieties, may be called different


names by different communities or individuals. In some cases more than one name may









be given for the same variety. Further study is called for on the diversity of manioc

varieties used by varzea farmers.

Table 3-4. The 19 most common useful species found in planted fields on high levees
with corresponding number of harvested fields
Planted Harvested
fields fields
Rank Local name English name n=82 % n=31 %
1 Mandioca Manioc 81 99 6 19
2 Banana Banana 57 70 20 65
3 Aail Ayai 14 17 3 10
4 Cacau Cocoa 13 16 2 6
5 Mulateiro Mulato wood 12 15 5 16
6 Samauma Kapok tree 11 13 3 10
7 Mamao Papaya 10 12 2 6
8 Cana Sugarcane 9 11 1 3
9 Jenipapo Genipap 6 7 3 10
9 Limio comum Lime 6 7 1 3
9 TaperebA Yellow mombim 6 7 5 16
10 Cedro Cedar 5 6 2 6
10 Jit6 Jit6 5 6 1 3
10 Ucuuba Virola 5 6 1 3
11 Cara New world yam 4 5 0 0
11 Jerimum Squash 4 5 0 0
11 Milho Corn 4 5 5 16
11 Muiratinga Muiratinga 4 5 2 6
11 Pimenta Capsicum pepper 4 5 0 0
* There are two species of a9ai but a9ai preto is most common on the virzea.

Nine varieties of banana were recorded in 74 fields on high levees (Table 3-7). In

this case, just over half (56%) of the fields held only one variety of banana (Figure 3-5).

The greatest banana diversity found in a single field was four types with an average of

1.5. The most common variety, banana prata, was encountered in 84% of all ro9as. The

second most abundant type was banana comprida, a variety of plantain or cooking

banana, found in 44% of ro9as. Certain varieties are more resistant to flooding, pests,

and diseases.











30
25
I 20
15
-o
E 10
z5
z 5
0
1 2 3 4 5 6
Nurnber of vManoc Varieties

Figure 3-4. Number of manioc varieties in fields on high levees


20

10


1 2


Nurmber of Banana Varieties


Figure 3-5. Number of banana varieties in fields on high levees









Table 3-5. Varieties of bitter manioc mandiocaa) growing in 70 fields on high levees
Rank* Variety Num. Fields %
1 Pacu 28 40
2 Valdevina 25 36
3 Olho roxo 8 11
4 Calai 7 10
5 Casca grossa 6 9
5 Ouro 6 9
6 Azulona 5 7
6 Pretona 5 7
6 Sisa 5 7
7 Antinha 4 6
7 Samauma 4 6
7 Tartarugao 4 6
7 Traira 4 6
8 Canjiru 3 4
9 On9a 2 3
10 Amendoem 1 1
10 Auraninha 1 1
10 Azulao 1 1
10 Colaia 1 1
10 Ourinho 1 1
10 Pelonia 1 1
* Rank is combined for varieties bitter and sweet manioc.

Table 3-6. Varieties of sweet manioc (macaxeira) growing in 70 fields on high levees
Rank Variety Num. fields %
4 Pao 7 10
5 Macaxeirao 6 9
7 Amarela 4 6
7 Branca 4 6
7 Pretinha 4 6
9 Preta 2 3
10 Cabral 1 1
10 Colombiana 1 1
10 Pagoa 1 1
* Rank is combined for varieties of bitter and sweet manioc.









Table 3-7. Varieties of banana growing in 70 fields on high levees
Rank Variety No. fields (N=70) %
1 Prata 59 84
2 Comprida 31 44
3 Maga 8 11
4 Nadja 4 6
4 Pacovao 4 6
4 Pacovinha 4 6
5 Guariba 2 3
6 Baiazinha 1 1
6 Urucuri 1 1

Crop Losses from Flooding, Pests, and Diseases

Locals view the yearly flood, especially the occasional high flood, as a major

difficulty of surviving on the floodplain. Even in normal years, losses can occur when

the erosive power of the current causes levees to collapse and fields are partially or

entirely swept away. Farmers usually attempt to harvest their crops just ahead of the

rising waters. The flood can be unpredictable, coming too fast or at the wrong time,

leaving farmers with insufficient means to harvest their crops. They usually cultivate

several fields in different locations and are sometimes faced with the decision of which

fields will be sacrificed to the river and which ones will be saved. When manioc fields

are drowned, the tubers quickly rot and farmers sustain additional losses when their

planting stock for the following season is destroyed. Families must depend on relatives,

friends, neighbors, and even neighboring villages to recover from these losses.

A number of families indicated that they had lost partial or entire crops during the

high flood of 1999. Families were not comprehensively interviewed concerning the loss

of crops but at least ten farmers reported that they had lost all or nearly all of their manioc

rogas to the flood the previous year. The example of one farmer in Vila Alencar may

clarify the magnitude of the losses that can accrue in a high flood event. Anselmo lost a









roya with the equivalent of about 200 kilograms of flour (farinha) because he didn't have

time to collect it. He also lost 500 banana plants and a good many fruit trees that he had

planted. In addition, at least 25 individuals of nine species were drowned in his home

garden. His troubles carried over into the subsequent year because he lost most of his

planting material with the loss of his crops and was only able to plant two small fields.

Although crop losses are potentially much greater during a high flood, farmers may

still lose parts of their fields during normal years when they are unable to harvest them in

time. At least two fields that were visited had small patches of un-harvested manioc

standing in water. One farmer stated that he had lost his crop of beans when the water

took it away.

At least two farmers from two different communities complained that corn they had

planted failed to sprout. In the community of Manacabi, one informant reported that his

family did not plant corn last season. He had planted some seed the year before that the

Mamiraua project had given them and it had not sprouted. He also confirmed that corn

could not be planted on the left bank of the channel because brown capuchin monkeys

(macaco prego, Cebus apella) destroy it. They only plant it on the right bank (actually an

island) where there are no monkeys. Another farmer explained that he normally sells a

lot of squash but he could not plant it this year because of the large amount of sand

deposited by the previous flood.

When a high flood occurs, farmers can suffer profound losses of perennial crops.

Bananas, an important source of income for many families in the region, can be wiped

out as many floodplain dwellers experienced in 1999. Bananas and plantains are a

convenient crop for floodplain farmers because they can harvest them year after year.









When they are drowned, however, residents must replant large areas and it may take

several seasons to build up their banana holdings once again. Young banana plants or

suckers are sometimes destroyed along with the mature plants leaving planting material

in short supply.

Banana losses depend on the length of time an area remains inundated and certain

varieties are more resistant to flooding than others. According to one informant the Prata

variety can resist up to three months, Ma a 2 months, and Comprida only 1 month.

Another resident reported that the variety Peruana does not die in the flood. Suckers,

locally called filhos (offspring), may stay under water more than three months and still

escape the flood but this also probably depends on the variety. Several farmers reported

that all of the banana suckers had been lost in the flood along with the crop.

During the interviews, at least 20 farmers reported that they had lost their banana

plantations in the high flood the previous year. Antonio of Vila Alencar reported that out

of 1000 banana plants, he was left with only 90 banana prata and 150 banana comprida.

Francisco of Jaraua lost his banana crop and no suckers survived for replanting. In Sitio

Forteleza, Antonio lost 110 of his 122 banana plants in a high levee plantation. He is one

of several farmers in the community who has decided to give up agriculture on the varzea

and move to terra firme. Francisco of Vila Alencar lost his entire crop of banana that he

had planted on a low levee this season without harvesting any. He plans to replant the

area again this year.

Perennial fruit trees that are not adapted to flooding are frequently killed in fields,

fallows and home gardens. Recently planted seedlings are especially susceptible to

drowning. Several villagers cited this reason for not planting fruit trees. They risk









having all of their work wasted with the next year's flood. Table 3-8 lists the perennials

that were drowned the previous season and the number of times they were reported. It

does not reflect the number of individuals that were killed. Avocado (Persea

Americana), lime (Citrus aurantifoha), and orange (Citrus sinensis) were the three most

commonly cited fruit trees that had drowned in fields and fallows during the 1999 flood

season.

An interesting comment was made by several informants concerning the

survivability of fruit trees in flooded conditions. These residents have observed that

many fruit trees can survive when flooded as long as they are not bumped by a passing

canoe or floating log. Apparently, the shock caused by the hit is enough to kill a tree that

would otherwise have survived a couple feet of water over its roots. Similarly, fruit trees

in calm water are much more likely to survive than those exposed to a strong current

especially since the current can carry floating objects that may bump the tree.

The economies of many communities have suffered from attacks on their crops by

pests and diseases. According to local informants, communities along the lower Japura

River relied heavily on bananas and made a good income until about five years ago when

mal (bad) struck. Mal is the generic term given to unknown pests or diseases that strike

their crops. Reports indicated that many families had formerly derived the major part of

their income from bananas before they were hit by disease in 1994. It is common for a

family's main source of income to shift depending on the season and also year by year.

As Swales (1999) observed for the lower Amazon, there are no aid programs for

farmers when crop losses occur and no extension workers to identify the diseases and

teach how best to combat them. It is possible that the banana crop is suffering from black









Table 3-8. Perennials that were reported lost in fields and fallows
Local name English Num. Local name English name Num.
name reported reported
losses losses
Abacate Avacado 12 Inga Inga 2
Limao Lime 11 Manga Mango 2
Laranja Orange 10 Puxuri Puxuri 2
Cupuacu Cupuacu 7 Abacaxi Pineapple 1
Abiu Abiu 5 A9ai A9ai 1
Caju Cashew 3 Cacao Cocoa 1
Jambo Malay apple 3 Camu camu Camu camu 1
Lima Lima 3 Carambola Carambola 1
Pupunha Peach palm 3 Seringa Rubber 1
Andiroba Andiroba 2

sigatoka, a disease that had spread south from Central America to Colombia by 1981

(Smith et al. 1992) and may have continued spreading down the JapurA River. Another

disease that affects banana in the region is Moko disease (febre da banana) (Pinedo-

Vasquez et al. 1996). Infected planting material can be brought into virzea plantations

when it is borrowed from the terra firme after a high flood leaves farmers in short supply.

Farmers have discovered that certain varieties of banana have greater resistance to

disease. Whether the bananas are grown in sun or shade also seems to make a difference.

Several farmers reported that bananas grown in shade were less liable to contract a

disease. Lemon is another important crop that has suffered from disease in the last

several years.

The history of one roca on a high levee in the community of Manacabi can

illustrate the impact of both flooding and disease particularly on banana production. The

farmer cleared the roca from 6-year-old secondary regrowth, burned, and planted manioc

and banana. He used the common strategy in which the field is left in banana after the

manioc is harvested. The high flood of 1994 drowned his banana crop. The next year he

planted it again with manioc and the Mayi variety of banana. Mal arrived and again his









banana crop was killed. According to him, the disease dries the leaves, the bananas do

not develop, and after about 3 months it dies. He replanted with a new variety, Prata,

with a greater resistance to the disease. Last year's flood destroyed that crop. Now he

says that the disease is beginning to affect the Prata variety.

Another example of the devastating effects of crop losses comes from the low-lying

terra firme community of Boa Esperanya located on the shore of Lake Amana. Some

residents reported earning a relatively prosperous living in previous years by selling

banana, lemon, and avocado. They had money in the bank and the harvest and sale of

fruit was comparatively easier work than growing manioc and processing manioc flour.

Then, in the mid-nineties, a new disease struck the bananas and greatly curbed

production. Shortly afterward, lemon trees were hit with a disease outbreak and most of

the trees were lost. They were left with their avocado trees that still brought in a good

income. A further blow was struck in 1999 when the community lost hundreds of trees

from its fields and orchards along with a substantial part of the manioc crop during the

unusually high flood. The high water killed most of the avocado trees and the community

had no choice but to intensify manioc production in the year 2000, reportedly initiating

an increase in forest clearing.

Storage and Loss of Planting Material

The semi-perennial crop manioc is reproduced vegetatively. On the varzea of the

middle Solim6es manioc is planted by cutting the stems into approximately 30 cm

sections, making a shallow hole, and horizontally placing the sections in the ground. At

harvest, leaves are stripped from some of the stems which are then saved for the

following season. Depending on their method of storage, a high flood can cause crop

loss not only for that year but also the following year. Planting material for the next









season is typically stored in the field by setting bundles of stems upright on the ground

(Figure 3-6). Any part that stays above water will be viable planting material but if

submerged it is ruined. A safer strategy is to place the bundles of stems on a raised

platform (Figure 3-7). In most years this method is successful, but the occasional high

flood can submerge the platform and destroy the planting stock. The surest way to

eliminate these losses is to store manioc stems on a floating raft but few people take the

trouble to do this as it would entail either building a raft near the field or transporting the

bulky bundles of stems to the village. Material for building the rafts, the lightweight

assacu (Hura crepitans) trunks used for floats, are becoming scarce as timber extraction

continues in the region (Albernaz and Ayres 1999).

When planting material is lost, people must turn to their relatives, friends and

neighbors for help. Fortunately, most residents are willing to share and divide up their

stock among those in need. Francisca explained that during the previous season there

was a large deficit of planting stock in the community of Vila Alencar. Only a little of

hers was lost and she was able to donate manioc stems to eight or nine families including

some in a nearby community. It was still not enough for everyone and another farmer

from the same village claimed that he had planted only part of one field and left another

large field fallow because he lacked manioc and banana stock. A resident of Manacabi

reported that he had planted manioc in less than half of his fields because he divided his

planting material with others who had none.

If farmers have connections with villages on the terra firme, they may take a trip

there to borrow some stock. One farmer from Barroso acquired stems from his church

brothers (irmaos da igreja) on the upland banks of the Japura River. He said that he was









lucky not to have a problem with tuber rot (podre de batata) as often happens when terra

firme varieties are planted on the floodplain. A farmer from Vila Alencar with a shortage

of planting material said he could have obtained some in Alvaraes, a nearby terra firme

town, but he did not have available transport. A problem that was repeatedly voiced by

varzea residents was the lack of adequate transport for situations like this and for getting

their produce to market.

In the year 2000, just after the high flood of 1999, many farmers claimed that they

were unable to plant all of the land that they would have normally cultivated because not

enough planting stock was available. Often fields were planted to only a third or a half of

their size. The survey documented 16 royas that had not been planted completely

because the farmer lacked planting material. The average area not planted was

approximately 1500 square meters per field.

Seeds of annual crops are dried and stored in bottles or cans for the following

planting season. Corn is hung out to dry in the sun and may be stored hanging under the

eves of the house where it can last for a year.

Bananas are planted from the suckers or young shoots that sprout next to mature

plants. The suckers must be taken from an existing field and brought to a new area for

planting. Once the suckers are taken out of the ground they must be kept shaded and

moist until planting. As was mentioned in the previous section, the suckers may also be

killed when a field is inundated for an extended period. They are usually able to survive

for periods of up to three months while immersed in floodwater.































Figure 3-6 Bundles of mamnoc stems placed upnght in resting alta field The water is
approximately 0 5 meters deep Community of Jaraua


Figure 3-7 Storage of mamnoc stems on a raised platform in a high levee field Bundles
of planting material are sometimes seen stored on the side of a floating house
or on a nearby raft Community of Jaraua















CHAPTER 4
CAPOEIRAS: ENRICHED FORESTS AND MANAGED FALLOWS

The popular perception that swidden agriculture entails two or three years of

cultivation followed by an abrupt transition to a long fallow period is not supported by

actual research. In many traditional swidden systems there is a gradual transition from

field to managed fallow or orchard in which production continues for years beyond

"abandonment" (Denevan and Padoch 1987; Eden and Andrade 1987; Hammond et al.

1995). There is often a steady progression of crops that in some ways mirror the natural

regeneration of the forest through the development of vegetation strata. The rural

population of the middle Solim6es region manage the development of fallow fields to a

great extent. Farmers also manage forests to enrich them with useful species including

fruits and construction materials. This chapter documents the rich agrobiodiversity

across a range of habitats and managed fallows in the middle Solim6es varzea.

In the Brazilian Amazon, the term capoeira means fallow or secondary forest.

Farmers in the middle Solim6es refer to some areas as capoeiras even when they contain

useful species and receive some form of management. When the concentration of useful

plants and/or degree of management is greater, they use other descriptive terms such as

bananal (banana grove) or frutal (fruit orchard) to refer to these work areas located

principally on high levees where flooding is occasional and less intense.

Documenting the size and management history of fallows is often more difficult

than gathering the same information for royas. The planted area of multiple ex-fields is

sometimes lumped together by owners and informants as one capoeira that can have parts









of different ages. In some cases a capoeira can be best thought of as a work area where

different activities are taking place through time. A piece of the fallow may be

occasionally used for a roca or bananal, or managed for other products. Small manioc

patches on the order of 10x10 meters or less within a fallow are also sometimes planted.

Partially as a result of this lumping together, the average size of capoeiras in table 4-4 is

almost 6000 square meters compared to rocas with an average of only 2600 square

meters. Management practices of forest enrichment with bananas and fruits also use

more extensive areas than a typical roca (some of these areas were not previously manioc

fields according to the knowledge of the informant). Information on specific field

divisions and management is forgotten over time and fallows can be left when residents

move away or be created by farmers from other communities. With strong kinship ties in

neighboring villages, farmers sometimes plant their fields in the territory of nearby

communities (the territories of communities are generally not rigidly divided). Close

relatives often plant their rocas side by side to take advantage of the collective labor in

clearing the forest and planting.

In the three Mamiraua communities studied by Inuma, Padoch, and Pinedo-

Vasquez, fallow management or agroforestry is a specialized activity used only by a few

households (Padoch et al. 1996, Pinedo-Vasquez et al. 1996). They reported that the

number of managed fallows was about 40% of the total. Managed fallows were divided

into two categories: 75% used the cria9ao-da-mata (forest creation) agroforestry system

which are developed from planted and protected volunteer species in a cleared field and

25% used the enriquecimento-da-mata (forest enrichment) system in which trees of

useful species were protected during field clearance. The researchers stress the point that









there is a tremendous variation in fallow management among individual farmers. They

detail 12 techniques that are used in managing the stand, species, and the individual plant.

These include thinning, girdling, selective cutting, removal of vines, leaving forest cover,

removing forest, planting, seedling selection, pruning, weeding, insect nest removal, and

bleeding.

Inuma, Padoch, and Pinedo-Vasquez (1996) also describe three classes of

techniques for managing forest stands with a high density of a specific valuable species.

Mata demarcada (demarcated forest) involves keeping a buffer of low shrubby vegetation

around the managed forest, mudan9a de mata (changing of the forest) entails the creation

of forest gaps to protect and encourage the growth of valuable trees and shrubs, and mata

mantida (maintained forest) refers to the maintenance of dense stands of fruits such as

bacuri-pari (Rheedia sp.) or camu-camu (Myrclaria dubia). The managed forest stands

are named according to the dominant economic species and include bacabais

(Oenocarpus bacaba), bacuriais (Rheedia sp.), camu-camuais (Myrclaria dubla), pau

mulatais (Calycophyllum spruceanum), and urucuriais (Attalea minor). The three

communities tend to specialize in one or two types of forest stands that they have within

their territory.

Restinga Baixa

Capoeiras on low levees restinga baixa) are generally not managed and contain

few useful species because they are flooded frequently and for a relatively long duration.

None of the more intensely managed land use classes were encountered on low levees.

Table 4-1 summarizes the data for size of fallows, number of useful species, and the

number of consecutive years that the area was cultivated before the current fallow period.

With a mean average of about 2900 square meters, fallows on low levees tend to be









Table 4-1. Summary of statistics for fallows on low levees
N Mean Min. Max. S Median
Size 49 2894 200 17,500 3009 2000
Num. of useful species 17 2.8 1 7 1.7 3
Age of fallow (years) 50 2.8 1 6 1.47 3
Num. of consecutive years
planted previously 12 3.75 1 8 2.2 3

smaller than those on high levees. This is probably because the more land extensive

fallow management techniques are little used on low levees. The size reflects more

accurately the smaller size of royas on low levees. Compared to the 2600 square meter

average high levee ro9as, a low levee roca is about 1600 square meters or just over half

the average capoeira size. This discrepancy between the size of roca and capoeira is most

likely a result of sometimes lumping individual roca "parts" into one capoeira. The

average age of fallows ranged from one to six with an average of just under three. The

number of consecutive crops planted previously varied from one to eight years with an

average of almost four. Only 34% of the fallows contained useful species. Seventeen out

of 50 low levee capoeiras held from one to seven species with an average of three.

Table 4-2 shows the 25 useful species with nine uses encountered in low levee

fallows. There is a large spread of useful species with low frequency counts. By making

more observations the list of useful species would certainly continue to grow. The two

most common, a9ai (Euterpe oleracea), used for making a beverage, and the Kapok tree

(Ceiba pentandra), a valuable timber, were each encountered in five out of 17 fallows. It

is interesting to note that these two highly esteemed trees, occur in low levee fallows in

only two communities each and tend to be found near one another. The different

environments where each community plant their fields partially explains why some

species tend to be found in higher concentrations in some communities and not others.









Table 4-2. The 25 useful species occurring in 17 capoeiras on low levees
Local name English name Use Num. Capoeiras (17) %
Agai Agai2 2 5 29
Samauma Kapok tree 3 5 29
TaperebA Yellow mombim 1 4 24
Jenipapo Genipap 1 3 18
Lauro inamui Lauro inamui 3 3 18
Mulateiro (Pau
mulato) Mulato wood 3 3
Cacau Cocoa 1,2 2 12
Castanha de macaco Castanha de macaco 7 2 12
Inga Inga 1 2 12
Muiratinga Muiratinga 3 2 12
Purui Purui 1 2 12
Taxi Taxi 3 2 12
Andiroba Andiroba 3 1 6
Bacaba Bacaba 2 1 6
Bacuri coroa Bacuri coroa 1 1 6
Bacuri liso Bacuri liso 1 1 6
Banana Banana 1 1 6
Cedro Cedar 3 1 6
Goiaba Guava 1 1 6
Jacareuba Jacareuba 3 1 6
Jauari Jauari 5 1 6
Limao comum Lime 1 1 6
Macacaricuia Macacaricuia 3 1 6
Paricarana Paricarana 3 1 6
Ucuuba Virola 3 1 6
Use categories: 1 Fruit, 2 Beverage, 3 Timber, 5 Fish bait, 7 Nut, 9 Medicinal, 10 Latex,
11 Vegetable, 12 Livestock feed
* There are two possible species of agai; agai do Para (E. oleracea), and agai preto (E.
precatoria).

Numbers three and four on the list, yellow mombim (Spondias mombim) and genipap

(Genipa Americana), are two large fruit trees that seem to thrive in swidden fallows on

the floodplain. They are rarely planted but often protected. In some cases however, they

will be sacrificed in the interest of clearing a manioc roya. The next two on the list, lauro

inamui (Ocotea cymbarum) and mulateiro (Calycophyllum spruceanum), found in three

fallows each, are both valuable timber trees that sprout easily in manioc royas and are

often protected and managed.









Restinga Alta

In contrast to low levee fallows that are flooded yearly for several months, fallows

on high levees restinga alta) are often managed and contain a much greater diversity of

beneficial plants. Middle Solim6es farmers use a variety of terms to designate work

areas on high levees. The terms range from the fallow with no useful plants to the sitio

(site), an area that is comparable to a home garden in the great diversity of useful species.

A list of the work area designations and their descriptions is shown in table 4-3. The

number of work areas visited gives an idea of the relative proportions in which they

occur. The majority (70%) are simple fallows that have a lower average number of

useful species than most of the other more intensively managed fallow types that made up

60 out of 202 (30%) non-roca work areas on high levees. Table 4-4 compares the size of

non-roca work areas.

The complexity of management practices and diverse land use histories and the

ambiguity of designations of land use types by different farmers makes it difficult to

precisely divide the work areas into discrete categories, resulting in some overlap. Of all

the areas not designated as ro9as, 75% contain useful species. The intensity of

management and number of species vary widely across these work areas (Table 4-5).

Some fallows contain only useful plants that were either spared when the roya was

cleared or sprouted spontaneously while others are planted with a few dominant crops

such as banana and ayai. Still others contain a diverse array of carefully tended fruits,

vegetables, timber, or medicinal plants. Management techniques such as weeding and

pruning vary in intensity from frequent to none.

The non-roca work areas can be split into two general groups; those that are

considered capoeiras and those that are not. A capoeira may be specified further by local









Table 4-3. Local designations of
description


work areas, number of areas visited, and their


English Num. of areas
Local Term translation visited Description
Banana plantation or forest enriched
Bananal Banana Grove 17 with banana
Cacaual Cocoa Grove 0 Area planted with cocoa
Capoeira Fallow 142 Fallow field
Capoeira Fallow enriched with banana or
bananal Banana Fallow 22 former bananal
Capoeira
cacaual Cocoa Fallow 2 Former cacaual
Capoeira
fruteira Fruit Fallow 12 Fallow enriched with fruit trees
Area enriched with a diversity of
Frutal Fruit Orchard 2 fruit trees
Area of intensive, long-term
management and/or a high
Sitio Site 5 diversity of useful plants
* No cacaual areas were encountered during the survey but a few capoeiras were said to
be former cocoa groves.

Table 4-4. Average size by type of capoeira
Mean size
Type N (m2) Min. (m2) Max. (m2) S (m2)
All
Areas1 188 5978 600 50000 7899
Bananal 16 4222 400 10000 3420
Capoeira2 133 5901 500 50000 7924
Capoeira
bananal 21 5362 400 40000 8362
Capoeira
cacaual 2 2525 1050 4000
Capoeira
fruteira3 9 10856 1400 48000 16630
Frutal 2 36000 12000 60000
Sitio 5 4760 2000 8000 2385
1 The five largest and smallest were dropped in calculating the average.
2 The two largest and smallest were dropped in calculating the average.
3 The largest and smallest were dropped in calculating the average.









residents as a capoeira bananal (banana fallow), capoeira cacaual (cocoa fallow), or

capoeira fruteiras (fruit fallow) based on the relative amount of work that has gone into

enriching them with valuable plants and the density of planted species. Three types of

land use, bananal (banana grove), frutal (fruit orchard) and sitio (site) are not considered

to be capoeiras by farmers. The owner has generally invested much more time and

energy into these areas and the density of planted species is relatively higher.

Table 4-5. Average number of useful species by type of capoeira
Mean num.
Type N species Min. Max. S
Bananal 17 2.4 1 7 2.2
Capoeira 88 3.5 1 18 3.1
Capoeira bananal 21 3.6 1 12 2.9
Capoeira cacaual 2 1.0 1 1
Capoeira fruteira 12 10.2 5 19 4.3
Frutal 2 16.0 13 19
Sitio 5 20.8 11 34
* Only capoeiras that hold useful species are included.

Bananal

Banana is found growing in all types of work areas on high levees including home

gardens. It is also planted on low levees where it is harvested after six months just before

the flood (or just after) and completely replanted the next season. Although banana may

still come in after manioc in importance as a regional staple, it was overall the most

frequently observed species in the survey with 325 total observations compared to

manioc which came in second with 210 observations. Its ubiquitousness across many

types of habitats and management areas underscores its versatility and adaptability to the

floodplain environment and its status as an important commercial and subsistence crop.

Bananas are produced in a monocultural grove or intercropped with manioc. They are

also found growing scattered here and there among ro9as, capoeiras, or home gardens.









The bananal is formed using one of three techniques. A common strategy is to

interplant banana with manioc in the roga on high levees. When manioc cultivation is

discontinued, the area is left as a bananal often for several years until a high flood

destroys the bananas at which time it may be cleared to plant a roga once again or left as

a capoeira. The second technique is to clear an area specifically for planting bananas.

The area may be cleared from either forest (mata) or secondary forest of any age. It may

or may not be burned. The third way is to clear only the undergrowth and select trees in a

mature forest or secondary forest leaving large trees standing to provide shade for the

crop. The area is often cleared strategically with undesirable trees removed and valuable

ones left in place. Several informants told us that, by growing bananas in the shade,

some disease can be avoided. In all three of these systems, various fruit trees may also be

added creating a bananal with several economically valuable species in the mix.

Out of 10 banana groves (bananais plural) where the information was collected,

half had been planted to roca at least one year at the beginning of the cropping sequence.

In the other five, only banana was planted. At least three of these five areas were not

completely cleared of trees. Instead, the underbrush was cut and the bananas were

planted under shade.

A particularly interesting bananal was visited on a very wide high levee near the

community of Vila Alencar. This area was referred to as mata bruta (mature forest or

regrowth) and was apparently very old secondary forest. Only a few trees were cut and

some trees were ringed but there was still sufficient canopy cover to shade out grasses.

Weeds of only a few families (Musaceae, Piperaceae, and Solonaceae), soft-stemmed and

easily weeded, are scattered throughout the field. According to the informant, after the









bananas are grown and weeded, the weeds do not come back and some disease is

avoided. In another field it was explained that roya was planted only the first year. After

harvest, the secondary forest was allowed to grow back and provide shade for bananas

and other fruit trees planted under the canopy. Besides banana, two other species (acai

and cocoa (Theobroma cacau) were planted and several timber species had been spared

including muiratinga (Maquira spruclana), virola (Virola surnamensis), and the kapok

tree. The area is occasionally weeded and other seedlings may be added. When bananas

are killed by drowning like in the 1999 flood season, surviving suckers are gathered, the

undergrowth is slashed, and the bananal is refurbished.

We visited 17 areas that were designated as bananal by informants. The average

number of species and average size was slightly lower than the other capoeiras but the

sample size is insufficient to conclude that there is a significant difference between the

two types of land uses (Tables 4-4 and 4-5). The average age of 11 bananais was 3.4

years with a maximum of 7 years (Table 4-7).

High floods necessitate the periodic refurbishment of a bananal which may occur

multiple times during the period of cultivation. Farmers sometimes change the location

of the bananal when disease arrives in the hopes of achieving a disease free crop for a

couple of seasons. Leaving a bananal fallow for a time can also help in controlling

disease.

Bananais contained the lowest average number of species of any capoeira type

except the capoeira cacaual. The planted species include acai, banana, cocoa, sugarcane,

cupuaqu (Theobroma grandiflorum), breadfruit (Artocarpus altihs), and manioc (Table 4-

6). Other species were either protected during forest clearance or sprouted (or









resprouted) spontaneously. One farmer informed us that he spared castanha de macaco

(Couroupita guzanensis), cedar (Cedrellafissils), and muiratinga (Maquira spruczana) in

his bananal at the time of clearing. A number of naturally occurring trees are

economically valuable timber species that are harvested in the region. These include

cedar, muiratinga, mulateiro, kapok, taxi, and virola (Alberaz and Ayres 1999). IngA

and yellow mombim (Spondius mombim) are two fruit trees that also occur spontaneously

on the floodplain. Itua (Gnetum nodiflorum) is a vine which bears fruit that can be

roasted and eaten and is also utilized as fish bait.

Varietal diversity was not found to be very great in bananais (Table 4-8). The

average number of banana varieties in 10 fields was 1.3 with seven containing only a

single variety. Only three varieties of banana were registered in these areas. The Prata

variety has replaced Mayi as the most popular variety in the region due to its elevated

resistance to disease. It was found in eight out often bananais. The Comprida variety

was second, occurring in four bananais while Mayi was found in only one.

Bananais were greatly affected by the previous flood season possibly lowering the

average number of varieties found in the fields but this may not be the primary reason

why there exists a relative dominance of a few select varieties. In his study of

agricultural intensification by a Tikuna Amerindian community in the Upper Solim6es,

Shorr (1999) revealed the increasing dominance of a few varieties of manioc and banana

in recent decades. According to his research, it was increasing market participation that

caused the shift in their crop repertoire. The Comprida variety was more frequently

planted due to its durability in transport, greater marketability, and higher price. It is









likely that a similar process has been occurring in the Middle Solim6es as farmers have

become increasingly integrated into local markets.

Table 4-6. The 22 useful species occurring in 17 banana groves (bananais)
Local name English name Use Num. Bananais %
Banana Banana 1 17 100
Agai* Agai2 2 3 18
Cacau Cocoa 1,2 2 12
Cedro Cedar 3 2 12
Muiratinga Muiratinga 3 2 12
Cana Sugarcane 2 1 6
Carapanauba Carapanauba 3, 9 1 6
Castanha de macaco Castanha de macaco 7 1 6
Cupuaqu Cupuaqu 1 1 6
Fruta pau Breadfruit 1 1 6
Inga Inga 1 1 6
Itui Itui 1,5 1 6
Mandioca Manioc 8 1 6
Mulateiro Mulato wood 3 1 6
Murumuru Murumuru palm 5 1 6
Namuirana Namuirana 3 1 6
Pau ferro Pau ferro 3 1 6
Samauma Kapok tree 3 1 6
Tanibuca Tanibuca 3 1 6
TaperebA Yellow mombim 1 1 6
Taxi Taxi 3 1 6
Ucuiba Virola 3 1 6
Use categories: 1 Fruit, 2 Beverage, 3 Timber, 4 Bowl, 5 Fish bait, 6 Thatch, 7 Nut, 8
Tuber, 9 Medicine
* There are two possible species of acai; acai do Para (E. oleracea), and acai preto (E.
precatona).

Table 4-7. Summary of statistics for banana groves (bananais)
N Mean Min. Max. S Median
Age ofbananal 11 3.4 1 7 2.0 3
Num. of species 17 2.41 1 7 2.24 1
Num. ofbanana
Num. of banana 10 1.3 1 2 0.48 1
varieties

Table 4-8. Banana varieties occurring in 10 banana groves (bananais)
Variety No. Bananais %
Prata 8 80
Comprida 4 40
Mayg 1 10









Capoeira

Farmers distinguish between two classes of secondary regrowth, old secondary

regrowth (capoeira velha) and young secondary regrowth (capoeira nova) but age

stipulations can vary by community or individual. For example, one informant may

designate a 1 to 3-year-old fallow as "young" and anything older than 3 "old" while an

informant from another community may consider fallows up to 5 or even 8 years "old."

There is never a sharp distinction between the two designations.

In the areas designated as capoeira, 88 of 142 (62%) were found to contain an

average of 3.5 useful species. Maintenance of swidden fallows consists of occasional

weeding around bananas and fruit trees by slashing encroaching vegetation with a

machete.

The capoeiras can be divided into two distinct types; those that were formerly rocas

and those that were formerly bananais. Information on previous land use was collected

for 111 fallows. Most of the areas designated as former bananais were not planted with

manioc at the start, agreeing with the information on management history of bananais

presented above. A few of the areas were planted with roya the first year and then

planted to banana (and sometimes other fruits) in the following years.

It is useful to compare these two types of fallows to see which form of management

may lead to a greater diversity of useful plants and the nature of the beneficial species

that are retained. Table 4-9 shows data on the number of useful species found, age, and

number of consecutive years formerly under cultivation for the two types of capoeiras.

Fifty-two of 79 (82%) former rocas contained an average of 3.5 species. Less of the

former bananais were found to contain useful species, only 20 out of 32 (60%). The

average number of species in the former bananais were slightly higher with 4.4 species.










Table 4-9. Summary of statistics for fallows (capoeiras) on high levees
N Mean Min. Max. S Median
Former Useful Species 52 3.52 1 18 3.33 3
roya Banana 20 1.3 1 3 0.57 1
Varieties
Age of Fallow 77 3.4 1 13 2.5 3
(years)
Years Planted 53 3.4 1 15 3.2 2
Before
Size1 (m2) 72 5196 800 18000 4770 4000
Former Useful Species 20 4.35 1 13 3.25 4
bananal Banana 9 1.1 1 2 0.33 1
Varieties
Age of Fallow 31 3.4 1 15 3.3 2
(years)
Years Planted 22 3.6 1 12 2.7 2
Before
Size2 (m2) 30 6891 1000 28000 7615 3950
1 Two largest and smallest values were dropped in calculating the mean.
2 Largest and smallest values were dropped in calculating the mean.

Table 4-10. Banana varieties occurring in fallows (capoeiras) on high levees
Num. Num.
Num. former former
capoeiras rogas bananais
Variety (41) % (23) % (13) %
Prata 31 76 17 74 10 77%
Maya 8 20 4 17 4 31%
Comprida 5 12 3 13 2 15%
Nadja 3 7 1 4 2 15%
Pacovinha 1 2 1 4 0 0%













Table 4-11. Twenty-five most frequently encoun
rocas and former bananais
Local name English name Use1
Banana Banana 1
Acai Acai2 2
Cacau Cocoa 1
Mulateiro Mulato wood 3
TaperebA Yellow mombim 1
Lauro inamui Lauro inamui 3
Ucufba Virola 3
Tacaca Tacaca 3
Assacu Assacu 3
Bacaba Bacaba 2
Limao comum Lime 1
Manga Mango 1
Muiratinga Muiratinga 3
Cedro Cedar 3
Cuia Calabash goard 4
Cupuaqu Cupuaqu 1
Jit6 Jit6 3
Cana de aqucar Sugarcane 2
Fruta pao Breadfruit 1
IngA Inga 1
Jambo Malay apple 1
Jauari Jauari 5
Jenipapo Genipap 1
Mamao Papaya 1
Urucuri Urucuri palm 6
Use categories: 1 Fruit, 2 Beverage, 3 Timber, 4
* There are two possible species of agai; agai do


Itered species and their use in capoeiras on high levees with a comparison of former


Capoeiras (88) % Former rocas (52) %
41 47 23 44
26 30 15 29
24 27 13 25
14 16 10 19
14 16 7 13
12 14 5 10
11 13 7 13
10 11 5 10
8 9 6 12
8 9 4 8
8 9 6 12
8 9 4 8
7 8 4 8
6 7 4 8
6 7 3 6
5 6 3 6
5 6 3 6
4 5 1 2
4 5 3 6
4 5 3 6
4 5 3 6
4 5 3 6
4 5 2 4
4 5 2 4
4 5 1 2
Bowl, 5 Fishbait, 6 Thatch
Para (E. oleracea), and a9ai preto (E. precatona).


Former bananal (20)
13
9
6
3
5
6
4
4
1
4
1
3
1
2
0
2
2
0
0
1
1
1
0
0
2









The results for both the average age of the fallow and the average number of years

formerly under cultivation were three and a half years; virtually the same for both land

use types.

A total of 68 useful species were recorded in 88 capoeiras on high levees. Table 4-

11 shows the 25 most frequently encountered and gives an indication of their relative

dominance in former rogas and former bananais. Although the sample of former

bananais is small, it can give us an indication of possible differences between the two

land use types. Banana is the most common valuable plant growing in fallow areas. The

percentage of former bananais containing banana was higher with 65% compared to

former rogas with 44%. A9ai also occurred 16% more frequently in former bananais.

The fact that yellow mombim occurs 12% more frequently in former bananais would

makes sense since it is a tree that would more likely be sacrificed when clearing land for

a roya because of the need for full sunlight in manioc production. It is more convenient

to spare fruit trees like yellow mombim and genipap or valuable timber in a banana

where large trees are often preserved to provide some shade and the banana plants are

often scattered over a wide area. Of the 25 most frequently encountered useful plants in

capoeiras, the majority (11) are fruit trees, eight are timber trees, three are used for

beverages, and one each are used for fish bait, thatch, or bowls.

Five varieties of banana were found growing in capoeiras on high levees (Table 4-

10). In 29 capoeiras, the average number of varieties was 1.2. Only six capoeiras (21%)

had more than one variety; five had two varieties, and one had three. Five of these are

former rocas and one is a former bananal. Again, prata is the variety that dominates in

these areas. This variety was found in 76% of the 41 capoeiras where information on









variety was collected. It occurred in almost the same percentage of former rocas and

former bananais. Maya is the second most common variety, found in 20% of 41

capoeiras. Comprida is third with 12 percent. Two relatively uncommon varieties, Nadja

and Pacovinha, were also found.

Capoeira Bananal

Local residents sometimes distinguish the capoeira bananal (banana fallow) from a

regular capoeira. It is relatively enriched with bananas and often with other fruits. It can

result from several possible origins. After a high flood kills much of a crop in a more

densely planted bananal the status of the work area is "downgraded" to capoeira bananal.

The area is often replanted to renew its status as a bananal. The capoeira bananal can

also result from a harvested roya that was interplanted with bananas. It may be only

partially planted with bananas. In that case it is more like a small bananal within a larger

capoeira. It can also be capoeira of any age that is enriched with scattered banana plants.

Out of 13 capoeira bananais (plural) surveyed, six were formerly rocas, four were mata

(forest), two were bananais, and one was capoeira.

The ages of capoeira bananais range from 2 to 18 years with an average of almost

four (Table 4-12). The number of years capoeira bananais was previously planted with

crops, reported in only three cases, was two.

The capoeira bananal is a range along a continuum of plant diversity in the

agricultural ecosystem of the middle Solim6es varzea. They contained from 1 to 12

useful species with an average of over three. The most frequently encountered species

are those that commonly sprout spontaneously in resting alta fields and fallows (Table

4-13). These include ayai, genipap, papaya, mulatto wood, and yellow mombim.










Table 4-12. Summary of statistics for banana fallows (capoeira bananal)
N Mean Min. Max. S Median
Age of capoeira
bananal 16 3.88 2 18 4.00 3
Num. of species 21 3.57 1 12 2.87 2.5
Num. of banana
varieties 17 2 1 4 1.06 2

Table 4-13. The 30 useful species found in 21 banana fallows (capoeira bananal)
Num. capoeira


Local name English name Use
Banana Banana 1
Mulateiro Mulato wood 3
Aqai* Ayai2 2
Mamao Papaya 1
TaperebA Yellow mombim 1
Cuia Calabash gourd 4
Jenipapo Genipap 1
Goiaba Guava 1
Manga Mango 1
Muiratinga Muiratinga 3
Samauma Kapok tree 3
Alfavaca Alfavaca 9
Assacu Assacu 3
Bacaba Bacaba 2
Bacuri coroa Bacuri coroa 1
Cacau Cocoa 1
Caju Cashew 1
Capim santo Lemon grass 9
Cubiu Cubiu 1
Goiaba araga Goiaba araga 1
Graviola Soursop 1
Inga Inga 1
Jacareuba Jacareuba 3
Jambo Malay apple 3
Jauari Jauari 5
Jit6 Jit6 3
Lauro abacate Lauro abacate 3
Limao comum Lime 1
Majiri9ao Majiri9~o 9
Mandioca Manioc 8
Use categories: 1 Fruit, 2 Beverage, 3 Timber, 4 Bowl, 5
Tuber, 9 Medicine
* There are two possible species of a9ai; a9ai do Para (E.
precatona).


bananal (21) %
20 91
6 27
5 23
5 23
5 23
3 14
3 14
2 9
2 9
2 9
2 9
1 5
1 5
1 5
1 5
1 5
1 5
1 5
1 5
1 5
1 5
1 5
1 5
1 5
1 5
1 5
1 5
1 5
1 5
1 5
Fishbait, 6 Thatch, 7 Nut, 8

oleracea), and a9ai preto (E.









Table 4-14. Banana varieties in banana fallows (capoeira bananal)
Variety Num. banana fallows (17) %
Prata 16 94
Comprida 9 53
Maga 5 29
Guariba 2 12
Peruana 2 12


























Figure 4-1. Banana fallow on a high levee on the Solim6es River. Banana, agai and
several other fruits are growing in partial shade in an area about ten meters
wide extending 200 meters.

Cuia, a planted and carefully tended shrub whose giant gourd fruits are used as bowls,

cups, and ladles was found in three of 21 capoeira bananais.

Of course, a capoeira bananal should have bananas. It is the presence and density

of bananas that determine the areas classification. They occur with up to four varieties in

all but one of the 21 surveyed with an average of two. Of five varieties, banana prata is

the most prolific, occurring in 94% of the fallows surveyed, Comprida (plantain) was